handbook of bolts and bolted joints

Simpo PDF Merge and Split Unregistered Version - http://www.simpopdf.com

title: Handbook of Bolts and Bolted Jointsauthor: Bickford, John H.

publisher: CRC Pressisbn10 | asin: 0824799771print isbn13: 9780824799779

ebook isbn13: 9780585158334language: English


subject Bolted joints, Bolts and nuts.publication date: 1998

lcc: TA492.B63H36 1998ebddc: 621.8/82

subject: Bolted joints, Bolts and nuts.


Page i

Handbook of Bolts and Bolted Joints

Edited byJohn H. Bickford

Middletown, Connecticut

Sayed Nassar

Lawrence Technological UniversitySouthfield, Michigan


Page ii

Library of Congress Cataloging-in-Publication Data

Handbook of bolts and bolted joints / edited by John H. Bickford,Sayed Nassar.p. cm.Includes index.ISBN 0-8247-9977-1 (alk. paper)1. Bolted joints. 2. Bolts and nuts. I. Bickford, John H.,II. Nassar, Sayed.TA492.B63H36 1998621.8'82--dc21 98-14949 CIP

The publisher offers discounts on this book when ordered in bulk quantities.For more information, write to Special Sales/Professional Marketing at theaddress below.

This book is printed on acid-free paper.

Copyright © 1998 by MARCEL DEKKER, INC. All Rights Reserved.

Neither this book nor any part may be reproduced or transmitted in any formor by any means, electronic or mechanical, including photocopying,microfilming, and recording, or by any information storage and retrievalsystem, without permission in writing from the publisher.

MARCEL DEKKER, INC.270 Madison Avenue, New York, New York 10016http://www.dekker.com

Current printing (last digit):10 9 8 7 6 5 4 3 2 1


PRINTED IN THE UNITED STATES OF AMERICA


Page iii

PREFACE

The bolted joint is a surprisingly complicatedaffair. Literally hundreds of factors affect theresults when we assemble a joint, and hundredsmore affect its behavior and life when we putthat joint to use. Because of this, the designs ofmost joints are based on "feel," supported by"past experience" or "custom." Also because ofthis most joints are overdesigned; the jointmembers are stronger and the bolts larger and/ormore numerous than would be the case if wecould control assembly and predict behaviormore accurately.

An earlier book by one of the present editors, AnIntroduction to the Design and Behavior ofBolted Joints (Marcel Dekker, Inc., New York)was an attempt to introduce the engineer to these


complexities by summarizing the various aspectsof design and behavior. That work, which wasoriginally published in a volume of 450 pages, isnow available in a third edition containing overtwice that many pages. But it is still, asadvertised, an "introduction" to the subject.

The time has now come to move beyond theintroductory level in a pair of hardcover texts.The present Handbook of Bolts and BoltedJoints is one of this pair; a sister volume,Gaskets and Gasketed Joints (1997, MarcelDekker, Inc., New York) is the other. Followingmodern practice, each book has been written by alarge number of authors, each contributing oneor more chapters on his specialty. We believethat each of our contributors is a recognizedauthority on his subject, and we have encouragedeach to provide basic, time-tested information aswell as dependable emerging research.

Although the material in this volume is moreauthoritative and contains greater detail than the


earlier Introduction, every effort has been madeto make it useful rather than just "scholarly."Whenever possible, each chapter includesstraightforward how-to-do-it or what-to-doinformation. Background theories are includedonly as necessary to give readers anunderstanding of bolt and joint behavior that willhelp them to select parts or materials, createdesigns, control the assembly process, andminimize or solve problems. Extensivereferences are provided, however, for those whowish to delve deeper into the theoretical workthat supports the information presented here.

Although truly complete coverage would requiremany more volumes, a serious attempt has beenmade to include chapters on all of the key issuesaffecting bolting. This includes all of the subjectscovered in the Introduction volume: bolt andjoint behavior, bolting materials, joint designprocedures, a wide variety of assemblytechniques, and all important failure modes.


Each such topic is covered here in greater detailthan in the Introduction and, as mentioned, iscovered by a specialist in his field.

In addition, the Handbook includes chapters onmany new subjects. For example, where theearlier work focuses on the design and behaviorof bolted joints, this volume includes manychapters devoted to the fastenerthe boltalone.There are chapters on metric fasteners,


Page iv

aerospace bolts, production techniques forfasteners, control of their quality, head markings,washers, and anchor bolts, for example. Othernew material deals with such diverse topics asBelleville springs, joining with aluminum alloybolts and nuts, thread lubricants, bolting sealantsand adhesives, the design of airframe joints,design rules for structural steel bolting, themulti-jackbolt tensioning system, selection offastening strategies for mass production, ways toinspect previously assembled joints, statisticaldesign and analysis of multivariable experiments,and the education of a fastener engineer.

Incidentally, even though gasketed joints arecovered more thoroughly in Gaskets andGasketed Joints, the design and behavior of suchjoints is important enough to demand inclusionhere, so these topics are also covered, by expertsand in some detail. Presumably only those


involved more or less full time with the designof gasketed joints will need to refer to the otherpublication.

By design the Handbook is not just a collectionof data. You will not find here, for example,reprints of existing government, ANSI, ASTM,ASME, SAE, IFI, or other specifications andstandards. Such documents are referred to inmany places, and small portions of a few of themare reprinted when necessary, but repeating themin their entirety would be redundantyou can findthem in other handbooks as well as in theiroriginal coversand, in our opinion, repeatingthem could be dangerous. Most specificationsand standards are subject to slow but more orless continual change, which, although oftenminor in appearance, has been consideredimportant by the originators. Using an obsoleteversion of a standard could be anything fromembarrassing to dangerous, and it would beimpractical for the publishers of a handbook


such as this to reissue it often enough to catch allsuch changes. Specifications and standards,furthermore, tend to "stand alone" and are readilyavailable. Our focus has been on material that isnot easy to find because it's new, or rare, orinmost caseswidely dispersed.

Because the Handbook has been written by avariety of authors, from different industries andfrom academia, with different types ofexperience and backgrounds to guide them, thereis some redundancy in the material covered, andsome differences of opinion. Bolted joints areaffected by literally hundreds of variables. Theexperiences of one engineer will often differ,sometimes substantially, from those of another,often for reasons not understood at the presentstate of the art. All the experiences, however, arevalid and have their place in the world of bolting.Seeing them combined in this way will warn theunwary that, especially if safety or abnormalexpense is involved, it is best to approach the


bolt and bolted joint with respect and withcaution. We believe that this volume will helpyou do that.

JOHN H. BICKFORD


Page v

CONTENTS

Preface iii

Contributors ix

Part I: Material Properties and Selection

1. Fastener MaterialsA. Craig Hood

1

2. Joining with Aluminum Alloy Bolts and NutsRussell E. Mack

35

3. Thread LubricantsGary J. Novak and Tom Patel

43

4. Adhesives and Sealants for BoltingJohn Cocco

57

Part II: Processing of Fasteners

5. Fastener ManufacturingJesse A. Phebus and Peter F. Kasper

69

6. Fastener CoatingsJohn Laurilliard

75

7. Fastener QualityChristopher B. Wackrow

107

Part III: Threaded Fasteners


8. Screw ThreadsEli Schwartz and John H. Bickford

121

9. Fastener Head MarkingsCharles J. Wilson

149


Page vi

10. Computing the Strength of a FastenerJoseph Barron

163

11. Computing the Stiffness of a FastenerJoseph Barron

179

12. Metric FastenersBruno Marbacher

187

13. Washers: A Guide to Their Function and SelectionMarc Levinson

233

14. Belleville SpringsGeorge Davet

251

15. Vibration-Resistant FastenersJoseph R. Dudley

281

16. Concrete AnchorsRobert L. Zink

293

17. Aerospace BoltsShahriar M. Sadri

309

Part IV: Design and Analysis of Bolted Joints

18. VDI Joint Design ProceduresAlex W. Heston

317

19. Design of Gasketed JointsBernard S. Nau

341

20. ASME Boiler and Pressure Vessel Code Design Rules for


Bolted FlangesGeorge Bibel

357

21. Design of Joints Loaded in ShearRichard T. Barrett

369

22. Design Rules for Structural Steel BoltingJohn H. Bickford and William A. Milek

399

Part V: Assembly of Bolted Joints

23. Behavior of a Bolted Joint during AssemblyJohn H. Bickford

425

24. Tightening Groups of Fasteners in a Structure and theResulting Elastic InteractionGeorge Bibel

451

25. Hydraulically Powered WrenchesGeorge A. Sturdevant

479

26. Hydraulic Stud TensioningWilliam L. Biach

501


Page vii

27. The Multi-Jackbolt Tensioning SystemRolf H. Steinbock

507

28. Selecting Fastening Strategies and Equipment for MassProductionJacob R. Bippus

535

29. Torque Control of AssemblyJohn H. Bickford

559

30. Stretch ControlRod Corbett

571

31. Increasing Joint Performance by Tightening Bolts to YieldPaul W. Wallace

591

32. TorqueAngle Tension ControlRalph S. Shoberg

603

33. Control with Direct Tension IndicatorsDavid Sharp

621

34. Use of Ultrasonics in Bolted JointsSayed Nassar

631

35. Selecting Preload for an Existing JointJohn H. Bickford

659

Part VI: The Joint in Service

36. Joint DiagramsJohn H. Bickford

687


37. Bolt FatigueFrancis R. Kull

699

38. CorrosionKingman E. Yee

709

39. The Susceptibility of Fasteners to Hydrogen Embrittlementand Stress Corrosion CrackingLouis Raymond

723

40. Vibration- and Shock-Induced LooseningDaniel P. Hess

757

Part VII: Testing and Inspection of Bolts and Joints

41. Statistical Design and Analysis of Multivariable ExperimentsBernard Clément

825

42. Field Testing of Structural Bolts and Installation Tools Usinga Bolt Tension CalibratorJohn W. Wilhelm, Jr.

871


Page viii

43. TorqueTension AuditsRalph S. Shoberg

881

Part VIII: Education and Training

44. Fastening Technology EducationBengt Blendulf

895

Index 901


Page ix

CONTRIBUTORS

Richard T. Barrett Barrett EngineeringConsulting, Olmsted Falls, Ohio

Joseph Barron Newport News Shipbuilding,Newport News, Virginia

William L. Biach Biach Industries, Inc.,Cranford, New Jersey

George Bibel Mechanical Engineering,University of North Dakota, Grand Forks, NorthDakota

John H. Bickford Consultant, Middletown,Connecticut

Jacob R. Bippus Hickok Incorporated,Cleveland, Ohio

Bengt Blendulf Clemson EduPro, Inc., Clemson,


Bengt Blendulf Clemson EduPro, Inc., Clemson,South Carolina

Bernard Clément École Polytechnique deMontréal, Montreal, Quebec, Canada

John Cocco Loctite Corporation, Rocky Hill,Connecticut

Rod Corbett Rotabolt, West Midlands, England

George Davet Solon Manufacturing Company,Chardon, Ohio

Joseph R. Dudley Nylok Fastener Corporation,Macomb, Michigan

Daniel P. Hess Department of MechanicalEngineering, University of South Florida,Tampa, Florida

Alex W. Heston SPS Technologies, Cleveland,Ohio

A. Craig Hood ACH Technologies, Wayne,Pennsylvania


Peter F. Kasper Vermont FastenersManufacturing, Swanton, Vermont

Francis R. Kull Consultant, Warminster,Pennsylvania

John Laurilliard Consultant, Philadelphia,Pennsylvania

Marc Levinson ITW Shakeproof, Elgin, Illinois

Russell E. Mack* Aluminum Company ofAmerica, Lititz, Pennsylvania

*Retired.


Page x

Bruno Marbacher Bossard International Inc.,Portsmouth, New Hampshire

William A. Milek Consultant, Glen Ellyn,Illinois

Sayed Nassar Mechanical Engineering, LawrenceTechnological University, Southfield, Michigan

Bernard S. Nau Consultant, BHR Group Ltd.,Cranfield, Bedford, England

Gary J. Novak Fel-Pro Chemical Products,Skokie, Illinois

Tom Patel Fel-Pro Chemical Products, Skokie,Illinois

Jesse A. Phebus A. Phebus and Associates, Inc.,Safety Harbor, Florida

Louis Raymond L.Raymond & Associates,Newport Beach, California


Shahriar M. Sadri Huck International, Inc.,Carson, California

Eli Schwartz Consultant, Philadelphia,Pennsylvania

David Sharp J&M Turner, Inc., Southampton,Pennsylvania

Ralph S. Shoberg RS Technologies, FarmingtonHills, Michigan

Rolf H. Steinbock Superbolt Inc., Carnegie,Pennsylvania

George A. Sturdevant Fastorq Bolting Systems,Inc., Houston, Texas

Christopher B. Wackrow MNP Corporation,Utica, Michigan

Paul W. Wallace Ingersoll-Rand Company,Flemington, New Jersey

John W. Wilhelm, Jr. Skidmore-Wilhelm,


Cleveland, Ohio

Charles J. Wilson Industrial Fasteners Institute,Cleveland, Ohio

Kingman E. Yee Lawrence TechnologicalUniversity, Southfield, Michigan

Robert L. Zink Hilti, Inc., Tulsa, Oklahoma


PART IMATERIAL PROPERTIES AND SELECTION

1Fastener Materials

A. Craig HoodACH Technologies, Wayne, Pennsylvania

1Introduction

The threaded mechanical fastener is anengineered structural product. It is controlled inselection, manufacturing, and processing by aseries of specifications under the jurisdiction ofsuch organizations as ASME, American Societyfor Testing and Materials (ASTM), Society of


Automotive Engineers (SAE), and theDepartment of Defense.

The use of specifications is voluntary in theUnited States. However, a purchase order andcontract for procurement make them subject tothe laws of business contracts, and they mustmeet these requirements. Fastener materials areselected either by the fastener manufacturer toconform to product requirements or by acustomer to meet the needs of the environmentto which the fastener will be subjected. Thepurpose of a bolt and nut is to provide clampingforce, and when you purchase a bolt you are, ineffect, buying clamping force. The nut, by itsdesign and application, enables you to achievethat clamping force. The selection of materialsfor bolts and nuts is based on maximum strength,high ductility, and high notch strength. Otherchoices are based on the environmental aspectsof bolt and nut applications, and these arediscussed here in separate categories.


Fastener materials are initially selected forengineering products. However, the fabricationand performance requirements of the fastenergovern the ultimate choice. One objective offastener design is to achieve a threaded productwith properties that match those of the basematerial.

Material selection is further based on capabilityof being fabricated and the material's response toheat treating and cold work strengtheningmechanisms.

In this chapter, consideration is given to theabove in discussing fastener selection. Alsoincluded are the reasons for selecting fastenermaterials for such environmental conditions ascorrosion, elevated temperatures, cryogenictemperatures, hydrogen embrittlement, and stresscorrosion. Coatings selection in conjunctionwith the base material is included as an adjunctto the basic fastener design. Fabricationcapability is covered in the appropriate section


on material classes that respond best to differentenvironments.


2Carbon and Alloy Steels

No other structural material offers the high strength at lowcost that can be achieved by steel. For this reason fastenersmade of steel constitute the majority of fasteners usedworldwide. The automotive industry, which uses about 25%of all fasteners used, uses nuts and bolts of carbon or alloysteel almost exclusively in automobile and truckmanufacture.

Carbon steels, containing essentially iron and carbon, are theleast expensive of the steels and are used whenever possible.A carbon steel, in addition to the carbon, contains less than1.60% manganese (Mn), 0.60% silicon (Si), and 0.60%copper (Cu). The balance, except for residual elements suchas sulfur (S) and phosphorus (P), is iron. Hot-rolled carbonsteels with a low carbon content are used for machine screwsand non-heat-treated bolts and nuts. They are at the bottom ofthe strength scale.

Steel is categorized according to grades established by theAmerican Iron and Steel Institute (AISI) and by the Society ofAutomotive Engineers (SAE).

Carbon steels range from SAE 1006 (0.08% max C) to 1095


(0.90% C). The low carbon range (SAE 10061025) is used inthe hot-rolled condition. Higher carbon grades (SAE10301040) are used in the heat-treated condition. It isrecommended that fasteners not be made with carbon levelsgreater than 0.45% because of the loss of ductility andsusceptibility to hydrogen embrittlement and stress corrosioncracking. Grades have been made with carbon greater than0.45%, but they have required very special fabricationtechniques.

The heat-treated grades are selected on the basis ofhardenability. This is a measure of the depth to which a steelcan be hardened in different quenching media during thehardening phase of heat treatment. All fasteners (with theexception of self-drilling and self-tapping screws) must bethrough-hardened. This provides for uniform properties inwhat is in many cases a small cross section.

The strength of the cross section is not only important intension. A large number of screws are used in shear, and theshear strength must be in a consistent ratio to the tensilestrength. This is not possible with a partially hardened crosssection.

Screw material is selected on the basis of hardenability andcross section. A typical selection of carbon steels forfrequently specified fastener grades is shown in thespecifications listed in Table 1.


One group of carbon steels are known as the high manganesesteels. They are the 1500 series of steels. These providehigher hardenability at low cost. One permitted in the Grade 8fastener up to 7/16 in. diameter is the SAE 1541 steel with0.360.45% C and 1.301.65% Mn.

Because of their limited hardenability, carbon steels arelimited in diameter as the strength increases. When thediameter becomes too large for the carbon steel to providethrough-hardening during heat treatment, a choice of alloysteels must be made. The selection of alloy steels is based ontheir ability to provide higher hardenability. A great manyalloy steels areTABLE 1 Strength Grades for SAE J429 Bolt MaterialsSAE J429[1] ASTM equivalent Diameter

(in.)Ultimate tensile strength (UTS)

(psi)Grade 2 A307 [2] 1/43/4 74,000

Over 3/411/2 60,000

Grade 5 A325 [3] 1/41 120,000

Grade 8 A490 or A354 BD[4,5] 1/41 1/2 150,000

Source: As noted.


TABLE 2 Strength Grades for SAE J995 Nut MaterialsASTM equivalent,a Proof load

stress (psi)Carbon

(SAE) (%)SAE J99579 [6] A56394 [7]a Diameter (in.) UNC UNF Min. Max.Grade 2 A 1/41 1/2 90,000 90,000 .47Grade 5 B 1/41 120,000 109,000 .30 .55

>11 1/2 105,000 94,000Grade 8 D 1/41 1/2 150,000 150,000 .30 .55aASTM A563, for the A and B grades of nuts, has only a 0.55% C maximumrequirement, whereas Grade D has 0.30% Mn in addition.

not used because their greater strength or higher cost is notjustified. Most of the alloy steels contain varyingpercentages of nickel (Ni), chromium (Cr), and molybdenum(Mo).

Strength levels and grades for bolts, screws, and steels areselected to conform to Table 1 in general usage. While thecurrent SAE J429 specification for Grade 8 permits SAE1541, a manganese carbon steel, to be used for diameters of7/16 in. and under, ASTM A354BD permits only alloy steelfor all diameters. As a rule, the user does not, in thiscategory, select the carbon or alloy grade but leaves it to thediscretion of the fastener manufacturer.

There is a family of special steels containing boron (B) thatprovide a high degree of hardenability to both low carbon


(0.20% C) and medium carbon (0.35% C) steels. Althoughthese steels have been permitted in special J429 grades,namely 5.2 and 8.2, they have not been allowed in thestandard grades. A newly amended J429 (still in committee)is expected to permit medium carbon boron steels to bepermitted for Grade 8 fasteners.

In fastener design considerations, the nut must be strongenough to break a fastener designed for use in tension. Thenut has an advantage the bolt does not have. When a boltfails, the failure is usually concentrated in the area of higheststress, which is the area of joint separation. The nut will fail,if insufficient threads are provided, by the phenomenon ofstripping when the shear strength of the total threads cannotcarry the load. Therefore, a nut designed with a greaternumber of threads does not need as high a material strength.As an example, in the fastener research laboratory where Iwas employed at the time, we broke a 300,000 psi bolt withan aluminum nut by merely increasing the number of threadsengaged.

There are limitations to the height of nut that can be used,and therefore some nuts must be heat treated to strengthlevels higher than that provided by low carbon steel.

Nut material requirements in the specifications forcommonly used grades are listed in Table 2. Carbon steel ispermitted even for 150,000 psi nuts. This is because of the


strength latitude permitted by nut dimensions and threadengagement. ASTM A563-94 lists a number of other nutstrength requirements based on nut dimensions and somegrades that have mild corrosion resistance to the weather.

In summary, steel bolt and nut material selection is basedprimarily on bolt and nut mass created by dimensions as wellas steel hardenability. Material should be selected on thebasis of the lowest alloy content that will provide thestrength required.

3Discontinuities

From a fabrication standpoint, except for those with largediameters, bolts are cold forged. Because of the severesurface deformation created during cold heading, boltingmaterials must have relatively seam-free surfaces. Surfacediscontinuities in the final product can come from


seams, forging cracks, and quench cracks thatoccur during heat treating. Various levels ofdiscontinuities are permitted in the final productbased on depth and location. Proprietary rawmaterial specifications issued by bolt and nutproducers contain allowable discontinuity depthlimits. Bolt and nut product discontinuity depthlimits are found in SAE specifications J123c andJ122a, respectively [8].

4Elevated Temperatures (Steel)

A question is often raised about the suitability ofcarbon and alloy steel fasteners at temperaturesabove room temperature. SAE J429 does notprovide for heat treatment of Grade 2 bolts.However, Grades 5 and 8 must be tempered at800°F min.

A fastener must provide reliable clamping force


at the in-service temperature. Clamping forcedepends, in part, on the modulus of elasticity,which is a measure of the elasticity orspringback. The modulus drops in value as thetemperature rises [9, pp. 115117] (see Table 3).As a result, loss of clamping force can occur atthe service temperature and must also be takeninto consideration.

The strength of the fastener is reduced at elevatedtemperatures, but this does not happen below thetempering temperature of the steel. However,there is a phenomenon known as stress relaxationthat is a form of elevated temperature creep thatreduces the clamping force as well. Stressrelaxation occurs when elastic deformationbecomes plastic or permanent and the loss ofelasticity lowers the clamping force. Bolts andscrews provide clamping force because the boltstretches during tightening. Its desire to go backto its original length provides the compressionload that holds the joint together. Lose this and a


loose joint connection occurs.

Unfortunately, very little data on stressrelaxation exist for carbon and steel alloys. Themost extensive data can be found in Ref. 9 (Table4.7 and Fig. 4.3, pp. 118, 120, 121). These arereproduced here as Table 4 and Figure 1,respectively. Note that the relaxation strength isanywhere from 200 to 400°F below thetempering temperature.

Some comparative data exist that show therelationship between carbon and alloy steel andboron steel fasteners. One study [10] shows thatfasteners made of 1340 steel (0.380.43% C;1.601.90% Mn) relax less than 10B22 fasteners.The same study with nuts shows that nutscontribute little to relaxation.

5Corrosion (Steel)

Carbon and alloy steel fasteners have little or no


corrosion resistance and must be coated to beused in even low moisture environments.Coatings are used to protect the fasteners, andthese present a set of problems not experiencedby uncoated fasteners. This is discussed inSection 7 on coatings.

6High and Ultrahigh Strength Steel Bolts

In general industrial usage, Grade 8 and A490are designed as ''high strength steel" bolts. Thereare, however, a number of steel bolts and screwswith higher strength. The earliest of these was thesocket head cap screw (SHCS) with a minimumtensile strength of 180,000 psi. Other socketscrew designs have 160,000 psi tensile strength.

The steel aircraft fastener category includes theMS20004 series of internal wrenching head boltsat 160,000 psi. At 180,000 psi are NAS624 andthe military version, MS21250. These have 12-point wrenching heads, as do all fasteners above


160,000 psi.

In the late 1950s a series of high strength aircraftbolts were developed by one manufacturer. Thematerial used was H-11, a hot work die steelwith excellent heat treatment response and atempering temperature of approximately 1000°F,making it capable of high tem-


TABLE 3 Modulus of Elasticity at Various Temperatures (×106 psi)Temperature °F (°C)

325 200 70 400 600 800 1000 1200 1400 1600 1800Specification Grade (198) (129) (20) (204) (316) (427) (538) (649) (760) (871) (982)ASTM A193 B5 32.9 32.3 30.9 29.0 28.0 26.1

B6 31.2 30.7 29.2 27.3 26.1 24.7B7 31.6 31.0 29.7 27.9 26.9 25.5B8-CL 1 30.3 29.7 28.7 26.5 25.3 24.1B16 31.6 31.0 29.7 27.9 26.9 25.5

ASTM A307 31.4 30.8 29.5 27.7 26.7 24.2ASTM A320 L7 31.6 31.0 29.7 27.9 26.9 25.5

L43 31.6 31.0 29.7 27.9 26.9 25.5B8 30.3 29.7 28.3 26.5 25.3 24.1

ASTM A325 Type 1, 2, 3 31.4 30.8 29.5 27.7 26.7 24.2ASTM A354 31.2 30.6 29.3 27.5 26.5 24.0ASTM A449 31.2 30.6 29.3 27.5 26.5 24.0ASTM A453 30.3 29.7 28.3 26.5 25.3 24.1ASTM A490 31.2 30.6 29.3 27.5 26.5 24.0ASTM A540 B21, B22 31.6 31.0 29.7 27.9 26.9 25.5

B23, B24 29.6 29.1 27.8 26.1 25.2 23.0SAE J429 GR 1, 2, 4 31.4 30.8 29.5 27.7 26.7 24.2

GR 5, 7, 8 31.2 30.6 29.3 27.5 26.5 24.0High Performance materials A286 29.1 23.5 21.1 18.7 18.9

Hasteloy-X 28.6 18.5H-11 30.6 23.0 16.0Inconel 600 31.4 23.1Inconel 718 29.0 20.0Inconel X750 31.0MP35N 33.6 30.8Nimonic 80A 31.2 22.7Ti 6AI-4V 16.5Rene 41 25.9 23.6Waspaloy 30.6 26.7 22.7

Table continues


ContinuedTABLE 3

Temperature °F (°C)325 200 70 400 600 800 1000 1200 1400 1600 1800

Specification Grade (198) (129) (20) (204) (316) (427) (538) (649) (760) (871) (982)Metric 4.6/4.8/5.8 31.4 30.8 29.5 27.7 26.7 24.2SAE J1199 8.8/9.8/10.9 31.2 30.6 29.3 27.5 26.5 24.0Joint materials Steels Low carbon 31.4 30.8 29.5 27.7 26.7 24.2 20.4

Med. Carbon 31.2 30.6 29.3 27.5 26.5 24.0 20.2Carbon-moly. 31.1 30.5 29.2 27.4 26.4 23.9 20.1Chrome-moly. 31.6 31.0 29.7 27.9 26.9 25.5 23.9Austenitic 30.3 29.7 28.3 26.5 25.3 24.1 22.82024 Aluminum 11.7 11.4 10.6 9.2

SAE J158 Malleable iron castings 2526SAE J434 Ductile iron castings 22

Cast aluminum 810Brass, cold-rolled 13.1Cast iron 1214Wrought iron 2629Magnesium 6.1Concrete 2.0Wood (with grain) 1.2

Source: Ref. 9.


TABLE 4 Stress RelaxationTesttemp.

Residual stress(ksi) Final

Bolt information °F °C Initial stress(ksi)

100hr

1000hr

10,000hr

Stress(ksi)

Time(hr)

ASTM A193 B7 1 in. diam. 850 454 30.0 13.2 (10.0) 1,920ASTM A453 GR 651 1 1/4 in. diam.; as recd plus 1200°F (650°C)for 48 hr 1200649 32 5 (2.85) 1,012

ASTM A453 GR 655: 2 in. diam. 1100593 19.9 14.315.9 16.0 3,769Type 431 stainless steel (AN6C) 700 371 95.8 68 67.8 150

900 482 95.8 23 20 150900 482 63.25 20 18.8 150

Austenitic SS DIN 17006: X5CrNiMo 1810 M12 bolt. Nut: DIN931 or 934: rolled threads

8421022

450550

50.650.6

46.534.1

44.722.9

A286 bolt: 5/16 in. diam. 1200649MS 9035-19 configuration 70 52.9 100HiR thread form per AMS 7478 70 34 100

Nimonic 80A: 3.5 in. diam. 1202650 39.5 30.424.2 17.8 14.2 21,1671292700 37.1 28.522.1 15.7 2,9001499815 29.6 9.8 5.7 290

Nimonic 80A: 1 1/8 in. diam. 1157625 33.1 28 24.2 18.6 15.7 27,200Bolt: AISI 8740 steel (M12) 662 350 45 39 31.5Nut: German aircraft spec. LN 1.4944 85.2 67.552.5Vascojet 1000 (modified H-11 steel) 1/220 studs 900 482 117 78MP35N: 1/428 bolt; nut of same alloy 700 371 132 113 150Rene 41 bolt 1400760Waspaloy nut

1032 threads 99.7 25.5 501/4 threads 73 32.9 501/220 threads 73 35.1 501/220 threads 46.4 28.3 50

Udimet 500 bolt, 5/16 threads 1100593 50 36.5 12Waspaloy nut and bolt, MIL-S-8879 thread form: 1/428 threads 1400760 100 7.7 50Source: Ref. 9.


Figure 1Stress relaxation of petrochemical bolting materials as a function of service temperature.

Explosure in each case was for 1000 hr at the temperature shown.(From Ref. 9.)

perature use to 900°F. The composition of H-11 is 0.40 C,0.30 Mn, 1.00 Si, 5.00 Cr, 0.04 V, and 1.30 Mo. From thismaterial a series of fasteners were developed at 220,000 psiand 260,000 psi tensile strengths.

All aircraft fasteners are available in a shear bolt version with


minimum shear strength at 60% of minimum ultimate strength.(See also Chapter 21.)

The highest strength commercially available bolt has aminimum tensile strength of 300,000 psi and is made fromT15, whose composition is 1.501.60 C, 0.100.40 Mn,0.100.40 Si, 4.504.75 Cr, 4.7555.00 V, 12.513.50 W, 0.50Mo, and 4.755.25 Co.

Further research on steel and fastener fabrication resulted infasteners with tensile strengths in the vicinity of 400,000 psi,hardnesses of Rc 6567, and sufficient ductility to produceacceptable properties. Fatigue was measured at an endurancelimit of 180,000 psi (R = 0.1). When threads were rolled afterheat treatment, the fatigue strength continued to climb, with noperceived upper limit. Unfortunately, at a hardness of Rc 66 thethread roll die life was only one bolt per die, and the projectwas abandoned with the assessment that it had no commercialpotential.

These ultrahigh strength steels were in service for perhaps 45years, when a field failure was observed that was determined tobe stress corrosion cracking. This prompted an extensive studyby one manufacturer over the next several years [11] thatencompassed a coatings development program and analternative materials program. The outcome of the research isdiscussed in Section 7. In summary, carbon and alloy steelfasteners offer a series of strength levels that will cover a greatmany application requirements. Their use at temperatures


greater than room temperature creates a number of designproblems that are often better served by going to fastenermaterials designed for elevated temperatures. If carbon or alloysteels are selected for economic reasons at elevatedtemperatures, these potential problems should be discussedwith experts familiar with high temperature joint requirements.

7Corrosion-Resistant Materials and Coatings

The carbon and alloy steels do an excellent job of providinglow to high strength fasteners at low to moderate prices. Forthis reason they dominate the choices in fastener materials.How-


ever, they will rust in normal atmospheres andwill corrode substantially in more corrosiveenvironments. For this reason corrosion-resistantalloys and coatings have been developed. Manyof these are applicable to fasteners. Some,because of their high nickel and/or chromiumcontent, are also used at elevated and cryogenictemperatures. They are discussed both in thissection and in Section 10.

The primary selection of fastener materials forcorrosion resistance is based on properties of thebasic material. Since the fastener must havecorrosion compatibility with the joint material, ithas been a challenge for the fastenermanufacturer to provide fasteners of the samematerial as that called for by the designmetallurgist.

Initially, corrosion-resistant fasteners were madeof pure nonferrous metals such as copper. Later,


the addition of zinc produced the brasses, and thefurther addition of tin created the bronzes. Nickeland nickel alloys were to follow; the mostpopular of these are the cupronickel alloys andMonel; the latter is used extensively in marineapplications. Aluminum has been used to someextent, primarily in nuts, where its light weightand lower strength requirement make it anexcellent choice (see Chapter 14).

Titanium came along in the 1950s after a processto produce it was developed. Although titaniumis often used where high strength and low density(i.e., a high strength/density ratio) are required, ithappens to be one of the most corrosion-resistantmaterials in the world. It rates just belowplatinum and gold in this regard as long as itsoxide surface is intact. Titanium has been widelyused in aircraft, first for military applications andlater in commercial aircraft. It was not availablein fastener form until the late 1950s. Titaniumbecame widely used in aircraft after the


development of the C-5A transport for themilitary. Today, based on C-5A technology, it isused extensively in alloy form (6A1-4V) for theskin fasteners of all the major commercial wide-bodied transports. Because of its light weight andbecause it develops an anodized surface tominimize corrosion, an aluminum alloy is usedfor the companion nuts. Many non-aircraft usesfor titanium are currently being developed as itsproperties become more widely appreciated.

Nonferrous structural fasteners are the subject ofASTM F467-93 [12] and F468-93 [13]. Tables5 and 6 (F467) and Tables 7 and 8 (F468) showthe chemical compositions and mechanicalrequirements of these specifications.

Since their development in the early part of thiscentury, stainless steels have become the mostpopular substitute for carbon and alloy steelswhere long-term corrosion resistance is required.Modern fastener manufacturing techniques havemade them more economical. The structural


stainless steel fasteners are covered in ASTMF593-95 [14] and F594-95 [15]. Tables 9 and10 (F593) and Tables 11 and 12 (F594] showthe chemical composition and mechanicalproperty requirements for these materials.

In spite of good corrosion resistance, thenonferrous and stainless steel fasteners could notmatch the strengths of the existing alloy steelbolts. Therefore, the fastener manufacturer waschallenged to provide high strength corrosion-resistant bolts and nuts. (It should be noted thatbolts and nuts are usually made of the samematerial in order to provide corrosioncompatibility.)

First the higher strength 400 series stainlesssteels were used. However, they lacked thecorrosion resistance of the 300 series andpresented fastener ductility problems in thehigher strength levels. The materials thatprovided good corrosion resistance and higherstrength were, in many cases, also high


temperature alloys. The earliest of these was A-286, a high Cr-Ni stainless steel with goodproperties to 1000°F. A-286 can be heat treatedto 125,000 psi. However, one manufacturer,using cold-working prior to aging, developed a200,000 psi version.

Another corrosion-resistant material with goodstrength and high temperature capability isInconel 718. This high nickel alloy can be heattreated to 180,000 psi. Again, this samemanufacturer developed, with cold working andaging, a 220,000 psi variety.


(table continued on next page)


(table continued from previous page)


TABLE 6 Mechanical Property Requirements, ASTM F 467Alloy Mechanical property

markingHardness,

minaProof stress, min

(ksi)Cu 110 F 467A 65 HRF 30Cu 270 F 467B 55 HRF 60Cu 462 F 467C 65 HRB 50Cu 464 F 467D 55 HRB 50Cu 510 F 467E 60 HRB 60Cu 613 F 467F 70 HRB 80Cu 614 F 467G 70 HRB 75Cu 630 F 467H 85 HRB 100Cu 642 F 467J 75 HRB 75Cu 651 F 467K 75 HRB 70Cu 655 F 467L 60 HRB 50Cu 661 F 467M 75 HRB 70Cu 675 F 467N 60 HRB 55Cu 710 F 467P 50 HRB 45Cu 715 F 467R 60 HRB 55Ni 335 F 467S 20 HRC 115N 276 F 467T 20 HRC 110Ni 400 F 467U 75 HRB 80Ni 405 F 467V 60 HRB 70Ni 500 F 467W 24 HRC 130AI 2024-T4b F 467X 70 HRB 55

AI 6061-T6 F 467Y 40 HRB 40


AI 6061-T6 F 467Y 40 HRB 40AI 6262-T9 F 467Z 60 HRB 52Ti 1 F 467AT 140 HV 40Ti 2 F 467BT 150 HV 55Ti 4 F 467CT 200 HV 85Ti 5 F 467DT 30 HRC 135Ti 7 F 467ET 160 HV 55Ti-38-6-44 F 467FT 24 HRC 120Ti 5 ELI F 467GT 35 HRC 125a For aluminum and titanium alloys, hardness values are for informationonly.bAluminum and titanium alloy 2024-T4 shall be supplied in naturally agedcondition. This material is not recommended for nuts in sizes greater than1/4 (0.250) in.Source: Ref. 12. Copyright ASTM; reprinted with permission.

8Stress Corrosion and Hydrogen Embrittlement

In the discussion of ultrahigh strength steels inSection 6, it was pointed out that these steels aresubject to stress corrosion cracking. Stress corrosionis a phenomenon whereby a corrosive materialattacks the surface of the metal (which is under highstress), creating a series of fine intergranular andoften transgranular cracks that initiate a failure. Asmentioned earlier, H-11, the workhorse of the


220,000 and 260,000 psi fastener steels, wassusceptible to stress corrosion, and a major effortwas undertaken to develop alternative materials aswell as coatings to combat this problem.


Before discussing this effort, it should beunderstood that stress corrosion cracking is nothydrogen embrittlement. (See Chapters 48 and49.) It is possible to have hydrogen releasedduring exposure to corrosive elements if agalvanic cell or differential aeration cell iscreated. This hydrogen can cause brittle failure,even in the absence of any significant corrosion.Hydrogen embrittlement will be encouraged ifthe fastener is exposed to hydrogen through acidcleaning and/or electroplating duringmanufacture. It can occur at hardness levelsabove Rc 35 in steel after the bolt is tightened toclamp the joint. It is interesting to note that H-11steel has a high resistance to hydrogenembrittlement but not to stress corrosioncracking (SCC).

Early research developed a test intended to detectstress corrosion cracking in the laboratory. The


results of this effort are described in Ref. 11. Thesolution to the problem did not involve reducingthe clamping force, as failure was observed withloads as low as approximately 30% of the tensilestrength.

Somewhat higher resistance to failure occurredwhen threads were rolled after heat treatmentwhen the failure occurred in the shank instead ofthe threads. This was not a solution to theproblem either. Some of the alloys that wereexamined for resistance to SCC included PH 13-8 Mo, Custom 455, PH 12-9 Mo, and Inconel718 [16]. The alloy steels 8740, 4037, and 4340,excellent for their strength capability, could notwithstand corrosion or stress corrosion.

Research continued in an effort to find a materialthat would replace H-11 for 260,000 psi bolts.An alloy containing 35% Ni, 20% Cr, 10% Mo,and 35% Co was offered to the fastenermanufacturer by the inventor. After almost fiveyears of research, it was produced as a material


and developed into a fastener grade material, thealloy MP35N. This was capable of strengths upto 260,000 psi with good ductility as well asexcellent resistance to corrosion and stresscorrosion [17]. In addition, this material offeredthe highest resistance to fatigue failure of anyfastener material yet developed (except for the400,000 psi steels) and is even used inapplications where SCC is not a problem, such asautomotive connecting rod bolts for racing cars.It is used in landing gear bolts for somecommercial aircraft where high strength and highfatigue resistance as well as resistance to SCCare needed.

An extensive study of the MP35N resistance tovarious environments was conducted [18]. Someof these data are shown in Tables 13 and 14.

Fasteners may also be subject to crevicecorrosion in which the corrosive material willinvade the cracks of the joint or under the nutand head of the bolt as well as along the shank.


Materials such as PH 13-8 Mo could not resistthis attack and were rejected.

The materials that showed the most resistance toSCC, had good fatigue life, and could befabricated were Inconel 718 and MP35N. Astudy including these alloys [19] presents anevaluation of these materials.

A major aircraft manufacturer, after experiencingSCC with H-11 standard bolts, switched toInconel 718 (220,000 psi) and MP 35N(260,000 psi). Other aircraft manufacturers soonfollowed.

9Coatings

An alternative to using corrosion-resistance boltmaterials is to use coatings. Sacrificial coatingssuch as zinc and cadmium will corrode inpreference to the base metal and provide limitedprotection in industrial and marine atmospheres.


One of the reasons the resistance is limited is thevery thin layer that must be controlled to avoidthread interference. Hot dip galvanizing is usedin structural bolting. It gives a heavier coating,but the nut must be retapped after coating toprevent thread interference.

Another problem with sacrificial coatings is thathydrogen is liberated during plating and stepsmust be taken to minimize hydrogenembrittlement. Field exposure may also provideenough hydrogen through the corrosionmechanism to occasionally cause hydrogenembrittlement.


(table continued from preivous page)


TABLE 8 Mechanical Property Requirements, ASTM F 468Full-size testsb Machined speciment tests

Alloy Mechanicalproperty marking

Nominaldiameter, in. Hardness Tensile

strength (ksi)Yield strength,

min (kis)Tensile strength

min (ksi)Yield strength

min (ksi)Elongation in 4D,

min (%)dCopper

Cu110 F 468A All 6590

HRF 3050 10 30 10 15

Cu270 F 468B All 5580

HRF 6090 50 55 50 35

Cu462 F 468C All 6590

HRB 5080 25 50 25 20

Cu464 F 468D All 5575

HRB 5080 15 50 15 25

Cu510 F 468E All 6095

HRB 6090 35 55 30 15

Cu618 F 468F

7095HRB7095HRB

8011075105

5045

8075

5045

3030

Cu614 F 468G All 7095

HRB 75110 35 75 35 30

Cu630 F 468H All 85100

HRB 100130 50 100 50 5

Cu642 F 468J All 7595

HRB 75110 35 75 35 10

Cu651 F 468K

7595HRB7095HRB

701005590

5540

7054

5338

88

Cu655 F 468L All 6080

HRB 5080 20 50 15 20

Cu661 F 468M All 7595

HRB 70100 35 70 35 15

Cu675 F 468N All 6090

HRB 5585 25 55 25 20

Cu710 F 468P All 5085

HRB 4575 15 45 15 40

Cu715 F 468R All 6095

HRB 5585 20 55 20 45

NickelNi335 F 468S All 2032

HRC 115145 45 115 45 35

Ni276 F 468T All 2032

HRC 110140 45 110 45 25



(table continued from previous page)Full-size testsb Machined speciment tests

Alloy Mechanicalproperty marking

Nominaldiameter, in. Hardness Tensile

strength (ksi)Yield strength,

min (kis)Tensile strength

min (ksi)Yield strength

min (ksi)Elongation in 4D,

min (%)d

Ni 400 F 468U75 HRB25HRC60 HRB25HRC

8013070130

4030

8070

4030

2020

Ni 400HFe F 468HF All 60 HRB95

HRB 70120 30 70 30 20

Ni 405 F 468V All 60 HRB20HRC 70125 30 70 30 20

Ni 500 F 468W2437HRC2437HRC

130180130180

9085

130130

9085

2020

AluminiumAl2024-T4f

F 468X All 7085 HRB 5570 36 62 40 10

Al6061-T6f

F 468Y All 4050 HRB 3752 31 42 35 10

Al7075-T73f

F 468Z All 8090 HRB 6176 50 68 56 10

Titaniumg

Ti 1 F 468AT All 140160HV 4070 30 35 25 24

Ti 2 F 468BT All 160180HV 5585 45 50 40 20

Ti 4 F 468CT All 200220HV 85115 75 80 70 15

Ti 5 F 468DT All 3039HRC 135165 125 130 120 10

Ti 7 F 468ET All 160180HV 5585 45 50 40 20

Ti-38-6-44 F 468FT All 2438

HRC 120150 115 120 115 15

Ti 5 ELI F 468GT All 2936HRC 125165 115 125 115 10

aWhere both tension and hardness tests are performed, the tension tests shall take precedence for acceptence purposes. Foraluminium alloys hardness tests are for information only.bThe yield and tensile strength values for full-size products shall be computed by dividing the yield and maximum tensile load by thestress area for the product size and threaded series as given in table on tensile stress areas.cYield strength is the stress at which an offset of 0.2% gage length occurs.dElongation is determined using a gage length of 4 diameters of test specimen with Test Methods E 8.e''HF" denotes a hot-formed product.fAluminium alloy temper designations are in accordance with ANSI H35.1gFull-size test mechanical properties apply to fasteners with a maximum diameter of 3 in. (76 mm). Mechanical properties of largersections shall be negotiated between the material manufacturer and the fastener producer.


Source: Ref. 13. Copyright ASTM; reprinted with permission.


TABLE 9 ASTM F 593 Chemical RequirementsComposition (% maximum except as shown)

Alloy group UNS designation Alloy C Mn P S Si Cr Ni Cu Mo OthersAustenitic alloys

1 S30300 303 0.15 2.00 0.20 0.15 min 1.00 17.019.0 8.010.0 0.60 maxa1 S30323 303Se 0.15 2.00 0.20 0.060 1.00 17.019.0 8.010.0 Se 0.15 min1 S30400 304 0.08 2.00 0.045 0.030 1.00 18.020.0 8.010.5 1.00

S30403 304L 0.03 2.00 0.045 0.030 1.00 18.020.0 8.010.0 1.001 S30500 305 0.12 2.00 0.045 0.030 1.00 17.019.0 10.513.0 1.001 S38400 384 0.08 2.00 0.045 0.030 1.00 15.017.0 17.019.0 1.752.25 0.05 maxa1 S20300 XM1 0.08 5.06.5 0.040 0.180.35 1.00 16.018.0 5.06.5 3.04.01 S30430 XM7 0.10 2.00 0.045 0.030 1.00 17.019.0 8.010.02 S31600 316 0.08 2.00 0.045 0.030 1.00 16.018.0 10.014.0 2003.00

S31603 316L 0.03 2.00 0.045 0.030 1.00 16.018.0 10.014.03 S32100 321 0.08 2.00 0.045 0.030 1.00 17.019.0 9.012.0 Ti 5 × C min3 S34700 347 0.08 2.00 0.045 0.030 1.00 17.019.0 9.013.0 Cb + Ta 10 × C min

Ferritic alloys4 S43000 430 0.12 1.00 0.040 0.030 1.00 16.018.04 S43020 430F 0.12 1.25 0.060 0.15 min 1.00 16.018.0 0.60 maxa

Martensitic alloys5 S41000 410 0.15 1.00 0.040 0.030 1.00 11.513.55 S41600 416 0.15 1.25 0.060 0.15 min 1.00 12.014.0 0.60 maxa5 S41623 416Se 0.15 1.25 0.060 0.060 1.00 12.014.0 Se 0.15 min6 S43100 431 0.20 1.00 0.040 0.030 1.00 15.017.0 1.252.50

Precipitation hardening alloy7 S17400 630 0.07 1.00 0.040 0.030 1.00 15.017.5 3.05.0 3.05.0 Cb + Ta 0.150.45aAt manufacturer's option, determined only when intentionally added.Source: Ref, 14. Copyright ASTM; reprinted with permission.


TABLE 10 ASTM 593 Mechanical Property RequirementsaFull-size tests Machined specimen tests

Stainless alloy group Conditionb Alloy mechanicalproperty marking

Norminaldiameter

(in.)

Tensilestrengthc

(ksi)

Yieldstrengthc,d

(ksi)

Rockwellhardness

Tensilestrengthc

(ksi)

Yieldstrengthc,d

(ksi)

Elongationin 4D (%)

1 (303, 304, 304L, 305,384, XM1, XM7, 303Se)

AFACW1CW2SH1SH2SH3SH4

F593AF593BF593CF593DF593AF593BF593CF593D

1/41 1/2,incl1/41 1/2,incl1/45/8,incl3/41 1/2,incl1/45/8,incl3/41, incl1 1/811/4, incl1 3/811/2, incl

85 max751001001508514012016011015010014095130

30654595756045

B85 maxB6595B95C32B80C32C24C36C20C32B95C30B90C28

80max7095801151059590

50 max30604090705550

4030202512152028

2 (316, 316L)


F593EF593FF593GF593HF593EF593FF593GF593H


85 max751001001508514012016011015010014095130

30654595756045


80max7095801151059590

50 max30604090705540

4030202512152028

3 (321, 347)

AFACW1CW2SH1HS2SH3SH4

F593JF593KF593LF593MF593JF593KF593LF593M


85 max751001001508514012016011015010014095130

30654595756045


80max7095801151059590

50 max30604090705540

4030202512152028

Table continues


ContinuedTABLE 10

Full-size tests Machined specimen testsStainless alloygroup Conditionb Alloy mechanical

property markingNominaldiameter

(in.)

Tensilestrengthc

(ksi)

Yieldstrengthc,d

(ksi)

Rockwellhardness

Tensilestrengthc

(ksi)

Yieldstrengthc,d

(ksi)

Elongation in4D (%)

Ferritic alloys

4 (430,430F)

ACW1CW2

F593NF593VF593W

1/41 1/2,incl1/45/8, incl3/41 1/2,incl

701007510570100

354035

B6595B7598B6595

707070

353030

252020

Martensitic alloys

5 (410, 416,416Se)

HHT

F593PF593R

1/41 1/2,incl1/41 1/2,incl

110140160190

90120

C2030C3445

110160

90120

1812

6 (431) HHT

F593SF593T

1/41 1/2,incl1/41 1/2,incl

125150180220

100140

C2532C4048

125180

100140

1510

Precipitation hardening alloys7 (630) AH F593U 1/41 1/2,

incl 135170 105 C2838 135 105 16

a Minimum values except where shown as maximum or as a range.b Legend of conditions:A Machined from annealed or solution-annealed stock thus retaining the properties of the original material, or hot-formed and solution-annealed.AF Headed and rolled from annealed stock and then reannealed.AH Solution annealed and age-hardened after forming.CW Headed and rolled from annealed stock thus acquiring a degree of cold work; sizes 0.75 in. and larger may be not worked and solution-annealed.H Hardened and tempered at 1050°F (565°C) minimum.HT Hardened and tempered at 525°F (274°C) minimum.SH Machined from strain hardened stock or cold-worked to develop the specified properties.cThe yield and tensile strength values for full-size products shall be computed by dividing the yield and maximum tensile load values by thestress area for the product size and thread series determined in accordance with Test Methods F 606 (see Table 4).dYield strength is the stress at which an offset of 0.2% gage length occurs.Source: Ref. 14. Copyright ASTM; reprinted with permission.


TABLE 11 ASTM F 594 Chemical RequirementsComposition (% maximum except as shown)

Alloy group UNS designation Alloy C Mn P S Si Cr Ni Cu Mo OthersAustenitic alloys

1 UNS30300 303 0.15 2.00 0.20 0.15 min1.0017.019.0 8.010.0 0.60 maxa1 UNS30323 303Se0.15 2.00 0.20 0.600 1.0017.019.0 8.010.0 Se 0.15 min1 UNS30400 304 0.80 2.00 0.045 0.030 1.0018.020.0 8.010.51 UNS30500 305 0.12 2.00 0.045 0.030 1.0017.019.0 10.513.01 UNS38400 384 0.08 2.00 0.045 0.030 1.0015.017.0 17.019.01 UNS20300 XM1 0.085.06.50.0400.180.351.0016.018.0 5.06.5 1.752.25 0.50 maxa1 UNS30430 XM7 0.10 2.00 0.045 0.030 1.0017.019.0 8.010.0 3.004.002 UNS31600 316 0.08 2.00 0.045 0.030 1.0016.018.0 10.014.0 2.003.003 UNS32100 321 0.08 2.00 0.045 0.030 1.0017.019.0 9.012.0 Ti 5 × C min; Cb + Ta 10 × C min3 UNS34700 347 0.08 2.00 0.045 0.030 1.0017.019.0 9.013.0

Ferritic alloys4 UNS43000 430 0.12 1.00 0.040 0.030 1.0016.018.04 UNS43020 430F 0.12 1.25 0.060 0.15 min1.0016.018.0 0.60 maxa

Martensitic alloys5 UNS41000 410 0.15 1.00 0.040 0.030 1.0011.513.55 UNS41600 416 0.15 1.25 0.060 0.15 min1.0012.014.0 0.60 maxa5 UNS41623 416Se0.15 1.25 0.060 0.060 1.0012.014.0 Se 0.15 min6 UNS43100 431 0.20 1.00 0.040 0.030 1.0015.017.0 1.252.50

Precipitation hardening alloy7 UNS17400 630 0.07 1.00 0.040 0.030 1.0015.017.5 3.005.00 3.005.00 Cb + Ta 0.150.45a At manufacturer's option, determined only when intentionally added.Source: Ref. 15. Copyright ASTM; reprinted with permission.


TABLE 12 ASTM F 594 Mechanical Property Requirementsa

Stainless alloy group Conditionb Alloy mechanicalproperty marking

Nominaldiameter,

in.

Proofstress, ksi,

(min)

Rockwellhardness

1 (303, 304, 305, 384,XM1, XM7, 303Se)


F594AF594BF594CF594DF594AF594BF594CF594D


70751008512011010085

B85 maxB6595,inclB95C32,inclC80C32,inclC24C36,inclC20C32,inclB95C30,inclB90C28,incl

2 (316)


F594EF594FF594GF594HF594EF594FF594GF594H

1/41 1/2,incl1/41 1/2,incl1/45/83/41, incl1 1/811/4, incl1 3/811/2, incl

70751008512011010085

B85 maxB6595,inclB95C32,inclB80C32,inclC24C36,inclC20C32,inclB95C30,inclB90C28,incl


3 (321, 347)


F594JF594KF594LF594MF594JF594KF594LF594M

1/41 1/2,incl1/41 1/2,incl1/45/8,incl3/41 1/2,incl1/45/8,incl3/41 incl1 1/811/4, incl1 3/811/2, incl

70751008512011010085

B85 maxB6595,inclB95C32,inclB80C32,inclC24C36,inclC20C32,inclB95C30,inclB90C28,incl

Ferritic alloys4 (430, 430F) A F594N 1/41 1/2,

incl 70 B6595,incl

Martensitic alloys

5 (410, 416, 416Se) HHT

F594PF594R

1/41 1/2,incl1/41 1/2,incl

100160

C2030,inclC3445,incl

Precipitation hardening alloys7 (630) AH F594U 1/41 1/2,

incl 135 C2838,incl

a Minimum values except where shown as maximum or as a range.b Legand of conditions:A Machined from annealed or solution annealed stock, thus retaining the properties ofthe original stock; or hot formed and solution annealed.AF Annealed after all threading is completed.AH Solution annealed and age hardened after forming.CW Annealed and cold worked. Sizes 0.75 in. and larger may be not worked.H Hardened and tempered at 1050°F (566°C) min.HT Hardened and tempered at 525°F (274°C) min.SH Machined from strain-hardened stock.Source: Ref. 15. Copyright ASTM; reprinted with permission.


Source: Ref. 15. Copyright ASTM; reprinted with permission.


TABLE 13 Corrosion of Alloys in Nitric Acid Solutionsa at 50°CCorrosion rate (mils/yr) at concentration (M)

Alloy 0.5 1 2 4 6 8 10MP35N 0.01 0.07 0.07 0.09 0.1Hastelloy C <0.01 0.03 0.08Incoloy 825 <0.02 <0.02 <0.02EN58Jb <0.02 <0.02 <0.02aTest time 2844 days.bEN58J similar to 316 stainless steel.Source: Data from Ref. 18.

Coatings are sometimes applied for lubrication, as some corrosion-resistance materials such as titanium, stainless steel, and nickel basedalloys will gall and seize during the tightening operation. Metals suchas graphite and petroleum jelly (MIL-T-5544), molybdenum disulfide(MoS2), and silver are used.

Part of the research effort to prevent stress corrosion cracking includeda study of singleand multiple-layer coatings [20]. Results are shown inFig. 2. The best of the coatings, nickel plus SermeTel W, was patented[21]. The SermeTel W is a proprietary coating but is covered bymilitary specification MIL-C-81751B [22].

Although these coatings do not match the performance of thecorrosion-resistant base metals that have been developed, they could beconsidered for lower strength applications, as SCC resistance isproportional to strength.

The selection of fastener materials and coatings is not the onlyconsideration when designing joints to resist corrosion. Not only mustthe materials be suitable to the environment, they must also becompatible with each other. One example of this possible problem is


the use of titanium fasteners with aluminum joints in wide-bodyaircraft. The superior corrosion resistance of the titanium causes rapidattack by galvanic and exfoliation (a form of crevice)TABLE 14 Corrosion of Alloys in Huey Testa (65% Nitric Acid, Boiling)

Corrosion rate (mils/yr)Alloy 48 hr 96 hr 114 hr 192 hr 240 hr Avg.MP35N

Annealed 13 21 36 52 63 37Cold rolled 14 23 39 50 85 42Cold rolled and aged 12 19 35 53 70 38As-welded 20 14 17 21 19 18Inconel 625, annealed 21 21 23 25 25 23Hastelloy C276 440 770 605

aData obtained in successive 48-hr periods on original specimens.Source: Data from Ref. 18. Copyright SPS Technologies.


Figure 2Effect of coating on stress corrosion cracking resistance of H-11 bolts.

(From Ref. 20.)

corrosion of the aluminum. Solutions have included the use of metallic


ceramic and aluminum coatings of the titanium fasteners.

Reviews of corrosion and stress corrosion are given in Refs. 23 and 24. Theycover the application of fasteners to joints and the precautionary measuresthat must be taken to prevent premature failures.

10Temperature-Resistant Materials

Structural materials are exposed to a wide range of temperatures both aboveand below the usual summer-winter exposure in buildings and bridges.Moderate temperatures are considered to be those between room temperatureand 900°F. Elevated temperatures range to 1600°F. Extreme elevatedtemperatures are those in the range 16003000°F. Moderately lowtemperatures range from room temperature to 50°F. Cryogenic (extremelylow) temperatures range from 50 to 423°F.

Fasteners have been developed to cover the range from 423 to 3000°F. Thetemperature ranges for various materials in terms of density versustemperature are shown in Fig. 3 [25].


Figure 3The useful temperature ranges of metallic structural materials.

(From Ref. 20.)

High temperature applications range from pressure vessels and ovens to heattreating furnacesand jet engines to space vehicles with both extremely low and extremely high temperaturerequirements. Moderately high and elevated temperature fastener materials are included inASTM specifications A193, A194, A437, and A453 [2629]. These include special steels andstainless steels. The presence of elevated to high elevated temperatures in aircraft jet enginesmandates the use of materials designed for that purpose. Materials such as A286 and Inconel718 were introduced as corrosionresistant materials in Section 7 but are also used in jetengines because of their good elevated temperature resistance. Other high nickel and cobaltbase alloys that are used in jet engines up to 1800°F include Waspaloy, Rene 41, and Astroloy.Waspaloy can be cold worked and aged to 260,000 psi, but its ductility is so low that it hasbeen rejected for stress corrosion applications.


Table 15 lists materials that are available as fasteners from aircraft fastener manufacturers inthe range from room temperature to 900°F. Table 16 covers fastener materials to 1600°F, andTable 17 lists fastener materials for use to 3000°F.

The refractory materials such as tantalum, columbium, and molybdenum alloys requireelaborate multilayer coatings to prevent failure by oxidation. Molybdenum alloys were used inonly a few applications because a pinhole break in the coating would cause the metal tovaporize.

Recent developments in jet engine alloys have included materials such as MP159, useful to1200°F, and a new alloy, Aerex 350 [30], which is useful to 1400°F. MP159 with room


TABLE 15 Ultimate Tensile Strength (Ftu) of Selected Alloysat Room to Moderately Elevated TemperaturesMaterial type Room temp. Ftu (ksi) Max. temp. (°F)Iron-base alloys

Vasco MA 300 900aH-11 260 900aMarage 300 260 900aPH 138 Mo 220 650Custom 455 220 650C. R. A-286 200 900B5F5 200 9008740 180 450a4340 180 450aPH 174 160 900155 PH 150 900

Nickel-base alloysMP35N 260 700C. R. Inco 718 220 900bInco 718 180 900b

Titanium-base alloysTi 1-8-5 200 300Ti 6-6-2 180 500Ti 64 160 500

aProtective plating is required.


bRolled threads after heat treatment.Source: Ref. 25

TABLE 16 Ultimate Tensile Strength (Ftu) of Selected Alloysat Elevated Temperatures (to 1600°F)Material type Room temp. Ftu (ksi) Max. temp. (°F)Iron base

A-286 200 1200A-286 140 1200

Nickel baseRene 95 230 1200Astroloy 190 1600Inco 718 180 1200Waspaloy 150 1400Rene 41 150 1400Source: Ref. 25.


TABLE 17 Ultimate Tensile Strength (Ftu) ofSelected Alloys at Extermely ElevatedTemperaturesMaterial type Ftu at max. temp.

(ksi)Max. temp.

(°F)Tantalum base

T-222 20 300090 Ta-10W 18 3000

Columbiumbase

Cb 752 35 2500C 1294 35 2500

Nickel baseTD-Ni 12 2000TD-Ni-Cr 18 2000

Cobalt baseHA 188 45 1800L 605 35 1800Source: Ref. 25.

temperature strength of 260,000 psi and Aerex350 with a room temperature strength of220,000 psi meet the need for high temperature


alloys that also have high room temperaturestrength.

A very extensive fastener material study wasdone for NASA in 1965 that was useful inselecting materials for Project Apollo and laterspace shuttles [31]. Materials in this study coverthe range from 423°F (the boiling point of liquidhydrogen) up to 1600°F. Other properties besidethe standard ones were evaluated, such asstrength/density ratio, tension impact, and stressrupture. The complete report is over 400 pageslong and contains extensive data on the effect oftemperature on both fastener materials andfasteners. Some examples of the data areincluded here to provide some sense of the typeof information that was developed.

Chart I shows notch properties of materials in therange of 423 to 1600°F.

Chart II shows stress relaxation of a René 41 boltand a silver-plated Waspaloy nut under different


preloads.

Table 18 shows tensile and angle block tensilestrength at temperatures 423, 320, and 70°F fora Ti6Al-4V bolt with an A-286 (160 ksi) nut.

As noted in Section 5, the selection of thefastener made from a particular material involvesa great deal more than the bolt and nut. Elevatedtemperature applications are even more complexthan those at room temperature and must takeinto consideration such things as lubrication,fastener material compatibility with jointmaterials, coefficient of thermal expansion,modulus of elasticity, and stress relaxation, notto mention the drop in yield and tensile strengthof the fastener material with a change intemperature.

11Summary

The selection of fastener materials starts with the


user's need to have a fastener made from thesame material as the joint or one compatible withit. The fastener material must have the capabilityto provide high tensile strength, notch strength,ductility, and yield strengthhigh enough to taketightening loads with minimum plasticdeformation.


Chart I(From Ref. 31.)


Chart II(From Ref. 32.)


TABLE 18 Mechanical PropertiesBolt NAS 127548; Ti6Al-4V (160 ksi)Nut FN 1216820; A-286 (160 ksi); 1/220Test no. Test temperature

(°F)Ultimate load

(lb) Ultimate stressa (psi)

A. Axial bolt properties1 423 32,500 203,0002 423 37,800 236,0003 423 32,400 202,6004 320 32,400 202,6005 320 30,000 187,6006 320 33,900 212,6007 70 29,500 184,4008 70 29,700 185,7009 70 29,600 185,100

B. Angle block bolt properties (3° Ð at nut bearing face)b10 423 27,000 157,20011 423 26,600 155,00012 423 28,010 163,50013 320 19,000 118,80014 320 21,060 131,70015 320 20,300 127,00016 70 25,800 161,40017 70 22,000 137,60018 70 21,000 131,300aStress calculated at the tensile stress area of 0.1599 in.2.bBolt grip length 3.0 in.Source: From Ref. 25.


Resistance to corrosion and stress corrosion is important,as is the capability to provide high fatigue resistance.Certain manufacturing techniques can provide thesecharacteristics. Temperature resistance to change inproperties over the widest possible range is valuable,although fasteners have been developed for specifictemperatures.

Finally, the fastener must be incorporated withoutdifficulty into the overall structural design of a product.The performance of fasteners from different materials maynot always be the same for the same material in each test.Part of this problem was recognized a number of years agowith the development of the MIL-STD-1312 series offastener test procedures. This is documented in Table 19along with tests for magnetic particle inspection, platingthickness, and metallography. Standardization of tests isnecessary in the presentation of any data on fastenermaterials and the fasteners made from them. (See alsoChapter 52.)


TABLE 19 Fasteners Test ProceduresHardness

MIL-STD-1312-6 ASTM-E3 ASTM-E18ASTM-A870 ASTM-E10 ASTM-E384SAE-J417

TensileMIL-STD-1312-8 ASTM-A470 SAE-J429MIL-STD-1312-18 ASTM-F606 SAE-J995MIL-STD-1312-23

ShearMIL-STD-1312-13 ASTM-F606MIL-STD-1312-20 ANSI/ASME-B18.8.2MIL-STD-1312-28

FatigueMIL-STD-1312-11 NAS 1069 (National Aerospace Standards)MIL-STD-1312-21

Stress durability (embrittlement testing)MIL-STD-1312-5 ASTM-F606MIL-STD-1312-14

VibrationMIL-STD-1312-7One of the tests in product specifications NAS 1675.

TorqueMIL-STD-1312-15 MIL-STD-1312-25 IFI-100/107MIL-STD-1312-24 MIL-STD-1312-31 IFI-125AS1310 (SAE) Fastener Torque for Threaded Applications,Definitions of.


Torque-tension/preloadMIL-STD-1312-15 SAE-J174 IFI-101MIL-STD-1312-16 ANSI-B18.16.2 IFI-124

CorrosionMIL-STD-1312-1 MIL-STD-753MIL-STD-1312-3 ASTM-B117MIL-STD-1312-19

Stress ruptureMIL-STD-1312-10

Stress relaxationMIL-STD-1312-17

Magnetic particle inspectionAMS 2640 (SAE) ASTM-E709 ASTM-1444MIL-STD-271 SAE-J420Federal Test Method Standard STD No. 1516.

Liquid penetrant inspectionAMS 2645 (SAE) MIL-STD-6866ASTM-E165 MIL-STD-271SAE-J426

Plating thicknessMIL-STD-1312-12 ASTM-B487 ASTM-B499ASTM-B504 ASTM-B530 ASTM-B568ASTM-B659 ASTM-B376Federal Test Method STD No. 151, Methods 520.1 and 523.

Table continues


ContinuedTABLE 19Metallography

ASTM-E3 ASTM-E1077 SAE-J123ASTM-E112 ASTM-F788 SAE-J418ASTM-E340 ASTM-F812 SAE-J419ASTM-E384 ASTM-J121 SAE-J423ASTM-E407 SAE-J122 SAE-J1061ASTM-E883

Source: Ref. 32.

References

1. SAE J429-83, Mechanical and MaterialRequirements for Externally Threaded Fasteners, SAEHandbook, Vol. 1, Society of Automotive Engineers,Warrendale, PA.

2. ASTM A 307-94, Standard Specification forCarbon Steel Bolts and Studs, 60,000 psi TensileStrength, ASTM Standards, Vol. 15. 08, AmericanSociety for Testing and Materials, Philadelphia.

3. ASTM A325-94, Standard Specification for


Structural Bolts, Steel, Heat Treated 120/105 ksiMinimum Tensile Strength, ASTM Standards, Vol. 15.08.

4. ASTM A 490-93, Standard Specification for HeatTreated Steel Structural Bolts, 150 ksi MinimumTensile Strength, ASTM Standards, Vol. 15. 08.

5. ASTM A354-95, Standard Specification forQuenched and Tempered Alloy Steel Bolts, Studs, andOther Externally Threaded Fasteners, ASTMStandards, Vol. 15. 08.

6. SAE J995-79, Mechanical and MaterialRequirements for Steel Nuts, SAE Handbook, Vol. 1.

7. ASTM A 563-94, Standard Specification forCarbon and Alloy Steel Nuts, ASTM Standards, Vol.15. 08.

8. SAE J 123c, Surface Discontinuities on Bolts,Screws, and Studs, SAE Handbook, Vol. 1.

9. Bickford, J. H., An Introduction to the Design andBehavior of Bolted Joints, 3rd ed., Marcel Dekker,New York, 1995.


10. Gill, F. L., Boron steels for mechanical fasteners,SAE Automotive Engineering Congress, Jan. 812,1968, Paper 680127.

11. Lin, C. S., J. J. Laurilliard, A. C. Hood., Stresscorrosion cracking of high strength bolting, in StressCorrosion Testing, ASTM STP425, ASTM,Philadelphia, 1967, p. 84.

12. ASTM F467-93, Nonferrous Nuts for GeneralUse, ASTM Standards, Vol. 15. 08.

13. ASTM F468-93, Standard Specification forNonferrous Bolts, Hex Cap Screws, and Studs forGeneral Use, ASTM Standards, Vol. 15. 08.

14. ASTM F593-95, Standard Specification forStainless Steel Bolts, Hex Cap Screws, and Studs,ASTM Standards, Vol. 15. 08.

15. ASTM F594-95, Standard Specification forStainless Stee Nuts, ASTM Standards, Vol. 15. 08.

16. Patel, S., and E. Taylor, New high strengthfastener materials resist corrosion, Metal Progress,September 1971.


17. Hood, A. C., MP35NA multiphase alloy for highstrength fasteners, Metal Progress, May 1968.

18. Taylor, E., Multiphase alloy environmentalresistance, SPS Laboratory Report 5817 (Revised),SPS Technologies, January 1, 1980.

19. Taylor, E., Stress-Corrosion Cracking Evaluationof Aerospace Bolting Alloys, Special TechnicalPublication 610, ASTM, 1976, p. 243.

20. Taylor, E., Stress-Corrosion Crack Protectionfrom Coatings on High Strength H-11 SteelAerospace Bolts, Special Technical Publication 518,ASTM, 1972, p. 131.

21. Hood, A. C., Protective coating for ferrousmetals, U.S. Patent No 3,897,222 (July 29, 1975).

22. Military Specification Coating, Metallic-Ceramic,MIL-C-81751B, Jan. 17, 1972.

23. Hood, A. C., Corrosion in threadedfastenersCauses and cures, Machine Design, Dec. 17,1964, p. 153.


24. Hood, A. C., Preventing stress-corrosioncracking in threaded fasteners, Metal Progress,September 1967.

25. Product Engineering Report, FastenerSeminar, Third Printing, SPS Technologies,1980.

26. ASTM A-193-95, Alloy Steel and StainlessSteel Bolting Materials for High TemperatureService, ASTM Standards, Vol. 01.01.

27. ASTM A194-95, Carbon and Alloy SteelNuts for High Pressure and High TemperatureService, ASTM Standards, Vol. 01.01.

28. ASTM A437-94, Alloy Steel Turbine-TypeBolting Material Specially Heat-Treated for HighTemperature Service, ASTM Standards, Vol.01.01.

29. ASTM A453-94, Bolting Materials, High


Temperature, 50 to 120 ksi Yield Strength withExpansion Coefficients Comparable to StainlessSteels, ASTM Standards, Vol. 01.01.

30. Buzolits, S. R., and L. A. Kline, Boltingalloy fills high temperature gap, AdvancedMaterials and Processes, February 1995.

31. Glackin, J. J., and E. F. Gowen, Jr.,Evaluation of Fasteners and Fastener Materialsfor Space Vehicles, NASA Contract No NASH-11125, Final Report, November 1963November1965, SPS Laboratories.

32. Borg, K., Specifications for fastener testing,Fastener Technology International, February1995, p. 44.


2Joining with Aluminum Alloy Bolts and Nuts

Russell E. Mack*Aluminum Company of America, Lititz,Pennsylvania

1Why Use Aluminum Bolts?

The characteristics of aluminum that cause it tobe chosen for a wide range of end productapplications are also those that provide the sameadvantages for bolts and nuts.

Resistance to atmospheric corrosion;nonstaining. Aluminum forms a protectiveoxide coating upon exposure to theatmosphere. The oxides of aluminum arecolorless, and such a corrosion product if it


occurs will not stain and mar the appearanceof the fastener or the assembled materials.

Weight/strength. Alluminum alloy fastenershave the highest strength-to-weight ratio ofany metals in common use for fasteners. Theirstrength is comparable to that of lowcarbonsteel fasteners, and aluminum fasteners weighabout one-third as much as steel and stainlesssteel fasteners.

Electrical conductivity. The high rate ofconductivity of aluminum makes it ideal forjoints intended to carry electricity. Dependingon the alloy choice, aluminum fasteners havea conductivity rating of 3843% InternationalAnnealed Copper Standard [at 68°F (20°C)]

Nonmagnetic. All aluminum alloy fastenersare nonmagnetic and nonsparking.

Compatibility/galvanic corrosion. The use ofaluminum fasteners in aluminum jointsminimizes the danger of galvanic corrosion.


Color identification/corrosion resistance.Inherent corrosion resistance can be increasedby anodizing the aluminum fastener forapplications that are considered to be highlycorrosive. The electrolytic anodic coating canbe dyed before it is sealed in a variety ofcolors suitable for identification or aestheticreasons. Anodic coatings and the process arecovered by Military Specification MIL-A-8625.

*Retired.


2Availability

Aluminum alloy bolts are domestically producedby a number of manufacturers in the types andsizes typically available in other materials.Several master distributors make it possible toobtain a wide range of popular sizes throughoutthe U.S. fastener distribution system.

While size range capability varies amongmanufacturers, the design engineer can expect toobtain bolts made to appropriate AmericanNational Standard specifications in types andsizes to meet any requirement.

Hex head bolts are available either as headed hexhead bolts (sometimes termed indented hex headbolts) or finished hex head bolts (sometimestermed cap screws). A comment about the headedhex bolt is appropriate as there exist


misconceptions about it. Since the head is forgedcomplete in the initial headed blank (i.e., notsheared like a cap screw), it has slightly roundededges on the top and underside of the hex as wellas on the corners. This in fact offers certainadvantages; the head will not mar the bearingsurface, and engagement of a wrench or socket isfacilitated.

2.1Size Ranges

Diameters 7/89, 18, and larger are produced byseveral specialty fastener manufacturers. Finepitch threads are also manufactured but are notreadily available and may entail substantialproduction runs. Metric equivalents are alsoproduced on special runs by a limited number ofmanufacturers. Lengths are typically 1/4 in.increments, and lengths in the ranges shown inTable 1 are readily available. Longer lengths arealso produced on special production runs.


The applicable standard for hex head boltsproduced in the United States is ANSI B 18.2.1,which is included in the Fastener StandardsHandbook published by Industrial FastenerInstitute, Cleveland, Ohio.

Hex head bolts for structural bolting ofaluminum transmission towers covered byASTM Specification F 901 are produced inappropriate aluminum alloys by selectedmanufacturers on special manufacturing runs.They are head marked and color anodized for sizeidentification.

The applicable standard for carriage boltsproduced in the United States is ANSI B 18.5,which is also included in the Fastener StandardsHandbook. Sizes tabulated in Table 2 reflectthose that are most readily available; lengths aretypically 1/4 in. increments. Other sizes, finepitch threads, and short neck carriage bolts areproduced by some manufacturers on specialproduction runs for which mill minimums are


required.

The bolts discussed above, hex and carriage, arethe ones most commonly used in applicationswhere the characteristics of aluminum are neededand appropriate. However, aluminum alloy boltscan be and are produced in virtually all headtypes including heavy hex head, hex flange, hexsocket, square head, and 12-point.TABLE 1 Common Sizes ofHex Head BoltsSize Length (in.)1/420 1/22 1/25/618 3/433/816 3/43 1/21/213 145/811 1 1/443/410 1 3/44


TABLE 2 Common Sizes ofRound Head, SquareNeck Carriage BoltsSize Length (in.)1024 1/221/420 1/22 1/25/618 3/433/816 3/441/213 14

2.2Nuts

Aluminum alloy hexagon nuts for all the boltsizes listed are readily available through the U.S.fastener distribution network. They aremanufactured in accordance with ANSI B 18.2.2.Generally they are cold forged and are chamferedand countersunk on both sides to facilitateassembly. Aluminum alloy nuts should always belubricated.


3Alloy Selection

The choice of aluminum alloys for male andfemale fasteners is relatively limited. Unless theproper alloy is used and is correctly heat treatedand aged, corrosion and strength problems can becatastrophic. Heat-treating procedures shouldadhere to the practices and controls of MIL-H-6088. Bolts that are roll threaded after heattreating and aging are found to have slightlygreater tensile strength than bolts that have cutthreads or are roll threaded before heat treating.

Bolts are most commonly produced in alloy2024-T4, which has strength, toughness, andcorrosion resistance suitable for mostapplications. Alloy 2024-T4 bolts are frequentlyanodized to enhance their inherent resistance toatmospheric corrosion.

Where the designer finds a need for slightlyhigher strength, bolts can be manufactured in


alloys 7075-T73 or 7050-T73, the latter beingvery special, probably difficult to obtain, andvery costly.

For limited applications where strength can besacrificed to gain further resistance to generalcorrosion and to some chemical corrosion, alloy6061-T6 can be specified, but it also is not asreadily available as 2024-T4 and still entailspecial production runs.

To develop the full strength of aluminum alloybolts and to avoid stress corrosion cracking ofthe aluminum nut, the alloy for nuts must beeither 6061-T6 or 6262-T9. Cold-forged nutsare manufactured in alloy 6061-T6. Nutsmachined from bar can be either 6061-T6 or6262-T9. It can be noted that 6262-T9 nuts arenot readily available and since they aremanufactured by machining will be more costlythan 6061-T6 forged nuts.

4


Mechanical Properties

The minimum mechanical properties and rawmaterial specifications for the aluminum alloysdiscussed here are given in Table 3. Chemicalcomposition limits and minimum strength valuesare covered by ASTM specifications F 467 fornuts and F 468 for bolts. Typical tensile andshear strengths are listed in Table 4.

The tensile strength of an aluminum bolt can beapproximated by multiplying the tensile strengthof the stock material by the cross-sectional areaof the bolt at the root of the threads. Dataconfirming this are shown in Table 5.


TABLE 3 Minimum Mechanical PropertiesAlloyandtemper

Tensilestrength

(psi)

Tensile yieldstrength (psi)

Elongation in 2in. or 4D (%)

Shearstrength

(psi)

ASTM materialdesignation

2024-T4 62,000 40,000 10 37,000 B 316

6061-T6 42,000 35,000 10 25,000 B 316

6062-T9 52,000 48,000 5 30,000 B 211

7050-T73 70,000 58,000 10 39,000 B 316

7075-T73 68,000 56,000 10 41,000 B 316

Axial tension and 10° wedge tension testing (per ASTMF 606) gives emphasis to the suitability of 2024-T4 and7075-T73 for bolts from the standpoints of strength andtoughness. Results are shown in Table 6.

The load-carry capacity in shear is calculated bymultiplying the shear strength of the bolt alloy by thecross-sectional area of the bolt in shear. If the threads arein the shear plane, the area at the root of the threadshould be based on the minimum root diameter.Calculated minimum tensile and single-shear loads are


given in Table 7.

5Installation Procedures

Although installation methods for aluminum alloy boltsare similar to those used for bolts in other materials, itshould nonetheless be recognized that, if overtightened,aluminum bolts are not as forgiving as are bolts in othermaterials. Consequently, more careful attention to theprocedure used is recommended. Also, as with fastenersof all materials, a quantitative measure of tightness isdifficult to obtain.

The torsion shear strength of the bolt thread area can beused as a guide in developing installation torque. Thetorsion shear strength can be calculated or found bypragmatically determining free breaking torque, thelatter by straining the bolt shank and applying acalibrated torque wrench to the head of the bolt. Table 8presents data for both methods for three common alloys.

Whether employing a predetermined torque value or ''theturn of the nut" from a specified snug position, it shouldbe noted that clamping load can vary widely due tovariables including thread fit, bearing surface condition,


and lubrication. While this is true with all bolt materials,aluminum bolts are particularly sensitive to thesevariables, especially the presenceTABLE 4 Typical Mechanical PropertiesAlloy andtemper

Tensile strength(ksi)

Shear strength(ksi)

2024-T4 68 416061-T6 45 306062-T9 58 357050-T73 75 457075-T73 73 44Mild steel 60 45


TABLE 5 Tensile Strength of 2024-T4 Bolts and Stock MaterialBoltsize

Lotstested

Totaltest

Tensile strength ofnominal root area

(psi)

Average tensilestrength of stocka

(psi)

Ratio of strength of rootarea to strength of stock

1/420 209 638 73,800 68,000 1.075/1620 78 234 70,100 69,100 1.023/816 54 161 69,900 69,100 1.011/213 68 205 69,000 68,900 1.005/811 34 101 67,300 69,200 0.973/410 5 16 67,000 68,000 0.98aTensile strength of raw material used to manufacture bolts heat treated with bolts.

or absence of lubrication. Consequently, experience hasshown that installation torque is more appropriatelydetermined by trial under actual installation conditions. Anumber of bolts can be installed and tightened until failureoccurs, with a calibrated torque wrench used to record thefailure torques. Subsequent job installation torque with asatisfactory safety factor would be 80% of the lowestrecorded failure torque.

Since job site installation may be difficult using a calibratedtorque wrench, a turn-of-a-nut method may be employedafter the number of turns has been correlated with either acalculated or empirical tightening torque as described above.

Proper lubrication in an aluminum bolted joint cannot beemphasized enough. Figure 1 graphically shows the salutary


effect of lubricant on clamping load relative to bolt tensilebreaking load. If tightening torque is applied to the nut, thecommon practice is to use a bulk lubricated nut so that thethreads and the bearing face are lubricated. Mostcommercially available nuts are waxed by the manufacturer,but the user should be certain this has been done. If the boltis to be tightened, then arrangements will have to be made tohave the commercially available bare aluminum boltlubricated, which again is best done in a bulk method to besure the threads and head bearing surfaces are lubricated.

If aluminum materials are being joined, it is appropriate touse aluminum fasteners. The alloys for aluminum bolts andother fasteners are usually as strong as or even stronger thanTABLE 6 Tensile Strength of Bolts in Axial and 10° Wedge Testa

Tensile breaking load(Ib)

Boltsize

Alloy andtemper

Axialtest

10° wedgetest

Strength reduction because ofwedge (%)

3/816 2024-T4 5,100 4,900 41/213 2024-T4 9,000 8,900 15/811 2024-T4 14,200 13,900 23/410 2024-T4 20,000 19,000 13/816 7075-T73 6,000 5,600 71/213 7075-T73 11,000 10,300 65/811 7075-T73 17,000 16,900 13/410 7075-T73 25,600 25,200 2aBolts tested had a minimum of 3/8 in. unthreaded shank.


TABLE 7 Minimum Tensile and Single-Shear Loads for 2024-T4 and 7075-T73 BoltsSingle-shear strength (Ib)

Tensile strength (Ib) Threads in shearplane Shank in shear plane

Normal size (in.) Basic major diameter(in.) Threads per inch 2024-T4 7075-T73 2024-T4 7075-T73 2024-T4 7075-T73

10 0.190 24 926 1,020 553 598 978 1,0601/4 0.250 20 1,710 1,880 1,020 1,020 1,720 1,8605/16 0.3125 18 2,880 3,160 1,720 1,860 2,710 2,9303/8 0.375 16 4,300 4,710 2,560 2,770 3,930 4,2401/2 0.500 13 7,950 8,720 4,750 5,130 7,060 7,6405/8 0.625 11 12,800 14,000 7,620 8,230 11,060 12,0003/4 0.750 10 19,100 20,900 11,400 12,300 16,000 17,3007/8 0.875 9 26,400 29,000 15,800 17,100 21,800 23,6001 1.000 8 34,700 38,000 20,700 22,400 28,500 30,800


TABLE 8 Breaking Torque for BoltsBreaking torque

(Ib.in.)Alloy andtemper

Shear strengtha(psi)

Screw or boltsize By testb Calc.c

6061-T6 29,500 3/816 193 1885/811 966 973

2024-T4 39,000 3/816 246 2505/811 1,148 1,280

7075-T73 48,100 3/816 316 3085/811 1,598 1,580

aDetermined by standard double-shear test of stock.bTest made in torsion machine.cTorque, T = pfsd3/13, where fs = shear strength of bolt material in psiand d = minor diameter of bolt thread in inches.


Figure 1Effect of lubrication on clamping load of 1/2-13 2024-T4 bolts. A, the effect of a

petroleum-based lubricant; B, the effect of a wax-based lubricant; C, the torque-loadrelationship obtained with a bare, unlubricated bolt.


the aluminum structures being assembled. Whileseveral cautionary caveats have been mentioned,by no means should the designer be reluctant tospecify aluminum bolts; they are easy to use. Ihave been acquainted with or aware of thetrouble-free use of literally hundreds of millionsof aluminum bolts and other fasteners.

Bibliography

Aluminum, Vol. 3, Fabricating and Finishing,American Society of Metals, Metals Park, OH.

Design considerations for aluminum fasteners,SAE Paper 80455, Society of AutomotiveEngineers, Warrendale, PA, 1980.

Fastener Standards, Industrial Fastener Institute,Cleveland, OH.


3Thread Lubricants

Gary J. Novak and Tom PatelFel-Pro Chemical Products, Skokie, Illinois

1Introduction

Lubricants are compounds that are applied tothreaded and other contacting surfaces onfasteners during the assembly of a bolted joint.Thread lubrication has a significant effect on thethreaded fastener during the assembly, in-service,and disassembly phases [1]. Thread lubricantsfacilitate the assembly of bolted joints byproviding smooth assembly and reducing torque-tension scatter. Fastener tension level, fastenertension variation around the joint, and successful


assembly without fastener failure are influencedby the thread lubricant. Without proper assemblyof the bolted joints, powerplants, papermills,refineries, chemical processing plants, internalcombustion engines, and numerous othermachines and process equipment systems wouldnot function properly. The results of improperfastener assembly can include fastener overloadleading to fastener failure, loose fasteners,nonuniform flange loads, leaking joints, andcrushed and overloaded gaskets. These types offailure lead to high plant maintenance cost andtotal system shutdown. Thread lubricants canhelp reduce the potential damage arising fromimproper assembly, which could range fromcustomer inconvenience through loss ofproduction, financial and bodily harm, hazardousemissions, and leakage of hazardous materials toloss of life. When removed or taken apart,fasteners that were assembled with threadlubricants are disassembled more readily. Someof the lubricant or a residue from the lubricant


remains on the threads and provides a layer thatlubricates and prevents galling duringdisassembly. During repairs, service, ordisassembly, cost, time, and damage to theassembled parts including fasteners and threadedcomponents are reduced through the use oflubricants.

Thread lubricants provide control of the threadfriction level and prevent metallurgical bondingduring the assembly of a fastener. Lubricantsaffect both the nominal level and the variabilityof the resultant tension that is developed in afastener when it is assembled to a target torquelevel. A lubricant that is properly selected willprevent galling, fretting, and seizing and reducefriction and wear of the fastener compared to adry-assembled fastener. Thread lubricants canalso provide antiseize function duringdisassembly or removal of the fastener. Thelubricant can protect the fastener from moistureor chemical exposure. Corrosion can be reduced


and in some cases prevented during storage andshipping and in service through the use oflubricants. Thread lubricants have a significanteffect on the torque-tension relationship of afastener.


2How Do Thread Lubricants Work?

When a fastener is under high load, the highcontact points on the mating surfaces of thethreads deform elastically until the contact areais sufficient to support the load. If the surfacesare clean (unlubricated) or if the lubricant breaksdown, the surfaces will gall. The threads mayweld together at several sites due to the loadand/or elevated temperature, and significantportions of the thread may be torn off as thefastener is assembled or disassembled. Toprevent this galling and potential thread failure,the two surfaces must be kept separated. A goodthread lubricant fills the spaces between contactpoints, thereby reducing the metal-to-metalcontact and preventing galling and seizing. Ineffect, it acts like a microscopic ball bearing tokeep the surfaces separated, ensuring quick and


easy disassembly and reducing equipmentdowntime.

3Selection of a Thread Lubricant

Thread lubricants are primarily selected tofacilitate the assembly and disassembly offasteners. However, the lubricant must becompatible with the environmental andfunctional requirements of the bolted assembly.Several important factors must be consideredwhen specifying a thread lubricant. The nominalvalue of the nut factor, K, is an importantparameter that is used to evaluate theeffectiveness of a thread lube as an assembly aid.The nut factor is related to the coefficient offriction on the surfaces lubricated with thesubject thread lubricant. Table 1 shows sometypical nut factors for a variety of boltingmaterials and lubricants. Low nut factor valueswill result in a greater bolt tension for a given


level of torque. In addition to nut factor, the typeof application and environment are alsoimportant issues when selecting a threadlubricant. Table 2 provides a thread lubricantselection guide for various applications andenvironments. Some of the issues that should beconsidered when specifying a thread lubricant areenvironment, fluctuating loads, storage andhandling, initial assembly or service application,and the assembly procedure.

The prevention of thread galling and theachievement of target thread friction or nutfactor are also important issues to be considered.The specification of a low friction lubricant mayresult in a nut factor that is too low andovertensioning of the fastener. The specificationof a lubricant that yields a high nut factor willresult in bolt tensions that are relatively lowwhen the fastener is assembled to a specifiedtorque. The selection of a thread lubricant for anexisting application should include an evaluation


of the assembly process and torque specificationsto ensure that final preloads in the assembledfasteners are within the target tension range. Thenut factor values in Table 1 include theminimum, mean, and maximum because there isa significant variability in the resultant nut factorthat is measured for each set of conditions. Themean value of the nut factor ranges from 0.086to 0.52. An inspection of the nut factorminimums and maximums in Table 1 shows thatsome lubricants are more effective than others inreducing the variability in nut factor.

The specification of a thread lubricant and thecorresponding nut factor will have a significanteffect on the state of stress in the fastener duringassembly to (or disassembly from) a desired levelof tension. The state of stress includes thefastener tensile stresses, a shear stress thatdevelops due to the applied torque duringassembly, and other stresses due to any externalforces and moments applied to the bolted joint.


The tensile and torsional shear stresses will bedirectly affected by the selection of lubricant.The amount of torsional stress retained in atightened fastener depends on the fastener under-head friction and the presence of any lockingdevices. The selection or lack of thread lubricantcan control bolt twist failures during assemblyand disassembly as well as the state of residualstress in the fastener after assembly. The residualstate of stress will affect other modes of failuresuch as direct tension, fatigue, or stresscorrosion cracking.

Thread lubricant not only affects the tension ofan individual fastener; it may also affectclamping load distribution across a bolted jointthat has many fasteners. In their study, Chaaban


TABLE 1 Nut Factors K for Various Fastener Materials and CoatingsaReported nut

factorFastener material and coatingb Min MeanMaxPure aluminum coating on AISI 8740 alloy steel [2] 0.42 0.52 0.62Electroplated aluminum on AISI 8740 alloy steel [2] 0.52 Mild or alloy steel on steel, as received 0.0580.2 0.267Stainless steel on mild or alloy steel, as received 0.3 A490, 1 in. diam. [3]

As received 0.179 Very rustyc 0.389 With Johnson 140 stick wax 0.275

Black oxidated 7/8 A325 and A490, slightly rustyd [4] 0.15 0.22Black oxide 0.1090.1790.279Cadmium plate (dry) 0.1060.2 0.328Vacuum cadmium + chromate [2] 0.21 Copper-based antiseize 0.08 0.1320.23Cadmium plate (waxed) 0.17 0.1870.198Cadmium-plated A286 nuts and bolts [5] 0.15 0.23Cadmium plate plus cetyl alcohol on A286 nuts and bolts[5] 0.11 0.16

Cadmium-plated nuts used with MP35N bolts [5] 0.18 0.29Dag (graphite + binder)e [6] 0.16 0.075Dicronite (tungsten carbide in lamellar form)e,f 0.045 0.075Emralon (PTEE + resin)e [6] 0.10 0.15Everlube 810 (MoS2/graphite in silicon binder)e [7] 0.09 0.115Everlube 811 (MoS2/graphite in silicate binder)e [7] 0.09 0.115Everlube 6108 (PTFE in phenolic binder)e [7] 0.105 0.13Everlube 6109 (PTFE in epoxy binder)e [7] 0.115 0.14


Everlube 6122 0.0690.0860.103Fel-pro C54 0.08 0.1320.23Fel-pro C-670 0.08 0.0950.15Fel-pro N 5000 (paste) 0.13 0.15 0.27Mechanically galvanized A325 bolts [4]

As received 0.35 0.49Clean and dry 0.46 Slightly rustyd 0.36 0.39

Mechanically galvanized A325 bolts; lubed with 1 partwater, 1 part Jon Cote 639 was [4] 0.11 0.26

Hot-dip galvanized 7/8 A325 [4]As received 0.14 0.31Slightly rustyd 0.09 0.17Clean and dry 0.09 0.37

Hot-dip galvanized 7/8 A325 lubed with 1 part water, 1part Jon Cote 639 was [4] 0.10 0.16

Gold on stainless steel or beryllium copper [8] 0.4 Graphite coatings 0.09 0.28Lube Lok 1000 (ceramic-bonded coating)e,f 0.275 0.31Lube Lok 2006 (silicone resin bonded MoS2 +graphite)e,f 0.075 0.25

Table continues


ContinuedTABLE 1

Reported nutfactor

Fastener material and coatingb Min MeanMaxLube Lok 4253 (high silver and indium content coating)e,f0.21 0.24Lube Lok 4856 (contains powdered tin + lead)e,f 0.21 0.25Lube Lok 5306e,f 0.04 0.18Machine oil 0.10 0.21 0.225Microseal 100-1 (graphite plus inorganic binder)e [7] 0.09 0.115Molydag (MoS2 + binder)e [6] 0.16 0.40Moly paste or grease 0.10 0.13 0.18Neolube 0.14 0.18 0.20Never-Seize (paste) 0.11 0.17 0.21N5000, threads only, not nut [9] 0.0970.1060.117Pepcoat 6122 on nut, threads, and washer [9] 0.0800.0850.089Pepcoatf [10] 0.09 0.11Phos-Oil 0.15 0.19 0.23PTFE plus bindere [6] 0.10 0.15SermaGard (aluminum particles in ceramic binder)f

#902 or 846; basecoat only 0.50 Basecoat + SermaLube 1000 0.09 0.15 0.40Basecoat + 751 wax 0.18 0.23 0.90Misc. Sermagard systems with modified silicone topcoat 0.17

Serme Tel-W (phosphate-bonded, aluminum-rich coating)[2] 0.30

Solid film PTFE 0.09 0.12 0.16Zinc plate (waxed) 0.0710.2880.52


Zinc plate (waxed) 0.0710.2880.52Zinc plate (dry) 0.0750.2950.53aSee Chapter 18 for list of coating suppliers and their addresses.bAll tests involve steel bolts in steel joints, unless otherwise noted.cExposed outdoors for 2 weeks.dDipped in water, air-dried 24 hr twice.eUse with extra caution. These nut factors were computed from publishedcoefficients of friction (µ) data using the equation K = µ(rt/D cos ß) + rn/D +P/2Dp. See Eq. (2) for terms and units. Threads ranging from 1/4;20 to 24-1/2; were assumed in the calculations.fFrom manufacturer's literature.

et al. [11] conclude that for some situations theselection of a thread lubricant generally leads to moreuniform bolt tensions. Also, this study points out that ajoint assembled with molybdenum disulfide basedthread lubricants demonstrated the most uniform gasketstress distributions of the four lubricants of their study,the other three being a machine oil, a nickel-basedantiseize compound, and a copper-based antiseizecompound.

In-service operating conditions may expose the fastenerand lubricant to elevated temperatures. The liquidcontents of the lubricant may evaporate, but the solidmaterials will remain on the thread contact areas toprovide release and antiseize function duringdisassembly. Stresses and deformation of the fastener


and related components may arise from these thermalloads. Differential thermal expansion between the bolt,flanges, and attached structures of the


joint can increase or decrease fastener preload,deformation, and stress. These loads should beconsidered in the analysis of fastener tension andpreload. Oxidation of the fasteners can alsooccur and will generally accelerate at elevatedtemperatures. The lubricant should be chemicallycompatible with the fastener materials as well asany materials that may contact the lubricantduring storage, handling, assembly, or in service.Corrosion can be reduced by using threadlubricants that prevent damage to the surfaceoxide or coating layers of the fastener duringassembly and provide a protective coating.

Stress corrosion cracking (SCC) and hydrogenembrittlement can be reduced by using threadlubricants. Hydrogen embrittlement can bereduced using a lubricant such as molybdenumdisulfide instead of electroplating. Bothhydrogen embrittlement and stress corrosion


cracking are less likely to occur when stresses atsites of potential failure are reduced through theuse of a thread lubricant.

The chemistry of the lubricants should becompatible with the conditions of assembly andservice. Molybdenum disulfide paste should beconsidered carefully when fasteners are operatingat temperatures above 500°F. In the presence ofhumidity, the lubricant can release sulfur to formsulfuric acid, which can lead to corrosion orstress corrosion cracking failures. The copper-based antiseize compounds should also bespecified with care. Copper-based lubricantsshould not be used in systems where they will beexposed to ammonia, where strong chemicalreactions may can take place. Copper should notbe used in the processing of edible oils because itmay accelerate spoilage. The copper in somelubricants will damage and may deactivateplatinum catalysts used in some refiningoperations. These are a few of the many


compatibility issues that should be consideredprior to the selection of a thread lubricant. A lessreactive nickel flake/graphite flake lubricant isavailable for applications where hightemperatures or chemical compatibility areissues of concern. The nickel and graphite flakelubricants are significantly more expensive thancopper or molybdenum disulfide lubricants.

Thread lubricants are used in regulated industriessuch as food processing, nuclear powergeneration, and military products. Lubricants thatmeet the specifications for performance andcomposition for these industries are available.Purity, particle size, and other quality factors aremonitored and certified to conform to variousspecifications including military (MIL)specifications, the federal Meat and PoultryProducts Inspection Program of the USDA, andnuclear equipment manufacturers' standards.Nuclear grade lubricants are supplied incontainers that identify the batch as well as the


compositional analysis. The analysis includeslevels of relevant impurities measured in partsper million. In the nuclear industry [1], theimportant elements to monitor include thehalogens fluorine and chlorine, sulfur, andcertain heavy metals such as lead, tin, zinc,cadmium, and mercury.

4Thread Lubricants

4.1Types of Thread Lubricants

Thread lubricants are essentially classified intothree categories: liquid lubricants (oils), pastelubricants (combination of oil and/or grease andsolid lubricants), and dry film lubricants(resinbonded solid lubricants).

Oil-based liquid lubricants include petroleum-based oils, for example, mineral oil or syntheticoils, that include silicone, esters, olefins, glycols,


and polybutenes. These lubricants have limitedapplication at moderately high temperaturesbecause oxidation occurs at temperatures around350°F (higher oxidation temperature for siliconefluids) and the thread surfaces are left in a stateof dry metal-to-metal contact. After exposure tothese temperatures, threads may seize; moisture,chemicals, or gases may get between threads; andcorrosion might occur. Oil-based liquidlubricants may squeeze out under extremepressure, creating sufficient metal-to-metalcontact to cause cold welding.


Paste lubricants are formulated with solid lubricantsdispersed in liquid oil and/or grease. They arecommercially known as antiseize lubricants. Most of thethread lubricants used for maintenance, repair, andoperation (MRO) applications fall into this category. Thesolid lubricants used in these products are metal powdersor flakes such as copper, nickel, aluminum, lead, or zinc.Nonmetallic lubricants include graphite, molybdenumdisulfide, boron nitride, polytetrafluoroethylene, mica,talc, and metal oxides, hydroxides, and fluorides, such ascalcium oxide, calcium fluoride, zinc oxide, titaniumdioxide, magnesium oxide, calcium hydroxide, bariumoxide, and tin oxide. Oil-based lubricants are based onmineral oil or synthetic oils from the silicone, ester,olefin, glycol, and polybutene families. Greases aremineral oil thickened with sodium, aluminum, calcium, orlithium, soaps or oil thickened with chemically treatedbentonite clays, silicas, or polymer-thickened oils such aspolyurea greases.

Paste lubricants are effective even when the oil boils off atelevated temperatures above 350°F. They leave a solidlubricating film on the threads, which allows fasteners tobe removed without seizing or galling. Under heavy loads,


they remain between metal surfaces and will not squeezeout like oil on the threads.

Dry film (or bonded) lubricants consist of dry lubricatingsolids dispersed in binders. The most commonly usedlubricating solids are molybdenum disulfide, graphite,tungsten disulfide, polytetrafluoroethylene, boron nitride,calcium fluoride, and metallic oxides such as antimonyoxide, calcium oxide, and magnesium oxide. Commonbinders are epoxies, silicones, silicates, phenolics,polyamides, polyimides, phosphates, and ceramics. Dryfilm lubricants can be cured at room temperature or atelevated temperatures depending on the formulations.These products perform as well as paste lubricants, butmany carry hazardous solvents. Once the solventevaporates, these products become a film that is dry to thetouch, clean, and easy to handle. Dry film lubricants areused in original equipment manufacturing becausefasteners can be coated in bulk before assembly.

4.2Application of Thread Lubricants

Thread lubricants are applied to clean, dry surfaces. Thelubricant may be applied with a brush, spray, dip, or otherprocess depending on the specific application. Thethreaded parts are then assembled. In some cases, spray-


applied lubricants are air-dried or oven-baked prior toassembly to provide a dry lubricating film. Threadlubricants are sometimes applied to the underside of thebolt head to reduce friction.

5Nut Factors and the TorqueTension Relationship

The applied torque and resulting bolt tension of a fastenerare most often modeled as a linear function for thepurpose of design. In most cases, the linear model is a verygood approximation of the actual behavior. The equationdetermines bolt tension from applied torque, fastenerdiameter, and the nut factor as follows:

where.

K=the nut factor (dimensionless)

D=the nominal diameter (in., m)

F=

the bolt tension force (in.-lb, N· m)

T=the torque (in.-lb, N · m)


Nut factors are determined experimentally. Assemblytorque and the corresponding bolt tension are measuredduring the assembly of a fastener. The torque-tensionrelationship is then established. Chaaban et al. [11] used astrain gaged bolt to monitor tension and a torque


wrench to apply bolt torque to prescribed values. In this study of six conditions thatincluded four different thread lubricants, the six torquetension curves were linear over thereported range of data. The slope, c, of the torquetension curve is used to determine thenut factor K. The nut factor K is calculated from the torquetension slope as

The bolt geometry and the coefficients of friction of the bolt thread and bolt head contactareas have a major effect on the nut factor. The performance of a lubricant is a function ofnumerous factors including thread geometry, material, hardness, heat treatment,temperature, finish, alignment, and cleanliness. These factors vary from application toapplication. Reported nut factor values should not be considered as absolute values butshould be used as reference values in selecting a lubricant. A torquetension study using astrain gaged bolt or ultrasonic tension measurements should be performed with actualfasteners under actual conditions to determine more accurate nut factor values forlubricants and fastener materials not specifically called out in the tables. The reported nutfactors in Table 1 should be used only as a guide.

5.1Coefficient of Friction

The assembly torque of a fastener is consumed by two mechanisms: friction-free boltstretching and overcoming friction. Useful work is done when the contacting surfacesslide relative to each other up the inclined planes formed by the thread surfaces. A portionof the applied torque on the bolt head creates normal forces on the threads to stretch thebolt. The balance of the bolt torque overcomes thread friction and under-head friction. Theassumed linear torquetension relation is a simplification of the actual behavior. Morecomplicated models reflect the effects of thread friction, bolt head friction, and the helixand thread flank angles. The following equations [12, pp. 758762] characterize theexternal bolt torque, T, required to tighten or loosen a fastener:

where

F


F=the bolt tension (lb, N)

Tt=the torque to tighten (in.-lb, N · m)

Tl=the torque to loosen (in.-lb, N · m)

q=the thread flank angle (deg)

µ=the helix angle (deg)

fn=

the friction between the bolt head and theflange (dimensionless)

dn=

the mean diameter of the head bearing face(in., m)

ft=

the bolt thread coefficient of friction(dimensionless)

dt=the mean thread contact diameter (in., m)

These equations can be rewritten in the form


A mathematical expression for the nut factor can then be written in the form

For small lead angle µ and low thread friction, the nut factor can be expressed as [13, p.224]

The torquetension equations show that the input torque is divided into three components.First, the applied torque must overcome friction between the bolt head and the flangeacting at diameter dn with a coefficient of friction fn. Second, the applied torque must alsoovercome the thread friction forces acting at mean thread diameter dt, with thread frictioncoefficient ft. The balance of the applied torque does the useful work of stretching the bolt.

The efficiency of the screw, e, is defined [12, pp. 817822] as the ratio of the torquerequired to develop a bolt load without friction to the torque required to develop the samebolt load with friction:

The efficiency indicates how much torque is being used to stretch the bolt and how muchis used to overcome friction. When the coefficients of friction at the bolt threads, ft, andunder the bolt head, fn, are zero, the efficiency is 100%, and all of the torque is used tostretch the bolt. When ft or fn is nonzero, the screw efficiency is less than 100%. The use ofthread lubricants with low coefficients of friction reduces the nut factor value and alsoincreases the bolting efficiency e. An analysis of the equation describing the torque toloosen, T1, shows that a low coefficient of friction at the bolt threads and under the bolthead will reduce the torque required to loosen. Fasteners that are assembled with lowfriction lubricants may be more liable to loosen than those with higher friction lubricants,because the loosening moments have less frictional torque to overcome.

6Combined Tension and Torsion Stress in the Fastener

The tightening of a fastener results in a state of stress that includes the uniaxial tensile


The tightening of a fastener results in a state of stress that includes the uniaxial tensilestresses and shear stresses due to twisting. The combination of these stresses along withthe stresses induced from external service loads should be used for failure analysis anddesign. In the case of a design evaluation against the maximum shear stress theory, thetensile stress and the shear stress are used to calculate the maximum shear stress due tothis biaxial loading [12, p. 769].


where

tmax=the maximum Shear stress (psi, MPa)

s =the shear stress due to pure tension (psi,MPa)

t =the shear stress due to pure torsion (psi,MPa)

A =the tensile area (in.2, m2)dr =the thread root diameter (in., m)dp =the thread pitch diameter (in., m)F =the bolt tension (lb, m)Q =the twisting torque (in.-lb, N ·m)

The twisting torque Q is the portion of the assembly torque T that is transmitted throughthe bolt shank. The lost torque overcomes the friction between the bolt head and theflange, and the transmitted torque overcomes thread friction and does useful work instretching and tensioning the bolt. The twisting torque is less than the total applied bolttorque:

where h varies from 0 to 1 depending on the values of fn and thread geometry.

When the torquetension relationship is included, the resulting expression shows the effectthat thread lubricant selection has on the maximum shear stress within the threadedfastener:

In most situations, the coefficient of friction between the bolt head and the flange is aboutthe same as the bolt thread friction. For typical lubricated and metal friction coefficientsand thread geometries, it can be shown that approximately 50% of the total applied torqueT is transmitted as twisting torque Q, and h is approximately 0.5.

When assembled to target levels of tension F, low nut factor lubricants will result inlower maximum shear stress values. Fasteners with lower maximum shear stress willfunction with a higher factor of safety when elevated against the maximum shear theory:


where Sy = the stress (psi, MPa) and N = the factor of safety (dimensionless).

7Fluctuating Loads

Gasket leakage, fastener overload, fastener loosening, and fastener fatigue failures canoccur if the proper preload is not developed at initial assembly. The specification of threadlubricants should be consistent with the assembly and torque specification to yield apreload that is in the ''proper" range to avoid these failures.

When external loads are applied to a joint, the fastener tensions and gasket loads willchange. Enough preload should be provided to prevent leakage when fluctuating loadscause the gasket to unload. If the bolt preload is too high and if gasket compression isincreasing during a condition of fluctuating loads, the gasket crushes and mechanicalfailure may occur. The selection of a thread lubricant must not change the initial preloadout of the design range or else gasket overload or underload may lead to failure. Forservice or maintenance applications, thread lubricants should not be specified that have anut factor significantly higher or lower than that specified by the factory without a detailedanalysis of loading, assembly methods, and torque specification. New applications shouldhave fastener size, lubricant, and as-


sembly procedures specified to meet applicablecode requirements and to provide suitableassembly bolt tension. The determination ofinitial bolt preload is discussed in Section VI ofthis handbook.

Mechanical systems are often exposed to cyclicloads that affect the performance of a fastener.These fluctuating loads can arise from sourcessuch as rotary imbalance, reciprocatingimbalance, cyclic external loads, and intermittentexternal loads that induce a structural ringing.When the frequency of cyclic loads is near any ofthe system's resonant frequencies, large cyclicfluctuations of the bolt loads, both axial andshear, can occur. These high cyclic loadings canaffect the performance of the fastener ifsufficient preload is not maintained. Two modesof failure that an occur under these conditionsare self-loosening and fastener fatigue failure.


7.1Self-Loosening

The fasteners in an assembled bolted joint mayloosen over a period of time when subjected tofluctuating loads. There are many theories as tothe mechanism of vibratory loosening. The topicof self-loosening is discussed in detail in Chapter60 in this handbook. The self-looseningcondition can lead to other modes of failure asthe preload is lost. Gasket failure or leakage,fastener fatigue failure, joint loosening, flangemisalignment, fastener misalignment, orcomplete failure of the bolted joint can resultfrom a condition of self-loosening. In general,the solution or preventive action for self-loosening is based on the concept that fastenerswill not self-loosen if the friction between thethreads is not reduced under the conditions offluctuating external loads [13, pp. 527563]. Areduced coefficient of friction and a steep threadlead angle will increase the tendency for


vibratory loosening compared to a fastener witha high coefficient of friction and/or a shallowthread lead angle. Self-loosening can becontrolled by proper preload and thread frictionand by mechanical devices. In situations wheresignificant fluctuating loads are anticipated, athread lubricant with a higher friction coefficientmay be specified. Devices that create a prevailingtorque, such as nuts without round holes, nylonlocking collars on nuts, nylon patches, or lockwashers can reduce the concern regarding lownut factor thread lubricants and self-loosening.Devices that mechanically lock the fasteners, forexample wiring or pinning, can be used toprevent self-loosening. See Section 8 of thischapter and Handbook Chapters 2123.

7.2Fatigue

Another issue concerning vibration andfluctuating loads is fatigue. Fasteners must havesufficient initial preload to minimize the portion


of the fluctuating external load that is carried bythe bolt and hence improve its fatigueperformance. A widely used fatigue failurecriterion, the Soderberg criterion, evaluates theminimum and maximum bolt stress to determinewhether the fastener can be expected to fail byfatigue. For most metals, a state of stress that isfluctuating between Smin and Smax is more likelyto result in a failure than a constant stress of Smax[12, p. 319]. The bolt preload and external loadsare used to determine the mean stress. Thefluctuating external loads on the bolted joint areused to determine the alternating stress level.Joint stiffness effects will split fluctuatingexternal forces between the bolt and the flange.The fluctuating bolt force is combined with thebolt stress area and any relevant stressconcentration factors to determine an alternatingstress level. A fastener design is expected tosatisfy the Soderberg criterion when the meanstress and the alternating stress fall below theSoderberg line. Some of the other criteria that are


used for the design of fatigue-resistant fastenersand other structures include the Goodman,modified Goodman, and Gerber [14, p. 299]. SeeChapter 54 for additional information on thefatigue failure of fasteners. When preloads aretoo low, the flanges can completely unload andcause all of the external load to be carried by thebolt, which results in excessive fluctuatingstress. Too much preload can result in either adirect, noncyclic tensile overload or a fatiguefailure due to the high mean stress. Eithercondition of improper preload, too high or toolow, can result in failure [12, pp.


817822]. The selection of a thread lubricant canhave a significant impact on the initial preloadapplied to a fastener. Care must be taken to keepthe preload at the proper level for both newdesign and repair and maintenance situations byspecifying lubricants that result in properpreloads to meet fatigue criteria.

8Related Devices

Mechanical thread-locking devices and liquidsealant thread adhesives create a prevailingtorque condition that prevents the loss of loadthrough self-loosening. These devices areimportant to the selection of a thread lubricantbecause the prevailing torque eliminates the self-loosening concerns when selecting a threadlubricant. Mechanical thread-locking devicesinclude lock washers, nylon thread inserts, cotter


pins, and other devices. See Chapters 2123 and60 for more information on thread-lockingfasteners. Liquid sealants are generally not usedin addition to thread lubricants because they areapplied to the same threaded surfaces.

Liquid sealants include retaining compounds,thread-sealing adhesives, and thread-lockingadhesives [15, pp. 130,137]. Retainingcompounds are normally anaerobic orcyanoacrylate adhesives. They fill the voids in thethreaded assemblies and then chemically harden.In many cases, the assembled fasteners maybecome permanently bonded. The threads aresealed, and they become vibration-resistant,protected from corrosion, and leak-free. Thread-sealing adhesives also fill the gaps between thethreads of an assembled fastener and provide alevel of thread locking. These materials are oftenused to seal threads on pipe joints. Thread-locking adhesives are intended to provide a levelof breakout torque. These adhesives come in a


variety of strengths and formulations dependingon the application. The popular formulations areone part solvent-based chemical curing, one partanaerobic, two parts curing, and "passivated"preapplied adhesives. The primary function ofthread-locking adhesives is thread locking.Thread-sealing performance will vary withmanufacturer and formulation. See Chapters 22and 23 for more information on thread sealants.

References

1. Schaefer, W. L., Thread Lubricants/JointCompounds, Fel-Pro Internal Communication,April 25, 1985.

2. Taylor, E., SermeTel W (TM) aluminumcoating for aerospace fasteners, SPS LaboratoryReport No. 5392, SPS Technologies, Inc.,Jenkintown, PA, Oct. 30, 1980.

3. Tests on 1 inch A490 conventional and twist-off high strength bolts and direct tension


indicators conducted in the as-received conditionand after weathering, Report No. 870702,Pittsburgh Testing Laboratory, Pittsburgh, PA,April 10, 1986.

4. Yura, J., K. Frank, and D. Polyzois, HighStrength Bolts for Bridges, PMFSEL Report No.83, University of Texas, May 1987.

5. Crispell, C., Torque vs. induced load of A-286and MP35N nuts and bolts with cadmium, dryfilm and cetyl alcohol lubricants, Request No. 8-1-8-EH-07694-AP29-B, NASA Contract No.NAS832525, Nov. 14, 1978.

6. Emrich, M., Corrosion protection forfasteners, Part 1, Assembly Eng., October 1982.

7. Gresham, R. M., Bonded solid-film lubricantsfor fastener coatings, Fastener Technol. Inc.,April/ May 1987.

8. Hung, N., Clamping force for electricallyconductive fasteners, Machine Design, October


1984.

9. Brenner, H. S., Results of special evaluationof Pepcoat 6122 and nuclear grade N5000antiseize lubricant systems on torquetensionperformance of ASTM A325 series hex-headbolts submitted by Power and EngineeredProducts Company, Inc. (PEPCO), SouthPlainfield, NJ, Report No. C 16034-1, AlmayResearch and Testing Corp, Los Angeles, Oct.25, 1982.

10. Pepcoat, a study in protection andperformance, G* Chemical Corp., Wayne, NJ,undated manufacturer literature.


11. Chaaban, A., M. Derenne, A. Bouzid, and W.Schaefer, Evaluation of torque coefficients andgasket stress distributions in a bolted flangedjoint using different types of lubricants.Presented at CETIM 3rd InternationalSymposium on Fluid Sealing, Biarritz, France,Sept. 1517, 1993.

12. Deutschman, A. D., W. J. Michels, and C. E.Wilson, Machine Design Theory and Practice,Macmillan, New York, 1975.

13. Bickford, J. H., An Introduction to theDesign and Behavior of Bolted Joints, 3rd ed.,Marcel Dekker, New York, 1995.

14. Shigley, J. E., and C. R. Mischke, MechanicalEngineering Design, 5th ed., McGraw-Hill, NewYork, 1989.

15. Brown, M. W., Seals and Sealing Handbook,3rd ed., Elsevier Science, Oxford, UK, 1990.


3rd ed., Elsevier Science, Oxford, UK, 1990.


4Adhesives and Sealants for Bolting

John CoccoLoctite Corporation, Rocky Hill, Connecticut

1Introduction

A structure of technology is like any otherstructure: foundations go in first. Thefoundations of adhesive technology are properplanning and a thorough understanding of thematerials being used. Knowledge of adhesivemethodology is essential for successful bonding.The most frequent causes for the failure ofbonded joints do not involve adhesive strength.Rather, they are inadequate preparation ofsubstrates and improper adhesive selection.


Adhesives are bridges between substratesurfaces, whether those surfaces are of the sameor different materials. The bonding mechanismdepends on (1) the bonding strength between theadhesive and the substrate, adhesion, and (2) thestrength within the adhesive, cohesion.

2Pretreatment of Surfaces to Be Bonded

Correct surface pretreatment is necessary foroptimum bonding. Bond strength is determinedto a great extent by the adhesion between thejoint surfaces and the adhesive. It is important tounderstand that adhesive joints are stronger themore thoroughly the surfaces are cleaned.Adhesion is improved by (1) removing unwantedsurface films by degreasing or mechanicalabrasion and, if needed, (2) building up a new,active surface by coating with primers.

Pretreatment methods for plastics are listed inTable 1.


2.1Degreasing Surfaces to Be Bonded

The complete removal of oil, grease, dust, andother residual dirt from the bond surface isrequired for the best possible adhesive joint.Solvents that evaporate without residues aresuitable for this. The most important solventsand their cleaning capacity are listed in Table 2.Alkaline or acid-based aqueous cleaning systemsalmost always contain corrosion inhibitors. Ifthese remain on the cleaned bond faces, they mayreduce the adhesion of the adhesive. If suchcleaning systems are to be used, then preliminarytests should always be carried out. In every case,substrates must be thoroughly rinsed or wipedoff.


TABLE 1 The Most Important Pretreatment Methods for Difficult-to-Bond PlasticsPretreatment method MaterialsFlame treatment Predominantly by PE, PPCorona porcess Plastics with low surface energyLow-pressure plasma Plastics with low surface energySurface primer/adhesion promoter Plastics with low surface energy

If special degreasing baths are used for large production runs, itis advisable to preclean very dirty surfaces so that the cleaningbath is not contaminated. Vapor degreasing systems are veryfrequently used. In this case, the solvent is heated to its boilingpoint and evaporated. When the cold substrates are brought intocontact with the evaporated cleaner, the cleaner condenses on thesubstrates. The liquid formed removes the remaining particles ofdirt and grease. Degreasing is often carried out within fullyenclosed machines with the use of degreasing solvents.

For many applications, pretreatment of the surfaces with a fast-acting cleaner is sufficient. It removes oils, greases, looseparticles of dirt, and other contaminants and thus prepares thesurfaces for bonding. When cleaning with solvents, it is possibleto assist the chemical degreasing process for separating dirt fromthe surface by mechanical action (rubbing with a cleaning rag,brushing), thus obtaining a better cleaning result. With gray castiron castings and nodular cast iron, additional mechanicalcleaning is necessary to remove graphite from the surface area.

2.2


Mechanical Pretreatment

Metal surfaces are often covered with an oxide coating thatcannot be removed by degreasing. In such cases, mechanicalsurface pretreatment such as grit blasting, grinding, or wirebrushing is necessary.

Grit blasting is a good way to clean large surfaces. The surfaceroughness achieved by this method provides very good bondingresults, provided not too coarse a grit is used. Grinding achievesequally good surface roughness. In this case, it is important touse a suitable grain roughness (e.g., 300600 for aluminum, 100for steel). After grit blasting, as well as after grinding orbrushing, parts should be degreased to remove all traces ofresidual grit. Very dirty parts should also be degreased beforemechanical treatment so that the grit or abrasive used does notjust smear the surface contaminants. Mechanical pretreatmentmethods are very simple to use and generally provide adequatebonding strength as well.TABLE 2 Solvents Used Most Frequently for Cleaning Bond SurfacesSolvent or cleaner Cleaning

capacityInflammable or

combustibleSurfaceresidue

Hydrocarbons (e.g.,isoparafins) Good Yes Sometimes

Ketones (acetone) Good Yes NoAlcohols (isopropanol) Moderate Yes NoWater-based cleaners Moderate/good No Yes, must

rinse


Paint should be stripped or ground from paintedparts before the actual pretreatment, so that bondstrength will not be reduced by the relatively lowadhesion of the paint to the substrate. If plasticsor rubber parts are to be bonded, the surface filmor vulcanization film should first be removedmechanically. For plastics, abrasives such asfused cast iron or aluminum oxide have provento be effective. Rubber surfaces require cleaningto remove mold release agents, either withsolvents or by grinding.

2.3Wettability Test

Cleaning processes can be evaluated with thewater break test. Several drops of pure water areapplied to the cleaned surfaces. On aninadequately cleaned surface, the spherical formof the drop is largely retained, and the surfacemust be cleaned once more. If the water runs on


the treated surface, then wetting has beensatisfactory; the bond face is sufficiently clean.

This method is not suitable for anodic coatingson aluminum and magnesium. The advantage ofthe water break test is the easy availability of the"test fluid"water. This advantage, however, islimited by the variation in water hardness, whichaffects surface tension. In some cases evendistilled water does not produce reliable resultswith the water break test. In such cases,comparable fluids, which are available withdefined surface tensions, are recommended. Notethat the test covers only wettability and notadhesive bonding capacity.

3Adhesive Joint Failure

Some important criteria on an adhesive jointfailure may be determined by visual evaluationof the bonded parts. Thus it may be possible todetermine whether adhesion or cohesion failure


has led to the failure of the joint or even whetherthe bonded parts have been destroyed.

Adhesion failure. The adhesive can beseparated completely from the face of onesubstrate.

Cohesion failure. The adhesive itself breaks.Remains of adhesive are to be found on bothsubstrates.

In adhesion failure, obviously the weak point ofthe bond is the boundary layer between thebonded part and the adhesive. Either the materialis unsuitable for bonding or the bonding face wasdirty. In either case the strength can be increasedby a suitable pretreatment of the surface.

In cohesion failure, the adhesive is overstressedthrough external action (e.g., stress spikes,temperature, aging). It may be necessary todesign changes in the bonding geometry and/orchoose a more suitable type of adhesive for thisapplication. Cohesion failure can also be caused


by insufficient curing of the adhesive.

The appearance of a joint failure only tells uswhere the weak point of the bonded joint is, notwhat caused the failure. To correct the problem itis essential to find the cause of the failure. Table3 lists the possible causes of adhesive jointfailure and how to deal with them.

4Choosing an Adhesive

4.1Types of Machinery Adhesives

Many adhesives are reactive materials. Reactiveadhesives are applied as liquids and react (cure)to form solids under appropriate conditions. Thecured adhesive is a plastic. The following is a listof common types of industrial adhesives.

Anaerobic adhesives react in the absence ofoxygen and contact with metal. The bond facemust be at least 5 mm wide to ensure the


absence of oxygen. Anaerobic adhesives haveproven their value in thread-locking,cylindrical assemblies, and liquid gaskets.


TABLE 3 Causes of Adhesive Joint Failures and Corrective ActionsPossible cause Corrective action

Faulty substrate Check tolerances, gaps, and materials, and monitormore carefully.

Dirty surfacesCheck pretreatment for suitability and modifyaccordingly (e.g., cleaning agents, cleaningprocedures, subsequent intermediate storage).

Faulty or incorrectassembly of the bond

Check all process parameters and techniques usedto assemble the bond; optimize type and durationof fixturing; check whether all curing conditionshave been met in fixtured state.

Insufficient curing of theadhesive

Check curing preconditions (e.g., gap, airtightness,temperature, humidity). Observe curing times inaccordance with data sheet. Check whether shelflife of adhesive was exceeded.

Mechanical overstress orunfavorable stress (peeling)

Enlarge bonding face and/or modify joint geometry.Check suitability of adhesive for type of stress(tensile, shear, etc.)

Thermal overstress Select adhesive with greater temperatureresistance.

Corrosion or infiltrationand destruction of theadhesive coating throughliquid and gaseous media

Protect gap at the contact faces to the medium witha suitable coating, or design the bonded parts insuch a way that there is no contact with themedium.

Modified acrylics cure in the absence of oxygen and incontact with an activator. During assembly, onesubstrate is wetted with adhesive, the other substrate


with activator. With these adhesives, there is noproblem with the mixing of components or the opentime for the joining process, for the adhesive reactsonly when the two bonding surfaces come into contactwith each other. The bonding surface requires aminimum width of 5 mm to exclude oxygen.

Ultraviolet (UV)-curable adhesives react whenexposed to ultraviolet light. One importantprecondition is that the UV light should reach theentire bond face. To achieve this, at least one of thebond faces must be transparent to the appropriatewavelength of UV light. Commercial UV lamps areadapted to these adhesives with respect to intensity andradiation spectrum.

Combination cure mechanisms incorporate two ormore mechanisms, e.g., anaerobic and UV. Ultravioletlight is used to cure the fillet fixture parts until theanaerobic mechanism takes over to cure the adhesivein the joint. UV and other cure mechanisms can becombined to surface-cure resins that may be in theshadow of the UV light during the cure process.

Cyanoacrylates (popularly known as "superglues")cure very quickly when confined between surfaces. A


trace amount of moisture on the substrate is sufficientto catalyze the curing reaction, which moves from thesubstrate surfaces toward the middle of the adhesivejoint. Cyanoacrylates are not true moisture-curingadhesives because water is not a direct reactant in thepolymerization process.


Single-component urethane adhesives curveto a flexible solid when exposed to ambientmoisture. The moisture reacts with anisocyanate complex to achievepolymerization.

Single-component (heat-cure) and two-partmix epoxy adhesives can be formulated tomeet a wide variety of end use requirements.Filler content can significantly affectperformance.

4.2Adhesive Durability

In selecting an adhesive for a particularapplication, one of the most importantconsiderations is the environment orsurroundings in which the adhesive joint will berequired to function. Of course, the force actingon the joint is of prime consideration, and the


adhesive joint must be capable of carrying themaximum expected load (without excessivecreep) and amount of fatigue or cyclic stresses.Cyclic stresses, particularly slow ones, are muchmore damaging to an adhesive joint than a steadystress. The adhesive selected for a particularapplication must be able to resist these loads andstresses not only initially but also after exposureto the most severe environmental factors to beencountered during the life of the adhesive joint.Heat and humidity are usually the most damagingenvironmental factors for most bonded joints.Thermal expansion stresses created betweendissimilar materials that have widely differentcoefficients of thermal expansion, e.g., a plastic-to-metal bond joint, require low modulus(nonbrittle) adhesives for the best performance.Other deleterious factors are solvents andultraviolet radiation. Always choose an adhesivethat is resistant to these factors; do not plan oncoating the adhesive joint with some "protective"coating that can possibly crack or eventually


become permeable to solvents or moisture.

5Functional Uses of Adhesives

5.1Thread Locking

Thread lockers guarantee more than functionalreliability. The cured liquid that fills the voids inthe threads not only prevents relative movement,it also seals the joint. Excellent chemicalresistance allows thread lockers to be used assealants with most gases and fluids used inindustry. They also seal out moisture andcorrosives that can shorten the life of theassembly. This sealing effect allows through-holes to be drilled instead of blind holes, aneasier and cheaper assembly method. A boltsecured with an anaerobic thread-locking agentwill last the life of the assembly without seizing.

5.1.1


Why Threaded Assemblies Fail

The water screw invented by Archimedes(287212 BC) led to the development of thethreaded fasteners now common to allconnection technology. Threaded assembliestoday are the most important detachable partsused in machinery construction, installation, andrepair. They are so common they are taken forgranted and their function is not seriouslyanalyzed. The two main causes for the failure ofa threaded assembly are relaxation of tension andself-loosening.

5.1.2Relaxation

A threaded assembly "relaxes" when a permanentchange in the length of the bolt occurs in thedirection of its axis or the substrate itself relaxes,as in gasketed surfaces. This reduces the bolttension and thus also reduces the residualclamping force. Permanent changes in length may


be produced by

Settling. The rough faces of the contiguousparts (e.g., nut, washer) become smootherunder the pressure of the bolt tension.


Creeping. The surface pressure on the bearing surface of the bolt or nut exceeds thecompressive strength of the material of the stressed part.

Preventing Relaxation

If the elasticity of the assembly can be increased so that the expected amount of settlingand creeping can be compensated for, then the drop in prestress force can be largelyprevented. This is made possible by the use of (1) bolts with a high l/d ratio (l = shaftlength, d = shaft diameter); (2) collar bolts and collar nuts as well as hardened andtempered washers, which reduce surface pressure and thus settling on the bearing surfaces;(3) bolts and nuts with pressed-on spring head washers or concave bearing washers; and/or(4) rigid conical spring washers or cup springs.

5.1.3Self-Loosening

The threaded assembly loosens itself when sliding movements occur between the contactsurfaces. These forced relative movements overcome the frictional forces in the threadedassembly, and the self-locking effect of the thread is eliminated. Only when the clampingforce is great enough to prevent such movements can the movement ML be overcome toloosen the assembly. Here, the following applies:

where

ML=self-loosening torque

FV=available prestress force

d2 =pitch diameter of threadf =helix angle of threadr =angle of friction of threadmA

=coefficient of friction of bearing surfaces

rA =lever arm of frictional force at the bearingsurfaces.

If the binding force of the threaded assembly cannot prevent relative movements between


the parts under stress, then the bolt makes a rocking movement: The flanks of the threadslide over one another and the bolt becomes almost frictionless. The loosening torque isthen

This loosening moment Mi, which is dependent only on the prestress force, the pitchdiameter, and the helix angle of the thread, acts against the direction of tightening andcauses the threaded assembly to loosen. Sliding movements between the contact surfacescan be caused by

1. Dynamic load in the direction of the axis. A pulsating axial overload leads to a relativemovement at the flanks of the thread

2. Dynamic load at right angles to the direction of the axis. If the materials of theassembled parts have different thermal expansion rates this can occur. Bending, repeatedimpacts, or vibrations can also overcome the forces of friction between the bearing parts.

Preventing Self-Loosening

The following design measures can prevent uncontrolled loosening of correctly loadedbolts:

1. The use of high tensile strength bolts allows prestress forces that are high enough toprevent relative movements.


2. Design that increases the l/d ratio (l =engagement length, d = bolt diameter) increasesthe elasticity of the assembly (historically, l/dratios ³6 have been optimum).

3. Friction can be increased by altering thesurface finish and structure of the bearingsurfaces of the bolt and nut.

4. The use of an adhesive eliminates the degreeof freedom for lateral movements due to the factthat the gaps are completely filled, and at thesame time thread friction is increased byinterfacial connection after the adhesive hascured.

5. Slip in the thread can be limited by creating apositive connection (e.g., body fit bolts, weldingspots).

5.1.4Anaerobic Thread Lockers


Anaerobic Thread Lockers

Devices to prevent self-loosening must meet thehighest standard with respect to thread locking.Single-component liquid adhesives have beendeveloped that completely fill the microscopicgaps between interfacing threads. They cure to atough solid when they come into contact withmetal in the absence of air. The adhesive createsan interfacial connection, keying to the surfaceroughness to prevent any movement of thethreads. The problem is thus solved where itarises: in the threads. This is why anaerobicthread lockers are among the most effectivemeans for locking fasteners.

It is important that the total length of the threadbe wetted and that there be no restriction to thecuring of the adhesive. (Certain oils or cleaningsystems can impede or even completely preventthe adhesives from curing by anaerobic reaction.)The liquid adhesive may be applied by hand orwith the help of a special dispensing device.Proper wetting of a thread is dependent on the


Proper wetting of a thread is dependent on thesize of the thread, the viscosity of the adhesive,and the geometry of the parts. If the parts are oflarger dimensions, then wetting both facesprovides the necessary reliability for the adequateapplication of the adhesive. With blind holethreads, it is essential that the adhesive be appliedto the bottom of the threaded hole. The quantitymust be such that after assembly the displacedadhesive will fill the entire length of the thread.

Some anaerobic thread-locking products have apositive influence of the coefficient of friction inthe thread. The values are comparable with thoseof oiled bolts. Prestress and installation torquecan be defined exactly. This property allowsthread-locking products cured by anaerobicreaction to be integrated into automaticproduction lines using existing assemblyequipment.

5.1.5Preapplied Fastener Coatings


If hand application of adhesive is not desired on acontinuous assembly line, or if dispensing unitscannot be used, then bolts precoated withadhesive are an alternative. Microcapsulescontaining an active ingredient are preapplied tothe threads in a dry film adhesive coating. Whenthe fastener is assembled, the capsules arecrushed, causing the chemical reaction (yieldinglocking strength). Self-loosening of the bolt isprevented. Precoated bolts are treated and storedas normal bulk material. This coating systemuses water (in most cases) as the carrier agent,making it user-friendly.

Precoating bolts also provides advantages forquality assurance. The quantity of adhesivedispensed for the coating is very consistent dueto constant quality checks performed byspecialized coating companies. Existingassembly equipment can usually be adapted touse precoated bolts without tooling changes.

5.1.6


A Comparison of Threadlocking Methods

The best way to evaluate a threadlockedconnection is to test its behavior under loadcycles in a dynamic test machine. The lower theloss in bolt tension, the more reliable theassembly.


Testing Clamp Load Retention Performance ofVarious Locking Devices

To determine clamp load retention curves forvarious locking devices, a bolt test stand similarto the Junker machine is used. The braced bolt isstressed vertically to the bolt axis by adisplacement unit that is adjusted with a cam.

Assemblies locked with anaerobic threadlockerseasily withstand the vibration that results fromthis type of analysis. Most mechanical methodsfail this test. This does not mean that thesemethods are not useful to a certain degree.However, when their functional operation iscompared with their cost, they are difficult tojustify. With the anaerobic thread-lockingsystem, there is no further effort or expenserequired for mechanical locking elementsbecause ''one size" fits all. Most thread-lockingproblems are thus solved in an economical way.


A further advantage of anaerobic threadlockingagents is that the liquid adhesive is effectiveregardless of bolt sizes and diameters. Whilespecial locking bolts require elaborate inventory,Loctite threadlocking adhesives fit every standardbolt. The need to search for or order speciallocking systems or bolts is eliminated.

A similarly favorable load cycle performance isshown only by the surface-compacting ribbedflange bolt. Its disadvantages include high costs,the relatively large amount of space required forflange-bearing surface, and the unavoidabledamage to the surface of the braced parts aroundthe bolt-bearing surface. Teeth on bolts with asawtooth flange penetrate the bearing surface ofthe braced material. Bearing surfaces of the headand the nut are damaged during loosening,limiting their possible applications. Parts withhardened surfaces cannot be reliably connected.

Breakloose Strength


In a comparative test, breakloose torque may beused as a measure of wetting, adhesion, anddegree of curing. In general, there is no directrelation between breakloose torque andselfloosening resistance. In practice, it is oftenrequired that threaded assemblies be capable ofbeing disassembled using normal tools.

Adhesive companies have developed productswith low and high breakloose torques to meetthese needs. For distinction, the definitions "maybe disassembled using normal tools" and"difficult to disassemble" were chosen. Becauseadhesive fills the joint, the fastener cannot seizebecause of corrosion. To reuse a bolt that hasbeen locked with adhesive, old adhesive isremoved before new adhesive is applied.

Breakloose strength depends on

1. Length of thread

2. Material match


3. Surface finish

4. Prestress torque

5. Bolt diameter

Note: Bolts should be replaced if they have beenstressed to the apparent limit of elasticity andthen loosened. If reused, the risk of a break existswhen the same prestress force is applied.

5.2Thread Sealing

Thread sealants prevent leakage of gases andliquids from pipe joints. All such joints areconsidered to be "dynamic" due to vibration,changing pressures, or changing temperatures.

5.2.1Sealant Types

Noncuring Pipe Drops

One of the oldest methods of sealing the spiralleak paths of threaded joints, pipe dopes are


pastes composed of oils and fillers. Theylubricate joints and jam threads but provide nolocking


advantage, squeeze out under pressure, and havepoor solvent resisstance. They do not work onstraight threads.

Solvent-Based Pipe Dopes

Solvent drying pipe dopes are also an old methodof sealing threaded joints.

Advantages: Provide lubrication and orificejamming, extrude less easily.

Disadvantages: Shrink during cure as solventevaporates. Fittings must be retorqued tominimize voids. They lock by friction.

Elastomer Tape

The best known elastomer tape is TeflonTM tape,which gives a good initial seal and resistschemical attack. It is the only organic sealantallowed for gaseous oxygen.


Advantages: Acts as a lubricant; allows hightorquing; resists solvents.

Disadvantages: Lubricates in off-direction aswell as on-direction; allows fittings to loosen.Tape residues contaminate filter systems.

Yielding Metal

While not in fact a "sealant," designers oftenspecify dry seal fittings that in theory do notrequire a sealant.

Advantages: Design is effective in carefullymachined joints.

Disadvantages: Machining costs are high, andtolerances are difficult to maintain.

Anaerobic Thread Sealants

Anaerobic sealants cure to insoluble toughplastic thread fillers that prevent leakageregardless of the torque applied.

Advantages: Lubricate during assembly; seal


regardless of assembly torque; seal to the burstrating of the pipe; provide controlled disassemblytorque, even years later; do not cure outsidejoint, easy cleanup; available without fillers forcritical hydraulic fittings; lowest cost per sealedfitting; easily dispensed on production lines;some grades are available as preapplieds.

Disadvantages: Not suitable for oxygen serviceor strong oxidizing agents; not suitable forsealing at temperatures above 200°C; typicallynot suitable for use with piping diameters greaterthan M80.

5.2.2The Key Factors of Sealing Threads

The many influences that pipe joints face duringtheir lifetime of service should be known andunderstood during the design stage when sealantsare selected. Sealants must be chosen forreliability and long-term quality. It is no longeracceptable for equipment to leak oil or any other


substance. Pipes must remain leak-free under theseverest vibration, chemical attack, heat, orpressure surges.

Leak paths go through the thread roots of eventhe most properly stressed threaded pipe joints.They also pass among the surface irregularitiesof thread flanks. This requires a sealant to havethe proper wetting ability to conform perfectly tothe thread surfaces. Size is a factor. A sealantdesigned for 8 mm fittings may not work on 800mm diameter fittings.

Many sealing materials do not totally fill theinner space in the threads. They seal only becausethe prestress is so high on the thread flanks thatthe materials are compressed into surfaceirregularities. Conventional pipe dopes are slowand messy to assemble and interfere with thetorques needed to obtain the proper prestress.Teflon tape requires skill to avoid overstressingand possibly cracking fittings or castings. Dryseal threads work when held in constant high


compression. However, it is often impossible toattain or maintain these required prestresses indynamic applications. There are three reasons forthis:


1. Pipe fittings are often backed off to achievespecial alignments for bends, elbows, or gages,sacrificing the proper prestress position.

2. Vibration can cause thread flanks to fret to thepoint of looseness. Flexible connections such ashydraulic hoses are especially vulnerable.

3. Relative motion can force tape sealant out ofthe joint.

5.2.3Anaerobic Sealants

The proper anaerobic sealant completely fills thevoids to create a seal. Good sealant selectionprovides the strength to eliminate relative motionin the joint and thus prevent the cause of mostleakage.

Anaerobic sealants can be applied by hand andwith automatic and semiautomatic equipment.


Excess material is wiped or washed away fromparts or hands.

Anaerobic thread sealants are quick, clean, andeasy to apply directly from containers orapplicators. When assembling straight or taperedthread connections, the sealant should be appliedto both male and female components. Anaerobicsealants fill threads thoroughly, making theprestressing of the joints less critical.

Anaerobic sealants also provide easy disassemblyin future maintenance because pipe joints cannotcorrode and seize. The filled threads prevent theentry of moisture or corrosive chemicals.Although varying locking strengths are available,all anaerobic sealants allow disassembly withconventional tools.

Contamination-Free Systems

Any sealant can enter its piping system. TeflonTMtape is especially dangerous to hydraulic systemsthat contain many small orifices. Uncured


anaerobic sealants fortunately dissolve withinmost fluid systems.

Chemical Resistance

After cure, anaerobic sealants resist mostindustrial liquids and gases. Fluid compatibilitytables are available from suppliers upon request.

Environmental Friendliness

Cured anaerobic thread sealants are low intoxicity and are often used in the food andbeverage industry. However, these uses aresubject to local regulations, which should alwaysbe followed. Many products have beenrecognized by the U.S. National SanitationFoundation for potable water systems (Article61).

Temperature Range

Anaerobic thread sealants operate from 55 to+150°C or from 65 to +300°F, continuoustemperature. Brief exposure to higher


temperatures will not impair the sealing effect.

Anaerobic sealants are not typicallyrecommended for long-term applications oncopper slip fit or other joints in combinationwith substrates with high copper content thatcome in contact with water over 40°C.

6Equipment

6.1How to Dispense Adhesives

The capital investment required for setting upproduction line bonding systems will pay foritself in adhesive and labor cost savings,particularly on higher production lines. In manycases the installation of suitable dispensing unitsis extremely cost-effective. Many suppliers offera


broad range of dispensing methods, from manualsystems for small batch production or repairs tofully automatic systems. The adhesive determinesthe choice of dispensing method.

Viscosity is one of the decisive factors in thechoice of a dispensing system. The flowproperties of the adhesive determine the controltechnique and the choice of valve for theapplication of the adhesive. The goal is tocompletely wet both bond faces, and for this theproperties of the adhesive and the surfaces mustbe controlled.

Chemical compatibility between the adhesive andequipment components is also quite critical.Therefore, adhesive suppliers who also provideequipment often have an advantage using asystems approach (one-stop shopping!).

One way to ensure a well-bonded joint is to


apply the correct quantity of adhesive in theproper place. Automatic dispensing systems areincreasingly relied on to attain this goal, whichcan statistically improve success rates.Automatic systems can be monitored and thework checked piece by piece. The correctpositioning and quantity can often be checked bymeasuring the fluorescence of the appliedadhesive.

Exact reproducible amounts of adhesive can bedispensed by using dispensing equipment. Theequipment available in most cases has beendeveloped with respect to the specificperformance of the products being dispensed.The amount of dispensed adhesive is determinedby a pressure-time system. The adhesive is putunder pressure, and the dispense valve is openedfor a certain time to provide the correct amountof adhesive. Because the dispensed productcontacts the pressurized air, air quality isimportant; the use of a filter-dryer with each


system is suggested.

6.2Equipment Specification

Dispensing and curing systems are used for awide variety of bonding, thread-locking,retaining, gasketing, and sealing applications. Forsome jobs, it is sufficient to dispense productdirectly from the bottle or tube onto the surfacesto be joined. In other cases, however, moreprecise and automated dispensing is required. Tomeet this need, many suppliers have developedequipment specially designed to makeapplication of adhesive products economical,fast, precise, and clean.

Currently available equipment technologyenables end users to apply beads, drops, orcontinuous rings of adhesives and spray, spin,and screen print chemical sealants. In addition tothe standard equipment available, mostengineering adhesive suppliers can design and


build equipment based on the customer'srequirements to provide unique applicationsolutions.

The end user should always consider methods ofapplication to be as important as the chemicalproducts themselves. And having developed theproduct, adhesive suppliers are in the bestposition to supply the type of equipment thatapplies the adhesive or sealant in the mosteffective way. Therefore, designers anddevelopers of any application should bear inmind the specific characteristics of the productsto ensure clean, precise, and reliable dispensing.

Reference

Loctite Worldwide Design Handbook,19961997 edition, Loctite Corporation, RockyHill, CT.


PART IIPROCESSING OF FASTENERS

5Fastener Manufacturing

Jesse A. PhebusA. Phebus and Associates, Inc., Safety Harbor,Florida

Peter F. KasperVermont Fasteners Manufacturing, Swanton,Vermont

1Introduction

To the design engineer, the primary concern isthe specific unit under design, its function, and


the components' ability to perform as intended,to achieve reliability, and meet and/or beat thetotal production cost parameters. Usually thefasteners that hold the product and itscomponents together are almost an afterthought.

When the design engineer gets to the final stages,the fasteners enter the picture. The designerdevelops the fastener package necessary to meetthe needs of the completed unit and to install theunit so it will perform the intended function.Rarely is much thought given to where and howthe fasteners are to be manufactured.

How important is the fastener manufacturingprocess? Knowing where and how fasteners aremade can affect the strength, reliability, cost, andserviceability of the entire unit.

Fasteners are one the few universal componentsthat must meet all the requirements of strength asthey function in the ground, under water, in theair, and in outer space. They perform their


intended function in any and all environmentalconditions encountered.

A vast variety of materials, processes, andsurface coatings come into play in fastenermanufacturing processes. The manufacturer'sknowledge and skills should be compatible withthe fastener type, style, strength, and application,and the machinery and processes used in theirmanufacture should be suited to the specificfasteners required. Manufacturers do not all havethe same capabilities, and few if any produce thefull range of fastener products. Most are nicheproducers. By locating those producers bestsuited to your fastener design requirements, youcan achieve the best quality, price, and service.

There are two categories of basic fastenermanufacturing methods: forming and threading.

The primary forming methods are

1. Cold heading


2. Warm heading

3. Hot forging or forming

4. Turning or screw machining


The primary threading methods are

1. Roll threading (forming or materialdisplacement)

2. Cut threading

3. Ground threading

2Forming Methods

2.1Cold Heading

Cold heading is a room temperature process. Notemperature elevation device is used eitherinternally or externally to the process. Theforming process is a force, and the expendedenergy allows the material to flow into the shapeof a fastener. The friction energy developed bythe forming process generates sufficient heat to


make the fastener feel hot to the touch.

2.2Warm Heading

Usually a header is equipped with an in-lineheater, commonly an induction heater. The rawmaterial or wire is preheated just prior to theforming process to a precalculated temperaturethat is dependent on the material alloycomposition and the part configuration. Thiselevated material temperature is generally in therange of 4251100°F. Thus the term "warmheading" has been applied to this process. Theelevated temperature allows the header, with lessforce than is required for cold heading, to formthe fastener material more easily and with lesschance of head bursts or cracking. This process iscommonly used when heading tougher materialssuch as stainless steels.

2.3Hot Forging


With hot forging the fastener product material istypically heated to 21002250°F. The heatedmaterial will develop a cherry red to cherry whiteappearance. This practice is used to formmedium carbon to exotic high temperaturealloys. It allows the forging of fastener blankswith fewer forming stations and with smallermachines of less forming tonnage. The hotforging method is employed when more exoticalloys are used to impart high strength tofasteners for nuclear applications and aircraft oraerospace use and/or when extremely low andhigh temperature fasteners are specified.

2.4Turning or Screw Machining

A turning or screw machine uses full-size barstock from which a fastener shape is generatedthrough removal of stock. Preformed shapes,such as hexagonal bar stock, are used from whichhex head bolts and hex-shaped nuts are turned,bored, or drilled.


The advantages of this process are itscompatibility with small runs, lower capitalcosts, and less variation in some productcharacteristics. The disadvantage is higher unitcosts due to lower production rates per machineand higher material losses from the turningsremoved from the full-size bar stock and thegreater number of unusable bar ends.

3Threading Methods

3.1Roll Threading

Roll-formed threads are commonly formed on anextruded blank diameter close to the thread pitchdiameter. Their strength is generally superior tothat of other threads. During the roll formingprocess the axial grain boundary flows andfollows the individual thread contour, thus


developing greater strength than a cut or groundthread. The thread surface finish is quiteacceptable, better than that from cut threadingbut not as good as the surface attained with theground threading method. The quality of the rollthread dies and the die alignment can result inthread flank, root, or crest discontinuities. Highstrength critical parts can be threaded after thehardening operation, resulting in improvedthread fatigue strength; however, this increasesmanufacturing cost.

3.2Cut Threading

Cut threading is one of the oldest methods usedto produce a thread. This process uses threadcutting chaser sets or a turning operation such asa lathe and a cutting tool approach. This methodis the least desirable from the point of view ofthread finish quality and thread strength. Many


small lot fasteners have cut threads. Cut threadsare produced on full body diameter blanks wherethe major diameter of the thread and the fullbody diameter are the same.

3.3Ground Threading

The ground threading method results in a highquality thread with respect to dimensionalcontrol, thread strength, and surface finish. Theground thread is usually created on hardened andcontour-ground fastener blanks. This method ismore costly than roll forming and cut threading.

4Production Support Elements

The factors discussed in this section contributeto the manufacturer's ability to produce cost-effective quality fasteners.

4.1Raw Materials


A wide range of materials are used to producefastener products, both ferrous and nonferrous.The material strength requirements are specifiedfor all products that are made to specificnational, international, and/or industry standards.In addition, special fasteners can be designed tomeet an application's requirements.

Generally the fastener is specified to meetspecific mechanical properties, and themanufacturer can use a range of materialchemistry in achieving these properties.

4.1.1Cold Heading Quality and Scrapless Nut QualityMaterial

Certain elements are necessary to produce cost-effective quality products. If a mill is to supplycold heading quality (CHQ) rod or scrapless nutquality (SNQ) material, this in itself defines aparticular specification; in most cases standardgrades are purchased. The material specification


controls the chemistry, surface discontinuities,grain size, and general quality for cold heading.

Wire processing plays a vital roll in the materialquality required for heading. The headingoperation requires close tolerance wire size and asmooth coated surface finish. Hence the cleanedand annealed green rod from the mill will be colddrawn into wire suitable for the cold headingprocess. There are three key factors in the rod-to-wire conversion:

1. Mill scale is removed by acid pickling ormechanical descaling.

2. Annealing or in most cases spheroidizingannealing is required to soften the material priorto cold heading.


3. The wire drawing, usually done with carbondies, results in size control and also provides thelubricated surface necessary to prevent galling inthe headerforming tools.

4.1.2Warm Heading

Warm heading materials are the same as thoseused for cold heading. Some exotic and high-endalloy materials that are not easily cold formedcan be cold headed through the use of warmheading.

4.1.3Hot Forging or Forming

For hot forging or hot forming, hot rolled rod,bar, or cold finished material is used in eithercut-to-length bars or coils. There is a significantcost savings compared to CHQ/SNQ materialsand screw machine bar stock.


and screw machine bar stock.

4.1.4Turning or Screw Machining

In most cases cold-drawn finished bar material isselected from specific material grades that areeasier to machine, sometimes called freemachining materials.

4.2Tool Control

In heading operations the forming tools controlmost fastener dimensional characteristics. Toolchanges contribute to most machine downtime;thus a sound tool control program is important.

4.3Machines and Equipment

The manufacture of fasteners has evolved overhundreds of years. As indicated below, there are agreat variety of machine types. In addition, eachnarrow range of fastener diameters requires a


machine of a specific size. There are othercharacteristics also that enter into the realm ofthe designer and purchaser for consideration.These are covered below for each of the primaryforming methods.

4.3.1Cold Heading

The equipment in use today for cold heading canrange from a vintage date of 1918 to state-of-the-art machinery. The modern equipment caninclude load-sensing or load-monitoring devices,tool quick-change or automatic changingfeatures, short feed detection, thread formmonitoring devices, etc. Table 1 presents a list ofequipment types and typical products for whicheach is used.

4.3.2Warm Heading

Warm heading technology has advanced totemperature-sensing devices that automatically


maintain the temperatures desired.

Heading Equipment

For warm heading, the equipment is the same asthat used to cold head the part headed blank.Depending on the material alloy and the partconfiguration, a slower cold header may result ina more effective forming process, one that willproduce better yields than high-speed headers.The heated raw material will generally respondbetter to the more gradual application of force,thus allowing the tougher metal alloys to flowwithout cracking and avoiding excessive toolbreakage and wear.


TABLE 1 Cold Heading EquipmentEquipment typea Typical productSingle-die, single-blow Nails, rivets

Single-die, two-blow Blanks with simple round head withphillips, struck slot etc.

Two-die, three-blowShank extrusion, square under head,more than 2.5 wire diameters headvolume, large thin heads

Boltmaker (or similar machines), fouror five dies and matching hammerswith integrated machine point and rollthread capability

Cap screws or similarly shaped parts

Nutformer (or similar machines) Hex, flange, and lock nut blanks

Progressive headers, three to eightdies with matching hammer stations

Complex multiple shank extrusions,backward extrusions; complex headshapes with up to 6 wire diameters ofhead volume material

Long-stroke headersTo run parts with approximately doublethe under-head length of standard orhigh-speed headers

Open-die and/or rod headers suitablefor basic parts, an under head fin ispresent, caused by split heading dies

Very basic parts of longer lengths such asappliance tie rods

Parts former; generally not a threadedproduct. Short-stroke machines withhigh tonnage and multiple formingstations.

Run part blanks such as gears, bearingrace, parts that may generate some roughfinished size, and surfaces that are latermachined or ground

aBoltmaker and Nutformer are trade names of National Machinery Company,Tiffin, Ohio.


Heating Equipment

There are a number of induction heating equipmentsuppliers. Most of the equipment generally performssatisfactorily. The equipment selection should be based onthe specific needs and the overall warm headingrequirements of the fastener producer.

4.3.3Hot Forging and Hot Forming

Hot forging and hot forming started with the blacksmith'stechnology for heating metals. Today the metals areheated electrically by a series of coils or heater boxesdepending on the specific hot forge manufacturingmethod employed. Bolts are generally forged by heating ininduction coils, the area heated depending on the volumeneeded to form the head. Nuts are forged either from cut-to-length bars or coils of mill-supplied green rod. In eithercase, the nut material is heated as the material is fed intothe machine. A few producers may be using gas-firedfurnaces to heat the metals for forging.

Hot Forging or Externally Threaded Fasteners

The types of machines and processes in use range fromone step above the blacksmith era to a semiautomatic in-


line process linking individual operations.


Equipment type Typical productSingle-die, two-blowsplit or solid die,no point

Square and hex head bolts; cut thread may bepointed for thread start

Single-die, two-blow; solid die,extrusion for thread, pointer

Cap screws; parts with washer face under head,pointed and roll thread start

The above equipment can be either horizontal headers or verticalpresses. These are hand fed or may be linked together with somedegree of automation. The quality and cost can vary significantlyfrom producer to producer.

More complex hot forged products require additional operationssuch as descaling of the hot-rolled bars for either surface peelingand turning or possible cold drawing. Sometimes the full bodydiameter and/or washer face is ground.

Hot Forming of Internally Threaded Product (Nuts)

Generally nuts from 3/4 to 1 7/8 in. in diameter are produced onhot nut formers from either cut-to-length hot-rolled bars or hot-rolled coils. Some smaller diameter and larger diameter parts suchas 5/8 in. and 2 in nuts are run on nut-forming machines. Themanufacturing methods in use can be quite varied. Most producersare using hot nut formers equipped with in-line heaters. Some usevertical presses to form and pierce the nut blanks; large nuts, 2 in.and up in diameter, are formed on presses that may or may nothave some degree of automation.

Internal threads are tapped. There are a number of OEMs whomanufacture and supply this equipment to the industry. Some large


nuts that are hot formed will have machined threads. Very largenuts, i.e., 612 in. in diameter, are machined complete including thethreads.

4.3.4Turning or Screw Machining

The turning or screw machine is used in the production of smallquantities of special nuts, cap screws, or boltsareas wheretraditional high volume fastener producers cannot compete inprice, service, or delivery.

The types of equipment employed to machine threaded fastenersfrom metal bar stock will depend on the type of product to beproduced.Machine type Typical productTurning lathe Small quantities of very special nonrepetitive externally threaded partsTurret lathe Small quantities of very special nonrepetitive nutsScrewmachines

This family of machines can generally produce any fastener that iseconomically feasible

Depending on the quantity and complexity of the product, thesemachines are manually operated or computer-controlled.

4.4Management Committed to Customer Satisfaction

Most cost-effective quality manufacturers subscribe to andpractice a management system that promotes the attitude thatcustomer satisfaction is of topmost importance to the company.For those seeking the best manufacturer of the fasteners requiredin their designs, the choice of a company certified to ISO/QS 9000


or an equivalent quality management philosophy is a good start inthe selection process.


6Fastener Coatings

John LaurilliardConsultant, Philadelphia, Pennsylvania

1Introduction

This chapter covers various standard platingsnormally applied to high strength precisionmilitary and aerospace threaded fasteners andbriefly cites other nonelectrodeposited coatingsalso commonly applied to this class of fasteners.

While the information supplied on coatings,techniques, process control, and testing appliesto high strength precision-threaded fasteners, it isalso applicable to less costly commercialfasteners. The same problems encountered when


coating precision fasteners also occur with theless precise variety but are often overlooked ornot detected because of economics and theirnoncritical applications.

2Selection of Coatings*

Threaded fasteners are such simple mechanicaldevices that they are often overlooked during theinitial design stages of the complex mechanicalequipment that they must hold together. Withoutadequate engineering information the choice ofcoating is often given even less consideration andis normally made on the basis of corrosionresistance alone. This simplistic approach doesnot consider the many critical physical,mechanical, and chemical properties threadedfasteners must have, especially the high strengthprecision variety. Therefore the selection ofelectrodeposited coatings requires specializedengineering knowledge.


Ideally, the fastener coating should provideuseful functional properties not provided by thebare alloy, but it should also not causesignificant degradation of the substrate or thejoint structure.

Some factors to consider when choosing afastener coating are listed in Table 1.

At least 24 metals and alloys (see Table 2) withwidely different chemical and physical propertiesare commercially electrodeposited. Theappearance and physical properties, and to alesser extent the chemical properties, of fastenersmay be varied by altering the plating bathformulation and the operating conditions such astemperature and current density. As an ex-

*This section draws heavily on Refs. 1 and 2.


TABLE 1 Factors Affecting the Choice of Fastener CoatingsCoating Factors

1 Corrosion resistanceenvironment2 Temperature limitation3 Effect on fatigue life4 Effect on torque vs. induced loadlubricity5 Effect on locking torque6 Stress corosion resistanceenvironment7 Galvanic compatibility with joint material8 Ability to relieve hydrogen embrittlement9 Abrasive resistance

10 Post-plate hydrogen embrittlement11 Stress alloying12 CostApplication Technique Factors

1 Degree of induced hydrogen embrittlement2 Effect on fatigue life3 Dimensional change or degradation of basis metal during preparation4 Thickness and distribution of coating5 Adhesion of coating6 Limitation of coating method because of fastener material, size, andshape complexity7 Cost

ample, nickel may be plated as a bright, highly stresseddecorative coating with a hardness of 650 Vickers or


decorative coating with a hardness of 650 Vickers oras a nonreflective, soft, low-stress or stress-freecoating with a hardness of 130 Vickers.

Additionally, special properties may be produced bythe codeposition of two metals that form alloys, someof which can be produced only by electrodepositionand not through standard metallurgical techniques.Nickel-zinc alloy coatings have chemical propertiesbetween those of the two component metals. Althoughnot as good as either metal alone, the alloy has betterhigh temperature oxidation resistance than zinc andbetter atmospheric corrosion resistance than nickel.However, nickel-tin deposits have properties unlikethose of either of the alloying metals alone. This alloyhas a pleasing bright pinkish appearance andapproaches the noble metals in corrosion resistance.

The ideal fastener coating would have goodatmospheric and marine corrosion resistance, a highmelting point, and good lubricity, have no adverseeffect on fatigue, and not cause undue galvaniccorrosion of the joint material. The technique ofapplication should produce a thin uniform adherentcoating and be capable of coating both large and small


complex shapes made of widely different metals andalloys. Most important, the process should cause littleor no hydrogen embrittlement, and the product shouldexperience minimal reduction of fatigue life due tosurface preparation.

Just as no one alloy or material can meet every boltapplication, no one coating will meet all therequirements of the ideal coating. But by knowing thefastener application and intended environment, anintelligent choice can be made.

Table 3 lists some of the more commonly usedplatings and auxiliary finishes and their more usefulengineering properties and limitations.


TABLE 2 Fastener CoatingsaMethods of Application

Metalic coatings

Method of application

Electrochemical Electrodeposits Chemical

Unalloyed Alloys Electroless Immersion Physical (variousb)

Cadmium Cadmium-titanium Nickel-phosphorus Copper Vacuum cadmium

Chromium Copper-zinc Nickel-boron Gold Ion vapor aluminum

Cobalt Copper-tin Cobalt-phosphorus Tin Hot dip aluminum

Copper Nickel-zinc Hot dip tin

Gold Nickel-iron Hot dip zinc

Indium Nickel-cadmium (diffused) Mechanical zinc-tin

Iron Tin-cadmium Mechanical cadmium

Lead Tin-lead Mechanical cadmium-tin

Nickel Tin-nickel Mechanical tin

Paladium Tin-zinc

Platinum

Rhodium

Silver

Tin

Zinc

Nonmetallic coatings

Inorganic Organic/inorganic

Method of application

Electrochemical (anodize) Chemical Physical

Oxide Phosphatec Chromate Lubricants Corrosion-resistant

Aluminum Steel Steel Cadmium Oils (various) Paints (various)

Magnesium Copper Cadmium Beryllium Long chain, alcohols Zinc chromate

Titanium Stainless steel Zinc Aluminum Molybdenum disulfide Fluoropolymers

Zinc (passivation) Aluminum Silver Graphite-grease Aluminum-ceramic

Titanium Zinc Zinc flake-organic

aConsult ASTM Standards Vol. 02.05 [3] for a complete list and cross index of specifications governing metallic and inorganic coatings.

bOther physical methods include flame and plasma spray, pack cementation, and sputtering of various metals and alloys.

cMn, Zn, Fe, F.


TABLE 3 Properties of Metallic CoatingsNo. Coatings Specifications Application

methodType ofsolution

Important properties of coatings ormethod Limitations Costs

1 AluminumMIL-C-83488 IVD

A. Atmospheric corrosionresistanceB. High melting point, 1220°FC. Galvanic compatibility withaluminum joints

D. Need lubricationE. High costF. Special equipmentG. Abrasive blast prior to coating

High

2 Cadmium

QQ-P-416ASM-2400;ASTM-B-766;ISO-2082

Ed Brightcyanide

A. 1A, C aboveB. Uniform depositsC. Coverage in recessesD. Precise control of dimensionsE. LubricityF. Thin coatingsG. Solderability

H. Use below 450°FI. Low MP (610°F)J. H2 embrittlement relief difficultK. Vaporizes in vacuumL. Stress alloying in titanium at lowtemp. (125°F)M. Stress alloying in steel above450°FN. Toxic; not for food use

Low




No. Coatings Specifications Applicationmethod Type of solution Important properties of

coatings or method Limitations Costs

3 CadmiumQQ-P-416; ASM-2401; MIL-STD-870

Ed Unbrightenedcyanide

A. Porous depositallows H2embrittlement reliefB. 1A, C; 2E above

C. 1G; 2I, K, L, M, N, aboveD. Carbon filtration for organicsremovalE. Causes H2 embrittlement duringuse from corrosionF. Less thickness uniformity thanNo. 2G. Little or no coverage in recessesH. Must be rack plated; barrelplating peens deposit, closingporous deposit

HigherthanNo.2

4 Cadmium-titanium

MIL-STD-1500;AMS-2419 Ed

Unbrightenedcyanide withtitanium additive

A. Low H2embrittlementB. Better corrosionresistance than No. 3C. 1A, C; 2E above

D. 1G; 2H, I, J, L, M, N aboveE. Elaborate titanium additivesystem and controlsF. Initial strike require for coverageG. Adhesion difficult on highchromium alloy steels

HigherthanNo.3



ContinuedTABLE 3No. Coating Specification Application

methodType ofsolution

Important properties ofcoating or method Limitations Costs

5 Cadmium NAS 672 Ed FluoborateA. Low H2 embrittlementB. High current densitiespossible

C. 2H, I, K, L, M, N; 3F, G above Medium

6 Cadmium MIL-C-8837 Vac A. No H2 embrittlement

B. No post-plate bakeC. 2H, I, K, L, M, N; 3F, G aboveD. Costly vacuum equipment High

7 Cadmium MIL-C-81562 Mech A. Low H2 embrittlement

B. 2H, I, K, L, M, N aboveC. Thin deposits on sharp edges andrecessesD. Possible nicking on large parts fromtumbling

High

8 Chromium, hard(thick)

QQ-C-320;AMS-2406;ASTM-B650

Ed Chromicacid

A. Hard (Rc70); wear-resistantB. Thick deposits, >0.001in.

C. Heavy buildup on high currentdensity areas.D. Often requires post-plate grinding fordeposit uniformityE. Normally on bolt shank and underheadF. Shot peening required to minimizefatigue lossG. Requires application of stopofflacquers

High




No. Coating Specification Applicationmethod

Type ofsolution

Important properties of coating ormethod Limitations Costs

9 Chromium,decorative ASTM-B456 Ed Chromic acid A. Bright cosmetic appearance

B. Applied over copper/brightnickel platingC. Normally racked, but smallparts can be barrel plated

Medium

10 Copper

MIL-C-14550;AMS-2418;ASTM-B734

Ed Cyanide oracid sulfate

A. High conductivityB. Flash plate for increased adhesionprior to silverC. Lubricity for extrusionD. Stop-off for nitriding

E. Can cause H2embrittlement Medium

11 Gold

MIL-G-45204;AMS-2425;ASTM-B488;ISO-4523

Ed Cyanide A. Oxidation resistantB. High infrared reflectivity C. 10E above Very

high

12 NickelQQ-N-290;AMS-2424;ASTM-B689

EdSulfamate,chloride,sulfate/chloride

A. Oxidation resistanceB. Low stressed deposits withsulfamate for minimal fatigue strengthlossC. Used under silver plate foradhesion

D. 10E aboveE. Less deposit uniformitythan cyanide cadmium

Medium



ContinuedTABLE 3No. Coating Specification Application

method Type of solution Important properties ofcoating or method Limitations Costs

13Nickel-cadimum,diffused

AMS-2416 EdSulfamate nickel +cyanide cadmium +630°F diffusion

A. Oxidation-resis-tant to 900°FB. Good atmosphericcorrosion resistance

C. Minimum nickel requiredin recessesD. Control nickel depositstressE. Loss of fatigue strength ifdiffusion is delayed beyond 1hr

High

14

Nickel-phosphorus(electrolessnickel)

MIL-C-26074;AMS-2404ASTM-B656

El Hypophosphite,sulfate

A. No electric currentrequiredB. Uniform deposits inrecesses and threadsD. Deposits can be hardenedto 1000 Vickers by a 1 hr,750°F heat treatment

D. Solution control critical fordeposit propertiesE. Low deposit ductilityF. Short solution life

High

15 Silver

QQ-S-365;AMS-2410;AMS-2411ASTM-B700

Ed Cyanide

A. Lubricant for hightemperature alloysB. Uniform thicknessdistributionC. High melting point(1761°F)D. High load carry capacityE. Heat and electricalconductivity

F. Requires nickel and silverstrike prior to main depositG. Prolonged expolsure totemperatures above 450°F inair reduces adhesionH. Oxygen penetrates silverat high temperature (>450°F)

High




No.Coating Specification Applicationmethod

Type ofsolution

Important properties of coatingor method Limitations Costs

16 TinQQ-T-435; MIL-T-10727; AMS-2408;ASTM-B545

Ed Alkalinestannate

A. LubricityB. Soft-seals threadC. Lower evaporation rate invacuum than Cd, Zn, Pb, andAgD. Uniform thicknessdistributionE. Good coverage in recessesF. Solderability

G. 450°F melting pointH. Converts to brittlegray tin below 40°F

Moderate

17 Zinc ASM-2402; ASTM-B633; ISO-2081 Ed

Cyanide; acidor neutralchloride

A. Good atmospheric corrosionresistance in industrialenvironmentB. Higher melting point thancadmium (787°F)C. Good thickness distributionD. Good coverage in recesses

E. Poor corrosionresistance in marineenvironmentF. Less lubricity thancadmium

Low (lessthancadmium)

aIVD = ion vapor deposition; Ed = electrodeposition; El = electroless; Mech = mechanical.


For greater detail on the properties and method ofapplication of various finishes consult Refs. 48.

3Principles of Electroplating

Electroplating is a common coating technique used onboth industrial and commercial products and may bedefined as an electrochemical technique of applying a thinmetal coating onto another metal. A short list of familiaritems would include small appliance items such astoasters and electric coffee pots, automotive trim,plumbing fixtures, silverware, jewelry, copper-clad potsand pans, printing rolls, diamond drills, tin-plated steel,printed circuits, bearings, and the ubiquitous nuts andbolts.

To the uninformed individual, plating and metal finishingare often thought of as a low tech dip-and-swishoperation. But in reality they are complex processesinvolving almost every field of the physical sciences:inorganic, organic, electro-, physical, and analyticalchemistry; physics; metallurgy; mechanics; and electronics.

Platers often use and are exposed to some of the most


deadly and hazardous chemicals known; cyanides, causticsoda, and sulfuric, nitric, and hydrofluoric acids, andothers, as shown in Table 4. A good plater must knowdetailed techniques, learned and acquired from years ofon-the-job experience, especially for manual operations.

3.1Mechanism of Electrodeposition

A simple example of an electroplating process is nickelplating. A standard nickel plating solution consists ofapproximately 48 oz/gal of nickel sulfate; 6 oz/gal ofnickel chloride, and 5 oz/gal of boric acid, with the pHadjusted to 3.03.5 and operated at a temperature of 130°F.

Nickel sulfate and nickel chloride, when dissolved inwater, form positively charged nickel ions and negativelycharged sulfate and chloride ions, respectively. The articleto be plated serves as the cathode, and the anode isnormally a slab of nickel metal placed opposite theworkpiece. The workpiece or cathode is electricallyconnected to the negative terminal of a 612 V source ofdirect current electricity such as a silicon diode powersupply. The anode is connected to the positive terminal.

When the electric circuit is closed, an electric field iscreated between the cathode and anode, with the cathode


becoming charged with electrons. The positively chargednickel ions are attracted to, and migrate to, the cathode,where they plate out as nickel metal.

Meanwhile, at the nickel anode nickel atoms are eachgiving up two electrons and are going into solution asnickel ions. This essentially maintains a constant nickelconcentration in the plating solution. The stream ofelectrons given up by the nickel atoms at the anode flowthrough the power supply to the cathode, where they arereunited with nickel ions plating out as nickel metal. If anickel ion is not readily available for the electrons theymay reduce water to hydrogen gas and a hydroxide ion.The following equations summarize the cathode andanode reactions.

Cathode reaction:

Anode reaction:


TABLE 4 Electrochemical Data for Plated MetalsDensityb Electrochemical equivalents

Metal, symbol Atomicweighta Valence g/cm3 g/in.3 g/(A·h)doz/(A·h)lb/(1000 A·h)A·h/ft2 to deposit 0.001 in.Typical plating solution

Cadmium, Cd112.41 2 8.65 141.72.097 0.07400 4.62 9.74 Cyanide, acid sulfate; fluoborateChromium,Cr 51.996 6 7.19 117.80.3234 0.01141 0.813 52.47 Chromic acid

Cobalt, Co 58.933 2 8.90 145.81.099 0.03878 2.42 19.10 Acid sulfateCopper, Cu 63.546 1 8.96 146.82.371 0.08364 5.23 8.92 Cyanide

2 8.96 146.81.186 0.04182 2.61 17.83 Acid sulfate; pyrophosphateGold, Au 196.967 1 19.3 316.37.350 0.2592 16.2 6.20 Alkaline, neutral, or acid cyanideHydrogen, H 1.00794 1 0.0899c 0.03761 0.42e Acid, alkaline, or neutralIndium, In 114.82 3 7.31 119.81.428 0.0504 3.15 12.08 Cyanide; acid sulfate fluoborateIron, Fe 55.847 2 7.86 128.81.042 0.0368 2.30 17.80 Acid chlorideLead, Pb 207.2 2 11.4 186.83.866 0.1364 8.52 6.96 FluoborateNickel, Ni 58.693 2 8.90 145.81.095 0.0386 2.41 19.18 Acid sulfate; chloride; sulfamatePalladium, Pd 106.42 2 12.0 196.61.985 0.0700 4.38 14.26 Alkaline amine dinitritePlatinum, Pt 195.08 4 21.4 350.71.820 0.0642 4.01 27.75 Amine, nitrito acid chloride

Rhodium, Rh 102.906 3 12.4 203.21.280 0.0451 2.82 22.86 Acid sulfate;phosphate

Silver, Ag 107.868 1 10.5 172.14.025 0.1420 8.87 6.16 CyanideTin, Sn 118.71 2 7.30 199.62.215 0.0391 4.88 7.78 Acid sulfate

4 7.30 119.61.107 0.0391 2.44 15.56 Alkaline stannateZinc, Zn 65.39 2 7.14 117.01.220 0.0430 2.69 13.81 Cyanide acid chlorideaFrom International Table of Atomic Weights (1989).bAt 300 K, 27°C, 81°F. Density of deposited metal may differ significantly from values shown because of porosity or occluded organic material.Values given in g/in.3 were calculated from values in g/cm3.cIn g/L.dAssumes 100% cathode efficiency.eIn L/(A.h) at STP.


3.2Faraday's Law

Equation (1) shows that each nickel ion, in order to deposit as a nickelatom, accepts two electrons at the cathode surface. Therefore, 1000 nickelions require 2000 electrons, 1 million nickel ions require 2 millionelectrons, etc. In other words, the amount of nickel deposited is in directproportion to the current that flows through the anodecathode circuit. Thisrelationship was enunciated in 1833 by Michael Faraday as one of twobasic laws of electrolysis. These laws, incidentally, were the firstquantitative demonstration of the electrical nature of matter and have longdefined the unit quantity of electricity.

Faraday's first law states that the quantities of substances set free at theanode or cathode are directly proportional to the quantity of electricity thatpasses through the solution. The quantity of electricity is equal to theproduct of amperes and the time for current flows. Therefore, it is possibleto deposit twice the thickness if either the time of plating is doubled or theamperes are doubled.

The second law, which today's plating industry uses to determine therequired current and time to produce a given deposit thickness, states thatthe amount of different substances liberated by the same quantity ofelectricity are proportional to their chemical equivalent weights. To depositone equivalent weight (atomic weight ÷ valence) requires 1 faraday (F) ofelectricity or 96,485 coulombs (C). One coulomb of electricity is equal to1 ampere (A) flowing for 1 second (s). One faraday of electricity, therefore,is equal to 96,485 ampereseconds (A.s), 1608 ampere-minutes (A.min), or26.8 ampere-hours (A.h). Platers normally use ampere-hours for largeamperage processes such as barrel plating and ampere-minutes for smallmanual plating tank operations.

To calculate the amount of metal deposited for any amperage and time,


Faraday's law is used in equation form:

where

Wt =weigth of metal depositedAW

=atomic weigth of metal being deposited

V =valence (number of electrons required to deposit one atom)i =current flowing through the plating solutiont =time that current flows (s, min, h)e =cathode efficiency (% of current that deposits the metal)

F =Faraday's constant = 96,485 A·s, 1608 A·min, or 26.8 A·h perequivalent weight

The thickness of the deposit (T) is then calculated as

where

D=

density of metal(g/in.3)

A =area of parts (in.2)

Combining Faraday's law equation and the thickness equation allows theplater to calculate the number of ampere-hours required to plate a giventhickness on a load of parts:


where

Metal Metal factor(A·h/(ft2·in.))

Cadmium 9,737Silver 6,180Nickel 19,180Zinc 14,300

3.3Cathode Efficiency

During the plating process, the solution chemistryin the cathode film, the 0.0050.010 in. thick filmin immediate contact with the cathode, controlsthe plating process and the nature and properties of


the metal deposited.

A distinguishing feature of the cathode film is itsmetal concentration gradient, which is the range inthe bulk solution concentration from the furthestfrom the cathode down to 2560% of that valuenearest the cathode depending on the currentdensity. The nickel ions diffuse through thecathode film, accept two electrons each at thecathode, and deposit as nickel metal.

If too much current is applied to the cathode, therewill be more electrons available than there arenickel ions diffusing through the cathode film toaccept them. The excess electrons will then reactwith water molecules either to release nascenthydrogen, which can penetrate the nickel and basismetal crystalline structures, or form hydrogen gasbubbles, which are released into the platingsolution.

As long as sufficient nickel ions are diffusingthrough the cathode film to accept the available


electrons, very little hydrogen is evolved. Whenhydrogen is evolved, the amount of metaldeposited is less than the amount that shouldtheoretically have deposited. The ratio of theamount of metal deposited to what theoreticallyshould have deposited is known as the cathodeefficiency:

In our nickel plating example, the highest cathodeefficiency and the least amount of hydrogenreleased are achieved by (1) keeping the nickelmetal above a minimum concentration in the bodyof the solution, (2) heating the solution to increasethe rate of diffusion of the nickel ions through thecathode film, (3) agitating the solution to decreasethe thickness of the cathode film, and (4) keepingthe current below the limiting current density forthe particular solution chemistry involved so thatonly enough electrons are being suppled to thecathode to plate out the nickel ions arriving to


accept them.

3.4Current Density

The number of amperes applied to a platingsolution through the cathode-anode circuit mustbe controlled and adjusted according to thesolution chemistry and the surface area of thework


being plated. Current density is the amperesflowing through the plating solution divided bythe surface area being plated and is usuallyexpressed as amperes per square foot.

Each type of plating solution requires thatcurrent density be within a particular range toproduce a quality deposit. This range of currentdensities increases or decreases as theconcentrations of the plating ingredients change.Usually the metal concentration is thecontrolling factor. As the metal concentrationincreases, a higher current density can be usedwithout deleterious effects. If the current densityexceeds its limiting value for a given solutionformulation, the amount of evolved hydrogenrapidly increases. In the case of nickel plating, the


pH in the cathode film increases to the pointwhere nickel hydroxide precipitates on thecathode, resulting in a ''burned," granular,defective deposit.

3.5Throwing Power

Ideally, the thickness of plating on a fastenershould be uniform. Unfortunately, platingthicknesses are not uniform over the surfaces ofa bolt or nut, or any other item for that matter.The plating tends to be thickest on prominentprojecting surfaces such as the point and head ofthe bolt and the outside of a nut, and the thinnestdeposits are found in recessed areas of a boltsuch as the shank-to-head fillet or the internalthreads of a nut. This ability or inability of aplating solution to plate a deposit of uniformthickness over the entire surface of a complexshape is known as its throwing power.

Cyanide and alkaline solutions tend to have


better throwing power than acid solutions.Chromium plating solutions have the leastthrowing power of all plating solutions. Thecause of good or bad throwing power iscomplex, but it is known to be influenced by theohmic resistance of the plating solution, whichdetermines the primary current distribution; thegeometry of the plating system; the shape of thepart; and the position of the part in the platingsolution relative to other parts and to the anode.

Polarization, or a reverse EMF, occurs in thecathode film at high current density areas andshifts the plating current to more recessed areas,thereby depositing more metal in those areas thanwould have been deposited with the primarycurrent distribution.

Throwing power is improved by using loweroperating temperatures, lower metalconcentrations, and lower current densities andby avoiding agitation of the solution.


Mechanically, throwing power may be improvedby keeping the anode as far away from the workas possible, by arranging parts so that their highcurrent density areas are adjacent to each other,and by using auxiliary "dummy" cathodes nearhigh current density areas to accept some of thecurrent.

3.6Solution Composition

Plating solutions can be broadly classified asacid, neutral, or alkaline. Solutions with a pH ofless than 5 can be considered acid; from 5 to 9neutral; and above a pH of 9, alkaline.

Acid and neutral solutions, except in the case ofacid cyanide gold, use either simple inorganicmetal salts such as sulfates or chlorides orcomplex inorganic salts such as fluoborates,pyrophosphates, or sulfamates. The majority ofalkaline solutions for plating cadmium, zinc,silver, copper, and gold contain cyanide


complexes. Tin is plated from non-cyanidecontaining alkaline stannate solutions.

Several variables must be controlled to produce ametallurgically sound deposit duringelectroplating: the solution chemistry, solutiontemperature, current density, and, to a lesserdegree, the basis metal.


Chemicals added to a plating solution serve atleast one of the following functions:

1. As a source of the metal being deposited.

2. To improve conductivity.

3. To provide anode corrosion.

4. To provide brightness or a smoother, finergrained deposit.

5. To buffer or maintain the proper pH.

In the case of our nickel solution example, thenickel sulfate supplies the nickel metal and alsothe conductivity because of its relatively highconcentration. The nickel chloride, besidessupplying some nickel, is added primarily toprovide chloride ions to promote anodecorrosion. It also contributes to conductivity.The boric acid acts as a buffer to minimize rapid


pH changes. Frequently a wetting agent such assodium lauryl sulfate is added to prevent pittingfrom hydrogen gas bubbles adhering to the platedsurface.

To achieve a decorative bright nickel deposit,several proprietary organic chemicals would beadded. Without these "brighteners," the nickeldeposits are smooth and fine grained but mattegray in color.

For a very low tensile stressed and ductile nickeldeposit, a sulfamate formulation is used insteadof the sulfate solution. If a high rate ofdeposition is required, a fluoborate formulationcan be used.

Although it is important that plating solutionshave closely controlled concentrations of majoringredients, it is also important that they be freeof certain impurities peculiar to each platingsystem. For example, all plating solutions, exceptchromium solutions, are sensitive to hexavalent


chromium, even in the parts per million range.Hexavalent chromium drastically reducescathode efficiency and in high enoughconcentrations will prevent depositionaltogether. Other solutions such as the alkalinetin solution must have all their metal ions of thesame valence. The tin must be present as thestannate (4+) without any stannite (2+) in orderto produce smooth, fine-grained deposits.

Just as small amounts of impurities aredeleterious in most plating solutions, smallamounts of certain chemicals are necessary insome plating solutions to deliver sound deposits.Chromium plating solutions, both those used fordecorative deposits and those used for thickdeposits, must contain a small amount of sulfate,usually in a concentration of 1% of the chromicacid concentration. Above and below this range,cathode efficiency rapidly decreases. In brightacid copper plating solution, chloride in a rangeof 2080 ppm is necessary to achieve the


maximum effect of the organic brighteners.

3.7Addition Agents

In addition to their major and minor inorganicingredients, most plating solutions must alsocontain a small amount of an organic material orchemical to produce an acceptable deposit. Thesematerials are referred to as addition agents andsometimes as brighteners.

Even though these additives are added to platingsolutions in small quantities and do not take partin the electrolysis (although they may beabsorbed into the deposit, oxidized at the anode,or reduced at the cathode) they have a profoundeffect on the character of the deposit. They refinethe crystal structure, resulting in denser, lessporous, and brighter deposits. More important,especially in cyanide solutions, they greatlyincrease throwing power and allow the use ofhigher current densities.


As an example of the intense effect of minuteamounts of addition agents, glue was once usedto prevent treeing in acid copper plating at aconcentration of only 3 mg/L (3 ppm). Althoughnot in common use today, glue is still included insome lead fluoborate plating formulations, but atthe much higher concentration of 200 mg/L.

Addition agents or brighteners wereserendipitously discovered by an accidentalcontamination of a silver cyanide solution withcarbon disulfide in 1843. This chemicalcombination is still used today by some platers.


It is interesting to note that the early developmentof brighteners centered on the small additions ofcommon organic chemicals found around thehome, especially the kitchen. As an example, thefirst brightener for cadmium cyanide solutionswas a caustic solution of dissolved wool in1922. During this period such common organicsas dextrin, gelatin, milk sugars, Instant Postum,molasses, and sulfonated castor oil were oftenused by cadmium platers.

Currently used commercially availableproprietary brighteners are materials whoseactive ingredients are identifiable chemicals withdefinite chemical structures and aremanufactured or refined by closely controlledprocesses.

4Plating Procedures


The essential features of a typical platingprocedure include

1. Preplating preparation

2. Electroplating

3. Postplating treatment

The preplating preparation phase removes anysoil, dirt, scale, or oxides that would interferewith the adhesion of the subsequent coating. Thepreplating step may consist of one or more, orall, of the following: vapor degreasing, chemicaldescaling, mechanical scale removal such asabrasive blasting, soak or electrolytic cleaning,and finally an activation acid drip.

The electroplating phase consists of immersingthe parts to be plated into the electroplatingsolution, positioning them properly between theanodes for optimum plating distribution,adjusting the electric current, and measuring thetime to deposit the required thickness. The parts


are then removed and water rinsed to remove anyresidual plating solution drag-out.

The postplating phase may consist, in the case ofcyanide cadmium plating, of a mildneutralizing/passivating dip in a weak chromicacid solution followed by one or more coldwater rinses and a final hot water rinse tofacilitate drying. The part is then baked at 375°Ffor 23 h to drive off any absorbed hydrogen. Thechromate conversion coating, if required, and anyother subsequent coating such as a wax ormolybdenum disulfide dry film lubricant, arethen applied. A typical cadmium plating cycle isoutlined below.

Step 1. Vapor degreasing: Removes bulk ofgrease, oil, and shop dirt prior to subsequenttreatments.

Step 2. Rust and scale removalAbrasive blastwith glass beads (no base metal removal) orfine aluminum oxide: Removes tenacious heat


treating scale and corrosion products.

Step 3. Soak or anodic alkaline cleaning:Removes last traces of soil. Produces chemicallyclean surface to promote good adhesion of thedeposit.

Step 4. Cold water rinse: Removes alkalinechemicals from parts.

Step 5. Acid dip (when permitted byspecification): Activates surface by removingoxides produced by alkaline cleaning. Lowconcentration, 30 s maximum. Promotes goodadhesion.

Step 6. Cold water rinse: Removes acid frompart.

Step 7. Cadmium plate: Depending on materialstrength, the solution is a cyanide solution eitherwith or without brighteners.

Step 8. Cold water rinse: Removes platingsolution from parts.


Step 9. Chromate dip: Removes or neutralizesresidual plating solution.

Step 10. Cold water rinse: Removes chromatesolution.

Step 11. Baking: Bake at 375°F for 4 or 23 h toremove hydrogen absorbed during plating.

Step 12. Chromate conversion coating: Whenrequired. Produces an iridescent or olive drabchromate coating that improves corrosionresistance.

Step 13. Cold water rinse:Removes residualchromate solution from part.


Step 14. Warm water rinse: Facilitates drying ofpart. Used with compressed air blow-off forlarge parts or a centrifugal spin dryer for smallparts.

Step 15. Inspection: Check for blisters, pits,nodules, adhesion, thickness, and salt spray.

These steps may not all be necessary in all platingprocedures.

The flowchart in Figure 1 illustrates this (waterrinses not shown).

4.1Rack or Barrel Plating?

Fasteners may be plated either in bulk in arotating barrel or individually on a fixture orrack. The decision to barrel or rack plate maydepend on the quantity of parts being plated,economics, part size, and the probability of


thread damage.

Generally, precision high strength bolts 3/8 in. indiameter or less by approximately 2 in. or less inlength are barrel plated. Beyond these limits,there is buildup beyond dimensional limits on theends before the minimum thickness is achieved atthe center of the part. But a more critical reasonfor restricting barrel plating to small bolts is thatthe larger the diameter and length, the morelikely it is that parts will suffer thread damagefrom handling and peening over of the threadcrests from prolonged tumbling.

The above restrictions are not generally ofconcern with lower grade fasteners. Theeconomics and vast volume indicate that they bebarrel plated.

Rack plating, although costly, is necessary forlarger diameter and longer precision fastenerswith close dimensional tolerances. Platingbuildup on the lead thread can easily be


prevented by making electrical contact at thethread end for part of the plating cycle. Normallymidway through the plating cycle the bolts arerepositioned on the rack by rotating 90° or 180°depending on the type of rack. This permits areasthat are low current density areas, and that haveminimum plate thickness, to become high currentdensity areas, making the final thickness moreuniform over the entire surface of the bolt.

Most people, including engineers, think thatelectroplating is a simple dip or dunk process. Itdefinitely is not. It is a science that employsvarious aspects of chemistry, physics,electronics, and electrochemistry along withmechanical, plumbing, and electrical tradesknowhow, combined with a great deal ofcommon sense and a bit of art.

Electroplating is a unique manufacturingoperation. Along with heat treating,electroplating is one of two manufacturingtechniques of special concern to users and


purchasers of high strength precision fasteners.

While other fastener manufacturing operationssuch as drilling, forging, tapping, grinding, andthread rolling impart obvious and easily verifiedshapes and dimensions, heat treating andelectroplating induce intrinsic properties thatallow a fastener to function as intended underspecific conditions and environments.

4.2Arc Burns

A common problem caused by electroplating isarc burns. An arc burn occurs, almost exclusivelyduring barrel plating, when there is too littlecontact area between the electrical connectorsinside the barrel and the bolts or nuts beingplated.

In the case of steel fasteners, when arcing occurs,the temperature of the burned area momentarilyreaches a temperature above the transitiontemperature at which martensite converts to


austenite. As the austenite cools it transformsback to untempered martensite. Theseuntempered martensite spots are areas that areconducive to fatigue or stress fractures. Tominimize the possibility of arc burning, theelectrical contacts inside the barrel should haveas large a contact area as possible. Rather thanuse the standard fist-size knob contacts at the endof the limp electric cables, large diameter discswith a large surface area should be used. Themore parts in contact with the discs, the lesslikelihood that arc burns will occur.


Figure 1A flowchart for the cadmium plating cycle.

Arc burns can be detected by stripping the plating from a sample of barrel-plated parts and using the standard nitol test for untempered martensite.

5Dimensioning Fasterners for Plating

There are few things as useless as a fastener that doesn't fit, being eitheroversize or undersize for the mating threads. Manufacturing unplatedfasteners to close dimensional limits is not difficult, but if the fastener is tobe plated and still meet the final blueprint dimensions it must bemanufactured unsize to allow for the buildup due to the added platingthickness.


5.1Tolerance

Tolerance is defined as the amount of variation permitted in the size or dimensions of a part. Ifa part is to be plated, there is not only a tolerance allowed in the size of the part before plating,but because the thickness of plating cannot be deposited exactly, a plating thickness tolerance isalso necessary. Plating thickness tolerances are necessary because the plating thickness willvary on the same part from point to point, the average thickness on each part will vary frompart to part, and finally the average thickness of each batch plated will vary from batch to batch.

If 0.0003 in. of cadmium plate is applied to only one side of a panel, then the dimension of thepanel will increase by 0.0003 in.; but if 0.0003 in. of cadmium is plated on both sides of apanel, the panel's dimensions will increase twice 0.0003 in., or 0.0006 in. total. When a threadis plated with 0.0003 in. of cadmium, the pitch diameter increases by 4 × 0.0003 in., or 0.0012in. These three examples are illustrated in Figure 2.

As shown in Figure 2, a plating thickness of 0.0003 in. increases the pitch diameter of thethread by 0.0012 in. because of the 30°/60°/90° triangle generated by the thickness of platingand the increased distance from the original unplated diameter to the pitch diameter of


Figure 2The effect of plating thickness on thread dimensions.


the plated thread. The hypotenuse, the increase from the unplated pitchdiameter, is twice the side opposite the 30° angle, the thickness of theplating, or 0.0006 in. Since this increase also occurs on the oppositeside of the thread, the total increase in pitch diameter is 4 times 0.0003,or 0.012 in.

5.1.1Establishing Tolerances

Tolerances for multiple manufacturing processes are additive. To goback to the example of plating a single surface of a panel, the totaltolerance of the plated panel is

or, transposing,

Example 1: If the before-plate dimensions of the panel are allowed tovary from 0.1000 to 0.1010 in., a tolerance of 0.001 in., and the platingthickness is held to 0.00020.00004 in., a tolerance of 0.0002 in., thenthe total tolerance of the plated panel is

However, if the plated panel must be held to the original unplatedtolerance of 0.001 in., the tolerance of the panel before plating must beheld to 0.0008 in.


Example 2: If a bolt body is plated on both sides, with the before-plateand plating tolerances as given at the beginning of Example 1, the totalafter-plate tolerance is now

However, if the after-plate tolerance must be maintained at 0.001 in. asin the first example, then the dimensional tolerance of the bolt bodypanel must be reduced to 0.0006 in. If the tolerance on the bodydiameter cannot be reduced that much, then both the preplate diametertolerance and the plating thickness tolerance can be reduced 0.0001 in.for a thickness of 0.00020.0003 in., or 0.000250.0003, 5 or0.00030.0004 in. or some combination that will satisfy the final 0.001in. tolerance. For example,

When plating threads, the tolerance equation is

or

The amount of preplate tolerance for threads requiring plating increasesas the nominal thread size increases. This increased tolerance is neededbecause lead error and thread drunkenness, caused by long thread rolldies, is more prevalent as the thread size increases. Some of thesetolerances are listed in Table 5.

5.1.2External Dimensions


The maximum material condition or maximum after-plate dimensionoccurs on parts that are at the maximum before-plate dimension andreceive the maximum specified plating thickness.


TABLE 5 Final and Preplate Pitch Diameter Tolerances for Some TypicalUNF ThreadsNominal threadsize

Final PDa(in.)

Final tolerance(in.)

Preplate PDa(in.)

Preplatetolerance (in.)

#640 0.11980.1218 0.002 0.11090.1202 0.00121/428 0.22430.2268 0.0025 0.22350.2252 0.00173/824 0.34500.3479 0.0029 0.34420.3463 0.00211/220 0.46430.4675 0.0032 0.46350.4659 0.00243/416 0.70560.7094 0.0038 0.70480.7078 0.0030112 0.94140.9559 0.0044 0.94070.9543 0.00361 1/212 1.44111.4459 0.0049 1.44031.4443 0.0040aPD = pitch diameter. These dimensions may not be valid for all fastenermanufacturers. The minimum or maximum preplate sizes may be greater orless to compensate for subsequent manufacturing operations. For example,the minimum preplate PD may be increased by 0.004 in. to compensate for0.0001 in. of scale formed during heat treatment. When this scale is removedduring the preplate cleaning and descaling operation the minimum PD justprior to plating will have decreased to that shown in the table.

Conversely, the minimum material condition or minimum after-plate dimension occurson parts that are at the minimum before-plate dimension and receive the minimumspecified plating thickness. The key words above are "receives the . . . specified platingthickness." If parts are within the required before-plate dimensions and the part receiveseither more or less than the specified thickness, then the parts will be dimensionally out oftolerance.

The rules, then, for calculating the before-plate dimensions are similar to those citedabove for determining tolerances, except that two calculations are made: one to determinethe maximum and one to determine the minimum dimension.

For the bolt body maximum before-plate diameter, subtract twice the maximum specifiedthickness from the maximum after-plate diameter. For the minimum bolt body before-plate diameter, subtract twice the minimum specified thickness for the minimum platediameter.

Similar calculations are made to determine the preplate pitch diameter (PD) dimensions of


bolt threads, except that four times the plating thickness is subtracted from the after-platepitch diameter dimensions.

5.1.3Internal Preplate Dimensioning

Calculating preplate dimensions for internal surfaces is not just the reverse of that forexternal surfaces. The minimum metal condition after plating occurs when the hole orpitch diameter is the greatest (largest hole) and has the minimum plate thickness applied.The maximum metal condition exists when the hole or pitch diameter is the least (smallesthole) and has the maximum plate thickness.


For holes:

For pitch diameters:

Note: Some manufacturers use a large maximum pitch diameter before plating to compensatefor the thicker plating buildup on the first thread at each end of the nut. Plating thickness ismeasured on the bearing surface of the nut per MIL-STD-1312 method.

5.2Plate Thickness Distribution

Electrodeposited coatings vary in thickness from area to area on a bolt or a nut. In general thegreatest thickness is found on the head and point of a bolt and on the outside of a nut, and thethinnest deposit is found on the middle of a bolt and on the inside of a nut. Figure 3illustrates typical plating thickness distributions on several sizes and lengths of bolts plated to0.0002 in. minimum according to QQ-P-416.


Figure 3Cadmium plating thickness distribution on 12-point fasteners.


Pitch diameter preplate dimensions for nonelectroplated coatings, suchas vacuum deposition of aluminum or cadmium or mechanicaldeposition of zinc, cadmium, or tin, must be determined by trialbecause the amount of metal deposited on the pitch diameter may beonly one-half that deposited on the shank.

All of the above recommendations and equations for dimensioningbolts and nuts apply to parts with 60° threads. For parts with a 75°thread, the pitch diameter increases by almost 8 times the platingthickness, or twice that of a 60° thread.

6Effect of Coatings on Fastener Performance

All fastener coatings affect fastener performance in some respect,although some effects may arise not from the physical presence of thecoating but from the method of application. Some of these effects arediscussed below.

6.1Hydrogen Embrittlement

Electroplating is unique in its ability to deposit a variety of thin,adherent uniform metallic coatings to suit the fastener's intendedapplication and environment. However, it has a serious side effect. Theelectroplating process can cause hydrogen embrittlement in highstrength fasteners.

Hydrogen embrittlement can be defined as delayed brittle failure duringstatic loading at stress levels well below the tensile strength of thematerial due to the presence of absorbed hydrogen. The process ofhydrogen embrittlement and the role of contributing factors are


complex. The degree of embrittlement, whether slight, moderate, orsevere, is related to the amount of hydrogen absorbed and retained inthe basis metal during manufacturing and chemical processing.

The many factors affecting the degree of hydrogen embrittlement inhigh strength fasteners are grouped in chronological order ofoccurrence in Table 6.TABLE 6 Factors Affecting the Degree of Hydrogen EmbrittlementFactors prior to plating1. Type, quality, and strength level of the material2. Microstructure of the material3. Surface condition of the material4. Mechanical, chemical, and electrochemical surface treatments prior to platingFactors during plating1. The metal deposited2. Nature of the deposit3. Thickness of the deposit4. Type and composition of the plating solution5. Addition agents and impurities present in the plating solution6. Temperature of the plating solution7. Voltage and current density during plating8. Length of plating timeFactors after plating1. Temperature and duration of the postplating embrittlement relief bake2. Stripping of defective deposits3. Corrosion while in service


Much of the practical research into the causes ofhydrogen embrittlement and its minimizationwas done during the period 19501970 by themilitary and the aerospace industry because ofthe increased use of high strength and ultrahighstrength materials during that period.

In the light of today's sensitive embrittlement testtechniques, some misleading conclusions weredrawn from some of the earlier researchprograms because of the use of some relativelyinsensitive stress methods such as bend tests andreduction of area tensile tests, which could detectgross embrittlement but not slight embrittlement.

It was not until the universal adoption of notchedtensile specimen testing that test data could bereliably compared with the results of otherresearch. However, even notched tensilespecimen data can vary widely because ofdifferences in material batches or heats, the notch


sharpness or Kt, stress concentration factor, andthe methods or techniques for manufacture andpreparation of the specimens.

From reported research on hydrogenembrittlement, the following general commentsappear to be valid concerning the variables listedin Table 6. Most of the emphasis of theseresearch efforts was concerned with cadmiumplating because it is commonly used as asacrificial corrosion-preventive metal coating bythe aerospace industry and the military.

The degree to which a material can be embrittleddepends on the intensity and longevity of theembrittling conditions, but the indicated degreeof embrittlement is determined by the type andseverity of the embrittlement test and the criteriachosen to indicate freedom from embrittlement.

6.2Factors prior to Plating

6.2.1


6.2.1Type, Quality, and Material Strength

Most iron-base alloys and steel are susceptible tohydrogen embrittlement but to different degrees.The susceptibility to embrittlement increases asthe strength of the material increases andgenerally manifests itself at strength levels at orabove approximately Rockwell hardness C-30[3].

Materials of the type AISI 4340 heat treated tohigh strength levels are the most susceptible tohydrogen embrittlement. This susceptibilityvaries among different production heats [4].

High strength alloy steels containing highpercentages of silicon [5,6] or tool steels similarto AISI H-11, which contains 5% chromium,show a relatively low degree of embrittlement[68].

Vacuum-melted steel has a lower susceptibilityfor hydrogen embrittlement than airmelted steel


[1].

High temperature alloys of the type Inco 718(180 ksi), Waspalloy (190 ksi), René 41 (200ksi), and U212 (180 ksi) have such a lowsusceptibility to hydrogen embrittlement thatthey do not fail when cathodically charged whilebeing stressed to 80% of their yield strength [2].

6.2.2Material Microstructure

Austenite has a higher solubility for hydrogenthan either ferrite or martensite and thereforetends to concentrate in retained austenite in steel.Baking does not remove this hydrogen unless thebaking also transforms the retained austenite.The hydrogen in the retained austenite does notcause embrittlement. However, if the austenitedecomposes to untempered martensite in serviceand releases the hydrogen, the steel may beembrittled.

6.2.3


Surface Conditions of Steel

Rough surfaces, because of the greater surfacearea, will absorb more hydrogen than a smoothsurface. Surface carburization produces a highsurface hardness and high surface stresses thatare conducive to embrittlement failure.


6.2.4Mechanical, Chemical, and ElectrochemicalSurface Treatment prior to ElectroplatingMechanical

Mechanical surface treatments such as colddrawing, grinding, and polishing have a tendencyto increase the susceptibility to embrittlement.These effects can be relieved by retemperingwithin 50°F of the original temperingtemperature or by anodic etching to remove thestressed surfaces.

Shot peening has been shown to be beneficialprior to cyanide cadmium plating.

Dry abrasive blasting, rather than wet vaporhoning, should be used to remove heattreatedscale because hydrogen is generated by thereaction of water with the oxide-free ironsurface.


Chemical

The degree of embrittlement from acid treatmentis influenced by the type of acid, concentration,temperature, and length of treatment.

Highly ionized acids such as hydrochloric (HCl),sulfuric (H2SO4), and hydrofluoric (HF) acidscause embrittlement because of increasedabsorption of hydrogen by materials. Lessionized acids such as phosphoric (H3PO4) causeless embrittlement.

Cadmium fluoborate plating according toNAS672 permits the use of a 10 s maximum dipin a 1.5% concentration of either nitric, sulfuric,or fluoboric acid for activation of bolts just priorto cadmium plating.

In general, the use of inhibitors in picklingsolutions reduces the amount of hydrogenabsorbed during pickling. However, they are notused for acid activation immediately prior to the


plating operation because the tenacious inhibitorfilm will interfere with adhesion of thesubsequent deposit.

Electrochemical

Alkaline cleaning either by immersion or byanodic electrocleaning does not causeembrittlement, but cathodic electrocleaning isseverely embrittling.

Thin (0.000125 in.) undercoats of bright cyanidecadmium, electroless nickel, gold, orpyrophosphate copper plus a bake have beenreported to minimize embrittlement from cyanidecadmium plating.

6.3Factors during Plating

6.3.1Type of Metal Deposited

All electroplating processes appear to causehydrogen embrittlement to some degree. Whether


or not a particular plating process causesembrittlement depends on the ability of thehydrogen absorbed during the plating process todiffuse out of the basis metal and through theplated metal during the post-plate bakeoperation.

Because hydrogen has a low solubility incadmium, zinc, and gold, it diffuses so slowlythrough dense deposits of these metals thatoutgassing of hydrogen cannot occur during thepost-plate bake to prevent hydrogenembrittlement. However, embrittlement due tochromium and nickel plating can be relieved witha suitable post-plate bake, because hydrogen candiffuse through these metals.

6.3.2Nature of the Deposit

A porous crystalline deposit of cadmium, platedfrom a cadmium cyanide solution without anyorganic brighteners, allows hydrogen to outgas,


while a bright dense deposit plated from a similarsolution containing organic brighteners causessevere embrittlement.


6.3.3Deposit Thickness

The thinner the deposit, the shorter the time forhydrogen to diffuse out through the coating. Amethod based on this concept was proposed thatconsisted of plating a thin layer of cadmium,approximately 0.0001 in., followed by a shortbake to outgas the hydrogen, and finally followedby plating to the required total thickness ofcadmium. But more sensitive embrittlement testsshowed that this method also causedembrittlement.

6.3.4Type and Composition of the Plating Solution

Cadmium is severely embrittling when platedfrom a cyanide solution but much less so whenplated from a fluoborate [1], sulfamate, orperchlorate solution. The cyanide ion acts as a


catalyst, promoting the absorption of hydrogeninto steel. Cadmium solutions with complexingagents such as pyridine, triethanolamine,methanol, and dimethyl formamide show a lowtendency to cause embrittlement.

Although seemingly contradictory, acid platingsolutions, with the exception of chromium,cause less embrittlement than alkaline solutions.

6.3.5Plating Solution Addition Agents and Impurities

Oxidizing agents such as hydrogen peroxide,nitrates, and titanium compounds have beenadded to cadmium cyanide solutions to reducethe degree of embrittlement. These materialsoxidize the nascent hydrogen before it candiffuse into the basis metal.

Organic brighteners or grain refiners and organiccontaminants produce dense cadmium depositsthrough which hydrogen cannot diffuse duringthe post-plate bake.


Sulfide has been a suspected impurity inunbrightened cadmium cyanide solutions, whileiron contamination nullifies the beneficial effectof titanium.

6.3.6Plating Solution Temperature

Temperatures in the range of 6090° for both zincand cadmium cyanide plating produce lesshydrogen than elevated temperatures.

6.3.7Voltage and Current Density

A reduction of over 1000 times in the hydrogeninput rate can be achieved by lowering thevoltage across the plating tank to 1 V maximum.In unbrightened cyanide cadmium, platingcurrent densities above 40 A/ft2 produce aporous deposit, allowing for a more efficientoutgassing of hydrogen during the post-platebake.


6.3.8Plating Time Duration

A shorter cathodic treatment such as platinglimits the amount of hydrogen migrating into thebasis metal. Also, higher current densities resultin shorter plating time for equal thicknessbecause most hydrogen is absorbed during theinitial part of the plating period.

6.3.9Current Fluctuations

Low current ripple is beneficial in hydrogenabsorption, and periodic reverse current is of novalue.

6.4Factors after Plating

6.4.1Temperature and Duration of the Post-PlateBake


Baking after plating relieves embrittlement bylowering the hydrogen concentration in the basismetal to a nonembrittling level in one or both ofthe following ways:


1. By outgassing the hydrogen through theplating

2. By dispersing the hydrogen uniformlythroughout the basis metal volume

For materials extremely sensitive to hydrogenembrittlement, relief must occur in the first way,while for materials less sensitive or of a lowerstrength level, relief is achieved in the secondway. In this mechanism the initial high hydrogenconcentration near the surface is reduced bydiffusion of the hydrogen into the body of thebasis metal to a concentration below thethreshold value necessary to causeembrittlement.

As the strength level of steels increase, the post-plate baking time must be increased. Formaterials below 190,000 psi tensile strength, a 3h bake at 375 ± 25°F is adequate. Above this


level, a 23 h bake is necessary. At strength levelsabove 280,000 psi, no amount of baking givesfull recovery.

The baking temperature should be as high aspossible without affecting the strength of thematerial or the integrity of the coating. Thebaking temperature for cadmium is generally375°F, whereas chromium can be baked at700°F and nickel as high as 1000°F.

6.4.2Stripping of Plating

The chemical stripping of metallic coatings canbe a source of hydrogen, especially if acidsolutions are used. A baking operation shouldfollow the stripping operation but precedereplating.

6.4.3Corrosion

When cadmium, zinc, or aluminum coatings on


steel corrode in service, hydrogen is liberated onthe steel surface at breaks in the coating wheresteel is exposed or at the steel/coating interfaceof porous deposits. This hydrogen can diffuseinto the steel, and if the part is under stress,delayed brittle failure can occur. Since brightcyanide cadmium is less porous than dullunbrightened cadmium there is less of a tendencyfor this type of failure with the former.

6.5Fatigue Strength

There are many factors other than plating thatinfluence fatigue strength, and they are coveredin another chapter of this book.

Most plated coatings and coating processesreduce the fatigue strength of steel bolts. Thedegree of decrease, or in rare cases increase,depends on the type and strength of the steel, theresidual surface stresses in the steel, themechanical and chemical properties of the


coating, the testing environment, and the methodof fatigue testing, whether by rotatingbending ortensiontension.

Any damage, no matter how slight, to theoriginal thread rolled surface will decrease thefatigue strength of a high strength bolt. A majorcause of surface damage is grit blasting or acidpickling to remove heat treat scale. The loss offatigue strength may be minimized bysubstituting alkaline descaling for acid picklingor by grit blasting using fine glass beads or finealuminum oxide, such as 350 grit, at as low anair pressure as possible.

The need for blasting or chemical descaling canbe eliminated if an inert atmosphere such asargon is used during heat treating and tempering;this will prevent scale formation. If blasting isnecessary, especially on large diameter bolts,automatic individual blasting machines will givea consistency of blasting time and coverage thatis impossible to achieve with a manual


operation.

6.6Stress Corrosion

It is important that high strength steel fastenersbe resistant to stress corrosion cracking.Electrodeposited nickel (0.0005 in.) or nickel(0.0003 in.) plus cadmium (0.0002 in.), eitherasplated or diffused, was twice as effective ascadmium (0.0005 in.) alone in preventing stress


corrosion failure in bolts of 290,000 psi tensilestrength loaded to 90% of their proportionallimit and exposed for more than 1000 h ofalternative immersion in 3.5% sodium chloridesolution.

The nickel, nickel-cadmium coatings were farsuperior to such coatings as vacuumdepositedcadmium, electrodeposited gold, or zincchromate primer. However, nickel plus SermeTelW, a phosphate chromate bonded aluminumcoating, surpassed almost all other coatingstested.

If resistance to stress corrosion cracking is ofprime importance as a bolt property, it is betternot to rely on plated coatings to provideresistance but to consider the use ofcorrosionresistant alloy bolts that require noprotective system.


6.7Stress Alloying

Cadmium-plated high strength steel shows somereduction in tensile strength, stress rupture, andfatigue strength at temperatures as low as 500°F.But at the melting point of cadmium, 610°F,there is a sharp decline in these properties.Therefore cadmium plate is limited for use onsteel parts that do not exceed 400°F in service.Cadmium plating is prohibited on titanium andtitanium alloy fasteners because stress alloyingor solid cadmium embrittlement can occur inservice at temperatures as low as roomtemperature.

6.8Corrosion Protection

Coatings that are electronegative or cathodic tosteel, such as nickel, chromium, or silver, protectsteel from corrosion by the barrier effect. Thesecoatings, being more corrosion-resistant than


steel, protect the steel by keeping the corrodantfrom contacting the steel. However, sincecoatings for fasteners are thin, in the range of0.00030.0005 in., some porosity is present. It isat these pores, or in mechanical scratches orscrapes if they are present, that a galvanic cell isproduced. These corrosion-resistant coatings,rather than protecting the steel, cause it tocorrode because the steel is now the anode in thegalvanic cell.

By contrast, coatings such as zinc, cadmium, andaluminum protect steel not only by the barriereffect but also by acting anodically to the steel.They will corrode in preference to steel atscratches and scrapes in the coating. However,this can be detrimental for high strength steels;because the steel will be the cathode in anycorrosion cell, hydrogen would be generated,which could diffuse into the steel and causehydrogen embrittlement.

6.9


Lubricity

Torque tension is one of the most importantfastener properties because it indicates thequantity of tightening or induced stress of thefastener and joint components for a givenamount of wrenching effort or energy. However,the amount of torque required to develop a giventension or load in a fastener depends on thefriction between mating surfaces of the bolt,joint, and nut. The fastener coating may eitherincrease or decrease this friction. Coatings suchas cadmium, copper, and silver will lowerfriction as is demonstrated in the use of thesemetals as lubricants during the manufacture ofnuts and bolts in various sizing and shapingextrusion processes.

One of two most popular plated coatings,cadmium requires less torque to develop aparticular preload than zinc. However, theadvantage of one over the other depends onwhether the plating is on the bolt or nut or both.


The well-known torquetension equation


where

T =torque, lb-ftK =friction factorD =bolt diameter, in.W =bolt tension, lb.

cannot be relied on for accurate torquetensiondeterminations because the handbook values ofK are only estimates. The true value of K must bedetermined for each particularfastenercoatingjoint combination at hand.

7Process Control

With electroplating it is not obvious whether theapplied coating has the specified properties ofthickness, adhesion, corrosion resistance, andlubricity or whether the plating process hasinsidiously caused hydrogen embrittlement.


These properties can be verified only by lengthyand costly and in some cases destructive testswith conclusions based on limited statistical dataand sampling. Accordingly, it is readilyrecognized that if good plating results are to beconsistently achieved, the plating process mustbe a precise and rigidly controlled operation.Many commercial job shop platers cannot berelied on to perform specified plating proceduresto meet current critical fastener platingspecifications. These specifications can be metonly by well-trained platers who follow andadhere to established, proven, documenteddetailed procedures, operating with process andplating solutions under the proper controls oftime, temperature, chemical composition, andcurrent density.

With ordinary industrial fasteners, the plater'schief concern is just to plate the minimumthickness. If plating builds up on the lead threadthere is still enough tolerance or allowance to


meet the specification gauge fit. But withprecision fasteners with narrow final tolerances,plating buildup is a problem. The plater mustmeet the minimum thickness on the body withoutbuilding up the lead thread beyond the acceptedgauge fit.

Close-tolerance plating can be accomplished by asystem of tight process control. The difficultywith which plating thickness can be held to a0.0002 in. tolerance increases as the minimumthickness increases. When the thicknessrequirement is 0.00020.0004 in., the maximumthickness is 100% higher than the minimum. Butwhen the thickness range is 0.00030.0005 in.,the maximum is only 66% greater than theminimum; and as the plating range moves to0.00040.0006 in., the maximum is only 50%greater than the minimum and requires extremelyclose process control. These controls areoutlined below.

7.1


Solution Control

The composition of all plating solutions shouldbe controlled to within ±5% of the setconcentration for each component. For example,a cadmium plating solution whose cadmiumconcentration is set at 4.0 oz/gal should not beallowed to vary beyond a range of 3.84.2 oz/gal.

All solutions should be analyzed as frequently asnecessary to prevent the individual componentsfrom varying beyond the set limits. The timeinterval may be once a week for solutions whereparts are plated on racks, the drag-in and drag-outis small, and production is low. However, theinterval between analyses could be as short as afew hours in the case of the pH of a barrel nickelplating solution where the ampere-hours pergallon value is high, drag-in and drag-out ofsolution is high, and production is high.Chemical additions should be made promptly.Consistent filtration of the solution willminimize roughness from particles, and periodic


solution purification using activated carbon willremove organic contamination and products ofthe organic breakdown of brighteners andwetting agents.


7.2Process Control

Ammeters and voltmeters should be standardizedevery 6 months, and ampere-hour meters shouldbe checked monthly. Temperature controlsshould be checked weekly. The post-plate bakeoven temperatures are checked weekly.

Mandatory detailed plating procedures should beadhered to when manual plating is done.Computer-controlled automatic plating machinesaccomplish this rigidly.

A controlled plating load zone of 55 ft2 isoptimum, with 65 ft2 a maximum for a 14 in. ×30 in. hexagonal plating barrel.

The plater should use a computer or nomographto determine the number of pieces per barrel loadand the ampere-hours necessary for the specifiedthickness.


thickness.

All plated parts must be in the post-plate ovenwithin 14 hr depending on the specificationrequirement.

7.3Post-Plate Inspection

The plating operator will inspect plated parts forappearance, blisters, thickness dimensions, andgage fit at the plating facility before sending partsfor the post-plate bake. Parts should bereinspected after the post-plate bake before theyleave the plating department.

8Terminology Related to Electroplating*

Addition agent: A material added in smallquantities to a plating solution to alter thecharacteristics of the deposit such as grain size,smoothness, brighteners, hardness, and stresscorrosion.


Ampere: Current that flows at the rate of 1coulomb per second and is capable of depositing0.0011180 g of silver in 1 s.

Anion: A negatively charged ion such as chloride(Cl) or sulfate ( ).

Anode: The electrode, usually made of the metalbeing deposited, connected to the positiveterminal of the electrical power supply. Theanode dissolves to replenish the metal beingdeposited at the cathode.

Anode bar: The metal bar, normally a 1 in. roundcopper rod, from which anodes or anode basketsare hung.

Anode basket: A metal basket hung from theanode rod that is filled with small anodes shapedas balls, cones, chips, buttons, or rods to providea large anode surface area.

Barrel plating: Bulk plating of small partswhere the parts are plated, while in a rotating


perforated cylinder.

Basis metal: The metal on which theelectrodeposited metal is applied.

Brightener: An addition agent that causes theplating to deposit bright rather than dull.

Burning or burnt deposit: A rough, dark,powdery deposit caused by using too high acurrent density.

Bus bar: A copper bar, 1/4 × 4 in. for carryingup to 1000 A from the power supply to a processtank in the plating system.

Cathode: The electrode or metal on which theelectroplating is deposited. It is connected to thenegative side of the power supply.

Cathode efficiency: The weight of metal actuallydeposited divided by what should havetheoretically been deposited according toFaraday's law. Expressed as a percent.


*See ASTM B 374, ''Standard Terminology Relating toElectroplating," for a complete list of terms used in platingand metal finishing.


Cathode film: A film approximately 0.0050.010in. thick that forms adjacent to the cathodeduring plating and through which the metal ionspass to be deposited. The metal concentration inthis film is significantly lower than in the body ofthe plating solution.

Cation: A positively charged ion such as sodium(Na+) or nickel (Ni2+).

Cathode bar: The metal bar, normally a 1 in.round copper rod, from which the cathode orwork to be plated is hung.

Chemical equivalents: The weight of an elementin grams equal to the atomic weight divided bythe valence of that element for a specificreaction. The number of atoms in a chemicalequivalent is equal to Avogadro's number (6.022× 1023) divided by the valence.

Combined cyanide: The cyanide in the plating


Combined cyanide: The cyanide in the platingsolution that exists in the cyanide complex. (SeeCyanide complex.)

Coulomb: The quantity of electric currentflowing through an electric circuit when 1 Aflows for 1 s; abbreviated C.

Covering power: The ability of a platingsolution to deposit metal in recessed areas ofparts at very low current densities.

Current density: The current per unit areaexpressed as amperes per square foot (A/ft2).

Cyanide complex: A negatively charged ionicgroup formed when a metal ion combines withtwo or more cyanide ions as in the reaction

.

Electrochemical equivalent: The weight ofmetal deposited or dissolved or of the gasevolved at either the cathode or anode during thepassage of a unit quantity of electric currentexpressed in faradays, ampere-hours, or ampere-


minutes.

Faraday: One faraday (1 F) of electricity isequal to 96,485 C or A · s. One faraday willdeposit one chemical equivalent of a metal.

Flash: A very thin plating, normally 0.0001 in.or less. (See also Strike.)

Free cyanide: The amount of cyanide present ina plating solution that is not included in acyanide complex.

High current density area: An area on a part thatis closest to the anode or a protrusion thatdevelops current densities much higher thanthose in other parts of the part.

Hydrogen embrittlement: Embrittlement of ametal due to absorption of hydrogen duringcleaning, activation, and plating. Hydrogenembrittlement may lead to failure when the partis stressed at levels considerably below its tensilestrength.


Ion: An atom or molecule that has gained apositive or negative charge by giving up oracquiring one or more electrons.

Limiting current density: The maximum currentdensity at which a satisfactory deposit can still beobtained.

Low current density area: An area on a part thatis farthest from the anode or a recessed area of apart such as a hole, which will develop currentdensities much lower than other areas of the part.

pH: The hydrogen ion concentration expressed asa negative logarithm. The pH scale ranges from 1to 14, with a pH of 7 being neutral, those below7 acid, and those above 7 alkaline.

Pickling: The removal of scale and oxides frommetal surfaces using an acid solution either aloneor with other chemicals or acids.

Plating rack: A spline or frame that articles tobe plated are attached to or suspended from so


that several parts can be plated simultaneously.

Plating range: The current density range overwhich a satisfactory deposit is obtained.

Strike: A very thin deposit usually plated out athigh current densities and short times to achievecoverage and adhesion.

Throwing power: The ability of a platingsolution to deposit a uniform thickness ofplating over the entire surface of a part.


Total cyanide: The total of the free andcombined cyanide concentrations of a platingsolution.

Work: The part or parts being plated.

References

1. Selection of Electrodeposited Coatings,American Electroplaters and Metal FinishersSociety, Orlando FL.

2. Pearlstein, F., Selection and application ofinorganic finishes: Part V, Metal deposits; PartVI, Barrier layer protective metal deposits; PartVII, Metal deposits for specializedcharacteristics, Plating and Surface Finishing,April 1979, p. 28; May 1979; June 1979, p. 42.

3. Cross Index of Standard Documents forMetallic and Inorganic Coats, Annual Book ofASTM Standards, Vol. 02.05, Metallic and


Inorganic Coatings; Metal Powders, SinteredP/M Structural Parts, 1991, p. 788.

4. Lowenheim, F. (ed.), Modern Electroplating,3rd ed., Wiley, New York, 1974.

5. Durney, L. (ed.), Electroplating EngineeringHandbook, 4th ed., Van Nostrand Reinhold, NewYork, 1984.

6. Safranek, W. H. (ed.), The Properties ofElectrodeposited Metals and Alloys, 2nd ed.,American Electroplaters and Surface FinishersSociety, Orlando, FL, 1986.

7. Lowenhein, F. A., Electroplating, McGraw-Hill, New York, 1978.

8. Metal Finishing Guidebook and DirectoryIssue, Metals and Plastics Publications, Inc.,Hackensack, NJ, 1995.


7Fastener Quality

Christopher B. WackrowMNP Corporation, Utica, Michigan

1Introduction

For many people, the word quality conjures upthoughts of craftsmanship, durability, style,price, a "name," etc. Unfortunately, such termsare perceptual measures and are often inadequatewhen considering the use of products such asfasteners for engineered applications.Nevertheless, quality is definable, and thischapter is intended to give an explanation ofwhat quality is and the means by which it comesabout. This should prove useful to those who are


(1) designing parts for an application, (2)evaluating the manufacture (or manufacturer) offasteners, or (3) examining parts in the field orthe laboratory.

In this chapter, we shall also see the importanceof process control and the causes of variationand how it manifests itself.

The selection of reference material is intendedfor those who are either new to this subject ormore widely read. From my own experience thevolume by Grant and Leavenworth [1] is a goodreference on the subject of quality, while theAIAG (Automotive Industry Action Group)references contain material that could applyequally well to all industries.

While not seeking to be a definitive work on thissubject, this chapter is written to address thoseaspects that can play a part in ensuring reliablefastener joints and seeks to show how variationcan enter in that can affect that objective.


Environmental factors can play a part too in thelong-term performance of fasteners, but thissubject is best researched by the readerelsewhere.

In my experience, the humble fastener is poorlyunderstood by most, especially with regard to itsproperties, standards, qualities, performance inassembly, and behavior in service. This issurprising considering the fact that almosteverything humans make uses fasteners of somekind. This lack of understanding affects correctpart selection.

Much needs to be done to educate users andspecifiers as to the requirements of fasteners, andmore people need to aspire to the capacity offastener engineers. The contents of this book, ofwhich this chapter is a part, will hopefully help inthis regard.

2Quality


The attainment of quality is involved with themeasurement and reduction of variation inproducts, services, and activities, such thatpredictability and stability result, unnecessarycosts


are reduced, and customer satisfaction isobtained. If variation is not controlled and isexcessive, it will cause losses to everyone in thechain. These can be losses in money, efficiency,performance, and, in certain cases, human lifeand property.

The practices of quality management and qualityengineering are concerned with how variationenters into products, services, and activities andhow these measured differences can be managed,controlled, and minimized. In these practices avariety of quality methods or tools are available.These methods can extend from the use ofstatistical techniques such as pareto charts,control charts, histograms, and capabilityanalysis, to employee involvement in processoperation, problem solving, and the operation ofthe business.

The reader should note that the scope of these


quality methods is not restricted to the making ofparts but can encompass other vital areas of thesupply chain such as design, testing,measurement, manufacturing process control,packaging, materials handling, office functions,and delivery, such as in the application of totalquality management.

In a wider sense, quality is key to the quality ofour lives, and the process of continuousimprovement in quality will involve everyoneand require much effort. These companies thatcan incorporate this and increase quality byreducing variation will have a distinct advantagein the marketplace.

2.1Quality Levels

It is probably assumed by most purchasers offasteners that fastener lots satisfy a "zero defect"criterion. Too frequently, the quality level offasteners is unspecified by the consumer, and


consequently he or she is exposed to varyinglevels of supplied quality. The purchaser willneed to inquire more closely as to the methodsused by the supplier to ensure that a faircomparison can be made between sources ofsupply. Price alone will be no indicator of thelevel of quality that will be supplied.

The consumer should note too that althoughmany systems are capable of providing highlevels of quality, the objective of zero defects isunverifiable without the inclusion of 100%piece-per-piece verification inspection andscreening techniques.

Certain industries have certain standardsregarding expected levels of quality, and thesecan vary widely. The automotive industries useparts per million measures when calculatingpieces shipped versus pieces rejected.Government industries often used AcceptableQuality levels, which could vary according topart type and the characteristic being checked.


The advent of automated assembly reduced thelevel of problems that could be tolerated,because the machine is not able to recognize aproblem at the point of use as a human can.Automated methods require the provision ofextremely high levels of quality. These qualitylevels will probably be achievable only with100% single or double sorting methods.

The best meeting point between supplier andconsumer is to specify the quality level desiredby stipulating the method of its confirmation.There are three methods available to do this, andthey can be used to qualify and/or certify lots.

2.1.1Acceptance Sampling

In using this section, refer to Section 6.

Method I involves the use of a statisticallyvalid sampling plan (sample size varies withlot size) and an acceptance criterion (thenumber of unsatisfactory parts that would be


acceptable in the sample).

Although there are a variety of samplingtables, a popular one is MIL-STD-105 [2],which, although a little outmoded, can be auseful sampling tool. This method requiresno calculations and permits ready selectionof sample sizes, according to a variety ofquality levels one is looking to confirm andthe lot size being examined. It is applicableto both variable (numerical readings), andattribute (go/no-go) inspections.


Method II involves inspecting a sample offixed size (30 pieces is sometimes suggested)and subjecting the data to statistical analysisby using an inexpensive computer, acalculator, or a longhand and paper method.The analysis can evaluate mathematically thedegree of statistical compliance to engineeringlimits. See Section 4.1.

Both methods are useful for those who want tovalidate a production process or validate a lotreceived from a supplier.

2.1.2Process Certification

Method III is to define the requiredexpectations of the production processes to beused and to rely on documented evidence ofcompliance as criteria for acceptance. Theexpectations can be supplemented by


stipulating expected levels of processperformance using the process indices shownin Section 4.1. In well-developed systems, thismethod can form a basis for reducing the needfor Method I or II acceptance sampling, andmay even negate it.

2.2Key Characteristics

In either acceptance sampling or processcertification it is important to define whichcharacteristics of the fastener the quality levelwill apply to. Since characteristics are not all ofequal importance, it is up to the design engineerto identify which characteristics must satisfy aspecified quality level. If the customer does notidentify them, the manufacturer must do so.

Although parts can vary widely in terms ofproperties and geometry, etc., it is recommendedthat they be evaluated to determine product fitand function parameters from the perspective of


both assembly and service. Sometimes wherestandard off-the-shelf items are involved, priorexperience may be quite adequate. However, areview of how the part could fail (as determinedby a failure modes and effects analysis [3]) willyield vital information in identifying thestructure of features to control and a suitablefrequency of in-process inspection and processsupervision. Those features, be they dimensional,metallurgical, etc., that are deemed to beparticularly important to fit and function mayalso benefit from audits to verify any effectsfrom subsequent upstream operations. Thisreview might indicate the need for testing that isnot routinely performed and that may have to beperformed by an independent laboratory.

3Quality Management

The supply of quality parts is the result ofeffective participation of individuals in the chain


of events extending from the point where afastener need is first determined to the placewhere it sees service. These persons, operatingwithin the procedures and policies of a qualitysystem, will ensure that variation is reduced inthe performance of prescribed duties. For thesake of brevity, the system can be broken intothree parts, as follows.

3.1Design and Quotation Control

It is important to determine whether a standardoff-the-shelf part is appropriate or whethersomething more specific to the needs of theapplication is required. This will need to bedetermined between the end user and thepersonnel assigned to obtain the fastener,preferably through the medium of a fastenerengineer and/or metallurgist. The degree towhich the part is defined by these personnel andthe information is communicated to themanufacturer will be crucial.


The buyer, or the buyer's engineer, shouldconsider

Key characteristicsQuality level expectations


Assembly and installation needs(automated/nonautomated assembly)Corrosion resistance needsTorquetension expectationsFatigue-producing conditionsEnvironmental conditions

To prevent the inadvertent introduction of errorsat the design stage and during manufacture, thedesigner and manufacturer can use tools such aserror proofing and checklists. Failure modes andeffects analysis (FMEA) [3] will be useful to theengineer to

1. Identify potential areas where the design orprocess may fail to satisfy requirements or causeit to not perform as intended

2. Bring about corrective actions

With these techniques, potential problem areascan be prevented from entering the


can be prevented from entering themanufacturing phase.

3.2Manufacturing and Supply Control

To be effective, the manufacturing quality systemmust identify those what, when, how, who, andwhy aspects will be required to be followed if therequired quality at production volumes is to beprovided on time.

Control plan [4] formats are effective methods toconvey this necessary detail.

What:

Quality level verification

Key part characteristics

Process control techniques

Gages and measurement

Sample sizes


Audits

Equipment

Tests

Materials, etc.

When: The frequency with which the "what"items are used, verified, and deployed.

How: Work instructions; methods andprocedures; techniques; etc.

Who: On-line personnel; off-line personnel withthe necessary skills, training, authority, and scopeof responsibility.

Why: Awareness training should explain toeveryone the reasons for the quality system andwhy their support, involvement, cooperation, andcompliance is important.

The quality plan should provide for the necessaryinspection of parts and audits of the qualitysystem itself to ensure that predetermined


requirements are being satisfied.

In addition, the system must have a way ofcorrecting problems quickly and a means tominimize and segregate any parts that are notsatisfactory. The system must ensure thatnonconforming items are correctly assessed anddealt with and that the knowledge learned fromresearching the causes is turned "inward" toimprove the system and educate the people in it.

3.3Lot Control and Traceability

In materials handling, it is important to avoidmixtures. Mixing material lots can compromisethe ability to provide quality by

1. Introducing variation that could affect processcapability at upstream operations


2. Affecting the ability to certify compliance tocustomer order and other lot-specific criteria

3. Affecting inventory accountability

To accomplish this, material in the system mustbe managed using a system that permits lotcontrol. This is usually accomplished byassigning numbers that reflect the uniquecharacter or origin of the material. Theidentification numbers are attached to thematerial by way of a tag, label, or other methodand stay with it as it passes through the system.

This identification is often supplemented byother numbers that provide other vitalinformation about the material and its origin.The degree of identification (and traceability)depends on the sophistication of the inventorycontrol system. These supplemental identifiersare sometimes used to uniquely identify portions


of a run, such as when a significant event hastaken or is about to take place. In this way,inventory can be moved through the system inmanageable lots.

With a discipline that reinforces segregation toidentity, lot numbers, and suitablecontainerization, undesirable mixing isminimized. In effect, the lot number becomesconnected to the unique properties of thematerial in that lot. Some typical examples ofsegregation are

1. Heat number segregation prior to and duringheat treatment

2. Segregation of heat-treated lots prior to andduring finishing

3. Segregation of finished lots prior to andduring packaging

4. Segregation of packaged lots prior to andduring sale and shipment


Lot control principles also apply to thecontrolled disposition of lots that are notsatisfactory. Such material should be speciallyidentified as to its nonconformity status, and thematerial control system must ensure that it doesnot come into contact with the lot from which itwas taken or with other lots sharing a "common"description.

Lot control also permits traceability tooperations and in-process records. This not onlyprovides a quality assurance verification and thevital linkage to test reports, but also permits thescope of any problems (and the reason for them)to be determined. The records should besufficient to enable the necessary reconstructionof past events so as to determine the extent of aproblem.

4Variation in Processes*

It is important to minimize variation in the


process output by ensuring that critical processfactors predictably influence part characteristicsin both the short and long term. The maincontributors to variation are (1) materials, (2)methods and procedures, (3) human factors, (4)process equipment, and (5) measuringequipment. In practice, actual results are due tothe interaction of two or more of these anddetermine the process's capability to satisfy anengineering specification.** The chief strategy inachieving this is to use process control andassess the nature and capability of the process.

4.1Process Capability

Although it is possible to measure processcapability from a handful of parts, systems usingcontrol charts enable both short- and long-termcapability to be determined as well as the

* This section relies heavily on Refs. 5 and 6.** Ideally, the designer of the specifications will have involvedthe processor in considering the capability of the process that


will satisfy the specifications. Since many specifications arewritten without any knowledge of the capability of the processused to produce them, the responsibility lies with themanufacturer to know what tolerances can be held by hisprocessing system. It is recommended that the purchaserascertain the supplier's knowledge of his process to verify whatquality levels the supplier is capable of supplying.


phenomenon called control. Control charts aretools that plot process output measurements. Thepositions of these plot points with respect tocontrol limits and each other give processbehavior "messages" to the user. Based on areaction plan, these messages can be acted on foron-the-spot or off-line corrective action, theadjustment or investigation of assignable reasonsor causes. Since it is unlikely that all possibleprocess variables and effects would showthemselves all at once, control charts are runover long periods of time. Once these "tools"have served their purpose, their use can besuspended.

If a process should show "out-of-control"behavior, then the control chart can be used tofind assignable causes. The removal of thesecauses will enable variation to be reduced,control to be established, and quality improved.


What remains is variation due to "commoncauses," and the process's best capabilitybecomes apparent. However, if this bestcapability (the process at its ''best") is not able toprovide the quality level required of the parts,then the following options are available,depending on the terms of company policy andgovernment or customer regulations:

1. Sort the products. The limitation here is thecapability of the sort. Sorts based on visualcriteria should be closely managed.

2. Change the engineering specification.

3. Revise engineering criteria. For example, apart not satisfying engineering tolerances may befully functional. The use of functional criteriacan improve ease of manufacture and obviatesorting.

4. Remove specification. The specification maybe unnecessary.


5. Repair or rework the part.

6. Accept the part as-is.

7. Scrap the part.

4.2Capability Measures and Statistical ProcessControl (SPC)

Variation is measured as the standard deviation s(sigma) and is expressed in the same units ofmeasure as the data.* Estimates are calculatedeasily from a control chart or from statisticalanalysis of raw data using software, calculator,or longhand and the equation found in Reference5 Typically, the capability is equal to 6s.

4.2.1Capability Ratio and Process Capability

When process capability is divided by thetolerance, a capability ratio (CR) is obtained.This ratio (multiplied by 100 to give percent)gives an idea of how much of the tolerance is


"consumed" by process variation.

For example, if s = 0.05 mm (so 6s = 0.3) andthe tolerance is 0.5 mm, the capability ratio is

An alternative measure is Cp, process capability(over the long term). This is the reciprocal ofCR.

Using the example above,

* The most accurate values of sigma are achieved as thenumber of inspected samples increases. Readers aredirected to the references in calculating "sample" and"population" sigma.


4.2.2Process Capability over the Long and Short Term

Although CR and Cp are useful indicators ofcapability, they do not give an indication of theposition of the data relative to the specification.Processes invariably look their best in the shortterm, they may not look much better over time.Since processes often need to be assessed in boththe short and long term, and process outputquality can depend on process "targeting"practices, the Cp and Pp factors were of limitedusefulness. To address the limitation, the factorsCp and Pp were supplemented with a centeringfactor, k, and a calculation that consideredprocess average position compared tospecification limits. Cpk and Ppk calculationsthen enabled users to evaluate process capabilityand process position simultaneously. The factorsCpk and Ppk are used to address this (Ppk


addresses short-term capability, and Cpk, long-term). In calculating either index, the arithmeticdifference (z) between the data average and thespecification limit that it is closest to is dividedby 3s.

Using the example above, and assuming a dataaverage of 0.3 mm,

4.2.3Capability Quality Standards

Process practitioners can elect to use the Cpk (orPpk)* values and/or the CR values depending onthe nature of the process, the rules set for them inthe end user's contract, or the level ofmanagement that the practitioner has determinedis right for their process. Any standards agreed on(especially Cpk) should be used with cautionespecially when die controlled dimensions, tool


wear, and narrow tolerances are involved. In suchcases, dimensions that approach specificationlimits as a normal consequence of operation canresult in Cpk values that suggest the presence ofnonconforming parts that in reality will not bethere.

4.2.4Comments on SPC

There are a variety of control chart methodsavailable, and their selection and application arebest guided by consulting Refs. 5 and 6. Themotive for using control charts is to study theprocess by plotting its data and to use the data todetermine its behavior. Control charts aredesigned to be sensitive to changes in processbehavior that are significant from a statisticalstandpoint. Statistically significant events arethose that signal the occurrence of some effectthat merits investigation and appropriatecorrective action. Some knowledge of processbehavior is needed to begin charting, in order to


space the sampling frequency and sample sizeapproximately.

The sample size should be adequate to reflectpart-to-part variation, and sampling should becarried out frequently enough to be sure that anysignificant events have a good chance of notescaping detection by falling betweeninspections. A suitable strategy will includemaking records and inspections when significantchanges occur to each of the contributing groups.

One cause of poor process control/capability isunnecessary adjustments or poor targeting of theprocess by the operator. Targeting is theintentional operation of a process to produceoutput dimensions (etc.) to a specified region ofthe dimensions tolerance. Where a dimension isadjustable, the personal preference of individualoperators in setting a process within thespecification can add unnecessary variance to theprocess output. This can contribute to out-of-control results or poor capability.


* The difference between Cpk and Ppk in this classificationdoes not mean a relaxing of standards but instead is areflection of the natural occurrences of sources of othervariation over longer periods of time.


Random Causes

Not all processes respond to control charting,because charting success is based on signalingsignificant events called "assignable causes."However, in some processes, random causes canoccur, and their presence can be particularlytroublesome. These events are not detectableusing a control chart because of

1. Low levels of occurrence

2. Infrequency of occurrence

Foreign material introduced during bulkhandling is an excellent example of this. Themagnitude of such causes is found only by 100%sorting prior to shipment or by screening on-lineat the assembly point.

4.3Metal Forming


In producing part features from die-dependentprocesses, it is the variables of tooling size andrate of tool wear that largely govern the amountof variation in the output. Control charts (suchas the bar X and R type) are not particularlysuited to these situations because

1. Trends are often natural to the process (trenddirection is predictable).

2. Piece-to-piece variation is small

Because of these, most metal-forming processcharts would be sending out "out-of-control"messages all the time and only cause confusionto the operator who observes minimal variationand the fact that no nonconforming parts arebeing made. These "messages" arise because thewidth of control limits is calculated from therange between the highest and lowest readings ina subgroup sample. When this range is small andapproaches 0 (as is often the case in dimensionsfrom tooling), control limits are prohibitively


narrow. The presence of any subsequent trendingis then signaled as an "out-of-control" message.

Improved quality in cold forming (and any tool-dependent process) is achieved by reducing thewear rate of the tool and using process controlsthat make use of the known behavior of the tool.As an added benefit, the lower the rate of wear,the less frequent will be the need for inspection.

4.4Thread Rolling, Roll Forming, CenterlessGrinding, and Machining

Thread rolling, roll forming, centerless grinding,and machining are subject to many of the samevariables as metal forming, with one majordifference. They can be adjusted on-line. In thesecases the operator can observe process variationand adjust the position of the tools to givedimensions at predetermined locations within theengineering tolerance.

4.5


Heat Treatment

The process of heat treatment is of particularimportance because it is the method whereby thestrength of parts is created. The greatest variablesaffecting heat treatment quality are

Material capability. The correct selection willbe the one that gives the desired structure andhardness of the core after soaking at thehardening temperature for the prescribed timeprior to quenching.

Within-heat chemical element variation.Generally this is a lesser concern than theinadvertent mixture of different heats of thesame grade.

The length of time at the hardeningtemperature. Parts must be at the hardeningtemperature for a certain minimum time. It isimportant that the thickness of the layer of


parts on the furnace belt or in a basket ortray be uniform. Soak times should bederived in conjunction with feed rates ofparts and apply to the thickest portion ofthe load.

Length of exposure to the required furnaceatmosphere conditions. Exposure for therequired time is critical from the standpoint ofcarbon loss (decarburization) or carbonabsorption (carburization) in steels duringneutral hardening. It is also important to avoidthe inadvertent pickup by steels of nitrogensuch as might occur in a furnace that isswitched from neutral hardening to nitridingand vice versa.

The speed with which the parts are subjectedto the quench media. This must be to obtainthe minimum asquenched hardness. This isimportant where quenching (hardening or


solution treating) is essential to thedevelopment of physical properties. The rateof quenching (and degree of media agitation)will depend also on the material'shardenability. Uniform quench mediatemperature can be an important factor toquenched material properties. Media selectionis important if quench cracking is to beavoided and must be appropriate to both thematerials and the engineering specification.

4.6Coating and Plating

4.6.1Coating and Plating Thickness

The purpose of most platings and coatings is toprovide corrosion resistance. Unfortunately, theyall add to the dimensional size of the part. Theeffect is small on most surfaces except thethreads, where at least 4 times the coatingthickness is added to the pitch diameters due to


the geometry of the thread angle. Both the designengineer and the manufacturer should becautious here because

1. Some finishes are inherently thick, andundersizing of the thread prior to coating will benecessary if maximum thread fit considerationsare critical.

2. The thickness of the coating at its highestpoint due to the finish's process capability needsto be considered, not the minimum specified forthe finish.

3. Coating uniformity depends on the method offinish application. Electrodeposited finishes aremore uniform than dip-drain (galvanized) or dip-spun organic ones.

4.6.2Corrosion Resistance

Quality corrosion resistance results aredependent on


Surface cleanliness of parts prior to applyingthe finish. The degree of cleaning necessarydepends on the amount of residue remainingfrom heat treatment. The tenacity of theresidue will depend on the type and amount oflubricant on the parts prior to heat treatmentand the effectiveness of the cleaning systemapplied in removing them prior to heattreatment.

Adhesion of the finish.

Amount of finish on the surface. Platings areusually measured by thickness, while organiccoatings are usually specified by coatingweight rather than thickness, due to thenonuniformity of the coating layer and thedifficulty of measurement.

Management of the corrosion test cabinet.Most corrosion testing is done in test cabinetsthat are miniature controlled environments(ASTM B117-based tests are popular).*


Although frequently specified, salt spraytesting is not regarded as a very effective toolto evaluate finish quality.

* There is often a temptation to try to equate ASTM B117test results to real life years but this is not exact andcomparisons should be avoided.


4.6.3Processing Effects

The exposure of parts to water-based cleaning,electroplating operations, and otherhydrogenbearing media (even outside moisture inassembled joints) presents the possibility thathydrogen will infiltrate into the parts. If it does,then, depending on the amount of hydrogenpresent, the form it has taken within the part, andthe hardness of the part, it can cause suddenfailure after assembly due to hydrogenembrittlement. Although this topic is covered indetail elsewhere in this book, the followingvariables are generally accepted as key factors inminimizing hydrogen embrittlement fromprocessing, when the hardness of the part somerits.

1. Length of time parts are in cleaning.


2. Length of time parts are in plating solutions.This is largely a function of the efficiency of theplating system, the maintenance of bathchemistry, and the thickness of plating.

3. Length of time parts soak in any water-basedmedium.

4. Length of time between the plating of partsand their introduction into the baking furnace.

5. Length of time between introduction into thefurnace and the speed with which parts at the"core" of the load reach baking temperature. Thefurnace circulating air temperature as well as thetemperature of the parts within the furnace mustbe monitored with a thermocouple. The use oftrays with air space between to reduce thethickness and bulk of a bake load and aid in hotair circulation is beneficial in increasing heat-upspeed.

6. Length of time parts at the core stay at thebaking temperature. The length of time necessary


will depend on the highest hardness of the partbeing baked.

4.7Surface Discontinuities

Surface discontinuities originate in the rawmaterials or are caused by processes carried outon the material.

4.7.1Raw Material

Seams and folds are discontinuous shallowfissures of variable length running along thelength of the raw material. Created during therolling of wire rods, they can be shallow enoughto not affect aesthetic appearance or specificationcompliance. They often cause visible splits orfissures during metal forming. Severe coldforming may require that the rod be shaved priorto forming or that rods be used that were made tospecific seam level standards.


4.7.2Processing

Fissures can occur if the material is overstressedin heading, with the yield stress of the materialbeing exceeded at some particular region on thesurface such as a seam or pit. Fissures at a 45°angle (so-called shear crack) signal such a cause.

Folds in certain critical areas, as identified in thestandards, must be avoided. Folds can be causedby poor upset shape or incorrect balancing ofworkpiece lengths on multistation machines.

Starting serration marks are the remnants ofindentations put into the profile of thread rollingdies. These indentations facilitate rapid pickup ofthe workpiece just prior to rolling. If thesuperficial marks are unacceptable, then partsproduced from circular roll die technology maybe preferred, or serration-free flat die rolling atlow rates of production.

Laps are sometimes confused with starting


serration marks and occur at thread rolling. Lapsthat are not starting serration marks are the causeof poor in-process control at the thread roller.There is a correlation between significantfunctional size/pitch diameter disparities and theproduction of severe laps.


Handling damagenicks, dents, and dingsdisplaysitself on edges, corners, and crests of threads.Damage that is severe enough to affectfunctional performance is of specific importance.Handling systems vary in their ability tominimize handling damage. Such a system'scapability will vary greatly as product size andproduct hardness variables enter in. Successiveprocessing and reprocessing provide cumulativelevels of surface damage, nicking, or thread crestrounding. Inspection methods will need to be ofa practical functional nature, since confusion ofacceptability can be a problem. In the event thatnormal levels of surface damage areunacceptable to an agreed-upon method ofinspection, secondary processing (methods that"chase" threads) probably will be required.

Quench cracks are cracks that occur as a result ofthe severe dimensional contraction stresses that


develop when certain steels are quenched into aquench medium that is too "fast" (too efficient inremoving heat from the quenched part). Suchcracks are irregular in shape and are probably themost severe kind of surface discontinuity. Thereis no corrective action possible for quench-cracked parts, and the choice of steel, hardeningtemperature, and quench medium will have to bereviewed.

5Variations in Gaging and Measurement*

Fasteners are often evaluated by instruments thatgive numerical information (variable-type gage)and those that give go/no-go results (attribute-type gage). In attribute gaging, no numericalinformation is obtained; the results arequalitative, not quantitative. Such gages,although useful for evaluating completed parts,should not be used for process control or forcontrolling key characteristics.


Gages can contribute errors that, if large, canconfound the ability to accurately measureprocess capability. Errors can also result fromvariations in the gaging method employed.Variations can be related to

Location within the part (within-piecevariation)Appropriateness of the gage for the inspectedfeatureThe environment of the gage or the conditionof the partHuman variables in using the gage (personalgaging technique)Other human-related errors

Other significant variables are

Ease of calibration, and the gage's ability tomaintain calibration under expected serviceconditions.

The operating range and usefulness of thegage at the extremes of its scale (linearity).


The gage's ability to discriminate anacceptable part from an unacceptable part. Thegage must be able to read to an accuracy equalto or better than the number of places to theright of the decimal point in the specification.

Gage cost and ease of use.

While many gage calibration systems typicallyaddress the calibration of gage instruments to areference standard, repeatability andreproducibility (R&R) gage error studies addressa gage's suitability with respect to the inspectedtolerance and include the human factors. Gageerrors can be calculated, and the amount of thepart tolerance that total gage errors can consumecan direct specific courses of action.

* This section is based on Ref. 7.


6Variation in Sampling

As one might expect, the larger a sample, thegreater the confidence in any conclusions drawnfrom it. In some cases, due to availability or cost,it may not be possible to obtain a large samplefor testing. However, with the use of statisticalanalysis, much can be learned from a smallsample provided some precautions are takenwhen analyzing the sample.

Sampling effectiveness is based on the theory ofprobability. If a batch of parts is obtained from acontrolled process where process control oranother prevention-based system is employed,then a very large sample will show favorableresults. However, if the process is not somanaged, then small samples could indicatenoncomplying parts. The presence of many parts


"hugging" a specification, for example, mightsignal sorted material or a process controlled to aspecific region of a tolerance for some specificreason (engineering fit preferences, readiness fora thick or heavy finish, tool size, etc.).

The origin of a sample and the results of sampleinspection will depend also on the history of thematerial as presented for inspection. The historywill show that the material presented is eitherstratified or homogeneous.

Layered (stratified): Results of sampling herevary depending on how far down one goes into acontainer to retrieve samples. In effect therecould be layers (strata) of varying quality present,as might happen, for example, if a process weretrending excessively as processed parts weregradually being discharged into the container.

Evenly mixed (homogeneous): In checking partsfrom an evenly mixed lot the probability offinding similar results from wherever one takes


the samples from within the container is greaterthan if one took parts from a stratified lot. Thiscase is most often encountered when parts aremixed as a natural part of downstream handling,feed, and processing. In effect, each step mixesup any strata put there at prior operations.

7Variation in Assembly and Service

Fasteners are designed to be subjected to stress.While stress is observed to be mostly in tensionresulting from the application of torque, shear orother resolved stresses are also often involved,largely as a result of part application.

The typical method of obtaining tension (clampload) in threaded fasteners is to apply torque. Thetorque develops tension through the medium ofthe thread. The efficiency of this process isdependent on friction, which is a reflection of theease with which the contacting surfaces moveagainst one another. This frictional coefficient is


dependent on the nature of the finish (its lubricityas well as its roughness) at the point of contactbetween the mating parts.

Variations in the coefficient of friction can arisefrom variations in

The choice of finishThe age or condition of the part finish at thejob siteThe type of finishing processThe choice of chemical processing system

Most specifications do not quantify or specifythe friction coefficient. It is recommended that itbe determined by testing prior to application.

The loss or degradation of all or part of the finishcan affect the coefficient of friction, which inturn affects the amount of clamp load stress.Unauthorized or unqualified changes to the kindof finish can have significant consequences onthe clamp load induced in the bolt. These effectsare either a reduction or an increase in the


friction coefficient.

A reduction of the friction coefficient has thepotential of taking the bolt beyond itsspecified lamp load. This can show itself asfastener fracture either at the moment ofassembly or sometime later, especially iftransient forces are experienced by the joint.


An increase in the friction coefficient has thepotential of not taking the part to its desiredclamp load. This exposes the part to thepossibility of fatigue failure sometime afterinstallation and service. Transient or normalservice loads cause loosening (due toinadequate joint tightness) and causes unevenstresses to be exerted on the fastener.

Power tools and manual forms of assembly canintroduce variation into the amount of appliedtorque also, with a corresponding variation inclamp load. Assembly systems that areelectronically controlled or provide actualreadings to the operator are valuable means toensure consistent application of applied torque.Greater confidence can also be attained bymeasuring both torque and angle, as well as byconducting audits of actual assembled joints.


References

1. E. Grant and R. Leavenworth, StatisticalQuality Control, McGraw-Hill, New York,1988.

2. MIL-STD-105 105 E, Department of Defense,Washington, DC, 1989.

3. AIAG, Failure Modes and Effects Analysis,Automotive Industry Action Group, Southfield,MI, 1995.

4. AIAG, Advanced Quality Planning, AIAG,Southfield, MI, 1995.

5. AIAG, Fundamental Statistical ProcessControl, AIAG, Southfield, MI, 1991.

6. Levine, Duncan, and McCune, All You EverWanted To Know About Control Charts, ASTMTranscripts STP 1209, American Society forTesting and Materials, West Conshohocken, PA,1994.


7. AIAG Measurement Systems Analysis, AIAG,Southfield, MI, 1995.


PART IIITHREADED FASTENERS

8Screw Threads

Eli SchwartzConsultant, Philadelphia, Pennsylvania

John H. BickfordConsultant, Middletown, Connecticut

1Introduction

The threads are obviously an important elementof the threaded fastener. They give this sturdy,industrial product its unique ability to beinstalled, removed, and reinstalled as many times


as we wish. They also affect fastenerperformance in a major way. Thread type, threadclass, thread configuration, the way in which thethreads are produced, and the fit between maleand female threads can affect not only threadstrengthand therefore fastener tensile strength butalso the resistance of the fastener to such thingsas self-loosening and fatigue. The amount ofpreload achieved for a given torque can beinfluenced by thread configuration and bywhether the threads have been cut or rolled. Allthings considered, it's worthwhile to take a closelook at threads.

2Thread Forms

2.1Thread Forms in General

Literally hundreds of thread forms have beendesigned, and many are still in common use in awide variety of applications. Fortunately, we


only have to worry about the few that arecurrently used in threaded fasteners. To clear thedecks, however, let's start by taking a quick lookat three other forms that are often mentioned andthat many newcomers to bolting assume theyshould know about. These forms are illustratedin Figure 1, along with a currently popular 60°form for comparison. The three we don't have toworry about are the ACME, buttress, andWhitworth.

The ACME thread is used for powertransmission, for example, to producetraversing motion on machine tools [1].

The buttress thread is used when the thrust onthe screw is in one direction only, forexample, for screws in airplane propeller hubsand columns for large presses [2].


Figure 1Three well-known thread forms that are not currently used with threaded

fasteners and, for comparison, one that is (the Unified form). The ACME isa wellknown machine tool thread used for traversing screws. The Whitworth

form was once used in the United Kingdom but has been replaced by a60° ISO form.

The Whitworth thread, which had a 55° includedangle instead of the now universal 60°, was fordecades a British standard form.

All of the modern fastener thread forms we need toknow aboutthe metric as well as inch series formsarebased on an arrangement of 60° angles. This basic


geometry is modified in several different ways, but it'sthe starting point for all contemporary fastener threadprofiles.

2.2Inch Series Thread Forms

In the United States our principal inch series threadform standards are ASME B1.1-1989 [3] and federalstandard FED-STD-H28/2B [4]. Both of these describethe basic Unified Thread form, identified by the codeletters UN or UNR.

A slightly modified version of the UN/UNR thread isthe UNJ form, which is defined in militaryspecification MIL-S-8879C.

The differences between these three formsUN, UNR,and UNJare shown in Figure 2. The differences appearto be slight, but, compared to the flat root UN thread,the rounded root of the UNR form has less stressconcentrations and greater fatigue life. Since the UNJform has an even larger root radius, stressconcentrations are further reduced and fatigue lifefurther increased. The internal thread minor diameterfor UNJ or MJ threads is slightly larger than that for


UN or M threads. In any event, the differences are that

1. The UN form has flat-bottomed or, optionally,slightly rounded roots.

2. The UNR form must have slightly rounded roots.

3. The UNJ form has generously rounded roots.

2.3Metric Thread Forms

Metric threads are identified by the code letters M andMJ. The basic geometries of metric and inch seriesthreads are identical, but the way we define metricthreads differs from the way we define the inch seriesthreads.


Figure 2These are the thread forms most commonly used in the western world atthe present time. They differ from each other in the roots of the external

(male) thread. The UN and metric M forms have flat or, optionally, slightlyrounded roots. The UNJ and metric MJ forms have generously rounded roots.

The UNR thread form and the M form used on metric fasteners with tensilestrengths of 115,000 psi (800 MPa) and stronger have

slightly rounded roots.

U.S. standards for metric screws threads include ASMEB1.13M-1995 for the general use M profile [6] andANSI B1.21M-1978 for the MJ profile threads [7]. TheM profile thread is similar to the Unified inch seriesthread and provides either flat or rounded external thread


roots. Rounded root external threads are recommended,and they are mandatory for metric fasteners with tensilestrength of 800 MPa (115,000 psi) and stronger. Mthread rounded roots have a slightly larger radius ofcurvature than those of the equivalent UNR inch thread.MJ threads are equivalent to the UNJ inch series threads.

3Thread Profiles

3.1Basic Profile

The design of the male and female threads starts with a''basic profile," which has been called the "permanentlyestablished boundary between the provinces of theexternal and internal threads" [8]. Figure 3 shows thebasic profile of both external and internal threads for theUN, UNR, and metric M thread forms. Figure 4 showsthe few differences that convert the standard UN or Mbasic profile to either a metric J or UNJ profile.

3.2Design Profile

The design profile, also called the "design thread form"[8], defines the maximum material profiles for the


internal and external threads of a given form. The designprofiles for internal threads are the same as the basicprofiles shown in Figures 3 and 4. External threaddesign profiles are shown in Figure 2. Figure 5 gives usa closer look at the differences between the UNR andUNJ (or rounded root M and MJ) profiles.


Figure 3The basic profile of the UN, UNR, and metric M thread forms. This is the basis forthe design of those threads. The design profile of internal threads is identical to the

basic profile. External thread design profile may vary from the flat base profile roots.

4Thread Series

Design profiles can be applied to threads of any size, and thisleads to what are called "thread series." In the Unified Threadsystem, for example [3],


1. Threads called just UN/UNR or UNJ constitute the"constant-pitch" series. Each thread in such a series has thesame number of ridges or grooves per inch. For example, inthe "8 pitch" series, each thread in the group has eightthreads per inch. The angle the helix of the thread makesaround the fastener varies with the diameter of the fastener,but the depth of the thread is constant, regardless ofdiameter, because of the rigid 60° geometry on which theform is based.

Altogether, there are eight constant-pitch thread series; 4, 6,8, 12, 16, 20, 28, and 32 threads per inch. These are allstandard for the UN and UNR forms, but only the constant-pitch series of 8, 12, and 16 threads per inch are standard forthe UNJ form.

2. In addition to the constant-pitch series, there are severalgroups classified by "coarseness." This refers not to theirquality but to the relative number of threads per inchproduced on a common diameter of fastener. For example,the codes UNC, UNRC, and UNJC identify "coarse-pitch"threads.

3. UNF, UNRF, and UNJF all designate "fine-pitch" threads.

4. UNEF, UNREF, and UNJEF designate "extra fine"threads.


Figure 4The basic profile for the UNJ and metric MJ threads differs from the profileshown in Figure 3 only in the dimensions illustrated here. Again, the designprofile of internal threads is identical to the basic profile and the external

thread design profile varies from the flat basic profile roots.

5. Provision has also been made for a "special" seriescalled UNS, UNRS, or UNJS, which havepitchdiameter combinations not found in any of thestandard series above.

As already mentioned, the coarseness designations


refer to the relative numbers of threads per inch. For afastener of a given diameter a UNC thread has fewerthreads per inch than a

Figure 5The design profiles of the UN/UNR and UNJ (or M and MJ) threads differ

primarily in the way the roots are shaped, as suggested by Figure 2. Thissketch shows a closeup of the difference for Class 2A UN threads. The

difference appears to be slight, but the UNJ and MJ threads have strengthand fatigue advantages over the UN/UNR or M threads.


UNF, which in turn has fewer than a UNEF. Forexample, here is the nomenclature used to definethe five standard threads currently specified for afastener having a nominal (body) diameter of 1/2in.

The UNC (coarse) thread for 1/2 in. diameterhas 13 threads per inch and is encoded as a1/2-13 UNC thread. Other options here are1/2-13UNRC or 1/2-13UNJC.

Next in line is a UN (constant-pitch) thread,which has 16 threads per inch and is called1/2-16UN or 1/2-16UNR, etc.

Next there's a UNF (fine) thread with 20threads per inch, called 1/2-20UN.

Then a UNEF (extra fine) thread with 28threads per inch, or 1/2-28UNEF.

Finally, there's another constant-pitch thread


for 1/2 in. fasteners, 1/2-32UN. As with allthe others in the group, it can have the UNRform too, but the constant 32 pitch series isnot standard for UNJ form threads.

These numbers all change as the diameter of thefastener changes. For example, the followinggroup of threads has been codified for 2 in.diameter fasteners:

2-4 1/2UNC 2-12UN2-6UN 2-16UN2-8UN 2-20UN

All the threads in this group are constant-pitchthreads except for the first, coarsest thread. Thisis true of most "large diameter" fasteners. In fact,above a 4 in. diameter there are nothing butconstant-pitch threads. And once again, each ofthese threads can have a UNR or UNJ forminstead of UN if you wish.

Metric threads generally duplicate the inch series


threads but with threads considered to be onlycoarse or fine. Instead of using threads per inchto classify them, however, we'll use the pitchdistance between two teeth, in millimeters. Forexample,

would specify a thread having a nominaldiameter of 6 mm and a pitch distance of 1 mm.

The metric coarse series consists of the coarsestpitches for most of the common diameters in therange of 1.668 mm. All other diameterpitchcombinations are considered fine. Table 1 showsa comparison of the standard metric fastenerseries and the UNC/UNF inch series threads. Allof the metric threads are coarse series throughsize M64.

5Thread Allowance, Tolerance, and Class

We still haven't finished classifying threads. The


next important consideration is the way the maleand female threads fit together. Is the fit looseand sloppy, or is it tight? Rough applicationsrequire the first, precision applications thesecond.

The fit between mating threads is determined bythe basic clearance between them, a clearancedetermined by the way the threads aredimensioned and by the tolerances placed onthose dimensions. Although the philosophy is thesame for both metric and inch series threads, thenomenclature used is different [3, 4, 8], so we'llconsider the two types separately.

5.1Inch Series Threads

5.1.1Allowance

The basic clearance is established by a "threadallowance," which determines the minimumclearance between threads of a given class.


Another way of saying this is that it determinesthe minimum distance between male and femalethreads when both nut and bolt are in their


TABLE 1 Coarse and Fine Inch Series and Metric Fastener SeriesCompared

Thread ThreadDiam. (in.) tpi Metric

series Diam. (in.) tpi Metricseries

0.060080UNF 0.750010UNC0.063 72.6 M1.6 × 3.5 0.750016UNF0.079 63.5 M2 × 0.4 0.79 10.2 M20 × 2.50.086056UNC 0.87 10.2 M22 × 2.50.086064UNF 0.8750 9UNC0.098 56.4 M2.5 × 0.45 0.875014UNF0.112040UNC 0.94 8.5 M24 × 30.112048UNF 1.00008UNC0.12 50.8 M3 × 0.5 1.000012UNF0.138032UNC 1.06 8.5 M27 × 30.138040UNF 1.1250 7UNC0.14 42.3 M3.5 × 0.6 1.125012UNF0.164032UNC 1.18 7.3 M30 × 3.50.16 36.3 M4 × 0.7 1.2500 7UNC0.164036UNF 1.250012UNF0.190024UNC 1.3750 6UNC0.190032UNF 1.375012UNF0.20 31.8 M5 × 0.8 1.42 6.4 M36 × 40.250020UNC 1.5000 6UNC0.24 25.4 M6 × 1 1.500012UNF0.250028UNF 1.65 5.6 M42 × 4.50.312518UNC 1.7500 5UNC0.32 20.3 M8 × 1.25 1.89 5.1 M48 × 5


0.312524UNF 2.00004.5UNC0.375016UNC 2.20 4.6 M56 × 5.50.39 16.9 M10 × 1.5 2.25004.5UNC0.375024UNF 2.2500 4UNC0.437514UNC 2.52 4.2 M64 × 60.437520UNF 2.7500 4UNC0.47 14.5 M12 × 1.75 2.83 4.2 M72 × 60.500013UNC 3.0000 4UNC0.500020UNF 3.15 4.2 M80 × 60.55 12.7 M14 × 2 3.2500 4UNC0.562512UNC 3.5000 4UNC0.562518UNF 3.54 4.2 M90 × 60.625011UNC 3.7500 4UNC0.63 12.7 M16 × 2 3.94 4.2 M100 × 60.625018UNF 4.0000 4UNC


"maximum material" conditions: the fattest boltand the thickest nut. The lower sketch in Figure 6shows external Class 2A (allowance) threadsmating with Class 2B (no allowance) internalthreads, separated only by the allowance.

5.1.2Tolerance

Manufacturing tolerances are now placed on theallowance. These are always in the direction ofless material; they always make the clearancebetween nut and bolt threads greater than theclearance determined by the allowance, neverless. The upper sketch in Figure 6 shows theclearance between external Class 2A and internalClass 2B threads, with full tolerances added tothe allowance.

In the Unified thread system the female threadsare always dimensioned to the basic profile, that


shown in Figure 3 or 4. The male thread isreduced a little (made smaller in

Figure 6


The lower sketch shows the clearance between maleand female Class 2A UN threads when only the basicallowance separates them. The upper sketch shows

how muchthis clearance increases when the fullmanufacturing tolerance is added to the allowance. Ineffect, the lower sketch shows the maximum material

condition for bolt and nut; the upper sketch theminimum material condition for both. The bolt thread

is reduced in diameter by the allowance and tolerance,and the nut thread is increased in diameter by the

tolerance.


diameter) by the allowance. Tolerances, againalways in the direction of less material, have tobe placed on both male and female threads, ofcourse.

5.1.3Class

Three basic fits or "classes" are defined forUnified threads. These are given the codes 1A,2A, and 3A for male threads and 1B, 2B, and 3Bfor female threads. The pair 1A and 1B define theloosest fit; 3A and 3B define the tightest.

Class 1A/1B threads are used for rough work,for example, where some thread damage canbe expected or conditions are very dirty.

Class 2A/2B threads are for general use.

Class 3A/3B fasteners are used forapplications requiring an extra degree of


precision.

Class 1A and 2A threads are assigned the sameallowance in the Unified system, but theallowance in Class 2A threads may be used forplating. Class 2A threads that may not use theplating allowance are designated as Class 2AG.Classes 1A and 1B have more generoustolerances than Classes 2A and 2B. Class 3A/3Bthreads are assigned a "zero allowance," so the fitcan be line-to-line. A small tolerance on eachthread makes assembly possible.

Internal threads (Classes 1B, 2B, and 3B) haveno allowances.

5.2Metric Threads

The clearance between male and female metricthreads is also determined by a basic allowanceand by tolerances in the direction of lessmaterial. The number of tolerance and allowanceoptions is greater with metric threads than with


inch series threads, and different names are usedto describe them [68].

5.2.1Tolerance Position (The Allowance)

The design clearance between a thread and itsbasic profile, called the allowance in inch seriesthreads, is called the tolerance position formetric threads and is identified by a lettersymbol:

G or H for internalthreads

e, f, g, orh

for externalthreads

H and h define zero allowance, no fundamentaldeviation from the basic profile. G and g providea small allowance that makes the thread ridges alittle less fat and may be used to accommodateplating or perhaps to provide a looser fit. Forgeneral-purpose M threads, the internal threads


have been standardized with no allowance (H)and the external threads with a small allowance(g), similar to standard UN inch threads. The MJthreads used in aerospace applications aredesigned with no allowance for either internal orexternal threads (H/h), similar to aerospace UNJinch threads. The larger external thread fitsprovided by positions f and e are not used exceptin special cases.

5.2.2Tolerance Grade (The Tolerance)

What is called the tolerance of inch series threadsis called the tolerance grade for metric threadsand is identified by a number symbol. Seventolerance grades have been established forexternal threads and are identified by the numbers39, with 9 defining the loosest fit (the mostgenerous tolerance) and 3 the tightest. Tocomplicate matters further, different groups ofoptions have been established for external andinternal threads, and tolerances have been placed


on several thread dimensions. The possibilitieslisted in ASME B1.13-1995 are shown in Table2.


TABLE 2 Tolerance Grades Assigned to Metric ThreadsDimension controlled Specified tolerance

gradesMinor diam., internal threads4, 5, 6, 7, 8Pitch diam., internal threads 4, 5, 6, 7, 8Major diam., externalthreads 4, 6, 8

Pitch diam., external threads 3, 4, 5, 6, 7, 8, 9

For general-purpose M threads, standardtolerance grade 6 is applied to the dimensions forboth internal and external threads. MJ aerospacethreads use tolerance grade 4 for pitch diametersand 5 or 6 for crest diameters.

5.2.3Tolerance Class (The Class)

The tolerance grade and tolerance positionsymbols are combined into a "tolerance class"code of alphanumeric symbols.

6g is a general-purpose callout for external


threads. It defines the tolerance position andtolerance grade for the major and pitchdiameters. Used in conjunction with a 6H nut,the fastener would be a reasonable substitutefor applications previously using a Class2A/2B pair.

4g6g is also a general-purpose callout forexternal threads, this time when a tighter fit isrequired. It defines the tolerance grade andposition for both the pitch diameter (4g) andthe major diameter (6g). Note that in inchseries practice these two diameters cannot betoleranced separately as they can here. Class4g6g, however, is considered to define anexternal thread that is an approximateequivalent to an inch series Class 3A threadwith an allowance available for plating. Theapproximate equivalent to the standard no-allowance Class 3A thread is designated 4h6h.

4H5H defines the tolerance grade and positionfor a common internal thread that could also


be used with the male threads defined above.The 4H classifies the pitch diameter and the5H the minor diameter. It is standard for MJprofile threads larger than 5 mm andapproximates an inch thread Class 3B.

5.3Inch Series and Metric Thread ClassesCompared

It is useful to know the relationships betweeninch and metric thread classes. Some exampleswere given above. Table 3 gives the approximateequivalents used in the United States.

5.4Coating Allowances

If fasteners are to be plated or otherwise coated,some allowance must be provided for thecoating. The way this is done is presumably ofinterest only to fastener manufacturers, whowould be guided by the applicable standards.Users (buyers) might find the following


summary of interest, however.

5.4.1Inch Series

The allowances specified for Class 2A externalthreads can accommodate coatings of reasonablethickness. Special provisions must be made forother classes of external thread, all internalthreads, and threads to be given heavy coatings.Major and pitch diameter limits before and


TABLE 3 Approximately EquivalentClassifications: Inch and Metric Practices

Bolt threads Nut threadsInch Metric Inch Metric1A 7ga 1B 7Ha2A 6g 2B 6H3A 4h6h 3B 4H6Hb

4H5Hca7H and 7g are not standard in the UnitedStates.bFor use with diameters of 5 mm and less.cFor use with diameters greater than 5 mm.

after coating must be specified on engineeringdrawings, for example. See ASME B1.1 [3] forUN/UNR threads or B1.15 [9] for UNJ threads.

5.4.2Metric Series

External thread tolerance classes 6g and 4g6gprovide allowances that can accommodatenormal coatings. For heavy coats, or if position h


or H tolerance positions are involved, one shouldconsult the standards ASME B1.13M [6] for Mthreads or B1.21M [7] for MJ threads.

Mechanically galvanized fastener threads providea thick coating of zinc. In the United States it isstandard practice (ASTM A563M [10]) toovertap the nut threads to accommodate thecoating. In some other countries, allowance ismade on the bolt threads, but this practice mayresult in a weakened joint.

5.5Tolerances for Abnormal Lengths ofEngagement

The allowances and tolerances specified in threadstandards assume normal lengths of engagementbetween male and female threads. The definitionof "normal" is spelled out in the specifications,but typically it means lengths of engagementranging from five pitches to one and a half timesthe nominal diameter of the thread [3]. For


metric M threads, normal length of engagementis defined in formulas based on both diameterand pitch [6].

If the length of engagement is to be abnormallyshort, then it is wise to reduce the clearancebetween male and female threads by reducing thetolerance. If the engagement is to be unusuallylong, tolerances must be relaxed, or pitchmismatch may make it impossible to assemblethe fastener or to run the bolt into a deep tappedhole.

The standards, again, define the procedures formodifying the allowance or tolerance. ASMEB1.13M is especially clear on this point. It saysthat for very short lengths of engagement thetolerance on the pitch diameter of the externalthread should be reduced by one number. Forexample, instead of 4g6g one might specify3g6g.

For extra long lengths of engagement, B1.13M


says that the allowance on the pitch diametershould be increased. A normal 4g6g wouldbecome 4g6g, for example.

6Inspection Levels

The three thread gaging systems defined inASME B1.3M [11] provide three levels ofinspection. Selection of a gaging system willdepend on the nature of the application, theconsequences of failure, and the process controlsused. Each screw thread specified must include adesignation of a gaging system either tacked onto the thread callout or in a note on a drawing,


standard, specification, or other contractual document. FED-STD-H28/20B [12] provides guidance in making a selection. If threadform is essentially perfect, any of the gaging systems will determineif the thread is too large or too small.

System 21 uses simple GO and NOT GO functional gages. Itchecks whether a fastener will assemble with a conforming matingthreaded part. It does not check as rigorously as the other standardsystems of gaging the minimum material conformance of a threadthat varies from perfect form.

System 22 also checks fastener assembleability. There is, however,more control of minimum material conformance possible thanthere is with System 21.

System 23 is similar to System 22, but greater control of flankangles, lead or pitch, taper, roundness, etc., is possible. Moresophisticated equipment is needed to check these characteristics.

7Thread Nomenclature

We can now put all of the foregoing together to give the completealphanumeric code or description of a thread.

7.1Inch Series

An example of an inch series external (bolt) thread "code" would be


where

1/4=nominal diameter in inches

20 =number of threads per inchUNCshows that this is a UN thread from the coarse series.

2A

shows that it is a loose-fitting external thread (A) with a finiteallowance available for plating up to the basic no-allowancedimensions. Tolerance on pitch and major diameters is greaterthan that of a Class 3A thread.

(21)shows that the thread is to be inspected with simple GO andNOT GO gages as explained in Section 6.

If that thread were used on a bolt with a 1 in. long body, the codeused to define the fastener would be

Coarseness and fit would not usually be added to the fastener code.The number of threads per inch gives the user coarsenessinformation (a quarter-inch UNF fastener has 28 threads per inch; aquarter-inch UNEF one has 32). A fit of 2A and inspection level 21would presumably be assumed for such a bolt.

Another example would be

0.2500-32UNJEF-3A, Safety critical thread

Here we see the quarter-inch nominal diameter of the externalthread given in decimal form followed by the number of threads perinch, 32; the series, UNJ Extra Fine; and the allowance andtolerance level, 3A. We're also told that this fastener is intended for


"safety critical applications." When the UNJ thread conforms toMIL-S-8879 [5], gaging system 23 provides the necessarymeasurements for the safety-critical application unless some otherrequirements are specified. If the UNJ thread conforms to ASMEB1.15, the required gaging system must be specified just as it is forUN threads.


7.2Metric Thread

An example of a complete code for an external metric thread wouldbe

MJ shows that the threads are metric and have controlled roundedroots with radii larger than the general-purpose M threads.

6 =nominal diameter (mm).1 =distance between successive thread crests (i.e., the pitch) (mm).

4h=

the tolerance grade (4) and tolerance position (h) for the pitchdiameter of the thread. (Position h specifies zero allowance;Grade 4 is used for standard MJ bolt threads per ANSI B1.21M[7]).

6h=

the tolerance grade and position for the major diameter. (Again hsignifies zero allowance; Grade 6 is also used for standard MJbolt threads per ANSI B1.21M).

(22)

shows that the thread is to be inspected using GO and pitchdiameter gages rather than GO and NOT GO gages; major andminor diameter and root radius inspections are also required.(Also see Section 6.)

8Coarse Vs. Fine Vs. Constant-Pitch Threads

Which is best, coarse-pitch, fine-pitch, or constant-pitch threads? Itdepends on your application. Each has advantages over the other [35,


13, 14].

8.1Coarse-Pitch Threads

Coarse pitch is generally recommended for routine applications. Suchthreads will have greater stripping strengths when used with weak nutor joint materials or when used on larger diameter fasteners. Somesay that bolts over 1 in. in diameter should always have coarsethreads; others put the crossover point at 1 1/2 in.

It is easier to tap brittle material if coarse-pitch threads are used. Suchthreads are also easier to use in most cases: easier to start, fasterrundown, etc.

8.2Fine-Pitch Threads

Fasteners with fine-pitch threads can have higher tensile strengthsbecause the thread root and pitch diametersand therefore the tensilestress area, Asare greater than they would be for a coarse-pitch threadon the same nominal diameter. This advantage can be obtained,however, only with a suitably long length of engagement betweenmale and female threads. This subject is treated in Section 9.

Since fine-pitch threads are stronger, they can be loaded to higherpreloads before yielding. They also resist self-loosening undervibration or shock better than coarse-pitch threads.

An extra fine series is used primarily where only thin nuts arepractical or where fine adjustment is necessary.

8.3Constant-Pitch Threads


The series of constant pitches available for inch fasteners range from4 to 32 threads per inch. They are used when the standard coarse, fine,and extra fine series are not satisfactory. The most common are the 8-, 12-, and 16-thread series.

The 8-thread series is used on large diameter fasteners and wasoriginally intended for bolts used in gasket joints containing highpressure. It is also widely used as a substitute for coarse seriesfasteners when the basic fastener diameter exceeds 1 in.


The 12-thread series is used as a continuation ofthe fine thread series when bolt diameters exceed1 1/2 in. It was also originally intended forpressure vessels but has since found wider use.

The 16-thread series is also used on large-diameter fasteners, again for those requiring fine-pitch threads. It is used as well for adjustingcollars and as a continuation of the extra finepitch series for bolt diameters over 1 11/16 in.

For metric fasteners, the series of constantpitches available range from below 1 mm to 8mm. Except for when a diameterpitchcombination is that of the coarse series, thesefasteners are considered to have fine-pitchthreads.

8.4Miscellaneous Factors Affecting Choice

Other thread characteristics that may affect our


Other thread characteristics that may affect ourchoice of thread are discussed in other chapters,but a few miscellaneous comments may be inorder here.

A tighter fit, such as a Class 3A/3B rather than aClass 2A/2B, will give a slight increase in threadstripping strengththe amount varies with threadsizebecause there is slightly more root crosssection to be sheared.

The UNJ and MJ threads have more resistance tofatigue than the UN, UNR, and M threads. Thereis, however, an even greater increase in fatiguestrength when the bolt thread is rolled after heattreatment.

The number of threads in the grip (between theface of the nut and the head of the bolt) affectsthe ductility and stiffness of the fastener. Sincewe (usually) want ductility and low stiffness (amore resilient spring for better energy storage), itwould seem that we would usually want fullythreaded fasteners. We will be especially


interested in ductility if we are using yieldcontrol to tighten the fasteners. Factors such asthe shear strength of the fastener and its fit withits hole, however, often argue instead for partialthreads and an unthreaded body of nominal orreduced diameter.

9The Strength of Threads

There is a surprising amount of disagreement onwhat parameters determine the strength of athread and on how best to evaluate the quality,including the strength, of a threaded fastenerbefore use. Let's take a look at someconventional wisdom concerning thread strengthand then consider some recent thoughts andconcerns about thread strength and quality.

9.1Basic Considerations

One principal design goal is to produce a fastener


strong enough to support the maximum preloadit might receive during assembly, plus themaximum additional loads it might see inservice, as a result of forces applied to the joint,differential thermal expansion, etc. The larger thenominal diameter of a fastener, of course, thestronger it will be. As far as static loads areconcerned, therefore, we would like the shank orbody of the bolt to be the full, basic, or nominaldiameter of the thread, or at least as great as theminimum pitch diameter of the thread.

We must then specify a length of threadengagement capable of developing the fullstrength of that body. This is just another way ofsaying that we want the bolt to break before thethreads strip, because a broken bolt is easier todetect than a stripped thread.

When the threads strip they do so by shearing inone of three ways. If the nut threads are strongerthan the bolt threads, then the bolt threads willstrip near the roots. If the bolt threads are


stronger, stripping will occur near the roots ofthe nut threads. If they have equal strengths, bothnut and bolt threads will strip simultaneously, attheir pitch diameters.


Studies made at the National Bureau of Standards manyyears ago showed that the tensile/shear strength ratio forcommon fastener materials varied between 1.7 and 2.0. Asa result, the stripping areas (AS) defined in the formulasbelow (on which the recommended lengths of threadengagement are based) are set at twice the tensile stressarea (As) of the same thread [4].

If the fastener is to be subjected to fatigue or impactloads, we would like it to be more resilient than a fastenersubjected to static loads. Some recommend a shank (body)diameter about 60% of that used for static loads if thefastener will see impact loads, or a shank diameter of 90%of the static diameter if it will experience fatigue loads(repreated load cycles).

9.2Tensile Strength of a Bolt

Experiments have shown that a general-purpose boltbreaks in tension through the threads, as if it werebreaking through an equivalent solid shank with adiameter between that of the pitch and minor diameters.The area of this equivalent shank is called the stress area,As, and the breaking load of the bolt is


where UTSs = ultimate tensile strength of the bolt (psi orMPa).

For UN or UNR inch threads, the standard value of stressarea is

where

D=

basic majordiameter

n =threads per inch.

For M-form metric threads, the standard value of stressarea is

where

D=

basic majordiameter

P =thread pitch.


For UNJ inch threads rolled after heat treatment, thestandard value of stress area is

where d2max = maximum pitch diameter.

For MJ metric threads rolled after heat treatment, thestandard value of stress area is

where

d2max=maximum pitch diameter

d2max=

maximum rounded root minordiameter

For UNJ inch and MJ metric threads rolled before heattreatment or not rolled at all, the standard value of stressarea is


9.3The Static Shear Strength of a Thread

We use a very simple equation to estimate the force required to strip (shear) the threads ofa bolt or nut:

where

USS=Ultimate shear stress of the nut or bolt materials (psi, MPa)

AS=

cross-sectional area through which the shear occurs (in.2,mm2) (This is not the cross-sectional area of the body, as weshall see in a minute.)

As mentioned earlier, thread failure will occur in either the nut or bolt threads, or in bothsimultaneously, depending on their relative strengths. A different expression must be usedto compute the shear stress area for each type of failure.

The following equations were taken from an appendix of FED-STD-H28/2B [4]. Table 4is provided for those not familiar with the current symbols used in screw thread formulas.

9.4Nut Threads Stronger Than Bolt Threads

Failure occurs near the root of bolt threads. The equations for shear area (ASs) and thelength of thread engagement (LE in inches or millimeters) required to develop fullstrength of the threads are as follows:

where

ASs =shear area at root of bolt threads (in.2,mm2


n =number of threads per inchAs =tensile stress area of bolt (in.2, mm2)

D1max=

maximum inner diameter (ID) of nut (in.,mm)

d2min=

minimum pitch diameter (PD) of bolt(in., mm)

Also, FED-STD-H28 gives simplified expressions for shear areas. The expression usedwhen the nut material is stronger than the bolt material isTABLE 4 Current vs Former Notation Used in ScrewThread FormulasCurrent symbol Former symbol Variable

d Ds Bolt thread, major diam.d2 Es Bolt thread, pitch diam.D1 Kn Nut thread, minor diam.D2 En Nut thread, pitch diam.LE Le Length of engagement


where d2bsc = the basic (or nominal) pitch diameter of the external thread (in., mm) (seecomments below).

Let's do an example to see how different the results are when we use the simple vs. themore complex expression and to see where to find the data required to use these equations.

Take as our example a 3/4-12 UN-2A thread and compute ASs for a length of engagementequal to one diameter (the thickness of a heavy hex nut). We get the necessary data from arecent edition of either ASME B1.1 or Machinery's Handbook [17]. These tell us that

D1max=0.678 in.

d2min =0.6887 in.

d2bsc =0.6959 in. (see notebelow)

Note that the basic or nominal pitch diameter of the bolt, d2bsc, is equal to the minimumpitch diameter of the internal (nut) thread, because a zero allowance is assigned to theinternal thread, which therefore conforms to the basic profile at the pitch line andelsewhere.

From Eq. (2), in calculator format,

From Eq. (4),

The calculated increase is 12% in this case, with the simplified expression giving a slightlyless conservative result (it would take more force to shear 1.025 in.2 than 0.917 in.2).

9.5Nut Threads Weaker Than Bolt Threads

The foregoing are the equations you would use to compute the strength of a tapped hole in


a joint material such as aluminum or cast iron. The use of aluminum is increasing inautomotive and military applications, for example, as designers struggle to reduce weight.

Failure occurs near the root of nut threads. The equations are

where

dmin =minimum OD of bolt threads (in., mm)D2max

=maximum PD of nut (in., mm)

UTSs=tensile strength of the bolt material (psi, MPa)

n =threads per inchAs =tensile stress area of bolt (in.2, mm2)

UTSn=ultimate tensile strength of the nut material

ASn =shear area of root of nut threads (in.2, mm2)

LE =length of thread engagement required to develop fullstrength (in., mm)


The simple expression for shear area when the bolt material is stronger than the nutmaterial, again from FED-STD-H28, is

where D2bsc = the basic (minimum) pitch diameter of the nut (in., mm).

9.6Nut and Bolt Threads of Equal Strength

Failure occurs simultaneously in both parts, at the pitch line. The equations are

where

D2bsc=nominal pitch diameter of the bolt (in., mm)

As =tensile stress area of bolt (in.2, mm2)AS =shear area at pitch line of both threads (in.2, mm2)

LE =length of thread engagement required to develop fullstrength (in., mm)

9.7Things That Modify the Static Strength of Threads

A number of factors can modify the anticipated tensile strength of a boltsuch things ashigh temperature, corrosion, torsion, or cyclic loading. These things can also modify thestrength of threads. So can some other factors that are not quite so obvious.

1. Nut dilation [18, 19]. If the walls of the nut are not thick enough, the wedging action ofthe threads will dilate the nut, partially extracting the nut threads from the bolt threads.This reduces thread engagement and therefore reduces the cross-sectional areas thatsupport the shear load, reducing shear strength. If the ratio between width across flats andnominal diameter is only 1.4:1, for example, strength will be reduced by 25% as shown inFigure 7. (The width across flats-to-diameter ratio of standard nuts is 1.5 or 1.6:1.) Note


that the reduction applies to both nut and bolt threads, the failure occurring in the weakerof the two.

2. Relative strength of nut-to-bolt threads [2022]. As we have seen, the relative strengthalways determines which members will fail. If there is too big a difference between thetwo materials, another factor must be considered: The weaker of the two threads willdeflect under the relatively stiff action of the other, creating a form of threaddisengagement that again reduces the area supporting shear stress. Note that it doesn'tmatter which threadnut or boltis substantially weaker than the other. The result is shownin Figure 8.

3. Coefficient of friction. The strength factor values shown in Figure 8 are for parts ''asheat treated" that have a relatively high coefficient of friction. If the coefficient of frictionbetween nut and bolt threads is lower, then both nut dilation and thread bending becomemore likely because the threads can pull apart more readily. A lubricant such as phosphateand oil, for example, is said to reduce resistance to thread stripping by as much as 10%[20].

4. Rotary motion [20, 23]. Dynamic friction is usually less than static friction. As we haveseen above, anything that reduces friction between nut and bolt threads makes


it easier for the nut to dilate and/or for the threadsto bend. This means that the threads are a littlemore likely to strip during torquing operationswhen the nut is moving relative to the bolt thanthey are under static loads. The reduction instrength is estimated to be approximately 5%.

To compute the modified potential strength of anut thread, therefore, you multiply the apparentstrength in pounds by the appropriate nut dilationfactor from Figure 7 and by the appropriate threadbending factor (for nuts) from Figure 8. If thethreads are lubricated, the computed strengthshould be reduced by an additional 10%; if torqueis used to tighten the nuts, a final 5% reduction isrequired.

Similar calculations are used to estimate thestrength of the bolt threads, the only differencebeing that the thread bending factor used (fromFigure 8) will be that for bolts rather than the one


for nuts. As an example, let us compute thestrength of the threads for the 3/4-12 UN-2A boltwhose thread stripping area we computed amoment ago to be 0.917 in.2 (using the longexpression for ASs). The mating Class 2B nut has acalculated thread shear area ASn, from Eq. (5),equal to 1.269 in.2. Let us assume that because ofspace limitations we are using a nut with a ratio ofwidth across flats to nominal diameter of 1.45:1, alittle less than normal.

We used the thread-stripping area formula for "nutmaterial stronger than bolt material." Now assumethat the nut material is stronger than the boltmaterial, with a shear strength of 105 ksi. The nutis to be used with an ASTM A490 bolt whose shearstrength is 96 ksi.

The bolts are to be lubricated with molybdenumdisulfide, an even better thread lube than phosphateand oil, and they will be tightened with a torquewrench.

We use Eq. (1) to compute the theoretical forces


required to strip the threads.

Figure 7Strength reduction factor for nut dilation as a function of the ratio of the

across-the-flats distance to the nominal diameter of the fastener.(Modified from Ref. 19.)


Figure 8Strength reduction factor for thread bending. The horizontal axis gives the ratio of nut strength to bolt strength.

(Modified from Ref. 20.)

where

USSs =96 ksi

ASs =0.917 in.2

and Fn = (USSn)(ASn) = 133,250 lb

where

USSn =105 ksi

ASn =1.269 in.2


Now we apply the strength factors as follows:

SF1 = strength factor for nut dilation for a 1.45:1 ratio = 0.8 (from Figure7).

SF2 = strength factor for thread bending where the strength ratio of the nutto bolt (Fn/Fs) is 1.5 = 1.1, a strength increase (from Figure 8).

SF3 = coefficient of friction factor. Let us assume a 15% loss of strength(which is probably conservative) because molybdenum disulfied is almost50% more lubricious than phos-oil. So SF3 = 0.85.

SF4 = rotary motion factor, a loss of 5%; so SF4 = 0.95.

The reduced estimate for the strength of our threads is now

A significant difference! Even if we assume a 10% reduction for lubricityinstead of 15%, the reestimated strength is only 66,230 lb. However, theultimate tensile strength of a 3/4-12 A490 bolt is only 59,670 lb (found bymultiplying As by the ultimate tensile strength of 170 ksi max from ASMEB1.1) so the bolt would presumably break before these threads stripped, whichis desirable. Not much margin for safety, however.

Is this analysis valid? The reduction factors we have just used come fromstudies made by E.M. Alexander for the SAE [20] and the results, commonlycalled the "Alexander model"


have been widely accepted and used. Conclusion:The estimate we have just made is valid. If weneeded more strength than that implied by theresults, we should increase the length ofengagement between the bolt and our abnormallythin-walled nut.

9.8Which Is Usually StrongerThe Nut or the Bolt?

You will find that the proof strength of astandard nut is generally greater than the proofstrength of the fastener with which it is supposedto be used. Designers would prefer bolt failure tonut failure because a failure of the bolt is moreobvious. For example, the amount of torque wecan apply to a bolt with a stripped thread is oftengreater than the torque we applied just beforestripping occurred. The increase in torqueindicates an increase in tension or preload in thebolt when, in fact, all preload is lost when the


thread fails. On the other hand, there is no chanceof misreading the situation when the body of abolt breaks; that's obvious. In an apparentcontradiction, nuts are made of weaker (softer)material than bolts. This encourages plasticyielding in nut threads to bring more threads intoplay in supporting the load. But the nut as a bodywill still withstand a higher tensile force than themating bolt. For the same reason designers willwant tapped holes to be deep enough to morethan support the full strength of the bolt.Equations (3), (6), and (9) will lead to thisresult. In fact, many people will use theequations only to find the length of tapped holes.They won't design nuts.

Note that standard nuts come in severalconfigurations. As far as hex nuts are concerned,a regular hex nut has a thread length equal to0.875 times the nominal diameter of the bolt.Thick and heavy hex nuts have a length equal tothe nominal diameter. All three types of nuts


should be able to develop the full strength of thebolt with varying factors of safety, but there canbe problems.

9.9Table of Tensile Stress and Shear Areas

Equations are nice, but it's often handy to have atable of "answers." Tables 5 and 6 give asummary of the various stress areas we'vediscussed: tensile stress (As) and the strippingareas [ASs and ASn from Eqs. (2) and (5)].Remember, we want the bolt, not the nut, to fail.

In Tables 5 and 6, the thread-stripping areas arecalculated for a length of engagement equal toone nominal diameter (the common length forthick or heavy hex nuts). For other engagementlengths, divide the listed area by the nominaldiameter of the thread and multiply by the actuallength of thread engagement.

10


Effects of Variations in Threads

Thread variations include those of size and thoseof form. We shall look at the effects of some ofthese. But first it is important to know that evenif a variation can possibly weaken a thread, thereis no well-documented history of failure causedby such variations.

10.1Pitch Diameter

Pitch diameter is the diameter of a cylinder thatcuts through the thread ridges and grooves at aposition where the width of the ridges is equal tothe width of the grooves. When this element iswithin tolerance, there is a correct amount ofmaterial in the thread at this location. If threadform is also proper, the full thread strength willbe attained and the thread will assemble with aconforming mating thread.

Both bolt tensile strength and the stripping(shear) strengths of bolts and nuts are affected by


pitch diameter. Let us look at an example of ajoint with a 5/8-18UNF-3A thread. The bolt hasan ultimate tensile strength (UTSs) of 150 ksiand an ultimate shearing strength (USSs) of 90ksi. Its mating nut engagement is equal to thenominal diameter, and its shearing strength(USSn) is 75 ksi.


TABLE 5 Selected UNC/UNF/8UN Thread Strees and Shear AreasaBolt thread strip area,

LE = 1 diam.Nut thread strip area,

Le = 1 diam.Thread Tensile

stress area 2A/2B 3A/3B 2A/2B 3A/3B

0.190024UNC0.0175 0.050 0.053 0.076 0.0800.190032UNF0.0200 0.052 0.055 0.074 0.0780.250020UNC0.0318 0.092 0.096 0.135 0.1410.250028UNF0.0364 0.093 0.101 0.130 0.1370.312518UNC0.0524 0.147 0.157 0.213 0.2220.375016UNC0.0775 0.216 0.232 0.310 0.3220.375024UNF0.0878 0.217 0.242 0.299 0.3140.437514UNC0.1063 0.296 0.321 0.428 0.4430.500013UNC0.1419 0.389 0.427 0.562 0.5810.500020UNF0.1599 0.400 0.444 0.541 0.5670.562512UNC0.182 0.503 0.548 0.717 0.7410.625011UNC0.226 0.624 0.683 0.891 0.9190.625018UNF0.256 0.624 0.708 0.853 0.8920.750010UNC0.334 0.910 1.00 1.29 1.330.750016UNF0.373 0.924 1.04 1.24 1.300.8750 9UNC 0.462 1.25 1.38 1.77 1.830.875014UNF0.509 1.26 1.43 1.71 1.781.0000 8UNC 0.606 1.66 1.82 2.33 2.401.1250 8UN 0.790 1.92 2.14 2.95 3.041.2500 8UN 1.000 2.65 2.91 3.65 3.751.3750 8UN 1.233 3.22 3.55 4.41 4.541.5000 8UN 1.492 3.86 4.26 5.25 5.41


1.6250 8UN 1.78 4.55 5.03 6.17 6.351.7500 8UN 2.08 5.30 5.87 7.15 7.361.8750 8UN 2.41 6.10 6.76 8.20 8.452.0000 8UN 2.77 6.96 7.72 9.32 9.612.2500 8UN 3.56 8.85 9.83 11.78 12.162.5000 8UN 4.44 10.96 12.19 14.53 15.012.7500 8UN 5.43 13.30 14.80 17.57 18.163.0000 8UN 6.51 15.85 17.67 20.88 21.593.2500 8UN 7.69 18.62 20.79 24.47 25.323.5000 8UN 8.96 21.63 24.16 28.37 29.363.7500 8UN 10.34 24.83 27.78 32.51 33.674.0000 8UN 11.81 28.28 31.65 36.96 38.30aAll areas given in square inches.

Bolt nominal tensile strength is 38,400 lb.

Bolt stripping strength calculated from formula (2)is 63,700 lb.

Nut stripping strength calculated from formula (5)is 66,900 lb.

This bolt will fail in tension before either it or thenut strips.

If the same bolt has an undersize pitch diameter(less material in the ridges) by an amount equal to


the full tolerance of 0.0035 in., then

Bolt tensile strength is reduced by 2.4% to 37,500lb.


TABLE 6 Selected Metric M Series Thread Stress and Shear AreasaBolt thread strip area,

LE = 1 diam.Nut thread strip

area, LE = 1 diam.Thread Tensile stress area 6H/6g 6H/4g6g 6H/6g and 6H/4g6g

M5 × 0.8 14.183 35.366 37.085 49.994M6 × 1 20.123 51.875 54.174 73.145M8 × 1.25 36.609 97.290 100.74 134.63

M10 × 1.5 57.990 155.80 160.73 214.77M12 × 1.75 84.267 227.72 234.86 313.41M14 × 2 115.44 313.94 323.24 434.47M16 × 2 156.67 417.55 429.92 569.12M20 × 2.5 244.79 665.58 682.06 907.76M22 × 2.5 303.40 814.67 1100.3M24 × 3 352.50 971.10 994.23 1320.2M27 × 3 459.41 1246.7 1674.3M30 × 3.5 560.59 1549.2 1582.5 2086.5M36 × 4 816.72 2271.5 2315.8 3026.2M42 × 4.5 1120.9 3120.2 3175.2 4168.3M48 × 5 1473.2 4119.5 4187.3 5482.0M56 × 5.5 2030.0 5680.3 5769.4 7510.2M64 × 6 2676.0 7483.6 7596.4 9847.3M72 × 6 3459.8 9574.2 9718.5 12478M80 × 6 4344.1 11922 12101 15419M90 × 6 5590.8 15217 15446 19534M100 × 6 6994.7 18856 19170 24059a All areas given in square millimeters.

Bolt stripping strength is reduced by 5.8% to 60,000 lb.


And if the nut has an oversize pitch diameter (less materialin the ridges) by an amount equal to the full tolerance of0.0045 in., then

Nut stripping strength is reduced by 6.5% to 62,600 lb.

The bolt will still fail in tension before it or the nut strips. Asmall reduction in strength will result, however.

10.2Lead (Including Helix)

The lead is the distance that a point on a thread will moveaxially when the fastener is rotated one turn in a perfectstationary mating thread. For a single-start thread, the lead isessentially the same as the thread pitch, i.e., the distancefrom a point on one thread ridge to the corresponding pointon the next thread ridge.

When the lead of a bolt thread is the same as the nominallead, the thread will engage fully into a mating thread withnear-perfect lead, such as in a GO thread gage. Each threadridge will fit into its mating thread groove over the length ofengagement. But when the bolt thread lead is significantlylonger or shorter, the mating thread will contact only someof the thread ridges and the thread ridges will not be able tofully engage all the grooves. (See Figure


Figure 9A bolt with a long lead thread that causes its functional diameter to be largerthan its pitch diameter. A short lead thread would have the same effect. Fornut threads with lead variations, the functional diameter is smaller than the

pitch diameter.

9.) For this condition, if we were to apply a GO threadgage with a variable pitch diameter, the GO gage pitchdiameter would have to be larger than the bolt threadpitch diameter in order to go on. The pitch diameter ofa GO gage that will snugly engage the product thread


is known as the functional diameter. As you can seefrom what we showed above, the functional diameterof a bolt thread with lead variations is greater than itspitch diameter. For a nut thread with lead variations,the functional diameter would be smaller than thepitch diameter.

If lead variation is excessive, it could prevent assemblywith the mating thread. Inspection with a GO gage orwith a functional diameter indicator will determine iflead variation is too great for assembleability.

10.3Flank Angles

The angles between the thread flanks and aperpendicular to the thread axis are known as flankangles. For a perfectly formed thread, the flank angle isequal to one-half the included angle of 60°, or 30°.

When the flank angle of a bolt thread is the same as thenominal flank angle, the thread will engage fully into amating thread with near-perfect angle, such as in a GOthread gage. Each thread ridge will fit into its matingthread groove over the length of engagement. Butwhen the bolt thread angle is significantly larger or


smaller, the mating thread will contact only near thethread crests or roots. Thread ridges will not be able tofully engage in the grooves. (See Figure 10.) For thiscondition, if we were to apply a GO thread gage with avariable pitch diameter, the GO gage pitch diameterwould have to be larger than the bolt thread pitchdiameter in order to go on. The functional diameter ofthe bolt thread with angle variations is larger than itspitch diameter, just as the bolt thread with leadvariations is. Similarly, for a


Figure 10A bolt with enlarged thread flank angles, which cause its functionaldiameter to be larger than its pitch diameter. Reduced thread angleswould have the same effect. For nut threads with angle variations, the

functional diameter is smaller than the pitch diameter.

nut thread with flank angle variations, thefunctional diameter would be smaller than thepitch diameter.


Just as a thread with excessive lead variationcould prevent assembly with its mating thread, anexcessive variation in flank angle could do thesame. Again, inspection with a GO gage or with afunctional diameter indicator will determine ifangle variation is too great for assembleability.

10.4What Happens to Thread Form under Load?

When torque is applied to a nut, the nut movesalong the bolt thread and is pressed against thesurface of the part to be fastened. This force at thebearing surface compresses the nut; the nut, inturn, transmits this force to the bolt and developsa tensile stress in the bolt. Since the nut iscompressed, its thread lead is reduced. Thetension in the bolt causes it to stretch, so itsthread lead is increased.

Before the load is applied to the nut, the threadleads of nut and bolt are the same. But now thatthe joint is loaded, the threads no longer match.


This puts an uneven load on the threads, with thehighest load concentrated at the threads nearestthe bearing surface of the nut. Localized yieldingresults in some of the load being transmitted tosubsequent threads. The problem is that highstresses are concentrated in just a few threads.Even if the threads do not strip, the fatigue life ofthe joint is reduced.

Under load, forces are transmitted from nutthread flanks to bolt thread flanks. When nut andbolt flank angles are the same, contact betweenthreads is across the full flanks and the effectiveload acts as if it were at the centers of the flanks.This force produces a moment


around the thread roots, like that of a cantileverbeam, and results in stresses concentrated at theroots.

If the flank angle of a thread ridge in the boltthread is smaller than that of the mating nutthread ridge, contact will be at the crest of thebolt thread and the cantilever beam stress at thebolt thread root will be approximately twice thatwhich was experienced when flank angles wereequal. But if the flank angle of the bolt thread islarger, contact between nut and bolt threads isnear the bolt thread root and cantilever beamstresses at the root are minimized. (See Figure11.) Minimizing the bolt thread root stresses isknown to reduce bolt fatigue. (It has beenrecognized that at the same time stresses at theroot of the bolt thread are minimized, stresses atthe root of the nut threads are increased. Nutsand other tapped parts are generally less likely to


fail from fatigue, however.)

In the 1960s, Standard Pressed Steel Company(now SPS Technologies, Inc.), on the basis of theknown effects of changes in thread form underload, developed modifications to the standardUNJ form thread (also a development of SPS) toimprove bolt fatigue life. One modificationreduced the lead of the bolt thread just enoughthat under load it would be the same as the leadof the loaded nut thread while still permittingassembly prior to application of the load. Theother modification increased the pressure flankangle of the bolt thread by 5° to ensure contact ofthe nut thread ridges near the bolt thread roots, tominimize bending stresses at the root. This threadis called the asymmetric thread and is used onbolts that must have the best possible fatigueperformance.


Figure 11Effect of flank angles on thread bendingmoment (F × arm). Thread deflection is

proportional to this moment.


Figure 12Thread form variations: tapered, out-of-round, crest/pitch diameter runout, and drunken (uneven helix).

10.5Other Thread Parameters

Deviations in other thread parameters may also have some effect on threadperformance. These deviations include excessive taper, out-of-round,crest/pitch diameter runout, and drunken (uneven helix) thread. They areillustrated in Figure 12.

References

Note: The standards listed below have been used for reference in thischapter. Each, however, is subject to periodic revision. Good engineeringpractice suggests that the most recently published version should always be


used for design or other projects.

1. ACME Screw Threads, ASME/ANSI B1.5-1988 (Reaffirmed 1994),ASME, New York, 1994.

2. Buttress Inch Screw Threads, ANSI B1.9-1973 (Reaffirmed 1985),ASME, New York, 1985.

3. Unified Inch Screw Threads (UN and UNR Thread Form), AmericanNational Standard ASME B1.1-1989, ASME, New York, 1989.

4. Screw Thread Standards for Federal ServiceSection 2, Unified InchScrew ThreadsUN and UNR Forms, FED-STD-H28/2B, Federal SupplyServices, General Services Administration, Washington, DC, Aug. 20,1991.

5. Screw Threads, Controlled Radius Root with Increased Minor Diameter,General Specification for, MIL-S-8879C, U.S. Government PrintingOffice, Washington, DC, July 25, 1991.

6. Metric Screw ThreadsM Profile, American National Standard ASMEB1.13M-1995, ASME, New York, 1995.

7. Metric Screw ThreadsMJ Profile, ANSI B1.21M-1978 (Reaffirmed1991), ASME, New York, 1991.

8. Nomenclature, Definitions, and Letter Symbols for Screw Threads,ANSI/ASME B1.7M-1984, Reaffirmed 1992, ASME, New York, 1992.

9. Unified Inch Screw Threads (UNJ Thread Form), ASME StandardASME B1.15-1995, ASME, New York, 1995.

10. Carbon and Alloy Steel Nuts, Metric, ASTM Specification ASTMA563M-1993, American Society for Testing and Materials, WestConshohocken, PA, 1993.


11. Screw Thread Gaging Systems forDimensional AcceptabilityInch and Metric ScrewThreads (UN, UNR, UNJ, M, and MJ), AmericanNational Standard ASME B1.3M-1993, ASME,New York, 1993.

12. Screw Thread Standards for FederalServicesSection 20, Inspection Methods forAcceptability of UN, UNR, UNJ, M, and MJScrew Threads, FED-STD-H28/20B, FederalSupply Service, General Services Administration,Washington, DC, Mar. 10, 1994.

13. Thread forms and torque systems boostreliability of bolted joints, Product Engineering,December 1977, p. 37.

14. Waltermire, W. G., A fresh look at a basicquestion: coarse or fine threads? MachineDesign, Mar. 17, 1960.

15. FastenersRecommended Tensile Stress Areas


15. FastenersRecommended Tensile Stress Areasfor External Threaded, National AerospaceStandard NAS 1348, Aerospace IndustriesAssociation of America, Washington, DC, 1979.

16. Areas for Calculating Stress or Load Valuesfor Metric MJ Externally Threaded Fasteners,Metric Aerospace Standard MA 1520, Society ofAutomotive Engineers, Warrendale, PA, 1987.

17. Machinery's Handbook, the 14th or anymore recent edition, The Industrial Press, NewYork.

18. Ellison, H. W., Effect of Nut Geometry onNut Strength, General Motors Corp., Warren,MI, 1970.

19. Formula for calculating the strippingstrength of internal threads in steel, Report toISO/TCI/WG4 by Sweden-Bultfabrike AB,1975.

20. Alexander, E. M., Design and strength ofscrew threads, Trans. Conf. Metric Mechanical


Fasteners co-sponsored by ANSI, ASME, ASTM,and SAE, Presented at American National MetricCouncil Conference, Washington, DC, 1975.

21. Gill, P., The Static Strength of ScrewThreads, G.K.N., UK.

22. Parisen, J. D., Length of ThreadEngagement in Nodular Iron, General MotorsCorp., Warren, MI, 1969.

23. Wiegand, H., and K. H. Illgner, Boltbarkeitvon Schraebenuer-bindungen mit ISO,Gewindeprofil, Konstruktion, 1967.


9Fastener Head Markings

Charles J. Wilson*Industrial Fasteners Institute, Cleveland, Ohio

1Introduction

Mechanical fasteners have a broad variety ofmarkings, and these markings may be indented orraised depending on the product configurationand the method of manufacture. Typically theyare used to identify the maker of the part and/orthe manufactured fastener capability.

When the manufacturer puts the company'sidentification symbol on a fastener, purchaserscan feel more confident about the product, for itmeans that the fastener maker stands proudly


behind what he has made. While many consensusstandards have required these markings forseveral years, the government and militarystandards often consider them optional. Theresult was to create a flood of fasteners thatcarried no manufacturer's marking and were thenintroduced into the stream of commerce in NorthAmerica. Although this of itself should not havecaused a problem, it did. The net result was theintroduction of millions of counterfeit, orsubstandard, fasteners that were not traceable toknown makers.

A major investigation by a number ofgovernment agencies determined that thecounterfeit fasteners that had been introducedeither had no markings or carried markings thatwere not traceable to a manufacturer. In somecases the performance markings did notcorrespond to the markings established byrecognized consensus performance standardssuch as those published by the ASTM (American


Society for Testing and Materials) or SAE(Society of Automotive Engineers). The result ofthe government investigation led to the passageof Public Law 101592 on November 16, 1990,which requires mandatory registration ofmanufacturer's markings and puts the force oflaw behind performance markings required byvarious consensus standards. In other words,products that are offered for sale with thesemarkings must meet the requirements of thestandards or specifications that define themeaning or engineering requirements for a givenperformance marking.

* With the Technical Staff of Industrial Fasteners Institute.


2Source Marking:Manufacturer's Identification

For many years, reputable fastener manufacturershave placed their company identification symbolson their products. Currently it is mandatory thatall externally threaded products 1/4 in. or largermanufactured to SAE, ASTM, IFI, or ASMEstandards carry markings of the manufacturingsourceunless, of course, the productconfiguration or hardness precludes reasonablemarking.

ASME B18.2.1, Square and Hex Bolts andScrewsInch Series, includes the followingparagraph:

2.8Identification Symbols


2.8.2Source Symbols

Each of the eight products included in thisstandard shall be marked to identify its source;manufacturer, or private label distributor.

There are many ways in which a manufacturer'sidentification symbol protects the best interestsof the purchaser.

Today's multibillion dollar North Americanfastener industry produces about 2 millionindividual types, sizes, and styles of fasteners. Alarge producer or distributor might actually carryan inventory of up to 45,000 standard line itemsat any given time. Thus, when it's time to replacean installed fastener, it is extremely convenientfor the purchaser to be able to identify themanufacturer by its source marking. Thepurchaser then knows with confidence that thismanufacturer has experience in making thatparticular fastener. And with all of the records


available, the distributor should be able to fill therequest in the shortest possible time.

The Industrial Fasteners Institute (IFI) has longmaintained a register of North Americanmanufacturer's markings to help facilitate theidentification of the manufacturer. Under theprovisions of the public law commonly known asthe Fastener Quality Act, all fastener makerswishing to introduce covered fasteners made tothe provisions of standards requiring such marksinto the stream of commerce in the United Statesmust register their respective markingidentification with the United States Patent andTrademark Office. Manufacturers that do nothave a mark will be assigned an alphanumericmark by that office. IFI has indicated that it willestablish an on-line service to help facilitateworldwide manufacturer identification.

3Grade Markings:


Fastener Capability

Grade marking identification indicates to theuser a known level of performance of a givenfastener. These marks may take many forms: adesign or symbol, a letter or letters, one or moreletters combined with one or more numerals, anumeral (or numerals) alone, a clockmark, orperhaps the unique part number.

In all cases, however, each marking is anaccepted indication by one of several nationallyor internationally recognized technical groupsthat a certain specific set of criteria have beenmet. These marks indicate that every fastener soidentified is manufactured in accordance with theapplicable specification. Technical societies andother bodies that develop fastener specificationsinclude

The American Society for Testing andMaterials (ASTM)The Society of Automotive Engineers (SAE)


The American Society of MechanicalEngineers (ASME)

ASME specifications are primarily dimensional;however, some of its performance specificationsalso define specific markings.


Other standards are published by the federalgovernment and some of its agencies specificallyfor fasteners used within the militaryestablishment (Department of Defense) and foraircraft and missiles (National AerospaceStandards Committee).

Whatever the source of the standard, however,the manufacturer is given specific requirementsto meet and the ultimate responsibility forplacing the performance capability mark at agiven location on the finished fastener. Thisresponsibility is correctly that of themanufacturer, because only the manufacturerknowns the fastener's historyraw material,manufacturing process (cold, warm, or hotheading), heat treating, plating, quality testing,acceptance, and packaging.

A fastener without a strength marking must beconsidered to be at the lowest common


denominator and at the lowest possible strengthlevel. Its capability may only be assumed to bethat of the basic material of manufacture.

The cost of a mechanical fastener is usually lessthan 5% of the in-place or assembled cost of theproduct. The largest fastener cost lies in theassembly time. A purchaser should alwaysexamine the total assembly time related to agiven fastener design to determine the overalleconomics, not just the cost of the fasteneralone.

For example, are there many material differencesthat will affect the fastener's performance,assembly, or installation? Is the fastener suppliera problem solver? Does he consistently look forsolutions to assembly problems? Purchasersmust make absolutely sure they consider allfactors related to the fastener for optimumeconomic assembly.

Once a manufacturer has properly marked the


product with regard to performance and source,the purchaser may buy with the assurance that ifthe fastener is correctly applied the assembledproduct joint will function as intended by itsdesign. Here is where the performance mark is asignificant contributor in helping to identify themechanical and chemical characteristics of agiven fastener. The fastener used in the jointassembly provides visual assurance of itscapability and origin.

4Strength Grades for Inch Series Fasteners

The primary standards that include strength grademarkings fall into four groups.

1. ASTM specifications include chemical andmechanical requirements for fasteners in all typesof engineering applications.

2. SAE publishes specifications that generallyinclude chemical and mechanical requirements


for fasteners used in automotive, truck, off-roadvehicle, and other engineering applications.These standards are under the jurisdiction of theSAE Fasteners Committee. Aircraft engine andother related SAE specifications are includedwithin a separate family under the jurisdiction ofits Committee E-25.

3. GM (General Motors) specifications arerepresentative of widely used corporatespecifications. Similar types of corporatestandards are published by such corporations asFord, Deere, Caterpillar, IBM, Chrysler, andXerox.

4. ISO (International Organization forStandardization) includes markings only forchemical and performance requirements formetric fastener standards. With the drive towardglobal standards accelerating, these offer thelong-term possibility of a single world system offastener standards in metric modules.


In addition to the markings required by nationaland international documents, some fastener usersrequire special identification markings on theproducts they purchase for their own use. Thereare more than 80 different strength grades in inchseries standards for ferrous fasteners alone. Eachis identified by a corresponding grade marking.To meet these individual specificationrequirements, manufacturers must modify theirmanufacturing process and material in each caseto produce the correct fastener mechanicalproperties.


TABLE 1 Basic Externally Threaded Inch Series Strength GradesSAE/GMspecificationnumber

Gradedesignation

Fastenerdescription

Gradeidentification

mark

Mfr. I.D.symbol

required?

Applicablematerials

SAE J429 GM260M

Grade 2260M

Bolts,screws, andstuds

Nonerequired

Yes, exceptfor studs

Low or medium carbonsteel

SAE J429 GM280M

Grade 5280M



Medium carbonsteel, quenched andtempered

SAE J429 GM300M

Grade 8300M



Medium carbon alloy steel,quenched and tempered

Three basic inch series grade markings are illustrated below.*

These three identify an externally threaded product made of a givenmaterial and processed to meet specific requirements. For example, forthe three inch series markings identified in SAE J429 and GMstandards, the requirements are those listed in Table 1. The basic gradesdefined in these standards are also related to a given size or diameterrange, which in turn is related to the ability of given materials torespond to forming and/or heat treatment to achieve mechanicalproperties such as hardness, tensile strength, and proof load. These areoutlined in Table 2.

These three strength grades are probably the most widely referred to in


the inch series of fasteners. However, SAE J429 includes 10 strengthgrades, which offer a range of materials including low carbonmartensite and high manganese steels in addition to those listed inTables 1 and 2. Some of these other grades are widely used for studsand screws with captive washers commonly referred to as sems.

For internally threaded inch series product, ASTM A 563 establisheseight grades of nuts, which are designated by a letter or letters. Theseinclude grade O, A, B, C, D, and DH. The numeral 3 is used withGrades C and DH nuts, i.e., C3 and DH3, to indicate a weathering gradeor characteristic of the steel used in manufacture that enhancescorrosion resistance. These nuts offer a range of proof loads from65,000 to 175,000 psi.

Grades C, C3, D, DH, and DH3 require mandatory grade markings.Correct matching of these nuts with externally threaded productrequires that the nut proof load equal or exceed the minimum bolttensile strength. For fail-safe applications, the nut proof load shouldapproximate the maximum tensile strength of the bolt or screw, whichis often related to a

*Note: IFI is used only for illustrative purposes to represent the manufacturer'sidentification.


TABLE 2 Basic Externally Threaded Inch Series Mechanical RequirementsStrength requirements

Size HardnessSAE/GMspecificationnumber

Gradedesignation

Nominaldiameter

range(in.)

Proofload(psi)

Yieldstrength,

min.(psi)

Tensilestrength,

min.(psi)

Rockwellhardness

requirementsRemarks

SAE J429 Grade 21/43/4over3/41-1/2

55,00033,000

57,00036,000

74,00060,000

B80B100B70B100

GM 260M requires stress relief of cold headed bolts andscrews at 875°F (470°C) min.

GM 260M 260M

#6 thru3/4over3/41-1/2

55,00033,000

57,00036,000

74,00060,000

B80B100B70B100

SAE J429

GM 280M

Grade 5

280M

1/41over

11-1/2#6 thru

1over

11-1/2

85,00074,000

85,00074,000

92,00081,000

92,00081,000

120,000105,000

120,000105,000

C25C34C19C30

C25C34C19C30

This is similar to ASTM A449, except that ASTM A449includes sizes up to 3 in. GM 280M requires identificationmarking on hex head bolts and screws only. SAE surfacehardness Rockwell 30N 45 max for sizes 1/41 and 30N50 max for sizes over 11-1/2.

SAE J429GM 300M

Grade 8300M

1/41-1/2 120,000130,000 150,000 C33C39

This is similar to ASTM A354 Grade BD, except thatASTM mandates only alloy steel while SAE allows carbonsteel by agreement for sizes thru 3/4 and high manganesesteel thru 7/16. SAE surface hardness Rockwell 30N 58.6max. GM 300M requires identification marking on hexhead bolts and screws only.


maximum hardness. Thus, a Grade C heavy hex nut that has aproof load of 144,000 psi is a good match for an ASTM A325bolt that has a maximum hardness of Rc34, which isapproximately 150,000 psi max tensile. With this relationship,the Grade C heavy hex nut will almost always break the bolt, theobvious failure, instead of stripping, which is never obvious andcan lead to catastrophic failure.

A summary of a few of the most popular nuts is presented inTable 3.

In metric, there are just seven property classes for carbon andalloy steel bolts, screws, and studs.

It is important to note that all externally threaded metricproducts in sizes M5 and larger will be marked with a newnumerical property class identification system. Because thesesymbols are completely different from inch markings, they serveas unique metric part identifiers.

The following section examines the performance markings thatare widely used for metric products.

5Property Classes:Metric Series Fasteners

All metric property class marks have their origin in ISO 898,which currently is divided into four parts as follows:


Number ofmarkings Standards

11 ISO 898-1 Mechanical properties of fastenersbolts, screws, and studs

9 ISO 898-2 Part 2: Nuts with specified proof load values

4 ISO 898-5

Part 5: Set screws and similar threaded fasteners not undertensile stress

6 ISO 898-6 Part 6: Nuts with specified proof load valuesfine pitch threads

In North America, many of these property classes wereeliminated because of strong economic motivation to establishsimplification and best economic choices for meeting designrequirements.

The basic standards used in North America for externallythreaded metric products are SAE J1199 and ASTM A568M. Inthe latter standard, the property classes are set forth for bolts,screws, and studs for general, structural, and mechanical service,as summarized in Table 4.

Property classes 3.6, 5.6, and 6.8 of ISO 898/1 are not includedin North American standards because no use is made of thesegrades. In the basic designation system, the digit(s) preceding thedecimal define one-hundredth of the approximate tensilestrength in megapascals (MPa). The first digit to the right of thedecimal is one-tenth of the approximate ratio of the minimumyield strength to the minimum tensile strength expressed inpercent. These designations are used for almost all metricexternally threaded fasteners throughout the world. The two


strength grade designations 8.8.3 and 10.9.3 are unique to NorthAmerican practice. The 3 on the right designates a weatheringsteel grade; the complete property classes are found in ASTM A325M and ASTM A 490M for metric high strength structuralbolting.

ASTM A 563M defines eight property classes of nuts that arewidely used with metric bolts. Table 5 is a summary of thestandard for hex and hex flange nuts for general applicationstructural and mechanical service. For full requirements, refer tothe complete standard.


TABLE 3 Common Inch Nut Grade Marking Requirementsa

Gradedesignation

Gradeidentification

mark

Mfr.'sI.D.

symbolrequired?

Applicablematerials

Nominaldiameter

range (in.)Proof load

(psi)

Hardnessrequirements(Rockwell)

Grade C Yes Carbon steelCoarsethreadheavyhex: 1/44

Non-zinc-coatedand zinc-coated:144,000

B78C38

Grade C3 Yes Atmospheric corrosion resistant and weatheringcharacteristic steel

Coarsethreadheavyhex: 1/44

Non-zinc-coatedand zinc-coated:144,000

B78C38

Grade D D Yes Alloy steel

CoarsethreadHex: 1/41-1/2Heavy hex:1/44Hexthick:1/41-1/2

Non-zinc-coatedand zinc-coated:135,000150,000150,000

B84C38

Fine threadHex: 1/41-1/2Heavy hex:1/44Hex thick:1/41-1/2

Non-zinc-coatedand zinc-coated:135,000150,000150,000

Grade DH DH Yes Alloy steel, quenched and tempered

CoarsethreadHex: 1/41-1/2Heavy hex:1/44Hex thick:1/41-1/2

Non-zinc-coated:150,000175,000175,000

Zinc-coated:150,000150,000175,000

C24C38

Fine threadHex: 1/41-1/2Heavy hex:1/44Hex thick:1/41-1/2

Non-zinc-coated:150,000175,000175,000

Zinc-coated:15,000150,000175,000

GradeDH3 DH3 Yes Atmospheric corrosion resistant and weathering

characteristic steel, quenched and tempered

CoarsethreadHex: 1/21Heavy hex:1/44

Non-zinc-coated:150,000175,000

Zinc-coated:150,000150,000

a See ASTM A 563 for a complete summary of the eight nut grades.


TABLE 4 Externally Threaded Metric ProductsaGradedesignation

Gradeidentification

mark

Mfr.'s I.D.symbol

required?Applicable materials

Nominaldiameter range

(mm)4.6 4.6 Yes Low or medium carbon steel 5100

4.8 4.8 Yes Low or medium carbon steel,partially or fully annealed 1.616

5.8 5.8 Yes Low or medium carbon steel,cold worked 524

8.8 8.8 Yes Medium carbon steel,quenched and tempered 1672

8.8 Yes Low carbon martensite steel,quenched and tempered

8.8.3 8.8.3 Yes Weathering steel, quenchedand tempered 1636

9.8 9.8 Yes Medium carbon steel,quenched and tempered 1.616

9.8 Yes Low carbon martensite steel,quenched and tempered

10.9 10.9 Yes Medium carbon steel,quenched and tempered 520

10.9 Yes Medium carbon alloy steel,quenched and tempered 5100

10.9.3 10.9.3 Yes Weathering steel, quenchedand tempered 536

12.9 12.9 Yes Alloy steel, quenched andtempered 1636

a See ASTM F 568M for complete requirements.



(table continued from previous page)Strength requirementsProofload

(MPa)

Yieldstrength,

min.(MPa)

Tensilestrength,

min.(MPa)

Hardnessrequirements

VickersRockwell Remarks225

310

380

240

340

420

400

420

520

120220

130220

160220

B67B95

B71B95

B82B95

Grades 4.6, 4.8, and 5.8: Bolts and screwssmaller than 5 mm diameter or with slotted orrecessed heads need not be marked. Studssmaller than 12 mm diameter need not bemarked.

600 660 830 255336C23C34 Bolts and screws with slotted or recessed headsneed not be marked.

600 660 830 255336C23C34Bolts and screws with slotted or recessed headsneed not be marked. Heads may also havespecial marks indicating weathering-type material.

650 720 900 280360C27C36

Product smaller than 5 mm diameter need not bemarked. Bolts and screws with slotted orrecessed heads need not be marked. Studssmaller than 12 mm diameter have a special grademark.

830 940 1040 327382C33C39Bolts and screws with slotted or recessed headsneed not be marked. Studs smaller than 12 mmdiameter have a special grade mark.

830 940 1040 327382C33C39

Bolts and screws with slotted or recessed headsneed not be marked. Studs smaller than 12 mmdiameter need not be marked. Head may alsohave special marks indicating weathering-typematerial.

970 1100 1220 372434C38C44

Product smaller than 5 mm diameter need not bemarked. Bolts and screws with slotted orrecessed heads need not be marked. Studs


smaller than 12 mm diameter have a special grademark.


TABLE 5 Metric Hex and Hex Flange Nut Grade Designations

Gradedesignation

Gradeidentification

mark

Mfr.'sI.D.

symbolrequired?

Applicablematerials

Nominaldiameter

range(mm)

Strengthrequirementproof load

(MPa)

Remarks

5 5 No Carbonsteel

Hex,style 1:

1.645681012162036

Heavyhex:

42100

520580590610630

630

Over-tapped

465470490500

500

Nuts in thread diameter of M4 and smaller need notbe marked. Proof loads are for bare or zinc-coatednuts. Zinc-coated nuts must be overtapped. SeeA563M Sec. 7.8.

9 9 No Carbonsteel

Hex,style 2:

34Hex,style 2 orhexflange:

56810121620

Hex,style 2:

2436Heavyhex:

42100

900

915940950920

920

920

Nuts smaller than 4 mm diameter need not bemarked.

10 10 YesAlloy steel,quenchedandtempered

Hex,style 1:


121620

Hex,style 1:2436

1040

10501060

1060

Nuts smaller than 4 mm diameter need not bemarked.



Gradedesignation

Gradeidentification

mark

Mfr.'sI.D.

symbolrequired?

Applicablematerials

Nominaldiameter

range(mm)

Strengthrequirementproof load

(MPa)

Remarks

12 12 Yes

Alloy steel,quenchedandtempered

Hex,style 2:


56810121620

Hex,style 2:

2436Heavyhex:

42100

1150

116011901200

1200

1200

Over-tapped

920930950960

960

960

Proof loads given are for bare or zinc-coated nuts.Sinc-coated nuts must be overtapped. See A563MSec. 7.8. Nuts smaller than 4 mm diameter need notbe marked.

8S 8S Yes Carbonsteel

Heavyhex:

12361075 Nuts smaller than 4 mm diameter need not be

marked.

8S3 8S3 Yes

Weatheringsteel,quenchedandtempered

Heavyhex:

1236 1075Head may also have special marks indicatingweathering-type material. Nuts smaller than 4 mmdiameter need not be marked.

10S 10S YesAlloy steel,quenchedandtempered

Heavyhex:

1236 1245

Over-tapped1165

Proof loads given are for bare or zinc-coated nuts.Zinc-coated nuts must be overtapped. See A563MSec. 7.8. Nuts 4 mm in diameter and smaller neednot be marked.

10S3 10S3 Yes

Weatheringsteel,quenchedandtempered

Heavyhex:1236 1245

Over-tapped1165

Head may also have special marks indicatingweathering-type material. Nuts 4 mm in diameter andsmaller need not be marked.


The S designations are unique and indicateintended structural applications, and the 3appearing as the last digit indicates a weatheringgrade of steel. Generally, the metric system bolt-nut combinations are based on the markingsystem, i.e., class 10 nuts may be used withproperty class 10.9 bolts or bolts of lowerproperty classes.

6Summary

While this chapter has been brief, it highlights thesignificance of property class and strength grademarkings and their relative importance in initialapplication and subsequent overhaul andmaintenance that may be required. Further, themanufacturer's identification mark is importantto ensure source identification and quality. Usersshould always obtain a complete copy of the


most current standard or specification toestablish all of the requirements.

Appendix:Sources of Fastener Specifications

AMS

Obtain aeronautical materialspecifications from:Society of Automotive Engineers400 Commonwealth DriveWarrendale, PA 15096

ANSI

American National Standards Institute1430 BroadwayNew York, NY 10018 (or from theASME)

ASME

American Society of MechanicalEngineersUnited Engineering Center345 East 47th StreetNew York, NY 10017


ASTM

American Society for Testing andMaterials100 Barr Harbor DriveWest Conshohocken, PA 194282959

BS

British Standards InstitutionSales BraanchNewton House101/113 Pentonville Rd.London N1, England

FF,QQ,GGG

Obtain federal specifications from:Commanding OfficerNaval Publications and Forms Center5801 Tabor AvenuePhiladelphia, Pa 19120

IFIIndustrial Fasteners Institute1105 East Ohio BuildingCleveland, OH 441142879

ISOAmerican National Standards Institute1430 BroadwayNew York, NY 10018


MIL

Obtain military specifications from:Commanding OfficerNaval Publications and Forms Center5801 Tabor AvenuePhiladelphia, PA 19120

NASNational Standards Association1321 Fourteenth Street N.W.Washington, DC 20005

SAESociety of Automotive Engineers400 Commonwealth DriveWarrendale, PA 15096


10Computing the Strength of a Fastener

Joseph BarronNewport News Shipbuilding, Newport News,Virginia

1Introduction

The strength of a threaded fastener is the mostbasic consideration in the design or analysis ofbolted joints. However, as with manyengineering problems, the first step in thecalculation procedure is to determine the actualor expected loading conditions. Once the loadingconditions are defined, the threaded parts can beevaluated for acceptability. This chapterdiscusses the issues that have to be considered in


evaluating the strength of a threaded fastener.The loading conditions are defined in Section IVof this handbook. The techniques presented hereare applicable to threaded connections as well asthreaded fasteners.

2Possible Failure Modes

Most threaded fasteners that are installed and putinto service never fail. However, many fastenersdo fail. It is the designer's responsibility to selectthe correct material and size combination for agiven application. The first step in the selectionprocess is to understand the possible failuremodes. Then it will be possible to developcalculation procedures to avoid these failures.There are several static failure modes for athreaded connection. The governing mode isdetermined by the combination of loadingcondition, thread dimensions, and materialproperties. For connections subject to axial


loading, failure can occur by

1. Stripping of external threads

2. Stripping of internal threads

3. Tensile failure of external thread

For connections subject to shear loading, failurecan occur by

4. Shearing of external fastener

These four failure modes represent catastrophicfailure of the threaded parts. The basic premiseof design is to preclude such failure. Moreover,there are usually criteria that define ''failure"


prior to the onset of actual part breakage. This could be thread yielding, which may makedisassembly and subsequent reassembly impossible. Or it could be a maximum allowablestress imposed by a design specification.

3Basic Design Rule

The preceding section discussed the various failure modes for threaded connections.Thread stripping failures should be avoided. In the worst case, such failures render theparts useless. In the best case, expensive and time-consuming repairs are usually required.Also, thread stripping failures may not be readily apparent during installation orsubsequent inspections. A broken bolt is much more noticeable than a bolt that isassembled in a stripped hole. Therefore, threaded connections should be designed using thefollowing rule:

Axially loaded threaded connections should fail by tensile failure of the external thread, not by thread stripping.

This rule requires that the stripping strength of both the internal and external threadsexceed the tensile strength of the external thread. However, as with many rules, there areexceptions. Some threaded connections are not intended to develop the full strength of theexternally threaded part. An example would be a bearing retaining nut used to secure a ballbearing on a shaft. Here, the nut must only withstand the maximum axial design load of thebearing, not the tensile load of the shaft.

4Tensile Loading

The tensile strength of a threaded fastener could be computed like that of any othermachine part. The cross-sectional area multiplied by the material strength would equal thestrength of the fastener. The tensile area of the body or unthreaded portion of a bolt orscrew would simply be its cross-sectional area.

where

AB=body area (in.2, mm2)

DB=

diameter of body (in.,mm)


However, the tensile area of an external thread is not readily apparent upon inspection. Onecould use the minor diameter. This cross section is defined as the root area (AR) and iscalculated as follows:

where

AR=root area (in.2, mm2)

D =nominal size (in.)d =nominal size (mm)p =thread pitch (mm)

n =number of threads per inch (in.-1)


Yet when threaded fasteners or threaded specimens are tensile tested, they exhibit strengthsthat are greater than that predicted by the material strength and the root area. These partsbehave as if they had a larger cross-sectional area. An empirical formula was developedthat accounts for this phenomenon. The tensile area is based on a diameter midwaybetween the minor and pitch diameters and is calculated as follows [1,2]:

where AT = tensile area (in.2, mm2).

While the tensile area is the most appropriate for tensile loading conditions, some designstandards and codes still specify that the root area be used. The use of this smaller area"builds in" an additional factor of safety. The above tensile area equations are applicable tothe standard Unified and metric thread forms. Thread forms with a defined root radius,known as J threads, require different equations because they have larger external threadroot diameters. The tensile area of J threads is calculated as follows [3,4]:

where ATJ = J thread tensile area (in.2, mm2).

SAE standard MA 1520 [4] provides equations for the tensile area of MJ threads thatdifferentiate between threads that are roll threaded before and those roll threaded after heattreatment. However, these values differ by only a few percent from those calculated usingEq.(7).

The preceding discussion presented equations for the tensile and root areas of an eternalthread. The equations presupposed the normal geometry of a solid externally threaded part.For parts that do not meet this definition, the cross-sectional areas would have to accountfor any holes, slots, flutes, or other changes in geometry.

The number of threads in the grip can affect the tensile strength of fasteners. There appearsto be a length effect that must be considered. Most commercial bolts and screws are testedwith six threads exposed in the grip. For this condition, the standard tensile area equation


produces acceptable results. However, most aerospace bolts and screws are tested with twoto three threads in the grip. This situation requires the use of a larger cross-sectional areaas defined by Eqs. (6) and (7) even if the fastener does not have J threads.

5Shear Loading

Shear is loading transverse to the axis of the fastener. This is illustrated in Figure 1. Theshear strength of a fastener would be equal to the shear area multiplied by the shearstrength.

where

SSS=shear strength of fastener (lb, N)

S =material shear strength (lb/in.2,N/mm2)

As shown in Section 4, fastener shear area is smaller than the tensile area. Also, the shear


Figure 1Bolt loaded in shear with a single shear plane passing through threaded region.

strength of typical fastener metals is approximately 60% of the tensile strength. Therefore,shear loading is not an efficient use of material. To avoid subjecting a fastener to shear, itis recommended that shear pins or another design feature be used to resist the shear forces.The preceding assumed a single shear plane. For applications with more than one fastenershear plane, the shear strength of the fastener would be equal to the total shear areamultiplied by the shear strength. An example of a fastener with three loaded shear planes isillustrated in Figure 2. The shear strength of this fastener would be

6Torsional Loading

Since most fasteners are preloaded by an applied torsional moment (torque), the externallythreaded fastener is subject to torsional stress. Examining the long-form torquepreloadequation described in Ref. 9, the input torque applied to the bolt head or nut (whichever isturned)

Figure 2


Bolt loaded in shear with two shear planes passing through body and onepassing through the threaded region.


is apportioned to three reaction torques. The three torques are bearing face friction, threadfriction, and bolt stretch (preload). The proportions are typically defined as 50%, 40%, and10%, respectively. Since the bolt "sees" the thread friction and stretch components, it issubject to 50% of the input torque. This torsion is resisted by the body and/or threads. Theshear stress due to torsion can be calculated as follows:

where

t =shear sterss (lb/in,2, N/mm2)T=torsional moment (lb-in., N·mm)

r =radius at point in question (in., mm)

J =polar second moment of inertia (in.4,mm4)

For fasteners with a central hole, the maximum shear stress reduces to

where

DO=

outside diameter (in., mm) (use external minor diameter forthreaded region)

DI=diameter of central hole (in., mm)

For solid fasteners, this reduces further to

Torsional failure is not uncommon. Many fasteners fail when they are "wrung off" atassembly or disassembly. Another troublesome instance of torsional failure is seen duringthe installation of interference fit threads. Here a stud that is slightly larger than normal isscrewed or "driven" into a normal size tapped hole. There is a definite interference at themating pitch diameters. The intent is to produce an assembly in which the stud becomes


semipermanently installed in the hole. Due to the thread interference, a driving torque isrequired to install the stud even though no preload is being produced. The shear stressdeveloped is defined by Eq. (12) except that the torsional moment would be equal to thestud installation torque. Approximate installation torques for medium carbon steel inchsize studs can be found in ANSI B1.12 [5]. Torsional failure of interference fit studsduring installation can be especially problematic when the stud is made of a low strengthmaterial or is a small nominal size.

7Combined Loading

So far, tensile and shear stresses have been discussed on an individual basis. Since mostfasteners are preloaded using an applied torque, it is necessary to examine the effect ofthese combined stresses on the fastener. Tests show that the breaking strength of a bolt isreduced 1520% during torque tightening [6]. This indicates that the torquing processapplies additional stress on the fastener above that produced by the preload. Fortunately,when the torque is removed, it is generally assumed that the torsional stress dissipatesquickly. Utilizing the distortion energy or von MisesHencky failure theory that isapplicable to ductile metals, the combined or effective stress can be calculated as


where

sc=

combined stress (lb/in.2,N/mm2)

s =tensile stress (lb/in.2, N/mm2)

t =torsional stress (lb/in.2,N/mm2)

This combined stress can then be compared to the material's yield strength or other definedallowable tensile stress. Another method of evaluating combined stresses is to useinteraction equations. Here, the shear and tensile stresses are compared with theirrespective allowable values. These load ratios, Rs and RT, respectively, are then combinedin the interaction equation. Satisfying the equation predicts a safe design. A number ofinteraction equations have been proposed [7]. These are illustrated in Figure 3. Byinspection, it is evident that some equations are more conservative than others. MIL-HDBK-5 [8] proposes the use of . By manipulation of Eq. (13), theinteraction equation for the distortion-energy theory incorporating a factor of safety is

where

RS =Shear stress ratio = t/tAllowRT =tensile stress ratio = s/sAllowtAllow

=allowable shear stress (lb/in.2,N/mm2)

sAllow=

allowable tensile stress (lb/in.2,N/mm2)

f =factor of safety

Regardless of which interaction equation is used, it is clear that an increase in the level ofone stress must be accompanied by a decrease in the level of the other stress.


Figure 3Plots of interaction equations for

combined shear and tensile stresses.Equation (a) is the most conservative.

(a) ; (b) ;(c) ; (d) .


8Bending Stress

In Section 7 it was shown how torsional moments increasethe stress on a fastener. There is another source ofadditional stressbending. Bending can be introduced intothe grip of a fastener by a number of conditions such as

Joint bearing surface nonperpendicular to hole axisFastener bearing surface nonperpendicular to fasteneraxisJoint members transversely misalignedFastener threads eccentric to bodyApplied loads such as flange rotation or joint prying

The amount of stress generated by an applied bendingmoment can easily be calculated from the familiar flexureformula

where

M=

applied bending moment (lb-in.,N.mm)

c =outer radius (in., mm)


I =second moment of area (in.4, mm4)

Inspection of this formula explains why bending should beavoided. Fasteners are typically slender. Therefore, themoment of inertia of a fastener's cross section is rathersmall. This situation produces a large stress for a givenmoment. The resulting bending stress must then be addeddirectly to the preload or service load tensile stress in thefastener. Reference 9 contains a thorough treatment ofeccentrically loaded joints and their effect on fastenerstress. While it is possible to compute the additionaltensile stress created by a bending moment, in practice thedesigner seldom has enough information to complete thecalculations. Therefore, the most practical method toaccount for bending stress in design is to reduce thesources of bending and their severity.

9Modified Bodies

Most externally threaded fasteners (bolts, screws, studs)have solid, full-size bodies or shanks. It is sometimesadvantageous to employ a reduced cross section. Forinstance, reducing the cross section will reduce thestiffness. (See Chapter 11.) This is sometimes useful toincrease the fatigue life of the fastener. A reduced cross


section can also be used to reduce the strength to apredetermined level. Therefore, the fastener would fail at aspecific location. There are a number of ways to modifythe body of a fastener. These include

1. Reducing the diameter of the body

2. Drilling an axial hole in the body

3. Machining flutes into the body

4. Machining slots into the body or threads

In all these cases, the actual dimensions of the fastenerneed to be considered in order to calculate the strength ofthe fastener.

10Dynamic Loading

When the stress level in a part is held constant, it is said tobe statically loaded. For tensile loading, failure occurswhen the tensile strength of the material is exceeded. Whenthe stress


level varies as a function of time, the part is said to be dynamically loaded. Failure underdynamic conditions is termed fatigue failure or just fatigue. It is important to predictfatigue failures. Experience has shown that dynamically loaded parts fail at average stresslevels well below the tensile strength or even the yield strength of the material. Thissection addresses why fasteners are especially susceptible to fatigue, methods to reducethis susceptibility, and how to determine whether a fastener is safe from fatigue failure.The reason that parts fail by fatigue while loaded to seemingly low stress levels is thatstress concentrations are present. At these locations, which are typically at changes in crosssection and sharp corners, the localized stress is many times the average stress. Therefore,the localized stress can easily exceed the yield strength. Over the course of many loadcycles, a crack initiates where there is a stress concentration and grows under this highstress until failure occurs. The thread roots, thread runout, and head-to-body fillet areas onexternally threaded fasteners are all areas of stress concentration. Standard threads havestress concentration factors between 4 and 10 [10]. The nonuniform nature of thread loaddistribution in engaged threads also contributes to fastener fatigue. The first thread at thenut bearing face is subjected to the majority of the thread load as well as the full cross-sectional load. With the combination of highest average stress and the stress concentrationof the thread root, the first thread is the location for most fastener failures. It would seemthat threaded fasteners are fatigue failures waiting to happen. However, there are methodsto improve the fatigue performance of fasteners. An obvious consideration is to reduce thestress concentrations. The standard UN and metric thread forms can have sharp corners atthe thread roots. The UNJ and MJ thread forms have a specified root radius. These threadforms have three to four times greater fatigue life than the other forms. There are alsoproprietary thread forms that are even more effective [11]. Since fatigue failure is due totensile stress, the presence of a residual compressive stress would reduce the magnitude ofthe tensile stress, thus improving fatigue performance. Most threads are produced bythread rolling. This cold working introduces a residual compressive stress in the threadroots, exactly where it is needed. However, many fasteners require heat treatment, whichwould reduce the residual stress. That is why most aerospace applications requiring longfatigue life use fasteners that are roll threaded after heat treatment. Nonperpendicularity ofbearing surfaces is another source of fatigue problems. Test results show that an angularityof only 0.5° reduces fatigue life by almost one-half while 2° reduces it to almost zero[11].

Whether or not a fastener is safe from fatigue failure can be determined through the use ofa fatigue diagram. Here, the mean stress (fastener preload) is plotted against the dynamicor alternating stress. Next is is necessary to define the failure criteria. There are severalfailure theories. The most commonly used is the Goodman criterion. Figure 4 is a fatigue


diagram including the Goodman line. Only points below that line are safe. The Goodmanline extends from the endurance limit at zero mean stress to the tensile strength at zeroalternating stress. This indicates that, just as in Section 7 for combined tensile andtorsional stresses, a decrease or increase in the alternating stress must accompany anincrease or decrease, respectively, in the mean stress. The likelihood of failure is evaluatedby plotting the mean and alternating stress conditions for the fastener. The mean stress isthe in-service fastener preload. Determination of this value is described in Chapters 18 and36. Fastener alternating stress is not the full alternating load applied to the joint; it is thestress due to that portion of the alternating applied load that the fastener experiences. Thisvalue is determined by using joint diagrams as described in Chapter 36. The Goodman linecan also be expressed using an interaction equation. Expressing the alternating and meanstresses as ratios of their respective allowable values, the interaction equationincorporating a factor of safety would be


Figure 4Fatigue diagram using the Goodman line for demarcation of safety. Points

below that line are considered safe.

where

Ra=alternating stress ratio = sa/Se

Rm=mean stress ratio = sm/ST

sa =alternating stress (lb/in.2, N/mm2)

Se =material endurance limit (lb/in.2,N/mm2)

sm =mean stress (lb/in.2, N/mm2)SY=

material yield strength (lb/in.2,N/mm2)

See Chapter 37 for a further discussion of fatigue.

11Thread Shear Area


Threaded connections transfer axial loads between the external and internal threads bycontact of the thread flanks (see Fig. 5). The thread shear area of a fastener is controlledmainly by the mating thread. The relative strengths of the two parts define where the shearforce will be applied. When the strength of the internal thread material greatly exceeds thatof the external thread, thread failure will occur by stripping of the external thread. Theexternal threads shear at a diameter established by the minor diameter of the internal thread(see Fig. 6a). It is possible to calculate the shear area of the external thread at this diameteras follows [1,2]:

where

ASs =external thread shear area (in.2, mm2)LE =length of engagement (in., mm)D1max

=maximum minor diameter of internal thread(in., mm)

d2min=

minimum pitch diameter of external thread(in., mm)


Figure 5Cross section of assembled

screw threads showing transferof axial load by contact of

mating thread flanks.

When the strength of the external thread material greatly exceeds that of the internal thread,the assembly will resist the axial load through shearing of the internal threads at a diameterestablished by the major diameter of the external thread (see Fig. 6b). It is possible tocalculate the shear area of the internal thread at this diameter as follows [1,2]

where

ASn =internal thread shear area (internal thread shear area(in.2, mm2)

dmin =minimum major diameter of external thread (in., mm)D2max

=maximum pitch diameter of internal thread (in., mm)

Inspection of Figure 6 reveals an important condition. The shear areas are different at eachof the failure locations. For example, given a 1/2-13UNC-2A/2B combination, theexternal and internal thread shear areas per unit length of engagement would be 0.779 and1.123


Figure 6(a) External thread shearing

at minor diameter ofinternal thread.

(b) Internal threadshearing at major diameterof external thread. Dashed

lines show location ofthread shear.


in.2/in., respectively. The internal thread shear area is over 40% larger than the externalthread shear area. Regarding thread dimensions, it is possible to produce more shear areaby decreasing the thread dimensional tolerances (i.e., using a higher class of fit). Forexample, the unit internal thread shear area for a 1/2-13UNC-3A/3B combination is 1.164in.2/in. But this is only 3.7% larger than the value for the previous 2A/2B example.Therefore, the use of higher thread classes is not an efficient method of producing morethread shear area. There is one application where altering thread dimensions does producea significant increase in thread shear area. For the case of external thread area, the minordiameter of the internal thread is a prime factor. The minor diameter of a tapped hole iscontrolled by the tap drill or reamer used to prepare the stock for tapping. Minor diameterstabulated for standard Unified and metric screw threads will produce internal threadheights of 6683% of the height of a sharp V thread. The height of a sharp V thread, H, isequal to 0.8660254p. Decreasing the tap drill or reamer size will decrease the minordiameter, thereby increasing the external thread shear area. Equations defining thepercentage of thread height for a given tap drill hole size are as follows [12]:

where

Z=

percentage of sharp V thread height(percent)h=tap drill hole size (in., mm)

Also, by rearranging terms, the above equations can be used to calculate the required holesize for a given percentage of thread height. However, this procedure has disadvantages.Increasing the percentage of thread height requires more power to turn the tap becausemore material is being removed. Also, the tap driving torque increases dramatically. Thisincreases the torsional stress in the tap, leading to tap breakage. For example, the MetalCutting Tool Institute found that for a 3/8-16 tapped hole in AISI 1020 steel, the tappingtorque doubled for an increase in thread height of 6072% and almost tripled at 80% [12].Tables of tap drill sizes for inch and metric screw threads that produce standard minordiameters can be found in many sources [12,13].

12Length of Engagement


Length of Engagement

The previous section provided equations to calculate the amount of thread shear area for agiven length of engagement. However, for most design applications, it is the minimumlength of engagement that must be determined. Referring back to the basic design rule inSection 3, the following equation can be written:

where

AS=

thread shear area (internal or external)(in.2, mm2)

ST=material tensile strength (lb/in.2, N/mm2)

This equation states that the shear strength of the engaged threads must be equal to orgreater than the tensile strength of the fastener. Therefore, by substituting the variousequations for AS, the length of engagement can be determined. This was the approach useduntil a more comprehensive approach was published by Alexander [6]. This methodprovides equations to calculate three failure loads: bolt breaking load, bolt stripping load,and nut stripping load. The


variables, of which length of engagement is one, are then selected until the basic designrule is satisfied (i.e., bolt stripping and nut stripping loads are greater than the bolt breakingload). The method considers such effects as internal thread dilation and bell-mouthing,thread bending, thread relative motion, and thread chamfering. The paper [6] also containsa Fortran computer program that computes the minimum nut height (length ofengagement). One useful feature of this program is the use of statistical theory. Instead ofusing worst-case values (i.e., minimum material condition), the program generates randomvalues consistent with typical production processes. This allows calculation of nut heightsthat are shorter than those produced by worst-case analysis but still statistically acceptable.The following non-computer-based treatment is adapted from the Alexander approach. Itconsiders nut dilation, thread bending, and thread chamfers. Since no statistical theory isemployed, the results are conservative with respect to answers generated by the program.

12.1Internal Thread Dilation

Due to the 30° flank angle of standard threads, axial loads produce a dilation of theinternally threaded part. An internally threaded part with thinner walls allows moredilation. This explains the strength advantage of heavy hex nuts over regular hex nuts. Adilation strength reduction factor is calculated as follows:

where

C1=

dilation strength reductionfactor

s =width across flats for nuts (in.,mm)

This equation is valid for values of s/D between 1.4 and 1.9 and is plotted in Figure 7.Note that no advantage is given for values of s/D greater than 1.9.


Figure 7Plot of dilation strength reduction factor.

The lower limit for s/D of 1.4 isestablished by the fact that the outside

diameter must be larger than the nominalscrew size in order to have an assembly.


12.2Thread Bending

The relative strengths of the internal and external threads determines which will bend. Asthe engaged threads bend, there is less effective thread shear area. Also, the bent thread hasa decreased flank angle, creating more dilation. The strength ratio is defined as

where

RS=strength ratio

Sn=

internally threaded part tensile strength (lb/in.2,N/mm2)

ss =externally threaded part tensile strength (lb/in.2,N/mm2)

Strength reduction factors C2 and C3 account for the bending of the external and internalthreads, respectively. These factors are calculated as follows:

where

C2=

external thread strength reductionfactor

C3internal thread strength reduction


=factor

12.3Tapped Hole Countersinks

Internal threads are usually chamfered or countersunk. While removing the feather edge ofthe thread reduces the possibility of cross-threading, it also reduces the amount of threadshear area in the length of engagement. Based on experimental results, Alexander foundthat for a nut with a 90° countersink, the length of threads in the countersink contributedonly 40% of the strength of an equal length without the countersink. Therefore,countersinks can be accounted for by modifying the length of engagement used to calculatethread shear areas as follows:

where

LE' =effective length of engagement (in.,mm)

LENom=

nominal length of engagement (in.,mm)

DCmax=

maximum countersink diameter (in.,mm)

D1min =minimum minor diameter (in., mm)

It is now possible to compare the three failure loads.


where

BBL=

bolt breaking load (lb,N)

BSL=

bolt stripping load (lb,N)

NSL=

nut stripping load (lb,N)

The factor of 0.6 in the stripping load equations denotes the common ratio between shearstrength and tensile strength for ductile metals. Comparison of the results of Eqs. (27)(29)indicates which failure mode will prevail. The variables are then manipulated until thedesired condition of bolt breakage is achieved.

13Depth of Engagement

Thread depth of engagement is defined as the radial overlap of the mating threads. Matedscrew threads are usually drawn with the axes of the internal and external threadscoincident. However, given sufficient tolerance and/or allowance, it is possible for themating threads to be eccentric enough to allow the threads to disengage on one side. Thisis important because the thread shear area equations of Section 11 assume concentricallylocated mating threads. A significant amount of eccentricity will reduce the strippingstrength of both the internal and external threads. The amount that mated threads can bedisplaced is [14]

where

ecc=

possible eccentricity of mated threads (in.,mm)

TD2=

tolerance of internal pitch diameter (in.,mm)

Td2tolerance of external pitch diameter (in.,


=mm)es =thread allowance (in., mm)

Maintaining sufficient depth of engagement is not a concern when standard threaddimensions are used. If it is necessary to specify a nonstandard thread, it is recommendedthat the following conditions be satisfied [14].

1. To provide for a minimum depth of engagement equal to 50% of thread height:

2. To provide for a minimum depth of engagement equal to 50% of the thread height onone side and zero on the other when the assembly is eccentric:

Loss of depth of engagement can also be caused by internal thread dilation and externalthread contraction. Obviously, these effects are more pronounced in thin wall parts.

14Proof Strength

The previous sections have discussed the calculation of strength for externally threadedparts as well as assemblies. These values are then compared with the expected joint loadingconditions. Given that most joints are assembled using standard nuts and bolts, it is notnecessary to calculate the strength of every fastener. For specified materials, nuts and boltsof standardized


dimensions will have specified strengths. Thisfact gave rise to the concept of proof strength.Proof strength is a rating. The proof strength of anut is not its axial stripping strength. The proofstrength of a bolt is not its tensile strength. Proofstrengths are specified loads or, in the case ofproof stresses, specified stresses that the fastenermust withstand without any permanentdeformation. For nuts, the axial load is appliedto the finished part using a hardened mandrel[15,16]. The load is equal to the proof stressmultiplied by the tensile area of the same sizeexternal thread. The nut must not strip or ruptureand must be removable from the mandrel withthe fingers. Failure to meet either of theseconditions indicates that permanent deformationhas occurred and is cause for rejection. Sincenuts come in various configurations, the proofload must be compatible with not only nominalsize and material but also nut height, width


across flats, slotting, and overtapping. Nutstandards typically specify the following relativeratings for nut proof loads:

Regular hex nut 100%Heavy hex nut 108110%Slotted hex nut 80%Overtapped hexnut 60%

Hex jam nut 60%

In order to provide a more severe proof load testto verify the integrity of hot-forged or heat-treated nuts, the cone proof test was developed.Here the load is applied to the bearing face of thenut with a truncated cone instead of a flat plate.The cone introduces a severe dilation of the nut.The test is used to verify that seams, laps, orcracks have not had a deleterious effect on theability of the nut to carry its rated load withoutstripping or rupturing. For bolts, the axial load isapplied to the finished part [15,16]. Again the


proof load is equal to the specified proof stressmultiplied by the tensile area of the bolt. The boltmust withstand the load without breaking orstripping. In addition, the length of the bolt ismeasured before and after loading. There can beno increase in length, which would indicatepermanent deformation. Failure to meet either ofthese conditions is cause for rejection.

References

1. ASME Standard B1.1, Unified Inch ScrewThreads (UN and UNR Thread Form), 1989.

2. ASME Standard B1.13, Metric ScrewThreadsM Profile, 1983.

3. National Aerospace Standard NAS 1348,FastenersRecommended Tensile Stress Areas forExternal Threaded, Aerospace IndustriesAssociation of America, May 30, 1979.

4. Metric Aerospace Standard MA 1520, Areasfor Calculating Stress or Load Values for Metric


MJ Externally Threaded Fasteners, Society ofAutomotive Engineers, Warrendale, PA, July1987.

5. ASME Standard B1.12, Class 5 Interference-Fit Thread, 1987.

6. Alexander, E. M. Analysis and Design ofThreaded Assemblies, SAE Paper No. 770420,1977.

7. Barrett, R. T., Fastener Design Manual, NASAReference Publication 1228, March 1990.

8. Military Handbook MIL-HDBK-5, MetallicMaterials and Elements for Aerospace VehicleStructures, Revision F, Nov. 1, 1990.

9. Bickford, J. H., An Introduction to the Designand Behavior of Bolted Joints, Marcel Dekker,New York, 1995.

10. VDI 2230, Systematic Calculation of HighDuty Bolted Joints, Joints with One CylindricalBolt, VDI Society for Product Development,


Bolt, VDI Society for Product Development,Design and Marketing, Committee for BoltedJoints, Verein Deutscher Ingenieure (VDI).Translation of German edition July 1986 by C.Junker and J. Newnham.

11. Product Engineering Report 4718, FastenerSeminar, SPS Technologies, Aerospace ProductsDivision, 1974.


12. Drilled Holes for TappingA Guide toSelection of Drills for Tapping, Metal CuttingTool Institute, 1986.

13. Federal Standard FED-STD-H28, AppendixA3, Screw-Thread Standards for FederalServices, Mar. 31, 1978.

14. Federal Standard FED-STD-H28/2B, Screw-Thread Standards for Federal Services, Aug. 20,1991.

15. ASTM Standard F606, Determining theMechanical Properties of Externally andInternally Threaded Fasteners, Washers, andRivets, 1986.

16. ASTM Standard F606M, Determining theMechanical Properties of Externally andInternally Threaded Fasteners, Washers, andRivets (Metric), 1995.


11Computing the Stiffness of a Fastener

Joseph BarronNewport News Shipbuilding, Newport News, Virginia

1Introduction

When threaded fasteners are used to assemble a joint, theyare usually preloaded so that there is a residualcompressive load placed on the joint members. Thefastener supplies this load by being stretched from its free-state length during the assembly process. This iscomparable to stretching a helical spring. Although thetypical fastener may not stretch as much as a spring, it doesbehave like a spring. Therefore, in order to characterize thebehavior of the fastener, it is necessary to determine themost important property of a spring; its stiffness or springrate. Also, it is possible to determine and control fastenerload by measuring the length of the fastener. Therelationship between fastener length and load is againdetermined by fastener stiffness.


2Effective Length

The process of determining fastener stiffness is based onsimple materials science principles. The standard tensiletest provides the necessary information. When an axialtensile load is applied to the ends of a straight cylindricalrod, it stretches. A graph of the applied tensile load versusthe resultant axial deflection for a given size specimen isshown in Figure 1. The linear portion of the curve can beexpressed as

Where

F=applied load (lb, N)

k =spring rate (lb/in.,N/mm)

d =deflection (in., mm)


Figure 1Diagram of loaddeflection curve for an elastic- plastic material.

The slope of the linear portion is the stiffness of the part.

This shows that the deflection is directly proportional tothe applied load. It should be noted that the loaddeflectioncurve is not universally applicable to all parts due togeometrical differences.

By converting the data to stress and strain, parts of any sizecan be analyzed. Load and deflection are related to stress


and strain as follows:

Where

s =applied stress (lb/in.2, N/mm2)A=cross-sectional area (in.2, mm2)

e =engineering strain (in./in.,mm/mm)L=original length (in., mm)

A graph of applied tensile stress versus resultant axialstrain is shown in Figure 2. Here, the linear portion of thecurve can be expressed as

Where

E=

elastic modulus (lb/in.2,N/mm2).


Figure 2Diagram of stressstrain curve for an elastic-plastic material. Theslope of the linear portion is the elastic modulus of the material.

The yield strength at 0.02% offset is defined at point C.

Of particular interest is the linear portion of the curve.Point A is defined as the proportional limit where thelinear relationship between stress and strain stops. Point Bis defined as the elastic limit. Beyond this point, thespecimen has been plastically deformed (strained) and will


not return to its original length when the stress is removed.Point C is defined as the yield point. This is typicallycharacterized by a specified amount of plastic strain,usually 0.2%. While many fasteners are intentionallypreloaded beyond the yield point, determination of thestiffness of a fastener is confined to the elastic region. Bycombining Eqs. (1)(4) and rearranging terms, an equationdefining the deflection can be written:

Further manipulation produces the equation defining thestiffness:

However, the preceding equation is applicable only toparts with a uniform cross section.

Threaded fasteners typically have significant changes ingeometry. Therefore, it is necessary to determine thedeflection of each segment and then add the values toobtain the total deflection. For example, given a fastenerwith three segments, the total deflection would be


It is sometimes easier to work in terms of the inverse of the stiffness, or 1/k. This is knownas the resilience or compliance and can be written

Therefore, the total resilience of the fastener can be written

where

f =resilience (in./lb, mm/N) =1/k.

It can be seen that the stiffness of a fastener is dependent on the number of segments, thearea and length of each segment, and the elastic modulus. Assuming that the fastener ismade of one homogeneous material, only the length and area are variables. Determinationof the length of each segment is straightforward except for those sections including thehead of the bolt or screw or the threaded area engaged with a nut or tapped hole. Here, thedesigner needs to know how much of the bolt head or depth of the tapped hole contributesto the length of the fastener being stretched. The amount of contribution of these areas tothe geometrical length affects the effective length of the fastener.

Experiments have shown that fasteners exhibit stiffnesses that indicate an effective lengthlonger than the grip length. VDI 2230 [1] specifies that a length equal to 0.4D, where D isthe nominal diameter, be added to the grip length to account for the contribution of thefastener head. It also specifies that a length equal to 0.4D be added to account for thecontribution of the engaged threads in nuts or tapped holes. Other sources recommendvalues from 0.3D to 0.6D. Bickford [2] recommends a value of 0.5D for both heads and


engaged threads. The amount of cross-sectional area in each segment is established asfollows:

where

D =nominal diameter (in.,mm)

d1=shank diameter (in., mm)

n =number of threads perinch


p =thread pitch (mm)

It is also necessary to account for any axial holes in fasteners. These are typically used toallow the insertion of stud heaters or length measurement rods.

where

dh=

diameter of hole (in.,mm)

The calculation of fastener stiffness, while based in part on experimental work, istheoretical. Since the stiffness is based on dimensional data, dimensional tolerances willhave an impact. For example, a 1/2-13UNC × 3 in. long hex head cap screw is assembledwith a hex nut to give a grip length of 2 1/2 in. Standard tolerances for the overall lengthand the length of threads for this combination would produce a variance of 4% in the valueof the stiffness. Another source of variation is in the elastic modulus. Many texts quote theelastic modulus of steel between 29 × 106 psi (200 GPa) and 30 × 106 psi (207 GPa). Thisis a 3% variation. Also, it has been found that fastener stiffness and hence deflection doesnot agree with experimental values when the fastener is ''short." In regard to stiffnesscalculations, fasteners should be considered "short" if the length-to-diameter ratio (I/D) isless than 5. However, these variations can easily be reduced if greater accuracy is required.It is a simple matter to produce load-deflection data for each lot of fasteners or even foreach fastener. Only you can determine if the added accuracy will offset the added costs.

3Example Stiffness Calculations

In these examples, a factor of 0.5 is used for the amount of contribution of bolt heads andthreaded areas to the effective length of the fastener.

Example 1: Hex head bolt and nut (see Fig. 3)


Figure 3A hex head bolt assembled with a hex nut. Thebody and thread lengths are shown along with

the effective length.


Figure 4A reduced body stud assembled with a nut at each end. Thebody and thread lengths are shown along with the effective

length.

Example 2: Double-ended reduced body stud andnuts (see Fig. 4)


Example 3: Socket head cap screw in a tappedhole (see Fig. 5)

Figure 5A socket head cap screw installed in a tapped hole. The bodyand thread lengths are shown along with the effective length.


Figure 6A stud with a heater hole installed in a tapped hole. The bodyand thread lengths are shown along with the effective length.

Example 4: Set stud with heater hole (see Fig. 6)

References

1. VDI 2230, Systematic Calculation of High Duty BoltedJoints, Joints with One Cylindrical Bolt, VDI Society forProduct Development, Design and Marketing, Committee forBolted Joints, Verein Duetscher Ingenieure (VDI).Translation of German edition July 1986 by C. Junker and J.


Newnham.

2. Bickford, J. H., An Introduction to the Design andBehavior of Bolted Joints, 3rd ed., Marcel Dekker, NewYork, 1995.


12Metric Fasteners

Bruno MarbacherBossard International Inc., Portsmouth, NewHampshire

1Introduction

Working with metric fasteners often dictates adifferent approach than working with inchfasteners. This is not to say that many things thatapply to inch fasteners do not apply to metricfasteners as well, but metric fastener standards,specifications, calculations, etc. have beendeveloped independently, taking into account themany viewpoints worldwide. At times there areclose similarities; at other times things are far


apart. Therefore one should acquire as muchinformation on metric fasteners as one possiblycan. In 1520 years from now we may be workingonly with metric fasteners.

1.1U.S. Customary Vs. Metric

In the 1970s U.S. companies skepticallyconverted to the metric system. Since the mid tolate 1980s the conversion process has beensteadily going forward. Nevertheless, consideringpublic resistance, the process will no doubtcontinue well into the next century.

1.2Converting to Metric Fasteners

Conversion to the metric system can be eitherhard or soft.

Hard conversions result in designs that are baseddirectly on the metric system. All aspects of thedesign, including the mechanical requirements,


dimensions and tolerances, testing specifications,and other functions are derived from the metricsystem. Hard conversions provide theopportunity to make use of all the possiblebenefits of metric product standards. The use ofstandards also makes it possible to introducesimplification and cost savings that can not onlyoffset the initial cost of converting to metric butalso reduce future costs for production, design,and purchasing. Hard conversions must becarefully planned and well coordinated,addressing the individual needs of the company.

Soft conversions maintain the original inchdesign but express measurements and data in themetric language, multiplying the inch units ofmeasurement by the appropriate metricconversion factors. Soft conversions should beapplied only when expensive tooling andproduction equipment cannot be replacedimmediately or when the period of transition tometric is limited. Using soft conversions for


fasteners is not a recommended practice.


Ultimately, everything should be "hard converted," because only thencan one take advantage of the numerous products that are available onthe market.

2The Metric System

2.1Metric Units and Terminologies

Working and designing with metric fasteners makes it necessary to useSI units (metric units of the Systéme International). Also, one shouldbe familiar with common symbols and abbreviations that are used withmetric measures. In this section we address only the metric units thatpertain to metric fasteners.

Glossary of Terms, Names, and Abbreviations Frequently Used inConnection with Metric Fasteners

newton (N) =unit for forceN/mm2 =newtons per square millimeter

pascal (PA) =unit of mechanical pressure/stressMPa =megapascal (1 MPa = 1 N/mm2)

degree Celsius(°C) =unit of temperature

meter (m) =SI unit for length

mega (M) =SI prefix for 1 million (106) units (e.g., 1 MPa = 1 ×106 Pa)

kilo (k) =prefix for 1000 (103) units (e.g., 1 kN (kilonewton)= 1000 N)

milli (m) =prefix for 0.001 (10-3) (e.g., 1 mm (millimeter) =


milli (m) =0.001 m)

micro (µ) =prefix for 0.000 001 (10-6) (e.g., 1 µm (micrometer)= 1 × 10-6 m)

pitch (P) =distance between crest of 2 adjoining threads in mmproperty class =equivalent to grade in inch terminology

Rp0.2 =stress at 0.2% at permanent set limit (yield stress)Rm =ultimate tensile strength

A1, A2, A4 =austenitic stainless steel grades for fastenersC1, C3 =martensitic stainless steel grades for fasteners

F1 =ferric stainless steel grade for fastenersSI =International System of Units

IT grades =international tolerance grades (fundamentaltolerances)

ANSI =American National Standards InstituteASTM =American Society for Testing and Materials

CEN =European Committee for Standardization EN-Standards

DIN =German Institute for StandardizationIFI =Industrial Fastener Institute

ISO =International Organization for StandardizationSN (VSM) =Swiss National Standards

2.2Linear Units

Although the metric system works with multiples and submultiples of10, in the technical field that includes fasteners, we mainly operatewith multiples and submultiples of 1000.


Figure 1Nominal dimensions of metric fasteners.

Metric fastener dimensions are expressedexclusively in millimeters, and it is standardpractice worldwide in the fastener industry toomit the millimeter unit symbol (mm). Whenwriting dimensions in millimeters it is sufficientto just state the dimension.

2.3Metric Dimensioning

Fastener nominal dimensions are mainlyexpressed in full millimeters or in increments of


onetenth of a millimeter (see Fig. 1). Metricdrawings, including fastener drawings, oftenshow only the nominal dimension with referenceto the applicable tolerance standard. Tolerancesthat divide for the overall tolerance specificationare normally listed right next to the nominaldimension (see Fig. 2).

3ISO Tolerance SystemTolerances for Limits andFits

Metric fastener and machine element tolerancesare frequently expressed in tolerance zones. Thetolerance zone symbols are composed of lettersand numbers, e.g., H7, h13, h8, m6. The letterindicates the tolerance position. Capital lettersare used to indicate tolerances on internalfeatures (holes). Lowercase letters are used toindicate tolerances for external features (shaft).The number indicates the tolerance range, theactual spread one can utilize.


Figure 2General linear tolerances.


Figure 3The ISO tolerance system

(limits and fits).

The ISO tolerance system is illustrated in Fig. 3.

There are a number of standards and publications available thataddress the tolerance for limits and fits in detail. These includeISO R/286, ANSI B 4.2, and DIN 7152.

The international standard ISO 4759 covers all the tolerancesthat are applicable to metric fasteners (Fig. 4 and Table 1). This


includes tolerances for limits and fits as well as geometrictolerances. These tolerances should be applied to nonstandardfasteners as well. The tolerances specified in ISO 4759 arefeasible and generally can be achieved without difficulty in themanufacture of fasteners. Specifying looser tolerances mayimpair the functionality of the fastener. Tighter tolerances willmake it more difficult to produce, driving up the manufacturingcost without affording any major benefits. ISO tolerances forsockets and slots are illustrated in Fig. 5 and listed in Table 2.

4The Metric Thread

Figure 6 illustrates the major dimensions of a metric thread.

The metric thread is identified by the capital letter M. The letteris followed by the nominal thread diameter. For example, M8indicates a metric thread of 8 mm diameter. (The letter Mshould be used only to identify a metric thread.)

The flank angle of a metric thread is 60° (Fig. 7). The "fineness"of the thread is then indicated by the pitch, defined as thedistance from the crest of one thread to the crest of theadjoining thread (see Fig. 8). In customary inch terms, onewould say that the finer the thread (smaller the pitch) the morethreads per inch it has.

The minimum metric thread root radius is specified as 0.125 ×the thread pitch.


Unless it is a fine thread, the pitch does not need to be spelledout in the thread designation. For example, a metric fine threadmight have the designation M8 × 1, whereas the equivalentmetric coarse thread would be designated M8.


Figure 4The ISO tolerance system as applied to metric fasteners.


Figure 5The ISO tolerance system applied to sockets, straight slots, and width across the flats.

4.1The Layout of the Metric Thread in Technical Drawings

In technical drawings the individual threads of a metric thread are not shown.

4.1.1External Threads

Figure 9 is an example of a technical drawing of an external metric thread. The major diameter isshown in bold lines. The minor diameter is then shown with thin lines. The end of the usefulthread is indicated by a bolt line. The incomplete thread, also called thread runout, is normally


shown with a thin line connecting the minor and major diameter lines at a 15° angle. The layoutof the thread runout may vary, depending on how the external thread is produced (see


TABLE 2 Tolerances for Sockets and SlotsNominal dimensionsover to C13 C14 D9 D10 D11 D12 EF8 E11 E12 Js9 K9

3 +0.20 +0.31 +0.045 +0.060 +0.080 +0.12 +0.024 +0.074 +0.100 ±0.0125 0+0.06 +0.06 +0.020 +0.020 +0.020 +0.02 +0.010 +0.014 +0.014 0.025

3 6 +0.24 +0.37 +0.060 +0.078 +0.115 +0.15 +0.028 +0.095 +0.140 ±0.015 0+0.06 +0.07 +0.030 +0.030 +0.030 +0.03 +0.014 +0.020 +0.020 0.030

6 10 +0.130 +0.19 +0.040 +0.115 +0.175 ±0.018 0+0.040 +0.04 +0.018 +0.025 +0.025 0.036

10 18 +0.20 +0.142 +0.212+0.05 +0.032 +0.032

18 30 +0.275+0.065

30 50 +0.33+0.08

50 80 +0.40+0.10


Figure 6Major dimensions of a metric thread.


Figure 7Metric thread flank angle.


Figure 8Metric thread pitch.

Figure 9External metric thread as shown on technical drawings.

Fig. 9). Looking at the face of the thread, the major diameteris shown in a full circle with a bold line, the minor diameterin a three-quarter circle with a thin line (Fig.9, right).

4.1.2Internal Threads

For the internal thread (Fig. 10), the major diameter is shownwith a thin line and the minor diameter is then shown with


bold lines. The end of the useful thread and the thread runoutare shown the same way as in drawings of the external thread.

Looking at the face of the thread (Fig. 10, right), the minordiameter is now shown in a full circle, whereas the majordiameter is shown in a three-quarter circle with a thin line.

4.2Thread Tolerances

The thread tolerance symbol is composed of a number and aletter, e.g., 6g, 6H. The number indicates the tolerance range,and the letter, the tolerance position. In simpler terms, theletter indicates the starting point of the tolerance (see Fig.11). In U.S. terminology this is referred to as the allowance.

Figure 10Internal metric thread as shown on technical drawings.


Figure 11Metric thread profile with tolerances.

Higher numbers indicate bigger tolerance zones, meaning a greaterspread between the minimum and maximum dimensions.

A capital letter defines the tolerance for internal threads (nut threads),and a lowercase letter defines tolerance for external threads (boltthreads).

The system for limits and fits and the thread tolerances are closelyrelated. The difference is that in the tolerance system for limits and fits,the letter precedes the number, whereas for thread tolerances thenumber comes first.

The common thread tolerances are

6gfor plain external threads6Hfor plain internal threads

The 6H tolerance is applicable even after the parts (nuts) are plated.


The 6g thread tolerance is comparable to Unified thread tolerance 2A.The 6H tolerance is comparable to Unified thread tolerance 2B. A6g/6H combination (Fig. 12) gives about the same amount of playbetween the mating threads as a 2A/2B combination.

Socket head cap screws per ISO standards have a thread tolerance of5g6g for property classes 12.9 and 10.9 and for the hardness class 45H.The 5g tolerance is tighter and applies to the pitch diameter; the 6gtolerance applies to the major diameter.

The ANSI standard specifies a 4g6g thread tolerance that tightens downthe tolerance on the pitch diameter even more.

4.3Thread Tolerances for Special Applications

There are many more thread tolerances. Most oft them have little use.Those used on a fairly regular basis are

6e

Thread tolerance 6e is used for screws requiring thicker platings orcoatings. The 6e tolerance is also applied for screws used forelevated and high temperature services. The added clearance allowsdisassembly of screws that have been exposed to high temperature.

8gAn 8g thread is applied to semifinished fasteners, mainly drop-forgedfasteners.


Figure 12Most common metric thread

tolerance combination.

4h6h.The 4h6h thread tolerance may be specified for threads on aircraftfasteners.

sk6.The tolerance sk6 causes the thread to be oversized. It is used forapplications requiring interference fits. It is commonly applied tothe tap end of double-end studs.

6G.

A 6G tolerance is used for internal threads requiring greaterclearance between the mating threads. Typical applications areweld nuts. The heat may cause distortions of the thread. The 6Htolerance would not offer enough clearance, and the mating boltcould possibly bind up.

7H. The 7H tolerance is applied for nuts of ISO product grade C(coarse finish).

For more details consult ISO 965.

4.4Preferred Metric Thread Sizes


To streamline the use of metric fasteners, the International Organizationfor Standardization (ISO) established preference classes (preferredsizes) for threaded fasteners. Table 3 lists the sizes included in each ofthe three preference classes.

Focus should be placed on preference class 1. These sizes are producedmore often and in bigger quantities and ultimately are more readilyavailable on the world market. Therefore, engineers ''should" specify inthese sizes.

The sizes listed under preference class 2 are not produced as often, andone should expect supply lead times to be longer. Unless these sizes areproduced in larger quantities one has to expect higher prices.

The sizes listed under preference class 3 (thread diameters of 7, 11, or15 mm) are to be avoided. These sizes are used mainly for shaft ends.The size M7 is still used on Japanese cars.

4.5Common Metric Fastener Lengths

Metric fasteners are commercially available in the following lengths:

Up to 6 mm620 mm2050 (eventually 75)mm50160 mm160300 mmAbove 300 mm

in increments of 1mmin increments of 2mmin increments of 5mmin increments of 10mmin increments of 20mmin increments of 40


mm


TABLE 3 Preference Classes forThreaded Fasteners

Preference class1 2 311.2

1.41.6

1.82

2.22.53

3.5456

7810

1112

1415

16


1820

Fasteners of lengths other than those listed arespecial. Engineers should work with the standardlengths as it greatly facilitates procurement ofmetric fasteners.

5Mechanical Properties of Metric ThreadedFasteners

Interpreting mechanical properties of metricfasteners requires the understanding of the metric(SI) units for mass, force, and stress (strength)and other force-related units.

5.1Metric Units for Mechanical Properties

5.1.1Metric Mass (Weight)

The metric unit for mass is the kilogram. Mass is


still very often described as weight. To theengineer weight means the force of gravity actingat a given mass; commercially it means theamount of a commodity (e.g., bread or sugar)one buys in a store. In the technical field oneshould always refer to mass, as this isindependent of the gravitational force of anygiven planet or location on a given planet.

Original definition of the kilogram: A cubewith each side measuring 100 mm (1 decimeter)filled with water at 4°C has a mass of 1 kilogram(kg).


Figure 13A mass of 1 kilogram (1 kg) was formerlydefined as the mass of water that would

occupy a cube of these dimensions at 4°C.

Current definition: 1 kilogram is the mass of theinternational prototype of the kilogram, which islocated in Paris, France.

5.1.2Unit of Force

Fastener load capacities are expressed in newtons


(N). Examples of items expressed in newtonsinclude proof loads, tensile loads, yield loads,permissible loads (load capacities), and shearforces.

Definition: One newton (1 N) is the force that,when applied to 1 kg of mass, will give thekilogram mass an acceleration of 1 meter persecond per second (1 m/s2).

Force on Earth

The earth's gravity gives 1 kg of mass anacceleration of 9.81 meters per second persecond (9.81 m/s2). To balance this out, onewould need 9.81 N of force depending on localgravity. Local gravity depends on latitudes

Gravity at Washington, DC = 9.801 m/s2Gravity at Seattle, WA = 9.809 m/s2

Strength of Metric Bolts


The strength of materials including bolt strengthcan be expressed in newtons per squaremillimeter (N/mm2).

In the United States the unit megapascal (MPa) ispreferred.

Since mega (M) signifies 1000 000 and 1 m2 =1000 000 mm2, we arrive at the followingconclusion:


Figure 14Metric strength in N/mm2 (MPa).

thus demonstrating 1 MPa = 1 N/mm2.


5.1.3Tightening Torques

The metric unit for tightening torque is thenewton-meter (Nm).

Definition: 1 Nm = 1 N of force applied at adistance of 1 m from the axis of the bolt to betightened.

This is illustrated in Fig. 15.

Tightening torque for bolt sizes up to andincluding M5 are generally expressed in newton-centimeters (Ncm) and for those above M5 innewton-meters.

Figure 15


A tightening torque of 1 newton-meter.


TABLE 4 Comparison of U.S. (Inch) Grades and Metric Property ClassesMetricpropertyclass

U.S.(inch)grade

Min. yieldstrength(Mpa)

Min. tensilestrength(Mpa)

Conversionfactor

Min. yieldstrength

(psi)

Min. tensilestrength (psi)

4.8 340 420 145 49,300 60,9005.8 2 420 520 145 60,900 75,4008.8 5<M16 640 800 145 92,800 116,000>M16 660 830 145 95,700 12,3509.8 >5 720 900 145 104,400 130,50010.9 8 940 1040 145 136,300 150,800

12.9 ASTMA 574 1100 1220 145 159,500 176,900

Torque wrench manufacturers sometimes show sizes indecinewton-meters (dNm). This is incorrect and notrecommended.

5.2Metric Property Classes

The mechanical characteristics of metric fasteners areexpressed in property classes, sometimes referred to asstrength classes. There are 10 ISO property classes for boltsand screws and nine for nuts. The ones most frequently usedare 4.8, 8.8, 10.9, and 12.9.


Standard hex cap screws are available in 8.8 and 10.9.

Standard machine screws are available in property class 4.8;some types of machine screws are also available in propertyclass 8.8.

Generally, only socket head cap screws are readily available in12.9. Button head and socket shoulder screws may also comein 12.9. The temperature range for these property classes is -50 to 300°C.

Table 4 lists the minimum yield and tensile strengths offasteners in various metric property classes and equivalentinch system grades.

5.3Property Class Markings

5.3.1Mechanical Property Marking Symbol

The property classes of metric fasteners are indicated bynumbers. The property class symbol consists of two digitsseparated by a decimal point. The digit to the left of thedecimal point indicates 1% of the tensile strength in MPa orN/mm2. The decimal portion of the number indicates the ratioof yield strength to tensile strength.

Calculation of Bolt Strength Based on Head Marking

A typical head marking is illustrated in Fig. 16. This bolt is in


the 10.9 strength class.

Tensile strength: The number before the decimal multipliedby 100 equals tensile strength in megapascals or newtons persquare millimeter. Therefore, for a property class of 10.9,tensile strength equals


Figure 16Property class 10.9 marking

of a metric bolt.

From Table 4, the conversion factor for psi =145.

Yield strength: The decimal portion of the classmarking (.9 in this case) is equal to the ratio ofyield strength to tensile strength:


Therefore, the yield strength of a 10.9 bolt is

5.3.2Proof Stress of Nuts

The proof stress of a nut is displayed as 1% ofactual proof stress expressed in megapascals.(See Fig. 17.)

For example, for a nut with a class marking of 8,the proof stress would be

Nuts must be mated with bolts of equal propertyclass. The property-class of the nut may be higherbut never lower than the bolt property class.


Figure 17Property class 10 marking

of a metric nut.


TABLE 5 Mechanical Properties of Major Property Classes8.8

Mechanical property 3.6 4.6 4.8 5.6 5.8 6.8 d<16 mm

d>16 mm 9.8 10.9 12.9

Tensile strength, Rm(MPa; N/mm2)

nom.min.

300330

400400

400420

500500

500520

600600

800800

800830

900900

10001040

12001220

Vickers hardness, HV, F > 98N

min.max.

95250

120250

130250

155250

160250

190250

250320

255335

290360

320380

385435

Brinell hardness, HB, F = 30 D2 min.max.

90238

114238

124238

147238

152238

181238 238 242 276 304 366

Rockwell hardness, HR min. HRBHRC 52 67 71 79 82 89 22 23 28 32

39

max. HRBHRC 99.5 99.5 99.5 99.5 99.5 99.5 32 34 37 39

44

Surface hardness, HV 0.3 max.

Lower Yield stress, (MPa; N/mm2

nom.min.

180190

240240

320340

300300

400420

480480

Proof stress, Rp02(MPa; N/mm2)

nom.min.

640640

640660

720720

900940

10801100

Stress under proof load Sp (MPa;N/mm2) Sp/RetSp/Rpo20.94

1800.94225

0.91310

0.93280

0.90380

0.92440

0.91580

0.91600

0.90650

0.88830

0.88970

Elongation after fracture, A min. 25 22 14 20 10 8 12 12 10 9 8

Strength under wedge loading The values for full size bolts and screws (not studs) shall not be smaller than theminimum values for tensile strength shown in 5.2

Impact strength, J min. 25 30 30 25 20 15Head soundness no fractureMinimum height of non-decarburized thread zone, E 1/2H1 1/2H1 1/2H1 2/3H1 3/4H1

Maximum depth of completedecarburization, G mm 0.015 0.015 0.015 0.015 0.015

Source: ISO 898/1.


5.4Mechanical Properties of Metric Fasteners

The mechanical properties of the 10 metric property classes for bolts and screws are listed inTable 5.

6Metric Stainless Steel Screws, Studs and Nuts

Stainless steels for metric fasteners are grouped into three material groups: austenitic, ferritic,and martensitic. The ISO system for designating fasteners within these three groups is outlined inFig. 18.

Austenitic stainless steels are used for 98% of metric fasteners. The austenitic stainless steels arecomparable to 300 series U.S. stainless steels.

The ISO austenitic group is subdivided into three alloy groups: A1, A2, and A4, which arefurther subdivided by strength classes.

6.1The A2 Alloy Group

The A2 alloy group typically contains: 1719% chromium and 813% nickel. For a stainless steelto be classified an A2, it must contain chromium and nickel within these above stated limits. Itmay, however, contain other elements also as long as they are within the specified limit of ISO3506 or ASTM F 738M if so specified.

A2 is often referred to as 18/8 stainless steel. It must be noted, however, that 18/8 is commercialterm with no clearly defined chemical composition.


Figure 18ISO designation system for stainless steel fasteners.


TABLE 6 Yield Strength of Group A2 and A4 Alloys in Percent of Yield Strengthat Room TemperatureaA2/A4 100° (212°F)

85200°C (392°F)

80300°C (572°F)

75400°C (753°F)

70aValues given are applicable to A2-70 and A4-70 alloys only.

The A2 alloy group offers good corrosion resistance.However, its members are not suitable for seawaterapplications. Their corrosion resistance to a given acidmust also be checked.

AISI 304/321 stainless steels meet the requirements ofthe A2 alloy group.


The A4 alloy group typically contains 1618.5%chromium, 1014% nickel, and 23% molybdenum. Themolybdenum is included to reduce the risk of pitting.

As with A2, for a stainless steel to be classified A4, itmust contain chromium, nickel, and molybdenum withinthe above-listed limits. Other elements may be added aslong as they are within the specified limits of ISO 3506 orASTM F 738M if so specified.


Stainless steels of the A4 group offer a high degree ofcorrosion resistance and are generally resistant toseawater corrosion.

AISI 316 stainless steel meets A4 requirements.


An A1 steel is basically an A2 with sulfur added to makeit free-machineable. The yield strength drops as thetemperature is raised (see Table 6).

6.4Other Alloy Groups

The ISO standard ISO 3506 is currently being revised andis expected to be published in the late 1990s. It will havetwo more alloy groups added, A3 and A5. A3 is derivedfrom the A2 group and comprises alloys resistant tointergranular corrosion above 400°C. The A2 groupalloys are not.

The A5 group is derived from the A4 group andcomprises alloys that should be used above 400°C wherethere is a risk of intergranular corrosion.

7


Conversion from Inch to Metric Fasteners

7.1Converting to Optimum Metric Fasteners

The safest approach to convert to metric fasteners is toexamine the application and consider all the possibleparameters. Then carry out calculations to determine theproper metric screw diameter (possibly using VDI 22301). Simpler, but not as accurate, is using the screwdiameter estimation chart from VDI 2230.

7.2Estimating Screw Diameter

The parameters that must be considered include

1. Operating force (direction of force; static or dynamic)

2. Material being clamped (surface pressure limit)

3. Operating temperature


TABLE 7Nominal diameter (mm)

Strength classForce(N) 12.9 10.9 8.8

250400630

1,0001,600 3 3 32,500 3 3 44,000 4 4 56,300 4 5 510,000 5 6 816,000 6 8 825,000 8 10 1040,000 10 12 1463,000 12 14 16

100,000 16 16 29160,000 20 20 24250,000 24 27 30400,000 30 36630,000 36


4. Corrosion

5. Amount of friction

6. Inaccuracy of tightening method

The following procedure enables an estimationof screw diameters, taking the most importantparameters into account. An operatingtemperature of 20°C (68°F) is assumed. Once adiameter is selected, one should double check theresults by either calculating or testing the boltedjoint.

Referring to Table 7:

1. Select in column 1 the next higher force to thework force FA acting on the bolted joint.

2. Find the required minimum preload FM min byproceeding from this number.

4 steps for static or dynamic transverse (shear)force


Figure 19


2 steps for dynamic, eccentric axial force

Figure 20

1 step for either dynamic and centric or static andeccentric force

Figure 21


Figure 21

Figure 22

0 step for static, centric axial force.

Figure 23

3. The required maximum preload force FM maxis found by proceeding from this force FM min by

2 steps: for tightening the screw with a power


screwdriver which is set for a certain tighteningtorque


or

1 step: for tightening with a torque wrench/orprecision power screw driver, which is set andchecked by means of dynamic torquemeasurement or elongation measurement of thescrew

or

0 step for ''turn of the nut" method or yield pointcontrolled method.

4. Once the preload (force) has been estimated,the correct screw size is found next to it incolumns 2 to 4 underneath the appropriatestrength class.

Example: A joint is loaded dynamically andeccentrically by the axial force FA = 8500 N.1The screw with strength class of 12.9 will beassembled with a manual torque wrench (see


Table 7).

1. 10,000 N is the next higher force to FA incolumn 1 of Table 7.

2. 2 steps for "eccentric and dynamic axial force"lead to Fmin = 25,000 N.

3. 1 step for "tightening" with manual torquewrench leads to FM max = 40,000 N.

4. For FM max = 40,000 N the proper diameter isM10, found in column 2 of Table 7, under 12.9.

7.3Light Duty Applications Fasteners

For light duty, not strength-critical, applicationsfasteners, one can simply convert to metricfasteners by using the appropriate conversionfactor.

Step 1: Multiply the inch dimensions by 25.4 toget the dimensions in millimeters.


Step 2: Once the major functional dimensions,for instance, thread size and length are converted,round up to the next commercially available size.

Step 3: Choose a material or strength class that iscommercially available.

Note that this procedure yields approximateequivalents and should not be used for strengthcritical applications. For applications where thebolt has previously been calculated and/or fieldtested and found to be adequate the loadcomparison charts of Tables 8 and 9 can be used.

8Proper Tightening Torque for Common PropertyClasses

Table 10 is a torque chart from which preloadsand tightening torques can be read directly for


and tightening torques can be read directly forproperty classes 3.612.9 and various thread sizes.

8.1How to Use the Torque Chart

The tightening torque is determined by followingthe steps below.

Step 1: Determine whether the screws to betightened are plain, zinc-plated, cadmiumplated,or lubricated with molybdenum disulfide (MoS2;e.g., Molylub).

1 The force FA as used in this example is related to 1screw.


TABLE 8 Tensile Load Comparison

Threadsize d

Major diam. d1(mm)

StressareaAs

(mm2)

Threadsize d

StressareaAs

(mm2)

SAEGrade 2 4.6 4.8 5.6 SAE

Grade 5 8.8 9.8 SAEGrade 8 10.9

12.9(alloysteel)

540 3.175 5.14M3 5.03 2.62 2.01 2.11 2.52 4.25 4.02 4.53 5.32 5.23 6.14632 3.505 5.86M3.5 6.78 2.99 2.71 2.85 3.39 4.85 5.42 6.10 6.06 7.05 8.27832 4.166 9.04M4 8.78 4.61 3.51 3.69 4.39 7.48 7.02 7.90 9.35 9.13 10.711024 4.826 11.31M5 14.20 5.77 5.68 5.96 7.10 9.36 11.36 12.78 11.70 14.77 17.321/420 6.350 20.50M6 20.10 10.46 8.04 8.44 10.05 16.97 16.08 18.09 21.21 20.90 24.528/1618 7.938 33.80M8 36.60 17.25 14.64 15.37 18.30 27.97 29.28 32.94 34.97 38.06 44.653/816 9.525 50.00M10 58.00 25.52 23.20 24.36 29.00 41.38 46.40 52.20 51.72 60.32 70.767/1614 11.112 68.60 35.01 56.77 70.971/213 12.700 91.50M12 84.30 46.70 33.72 35.41 42.15 75.72 67.44 75.87 94.65 87.67 102.859/1612 14.288 117.00M14 115.00 59.71 46.00 48.30 57.50 96.83 92.00103.50 121.03 119.60 140.305/811 15.875 146.00M16 157.00 74.51 62.80 65.94 78.50 120.83125.60141.30 151.03 163.28 191.54

M18 192.00 76.80 80.64 96.00 159.36172.80 199.68 234.243/410 19.050 216.00M20 245.00 110.23 98.00102.90122.50 178.76203.35220.50 223.45 254.12 298.907/89 22.225 298.00M22 303.00 23.31121.20127.26151.50 246.62251.49272.70 308.28 315.12 369.6618 25.400 391.00M24 353.00 161.79141.20148.26176.50 323.58292.99317.70 404.48 367.12 430.66

M27 459.00 183.60192.78229.50 380.97413.10 477.36 559.981 1/87 28.575 492.00 203.58 356.27 508.96

M30 561.00 224.40235.62280.50 465.63504.90 583.44 684.421 1/47 31.750 625.00 258.62 452.58 646.551 3/86 34.925 745.00 308.27 539.48 770.69

M36 817.00 326.80343.14408.50 678.11735.30 849.68 996.741 1/26 38.100 907.00 375.31 656.79 938.27

M39 976.00 390.40409.92488.50 810.08878.40 1'015.04 1'190.72As = stress area; Rm = minimum tensile strength.1 kN = 224.8 lb.


TABLE 9 Yield Load Comparison

Threadsize d

Major diam. d1(mm)

StressareaAs

(mm2)

Threadsize d

StressareaAs

(mm2)

SAEGrade 2 4.6 4.8 5.6 SAE

Grade 5 8.8 9.8 SAEGrade 8 10.9

12.9(alloysteel)

540 3.175 5.14M3 5.03 2.02 1.21 1.71 1.51 3.26 3.22 3.62 4.61 4.73 5.53632 3.505 5.86M3.5 6.78 2.30 1.63 2.31 2.03 3.72 4.34 4.88 5.25 6.37 7.46832 4.166 9.04M4 8.78 3.55 2.11 2.99 2.63 5.74 5.62 6.32 8.10 8.25 9.661024 4.826 11.31M5 14.20 4.45 3.41 4.83 4.26 7.18 9.09 10.32 10.14 13.35 15.621/420 6.350 20.50M6 20.10 8.06 4.82 6.863 6.03 13.01 12.86 14.47 18.38 18.89 22.115/1618 7.938 33.80M8 36.60 13.26 8.78 12.44 10.98 21.45 23.42 26.35 30.30 34.40 40.263/816 9.525 50.00M10 58.00 19.66 13.92 19.72 17.40 31.72 37.12 41.76 44.83 54.52 63.807/1614 11.112 68.60 26.97 43.53 61.501/213 12.700 91.50M12 84.30 35.97 20.23 28.66 25.29 58.05 53.95 60.70 82.03 79.24 92.769/1612 14.288 117.00M14 115.00 45.99 27.60 39.10 34.50 74.23 73.60 82.80 109.40108.10 126.505/811 15.875 146.00M16 157.00 57.39 37.68 53.38 47.10 92.63100.48113.40 130.90147.58 172.70

M18 192.00 46.08 65.28 57.60 126.72138.24 180.48 211.203/410 19.050 216.00M20 245.00 84.91 58.80 83.30 73.50 137.05161.70176.40 193.65230.30 269.507/89 22.225 298.00M22 303.00 73.98 72.72 103.02 90.90 189.08199.98218.76 267.17284.82 333.3018 25.400 391.00M24 353.00 97.07 84.72 120.02 105.99 248.08232.98254.16 350.55331.82 388.30

M27 459.00 110.10 156.06 137.70 302.94330.48 431.46 504.901 1/87 28.575 492.00 122.15 274.84 441.10

M30 561.00 134.64 190.74 168.30 370.26403.92 527.34 617.101 1/47 625.00 155.17 349.14 560.34

31.7501 3/86 34.925 745.00 184.96 416.17 667.93

M36 817.00 196.08 277.78 245.10 539.22588.24 767.98 898.701 1/26 38.100 907.00 225.18 506.67 813.17

M39 976.00 234.24 331.84 292.80 644.16702.72 917.44 1'073.60As = stress area; Rp = 1 kN = 224.8 Ib.


Step 2: When the surface condition (finish) is defined,the friction coefficient m can be found from Table 11.

Example: For a dry zinc-plated screw or nut, m =0.125.

Step 3: Obtain the proper tightening (seating) torquefrom the tightening torque chart (Table 10).

Example: Hex cap screw DIN 933, property class 8.8,zinc-plated, M6 × 20 dry.


1. Look for the thread size M6 (left most column)

2. Choose friction coefficient (2nd column), in thiscase 0.125.

3. Move right to the tightening torque columns andlocate property class 8.8.


The maximum torque is 9.9 Nm.

Step 4: Deduct the scatter (tolerance) of thetightening method. (When tightened with acommercial torque wrench, a tolerance of ±1723%can be taken into consideration.)

Note: The scatter must be considered to ensure thatthe applied torque does not induce a preloadexceeding 90% of the yield strength. To allow forthe scatter, calculate the mean torque. The torque tobe set on the torque wrench is the calculated meantorque. With this setting, with maximum scatter,90% of the yield strength would not be exceeded.

Example: Scatter = ±17%.

1. Torque from chart = 9.9 Nm = 117% of meantorque. (For ±23% scatter, max. torque = 123%.)

2. Mean torque if scatter is ±17% = 9.9/1.17 =8.46.

3. Torque to be set on torque wrench = 8.46 Nm.

Step 5: Check preload. To do this follow a


procedure similar to step 3, above, but using thepreload column for the appropriate class.

Example: Preload of class 8.8 with a frictioncoefficient of 0.125 = 9290 N.

Considering total scatter of ±17%, the minimumtorque equals 8.46 (mean torque) × 0.83 = 7.02(min. torque) Nm.

Is the minimum preload adequate (preventingslipping or separating of joint members)? If not, usea bigger screw diameter with the same propertyclass (repeat steps 35) or choose a screw with ahigher property class.

Caution: Higher strength bolts cause higher surfacepressure underneath head/nut.

Step 6: To get the torque in inch-pound units, usethe following factors:


Exampe: 8.22 Nm × 0.7376 = 6.06 ft-lb


TABLE 12 Thread EngagementProperty class 8.8 10.9 12.9Ratio thread diam./pitch (d/p) <9 ³9 <9 ³9 <9Low carbon steel 1.0d 1.25d 1.4d Medium carbon steel 0.9d 1.0d 1.2d 1.5dHeat treated steel (approx. 1045) 0.8d 0.9d 1.0d 1.25dCast (gray) iron 1.2d 1.2d 1.4d aAluminum alloy (»400 MPa) 1.5d 1.5d a aaUnsuitable.

Step 7: Check the surface pressure limit of the materialclamped (Table 13) and make sure that there is sufficientthread engagement (Table 12). Exceeding the surface pressurelimit of the clamped material could result in embedding andconsequent loosening of the bolted joint (see Tables 14 and15).

Notes:

1. Any tightening method involves certain inaccuracies that arethe result of such things as

a. Estimating the friction coefficientTABLE 13 Surface Pressure Limits of Commonly Used Materials (Reference Values)Material of partsbeing clampeda Tensile strength in MPaMax. surface pressure limitc (MPa)

AISI 1018 370 260AISI 1050 500 420AISI 1045 (heat treated) 800 700AISI 4140 (heat treated) 1000 850


AISI 4140 (heat treated) 1000 850AISI 4340 (heat treated) 1200 750AISI 304/316 500700 210Precipitation hardened 12001500 10001250Stainless steels (177 PH)Pure titanium 390540 300Ti-6 AI-4 V 1100 1000Gray iron class 25b 150 600Gray iron class 35b 250 800Gray iron class 50b 350 900Gray iron class 60b 400 1100Aluminum

Die castings 300 (200) 220 (140)Permanent mold casings 200 (300) 140 (220)

Wrought aluminum (hard) 450 370Pure aluminum 160 140aSurface pressure limit are based on the listed tensile strength values.bASTM A48.cBy motorized tightening, the surface pressure limits may be up to 25% lower.


TABLE 14 Surface Pressure Under the Head of Hex Cap Screws and Bearing Area of Hex NutsSurfacepressureunder-

neath head(MPa;

N/mm2)

Threaddiameter

Width across theflats Smax. (mm)

Washer face diameterdw min. (mm)

Clearancehole

(ISO 273)dn (mm)

Bearing surfaceAp (mm2)

StressareaAs

(mm2)

8.8 10.9 12.9

M3 5.5 4.6 3.4 7.54 5.03307 434 520M4 7 5.9 4.5 11.4 8.78352 496 596M5 8 6.9 5.5 13.6 14.2 484 680 799M6 10 8.9 6.6 28 20.1 332 466 659M8 13 11.6 9 42 36.6 406 571 686M10 16 14.6 11 72.3 58 375 528 633M10 17 15.85 11 96.1 58 282 398 477M12 18 16.6 13.5 73.2 84.3 541 760 913M12 19 17.4 13.5 94.6 84.3 419 588 706M14 21 19.6 15.5 113 115 481 676 812M14 22 20.5 15.5 141 115 385 529 650M16 24 22.5 17.5 157 157 476 669 803M18 27 25.3 20 188 192 484 681 816M20 30 28.2 22 244 245 480 672 807M22 32 30 24 254 303 575 807 969M22 34 31.7 24 337 303 433 608 730M24 36 33.6 26 356 353 471 663 798M27 41 38 30 427 459 518 728 876M30 46 42.7 33 576 561 467 656 788


TABLE 15 Surface Pressure Under Head of Socket Head Cap ScrewsSurface

pressure underhead (MPa;

N/mm2)

Threaddiameter

Head diameterdk (mm)

Diameter ofbearing areadw min. (mm)

Clearance holedn (mm)

Bearing surfaceAp (mm2)

StressareaAs

(mm2)

8.8 10.9 12.9

M3 5.5 5.07 3.4 11.1 5.03 209 295 353M4 7 6.53 4.5 17.6 8.78 228 322 386M5 8.5 8.03 5.5 26.9 14.2 245 344 413M6 10 9.38 6.6 34.9 20.1 266 374 449M8 13 12.33 9 55.8 36.6 306 430 516M10 16 15.33 11 89.5 58 303 427 512M12 18 17.23 13.5 90 84.3 440 618 742M14 21 20.17 15.5 131 115 415 583 700M16 24 23.17 17.5 181 157 476 580 696M18 27 25.87 20 211 192 431 607 727M20 30 28.87 22 274 245 427 599 719M22 33 31.81 24 342 303 427 599 719M24 36 34.81 26 421 353 399 561 675M27 40 38.61 30 464 459 476 670 608M30 45 43.61 33 638 561 422 592 712


b. Manipulation errors of torque wrench (operator errors)

c. Torque wrench tolerance

Depending on how much these factors can be measured and/or controlled in the inhouseassembly or field assembly, either a higher or lower scatter must be considered.

2. Adequate thread engagement must always be considered (see Table 12).

8.2Calculating the Tightening Torque

For load-critical applications it may no longer be sufficient to use the torque chart. In suchcases it is advisable to actually calculate the tightening torque.

8.2.1Tightening Torque Calculation for Metric Fasteners

Working with metric fasteners often requires a different approach than the one taken withinch series fasteners. Applying the metric approach allows engineers to use the manycalculation formulas and charts that available.

The following formula has long been used to calculate tightening torque for inch fasteners:

where

T =torque (Nm)K=nut factor

D=nominal diameter (mm)

F =clamp force (preload)(N)

The same formula can be applied to metric bolts with a slight modification:

8.2.2Original Formula for Torque


The following equation addresses friction in the thread as well as the friction in the bearingarea of the bolt and/or nut. The torque can be calculated very accurately.

where

r =friction angle under the headr¢ =thread friction angle

j =thread helix angleDkm

=effective diameter of bearing area (bolt head ornut) (m)*

d2 =pitch diameter (m)*FM

=preload (N)

MA=tightening torque (Nm)

* Note: Pitches, pitch diameters, and effective diameters are normally listed in millimeters; they must be converted tometers for the torque calculation.


8.2.3Simplified Formula for Tightening Torque

For everyday use the formula below may be used:

where

MA=torque (Nm)

P =pitch (m) (See Table 16.)

mk =coefficient of friction in bearing area of bolt or nut (SeeTable 17.)

mG =coefficient of friction in the thread (See Table 18.)d2 =pitch diameter (m)*Dkm

=effective diameter of bolt head or nut bearing area (m) [SeeEq. (5).]

Fm =preload (N) [See calculation for preload, Eq. (8).]

where

dW=head diameter with cylindrical head

=washer head diameter of hexagon screws (SeeTable 19.)

dh =clearance hole diameter (See Table 20.)

8.2.4Calculating the K Factor for Metric Fasteners

TABLE 16 Thread PitchesMetric Pitch Metric Pitch


thread (mm) thread (mm)1 0.25 10 1.51.2 0.25 12 1.751.4 0.3 14 21.6 0.35 16 21.8 0.35 18 2.52 0.4 20 2.52.2 0.45 22 2.52.5 0.45 24 32.6 0.45 27 33 0.5 30 3.53.5 0.6 33 3.54 0.7 36 45 0.8 39 46 1 42 4.57 1 45 4.58 1.25 48 5


The K factor is calculated by dividing the term in brackets by the screw's nominal diameter.

This allows the use of the old favorite formula:

8.2.5Preload and Clamp Load

The preload must be determined before the tightening torque can be defined. Proper


preload is difficult to determine, whether for an inch or metric fastener. The followingcalculations will help you to come up with a close approximation. The necessary minimumpreload (clamp load) can be measured (tested in the actual application) or can be calculatedwith the following formula. At any rate, the preload should be determined in such a waythat there is always a minimum clamp force in the joint to prevent slipping or separating inthe interfaces of the joint members when the operation forces are applied.

Even if the preload is calculated, it must be double-checked in the application to determinewhether it is suitable for the particular industrial application.


TABLE 19 Clearance Hole Diameter dh(Medium)Thread diameter Clearance hole dh (mm)M3 3.4M4 4.5M5 5.5M6 6.6M8 9M10 11M12 13.5M14 15.5M16 17.5M18 20M20 22M22 24M24 26M27 30M30 33Source: Derived from ISO 273.


TABLE 20 Diameter of Bearing Area dwmin of Hex Cap Screws

Threaddiameter

Width across theflats Smax (mm)

Diameter ofbearing areadw min (mm)

M3 5.5 4.6M4 7 5.9M5 8 6.9M6 10 8.9M8 13 11.6M10 16 14.6M10 17 15.85M12 18 16.6M12 19 17.4M14 21 19.6M14 22 20.5M16 24 22.5M18 27 25.3M20 30 28.2M22 32 30M22 34 31.7M24 36 33.6M27 41 38M30 46 42.7

Preload Formula

The maximum preload FM max is calculated as

where

FMmax =maximum preload (N)

a A =tightening factor (see Table 21)FC req.

=minimum required clamp force (N) (Consult Examples 1 and 2, Figs. 24,25)

(1F)FA

=

portion of operation force that relieves contact pressure in the interfacesof the joint members (N) (See Fig. 26)

FA=Operation force (N)FZ=preload loss because of relaxation (N) [See Eq. (12).]


The Tightening Factor aA

The tightening factor aA is the ratio between the maximum possible preload resulting frominaccuracy of the tightening method and the minimum necessary preload to preventslippage and separating in interfaces of joint members (see Table 22).


TABLE 21 Diameter of Bearing Areadw min of Socket Head Cap Screws

Threaddiameter

Headdiameter dk

(mm)

Diameter ofbearing areadw min (mm)

M3 5.5 5.07M4 7 6.53M5 8.5 8.03M6 10 9.38M8 13 12.33M10 16 15.33M12 18 17.23M14 21 20.17M16 24 23.17M18 27 25.87M20 30 28.87M22 33 31.81M24 36 34.81M27 40 38.61M30 45 43.61Source: Derived from ISO 4762.


Figure 24Bolted joint in a hydraulic cylinder.


Figure 25Hole-to-hole diameter Dt.

Minimum Required Clamping Force FC req.

The value of the minimum required clampingforce depends on the application and the designlayout. There are obviously a tremendous variety


layout. There are obviously a tremendous varietyof applications and designs out there. VDI 2230shows a number of different approaches tocalculating the minimum required clampingforce in the interfaces.

It is important to have a minimum clampingforce in the joint at all times to

1. Ensure sealing

2. Maintain a frictional grip in the interfaces ofrotation clutch flanges

3. Prevent separating of clamped plate (jointmembers), etc.

The following examples afford some insight intohow to go about it. These two examples arederived from the VDI 2230 publication. Thepublication itself addresses even more examples,in particular for eccentric applied loads.

Application Examples

Example 1: A hydraulic cylinder is part of a press


exposed to x cycles/hour (see Fig. 24). Theremust be no leakage at any time. Therefore theremust always be a minimum required clampingforce FC req., even when the axial operation forceFA is at its maximum. The sealing force isdictated by the individual application (size ofcylinder, internal pressure, etc.).

Example 2: For the disc clutch (Fig. 25), theremust be no slippage at the interfaces of the jointmembers. (Frictional grip must be ensured at alltimes.)


Figure 26Force ratio nomogram.


where

FQ=shear force acting on one boltss

mT=

friction factor of joint memberinterfaces


Mt max=torque acting on the bolts


i =number of bolts used

Dt=diameter between bolt holes (seeFig. 25)

Since there is no axial operation force FA, FA = 0, the component (1 F)FAdoes not have to be included in the preload calculation.


Figure 27Load introduction factors.

The Force Ratio F

The force ratio F changes are based on joint layouts, joint member stiffness (hardness), andbolt stiffness and is influenced by how the load is applied (centric or eccentric to the boltaxis) and similar factors. It is quite time-consuming tso calculate F.

The force ratio can be taken from Fig. 26, which is based on the VDI 2230 calculation.

Eccentric loading:

where n = As load induction factor one can set one of the listed load factors fromexamples. For load-critical applications, however, you should review VDI 2230. (See Fig.27.)

Operational Force FA

The axial operational force FA is given by the application, the amount of axial operationalforce for a given bolt has to be divided by the number of bolts used.

Preload Loss because of Relaxation

Preload loss FZ is the result of setting (relaxation) of the bolted joint. It can be determinedby using Eq. (12) and Figures 28 and 29.

9Metric Standards and Standards Organizations

To take advantage of all the metric standard fasteners that are available worldwide, one


should acquire standards from the following standards organizations. There are a widearray of types of fasteners available. There is no need to design one's own fasteners, unlessthere is a particular design requirement that cannot be addressed by using standardfasteners. The use of standard fasteners ensures much greater interchangeability of fastenersand offers a great deal of cost savings. To fully utilize all the standard fasteners that areavailable on the world market one may have to choose fasteners that were made to astandard other than an American standard.


Figure 28Joint member stiffness factor nomogram.


Figure 29Joint setting factor.


9.1International Organization for Standardization

The International Organization forStandardization (ISO) consists of roughly 100members.* The intent of the ISO is to promotestandardization to facilitate the exchange ofgoods and services throughout the world. TheISO has published more than 7500 standards.

The standardization work is carried out by 169technical committees and 645 subcommittees.The committees that deal with fastener-relatedmatters are TC 1 and TC 2.

Originally, ISO only published recommendationsconcerning what and how things should bestandardized. The majority of ISO members(national standards associations) incorporatedthese recommendations into their national


standards.

In the 1970s, ISO started to publish worldstandards. However, acceptance of the individualproduct standards has been slow. This has beengradually changing since some major Europeanmanufacturers started to use ISO standards.

ISO standards for material and mechanicalrequirements for fasteners are generally wellaccepted and applied worldwide. One of thegreatest accomplishments of the ISO is theintroduction of an international universal screwthread.

* The short form of the name, ISO, is not an acronym butwas chosen by the organization for its Greek meaning,''equal."


9.2German Institute for Standards

The German standards body is the DeutschesInstitut für Normung (German Institute forStandards). DIN has published roughly 20,000standards on virtually every item one uses everyday.

Although DIN began as a national standardsorganization, DIN standards are now recognizedand accepted internationally. DIN fastenerstandards have long been used in the industryworldwide, including the U.S. Fasteners made toDIN standards can be purchased from almost anycountry that produces metric fasteners andmachine elements.

When the ISO introduces new fasteners standards


When the ISO introduces new fasteners standardsfor items for which DIN standards already exist,then frequently those DIN standards areconverted to ISO standards. Many of the olderDIN standards that do not agree with current ISOstandards are being harmonized to fall in linewith ISO.

DIN standards cover more types of fasteners thanany other standard.

9.3American National Standards Institute

ANSI is the U.S. representative to ISO, theInternational Organization for Standardization.

ANSI metric fastener standards are almost inagreement with ISO fastener standards. ANSI,however, includes some amendments andadditions.

ANSI metric fastener standards are not yet


ANSI metric fastener standards are not yetrecognized worldwide. (Standards documents arerarely available in European and Far Easterncountries.)

Metric fasteners are produced in many differentcountries around the world, so if one specifies afastener per an ANSI standard one should allowfor DIN and ISO standards to be permissiblealternatives.

9.4American Society for Testing and Materials

ASTM metric standards covering the commonmetric property classes and stainless steel arebeing harmonized with ISO standards. There are,however, stipulations in the standards that maymake enough of a difference to impairinterchangeability. Therefore ASTM and ISOstandards should be reviewed for compatibility.


Generally, metric fasteners made to an ASTMstandard are not yet readily available in the worldmarket. Certain ASTM standards require heavyhex bolts and nuts that are not availableworldwide.

Some ASTM specifications are recognized inother countries, e.g., A 193, A 194, A 320, and A325. However, only regular hex cap screws andstuds are ordinarily available and in a limited sizerange.

If general-purpose fasteners are specified toASTM standards, then DIN and ISO standardsshould be permissible alternatives.

9.5European Committee for Standards

As a result of the European community


(European Union) the European Committee forStandards (CEN) was founded. CEN publishesEN standards (European norms).


The CEN technical committee for fastenersreviewed and discussed the available ISOfastener standards. The general consensus wasthat whenever there are existing ISO standards,they should become the basis for the ENstandards. The result is that the ISO mechanicalproperty standards for screws, studs, and nutshave been accepted and will become ENstandards. CEN has also accepted ISO fastenerproduct standards. After the introduction of ENstandards, European national standards are to bephased out. The phasing out of nationalstandards, however, may take a number of years.

9.6Availability of Metric Documents andStandards

ANSI and ASTM standards are available fromtheir respective organizations. Addresses aregiven in the Appendix to Chapter 9.


DIN standards and VDI 2230 are available fromBeuth Verlag, Berlin, Germany.

EN and ISO standards are available from ANSI.They may also be obtained directly from theInternational Organization for Standardization(ISO), Geneva, Switzerland.

There are also a number of documentdistributors in the United States, where one canobtain national and international standards anddocuments.

The DIN handbooks covering fastener standardsare available from Bossard International Inc.,Portsmouth, NH.

10Ground Rules for Working with Metric

Following are a number of "rules" that oneshould apply.

1. Think in global terms. The metric system is an


international system that addresses the needs ofthe international market.

2. Set up a metric library. This will facilitatedesigning and working with metrics.

3. Design in simple, readily available fastenersand components. This will give you access toless costly quality fasteners.

4. Become a member of ANSI. You will getregular updates on new ANSI, ISO, and ENstandards.

5. Obtain and use the appropriate documents foryour projects. They will help with the properdesigning of metric fasteners.

6. Do not intermingle standards. Most metricstandards are coherent; that is, they are closelyinterrelated and related to basic standards.Specifying anything other than what the standarditself refers to turns a standard fastener into aspecial fastener, leading to higher costs with


little or no benefits.

7. Do not soft convert interchangeable parts.Doing so introduces errors, leading to the use offasteners that are not interchangeable.

8. Get metric instruments. This will makeworking with metric parts much easier.Measuring in inches and converting to metricdimensions is time-consuming and often leads toerrors.

9. Do not specify material, unless the applicationleaves no other alternative. Instead specify aproperty class (e.g., 10.9) or the desired physicaland mechanical properties (e.g., yield strength of640 N/mm2). Metric fasteners are produced allover the world. The manufacturer may not havethe requested material on hand but may have analternative material with the necessary physicaland mechanical properties.

References


Bossard International Inc.: "Metric is simple."

Bossard International Inc.: "Securely FastenedJoints."


Bossard International Inc.: Excerpts of technicalsection of fasteners catalog.

Bossard AG: Bossard Handbuch derVerschraubungstechnik (Bossard-manual offastening technology).

International Organization for Standardization(ISO): Data on standards.

Verein Deutscher Ingenleure: VDI 2230(systematic calculation of high duty boltedjoints).


13Washers:A Guide to Their Function and Selection

Marc LevinsonITW Shakeproof, Elgin, Illinois

1Introduction

Washers play an essential role in generating andsustaining integrity in bolted assemblies. Theyare active contributing components from themoment torque is applied in the assemblyprocess.

When the primary purpose of the fastener is toresist shear forces, a washer may be added to theassembly to


1. Prevent damage to the bearing surface adjacentto the fastener or to prevent possible embedmentof the fastener by distributing the clamp load

2. Function as a locking, or backoff-resistant,device

When fasteners are used to provide a clampingforce to ensure the integrity of an assembly, awasher may be added to

1. Prevent damage to the surface adjacent to thefastener

2. Prevent head embedment

3. Span oversize holes

4. Generate and/or maintain tension in thefastening system

5. Function as a locking, or backoff-resistant,device

Still other washers are used to compensate fortolerance variations and eliminate end play or


rattles in assemblies. In many assemblyapplications, a combination of two or more ofthese functions may be required in a washer.

2The Loosening Phenomenon

One of the important contributions that a washermakes to the integrity of an assembly is ininhibiting the loosening phenomenon that leadsto assembly failure.

The integrity of an assembly can be negativelyaffected in many ways. The fundamental natureof the threaded assembly, which involves themating of inclined planes under load (the


Figure 1Transverse sliding of threaded assembly.

thread lead and flank angle of the screw or boltand the angle of the mating thread in the nutplate), creates an inherent tendency for themating pieces to slide "downhill" until there isno longer any load being applied to the inclinedsurface. (See Fig. 1.)

This transverse sliding/loosening tendency isaccelerated when joined materials are soft orductile and yield under load rather thangenerating tension through resistance.Thermocycling, with its resulting expansion and


contraction of materials, also dissipates neededassembly tension.

Only by generating and retaining needed tensionin the assembly can equilibrium be sustained inthe inclined surfaces of the mating threads toprevent transverse sliding and loosening. In theideal assembly environment, a bolt will stretchenough, when engaged in a nut plate, to generatethe tension needed to offset the effects of headembedment or any yielding of the clampedmaterial and by doing so will sustain a secureassembly.

A steel bolt elongates approximately 0.002 in.per inch of effective length when it is stressed to60,000 psi, assuming that the yield strength ofthe material has not been exceeded. A long bolt,properly torqued, can therefore serve as a veryeffective tension-generating spring.

If a bolt can be long enough and if sufficient loadcan be applied, the only reason for adding a


washer to the fastening system would be toprotect the bearing surface adjacent to the bolt ornut by distributing the load over a wider area.Unfortunately, in the majority of commercialapplications, space and weight considerationslimit acceptable bolt diameter and length, and,without a supplementary washer, loosening of anassembly can occur. For example, the head of aGrade 5 fastener with an effective length of 1/2in. will have to embed only 0.001 in. due toburrs, paint, or dirt on the faying surfaces of thejoint to lose 100% of a 60,000 psi tensile stresspreload.

Assemblies are often subjected to cyclic loads(shock and vibration), and that too can producemomentary release of applied tension and permitslippage on the inclined plane.

3Washer Function and Selection

Needed tension in an assembly can be most


efficiently generated by adding a washercomponent. A wide variety of washer types havebeen developed to offset specific looseninginfluences. In selecting a washer appropriate toan assembly application, the purpose andfunction of the many washer options availablemust be understood.

3.1Washer Option No. 1:Distribute Clamp Load to Prevent Screw andBolt Head Embedment and Surface Damageand Span Oversized Holes

A flat washer is used to protect faying surfacesby distributing clamp load over the largesteffective bearing area. It is placed under thefastener component that is rotated duringassembly.


TABLE 1 Nominal Dimensional Comparison of Type B Flat Washer StandardsNarrow Regular Wide

Screw diameter OD Thickness OD Thickness OD Thickness#6 0.312 0.032 0.438 0.040 0.562 0.040#8 0.375 0.040 0.500 0.040 0.625 0.063#10 0.406 0.040 0.562 0.040 0.734 0.063#12 0.438 0.040 0.625 0.063 0.875 0.0631/4 0.500 0.063 0.734 0.063 1.000 0.0635/16 0.625 0.063 0.875 0.063 1.125 0.0633/8 0.734 0.063 1.000 0.063 1.250 0.1007/16 0.875 0.063 1.125 0.063 1.469 0.1001/2 1.000 0.063 1.250 0.100 1.750 0.100

The flat washer is also used to span oversized holes. Its onlyclaims to contributing to needed tension in an assembly are that itprevents embedment of the bolt head or nut and it adds toeffective bolt length by its thickness.

Flat washers are available in a wide variety of shapes, sizes, andmaterials. Let's examine the three basic styles of the widely usedType B flat washer as described in Industrial Fastener Institute(IFI) standards; they are listed in Table 1.

It might appear that if space limitations were not a factor, the flatwasher with the largest outside diameter should always beselected for maximum distribution of clamp load over the largesteffective bearing area. This is not always the case.

Load transmission is effective only through a washer for adiameter equal to the bearing surface of the bolt head plus twice


the thickness of the washer. This is illustrated in Figure 2.

Figure 2Flat washereffective load transmission.


TABLE 2 Optimum Washer DiametersHex heads

Type B washer thickness Effective bearing diameterScrew diam. Hex across flats Narrow Regular Wide Narrow Regular Wide Matching washer#8 0.250 0.040 0.040 0.063 0.330 0.330 0.380 Narrow#12 0.312 0.040 0.063 0.063 0.390 0.440 0.440 Narrow3/8 0.562 0.063 0.063 0.100 0.690 0.690 0.760 Narrow

Hex washer headsType B washer thickness Effective bearing diameter

Screw diam. Hex washer flats Narrow Regular Wide Narrow Regular Wide Matching washer#8 0.348 0.040 0.040 0.063 0.430 0.430 0.470 Regular#12 0.432 0.040 0.063 0.063 0.510 0.560 0.580 Regular3/8 0.780 0.063 0.063 0.100 0.910 0.910 0.980 Regular


In Table 2 charts we have applied the principleillustrated in Figure 2 to select the optimum sizeflat and bearing washers for typical hex head andhex washer head bolts or screws.

It is evident from the dimensional comparisonsthat the narrow series of flat washers is actuallyoptimum for the hex heads and the regular seriesis optimum for the hex washer heads.

The question that logically follows is: Why is thewide series needed at all? If the bearing surface isa particularly soft or thin material that will tendto yield under load, the bearing stresses will bedistributed toward the outside of the washer asthe material under the head of the screw yields. Insuch a case, the wide diameter washers willcontribute by reducing the bearing stresses in thesoft or thin materials. In other cases,manufacturing considerations may override thepurely engineering considerations and result in


the selection of a larger diameter washer than thetables suggest.

A Note of Caution: Flat washers need to be trulyflat to provide effective load distribution. IFIdefines acceptable flatness for flat washers asbeing 0.005 in. in washers with diameters up to0.875 in. and 0.010 in. for larger diameters. Thislevel of flatness cannot be obtained inconventional pierce-pilot-pierce progressivestamping dies due to a tendency to crown thewasher in the final piercing station.Manufacturers of flat washers should use asecondary flattening station or compound toolingto achieve acceptable flatness.

3.2Washer Option No. 2:Generate and/or Sustain Tension in theFastening System

The conical washer, with all of its variations, isone of the more widely used types of tensioning


washers. Conical washers, while absorbingdriving torques and sustaining needed tension,also span alignment holes and effectivelydistribute bearing loads over soft or thinmaterials.

3.2.1Belleville Washers

The Belleville conical washer shows an almostlinear rate of load vs. compression from itscrown height to the flat condition. (See alsochapter 15.) The yield strength will not beexceeded even when the washer is compressed toflat, and the washer will exhibit 100% springreturn when the compression force is removed.This optimum performance is achieved bymaintaining very specific relationships betweenthe washer's inside diameter, outside diameter,thickness, and crown height. The ratio ofmaterial thickness to rim width, for example, isheld to about 1:5. Crown height should notexceed 40% of material thickness.


To obtain required load deflectioncharacteristics, Belleville washers are oftenstacked as shown in Figure 3.

When stacked in parallel, the force required toflatten the washers is multiplied by the numberof washers in parallel. If one washer flattens at100 lb, three similar washers stacked in parallelwill flatten at 300 lb.

When Belleville washers are stacked in series, itis the deflection that is multiplied by the numberof washers. For example, if the washer thatrequires 100 lb to flatten has a crown height of0.025 in., then four washers stacked in serieswill have a total deflection of 0.100 in. under a100 lb load. By alternating parallel and seriesstacking, a wide range of load and deflectionneeds can be met.

3.2.2Conical Washers


Although Belleville washers are produced tospecifications designed to optimize load-deflection performance, there is a wide range ofconical washers produced to less stringentspecifications. Most of the conical washers usedin preassembled screw and washer assembliescalled "Sems," for example, have crown heightsfar greater than 40% of the washer thickness, yetthese washers are intended to be flattened whenused. In the case of conical washers for Semsfasteners, the increased crown height is fororientation during washer-screw assembly.


Figure 3Alternative stacking arrangements for Belleville washers.

When a conventional conical washer is flattened the firsttime, the steel is stressed beyond the yield point.Consequently, the washer is permanently deformed andtakes a set. (See Fig. 4.)

As the washer is initially compressed, it exhibits a linearload deflection characteristic, but eventually the stresses inthe washer become high enough for the material topermanently yield. When this occurs, the load begins to drop


and the crown height reduces to a height that is compatiblewith the induced stress, which the material is then capableof absorbing without further yielding.

It is probably more correct to say that the washer has beenre-formed rather than deformed, for if it is unloaded andthen reloaded it will act in a consistent elastic manner overits reduced crown height.

Conical washers are frequently used in applications wherethey function entirely within the elastic range and are neverfully compressed to a flat condition. It is quite common tosee a conical washer that would yield if completely flattenedused in an application where the deflection and inducedstress never approach the yield point.

By varying thickness (t) and crown height (h) relationships,conical washers can be designed to meet a wide range ofloaddeflection requirements.

Reflecting the Belleville principle, when crown height (h)equals 40% or less of thickness (0.4t), an elastic conditionis maintained. As h increases in relation to t, the washerexceeds the yield point of the material when flattened. Whenh is in the range 0.4t0.8t, a fairly constant spring rate ismaintained. An h/t ratio of up to 1.4 will sustain a positiverate of increase of the load up to 100% deflection. An h/tratio of 1.5 sustains a constant load over a fairly large


deflection, making washers of this type useful inapplications where they must absorb an


Figure 4Stressing a conical washer beyond the yield point.

extreme wear condition. When the h/t ratioexceeds 1.5, yielding will occur; in extremeconditions, oil canning or even inverting canoccur.

3.2.3


Square Cone Washer*

A significant improvement on the basic conicalwasher, the high performance Square Conewasher, was developed and introduced by ITWShakeproof a number of years ago. This uniquewasher has two distinct conical sections andconsequently two separate load deflection rates.(See Fig. 5.)

The outer cone, or first stage, of the Square Conewasher flattens first and has a relatively lowspring rate. Once the first stage has flattened,additional loading begins to flatten the innercone, which, due to its geometry, has asignificantly higher spring rate.

The Square Cone washer is capable of acceptinga high tension load while maintaining a highspring action and generating a constant load,even with variations in installation torques. Itprovides a higher retained clamp load andimproved compensation for thermal cycling and


vibration.

The relationship between torque and tension in afastened joint is very difficult to predict.Theoretical methods exist for calculatingtorquetension relationships, but in the real worldvariations as high as 30% can occur in a typicaljoint design. In spite of this, in most applicationstorque is used as a method of determiningwhether adequate clamp load has been generatedin a joint. As can be seen in Figure 6, the SquareCone washer provides for more uniformity in thetorque/tension ratio. With the Square Conewasher, torque variation is just over 10% andtension variation 14%. Torque variation on theconventional conical washer is 22%, and thevariation in tension is 33%.

* ''Square Cone" is a registered trademark of Illinois ToolWorks, Inc.


Figure 5The Square Cone washer has two

load deflection rates.

The reason for the more uniform relationship between torque and tension in a joint thatincorporates the Square Cone washer is demonstrated in Figure 7, which compares thetorquetime relationship during the driving of a screw with a 7/8 in. outside diameter SquareCone washer and a screw with a regular conical washer having the same 7/8 in. outside diameter.

In achieving the same clamp load, the standard conical washer required about 30% more torque,and the torque was developed over a much shorter time, which can transmit a shock

Figure 6Torque distribution for the Square Cone washer compared to that of a standard conical washer.


Figure 7Torquetime relationship for the Square Cone and standard conical washers.

load to the operator. Shock loading has been shown to cause carpal tunnel syndrome, a physicalproblem resulting in serious operator discomfort, lost time, and high medical costs.

The Square Cone washer produces a more uniform torquetension relationship because so littlechange in tension occurs over a wide deflection range. This is apparent in Figure 8, whichdepicts the typical Square Cone tensiondeflection curve. This characteristic provides a large"window" for accommodating screw gun torques, resulting in more predictable clamping forceand reducing shock loading on the drive tool.

3.3Washer Option No. 3:Resist Loosening Effects of Shock and Vibration

Certain washers are designed specifically to resist loosening or backoff rotation whileenhancing and sustaining tension in an assembly. Basically, these are the helical spring andtooth lock washers in all of their variations.

3.3.1


Helical Spring Lock Washers

In 1887 the helical spring lock washer was the first to be introduced as a loosening-resistant orlocking washer. (See Fig. 9.) It has proven its claim through the years. This device is essentiallya single coil spring, manufactured from trapezoidal wire, with the maximum thickness being onthe inside diameter. A single coil spring is not capable of providing very much reactive tensionin a joint, even when flattened.

Figure 9 shows that the maximum tension achieved from a 7/16 in. washer is about 450 Ib,which compares to a proof load of 9050 Ib and a proof load stress of 85,000 psi for a


Figure 8Tension-deflection relationship for the Square

Cone washer.

Grade 5 bolt of this size. This would seem toindicate that the helical spring washer does notcontribute much to joint integrity. Yet it hasperformed the job expected of it for over 100years. Why?

Tests conducted at Lawrence TechnologicalUniversity show that these washers have a


definite spring rate that develop after the singlecoil is closed and the washer is visually flat.Further loading, it has been shown, results inadditional deformation of the washer. Thetrapezoidal section apparently twists so that theface of the washer lies flat. The maximumoutside diameter of the washer is actuallyincreased. (See Fig. 10.)

On a 7/16 in. diameter helical spring lockwasher, this deflection provides a spring rate of6240 lb/0.001 in. at 60% of the proof load ofthe fastener. This is the equivalent of adding 0.72in. to the effective length of the bolt plus thethickness of the washer for a total increase ineffective bolt length of 0.85 in. As previouslyexplained, the longer the effective length of thebolt, the less likely it is that embedment of thefastener will result in dropping the prestress levelbelow the induced stress. Therefore, it is lesslikely that the frictional balance between thescrew and the nut plate will lose its equilibrium.


The section width and thickness of helical springlock washers can be modified to meet applicationneeds. Standard sections include regular, heavy,extra duty, and light. They are also available inspecial types, such as the Hi-Collar washer foruse with socket head cap screws

Figure 9Helical spring lock washer.


Figure 10Helical spring lock washer under load.

and in other limited space applications and the DoubleCoil, which provides the much greater reactive rangeneeded in power line applications and in other wood-to-metal assemblies used in construction.

3.3.2Spak Washers

The Spak washer is the newest variation of the split lockwasher. (See Fig. 11.) This unique washer adds thereactive tension and torque absorption qualities of thewave or spring washer to the sustained locking tension ofthe split lock washer. It is produced as a split lock washerand is then stamped to form a wave. It has the ability to


absorb shock and vibration and sustain tension. Inventedand patented in Germany and introduced in Japan in the1960s, Spak washers, traditionally preassembled toscrews in Sems form, are the fasteners of choice in Asiafor electronic and automotive assemblies.

3.3.3Twisted Tooth Lock Washers

The twisted tooth lock washer functions by biting intothe head of the fastener and into the bearing surface ofthe component being clamped. The twisted teeth areoriented so that as the screw is tightened, the teeth slideon the rotating surface and are flattened, therebyproviding a flexing feature that adds tension to the joint.

Any tendency for the screw to loosen causes the teeth totry to return to their original form, resulting in the sharpcorners biting into the mating surface. The teeth thenfunction as struts, resisting further backoff rotation orloosening. (See Fig. 12.)

Twisted tooth lock washers, which are hardened into therange (per ASME) of 4050 HRC, work best when theyare about 10 HRC points harder than the other materialsin the joint.


Two basic tooth forms, the Type A and the Type B, areavailable in the United States. (See Fig. 13.) Both ofthese tooth forms work well, with the Type A having anadvantage in light gage metal applications and the Type Bhaving an advantage in harder materials or casting. TheType B form also functions well when an electricalground connection is required. The

Figure 11Spak "split-wave" washer.


Figure 12Tooth lock washerresistant strut action.

Type A tooth form has a good spring return, whilethe Type B, with its rigid tooth form, has almostno spring action.

Even with the Type A tooth form, spring action islimited compared to a conical washer. Generally,tooth washers are smaller in diameter than conical


washers and may be preferable in applicationswhere space is limited or the joint is visible.

The basic external and internal tooth washers arethe widely used standards in the majority of OEMapplications. (See Fig. 14.) The external toothwasher provides maximum resistance to backoff,as the tooth bite is positioned under the outer edgeof the standard screw head.

Exposed applications are sometimes not wellsuited to the external tooth washer. It may bedesirable to avoid visual marring around thecircumference of the screw head, or there may be adanger of snagging something on the exposedteeth. The internal tooth washer was developed toaccommodate these applications. The internaltooth washer also works well in closefittingcounterbored holes where the fastener head isrecessed.


Figure 13Type A (left) and type B (right) tooth forms.


Figure 14External tooth lock washer (left) and internal tooth lock washer (right).

With the external/internal tooth lock washer,twisted teeth are formed on both the insidediameter and the outside diameter of the washer.(See Fig. 15.) This type of washer is particularlyeffective between two components to preventshifting and retain relative position. It is ideal forelectrical grounding and is also used where a largebearing surface is required for oversized orelongated holes.


All of these tooth washers are available in varyingrim widths and thicknesses to meet applicationrequirements.

The design versatility inherent in the tooth washerprinciple has also made it possible toaccommodate flush-fitting screws in countersunkholes. Today, the more popular sizes ofcountersunk tooth washers are used with 82°countersunk screw heads. (See Fig. 16.)

3.3.4Combination Locking/Tension Washers

With the inherent adaptability of the tooth form, itwas inevitable that it would be combined withvariations of the conical washer to provideoptimum locking and tension in a single washerthat could also span oversized holes and distributeload. Two types, the dished tooth lock washer andthe domed tooth lock washer, were introducedmany years ago.

The dished type tooth washer is resilient and is


excellent for distributing a load over a soft or thinmetal and for spanning oversized or elongatedholes. The tooth periphery version (Fig. 17) isdesigned to bite into an upper sheet to preventshifting when an oversized hole is

Figure 15External/internal tooth lock washer.


Figure 16Countersunk tooth lock washer.

Figure 17Dished tooth periphery lock washer.


Figure 18Domed plain periphery lock washer.


Figure 19Pyramidal lock washer.

needed for adjustment purposes. A flat rimversion is available for bearing against hardplastic surfaces or to avoid marring of the uppersheet.

The dome type tooth lock washer (Fig. 18) isdesigned to function in the same type ofapplication as the dished type of washer, but itprovides added rigidity to support greater loadswithout oil canning or inverting even whenspanning large holes.


The pyramidal type washer (Fig. 19) is designedfor applications that require very high tighteningtorques. It holds up under heavy loads whileproviding resiliency and optimum holdingpower. The washer is produced from heavy gagematerials, and the crown is strengthened with sixradial creases embossed into the material that runfrom the inside diameter to the corners of thehexagonal outer diameter. The sharp points at theoutside corners of the washer prevent shifting,and the rigid crown forces the teeth on the innerdiameter to bite deeply into the bolt or nut formaximum holding power.

3.4Washer Option No. 4:Compensate for Tolerance Variations

3.4.1Spring and Wave Washers

Spring or wave washers are specifically designedto provide a compensating force and sustain a


load or absorb shock. (See Fig. 20.) Many designvariations have evolved to best serve one or theother of these functions or to optimize bothfunctions in a single part within specific insideand outside diameter limits.

Variations in washer thickness, number ofwaves, and wave height directly affectloadbearing and deflection capabilities. Single-wave washers are generally best in applicationsinvolving low loads and requiring highdeflection. The loaddeflection curve for this typeof washer is virtually linear. Typical applicationswould include the elimination of end play inelectric motors, the take-up of large tolerancevariations, the elimination of rattles, and thecushioning of light loads.


Figure 20Multiple-wave spring washer.


When considering multiple-wave washers (traditionally threeor four waves) it should be remembered that the greater thenumber of waves the greater the load-bearing capability andthe less the deflective capability. Multiple-wave washers havewide application as take-up springs.

In specifying spring washers, in addition to the amount ofload that may be applied to the washer, the type of load (staticor dynamic) must also be considered. In a static loadenvironment, the basic function of the spring washer is toretain load. In this type of environment, the elastic limit of thematerial may actually be exceeded.

In a dynamic load environment, the washer performs as aregularly functioning spring, and the elastic limits of thematerial must therefore not be exceeded. Loading the washerbeyond its yield strength results in permanent distortion of thewave height, and consequently its responsiveness to dynamicloads is reduced.

The type and magnitude of the load to which a spring washerwill be subjected and the reactive range required are theprimary factors in determining the type (single-wave,threewave, four-wave, etc.) and physical characteristics(material thickness, wave height, etc.) that should be specifiedfor the washer.


Formulas for determining load and deflection characteristicsfor single-wave and multiple-wave washers are given inSection 3.4.2.

Although their effectiveness is often overlooked, the properspecification and use of spanning, locking, tensioning, andcompensating washers can make the difference between areliable, secure assembly and an assembly that causesproblems in field service.

3.4.2Load and Deflection Formulas

Single-Wave Washers

Load equation:

Deflection equation:

where

P =load (lbf)

E =modulus of elasticity of material (30,000,000 psi forsteel)

t =material thickness (in.)f =deflection (in.)


d =inside diameter (in.)D=outside diameter (in.)

S =maximum allowable stress (200,000 psi for steel)

Multiple-Wave Washers

Load equation:

Deflection equation:


where

N=number of waves

P =load (lbf)

E =modulus of elasticity of material (30,000,000 psi forsteel)

t =material thickness (in.)f =deflection (in.)d =inside diameter (in.)D=outside diameter (in.)

S =maximum allowable stress (200,000 psi for steel)

Note: These equations are presented only as a guide.Hardness, for example, will influence the reaction of aspring washer to load but is not factored into theseequations. They are presented to guide a specifier togeneral design parameters, and calculated values may varyby as much as 30% from actual values in some cases.


14Belleville Springs

George DavetSolon Manufacturing Company, Chardon, Ohio

1What Is a Belleville Spring?

A Belleville spring is a conical disc that willdeflect (flatten) at a given spring rate whensubjected to an axial load Fp (Fig. 1). This rate isusually very high relative to that of a coil spring,which makes a Belleville an excellent candidatewhere large loads must be delivered through ashort movement. Some applications whereBellevilles are commonly found are clutch andbrake mechanisms in heavy equipment, punchand die sets, bearing assemblies, switchgear, and


anywhere bolt preload must be maintained overtime. This chapter focuses on Bellevilles used inbolting applications.

A Belleville spring's geometry can becharacterized by four dimensions:

OD=outside diameter

ID =inside diameter

t =materialthickness

h =deflection to flat

Some manufacturers use the parameter H (freeheight » t + h) in lieu of h. In a bolted joint thespring is normally loaded at the upper inside edgeby the nut or bolt head and the lower outsideedge by the joint.

Manufacturers also use a parameter called ''flatload" or "load to flat" (Fw) in their description ofa Belleville spring. This is the force required to


"push" the spring into the flat position. As thenut in a bolted joint is torqued, the spring startsto deflect (flatten) and the bolt begins to stretch.The Belleville will be in its flat position whenthe preload in the bolt equals the flat load. Sincethe spring will be clamped between the flat jointand the nut, any additional torque applied to thenut will only stretch the bolt. Loads higher thanFw will normally not damage the spring if theload cannot deflect it past the flat position (as inmost bolting applications). Therefore, themaximum force that can be applied to aBelleville mainly depends on the limitations ofthe bolt or joint designs.

The spring rate of a Belleville depends ongeometry, material, and loading conditions. Aplot of load versus deflection will show that theBelleville's characteristic curve is linear up


Figure 1A section of a Belleville spring showing load Fp applied at the

"upper inside" and "lower outside" corners.

to about 90% of the flat load.* After 90% the


load increases progressively because the springbegins to "bottom out." Figure 2 shows a typicalloaddeflection curve of a single Belleville. Notethat for a given deflection the decreasing loadsare lower than those increasing. This hysteresis iscaused by the friction between the spring and theloading surfaces.

The loaddeflection relationship can be altered byusing multiple springs in series or parallel or acombination of the two (Fig. 3). Bellevilles aresaid to be in series when the concave and convexsurfaces face in alternate directions [1]. In otherwords, they are stacked "cup to cup" and "crownto crown." Springs stacked in parallel face thesame direction. The convex side of one spring is"nested'' into the concave surface of the next.Bellevilles may also be arranged in "series-parallel" by alternating sets of springs stacked inparallel.

Stacking two springs in parallel doubles the loadrequired to flatten the springs with no increase in


deflection. Two Bellevilles stacked in series willproduce twice the deflection for the same load.Figure 4 shows the loaddeflection curves forvarious combinations of springs. The frictionalhysteresis increases as more springs are used inparallel because the larger contact areas result inhigher frictional forces. These forces aredecreased as more washers are used in seriesbecause some of the contact surfaces will "roll"rather than "slide."

* This applies to most Belleville springs that are used inbolting applications; however, springs can be designed tohave a nonlinear characteristic curve.


Figure 2Loaddeflection curve for a single Belleville spring. The upper curve represents an increasing load; the lower curve shows unloading.

Figure 3Flat load or deflection or both may be altered by stacking Belleville springs in

various arrangements.


Figure 4Load deflection curves of several Belleville spring arrangements. Adding springs in series increases deflection without increasing

load, while using springs in parallel multiplies load without changing deflection.

2Belleville Springs in Theory

2.1A Joint with No Belleville Springs

One of the keys to effective bolting is to understand the various loads and deflections present ina joint. If this is accomplished, then the engineer stands a better chance of getting the variousdeflections to work in his favor. Modeling the joint as a system of springs makes predictingchanges in preload relatively easy. A joint diagram is simply an illustration of the various springrates present in a bolted joint.* This tool is designed to allow the engineer to predict the changein preload through external loads or relaxations [2].

In order to construct a joint diagram, several assumptions will be made. Assume that a bolt hasa spring rate that is 40% the rate of the joint. In other words, if the bolt stretch at a givenpreload is 0.005 in., then the joint compression at the same load would be 0.002 in. A plot ofload versus deflection for a bolt is shown at the upper left portion of Figure 5. If dL is theamount of stretch in the bolt at an aim preload Fp, then the spring rate of the bolt KB can becomputed:


Furthermore, if the amount that the joint compresses at Fp is dT, then the spring rate of the jointKJ is

When the two plots are combined, a joint diagram is formed. At this point, the bolt load is equalto the clamping load.

* For more on joint diagrams please refer to Section IX of this Handbook.


Figure 5A joint diagram where Fp = aim preload; dL = bolt stretch = 0.005 in.; dT = joint deflection = 0.002 in.

Here the joint is very stiff compared to the bolt.

Now consider an external load applied to the system. For instance, assumethat a tensile load LX is applied to the bolt where the face of the nutcontacts the joint (Fig. 6). This interface is the "loading plane." As the boltis pulled, the load on the bolt, FB, is increased while the clamp load on thejoint, FJ, is decreased. When the system was first assembled, these two


loads were equal to the preload Fp. As the external load is applied to thebolt, the joint starts to return to its unloaded thickness along the elasticcurve. At the same time, the bolt is elongated by the combined force of thejoint "pushing" out on the nut and the external load. Note that since the nutstays in contact with the joint while the external force is applied, thechanges in deflection in the bolt and the joint are equal. In other words, thebolt gets longer by the same amount the joint becomes thicker.

Figure 7 is the joint diagram showing the effects of an external load Lx.Since the bolt is more elastic than the joint, any increase in load on the boltdFb, results in a larger decrease in clamp load on the joint dFj. Another wayto look at this is in terms of how deflection of the bolt affects clamp load.As the bolt is stretched from dL to dL', a significant amount of clamp loadis lost. In Figure 7 the bolt is stretched an extra 0.001 in. The result in thisexample would be a 50% decrease in clamp load Fj.

2.2Using Belleville Springs to Maintain Bolt Preload

The spring rate KW of a single spring is about one-third to one-seventh ofthe bolt's rate (Kb). This is because a Belleville will deflect three to seventimes as much as a bolt will stretch at a given preload.* This is true untilthe preload begins to exceed the flat load of the spring (or

* These figures are based on preloads that are below 60% of yield and length-to-diameterratios lower than 12.


Figure 6A section of a joint showing the application of

external load Lx at the loading plane (where thenut and bolt head contact the joint surfaces).


spring assembly). Remember that after theBelleville has been flattened, any additional loadwill only stretch the bolt and compress the joint.For this reason, springs should be selected withflat loads that are close to the preloads to be used(for bolting applications). If the flat load is muchgreater than the preload, then the Belleville willbe only partially compressed. Since only afraction of the spring's potential deflection isused, the engineer is not maximizing its benefits.If the preload is greater than the flat load, thenthe Belleville will remain in its flat position untilthe residual preload falls below the flat load.

The joint diagram discussed in Section 2.1 looksquite different if Belleville springs are placedunder the head of each nut. In order to properlyanalyze the joint the following assumptions aremade:

1. Assume that the loading plane is at theinterface of the nuts and the Belleville springs.Therefore, the springs are part of the joint for


this analysis. This assembly is called the jointsystem. By adding the movement of theBellevilles to the deflection in the joint, thepreload retained for a given external load can bedetermined.

2. The spring rate of each Belleville will be one-fourth the rate of the bolt. The rate of the jointwill be consistent with the preceding example.Therefore, if the bolt stretches 0.005 in. at agiven preload, then the joint will compress 0.002in. and each spring will deflect 0.020 in.

The new joint diagram is shown in Figure 8.Note that the slope on the bolt side of thediagram is very high compared to that on thejoint side. With the Belleville springs the joint


Figure 7Joint diagram where Lx = external load; dL' = new bolt stretch; Fj = load remaining on the joint = 0.5 × Fp.

system is now 1100% as elastic as without. This is because the total deflection of the joint system isequal to the movement in the joint sections plus two times the deflection of one spring (h).

An external tensile load is now applied to the bolt by pulling on the nut and bolt head. The change inload on the joint system dFj, is very small compared to the external load (see Fig. 9). Without theBellevilles the external load that resulted in a 50% loss of load on the joint now causes a 2.4% loss.This is because originally the joint was very stiff compared to the bolt. By adding springs the jointsystem has become 21 times as elastic as it was.

2.3Movement in the Joint

In the real world, it is very rare to see external loads applied to a bolt once it is in service. A morelikely situation is one in which there is some "movement" in the joint. In other words,


Figure 8Joint diagram using Belleville springs where dL = bolt stretch = 0.005 in.; dT = joint system deflection = 0.002 in. + 2 × 0.020 in. = 0.042 in.

Now the joint is highly elastic compared to the bolt.


Figure 9Joint diagram using Belleville springs where Lx = external load; dL' = new bolt stretch; Fj = load remaining on the joint = 0.976 × Fp.

the joint faces are allowed to come closer together because of unloading of the joint (such as creepor yielding of the gasket material). (Movement in the joint might also refer to unloading due tocreep or relaxation of the joint or bolt materials if there is no gasket. An example of a gasketedjoint is used here because the amount of movement is typically more severe. Use of Bellevilles in anongasketed joint is reviewed in the next section.) To use the joint diagram to analyze this scenarioit is important to think in terms of movement rather than changes in load. In Section 2.2 the jointdiagram was used to determine clamp load on the joint when an external load is applied to the bolt.In this section clamp load is calculated based on how much the joint or bolt moves (due to creep,yielding, thermal expansion, etc.). If the amount of movement in the joint can be predicted, then thechange in force applied to the joint can be determined. The following assumptions are made:

1. There is some sort of elastic material, such as a gasket, between the two sections of the joint.Although the loading curve of most gasket materials is nonlinear, the unloading rate is fairlyconstant [3]. For the purpose of this analysis, only the way in which a bolt's preload is affected byan unloading (creep, yielding, etc.) of the gasket material is considered. Also, this unloading rate isassumed, by coincidence only, to be equal to the spring rate of the bolt. The spring rates of the jointand Bellevilles are consistent with the preceding example.

2. Deflection of the gasket is measured between the faces contacting the surface of the gasket. Theload at this interface is critical because this is where leaks normally occur. For the purposes of thisdiscussion, this is called the loading plane (even though no external loads will be applied here*).Keep in mind that gasket relaxation or yielding will cause unloading rather than loading [4].

3. Since the loading plane is now at the face of the joint (contacting the gasket), the deflection ofthe Belleville's and the joint sections must be added to the bolt stretch. The sum of thesedeflections will make up one side of the joint diagram and the gasket unloading rate the other.

Figure 10 illustrates what happens in theory when there is relaxation in the joint that uses noBellevilles. For this joint diagram,


Bellevilles. For this joint diagram,

* The term "loading plane" may also refer to the plane at which changes in load are measured rather than applied.


Figure 10Joint diagram where dLf = fastening system deflection = 0.006 in.; dLg = gasket deflection = 0.003

in.; Fg = load on the gasket; R = gasket relaxation = 0.003 in.; F'g = residual gasket load = 0.5 × Fg.

dLf=

fastening system deflection (includes bolt andjoint)

dLg=elastic portion of gasket deflection

Fg =original load on gasketF'g=present load on gaksetR =gasket relaxation

The key to this analysis is understanding the dimension for gasket relaxation R. As load isapplied to the fastener, the gasket and joint sections are compressed and the bolt is stretched.When the desired preload is reached, the gasket material will have a given thickness (equal tothe distance between the inside faces of the joint sections). This dimension will change overtime due to many phenomena such as elastic interactions, gasket creep, and differentialthermal expansion. (The dimension normally becomes smaller because most of thesephenomena rob the fastener of preload.) The amount by which the gasket thickness decreasesis R. Note that R is on the fastening system side of the joint diagram (Fig. 10). This is


because as the faces of the joint move together by the dimension R, the fastening system isunloaded. Since R is strictly a change in dimension, it should have no effect on the slope ofthe diagram. It is as if the point of origin of the fastening system side of the diagram has"slipped" or moved by R. The entire joint diagram becomes smaller, resulting in a lowerresidual preload. Preload lost is strictly a function of the amount of the relaxation and thespring rate of the fastening system.

The example in Figure 10 might have the following parameters [5]:

Bolt stretch = 0.005 in.

Joint deflection = 0.001 in.

Gasket deflection (dLg) = 0.005 in.

Relaxation (R) = 0.003 in.


Figure 11Joint diagram where dLf = fastening system deflection = 0.046 in.; dLg = gasket deflection = 0.003 in.; Fg = load on the gasket; R =

gasket relaxation = 0.003 in.; F'g = residual gasket load = 0.935 × Fg.

Since the amount of preload lost due to a give relaxation R is equal to R/dLf,

of the original preload would be lost. This explains why so many gasketed joints that aresubjected to thermal cycles begin to leak after a short time. This analysis reveals that a relativelysmall relaxation results in a substantial decrease in preload.

Figure 11 is a diagram of the same joint with Belleville springs added. If two Bellevilles areused that have a total deflection of 0.040 in., then dLf increases from 0.006 to 0.046 in. If thesame figure for R is used, then

of the original load would be lost to relaxation. To lose as much load as was lost without springrelaxation, R would have to increase by 460% (0.014 in.). Using Bellevilles in this manner toimprove retained preload is commonly known as "live loading" the joint. Live loading iseffective because most gaskets do unload after installation. Increasing the fastening systemdeflection (dLf) reduces the effect of this phenomenon. The ratio of total R to dLf simplybecomes smaller.

2.4Bus Conductors and Differential Thermal Expansion

Bellevilles are not only used in gasketed joints. Using springs to compensate for unloading dueto differential thermal expansion in electrical bus conductors is another very commonapplication. Bus conductors are used for electric circuits that usually carry high voltage andcurrent [7]. This can be very similar to a gasketed joint. When bus conductors are joined (see


Fig. 12), a joint system consisting of two or more sections of bus bar made of soft aluminum istypical. The fastening system consists of bolts, flat washers, and (usually) Belleville springs.These components are often made from nonmagnetic and corrosion-resistant materials such asstainless steel.

In a bolted bus connection, heat is produced from two sources: the resistivity of the bus barmaterial itself and that generated by the contact resistance at the interface of the bus bar sections.In a well-designed joint, the heat will be dissipated by radiation and convection to theatmosphere. If more heat is generated at the contact interface than can be effectively dissipated, a"hot spot" temperature rise will occur. This originates in an area where the joint pressure ishighest because this is where the majority of the current is carried. When pressure


Figure 12Components of a typical bus conductor joint.

is concentrated around the bolt hole, the soft busbar material is relatively free to flow in towardthe bolt. The increased temperature causes ahigher rate of creep. As the bus bar materialcreeps, contact pressure falls and resistanceincreases until the joint fails. To prevent this, the"efficiency" of the joint must be maximized.


"efficiency" of the joint must be maximized."Efficiency" is defined in this case as the ratio ofthe bolted joint resistance to the resistance of anequivalent length of solid bus bar material. First,a sufficient number of fasteners must be used toproduce enough contact pressure. Next, evendistribution of fastener preload will prevent highconcentrations of stresses in the vicinity of thebolt holes. This can be accomplished by usingflat washers on both sides of the joint (it is notrecommended that a Belleville spring be usedwithout flat washers because its lower outsidecorner will tend to "dig" into the soft busmaterial). Finally, a minimum residual preloadmust be maintained when creep or relaxationoccurs. An effective way to do this is to employBelleville springs.

The benefits of using Bellevilles in bolted busconductors can be demonstrated with the jointdiagram. The following assumptions are madefor this analysis:

1. The bus bar is fabricated from two 1/4 in.


1. The bus bar is fabricated from two 1/4 in.thick sections of EC-H13 aluminum.

2. For the first analysis (where no springs areused), the fastening system will consist of a 1/2in. stainless steel bolt and two 1/8 in. thickstainless steel flat washers. In the second case,two 1/8 in. thick stainless steel Bellevilles areused along with the flat washers. The flat load ofthe springs will be 7100 lb.

3. The temperature at assembly will be 70°F.With the conductor in service the bustemperature in the vicinity of the bolt will reach220°F. Since the bolt carries no current, itstemperature will reach only 150°F.

4. For simplicity, the assembly preload will beequal to the flat load of the Belleville, 7100 lb.*

If the loading plane is the joint interface betweenthe washer face and the bus bar, the diagramwould look like the one shown in Figure 13a.(This is similar to the joint discussed


* The assembly procedure for bus conductors in manyplants calls for the mechanic to tighten the bolts until theBellevilles are flattened, as opposed to using torquewrenches.


Figure 13Joint diagrams with different loading plane positions. When the loading plane is moved to the

center of the joint, the diagram has only one side because a change in Lx has the sameeffect on all of the joint components. In other words, as Lx increases, the bolt, joint, washers, and

Bellevilles are loaded.

in Section 2.1.) On the left-hand side is the joint with a deflection ofdT at load Fp, while the right-hand side is the bolt with stretch dL atFp. The dashed lines represent the continuation of the elastic curvesfor the bolt and the joint at higher loads. Note that the elastic curve forthe joint begins to flatten out above Fp. This is because the soft jointmaterial yields at relatively low levels of stress. The drawing to theright of the joint diagram shows a load (Lx) that is applied at theloading plane. As Lx is increased, the joint is unloaded while the bolt isloaded. In other words, Lx will reduce joint deflection and increasebolt stretch.

Now, for a bus conductor connection, the load at the joint interface ofthe two sections of bus bar is of greatest interest. This load is directlyrelated to the contact resistance (and efficiency) of the joint.Therefore, the loading plane used for this analysis will be shifted to thecenter of the joint (see Fig. 13b). There will no longer be two ''sides"to the joint diagram. This is because any load Lx will increase load onboth the joint and the bolt. Since Lx increases


bolt stretch and joint deflection, these values should be on the same side of the jointdiagram .When the preload Fp is applied, the horizontal leg of the diagram will equal thesum of the deflection in the joint and the stretch in the bolt, dT + dL.

For this example, the assembly preload is 7100 lb. Since the joint has hardly yielded at thispoint, the diagram is basically a right triangle. At this preload, the deflection of the joint is0.0052 in. and the bolt stretch is 0.0017 in.* Therefore, the horizontal leg of the triangle is0.0052 + 0.0017 = 0.0069 in.

As stated earlier, when current begins to run through the conductor, the assembly begins toheat up. Using the assumed service temperatures and material properties, the change inlengths DL of the bolt, joint, and flat washers can be determined using the equations

where

rB =coefficient of thermal expansion of the bolt material=6.5 × 10-6 in./(in.·°F)

rJ =coefficient of thermal expansion of the joint material=12.8 × 10-6 in./(in.·°F)

rW=coefficient of thermal expansion of the washer material=6.4 × 10-6 in./(in.·°F)

LB =grip length of the bolt = 1.25 in.LJ =thickness of the joint = 1.00 in.Lw =thickness of a washer = 0.125 in.DT1

=change in temperature of bolt and washer = 150°F70°F

DT2=change in temperature of joint = 220°F 70°F

Therefore,

DLB =0.0006 in.DLJ =0.0019 in.DLW=0.00005 in.


Note that the joint expands more than the bolt does. This will cause an increase in preload.The change in load caused by the differential thermal expansion FT can be found using theequation [2]

where KB and KJ are the spring rates of the bolt and joint, respectively. The spring rate of aflat washer is negligible. Using the figures for the spring rates and expansions calculatedearlier, the increase in preload FTis 1425 lb. This increase is reflected on the joint diagramin Figure 14. Remember that the elastic curve is nonlinear above the preload because thejoint begins to yield at 7000 lb.

Equation (4) is based on a linear elastic curve. The actual increase in load may be slightlyless because of yielding of the joint. On the other hand, loads may be increased because the

* Joint deflection was determined empirically while bolt stretch was calculated using Hooke's law. The yield point ofthe bus bar was found through testing to be 7000 lb.


Figure 14Joint diagram showing the yielding of the aluminum as load is

increased by 1425 lb due to differential thermal expansion. Thedashed line reveals that residual preload will fall to 5500 lb as

temperature returns to 70°F.

yielding will cause temperature (and differential thermal expansion) to increase. Thesephenomena should offset each other, which would lead to fairly accurate results.

When the temperature returns to 70°F, the residual preload will be lower than at assembly(represented by the dashed line in the joint diagram). Note that load on decrease is parallelto the linear portion of the elastic curve. This is because yielding in the bus bar materialeffectively shifted the joint diagram. For this example, an increase of 1425 lb will result ina yield (based on empirical data) of 0.0015 in. A 0.0015 in. shift will cause residualpreload to fall to 5550 lb (a 22% decrease). Because the lower preload will increasecontact resistance, as current runs through the conductor, more heat will be generated. Thiswill not only increase the differential thermal expansion but may also cause the jointmaterial to unload even more due to creep. Each time the bus conductor is cycled, moreload is lost until the connection eventually fails.

Now consider the case where Belleville springs are used. Assume that two springs in serieswith 0.019 in. of deflection (h) are used. The load applied to the bolt is the same as whenno springs were used. Therefore, the vertical leg of the joint diagram in Figure 15 is thesame (7100 lb). However, the two Bellevilles have added 2 × 0.019 in. = 0.038 in. to thehorizontal leg. This decreases the slope of the elastic curve by a factor of 6.5. Since theirmaterials and thicknesses are the same, the change in length of the Bellevilles will be thesame as the flat washers (0.00005 in.). Now the change in load can be computed:


Note that this formula is virtually the same as the one used for no Bellevilles. The onlydifferences are that the change in length of the washers is multiplied by 4 rather than 2 andthe bolt length is 1/4 in. longer. This accounts for the two Bellevilles. The spring rates ofthe Bellevilles are not in the equation because they are in the flat position when the preloadis


Figure 15Joint diagram of two Belleville springs with the loading plane at the center of the joint. Note that the elastic curve changes slope

(dashed line) if load is increased beyond 7100 lb. This is because the springs are flat at this point.

7100 lb, and a Belleville in its flat position has a practically infinite rate because any increase inload will fail to produce a deflection. Plugging in all of the numbers yields an increase in preloadFT of 1347 lb. The increase in load is shown by the solid line on the joint diagram in Figure 16.Note that when load is raised above 7100 lb, the slope of the elastic curve increases. This isbecause the springs will no longer deflect beyond their flat load. The joint material will yield asif there were no Bellevilles. However, as the components return to their original temperature,preload falls quickly until the flat load of the springs is reached. Then the Bellevilles begin tounflatten slightly to "absorb" some of the change in load. This is why the unloading line changesslope (see the dashed line in Figure 16) at the flat load of the springs. For this example, thedifferential thermal expansion resulted in only a 3.4% decrease in preload. This is a substantialimprovement over the 22% lost when no springs were used. The next time the bus conductor iscycled, the increase in preload will be much smaller.

Because the Bellevilles are no longer flat, their spring rate can be incorporated into the formulafor change of load due to differential thermal expansion:

where Ks is the spring rate of two Bellevilles. In this case, the differential thermal expansion willcause an increase of 237 lb. Such a small increase in preload should not lead to any yielding ofthe bus joint. Therefore, when the assembly cools to ambient, preload will return to the samelevel. This is why many plant procedures call for the technician to tighten the bolt


Figure 16With Bellevilles, the differential thermal expansion causes a 1347 lb increase in load. However, when temperature returns to 70°F, the

joint unloads (dashed line) at a steep rate until the flat load of the springs are reached. Then the springs begin to"absorb" some of the change in load so that residual preload only falls to 6860 lb.


until the Bellevilles become flat and then "backoff" 1/4 turn. Backing off allows the spring tounflatten by a small amount so that anydifferential thermal expansion will be "absorbed"by the Belleville. The example reveals that thispractice is unnecessary. After a single thermalcycle, the spring unflattens slightly anyway.

2.5Test of Bellevilles Used on a Bolted Joint

The fixture shown in Figure 17 was designed totest the utility of Bellevilles in an actual joint. Apiece of elastic gasket material was placedbetween two blocks of steel. Using the load cells,each stud was tightened to a preload of 4000 lb.With a hydraulic press an external load F wasthen applied at the outer surfaces of the joint. Asthe external load was increased, the movement ofthe joint was recorded using a dial indicator. Ateach respective increment of deflection the bolt


preload was recorded. In other words, bolt loadwas observed as the joint sections moved closertogether.

The test was performed with and withoutBelleville springs on the studs. The results areshown in Figure 18. In both cases, as deflectionincreased (joint sections moved together) thebolt preload decreased. However, the amount ofload that is lost was greatly affected by the springrate of the fastening system. Note that the studsalone became completely unloaded at 0.0095 in.(Bolt stretch at 4000 lb was only about 0.005 in.The reason it deflected 0.0095 in. to unload thestuds is that there was a total of 0.0045 in. ofmovement between the fixture components andthe hydraulic press that was used.) WhenBellevilles were used, over 62% of the originalpreload still remained at 0.0095 in. deflection.

3Belleville Springs in Practice


3.1Tests of Bellevilles Used on a Flanged Joint

So far this chapter has dealt with how Bellevillesprings work in theory and on a two-bolt testfixture. The flanged system presents someadditional problems to the engineer. Factors suchas elastic interactions, embedment relaxation,and differential thermal expansion have negativeimpacts on preload. To properly evaluate the useof Bellevilles for live loading of flanges thefollowing experiment was developed.

Two separate tests were performed for thisexperiment. The first test was done with noBellevilles and the second with two springs inseries. Variables such as temperature,lubrication, gasket type, pressure, and assemblymethod were held constant. The fixture in Figure19 was designed to simulate a field situation. Atone end of the pipe a plate was welded to allowthe fixture to be bolted to the floor. At the otherend was the flange system: a slip-on flange and a


blind flange. Inside the pipe, furnace elementswere used to simulate a high temperatureoperation.

It was important to have the heating elementsinside the test fixture so that realistic resultscould be obtained. Several testing labs haveattempted to simulate a high temperature processby placing a flange assembly in a furnace. Whenthis is done the bolts heat up before the flangesections and the gasket stress is reduced. This isthe opposite of what normally occurs in a hightemperature process. When the fixture is heatedfrom the inside, the gasket stress is increasedbecause the flange heats up faster than the bolts;this produces more realistic results.

Standard parts and components that wereemployed include

1. Class 300 (ASME B16.5) slip-on and blindflanges

2. (16) 1 in. stud-bolts (A193 B7) with ball


bearings embedded into ends

3. 10 in. SCHD 80 pipe

4. Spiral-wound graphite-filled gasket

5. 16F30 flange washers (Belleville springs)


Figure 17Drawing of a two-bolt test fixture used to test how preload falls as a gasket unloads.

6. Aluminum/copper-based antiseize compound

7. 89 in. micrometer

8. Calibrated dial-type torque wrench

With the gasket between the flange faces the studs were assembled tofinger tight. The length of each stud was then carefully measured using a


micrometer. To provide fast, consistent readings, ball bearings wereembedded into the ends of the studs. Anvils with coined indentationswere also installed over the micrometer tips. The studs were numbered1 to 16, and they were arranged in order around the flange (Fig. 20).Their position in the bolt circle was maintained throughout the tests forcontrol.

The manufacturer of the gasket recommended that a bolt stress of30,000 psi be used. This equates to a stud preload of 18,180 lb. A finaltorque of 227 ft-lb was calculated using a nut factor of 0.15 [8].* Thefirst pass was made using a crossing pattern at 30% of the calculatedfinal torque. The stud length was again recorded. The next two passeswere at 60% and 100% of final torque [9]. Using the changes in studlengths, preload was calculated using Hooke's law. Graphical results ofresidual preload for the first three passes are shown in Figures 21 and22 (stud lengths were recorded after each pass). On the x axis the studposition is displayed from left to right in the order in which they weretightened (stud 1 was tightened first, stud 12 was tightened second,etc.). The y axis is the calculated residual preload (based on the changesin their length). Note that residual preload is the load remaining in eachfastener after the pass had been completed. Next, a reverse pass at 100%of the desired torque was executed.

* This was the nut factor recommended by the manufacturer of the antiseize. This figurelater proved to be low.However, in order to make use of all of the test results, 0.15 wasused through the entire experiment.


Figure 18Graph of decrease in preload as the gasket unloads. (¨) Average preload with two Belleville springs;

(D) average preload with no Belleville.

(shown in Figs. 23 and 24). The same procedure was followed for the tests that employed twoBellevilles.

The next step of the experiment was designed to demonstrate the effects of differential thermalexpansion. After the final lengths were recorded, the heating elements inside the pipe wereturned on. The temperature within the pipe was raised to 1000°F within 1 h. That temperaturewas held for an additional 2 h. The heater was then turned off, and the assembly was allowed tocool to ambient. Stud lengths were recorded. Figures 23 and 24 show preloads after the reversepass and after the assembly was exposed to high temperature.

3.2Conclusions of Flange Tests


The results of the first three passes showed a marked reduction in preload scatter as the springrate of the fastening system was reduced. This is because the effects of the elastic interactionswere reduced. In other words, as the joint is pulled together by the tightening of each stud,adjacent studs are allowed to relax. Going back to the joint diagrams in Figures 10 and 11, thejoint moves together by a dimension R each time a stud is tightened to a given torque. At thesame time, the fastener has been stretched (and springs deflected) by a dimension dLf. Theamount of preload lost in adjacent studs is proportional to the ratio R/dLf. If dLf is largecompared to R, then elastic interactions will have a small impact on preload. Using Bellevillesprings to live load the joint effectively increases dLf by increasing the deflection of thefastener


Figure 19Drawing of flange fixture designed to show how preload is affected by assembly procedure, use of Bellevilles, and thermal cycles.

system. When springs were not used it is apparent that at times the R dimension approacheddLf (or even exceeded it). This explains the very small preloads toward the beginning of eachpass. Any negative loads are obviously due to measurement error.

Next, the flange was thermally cycled. This caused additional stress to be applied to thegasket. The increase in load results in a small amount of gasket set, which may also be calledR. Now that the flange faces have come together by R, the same R/dLf ratio applies to preloadloss. The case where no Bellevilles were used (Fig. 23) displayed a definite downward shift inpreload. This is the reason that many plant procedures call for bolts to be retorqued afterstart-up. When the joints were live loaded the preload lost was immeasurable.

Table 1 summarizes the average preloads and standard deviations for the 16 studs. In general,these results show that reducing the spring rate of the fastening system by live loading canreduce the effects of most of the phenomena that rob the bolts of preload. It must be notedthat this tells us very little about their effectiveness in preventing leaks. Every one of thejoints we tested may have been "leak-free" in the field. On the other hand, each one may havefailed under pressure. It is safe to say, however, that achieving and maintaining proper preloadwill greatly increase the likelihood of obtaining a sound joint.


Figure 20Stud position and tightening sequence used for this experiment.

Figure 21Residual preload for first three passes when no springs were used. ( ) First pass; (·) second pass; ( ) third pass.


Figure 22Residual preload for first three passes when two Bellevilles were used. ( ) First pass; (¨) second pass; ( ) third pass.

Figure 23Residual preload ( ) for the reverse pass and (à) after a thermal cycle when no springs were used.


Figure 24Residual preload ( ) for the reverse pass and (à) after a thermal cycle when two Bellevilles were used.


TABLE 1 Results of Flange Tests With and Without Belleville SpringsNo Bellevilles Two Bellevilles

Avg. Preload Std. deviation Avg. Preload Std. deviation1st pass 3726 2243 3905 13082nd pass 5613 4745 5206 31223rd pass 7999 6886 9338 4361Reverse pass 8862 6486 11693 2123After heatup 7158 6100 11776 2116

4Joint Design Using Bellevilles

4.1Designing a Live-Loaded Joint System

The key to live loading a joint is to reduce the spring rate of thefastening system to a point where a threshold preload will bemaintained. By maintaining this preload limit, the engineer shouldreduce the probability of joint failure. This will ultimately increasesafety and reduce maintenance costs. A discussion of some of thecriteria that should be observed follows.

4.1.1Suitability for Live Loading

Is the joint even a candidate for live loading? Most are not. The aimhere is to save money. Blindly live loading every joint in a plant willcost a great deal and is probably unnecessary. The following questionsshould be asked before deciding to live load a joint system:

1. Is joint failure a particular safety concern?

2. Will the joint be subject to large temperature fluctuations?


3. Is the length-to-diameter ratio of the bolt less than 3?

4. Are any of the factors that rob a bolt of preload (elastic interactions,creep, etc.) prevalent in the system?

5. Does the joint have a history of maintenance problems?

Keep in mind that the use of a Belleville spring may correct only one ofmany problems with the design. Live loading only improves the amountof preload maintained over time. Since low residual preload is one ofthe main reasons for joint failure, the use of springs will reduce the riskof a problem. A case in point is an extensive study conducted by a majoroil company on the prevention of leaks in heat exchanger joints. Overseveral shutdowns, they changed out the gaskets on a number of theirhistorically most troublesome exchangers. Variables such as use ofBellevilles, gasket design, assembly method, preload controlprocedures, and bolt condition were tested for their effectiveness inpreventing leaks. Results revealed that of 62 joints that used springs, 10had developed leaks after 7 months (16%). Of the 22 joints thatemployed no Bellevilles, 9 were leaking (41%). The study also statedthat only 1 of 35 (3%) joints that used Bellevilles and also upgradedgasket designs* leaked compared to 5 of 20 (25%) for the same gasketswith no springs. The use of springs clearly had a positive overall impacton the performance of the joint; however, using them with a bettergasket all but eliminated the problem of leaks.

* These gaskets proved to be very resistant to unloading after thermal cycling.


4.1.2Choice of Material

A Belleville material must be chosen. Theobvious answer would be to employ the samematerial as the one used for the bolt. However,since Bellevilles are normally made frommaterials that make good springs, this is notalways possible. The manufacturers are alsolimited to materials that are readily available instrip form, although some Bellevilles aremachined from bar or slugged from plate. Thecriteria for choosing an appropriate springmaterial are basically the same as those for a bolt(discussed in Section II of this handbook).However, since the stresses in Bellevilles arenormally higher than those in the bolts, theengineer should be more conservative whenchoosing a material. Susceptibility to stresscorrosion cracking, creep, hydrogen


embrittlement, etc are increased at higher levelsof stress. Some of the material properties andapplications are listed below.

1. Carbon steels (AISI 6150 and AISI 1074).Operating range is from 40 to 350°F. Thesematerials are used on indoor and outdoorapplications at ambient temperatures. Their highstrength and ductility leads to excellent springproperties. However, this also makes themsusceptible to environmentally assisted cracking.For this reason, a corrosion protection such asmechanical plating is suggested.

2. Corrosion-resistant materials (Types 301 and316 stainless steels). Operating range is from400 to 550°F. These materials are extremelyresistant to rust and acids. They are slightlymagnetic and obtain their strength through workhardening. For this reason, the thickness of thesprings is limited.

3. Antimagnetic and corrosion-resistant materials


(510 phosphor bronze and beryllium copper).Operating range is from 400 to 300°F. Phosphorbronze is used with a copper bus and siliconbronze bolts and nuts. Both materials have goodelectrical conductivity. Since their elastic moduliare much lower than those of carbon steels, flatloads will tend to be reduced.

4. Precipitation-hardened stainless steels (Types17-7PH and 17-4PH). Operating range is from400 to 550°F. These corrosion-resistant, highstrength materials are used in a wide variety ofindoor and outdoor applications such as liveloading of flanges in cryogenic and hightemperature service and live loading of valvestems. Since they are susceptible to cracking inchloride and fluoride atmospheres, a surfaceprotection such as sulfamate nickel plate isrecommended if the springs are to be used near aseacoast.

5. Hot worked tool steels (H13 and H11 toolsteels). Operating range is 250 to 1000°F. These


Bellevilles are typically used in flange liveloading where fasteners are subjected to highloads and temperatures. Due to their thickness,these parts are usually machined.

6. High temperature materials (Inconel 718 andX 750). Operating range is from 400 to 1100°F.These Bellevilles are used in high temperatureand corrosive service. They are also extremelyresistant to environmentally assisted cracking.Due to their higher cost, however, their use isusually reserved for critical applications.

4.1.3Inside and outside Diameters

The inside and outside diameter limits should beestablished. First, the side diameter of theBelleville should obviously coincide with thebolt size. Second, the outside diameter shouldcontact on a flat, even surface. If there are anyirregularities or if the spring overhangs the jointsection, the loading and deflection characteristics


of the Belleville will be altered.

Some manufacturers offer outside diameters thatare equal to that of the points on a standard nut.These ''flange washers" prevent any interferenceproblems on large heat exchanger flanges wherethe bolts are typically close together. They alsoare designed to have higher flat loads so thesprings rarely have to be combined in parallel.


4.1.4Flat Load

A flat load for the Belleville spring must be selected. Most manufacturers offer more thanone flat load for a given bolt size. Furthermore, multiple washers can be stacked in parallelto achieve varying flat loads. The methods outlined in Section VIII.B of this handbook maybe used to determine preload as if no Bellevilles were employed. Next, a springmanufacturer's literature is used to find a spring or combination of springs with a flat loadthat is close to this preload figure.

The preload that is calculated will not always equal the flat load of any one spring. That'sokay; the Belleville will still work if the loads are not exact. If preload is halfway betweenthe flat loads of two springs, then choose the lower of the two flat loads, providing that theflat load of the lighter Belleville is higher than the calculated minimum preload. This mayrun contrary to intuition, because engineers normally move up to the heavier sizes whendesigning a bolted joint. However, it is important to find a spring that will produce as muchdeflection as possible at a selected preload. Using a lower flat load yields two benefits.First, lighter washers will normally have more deflection available* (assuming equal ODs).Second, the lighter Belleville will fully flatten at the selected preload, whereas the heavierspring will only partially deflect. Remember that as the fastening system deflectionincreases, residual preload also increases. Additionally, overloading the spring will notdamage it, as would be the case with a bolt. It merely remains flat until the joint begins tomove. Therefore, increasing the fastening system deflection by using lighter springs willresult in a lower R/dLf ratio and less loss of preload.

When Bellevilles are not used, the formulas used to determine preload compensate for lossof load due to elastic interactions, etc. [2, 6]. In other words, the engineer adds in extrapreload in anticipation of the gasket unload after the joint is put into service. The Bellevillecounteracts many of these losses by maintaining a higher degree of the original preloadthrough a given movement in the joint. Therefore, the changes in preload due to elasticinteractions, embedment relaxation, differential thermal expansion, etc. will all besubstantially reduced in a live-loaded joint. Thus, if Bellevilles are used, the requiredoriginal preload can be lower.

4.1.5Number of Springs

The number of springs to be used in series on each bolt must be determined. A good rule ofthumb here is to use two, one at each end of the bolt. This will increase the deflection ofthe fastening system by a factor of 715, for a typical 300 lb flanged joint. This increase in


performance should be ample for almost any bolting application.

There are some cases where more than two Bellevilles should be used. Excessivemovement in joint materials, short active bolt length, and high differential thermalexpansion could demand even more deflection from the fastening system. The equation forthe number of Bellevilles required in series is as follows (for a derivation of this formula,see the Appendix):

where

R =the sum of all the factors that loosen the joint (in.). These mayinclude differential thermal expansion, gasket relaxation, bolt creep,elastic interactions, etc.

FP=original preload (lb)N =number of Bellevilles (or Belleville sets) required

* Due to permissible stresses in Belleville springs the deflection h must decrease as the thickness t increases (assumingthat material and other dimensions are held constant).


h =published deflection-to-flat of the Belleville (in.)Fw =published flat load of the Belleville (lb)LB =active length of the fastener (in.)A =root area of the fastener (in.2)

PR =percentage of load we wish to retain

This equation is accurate if the flat load of the Belleville, FW, is greaterthan or equal to the preload Fp (because the slope of load vs. deflectionchanges after the spring becomes flat). Note that the number that iscomputed should be rounded up to the next integer. If the answer isnegative, then no washers are needed.

4.1.6Installation Procedure

An installation procedure for the Bellevilles should be developed. Thesesprings must be used correctly to maximize their benefit. There areseveral important points when live loading a joint:

1. Be sure the bolts are long enough to account for the thickness of theBellevilles.

2. The OD corner of the spring should contact the joint while the IDcorner interfaces with the bolt head or nut (see Fig. 25). If the crown ofthe Belleville is allowed to extend into the hole in the joint, then itsloading characteristics will be altered. The same is true for a nut thatcontacts the spring's bottom surface. Since deflections are relativelysmall, it is difficult for field personnel to see which side is up. Somemanufacturers have incorporated chamfers or painted top surfaces tohelp field personnel determine the orientation of their springs.


3. If a tensioner is used to preload the bolts, then the Bellevilles must beon the opposite side of the joint. It is impossible for the tensioningdevice to apply load to a spring if it is located on the side that is beingpulled.

4. There are rare circumstances where Bellevilles of differentthicknesses and flat loads should be used in series on the same bolt. Forinstance, when differential thermal

Figure 25Examples of proper and improper ways to install two Belleville springs in series.


Figure 26Drawing of how two lighter springs should be combined in series

with one heavy spring.

expansion is extreme, a heavy spring that is onlypartially deflected will prevent the gasket or boltsfrom being overloaded while the system is


heating up (Fig. 26). On the same joint, severallighter springs should be flattened to compensatefor relaxation when the system cools. If this isthe case, then care must be taken not to allow anyof the lighter springs to be "pushed" past theirflat position. This could result in yielding oreven fracture of the spring material.

4.2A Practice Problem

The following scenario is an example of a typicallive-loading problem for a flanged joint. A heatexchanger joint in a petroleum refinery (near anocean) has a history of leaking. The maximumoperating temperature is 750°F. A flange withsixteen 1 in. B7 studs is used along with a spiral-wound gasket with a graphite filler material.Spot faces on the flange surface are slightlylarger than the points on the nut. The gasketmanufacturer has recommended a bolt stress of30,000 psi to seat the gasket. It is estimated thatthe cumulative effects of the factors that rob the


bolt of its preload amounts to 0.010 in.

The first step is to determine whether the joint isa candidate for live loading. Since the flange hashad problems in the past with leaks, it is decidedthat it is. Next, a Belleville material must bechosen. If the exchanger is not insulated, thetemperature at the bolt will probably be muchlower than 750°F. There are several materialsthat will work within these constraints. However,precipitation-hardened stainless steels should beeliminated because of their susceptibility tostress corrosion cracking in marineenvironments. Since a B7 bolt material is used, asimilar H-13 tool steel is selected rather than anexpensive Inconel spring.

Next, a spring with an OD that is smaller than thespotface must be found so that it will interfacewith a flat surface. For a 1 in. bolt, the ODshould be less than 1 7/8 in. Flat load for thespring should also be around 18,180 lb (based onthe gasket manufacturer's suggestion).


With these constraints a flange washer with an OD of 1.810 in. and a flatload equal to the recommended preload is chosen.

The equation to determine the number of Bellevilles to use in series isemployed to check that two springs per bolt is adequate. The assemblyparameters are as follows:

R =0.010 in. L = 5Fp =18,180 lb A = 0.606h =0.020 in. PR = 0.75 (assumed)

Fw =18,180 lb

Solving for N:

Therefore, two washers must be used. Bolt stretch at 18,180 lb is foundto be 0.005 in. Since R exceeds the total bolt stretch, N will never be lessthan 0. That is because without springs the bolts will become completelyunloaded. Also note that N can become very large if the value for PRincreases. For instance, if 95% retained preload were used, then 10springs would be required.

Appendix:Determining the Number of Flange Washers Required Based on aKnown Gasket Relaxation


A.1Definition of Variables

R =anticipated relaxation of the gasket afterassembly

Fp=original preload

N =number of washers required

h =published deflection-to-flat of thewasher (in.)

Fw=published load-to-flat of the washer (lb)

LB=active length of the fastener (in.)

A =root area of the fastener (in.2)PR

=the percentage of original preloadrequired

A.2Derivation of Formula

If Kw is defined as the spring rate of the washer system and the washersare used in series, then

Furthermore, if KB is the spring rate of the bolt, then


These springs are used in series. That is, at any given load theirdeflections are added to obtain the spring rate. The formula forthe spring rate of a system of springs in series is

Therefore, if KF is defined as the spring rate of the fastenersystem, then

and

Now, it is known that a given relaxation of the gasket R willcause some loss of load on the fastening system. The amount ofload lost is directly related to KF. The goal is to reduce KF (byadding flange washers*) to a point where the anticipatedrelaxation of the gasket R will not cause the preload on the boltto fall below some minimum value. The minimum preload is thepercentage of original preload required (PR) multiplied by thebolt preload at the start of the test, Fp.

The quantity of a given washer that will be needed can be


established by using a maximum amount of load that can be lostPLOST,

PLOST can also be defined as the product of the amount ofrelaxation in the gasket R and the spring rate of the fastenersystem.

Therefore,

And solving for N,

References

1. Manual on Design and Manufacture of Coned Disk Springs(Belleville Springs) and Spring Washers, SAE HS 1582, June1988.

2. Bickford, J., An Introduction to the Design and Behavior ofBolted Joints, Marcel Dekker, New York, 1995.


3. Payne, J., Bolted Joint improvements through gasketperformance tests, NPRA Maintenance Conference, SanAntonio, TX, May 1992.

* Note that the spring rate of the fastener system can also be reduced bydecreasing Fw or increasing h. This means that a different fange washer must bechosen. If preload is larger than the load-to-flat of the washer, then thisequation will not work.


4. Bickford, J., That initial preloadWhat happensto it?, Mech. Eng., October 1983, p. 57.

5. Winter, R. J., Gasket selectionA flowchartapproach, Presented at the 2nd InternationalSymposium on Fluid Sealing of Static GasketedJoints, LaBaule, France, September 1990.

6. Keywood, S., Testing and evaluation ofPTFE-based gaskets to chemical plant service,presented at the 5th Annual TechnicalSymposium of the Fluid Sealing Association,Fort Lauderdale, FL, October 1994.

7. ALCOA Aluminum Bus ConductorHandbook, Aluminum Company of America,Pittsburgh, PA, 1957.

8. Bickford, J., Bolt torque: Getting it right,Machine Design, June 21, 1990, p. 67.

9. Bickford, J., Tightening a group of bolts, The


Distributor's Link.

10. Payne, J., Traditional vs. new bolt loadcalculations, Presented at the ASME, PVPConference, San Diego, CA, June 1991.

11. Roberts, I., Gaskets and bolted joints, J.Appl. Mech., 1950, p. 102.


15Vibration-Resistant Fasteners

Joseph R. DudleyNylok Fastener Corporation, Macomb, Michigan

1Introduction

Most bolted joints are exposed to motion,vibration, or shock, which can cause amomentary loss of pressure between the threadflanks. This loss of pressure will cause the jointmembers to slip sideways and loosen. A widevariety of locking fasteners resist loosening.Some fasteners use an accessory part to provide amechanical lock.Others resist loosening throughthe use of prevailing torque.

This chapter presents data on vibration-resistant


bolts and screws, which are fasteners withexternal (male) threads, and nuts, which haveinternal (female) threads. Prevailing torque andlocking type fasteners are covered here; othertypes are discussed in Chapter 41.

Following some basic definitions of terms inSection 2 and a general description of types ofvibration-resistant fasteners in Section 3, Section4 presents data on fasteners with external threadswhose locking mechanisms are nylon patches orchemical adhesives. Section 5 concludes withdata on fasteners with internal threads. Inaddition to 180° and 360° patch and adhesive-locking nuts. Section 5 includes data forprevailing torque type steel hex locknuts. In allcases, tables are included for both inch andmetric series fasteners.

2Definitions

Prevailing torque. Force necessary to rotate a


fastener relative to its mating componentagainst a frictional resistance.

Initial installation torque (maximum). Forceproduced the first time the locking fastener isinstalled, measured during the first fiverevolutions after engaging the lockingelement.

Prevailing on-torque. Torque force read inthe first 360° of rotation in the installation(on) direction after the component thatcontains the frictionally resistant member isfully engaged and after the initial installationtorque has determined.


Prevailing off-torque. Torque read during thefirst 360° of rotation in the off direction afterthe break-loose or breakaway torque has beenmeasured. Both high and low torque shouldbe recorded.

Seating torque. Tightened torque or clampload. The maximum value in lb-ft or N · mthat occurred at the final assembly of the joint,as the joint became fully seated in thetightening direction.

Break-loose torque. The force required toinitiate reverse rotation to a fastener that hasbeen fully seated, clamped, or tightened.

Breakaway torque or static torque. The forcerequired to initiate reverse rotation to afastener that has not been seated, clamped, ortightened.


3Types of Vibration-Resistant Fasteners

Prevailing torque fasteners that use physicalinterference function on the wedge principle.Normally a polymer compound, usually nylon inthe form of a pellet, strip, or patch, ispermanently attached to either an external orinternal thread. This nylon pellet, strip, or patchforms a wedge between the mating threads,creating friction, which is measured as torque.The nylon patch can be applied circumferentiallyfrom partial coverage up to 360° according totorque requirements, to either internal or externalthreads. 180° patch coverage is normallyspecified. The nylon can be reused a minimum offive times; it is nonreactive and can be used at atemperature of up to 250° continuous. There isalso a high temperature version that functions upto 430°F.* Nylon material is generally specifiedwhen a joint has to be disassembled andreassembled for service and repair. See Tables 1,


2, and 58.

Another type of prevailing torque fastener is theall-metal deformed thread internal threadedproduct. This product, since it uses threadinterference to produce torque, has limited reuse.Its temperature range is the same as that of steel,and it is nonreactive. Its use is usuallypermanent, and when the bolted joint isdisassembled a new nut should be used forreassembly. See Tables 11 and 12.

External threads are also specified with deformedthreads, usually a thread forming styleconfiguration. There are many types of threads,and as the joint comes to clamp load the metal inthe thread form is displaced or destroyed. Reuseis generally not recommended.

A third type of vibration-resistant uses achemical adhesive on either the internal orexternal thread to provide a locking action. It isusually not reusable and is not used in a joint


that would normally be disassembled. It can beused at temperatures up to 340°F. See Tables 3,4, 9, and 10.

The torque achieved with either a fine or coarsemetric series or inch series thread is quitecomparable within a size. There are many factorsthat will affect the friction and thus the torque ina bolted joint: the hardness of the parts, type ofmaterial plating, finish, lubricants, tighteningspeed, thread fit, clearance hole size, surfacepressure, and many others. One engineerestimated that over 200 factors can affectfriction.

4Fasteners with External (Male) Threads

The performance standards for various types offasteners with external threads are presented inTables 14.

The standard procedure in testing is that similar


materials and finishes are preferred for the nutand bolt. The most common materials for theinch series is a Grade 5 bolt with phosphatecoating and plain finish. Nuts are class B steeland zinc-plated. Metric fasteners use

* NYTEMP, Nylok Fastener Corporation, Macomb, MI.


TABLE 1 Performance Standards for Nonmetallic Resistant Element Type PrevailingTorque ScrewsNylon PatchInch Seriesa

Clamp load (lb)Strength grade of screw

Nominal threaddiameter andthread flush

SAEGrade

5

SAEGrade 8

ASTMA574

Prevailingon-torque,maximum(lb-in.)

First removalprevailingtorque,

minimum (lb-in.)

Fifth removalprevailingtorque,

minimum (lb-in.)

No. 2-56 234 330 390 2 3b 1.5bNo. 2-64 250 355 410No. 4-40 385 540 635 5 1 0.5No. 4-48 420 595 690No. 6-32 580 820 950 8 2 1No. 6-40 650 910 1,060No. 8-32 900 1,260 1,470 12 2.5 1.5No. 8-36 940 1,330 1,540No. 10-24 1,120 1,580 1,840 18 3 2No. 10-32. 1,280 1,800 2,1001/4-20 2,020 2,850 3,220 40 5 31/4-28 2,320 3,260 3,6805/16-18 3,340 4,720 5,300 85 8 55/16-24 3,680 5,210 5,8703/8-16 4,950 6,980 7,880 110 14 93/8-24 5,590 7,880 8,9007/16-14 6,790 9,600 10,800 150 20 127/16-20 7,580 10,600 12,0001/2-13 9,080 12,800 14,400 220 26 161/2-20 10,200 14,400 16,2009/16-12 11,600 16,400 18,400 270 35 22


9/16-18 13,000 18,300 20,6005/8-11 14,400 20,300 22,900 350 45 305/8-18 16,400 23,000 26,0003/4-10 21,300 30,100 33,800 460 60 453/4-16 23,800 33,600 37,8007/8-9 29,500 41,600 46,800 700 95 657/8-14 32,500 45,800 51,5001-8 38,600 54,500 61,400 900 130 851-12 42,300 59,700 67,10011/8-7 42,400 68,700 77,200 1050 150 11011/8-12 47,500 77,000 87,00011/4-7 53,800 87,200 98,200 1150 200 14011/4-12 59,600 96,600109,00013/8-6 64,400104,000117,000 1300 240 15013/8-12 73,000118,000134,00011/2-6 78,000126,000142,000 1500 280 18511/2-12 87,800142,000160,000aThe clamp torque applied in each case was equal to 75% of the proof loadspecified for the property class.boz-in.Source: IFI Specification 124.


TABLE 2 Performance Standards for Nonmetallic Resistant Element TypePrevailing Torque ScrewsNylon PatchMetric Seriesa

Clamp load 1(kN)

Property classof

screw

Nominal threaddiameter andthread pitch

8.8 9.8 10.9Prevailing on-

torque,maximum (N ·

m)

First removalprevailing off-

torque, minimum(N · m)

Fifthremoval

prevailing off-torque,

minimum(N· m)

M1.6 × 0.35 0.6 0.1 0.01 0.004M2 × 0.4 1 0.2 0.02 0.01M2.5 × 0.45 1.6 0.4 0.05 0.03M3 × 0.5 2.4 0.6 0.14 0.06M3.5 × 0.6 3.3 0.9 0.22 0.11M4 × 0.7 4.3 1.2 0.26 0.16M5 × 0.8 6.9 8.8 2.3 0.36 0.23M6 × 1 9.8 12 3 0.45 0.3M8 × 1.25 18 23 10 0.9 0.58M10 × 1.5 28 36 14 1.8 1.1M12 × 1.75 41 52 21 2.6 1.5M14 × 2 56 72 30 3.6 2.3M16 × 2 7176 98 40 5 3.4M20 × 2.5 110 150 60 8 5.5M24 × 3 160 220 90 13 8.5M30 × 3.5 250 350 120 19 13


M36 × 4 370 510 150 28 18aThe clamp torque applied in each case was equal to 75% of the proof loadspecified for the property class.Source: IFI Specification 524.

TABLE 3 Performance Standards for Chemical Locking Externally ThreadedFastenersInch Series

Breakawaytorque (lb-in.)

Prevailing off-torque, (lb-in.)

Nominal size andthread pitch

Prevailing on-torque,max (lb-in.) Cured Min Cured Min

1/4-20 15 20 125/16-18 20 40 153/8-16 45 100 457/16-14 60 150 601/2-13 80 200 909/16-12 100 250 1205/8-11 125 350 1503/4-10 150 450 2007/8-9 180 550 250Source: Data from Nylok Fastener test laboratory.


TABLE 4 Performance Standards for Chemical Locking Externally ThreadedFastenersMetric SeriesNominal size andthread pitch

Prevailing on-torque, max (N · m)

Breakawaytorque, min (N ·

m)

Prevailing off-torque, min (N · m)

M6 × 1 0.6 1 1.7M8 × 1.25 1.5 8.5 5.6M10 × 1.5 2 22 10.2M12 × 1.75 3.5 27 14.1M14 × 2 5.5 62 24.9M16 × 2 8 88 37.6M20 × 2.5 11 110 49.5M24 × 3 14.5 137 61.9Source: Data from Nylok Fastener test laboratory.

10.9 bolts with phosphate coating and class 10 steelnuts with plain finish unless otherwise specified.

No table has been developed or published for externallythreaded thread forming screws. Because of their vastvariety and the fact that they are not a prevailing torquetype of fastener, it is very difficult to determine orpublish torque values for their installation.

The following rules apply in testing chemical lockingexternally threaded fasteners (Tables 3 and 4).


1. Performance of Test. During the completeperformance of the test, the nut shall be turned whilethe test bolt shall be restrained from turning. If the nutis restrained so that the torque reading is not adverselyaffected, it is permissible to turn the test bolt.

2. Prevailing On-Torque. The test nut shall beassembled onto the preapplied adhesivecoated area ofthe bolt until the full thickness of the nut is engagedinto the adhesive. The maximum prevailing on-torqueshall be measured and recorded during the next 360° ofrotation, while the fastener is in constant motion. Thespeed of rotation shall be approximately 1020 rpm.

3. Breakaway/Prevailing Off-Torque. Assemble theadhesive-coated bolts and nuts. The activatedadhesiveboltnut assembly shall be allowed to cure at 25±2°C for 24 hr. The fastener shall then be rotated in theoff direction with a torque-measuring device of suitablecapacity and sensitivity. Torque readings shall bemeasured and recorded at breakaway torque and within360° of further rotation for the maximum prevailingoff-torque.

5


Fasteners with Internal (Female) Threads

Performance standards for fasteners with internalthreads are presented in Tables 512.

Caution is urged when selecting prevailing torqueelements or locking elements in internally threadedproducts. Depending on the style of nut specified,especially when using standard finished or metric hexnuts, style 1 (no orientation required), there may not beenough threads inside the nut to coat with nylon oradhesive to achieve the amount of torque specified.Table 13 shows the limits and the number of threadsavailable with standard nut thickness.


TABLE 5 Performance Standards for 180° Nylon PatchNuts With Internal ThreadsInch Seriesa

Size Initial installationmaximum (lb-in.)

Prevailing1st off

minimum(lb-in.)

Prevailing 5th offminimum (lb-in.)

10-32 5.5 3 1.8

1/4-20 18 7 2.5

5/16-18 30 12.5 3.5

3/8-16 35 14 8.5

7/16-14 55 21 16

1/2-13 60 30 20.5

5/8-11 165 60 33

3/4-10 190 80 60

7/8-9 220 110 701-8 60b 30b 12baNut: Grade B steel HRC 28 max. Bolt: Grade 5 steelHRC 2534. Test procedure per IFI 100/107.


blb-ft.Source: Data from Nylok Fastener test laboratory.

TABLE 6 Performance Standards for 180° Nylon PatchNuts With Internal ThreadsMetric Seriesa

SizeInitial installationmaximum (N ·

m)

Prevailing 1stoff minimum (N

· m)

Prevailing 5thoff minimum (N

· m)M4 ×0.7 0.6 0.2 0.1

M5 ×0.8 1.3 0.4 0.0

M6 ×1.0 2.2 0.8 0.4

M8 ×1.25 4.0 1.7 1.0

M10× 1.5 12.0 5.0 2.5

M12×1.75

17.0 5.5 3.0

M14× 2.0 24.0 11.0 4.5

M16× 2.0 35.0 18.0 10.0

M20× 2.5 70.0 35.0 20.5

aNuts: Class 10 zinc finish HRC 2636. Bolts: Class


10.9 phos and oil finish HRC 3339. Test procedureper ASME B 18.16.1M. Program B, Table 4.Source: Data from Nylok Fastener test laboratory.


TABLE 7 Performance Standards for 360° Nylon PatchNuts With Internal ThreadsInch Seriesa

SizeInitial installationmaximum (lb-

in.)

Prevailing 1st offminimum (lb-in.)

Prevailing 5th offminimum (lb-in.)

10-32 11.5 5.0 2.5

1/4-20 18.0 9.0 3.5

5/16-18 30.0 16.0 5.0

3/8-16 42.0 15.0 10.0

7/16-14 60.0 21.0 14.0

1/2-13 65.0 24.0 17.0

5/8-11 95.0 55.0 20.0

3/4-10 200.0 95.0 45.0

7/8-9 240.0 80.0 55.01''-8 44b 33b 15baNut: Grade B steel HRC 28 max. Bolt: Grade 5 steelHRC 2534. Test procedure per IFI 100/107.blb-ft.


Source: Data from Nylok Fastener test laboratory.

TABLE 8 Performance Standards for 360° NylonPatch Nuts With Internal ThreadsMetric Seriesa

SizeInitial

installationmaximum (N ·

m)

Prevailing 1stoff minimum (N

· m)

Prevailing 5thoff minimum (N

· m)

M4× 7 (1) 0.5 0.3

M5× 0.8 (1) 0.7 0.4

M6× 1.0 4.0 2.0 0.5

M8×1.25

6.0 2.0 1.5

M10× 1.5 14.0 4.0 2.0

M12×1.75

23.0 7.0 3.5

M14× 2.0 27.3 10.0 5.0

M16× 2.0 33.0 17.0 8.0

M20× 2.5 50.7 22.0 12.0


M24× 3 55.0 25.1 15.0

aNuts: Class 10 zinc finish HRC 2636. Bolts: Class10.9 phos and oil HRC 3339. Test procedure perASME B 18.16.1M.Source: Data from Nylok Fastener test laboratory.


TABLE 9 Performance Standards for Adhesive LockingInternally Threaded FastenersInch Series, Coarse Threadsa

Size(in.)

Initialinstallation

maximum (lb-in.)

Breakawaytorque

minimum (lb-in.)

Prevailing off-torque 1st360 after breakaway

minimum (lb-in.)

1/4 3 20 155/16 5 28 203/8 7 40 287/16 9 90 1201/2 16 125 1459/16 18 151 1635/8 20 260 2553/4 45 470 4057/8 50 500 4801 72 560 530aNuts: Grade B steel, commercial zinc-plated, HRC 28maximum.Bolts: SAE Grade 5 steel, HRC 2534.Source: Data from Nylok Fastener test laboratories.

TABLE 10 Performance Standards for Adhesive LockingInternally Threaded FastenersMetric Series, Coarse Threadsa

SizeInitial

installationBreakaway

torquePrevailing off-torque 1st

360 after breakaway


Size maximum(N·m)

minimum(N · m)

minimum(N · m)

M4 ×0.7 0.1 1.0 0.8

M ×0.8 0.1 1.0 1.0

M6 ×1.0 0.2 2.5 2.0

M8 ×1.25 0.5 4.5 3.0

M10× 1.5 0.7 6.0 4.5

M12×1.75

2.0 18.5 10.5

M14× 2.0 3.0 30.0 13.0

M16× 2.0 3.5 35.0 18.0

aNuts: Class 10 steel commercial zinc-plated, HRC 2636.Bolts: Steel class 10.9, commercial zinc-plated, HRC 3339.Source: Data from Nylok Fastener test laboratory.


TABLE 12 Prevailing Torque Type Steel Hex LocknutsMetric SeriesPrevailing torqueb(N · m)

Class 5 and 9 Class 10Proof load (kN) Clamp loada (kN)

Nominal nut diam. and threadpitch

Class5

Class9

Class10

Class5

Class9

Class10

Firstinstallation

Firstremoval

Fifthremoval

Firstinstallation

Firstremoval

Firstremoval

M3 × 0.5 2.62 4.53 5.23 1.43 2.45 3.13 0.43 0.12 0.08 0.60 0.15 0.10M4 × 0.7 4.57 7.90 9.13 2.50 4.25 5.47 0.90 0.18 0.12 1.20 0.22 0.15M5 × 0.8 8.23 13.00 14.80 4.05 6.92 6.64 1.80 0.29 0.20 2.10 0.35 0.24M6 × 1 11.70 18.40 20.90 5.73 9.80 12.50 3.00 0.45 0.30 4.00 0.55 0.40M6 × 1.25 21.60 34.40 36.10 10.40 17.80 22.80 7.00 0.85 0.60 9.00 1.15 0.80M10 × 1.5 34.20 54.50 60.30 18.50 26.30 36.10 10.50 1.50 1.00 14.00 2.00 1.40M12 × 1.75 51.40 80.10 88.50 24.00 41.10 52.50 15.50 2.30 1.60 21.00 3.10 2.10M14 × 2 70.20 100.00121.00 32.80 56.10 71.60 24.00 3.30 2.30 31.00 4.40 3.00M16 × 2 96.80 149.00165.00 44.80 76.50 97.50 32.00 4.50 3.00 42.00 8.00 4.20M20 × 2.5 154.00225.00280.00 69.80 110.00152.00 54.00 7.50 5.30 72.00 10.50 7.00M24 × 3 222.00325.00374.00101.00159.00220.00 80.00 11.50 5.00 106.00 15.00 10.50M30 × 3.5 363.00516.00595.00 94.50 252.00349.00 108.00 16.00 12.00 140.00 19.00 14.00M36 × 4 515.00752.00866.00138.00366.00509.00 136.00 21.00 16.00 180.00 24.00 17.50aThe clamp loads for class 5 nuts are equal to 75% of the proof load of property class 5.8 bolts for diameters M3M24 and 75% of the proof load ofproperty class 4.8 bolts for diameters larger than M24. The clamp loads for class 9 nuts are equal to 75% of the proof load of property class 9.8 bolts fordiameters M3M16 and 75% of the proof load of property class 8.8 bolts for diameters larger than M16. The clamp loads for class 10 nuts are equal to75% of the proof load of property class 10.9 bolts. Proof loads of bolts are given in SAE J1199.bFirst installation torque is the torque occurring while the nut is being assembled on the test bolt for the first time. The torque is measured through the first360° of nut rotation after two full bolt threads protrude through the nut. For the nut to be acceptable, the maximum torque reading shall not be greater thanthe tabulated value. First and fifth removal torques are the torques occurring during the first and fifth times the nut is disassembled from the last bolt. Thetorque is measured through the first 360° of nut removal rotation after (for first removal only) the clamp load has been fully relieved. For the nut to beacceptable, the maximum torque reading shall not be less than the tabulated value.Source: ASME Specification B18. 16. 1M.


TABLE 13 Coverage of Locking or Prevailing Element in Standard MetricHex Nut Threadsa

Standard hex, style 1 Standard hex, style 2bThread size(coarse)

Totalthreads

Availablethreadsc

Totalthreads

Availablethreadsc

M5 × 0.8 4 2 5 3M6 × 1.0 4 2 5 3M8 × 1.25 4 2 5 3M10 × 1.5 5 3 6 4 1/2M12 × 1.75 6 4 6 1/2 4 1/2M14 × 2 5 3 6 4M16 × 2 5 3 6 4M20 × 2 5 3 6 4aOne starting thread each side, no orientation.bStyle 2 nuts, not stocked, expensive, mostly by special order only.cFor hex flange nut, must be oriented, add two threads. For cap (acorn)nuts, add one thread.

TABLE 14 A Comparison of 180° Patch, 360° Patch, and Collar Nuts180° Patch 360° Patch Nylon insert collar nut

A. External threadcondition

1. Thread chatter Affects torqueregistrya No effect Affects torque registry

2. Plating finishes Affects torque No effect Affects torque3. Foreignmateria deposits Affects torque No effect Affects torque

4. Long run-down Good Better Good

B. Nut1. Weight Standard Standard Increases 1525%2. Size Standard Standard Increases by 2 threads minimum

C. Mating bolt Standard Standard Increases by 2 threads minimum tooverall height


D. Manufacturingeffort Good Good Extremely critical in skirt area and

nylon placement

E. Cost Standard Standard Increases up to 50% over patchnuts

F. Assembly forces Horizontalagainst threads

Horizontalagainst threads

Vertical against nylon ring andcrimped area

G. Plating No effect No effect Plating cracks and splits aftercrimping

aThe actual torque/tension on every automotive critical and/or safety joint is recorded on acomputer. These printouts are the source for recalls. Chatter destroys the computerprintout.


Adhesive patched nuts show low installationtorque-on values as the compound is soft andmixed during installation. Breakaway torquescan be as strong as the breaking point of somesoft steels and can be modified to meet customerrequirements. Prevailing off-torque of adhesiveparts, once the bond is broken, is less than that ofplastic patch parts. Some second time hardeningoccurs, as not all of the encapsulated beads crushand mix the first time, but the values obtainableare happenstance numbers. Reuse of adhesive-coated fastener joints is generally notrecommended. Breakaway values, as stated inTables 9 or 10, could be quite high and in thecase of larger sizes (5/8 in. M16) make the partsdifficult to remove. The time required foradhesive hardening might limit adjustment of theinstalled fastener. Heat and humidity can causepremature hardening of the adhesive, therebylimiting the shelf life and storage conditions of


preapplied adhesive fasteners to inside areasaway from furnaces and heaters. Because theadhesive is soft and fills the thread interspacescompletely when smeared and mixed by thefasteners, long rundowns are not recommended.

The ideal condition is for the bolt end to betightened to flush with the nut face so theadhesive coating is in full contact within thethreads. It is suggested that no more than one ortwo threads be used for normal designs and atfour threads of protrusion the joint strength maybe reduced as much as 90%, the reduction beingdue to the wipeout of the adhesive by theexternal fastener threads during installation.Another design consideration when usingadhesive locking features is the number ofthreads available for coating. Standard nuts havefew internal threads available for adhesive. Astandard 5/16 in. (M8) nut has four internalthreads, and a 1/4 in. (M6) nut has six. Sincemany specifications require one thread to be free


of adhesive for assist in starting, the two threadsof the 5/16 in. and four of the 1/4 in. nuts do notleave much room for adhesive and highbreakaway torques.

Nylon Patch Nuts, 180° patch coverage vs. 360°patch coverage compared to the collar nut. Thereare distinct differences in the applications ofthese three nuts. Whereas the various torquevalues are relatively the same, the 180° standardpatch exposes 180° of bare metal to bare metal,while the 360° patch and the collar nut use acircumferential nylon surface against the externalbare metal thread. The differences are shown inTable 14.

Commercial zinc plating is very smooth andpresents a difficult base for adhesive bonding.When using adhesives on very smooth coatings(including chrome and nickel plates as well aszinc), consult Nylok for technical assistance.Smooth coatings may produce minimal torquevalues. Other finishes, such as dry phosphate,


may double the locking forces due to itsexcellent bonding surface, and may generate veryhigh breakaway torques.

Small internal threads will have low strengthsdue to limited area for adhesive adhesion.


16Concrete Anchors

Robert L. ZinkHilti, Inc., Tulsa, Oklahoma

1Introduction

1.1Anchor Types and Applications

In construction, supporting structuralconnections are made with bolts and welds.When fastening to concrete, bolts may be cast inplace or postinstalled. Cast-in-place fasteningmechanisms such as J-bolts or slotted receptaclesare selected at the blueprint stage, and theprocess is fairly straightforward. In retrofit orremodeling, or when the anchor placement


cannot be predetermined, postinstalled fasteners(often referred to as concrete anchors) providethe desired flexibility. Their selection requires aninitial determination of the type of anchor,followed by a determination of the loadingcapacity of the anchor or anchor group.Consideration must be given as well to dynamicloading, corrosion environments, and high orlow temperature service.

Concrete bolts known as mechanical expansionanchors operate by expansion in concrete,categorized into two mechanisms: torque-controlled and force-controlled. In both types,the anchor is lightly hammered into a hole drilledin the concrete or masonry. Then a wedge orsleeve is expanded against the side of the hole.

Torque-controlled expansion anchors, such asthe one shown in Figure 1, develop their holdingpower by applying a specified torque to the boltor nut. The tensioning force reacts to press outwedges into the base material. Loads later


applied will not affect the anchor if they are lessthan the tensioning force. If the applied load doesexceed the tensioning force, followup expansionoccurs, further setting the anchor up to the pointof failure.

Force-controlled anchors are also known asdisplacement-controlled, deformation-controlled, drop-in, or flush anchors. A typicalexample can be seen in Figure 2. They are set bydriving a plug into the anchor body, forcing theshell of the anchor to expand into the concrete. Asetting punch moves a plug to the bottom of theanchor, providing an expansion that isdimensionally determined. Unlike torque-controlled expansion anchors, the applied loaddoes not affect any follow-up expansion of theanchor. Friction and keying, however, providethe holding power in both types.

Heavy duty anchors are also available thatcombine internal threading with follow-upexpansion capabilities. An example is shown in


Figure 3.

In designing with expansion anchors, aconservative rule of thumb to obtain maximumperformance is to allow a spacing betweenanchors of 10 anchor diameters and a distance to


Figure 1A stud anchor.

the edge of the concrete of five anchor diameters.These are known as "critical spacing" distances.Adhesive anchors, which use chemically reactivepolymer resins to secure the bolt, can be set closerto the edge and to each other, since no settingstresses are introduced. With these, as well as withmetal proprietary anchors, it is wise to follow themanufacturer's instructions as to allowable spacing.Modifications to allowable loads to account for lessthan critical spacing can be made using the methodsset out in Section 2 of this chapter.

1.2Available Data

The requirements of the Nuclear Regulatory


Commission, regional and local code bodies,departments of transportation, and seismicapplication considerations have brought aboutextensive testing by major anchor manufacturers toqualify their anchors. Independent laboratory testingor witnessing is required for many applications, andthe reports generated are made available todesigners and users. In addition to published data onholding values, some manufacturers also have fieldengineering support available to designers andusers. When on-site testing is required by anapproval agency or project engineer, themanufacturer's engineers can furnish guidance onthe general procedure and specific methodology.

2Design of Connections

2.1History

Tying down today's modern structures requiresmodern concrete anchoring techniques. Reliance onJ-bolts cast in concrete was accepted until the


arrival of nuclear plants and svelte bridges and otherlean infrastructure. The end of monolithic cubes andthe arrival of fast-track scheduling brought a needfor flexibility and efficiency in design.

Not until the close of the 1960s did engineers paymuch attention to concrete anchoring technology.Manufacturers of postinstalled anchors publishedtest results for their proprietary systems, but therewas no generally agreed-upon technical approach.Creative efforts since the mid-1980s in Europe(University of Stuttgart) and in the United States(University of Texas at Austin) have resulted intheoretical design models that agree quite well witha large body of test results [1].

Figure 2


A flush anchor.


Figure 3A heavy duty internally threaded expansion anchor.

2.2Failure Modes

A concrete anchor may fail in one of four ways.(1) The load may break out a chunk of concreteor (2) it may break the body of the anchor. (3)The anchor may pull out without any majorfailure of the concrete. (4) The concrete maysplit in a plane beneath the anchor andperpendicular to the anchor axis.

The fourth mode is usually the result of closelyspaced reinforcing bars and can be avoided byproper design. Splitting of the concrete caused byanchors spaced too close to an edge or too


closely together is prevented by the overalldesign approach developed in this chapter.

In design, always start with test results that arepublished by the anchor manufacturer. Somemanufacturers also publish computer programsthat facilitate anchor selection [2]. Also check tosee if the manufacturer has a design load capacitytable approved by the appropriate code-enforcingagency.

It is also important to refer to the ASTMstandards for postpour-installed anchorperformance in concrete (Z5818Z and Z5819Z),which provide for anchor reliability classified asto sensitivity to installation technique [3,4].These affect the pullout mode valves, asdescribed later in this chapter, by applying ffactors. Also, seismic compensation testqualifications are addressed in the ASTMstandards.

2.2.1


Concrete Failure

When the anchor fails and pulls a chunk ofconcrete with it, the attached concrete appearssomewhat like a 45° cone. This led to the designapproach that is incorporated as Appendix B tothe American Concrete Institute (ACI) standard"Code Requirements for Nuclear Safety RelatedConcrete Structures (ACI 349-90, Appendix B-Steel Embedments)." If the conical concretefracture surface area is recognized as a limitingfactor furnishing the constraint, then the anchorcapacity is directly related to this surface area(Fig. 4). Since the flat-top unbroken surface areaof the 45° cone varies directly with the brokensurface area, this area can be used as an indicatorof the concrete tensile strength available to theanchor.

Mapping this exposed surface area surroundingthe anchor will indirectly provide the concretebreakout strength. This model provides a surfacearea roughly equal to a circle with a diameter


twice the depth of the anchor. If this mappedcircle overlaps an edge, it is clear that thesubmerged cone surface is similarly limited, andthe anchor capacity is reduced accordingly (Fig.5a). In an analogous manner, if multiple anchorsare installed so that their surface circles overlap,the total load capacity of the anchor set isreduced (Fig. 5b).

This mapping allows the designer to calculate thereduction in the anchor capacity for edge andspacing conditions by calculating the boundedarea of the overlapping circles and has beenaccepted in practice. The pioneering work atStuttgart and Austin has generated anothermodel, however, that is easier to work with andprovides theoretical values that agree moreclosely with test results. This new approach iscalled concrete capacity design (CCD).


Figure 4Conical breakout under tension.

In CCD, instead of a cone model, a rectangular prism (apyramidal "stress block") is used; an approximately 35°angle from the surface has been adopted instead of 45°.*The resulting projected surface square is then taken to beequal in area to 9 times the square of the effective anchor


embedment (Fig. 6). Empirical formula factors areintroduced to make the theoretical results match testresults. The exact shape of the prism is not important inanalyzing the capacity of a single anchor implanted in avast expanse of concrete. The corresponding strengthformula can adjust for changes in area between prismmodels. What is newly apparent, however, is that

1. Calculating rectangular areas is significantly easier andmore user-friendly than calculating partial and overlappingcircle areas.

2. The CCD model is more sensitive to anchor edgedistance and spacing than the cone model. The generatedsurface is larger, so overlaps and edge-area "chop-offs"will be more likely.

In addition to this prism manipulation, the CCD approachintroduces size effect factors that in combination providecalculated results that agree with test results through abroader range of geometric conditions than the conemodel. The expected universal acceptance of the CCDmethod arises from the happy result of coupling easier usewith greater accuracy.

* The Precast/Prestressed Concrete Institute (PCI) adopted arectangular prism while maintaining the 45° angle [5].


Figure 5(a) Conical breakout close to edge.


(b) Conical breakout with centerline spacingless than one projected diameter

(surface overlap).

The CCD method, like the 45° conical method, isapplicable to both tension and shear calculations.The actual calculation process is detailed below,adding modifying factors gradually, along with arudimentary explanation of the effect on the baseformula.

Single Anchor in Tension


Concrete anchors (or internally threaded insertsor slotted receptacles) will have an effectiveembedment depth, hef, that represents theconcrete prism depth that will be pulled. Forcastin-place anchors, this will be the depth abovethe embedded head of the anchor or above anysubmerged anchor plate. For wedge-typeexpansion anchors or undercut anchors (thosethat use a special tool to cut out a submergedannulus for the anchor's expansion), the effectivedepth will be measured from the largestexpanded diameter.


Figure 6Pyramidal breakout model.

The formula for the concrete capacity of an anchor spaced remotely from an edge and wellaway from other anchors is given by the formula

where

Nb=concrete tensile breakout strength (lb)

k=a calibration factor empirically derived from tests

l=

a coefficient for concrete density (l = 1.00 for normal weightconcrete; 0.85 for ''sand-lightweight";0.75 for "all-lightweight")

¦'c=concrete compressive strength (psi)

hef=effective embedment depth, as described above (in.)

The calibration factor k allows the use of any unit system by the manufacturer publishingthe data. The formula enables the designer to take the manufacturer's published values andadjust for depth and different concretes. This k coefficient uses the statistical "scatter" oftest mean values, giving a deserved advantage to those fasteners that have low variability as


demonstrated in testing. It is set at the mean value for specimens tested in cracked concrete,less 1.67 standard deviations (one-sided 95% tolerance limit with 75% confidence for theappropriate sample size).

For inch-pound systems, the k factor may generally be taken as equal to 21 for headed boltscast in place and 18 for steel expansion and undercut anchors. It wraps up the "fit" to thetest results.

The depth influence relates to the 1.5 power instead of the power of 2 for the 45° conemethod, so we see a weaker influence of embedment change when using the CCD method.

The concrete compressive strength square root factor, relating to the effective concretetensile strength, conforms to the 45° cone model.


Nb may be reduced by anchor position, manner of loading, and condition of concrete:

where Nn = the nominal tensile strength of a fastener or group of fasteners and the otherfactors are as defined below.

Edge and spacing effects: The modification factor is

where

AN=

projected surface area of all anchors, excluding truncations by edgeproximity or spacing (equals the bounded area) (see Fig. 7.)

ANo=

the projected surface for one anchor remote from an edge (nine timesthe square of the effective anchor depth)

Dimensional units of these factors cancel, so any consistent dimensional system can beapplied.

Effect of eccentric loading: Nb is further limited by modification factors relating toeccentric loading, y1:

where

e'N=

offset of the load vector from the anchor systemcentroid (in.)

hef=effective embedment depth (in.)

This y1 modification is valid only if the eccentricity resultant vector is between anchors. Ifthe resultant of the load indicates that one or more anchors do not contribute in tension,


Figure 7Pyramid model:

projected surface areas overlappingeach other and extending beyond edge of slab.


neglect their contribution totally. If there is eccentric loading about two axes, compute foreach axis and use each result as a factor.

Edge distance modification: An additional modifier for edge distance, y2, is introduced:

where c1 is the closest concrete surface edge distance, in inches (Fig. 8). Use 1.0 asmodifier y2 if the bounds of the projected area do not overlap any edge.

If there are three or more edges involved, it is easier to just reduce the effective depth to thelargest involved edge distance that is less than c1.

See Section 2.2.6 for minimum allowable edge distances.

Concrete cracking factor: The capacity of the anchor is reduced if there is contemplatedcracking at the anchor location.

The concrete cracking factor, f3, equals 1.0 unless there is no expectation of flexuralstructural cracking or if cracking is controlled by reinforcement to within 1/64 in., in whichcase a factor of 1.4 may be used. Concrete as a structural member is often designed toaccommodate flexural cracking, which may be controlled by steel reinforcement. Thus, thestudy of anchor performance in cracked concrete was a necessary step in the technology [6].

Single Anchor in Shear

Since an anchor failing the concrete in shear will be near an edge, the basic formula forconcrete breakout shear strength Vb includes an edge factor at the formula's end (Fig. 9).

The factor 7 is the k coefficient, which adjusts the calculated load to test results.

l is the load-bearing length of the fastener in inches (limited to 8 times the outside shaftdiameter; also, for expansion anchors with a sleeve surrounding and spaced radially awayfrom a central anchor shaft, limit l to twice the outside diameter).

d0 is the outside anchor shaft diameter (in.)

l is the coefficient for the concrete composition:


Figure 8Pyramid model:

bolt close to edge.


Figure 9Failure in shear, near edge.

l =1.00 for normal concretel =0.85 for "sand-lightweight" concretel =0.75 for "all-lightweight" concrete

¦'c is the concrete compressive strength (psi)

c1 is the distance to the edge (in.) that the load attempts to break out

If one is faced with the combination of a narrow, thin member, it is easier to use theformula by reducing the effective anchor depth hef to the maximum involved edge distance.

Similarly to the effects on tensile capacity, concrete breakout strength Vb, may be reducedby anchor position, manner of loading, and condition of concrete.

This formula's limitation factors are defined in the following four sections.

Edge and spacing effects: The modification factor is

As with the tensile model, for this ratio apply the rectangular prism. Because the slab edgesurface projected area in the direction of the shear load for a single anchor is only half thefull surface used in the tensile example, the area AV0 is numerically equal to only 4.5 timesthe anchor-to-edge distance (Fig. 10).


Again, as with the tensile model, there are three y modification factors:

Effect of eccentric loading: For eccentricity correction:

This is similar to the equation used in the tensile model [see Eq. (3)]. The same commentsas to anchor contribution apply.

As with the tensile calculation, to use the formula the eccentricity vector must not belocated outside anchor spacing. If it is, use standard structural analysis.


Figure 10Failure in shear, pyramid model projected area.

Edge distance modification: For edge and spacing effects,

where c2 is the edge distance to the closest edge parallel to the shear direction and c1 is thedistance to the edge that the load attempts to break out (Fig. 11).

Concrete cracking factor: The shear capacity is reduced if structural flexural concretecracking is contemplated at the anchor location. The modification factor is y6. For crackedconcrete, use y6 = 1.0; for straight reinforcement along the edge, use y = 1.2; for edgesreinforced with stirrups spaced at no more than 4 in. or if no structural cracking isexpected, use y = 1.4.

Figure 11Shear:

edge distance modification.


Short anchors: One must pay attention to stubby fasteners, which are capable of prying outthe concrete at the back of the anchor. Design for anchors with a depth-to-diameter ratio ofless than 3 by using a moment analysis or by testing.

2.2.2Steel Shaft Failure

Generally a product should be optimally designed so that all components wear out at thesame time. When safety is involved, however, the failure order becomes important. Thenuclear industry has indicated a preference for fastening designs that cause the anchor bodyto fail well before any concrete breakout. Several advantages are proposed:

1. The mechanical properties of steel have lower variability than those of concrete, andtherefore the failure load can be more accurately predicted.

2. The ductile failure of steel is less catastrophic, and it may be detected before completefailure.

3. There would be some energy absorption due to plastic deformation, which will defusethe failure load through stress redistribution.

The drawbacks of this approach are that

1. The anchors must have high embedment to ensure that it fails before concrete breakout,which adds cost and may not be feasible in thin slabs.

2. Unless the anchor is highly embedded, an analysis must still be performed on theconcrete failure mode to ensure that ductile failure of the steel anchor will occur first.

3. Relative to overloading of the overall structure there is not much energy absorption in aconcrete anchor being stressed to ultimate, and the elongation is not great. Plasticaccommodation and energy absorption must be located elsewhere in the structural linkage.

4. Because of edge, corner, or anchor spacing constraints, it may not be possible to expectductile failure no matter how deep the anchor is.

Whether or not the "ductile anchor" approach wins you over, the shaft failure calculationmust still be made as part of a "weakest link" analysis.

The effective cross-sectional area Ase of a threaded anchor runs to about 65% of thenominal cross-sectional area, more accurately given by the formula


where n1 is the number of threads per inch and d0 is the smallest nominal solid crosssection of the anchor. In a wedge-type expansion anchor, Ase is the diameter of the narrowend of the expansion cone.

Multiply Ase by the steel yield strength ¦y to obtain Ny, the nominal material tensile strengthof a single fastener in pounds, and by modifiers known as phi (f) factors, defined below.These factors are developed to produce adequate local safety conditions in normal use.

For steel that will fail in a normal ductile manner, f is set to 0.90. For instances where thesteel may fail in a brittle manner, f is set to 0.75.

Functionally, f factors are necessary to account for sensitivity to installation techniques,where reliability may be affected. Classes of sensitivities will be found in ASTM Z5818Z[3] and ASTM Z5819Z [4]:

Class 1 0.75Class 2 0.65Class 3 0.55


Generally, for cast-in-place anchors, the majority of torque-controlled expansion anchors,and most undercut anchors, f will be 0.75. Sleeve-type anchors and displacement-controlled expansion anchors may be Class 2 or Class 3, depending on proprietary design.

The final formula for the fastener tensile design strength fNy is

If there is no well-defined yield point for the material, use the ultimate strength ¦ut times0.8. Neither 0.8¦ut nor ¦y, should be taken to exceed 120,000 psi.

Metal shear failure: In a similar manner to the tensile calculation, the allowable shearstrength of the fastener in pounds is given by

where Vy is the nominal material yield shear strength of a single fastener. This formulaassumes that steel shear strength is 60% of yield strength.

As with calculations of the allowable tensile load, if there is no well-defined yield point forthe material, use the ultimate strength of the steel, ¦u, times 0.8. Again, neither this resultnor the material tensile strength should exceed 120,000 psi.

Because installation technique is not as critical in shear applications, set reduction factor f= 0.75 for all fasteners in shear.

2.2.3Seismic Applications

The above-mentioned ASTM standards also set out the simulated seismic tests forqualifying fasteners in seismic design zones. In moderate and high seismic risk zones, anadditional reduction factor of 0.75 is applied to the allowable loads, fNn and fVn,calculated above.

If anchors are not qualified under the simulated seismic tests of ASTM Z5818Z andZ5819Z, the attachment itself can be designed to yield in a ductile manner at a load nogreater than 75% of the anchor design strength.

2.2.4Pullout

The pullout failure mode is very dependent on the proprietary design of the anchor


manufacturer. If the anchor head is only slightly larger than its shaft, the head may pullthrough. Also, because expansion anchors rely primarily on friction and wedging, properinstallation is critical. This is especially true for displacement-controlled anchors, whichhave no follow-up expansion. For torque-controlled expansion anchors, the cone may bepulled through the wedges.

Comprehensive prequalification tests are required under the ASTM standards, since thesetypes of failures cannot be predicted theoretically. The loaddeflection characteristics shouldpreferably be linear, as with cast-in-place anchors, so that some predictability is provided.

Pullout tensile strength Np for a headed cast-in-place bolt may be obtained using theformula

where Ab is the net annular area of the bolt head or its attached rigid plate (Fig. 12), and l isthe coefficient for concrete composition.

If there is no potential for cracking from structural loading, a magnification f factor of 1.4can be applied to Np.

2.2.5TensileShear Interaction

For simplicity, a linear interaction is acceptable. For a more conservative solution, use a"trilinear" approach (Fig. 13). Neglect the interaction if the amount of either the shear orthe tension is less than 20% of capacity. If this is not the case, use the formula


Figure 12Net annular area of bolt head.

where Nn and Vn are the design load vectors, and fNn and fVn are the allowable pure tensileload and the allowable pure shear load.

2.2.6General Limits of Anchor Arrangement Geometry

In addition to the foregoing design considerations, the following limitations should beobserved:

Center-to-center spacing no less than 4 times the anchor diameter.

Minimum slab thickness 1.5 hef (preferably 2.0hef);

Minimum edge distance to avoid side blowout with deep embedment for cast-in-placeanchor,


Figure 13Trilinear tensileshear interaction.


where c is an edge distance < 0.4hef and center-to-center spacing is at least 6 timesthe anchor diameter; Nb and Ab are defined in Sections 2.2.1 and 2.2.4, respectively.

For postinstalled anchors, determine the minimum edge distance according to ASTMZ5818Z and Z58189Z; if not so determined, the edge distance must be at least 4 timesthe anchor diameter.

2.3Adhesive Anchors

An adhesive anchor should ideally be designed to exhibit simultaneous tensile failure inbond breakage and in shaft ultimate strength. The International Conference of BuildingOfficials (ICBO) Evaluation Service has published acceptance criteria for evaluationunder the Uniform Building Code of adhesive anchors in uncracked concrete and newmasonry [7]. Although current ACI standards do not address adhesive anchors, theselection and evaluation process should be similar to the one for a cast-in-place steelanchor. Extra care must be taken at installation to ensure use of the proper curingenvironment according to the manufacturer's specifications.

3Summary

Steps in anchor selection and design:

1. If designing cast-in-place or grouted anchor systems, go directly to the applicablebuilding code and work with the design approach outlined in this chapter if permitted.

2. If selecting a postinstalled system, and the application requires an agency approval or isone in which human life and safety are involved, contact the field engineering office of ananchor manufacturer to obtain information on the test data and approval informationrelating to their proprietary designs.

3. If only a typical application is involved, the manufacturer's published data may be used,and the approach set out in this chapter used for minor modifications.

When designing the anchor system, the design load must not exceed the smallest designvalue of

1. Anchor shaft breakage


2. Anchor pullout or pull-through

3. Concrete breakout, whether by stress block breakout or concrete splitting

Consider the following as well: seismic areas, importance of good installation procedures,combined loading, and local concrete aggregate characteristics.

Dynamic and impact loading and ductile or plastic design as applied to the attachmentsmay influence the anchor design or selection.

Site-specific testing of sample installed anchors and proof loading may also be required bythe project structural engineer or may be advisable in unusual conditions.

References

1. Fuchs, W., R. Eligehausen, and J. E. Breen, Concrete capacity design (CCD) approachfor fastening to concrete, ACI Struct. J., 92(1): 7393 (1995).

2. HILTI Anchor Program (HAP), Hilti, Inc., Tulsa, OK.

3. ASTM Z5818Z, Performance of Anchors in Cracked and Uncracked Concrete Elements,American Society for Testing and Materials, West Conshohocken, PA, Draft, March 1993.


4. ASTM Z5819Z, Performance of Anchors inUncracked Concrete Elements, American Societyfor Testing and Materials, West Conshohocken,PA, Draft, March 1993.

5. PCI Design Handbook, 4th ed.,Precast/Prestressed Concrete Institute, Chicago,1992.

6. Eligehausen, R., and T. Balogh, Behavior offasteners loaded in tension in cracked reinforcedconcrete, ACI Struct. J., 92(3): 365379 (1995).

7. ICBO, Acceptance Criteria for AdhesiveAnchors in Concrete and Masonry Elements,AC58, ICBO Evaluation Service, Inc. Whittier,CA, April 1995.


17Aerospace Bolts

Shahriar M. SadriHuck International, Inc., Carson, California

1Introduction

Hundreds of different bolt designs with varioussizes, strength levels, and materials are used inassembly of an aircraft. On the average, forexample, 2.4 million fasteners are used toassemble a Boeing 747 aircraft. Of this total,22% are structural bolts (mostly titanium) andthe rest are rivets.

Basic aircraft fastener types are

Solid rivets


Bolts

Threaded nut/collarSwaged collarUpset button

Recess drive screws/nutplatesBlind RivetsBlind bolts

Introduction to all of these fastening systems,although a very interesting exercise, would bevery extensive and beyond the scope of thischapter. An attempt is made, however, to coverthe most commonly used aerospace structuralbolts used in modern aircraft manufacturing.

Aerospace structural bolts in general areclassified under two categories:

1. Two-piece grooved pin/swage collar(lockbolt). This type of product, shown in Figure1 and its mating collar component, comes ineither a pull version, installed with a manual


pneudraulic or hydraulic tool, or a stumpversion, which is mostly suitable for automaticinstallations.

2. Two-piece threaded pin with threaded collar orlocknuts. As the name suggests, in this casefastening is done by engagement of the threadedcomponents (Fig. 2). Normally a hex recess isbuilt into the bolt at the bolt end (opposite thehead end). The parts are normally installed byusing manual pneumatic installation toolingequipped with the proper nose assembly. Thenose assembly consists of a stationary


Figure 1Pull-type and stump-type lockbolts.

hex key and a 12-point rotating outer socket. Theformer engages the pin recess while the latterengages and rotates the outer hex (or 12-point)portion of the nut.

2Structural Bolt Designs and Properties

2.1


Head Styles

The most common bolt head styles used inaircraft structures are shown in Figure 3. The100° flush head styles, for both the lockbolt typeand threaded type, are used for aerodynamicflushness in areas such as skin attachments.

Figure 2Threaded pin with hex or 12-point locknut.


Figure 3(a) 100° flush shear head

(b) 100° flush tension head

(c) protruding head.

2.2


Two-Piece Grooved PinSwage Collar System (Lockbolt)

During installation of a swage type fastener, a soft componentcalled the collar is swaged onto the annular grooves of a pin bythe action of an installation tool. Since the swage actioninvolves pure tensioncompression, as compared to the torqueaction in a threaded system, the final joint clamp variations areheld to an absolute minimum. There are a number of differentgroove forms used in conjunction with dedicated collars, eachintended for a specific application or design requirement. Ingeneral, the strength of a lockbolt is achieved through thecontrol of a few factors:

1. Groove design and number of annular grooves.

2. Collar packing: the amount of interference between theswage cavity, the collar, and the pin.

3. Pincollar material selection.

There are two methods used for installation of lockboltsystems, depending on whether the lockbolt is a pull type orstump type.

A pull type lockbolt has a tail portion, the pintail, which isbroken off upon full setting of the fastener assembly. Figure 4is a sketch of the installation sequence of a pull-type lockbolt.Today more than 60% of lockbolts are installed in this fashion.


Figure 4Installation sequence of a pull-type lockbolt.


Figure 5A modified drill riveter for automatic installation of lockbolts. ACFD = automatic collar feeding device.

(Photo courtesy of Huck International, Inc.)

The stump-type lockbolt does not have a pintail and is installed automatically by a modifieddrill riveter such as the one shown in Figure 5. The riveter will follow an automaticpreprogrammed sequence for installation of stump-type lockbolts. Typical steps taken duringan automatic installation cycle are as follows:

1. The workpiece is held in position and securely clamped.

2. The hole is drilled.

3. A collar is advanced to a position directly under the hole.

4. A bucking bar in the upper ram pushes a pin, delivered by the pin feeder, into the drilledhole and through the collar.

5. The lower ram moves upward, forcing the swage anvil over the collar, thereby swaging thecollar material into the grooves on the fastener.

6. The anvil is ejected from the collar, and the next cycle begins.

2.3Two-Piece Threaded System


Typical threaded pins are shown in Figure 6. Like the lockbolt systems, the head styles andthread lengths are selected to meet the application requirements for shear, tension, etc. Asketch of an installed threaded pin and nut is shown in Figure 7.

The pins are manufactured with a hexagonal drive recess. Generally, the pin threads areoptimized so that a certain minimum strength in a predefined grip range (normally 1/16 in.) isachieved with the smallest number of full thread engagement. On some of the high endsystems, in order to maximize the strength-to-weight ratio, the pin threads are rolled using aspecial technique that would minimize the thread runout (partial thread) portion.

The mating nuts for threaded systems are categorized into two groups: frangible andnonfrangible. Full setting and installation of a frangible nut and collar (Fig. 8) is achieved byshearing a hexagonal drive section of the nut at a certain torque established by a break groove.

A nonfrangible nut (Fig. 9) is installed by using a torque-controlled tool. The installation toolis calibrated and will shut off when a certain predetermined torque level is reached.


Figure 6Threaded pins.


Figure 7Installed threaded

pin.

Figure 8Frangible type nut

and collar.

Figure 9Twelve-point

nut.


TABLE 1 Lockbolt Materials and Finish OptionsLockbolt Collar

Pin material Pin finish Collar material Collar finish

Titanium alloy 6AL-4V, BETA-C Aluminum coating

Titanium alloy 3AL-2.5V, CPAluminum alloy2024

Al coating, dry filmlubricantAnodize

Aluminum alloy 7075 Anodize Aluminum alloy3000 series Anodize

Stainless steel A-286 Passivate, cadmiumplate

Stainless steel A-286Aluminum alloy2024

Dry film lubricant

Anodize

Alloy steelAISI 8740, 4037

Cadmium plate ornickel-cadmium

Stainless steel A-286 Dry film lubricant

Due to the high vibration environment of the application, the nutsused with the structural threaded bolts are required to have alocking feature. There are five methods commonly used forgeneration of locking torque for nut retention:

1. Deformed nut (oval, triangular crimp)

2. Deformed thread (localized crimp)

3. Installation swaging

4. Encapsulated adhesive (limited approval)

5. Lock wires or cotter pins (limited use)

3


Bolt Materials and Coatings

The most widely used materials in the manufacture of aerospacebolts are listed in Tables 1 and 2. In these tables, the bolt(lockbolt or threaded) is listed together with the most commonTABLE 2 Threaded Bolt Materials/Finish Options

Threaded bolt CollarPin material Pin finish Collar/nut

material Collar/nut finish

Titanium alloy 6AL-4V Aluminum coating

Stainless steelA-286, 300seriesAluminum alloy7075, 7050

Dry filmlubricant

Anodize

Stainless steel A-286 Passivate

Stainless steelA-286Aluminum alloy2024

Passivate

Anodize

Alloy steel AISI 8740,4037

Cadmium plate, nickel-cadmium

Stainless steelA-286 Passivate


collar or nut material. For example, bolts madefrom titanium alloy are used with collars madeof aluminum or titanium alloys. The mostcommon boltcollarnut finishes are also listed inTables 1 and 2.

4Applications

4.1Basic Applications

Most common aerospace structural bolts,lockbolt or threaded, are made from 6AL-4Vtitanium heat treated to 95 ksi shear strength andcoated with a low friction aluminum-pigmentedepoxy coating.

The basic applications (or needs) for structuralbolts are

1. Shear (predominate)


2. Tension

3. Fatigue (primarily wing)

4. Fuel tightness (wing)

5. High temperature

6. Corrosion control

4.2Aircraft Structural Materials and Bolt Selection

Mechanical fasteners are used in all areas of acommercial aircraft: primary, secondary, andnonstructural. Aluminum, in various alloyforms, is the predominant material used in themanufacture of commercial aircraft. Forexample, Table 3 lists the typical aluminumalloys used in commercial aircraft.

Many factors affect the selection of a specifictype of bolt for a specific area or application inan aircraft. Some of the more importantconsiderations are listed on the following page.


TABLE 3 Example of Aluminum Alloy Use onCommercial AircraftStructure Portion MaterialFuselage Skin 2024

Stringers 7075Wing panels Upper panel 7075

Upper stringers 7075Lower panel 2024Lower stringers 2024

Wing spars Web 7075Upper chord 7075Lower chord 2024Stiffeners 7075


Installed cost Part pullupRate Final clamp-upWeight Tooling requirementsShear, allowable Skin quality controlSkin thickness Edge margin controlAccess Fuel tightnessNoise Air leakageStructural fatigue performance Aerodynamic flushnessCorrosion

For the selection of primary fastening systems, thedesigner will look at all of the applicable factors andassign the appropriate weighting to them. Thecandidate systems are then compared, and a system isselected accordingly.

Bibliography

1. Huck International Inc., Design Guide for LockboltFasteners.

2. Huck International Inc., Aerospace Fastener Manual,1996.


3. Huck International Inc., Installation ToolingManual.

4. Huck International Inc., VERI-LITER FasteningSystem Product Catalog.

5. Huck International Inc., Fatigue Rated FasteningSystems, 795.

6.VOI-SHANR, Division of Fairchild Corp.,Aerospace Fastener Design Manual, 1975.


PART IVDESIGN AND ANALYSIS OF BOLTEDJOINTS

18VDI Joint Design Procedures

Alex W. HestonSPS Technologies, Cleveland, Ohio

1Introduction

VDI is the German society of engineers, VereinDeutscher Ingenieure. VDI 2230 is one of thissociety's published guidelines for bolted joints.The full title for this document has beentranslated into English as ''Systematic


Calculation of High Duty Bolted Joints: Jointswith One Cylindrical Bolt."

The first step in predicting how a bolted structurewill respond to external loads is to identify asingly bolted joint for analysis. Loads on thejoint are presumed to be known or to bedeterminable from statics or other forms ofanalysis. These loads are shared by the bolt andthe clamped (or jointed) material. The designer ispermitted to use any analytic method, includingfinite element or boundary element methods, orexperimental methods (strain gages, ultrasonics)to determine how the external loads aredistributed. Equations based on beam theory areprovided for the default method of analyzing thebehavior of the joint under loading. Multiboltedjoints are handled by discretizing the system intoseveral singly bolted joints and analyzing each ofthem separately.

2


Units

International (SI) units are used in VDI 2230 [1],as it is a European document, and are usedexclusively here. There is nothing inherent to theanalytic procedures to preclude the use ofEnglish units. Table 1 shows the mostconvenient metric and English units to use formost analyses. Note that the correlations givenlater in Eqs. (15), (25), and (26) must becalculated using metric units.

3Modes of Failure

The externally loaded joint is systematicallychecked against a list of criteria for failure. Thefive primary modes of failure that should beexamined for all critical joints are

1. Insufficient preload

2. Tensile overload of the bolt


TABLE 1 Preferred Units for Use in Joint AnalysisFeature Metric unit English unit Multiplying factor for value in English unitsLength mm inch 25.40Force N lb 4.448Stiffness N/mm lb/in. 0.175Compliancemm/N in./lb 5.710Stress N/mm2 (MPa) psi 6895

3. Fatigue failure of the bolt

4. Yielding of the joint material under the bolt head or nut

5. Vibration loosening of the bolt

Guidelines are also given to guard against other design-specific modes of failure such as thread stripping resultingfrom insufficient thread engagement and head failures insocket head cap screws resulting from insufficientunderbroach height.

4Identifying the Bolted Joint (Characteristic Diameter)

When a bolt is tightened, the clamped material compresses,with the highest compression levels generally occurringnear the bolt hole. The joint faces will remain in contactwithin a fixed zone around the bolt until either the preload


is removed (preload loss or untightening) or external loadsof sufficiently high levels are applied. Classical jointanalysis generally isolates some axisymmetric section(usually conical or cylindrical) of the clamped materialaround the bolt for the purpose of making a stiffnessestimate. The bolt and this clamped material make up thebolted joint. VDI 2230 presumes that the designer canreadily identify the bolted joint and provides little in theway of guidelines for doing so.

Generally, a conservative approach for determining thejoint diameter is taken by using the smallest characteristiccircle for the joint. Figure 1 shows a flange connection andthree characteristic circles concentric to the bolt hole. Onecircle is tangent to the flange inner diameter, the second istangent to the flange outer diameter, and the third is tangentto a circle of the same diameter around the next nearest bolthole on the bolt circle.

A basic assumption used for the joint stiffness is that thecompressed region spreads out from under the bolt head,from the nut face, or from the mating threads at no morethan a 45° angle. This leads to a second criterion for thesmallest characteristic circle. Both criteria must beexamined to determine the smallest characteristic circlediameter.


5Estimating Loads on the Bolted Joint

VDI 2230 presumes that the designer can determine theapplied loads for the joint. Typical loading is shown inFigure 2.

There is an applied axial force FA, an applied shear forceFQ, and an applied bending moment MB. Most of thetreatment is for the special case when MB = 0. Even for thisspecial case, there may still be a bending moment Mbintroduced into the analysis because the lines of action ofthe axial force and bolt clamp load are offset from thecentroid of the joint face.

In some simple joints, the bending moment introduced byeccentric loading is statically determinant and can bedetermined from the axial load and local geometry. In manytypes of


Figure 1Characteristic circles for the joint diameter.

joints, the loading across the joint face is statically indeterminate. In the latter case, energymethods must be used to determine the bending moment and shear load associated with theaxial loading. Energy method solutions for many types of geometries are published instandard handbooks. A finite element model or experimental test program may be the onlyrecourse for some geometries. Figure 3 illustrates both statically determinate and staticallyindeterminate loading conditions. The righthand side of Figure 3 illustrates how bending isintroduced into a conrod (connecting rod) joint.

Grotewhol [2] used energy methods to determine the ratio of the effective moment arm tothe ring radius for a number of rings representing conrod geometries. a is used for themoment arm, and r0 for the distance to the joint centroid. (See Table 2.)

Hagiwara [3] conducted additional research on the statically indeterminate conrodgeometry.

Sometimes estimation of the external loads themselves is not straightforward. McNeill [4]developed and published a procedure for estimating the separating loads in a conrod basedon the geometry, mass, and speed of the components.

where


MREC =reciprocating massMROT =rotating massMCAP =mass of the capMBLOT

=mas of one bolt

MSHELL=mass of one shell

w =engine speed in radians per second (1 rpm =0.1047 rad/s)


Figure 2Free-body diagram for a joint.

Figure 3Statically determinate and indeterminate joints.


TABLE 2 Grotewhol Ring SolutionsCondition Ratio a/roThin uniform ring 0.36Bottom half twice as rigid 0.28Bottom relatively rigid 0.16Bottom ring rigid between 225° and 315° 0.16Actual conrod example 0.20

L=

center-to-center distance from the crankshaft bore to the piston ringbore

r=

crankshaft throw, i.e., eccentricity of crankpin axis to crankshaft axis.Half this load goes to each joint.

Often the external load can be determined more readily. In a cylinder head joint,adjacent cylinders do not fire simultaneously. Figure 4 shows a simplified cylinderhead joint. Four bolts surround each cylinder of diameter DBORE, and the combustionpressure is P0. The external load is shared equally by all four bolted joints. The loadper joint is given as

6Axially Loaded (Concentrically Loaded) Joints

6.1Springs in Series and Springs in Parallel

The classical theory of bolted concentrically loaded joints is based on a model of twolinear springs representing the bolt and the clamped material. The springs may be inparallel, in series, or partially in parallel and partially in series, as discussed later inSection 6.6.

Figures 5a and 5b illustrate the concepts of parallel and serial springs. Parallel springsstretch (or contract) by equal amounts, whereas serial springs carry equal loads.Parallel spring models are used to show how external load is shared by the bolt and theclamped material. Serial spring models are used to show how clamp load is affectedby relative axial movement between the bolt and the clamped materials at theinterfaces such as occurs during tightening.


Figure 5c shows a more generalized model where the spring representing the jointmaterial is subdivided into three springs. The middle one of these is in parallel withthe spring

Figure 4Axial loading in a cylinder head joint.


Figure 5Spring models. (a) Parallel

(b) series


(c) combination.

representing the bolt, and the other two are in series with the middle spring. This model isdiscussed in more detail in Section 6.6.

When a load FA is applied to two springs in parallel, the portion of the load carried by eachspring is directly proportional to its stiffness, K. The subscripts B and J correspond to boltand clamped (joint) material.

Similarly, when springs in series are stretched, the portion of the elongation UA carried byeach spring is directly proportional to the inverse of its stiffness, which is known as itscompliance, d = 1/K.

6.2Estimation of the Bolt Compliance dB

The bolt is modeled as a set of springs in series. The compliance of each section iscalculated, and the sum of the compliances computed. Only the portion of the bolt withinthe grip length is included in those calculations. The grip length lk used for thiscomputation is the distance from the first engaged thread to the underhead of the bolt (orscrew) or to the other first engaged thread for a stud. Figure 6 shows a bolt passingthrough a 28 mm thick top plate and a 4 mm counterbore in the bottom plate beforeengaging in the tapped threads in the lower plate. The grip length and stiffness of the boltdepend not only on its dimensions but also on the dimensions of the joint, including the


thicknesses of any washers used.

The compliance d and stiffness K of a cylinder of diameter d and length l made of amaterial with a Young's modulus E are given by the equation


Figure 6Bolt in joint for sample bolt compliance calculation.

The general formula for the compliance of a bolt or screw is given in the equation

where

The factors 0.4 and 0.5 for the compliances of the head, dK, engaged external threads, dG,and engaged internal threads, dM, are empirical values for socket and hexagonal head boltsof nominal diameter dn in steel joints. dMINOR in Eq. (7) denotes the thread minordiameter. An additional dM would be added for a stud in which both threaded ends are


stretched during


tightening. The summation in Eq. (6) is over the N cylindrical sections of the fastenerwithin the grip length (N can be 1 for a thread-to-thread part), with i being the dummyindex.

Table 3 presents an example of compliance calculations.

6.3Estimation of the Compliance of the Clamped Parts, dJ

The determination of the characteristic joint diameter DA in Section 4 plays an importantrole in determining the corresponding axial compliance. An equivalent area, Aers, is definedso that a cylinder of length lk, with an area Aers and material properties corresponding tothose of the clamped parts also has the same compliance as the clamped parts.

The compressive region of the clamped material spreads radially from the bearing face ofthe bolt head or nut to the center of the joint at an assumed limiting angle of 45°. Thecharacteristic joint diameter, DA, is limited accordingly: dw < DA < lk + dw. dw is taken asthe nominal flange or underhead bearing diameter for the bolt or nut. In the event that thereare two possible choices, such as a bearing diameter for a bolt and a bearing diameter for anut, dw should be the smaller of the two.

The calculation sequence is shown in the following two equations, which are evaluated inTable 4 for the joint shown in Figure 7.

where x is a parameter used to interpolate between clamped shapes corresponding to acylindrical sleeve and a 45° cone.

An equivalent average elastic modulus can be estimated when more than one material isclamped.


6.4Assembly and Settling Losses

When a fastener is tightened, the loadelongation curve is initially nonlinear because themating surfaces are not fully in contact. Other sources of nonlinearity may be seating ofgaskets, bearing shells, and nonparallelism. This nonlinearity is reflected in the preloadangle curve. In many joints, the preload angle curve becomes linear and remains so untilyielding takes place in the bolt or clamped parts. During this process, the bolt stretches andthe joint material compresses. An idealized linear model is shown in Figure 8.

The bolt and clamped parts are seen to be acting like springs in series because the relativeelongation, UV, is the sum of the stretch of the bolt and the compression of the clampedparts. The clamp load in the bolt and in the clamped parts is the same.

The stiffnesses, KB and KJ, are in units of newtons per millimeter, and the displacements,UB


TABLE 3 Compliance Calculations for Steel Bolt in Figure 6Bolt part Length used (mm) Diameter used

(mm)Compliance

(mm/N)Head 0.4dn = 4 10 2.462 × 107Shank l1 = 22 9 16.718 × 107Exposed threads l2 = 10 8.159 9.246 × 107Engaged threads 0.4dn = 4 10 2.462 × 107Mating threads 0.5dn = 5 8.159 4.623 × 107

KB = 1/dB = 280,200N/mm; dB = 35.511 × 107 mm/N

and UZ, are be in millimeters. As the compliances, dB and dJ, are the reciprocals of thestiffnesses, their values are in millimeters per newton.

The bolt threads turn 360° with respect to ground when they have advanced one pitch, p(mm), with respect to the mating internal threads. The relationship between the relativeaxial displacement between the external and internal threads, UV, and the angle q throughwhich the bolt threads turn with respect to ground is given by the equation

Substituting Eq. (13) into Eq. (12) and making use of the fact that the product of the boltstiffness and bolt compliance is 1 (KBdB = 1) leads to the following relationship betweenclamp load and tightening angle:

Normally, some settling takes place in a bolted joint over time and results in a loss in clampload. This amount may be determined by theoretical or experimental means. If no otherinformation is available, the following VDI 2230 empirical correlation based on griplength lK and nominal thread diameter dn may be used to determine the settling inmillimeters, UZ. Note that lk and dn must be in the same units (mm).

When settling occurs, the combined physical settling of the bolt and clamped parts, UZ,leads to an identical drop in the bolt clamp load and the joint compressive load, FZ. This isanalogous to F = FZ in the series spring model of Figure 5b, and so Eq. (4) applies.


TABLE 4 Compliance and Stiffness Calculations for Figure 7Data DA = 26 mm dw = 20 mm dh = 11 mm lk = 32 mmParameter x = 0.982 Aers = 357.1 dJ = 4.332 × 107 KJ = 2.308 × 106Units mm2 mm/N N/mmEquation (8) (9) (10) (5)


Figure 7Bolt in joint for sample joint compliance calculation.

6.5Thermal Loading and Expansion

The difference between the elongation of the clamped material and the elongation of theportion of the bolt within the grip length resulting from temperature variation iscalculated, and then the rule for springs in series is applied to this differential elongation todetermine the corresponding change in clamp load.

If the bolt material has a coefficient of thermal expansion xB, the clamped material has acoefficient of thermal expansion xJ, and both materials are heated (or cooled) from roomtemperature to a higher (or lower) operating temperature, the resulting change in clampload is calculated according to the equation


Figure 8Tent diagram showing preload and settling losses.

Heating of the bolted joint is not always uniform, in which case average values may have tobe used. Steady-state temperatures for the bolt and clamped material may also differ. Anaverage coefficient of thermal expansion can be determined using the approach of Eq. (11)when more than one material is clamped.

6.6The Loading Plane Factor, n

The applied axial load is generally not introduced at the interfaces between the bolt and theclamped material. Figure 9 shows an idealized case of load being applied at two distinctpoints within the clamped material. The bolt is seen to be a spring acting in series withsprings corresponding to the portions of the clamped material above and below the pointswhere the load is applied. This combined spring is in parallel with a spring corresponding tothe portion of the clamped material between the points where the load is applied. The loadin the bolt spring, FB, resulting from an applied load F is calculated using a formula similarto that for parallel springs, Eq. (3). Equation (18) has an additional cofactor, KJ/KJ2 = n. nis referred to as the loading plane factor. Note that 0 < n < 1.

For joints with relatively little clamped material extending beyond the bearing diameter dw,the ratio KJ/KJ2 is approximately the ratio of the length of the portion of the joint materialbetween the points of applied loading to the grip length. This can be recognized easily bynoting that Aers has little sensitivity to lk when DA» dw because the second term on the rightin Eq. (9) is then approximately zero.


Cylinder head joints are often assumed to have a loading plane factor of zero. In this case,none of the combustion load is carried by the cylinder head bolts.

6.7The Force Ratio f

The force ratio f is defined as the portion of the external load carried by the bolt.Mathematically, f is the computed cofactor for F in Eq. (18):


Figure 9Illustration of the loading plane factor n.

If the applied load does not overcome the preload, then the portion of the applied load FAcarried by the bolt is fFA and the clamp load at the joint face is reduced by (1 f)FA.

VDI 2230 uses several different subscripts for f to indicate that the force ratio accountsfor eccentricity, loading plane factors, and/or bending moments. This approach is not usedhere; instead, the different formulas for f are kept distinct by subsection.

6.8Determining Preload Requirements

The minimum possible preload developed by the assembly specification should leavesufficient clamp load to keep the joint closed after settling losses, thermal losses, andother miscellaneous losses have occurred. In the case of gasketed joints, keeping the jointclosed may translate into maintaining a minimal average sealing pressure over a region. Ifthe joint also experiences shear loading, then the remaining clamp load must maintainsufficient friction across the joint faces to prevent one from sliding past the other.

FKR is used here to denote any special force required for sealing, crushing (or seating) ofbearing shells into conrods or main bearings, etc. FKR is zero unless some of the boltclamp load is required to compress an element in parallel with the clamped material or thenominal compression across the joint face must always exceed zero for functional reasons


such as sealing. The minimum preload required to withstand the axial loads is calculatedas follows.


where

FZ=settling loss

f =force ratio (number between 0 and 1 representing the fraction ofapplied axial load carried by the bolt)

FA=applied axial force

In cases of random or cyclical loading, the value used for FA should correspond to themaximum possible axial load.

A similar calculation is made to determine the preload required to prevent sliding of thejoint faces caused by a shear force FQ.

mJ should be the minimum established coefficient of friction between the joint faces. Thevalue of FA used in Eq. (21) should correspond to the axial load at the time the shear forceFQ is applied. It will differ in value from the value of FA used in Eq. (20) if the shearforce, usually the maximum shear force, does not occur at the same time as the maximumaxial load. In cases of random or cyclical loading, it may be necessary to evaluate Eq. (21)under more than one set of loads to determine the most severe condition. FZ is again thesettling loss.

The minimum required preload is the larger of the two values of Pmin calculated from Eqs.(20) and (21). If thermal loading can lead to a decrease in clamp load, that drop should beaccounted for by including an additional term in Eqs. (20) and (21).

The maximum allowable preload is limited by the bolt strength, the thread strength, or thecompressive strength under the bolt head or nut face. The maximum allowable preload isestimated assuming that no settling takes place.

The value for FMAX used in Eq. (22) should be the smallest of the three following values:the bolt yield strength, the stripping strength of the mating threads, and the embedmentlimit of the clamped material.

If the bolt is yield-tightened by a torque plus angle or gradient strategy, it will work harden,


as illustrated in Figure 10. In my opinion, the value used for the tensile yield strength maybe increased to a level a few percent higher than the maximum level specified on the boltprint. Ideally, tensile testing of yield-tightened parts will be used as a basis.

Generally, the bolt material is at least as strong as, if not stronger than, the material usedfor the internal threads. So thread stripping of the bolt material is usually not a concern. Acheck should be made to ensure that there is sufficient thread engagement for the internalthreads to carry the full tensile strength of the bolt, preferably without significant yielding.Equation (23) gives the stripping strength of internal threads.

where

tU =ultimate shear strength of the mating material, about5060% of its ultimate tensile strength

P =thread pitchLENGAGED

=thread engagement length

DSMIN =minimum major diameter of the bolt threadENMAX =maximum pitch diameter of the mating thread

Values for DSMIN and ENMAX may be found in most reference books on threadedfasteners. The


Figure 10Increase in yield load from work hardening.

minimum value of LENGAGED based on the bolt print and the print of the part with themating threads should be used in Eq. (23).

Example: Determine the required 6H thread engagement in 310 aluminum for a 10.9metric strength level M10 × 1.56g bolt. The maximum bolt strength for a 10.9 metricstrength level bolt may be taken as the nominal strength of a 12.9 metric strength levelbolt, 1200 MPa. The tensile stress area of the threads is 58 mm2. The maximum bolt loadis 1200 × 58 or 69,600 N. The ultimate shear strength of the aluminum is 130 MPa.Tabulated thread specifications show DSMIN = 9.968 mm and ENMAX = 9.206 mm.Substituting 69,600 for FSTRIP in Eq. (23) and using the data provided here leads toLENGAGED = 21.6 mm.

Tests with hex nuts have shown that dilation may reduce the stripping strength of internalthreads if there is insufficient wall thickness. The data have been correlated in terms of theratio of the across-flats dimension w and the nominal thread diameter dn. The followingequation is applicable for w/dn ratios between 1.4 and 1.9

This factor, which is between 0.75 and 1.0, is multiplied by the stripping strength


calculated in Eq. (23) to calculate the reduced level for parts with thinner walls.

The bolt head and nut face, if any, bear on clamped material. If the clamp load is too high,the bearing surfaces will embed in the clamped materials, resulting in a loss in clamp load.The limiting clamp load is the product of the clamped material yield stress and the bearingarea.

The underhead bearing area in Figure 11 is an annulus. The area is


Figure 11Definition of bearing area.


The underhead bearing area for football head andD-head bolts may be somewhat more difficult toevaluate. If a nut is used, the bearing area for thenut face must also be calculated. Work hardeningof the clamped material during tightening andconstraint by the unloaded material around itmay increase the apparent yield strength abovepublished levels. If operating conditions lead toan increase in clamp load resulting from thermalloading, then the right-hand side of Eq. (22)should be reduced by that amount.

The range of preload from PMIN to PMAXdictates what type of tightening strategy issuitable. The tightening strategy will produce amaximum preload of FSP and a minimumpreload of FSP/aA, where the tightening factor aAis about 1.0 for strain gage or ultrasonicmeasurements, 1.22 for yield tightening, andanywhere from 1.4 to 2.7 depending on coatingsand equipment for other less accurate elastic-range tightening strategies such as torque


control, torque plus angle control, and hydraulictensioning.

6.9Fatigue Analysis

VDI 2230 uses the approach that the appliedfatigue cycle on the bolt can be characterized by acalculated applied alternating stress. Thisalternating stress is compared to the endurancelimit for the bolt, which is the maximumalternating stress under comparable mean stressconditions that will not cause the bolt to fail.


The alternating force is one-half the difference between the maximum and minimumapplied forces. The alternating stress is the alternating force divided by the thread diameterminor area, AMINOR. The mean or average force is simply the average of the minimum andmaximum applied axial forces. The mean stress is the mean force divided by the threadtensile stress area. (The traditional use of thread (TSA) for calculating the mean stress andof thread minor diameter area for calculating the alternating stress has been a cause ofunresolved contention.)

The alternating stress on a bolt should not exceed its endurance limit. Figure 12 can beused to estimate the endurance limit of an alloy steel bolt, metric strength class of 8.812.9,with rolled threads.

The endurance limit sA for alloy steel bolts with threads rolled before heat treatment(RTBHT) is relatively independent of the mean load over a large range. This is a somewhatunexpected trend, but a theory has been proposed by Yoshimoto [5] to explain it. sA inmegapascals (MPa = N/mm2) is correlated solely in terms of the nominal thread diameterdn in millimeters and does not depend on the metric strength level.

The endurance limit of alloy steel bolts with threads rolled after heat treatment (RTAHT)is normally somewhat higher and depends on the ratio of the mean load Fm to the 0.2%yield load F0.2 over most of the range.


Figure 12Fatigue diagramendurance limit vs. bolt nominal diameter.


Equations (25) and (26) are adequate for mean stresses ranging from 20% to 90% of the0.2% yield stress, based on the tensile stress area of the threads. Note that Eq. (26) usesthe ratio of the mean force to the yield strength, which eliminates any ambiguity regardingthe appropriate stress area for calculating the mean stress.

Example: Predict the endurance limit of a 10.9 M10 bolt with threads rolled after heattreatment when the minimum load is 10% of the maximum load FMAX. The mean load is(FMAX + 0.1FMAX)/2 = 0.55FMAX, and the alternating load is (FMAX 0.1FMAX)/2 =0.45FMAX. Substituting these expressions into Eq. (26) leads to the equation

where

AMINOR=

thread minor diameterarea

dn =thread nominaldiameter

F0.2 =yield strength

Using dn = 10 mm, s0.2 = 940 MPa, TSA = 58 mm2, F0.2 = 940 × 58 = 54.52 kN, andAMINOR = 52.2 mm2 leads to a value for FMAX of 11,480 N. The endurance limit underthis condition is then 11480/52.2 = 220 MPa.

Bolts may be designed such that the shank diameter is sufficiently necked down to thepoint where fatigue failures will occur in the shank rather than in the threads. Fatiguefailures may also initiate within knurled sections, in the bolt underhead fillet, and throughother regions of geometric transition. Figure 12 overestimates the endurance limit forthese cases.

The endurance limit may also be increased by using larger radius thread roots, i.e., an MJprofile, using specialty materials such as MP159, and by various methods of adjusting theload distribution between mating threads. In contrast, the endurance limits of bolts withcut threads are lower than those for corresponding parts with rolled threads.


For these reasons, experimental determination of the endurance limit is preferable to theuse of Figure 12 when time and resources allow.

7Eccentrically and Nonlinearly Loaded Joints

The primary difference between the analytic methods for concentric joints and eccentricjoints is that beam bending is also taken into account. Generally, the bending stiffness ofthe clamped parts is much greater than that of the bolt. Thus, a simplifying assumption canbe made that the curvature of the loaded joint is a function of its geometry and the bendingmoment. This bending moment is the sum of the external bending moment applied to thejoint, MB, the bending moment caused by the applied axial load, Mb, and the bendingmoment produced by the bolt because its axis is offset from the centroid of the joint face.Closed-form solutions can be derived for these conditions.

The same approach can be used for nonlinear joints, but graphical methods or iterativecomputer techniques are necessary to arrive at a solution.

The behavior of the joint generally becomes nonlinear if sufficient loads are applied tocause localized separation of the clamped material at the joint face. While VDI 2230 doesprovide some equations for analysis of this condition, the primary recommendation is thatthere be sufficient preload to prevent the condition from occurring.


7.1Geometric Properties of the Joint Face

The section modulus is used in beam bendingproblems for determining the maximum tensilestress in the body. The same approach is used inVDI 2230. The process of drilling the bolt holethrough the joint face shifts its centroid, and themoment of inertia is calculated using the parallelaxis theorem. Figure 13 shows a generic jointdrawn as a rectangle for expediency. CG denotesthe centroidal axis.

The bolt hole may be offset from the centroid ofthe joint face. s is used for this dimension, and theconvention is to use a positive sign for s when thebolt preload counteracts the applied moment.Generally, the axial load is tensile, and s is takenas positive when the bolt axis is on the same sideof the centroid as the line of action of the appliedforce.


The following geometric characteristics must bedetermined.

a Equivalent moment arm of the applied force(mm)

u Distance from the centroid to the separableside (mm)

v Distance from the centroid to the opposite side(mm)

s Vector (size + sign) from centroid to bolt axis(mm)

AB Planar area of the predrilled joint face (mm2)ADPlanar area of the drilled joint face (mm2)

IB Moment of inertia of the predrilled joint face(mm4)

IBTMoment of inertia of the drilled joint face(mm4)

The values of AB and IB are not used directly butonly in intermediate steps for calculating AD andIBT.


The moments of inertia are taken about centroidalaxes that are parallel to the axis of the appliedmoment.


Figure 13Definition of joint face geometric parameters.


u is the distance from the centroid of the joint to the edge of the joint that will separate ifthe joint is overloaded. v is the distance from the centroid of the joint to the opposite end.The value of v is used only for prying joint calculations, which are not covered here.

If the applied axial force in Figure 13 were in the opposite direction (compressive), thenthe values used for u and v would be interchanged, and the sign of s would be negative.

7.2External LoadingConrod Example

A top view of the two joints is shown in Figure 14.

7.3Vibration Loosening

VDI 2230 provides no quantitative calculations relating to vibration loosening. Thegeneral rules for eliminating vibration are long bolts, i.e., high lk/dn, ratios, and highpreload. If a locking feature is needed, the function must be properly understood.

Serrated bearing faces on the nut or bolt and adhesives are the only tabulated functionalantirotational locking features. Deformed bolt threads, plastic patches or inserts, all-metalprevailing torque nuts, castellated nuts, tab washers, and cotter pins (wire) are all listed as


being primarily functional for preventing loosening following clamp load loss caused byembedment.

Figure 14Geometry of joint face geometry for sample calculations.


7.4The Force Ratio f

The bending compliance b of a body is calculated using a formula similar to Eq. (5) withthe moment of inertia used in the denominator rather than the area.

As l is in mm, Young's modulus E is in N/mm2 (MPa), and I is in mm4, b is in units of (N ·mm)1.

If the joint cross section is fairly constant, the value of IBT calculated for the joint face maybe substituted into Eq. (28) to determine the bending stiffness of the clamped material. Inthis case, the moment of inertia for the joint face is also the equivalent moment of inertiafor the clamped material, IBers (IBT = IBers).

When the cross section of the joint varies through the grip length, the equivalent momentof inertia is calculated by using an averaging scheme.

The following equation is the generalized formula for the force ratio. Note that VDI 2230does not explicitly account for the hole in the moment of inertia, IBers for this equation.This was done in earlier revisions, however, and would appear to more consistent with theoverall beam bending formulation.

n again denotes the loading plane discussed in Section 6.6 and should be a value between 0and 1. The ratio of the equivalent area for the clamped parts, Aers, to the equivalentmoment of inertia of the clamped parts, IBers, appears in both the numerator and thedenominator of Eq. (30). VDI 2230 makes a point of relating this ratio to the radius ofgyration [(IBers/Aers)0.5] for the clamped parts. Note that Eq. (30) reduces to Eq. (19) whenthere is no geometric eccentricity, i.e., s = 0.

7.5Determining Preload Requirements

There is a certain amount of preload, FKerf, required to keep the joint from one-sided


opening when the axial force FA is applied.

This term is added to the others in Eq. (20).

Equation (21) remains unchanged and its calculated value compared with that from Eq.(32) to pick the larger PMIN as the minimum required preload. The limitations for themaximum allowable preload discussed in Section 6.8 are still applicable.

7.6Fatigue Analysis for Combined Axial and Bending Loads

Fatigue data for combined tensile and bending stresses are almost always unavailable andare difficult to generate. The calculated largest alternating combined bending and tensilestress at


the thread root, or other weakest section, is calculated and compared to the endurance limitfor the bolt for the same mean load.

If MB is constant (zero is a constant) and the bending stiffness of the bolt is much less thanthat of the joint, then the maximum alternating stress at the thread root corresponding to arange of axial loads FAU and FAO can be shown to have the form

VDI 2230 explicitly accounts for the drilled hole in the calculation of IBers used for Eq. (33)and uses a bar over the symbol to distinguish this moment of inertia from the moment ofinertia for a predrilled body.

lers is the equivalent bending length of the bolt based on the thread minor diameter.Procedures analogous to those used for determining the bolt axial compliance in Section 6.2are used for determining its bending compliance. lers is the length that would be substitutedinto Eq. (28) with the moment of inertia corresponding to the thread minor diameter toarrive at the same bending stiffness or compliance as that of the bolt.

The factor 0.8 represents the sum of two factors of 0.4.

As discussed in Section 6.2, the factors 0.4 and 0.5 correspond to empirical values for hexhead and socket head bolts and represent flexure in the head and mating threads. A second0.5 term would be included for a stud in which both threaded ends are stretched duringtightening.

The summation in Eq. (34) is over the N cylindrical sections of the fastener within the griplength (N can be 1 for a thread-to-head part). i is a dummy index for the summation.

8Calculation of the Factors of Safety

Factors of safety are calculated as ratios of acceptable levels of loading to calculated levelsof loading.


where

FSP=target clamp load

a =preload scatter factor representng the ratio of the maximum preloaddeveloped by the tightening strategy to the minimum preloaddeveloped by tightening stratety

PMIN=calculated minmum required preload for a functional joint


TABLE 5 Table for Estimating the Appropriate Bolt SizeBolt size for corresponding

strength levelLoad (KN) 12.9 10.9 8.8

0.250.400.631.01.6 3 3 32.5 3 3 44.0 4 4 56.3 4 5 5

10.0 5 6 816.0 6 8 825.0 8 10 1040.0 10 12 1463.0 12 14 16100.0 16 16 20160.0 20 20 24250.0 24 27 30400.0 30 36630.0 36

Again, the denominator in Eqs. (37) and (39) may have to beincreased to reflect thermal loading. The numerator in Eq. (37)may also be increased based on experimental data relating to workhardening of the corresponding material.

9Fastener Sizing Guidelines

The entire process of designing a joint and calculating its safety


factors may have to be repeated more than once to arrive at asatisfactory design. Table 5 provides a basis for initial sizing of thefastener.

To use Table 5:

1. Pick the row with the lowest load that is greater than or equal tothe applied load.

2. Estimate the minimum required preload.

a. Move four rows down for shear loading.

b. Move two rows down for eccentric fatigue loading.

c. Move one row down for eccentric static or concentricfatigue loading.

3. Estimate the largest required preload.

a. Move two rows down for straight torque tightening.

b. Move one row down for elastic-range tightening usingelongation measurements or torque plus angle monitoring ora calibrated torque wrench.

4. Pick the 8.8, 10.9, or 12.9 bolt size on that row.

Example: What size 10.9 bolt should be used for a shear load of 9kN? Pick 10 kN as next biggest load in the leftmost column.Increase by 4 steps for shear to 63 kN. Increase by 1 step for atorque wrench to 100 kN. M16 would be used for 10.9 (or 12.9).


List of Symbols

a Effective moment arm for axial force FA; Mb = aFA (mm)AB Gross area of joint face before bolt hole is drilled (mm2)AD Net area of joint face after bolt hole is drilled (mm2)Aers Equivalent area for clamped material (mm2)Ai Cross-sectional area of ith section (mm2)AH Hole areaAMINOR Cross-sectional area of thread minor diameter (mm2)

d Generic diameter (mm) disthread nominal diameter (10 for M10,12 for M12, etc.)

dh Hole diameter (mm)dMINOR Thread minor diameter (mm)dn Thread nominal diameter (mm)dw Bearing or washer diameter (mm)DA Characteristic joint diameter (mm)DSMIN Minimum external thread major diameter (mm)E Generic elastic (Young's) modulus (MPa; N/mm2)EB Elastic modulus of bolt material (N/mm2)Ei Elastic modulus of ith section of clamped material (N/mm2)EJ Elastic modulus of clamped material (N/mm2)ENMAX Maximum pitch diameter of internal thread (mm)F Generic axial force (N)F0.2 Yield strength of bolt based on 0.2% yield stress (N)FA Applied axial force (N)BB Force in bolt (spring) arising from applied axial force (N)

FJ Reduction in compressive force between clamped parts (spring)caused by the applied axial force (N)

FKR Additional force requirement for seating, sealing, etc. (N)FKerf Minimum clampload to prevent any separation of the joint face (N)FM Mean axial bolt load in fatigue cycle (N)FQ Applied shear force (N)


FSTRIP Axial load that will strip internal threads (N)FV Preload or clampload (N)i Dummy summation indexI Generic moment of inertia (mm4)IB Moment of inertia of predrilled joint face (mm4)IBers Effective moment of inertia of clamped parts (mm4)IBT Moment of inertia of joint face with drilled bolt hole (mm4)K Generic axial stiffness (N/mm)KB Axial stiffness of bolt (spring) (N/mm)KJ Axial stiffness of clamped parts (spring) (N/mm)l Generic length (mm)

lers Equivalent bending length of bolt within clamped length based onthread minor diameter (mm)

lk Grip length between bearing surface and first engaged thread oropposite bearing surface for studs (mm)

LENGAGEDLength of mating external and internal threads (mm)MB Applied pure bending moment (N · mm)Mb Bending moment resulting from eccentricity of FA (N · mm)n Loading plane factor (0 < n < 1)N Number of sections


p Thread pitch (mm)PMIN Minimum required preload for a functional joint (N)PMAX Minimum required preload for a functional joint (N)s Distance from bolt axis to centroid of joint face (mm)T Temperature (°C)TROOM Room temperatureTOPERATINGOperating temperature (°C)

TSATensile stress area. Tabulated values can be found for standardthreads. For metric threads (p/4)(d 0.9382p)2. For Englishthreads (p/4)(d 0.9743p)2.

u Distance from centroid of joint face to portion of clamped partsthat will separate first when overloaded (mm)

UV Initial combined bolt stretch and compression of clampedmaterial developed when bolt is preloaded (mm)

UA Applied axial displacement (elongation) to joint (mm)UB Elongation of bolt (spring) resulting from axial load (mm)UJ Decompression of clamped parts (spring) from axial load (mm)

v Distance from centroid of joint face to edge in opposite directionto that used to measure u (mm)

w Width across flats for hexagonal nuts (mm)x Interpolating parameter in Eqs. (7) and (8)aA Ratio of maximum to minimum preload developed in assemblyb Generic bending compliance [(N · mm)1]d Generic axial compliance (d = 1/K) (mm/N)dB Axial compliance of bolt (spring) (mm/N)dG Axial compliance of engaged external threads (mm/N)dJ Axial compliance of bolt (spring) (mm/N)dM Axial compliance of bolt head (mm/N)dM Axial compliance of engaged internal threads (mm/N)q Angle bolt turns with respect to internal threads (degrees)

mJ Coefficient of friction between clamped materials at the joint


mJface

xB Coefficient of thermal expansion for bolt material [(°C)1]xJ Coefficient of thermal expansion for clamped material [°C)1]s0.2 Axial yield stress based on 0.2% offset (N/mm2)sA Endurance limit of bolt for fatigue loading (N/mm2)sa Peak alternating stress in bolt thread root (N/mm2)tU Ultimate shear strength of internal thread material (N/mm2)f Force ratio, fraction of external axial load the bolt carries

References

1. VDI 2230 B1.1, Systematische Berechung von hochbeanspruchtenSchraubenverbidungen, zylindrische Einschraubenverbindungen, VDI,Düsseldorf, Germany, 1986.

2. Grotewhol, A., Calculation of a dynamically and eccentrically loadedbolted conrod connection according to VDI 2230, SAE Paper 750882, 1975.

3. Hagiwara, M., A method to simplify the strength design of boltedjointsCase of connecting-rod bolts, Exp. Mech., March 1984, pp. 2832.

4. McNeill, W., Dynamic analysis of connecting rod bolts, SAE Paper881287, 1988.

5. Yoshimoto, I., A hypothesis concerning fatigue strength of bolt nut joints,Bull. PME (TIT), 51:4346 (1982).


19Design of Gasketed Joints

Bernard S. NauConsultant, BHR Group Ltd., Cranfield,Bedford, England

1Introduction

1.1Why Gasketed Joints Are Special

The reasons for treating the design of gasketedjoints separately from that of other boltedconnections are quite fundamental:

1. A pressurized gasketed joint is staticallyindeterminate.

2. Control of leakage is a crucial requirement.


The static indeterminacy arises from the fact thatthe gasket, the two flanges, and the boltingbehave like an assemblage of coupled springs.The gasket resembles a compression spring(nonlinear), the bolting a tension spring, and theflange a rotational spring. Also, the lines ofaction of the ''springs" are offset relative to oneanother, with the result that moments as well asforces act on the flanges. In general, the twoflanges can differ in geometry and material.

In consequence of the static indeterminacy,iteration is needed to calculate the equilibriumdisplacements of the joint.

The general objective of the design process is toestablish, for given gasket characteristics, safeeconomical dimensions of the flanges andbolting while ensuring adequate control ofleakage.

1.2Scope of the Chapter


Gasketed joints come in many geometries basedon a variety of design concepts. Although mostjoints are circular, as in piping joints, examplesof other shapes are a rectangular joint between asump and cover plate; the complex joint betweenthe two halves of a split-casing pump, valve, orturbine; and the cylinder head joint of an internalcombustion engine. The contained fluid may beliquid or gas, pressure may range from vacuumto thousands of bar (applied internally orexternally), and temperature may range fromcryogenic to 5001000°C or more. Operatingconditions may be steady or cyclic or transient ora combination of these. Life and leak raterequirements also vary widely.


Here we concentrate on internally pressurized circular bolted flanged gasketed joints toillustrate approaches to the design of joints more generally.

1.3General Design Considerations

Regardless of geometry and operating conditions, the design of a gasketed joint mustaddress two primary concerns:

1. Stress throughout the structure must always be within safe limits.

2. Gasket stress must be sufficient to keep the leak rate below an acceptable limit (seeChap. 7).

These requirements must be met for all load conditions. Steady-state load conditions forconsideration at the design stage include

1. Initial assembly, when bolts have been tightened.

2. Hydrostatic test at, for example, 1.5× normal working pressure.

3. Normal working conditions (e.g., elevated temperature and pressure).

Depending on the application it may also be necessary to consider

External loads, axial or bending, e.g., pipework weight and thermal expansion.

Bolt-tightening inaccuracy, depending on method used (Chaps. 3854).

Transient conditions such as heating up or cooling down.

Fatigue due to cyclic operation.

Corrosion, which reduces metal thickness, necessitating incorporation of a corrosionallowance on affected dimensions.

Creep of flanges, bolting, and gasketall can reduce gasket compression stress.

Many applications can be met by choosing a standard joint design; this approach isdiscussed in Section 3. When a standard joint cannot be used, the next possibility toconsider is the use of a "code" joint as described in Section 5.

To provide an understanding of the complexities of the design and behavior of boltedgasketed flanged joints, we first consider in some detail the workings of gasketed joints.This is sometimes done geometrically, using a rather complex loaddeflection diagram (e.g.,in the German draft standard DIN V2505 [1]); an algebraic approach is more explicit and is


adopted here.

2Forces, Moments, and Interactions

2.1Simplified Model of a Joint

Here we focus on joints using taper-hub weld-neck flanges, which are commonly chosenfor high duty applications. To understand the interactions of the loads acting on a joint andthe resulting displacements, we can view the joint as a system of "spring" elements. In thissimplified approach, the gasket, bolting, and flanges are each treated as some form ofspring (linear), an approach first adopted by Wesstrom and Bergh in 1951 [2].

The spring rates can be written down directly for the gasket and the bolting.

Compressive stiffness of the gasket:

where

Eg=compression modulus of the gasket, unloading condition(psi, N/mm2)

Ag=an appropriate gasket contact area (in.2, mm2)


tg=an appropriate gasket thickness (in., mm)

kg=compressive stiffness (lb/in., N/mm)

Axial stiffness of the bolting:

where

Eb=Young's modulus of the bolts (psi, N/mm2)

Ab=effective cross-sectional area of the set of bolts(in.2, mm2)

lb=effective clamping length of the bolts (in., mm)kb=axial stiffness of the bolting (lb/in., N/mm)

In reality, the calculation of the deflections of the flange can be complicated by theexistence of several factors affecting axisymmetric deflection:

1. Ring-mode rotation of the flange cross section about its centroid without change ofshape of the cross section

2. Radial bending of the flange proper

3. Axial bending of the hub

4. Bending of the bolts

5. Radial stiffness of the wall of the adjoining piping or vessel

6. Angular stiffness of the wall of the adjoining piping or vessel

A simple approximation is to model only the first of these, although a rigorous analysisrequires consideration of all modes. Bending of the flange is important when the radialwidth is large compared with axial thickness; its neglect leads to underestimation ofdeflections. Neglect of bending of the hub also causes an underestimate. However,neglecting bolt bending is conservative, as it resists flange rotation.

The connection to the pipe or vessel wall constrains flange deflection but also introducesextra loads, due to geometric discontinuity, which appear as a moment and radial shearforce at the junction.


Considering one flange of the pair, the major forces acting are shown in Figure 1, whichalso defines the coordinates x and f, describing flange deflections relative to the matingflange.

Figure 1Major forces acting on a flange.

(From Ref. 3.)


In Figure 1 the notation is as follows:

B=total bolt load (lbf, N)

G=gasket reaction (lbf, N)

P=

fluid force due to pressure acting over the end of the pipe orvessel (lbf, N)

x=

coordinate for axial translation of the flange centroid relative tothat of the second flange (in., mm)

f=

coordinate for axisymmetric rotation of the flange as a ring(radians)

b=lever arm of the bolt force B about the centroid (in., mm)

g=lever arm of the gasket force G about the centroid (in., mm)

r=

lever arm of the fluid end force P (initially zero) about theflange centroid (in., mm)

As operating conditions change, for example, due to pressurization of the fluid, the threeforces B, G, and P interact due to changes in

1. Axial displacement x of the flange, as the gasket compresses or relaxes, and bolt stretchchanges

2. Rotational displacement f of the flange about its centroid, as moments vary with thechanging forces

These flange displacements are defined relative to the initial assembled condition. As wesee later, these interactions also depend on three stiffness coefficients:

1. Compressive stiffness of the gasket, kg (actually a nonlinear function of compressivestress)

2. Axial stiffness of the bolting, kb

3. Rotational stiffness of the flange, kf, (actually a nonlinear function of rotation if localyield occurs in the flange)

The mating flange might not be of the same design and material. Here for clarity weassume that the mating flange is rigid. Regardless of this assumption, the values of B, G,


and P acting on the second flange are the same as those on the first, but moments andflange stiffness may differ.

When the joint is assembled, before the fluid is pressurized (fluid pressure p = 0), we canwrite an axial force equilibrium condition:

and after pressurization, fluid pressure p > 0:

Subtracting Eq. (3) from Eq. (4) gives an expression relating changes in force:

These force changes are proportional to flange displacements (Dx, Df):

Also, internal flange stresses are in equilibrium with the moments acting; therefore,

where Mext(p) represents pressure-dependent external moments (axisymmetric) due, e.g., tothe adjoining pipe wall as it expands radially under the influence of the fluid pressure.

To simplify this equation define a coefficient C such that


where Dp º p. Then

We now have four equations [(5), (6), (7), and (8b)] to solve for four unknowns: DG, DB,Dx, and Df. Rather than use algebraic manipulation, it is more convenient to write theequations in matrix form and solve using a computer matrix solver. The matrix equationtakes the form

in which DP has been replaced by (ApDp), where Ap is area exposed to axial fluid pressure.

If the second flange is not rigid, Eqs. (6) and (7) have to be generalized by replacing

and

with appropriate coordinate system, and writing Eqs. (8) separately for each flange. Thisgives five equations, which can be solved for the five unknowns.

2.2Effects of Relative Stiffness of Gasket and Bolting

The special case of a joint having rotationally rigid flanges (Df = 0) gives useful insightinto the significance of the relative stiffness of gasket and bolting in a joint. It can beshown, from Eqs. (5)(7), that when fluid pressure is applied to the assembled joint, thegasket force and bolt force change as follows:

Thus the gasket stress decreases and the bolt force increases to maintain the axial forceequilibrium as the fluid pressure is applied. There are two limiting cases of interest toconsider.


Case (i): If the gasket is stiff compared with the bolting, kg >> kb, then

Case (ii): If the bolting is stiff compared with the gasket, kg << kb, then

Case (ii) gives constant gasket stress (DG = 0), a desirable situation as the gasket stressmaintains its design value. It might therefore seem that the gasket should be chosen to bevery compliant; in practice, however, excellent gasket performance is obtained, at highpressure and temperature, when the gasket is stiff compared with the bolting. Why shouldthis be? In fact, our simple model omits flange rotation, and when a flange rotates a stiffgasket applies a restoring moment to the flange, opposing rotation. A more completemodel would bring out such effects. It might seem "obvious" that sheet gaskets and spiral-wound gaskets would be appreciably stiffer than the bolting in joints, but actually this isusually not the case, especially for large sizes [4]. This slightly surprising fact is due to thecombination of large gasket area with fairly high modulus (which increases at highcompressive stresses), which together give a high stiffness. If the flanges can also rotate,then it is possible for the bolt force to either increase or decrease. Either way the gasketstress decreases.


TABLE 1 Examples of Specifications for Standard Steel FlangesInternational Standard ISO 7005, Metallic Flanges. Part 1: Steel Flanges.ASME B16.51988, Steel Pipe Flanges and Flanged Fittings (formerly ANSIB16.5 and then ASME/ANSI B16.5).ASME B16.471990, Large Diameter Steel Flanges: NPS 26 Through NPS60.European Draft Standard CEN prEN 10921, Flanges and TheirJointsCircular Flanges for Pipes, Valves and Fittings. Part 1. SteelFlangesPN Designated.European Draft Standard CEN prEN 17591, Flanges and TheirJointsCircular Flanges for Pipes, Valves, Fittings and Accessories, ClassDesignated. Part 1. Steel Flanges, DN 15 to DN 600 (NPS 1/2 to 25).Source: Ref. 3.

3Standard Joints

Many applications can be met by choosing a standard jointdesign. This is usually the quickest and most efficientapproach. Standard flanges are commercially available, asare gasket sizes to match. Standard joints are specified indetail in national and international standards, such as thosein Table 1.

3.1Flange Types

Standard flange sizes range up to 600 mm and more.


Typical design types include those listed in Table 2 andillustrated schematically in Figure 2.

The weld-neck taper-hub flange (Fig. 2A), with itsrelatively narrow raised face to accommodate the gasket, isa common choice for high duty applications. Its reinforcinghub gives it extra rotational stiffness, which reducesstresses in the metal and gives a better environment for thegasket. It is usually less expensive to extend or thicken thehub than to increase the axial thickness or diameter of theflange proper.

Flanges with full-face gaskets (e.g., Fig. 2D) are regardedas light duty designs. This is because the compressivestress applied to the gasket by the given bolting is lowcompared with that for a narrower gasket. Threaded flanges(not illustrated) are also intended for light duty. In thiscase, because the threaded construction provides littlereinforcement and the flange is essentially free to rotateunder the moments generated by the bolting force, gasketreaction and fluid end-load.

The controlled compression gasket concept (Fig. 2F) is notwidely used, as its advantages are not generallyappreciated. The chief of these is that the gasket does notsuffer changes in compressive load as fluid pressure andtemperature change, so it is maintained in an optimum


state for leakage control. Also, provided that the bolts aretightened well beyond what is required to seat the flangesmetal-to-metal, the bolts do not suffer cyclic loading,which reducesTABLE 2 Examples of Standard Flange TypesWelded construction Nonwelded constructionWeld-neck taper-hub flange Threaded flangeSlip-on weld plate flange Lap flange (engages abutment at pipe end)Weld-socket plate flange Blank flange ("end closure")Source: Ref. 3.


Figure 2Examples of standard flanges.

(From Ref. 3.)


the risk of fatigue or loosening. Furtheradvantages result from reduced flange rotation asa consequence of

1. The rotational constraint due to the pivotreaction force acting at the periphery

2. The direction of the bolting force moment,which is reversed due to the bolt circle beinginboard of the flange pivot (unlike a raised-facejoint)

The tendency of such flanges to bend about thebolt circle results in radial bending stresses herethat have to be taken into account whenevaluating design stresses. Consideration mayalso have to be given to possible stress corrosionproblems resulting from the bolting being morehighly stressed than in other types of joints.

Spiral-wound gaskets are commonly fitted withan outer ring (Fig. 2E) to limit compression of


an outer ring (Fig. 2E) to limit compression ofthe windings. This does not behave in the sameway as the controlled compression concept justdescribed. Because the limit ring is inboard ofthe bolt circle it is possible for overtightening ofthe bolts to cause the flange and ring to pivot onthe outer edge of the flange's raised face. Thisunloads the inner region of the gasket and canactually increase leakage if the bolts areovertightened.

The lapped plate flange (Fig. 2G) is anotherdesign that has not been fully exploited. With thisarrangement, flange rotation is largely decoupledfrom the pipe-end hub and the gasket. This againbenefits gasket performance.

The blanking flange (Fig. 2H) is a standarddesign. In demanding applications, stresses canbe reduced if the central disc is axially offsetfrom the flange proper by inserting a length ofpipe. This isolates the flange region from thelarge bending moment due to fluid pressurebowing the central region of the disc.


bowing the central region of the disc.

There may be limitations on the use of certaintypes of standard joints. For instance, socket-weld flanges and threaded flanges are notrecommended for use at extreme temperatureswhen there is temperature cycling (below 40°Cor above 260°C) or if large temperaturegradients will exist in the joint.

3.2Information in Standards

Flanged joint standards typically tabulate suchdetails as

Dimensions of flanges and bolts.

Face details, including dimensions, finish, andpermissible defects.

Bolt-hole templates.

Weld geometry and dimensions.

Material specifications.


"Ratings"; i.e., maximum-allowable non-shockpressuretemperature limits, for each size andduty class, for various groups of constructionmaterials. The materials are grouped accordingto their stressstrain properties over relevanttemperature ranges.

Gasket dimensions may be included in the mainflange standard (e.g., ASME) or in separategasket standards (e.g., ISO and CEN).

3.3Terminology

Class and PN

Standard flange designs are grouped according tothe severity of the duties for which they aresuitable. For example, in ASME B16.5 [5], Class150 joints are the lightest duty designs and Class2500 are the heaviest duty designs. In ISO 7005[6], designations range from PN 2.5, the lightest,to PN 420, which is the heaviest duty design.


The designation "Class" is used by ASME (andCEN in standards derived from ASME), and thedesignation "PN" is used by ISO and CEN. Bothdesignations are simply numerical identificationlabels; they do not have physical units.

Note that ASME B16.5 also uses the numericalvalue of the class label in deriving its tabulatedpT ratings. Thus for Class 300, the primaryrating pressure used in calculating ratings is 300psi; for Class 600 it is 600 psi; and so on, withthe exception of Class 150, for which it is only115 psi. However, the primary rating pressurevalue does not appear explicitly in the ratingtables.

NPS and DN

These are analogous designations to indicatediametral size. Thus ASME B16.5 uses NPS##,i.e., nominal pipe size; and ISO and CEN use


DN### (CEN uses both for class-designatedflanges). Again, such designations are intendedonly as an identification label, in this case toindicate a general diametral size category for usein identifying items of equipment of consistentsize.

When items of equipment are combined togetherin a system, they are normally chosen so that allthe items in the system have matching numericaldesignations for duty and similarly for size. Forsafety reasons one would not use a Class 300flange on a Class 600 valve, for example, and foreconomic reasons one would not do the reverse.

4Proprietary Joints

There are many special joint designs, and evenmore special gasket types, the latter being outsideour scope. These are developed and/or marketedby specific companies and are usually the subjectof patents. Examples of proprietary joints are


given in Table 3 and Figure 3 to illustrate designfeatures of interest. Most are more compact andweigh less than a standard flange, while retainingor improving on pT rating. They can be groupedaccording to whether they have axial bolting or aclamping arrangement.

Wedge gaskets (designs A, B, and E in Fig. 3)generate high stresses at the sealing contactwithout requiring very high bolt loads, giving thepotential for good high duty performance. Clampdesigns E and F have the advantage of fewerbolts, simplifying assembly and dismantling. Onthe other hand, some axially bolted designs (e.g.,A) conveniently require only slight modificationof standard flanges, avoiding costly specials.Design F uses a molded elastomeric seal ratherthan a more conventional gasket. Although thislimits thermal and chemical compatibility, itavoids the need for any welding to the pipe.Design D aims at the ultimate in compactness byusing small bolts located on a small pitch circle;


it also minimizes flange rotation by arrangingthat pivoting is at the outer perimeter, and it doesnot require a gasket if the faces are suitablyfinished. However, like other compact flanges, itdoes not mate with standard flanges, which limitsthe scope of application. Design C achievescompactness by a different approach, again usingsmall bolts on a small pitch circle but reducingflange rotation by using loose flanges. Thisdecouples flange rotation from the gasket. Theflanges are stiffened against rotation by virtue oftheir axial length, which is greater than in astandard loose flange; radial alignment of thepipes is ensured by a sleeve sliding over the endhubs.

5Code Design Joints

5.1What Is a Code?

The purpose of a joint design code is to provide a


simple "recipe" or procedure that enables theuser to design a gasketed joint that is structurallysound and meets leakage control requirements.Typically such codes comprise a series offormulas and charts. Key data, including


TABLE 3 Some Proprietary Joint DesignsaDesign characteristics Examples (trade names)Axially bolted, wedge-seal

Modified standard flanges (Fig. 3A)

Destec (Destec Engineering Ltd,Lincoln, UK)Flangeplus (Techlok Ltd, Port Talbot,UK)Grayloc (Gray Tool Co., Houston,TX)

Special flanges (Fig. 3B) Taperlok (Taper-Lok Corp., Houston,TX)

Compact

Lap joint (Fig. 3C)Integral (Fig. 3D)

UKAEA (UAKAEA, Warrington,UK)Verax (Steel Products Offshore A/S,Drammen, Norway)

Clamped

Hubbed pipe, controlled-compressionwedge gasket (Fig. 3E)

Destec G (Destec Engineering Ltd,Lincoln, UK)Flexitallic Clamp Joint (Flexitallic Ltd,Cleckheaton, UK)Grayloc (Gray Tool Co., Houston,TX)Techlok (Techlok Ltd, Port Talbot,UK)

Hubbed pipe, quick-disconnect, singlebolt Cefilac (Cefilac, Saint-Etienne, France)

Grooved pipe, elastomeric seal (Fig. 3F)Victaulic (The Victaulic Company,London, UK)

Plain pipe, ball swaged Morgrip (Hydra Tight, Walsall, UK


aSee Figure 3.Source: Ref. 3.

dimensions, gasket characteristics, operating conditions,and limiting stresses of the materials of construction, arefirst collated. Forces and moments acting on the joint arethen evaluated; this enables the bolting requirements, i.e.,number and diameter of bolts, to be established. Next, themagnitudes of stresses that have to be checked arecalculated and are tested against designated maximumvalues.

The procedure requires that you first assume plausibleflange dimensions, to be improved subsequently in light ofthe results of the code calculations. This sequence isinconvenient, but the nature of the problem is such that it isnot possible to work in the preferred direction, i.e., todirectly calculate required flange thickness for given stressand leakage requirements.

Table 4 compares some features of major design codes.

5.2Stress Analysis

The longest established code is the ASME Boiler andPressure Vessel Code for Unfired Pressure Vessels [7].This originated in an analytical stress analysis of taper-hub


flanges developed during the 1920s and 1930s [8], nowcalled the Taylor Forge method. It is the basis of nationalcodes in a number of other countries, including the Britishcode BS 5500 and the French CODAP. The approachfollowed in the German code DIN 2505 is quite different.


Figure 3Proprietary joint designs in Table 3.

(From Ref. 3.)

An international code currently under development is thedraft European code CEN/prEN 1591 [9]. This differsfundamentally from all those listed above, which use linearelastic analysis. The draft CEN code is based onelastoplastic limit load analysis. The basis of this code isone used successfully in the former German DemocraticRepublic (TGL 32903/13 [10]).

A limit load analysis is more realistic because it recognizeswhat has long been known, namely that many jointsexperience local yield and yet perform satisfactorily (see,e.g., Ref. 11). Local yield allows high stresses toredistribute themselves safely, behavior that has long


TABLE 4 Features of Some Design Codes for Integral Raised-Face Flanged Joints

Subject ASME Section VIII, Div 1,1995a

DIN 2505,1990b

CEN prEN 1591,1994

Basis flange model Taylor Forge annular plate Ring TGLRing

Hub model Cylindrical shell, variablethickness

Equivalent cylindricalshell

Equivalent cylindricalshell

Wall model Cylindrical shellAllowance for boltholes No Yes Yes

Reference state Elastic Elastic Plastic collapseReference stress Sallowable(T) syield(T) or sult(T) syield(T)Egasket No E = E(Sg, T) E = E(Sg, T)Gasket stress basis m, y m, svu optionsPressure inflation No No NoExternal moment No No YesExternal thrust No Yes YesTemperaturedifferential No (Yes) Yes

Flange rotation No? Yes YesLoaddeflectioninteraction No Yes Yes

Gasket width Arbitrary reduction Empirical CalculatedBolt bending No No NoFlange radialconstraint Hub junction Natural Natural

Reaction at hub/flangejoin Mid-hub Mid-wall Mid equivalent shell

aAlso BS 5500, CODAP, and others.b(Yes) denotes partial treatment.Source: Ref. 3.


been used in the design of civil engineering structures and withoutwhich most bridges and skyscrapers could not exist. A drawback ofthis approach is that it introduces nonlinear behavior into an alreadycomplicated analysis; however, this is simply recognizing reality.

5.3Flange Rotation

Flange rotation is of particular importance to the gasket and to leakagecontrol. Until recently, ASME-type codes did not take account of this(except by a crude designation of an ''effective width" of the gasket)even though the original theoretical development included an explicitformula for flange rotation [12]. The German DIN 2505 and draftCEN code do take direct account of flange rotation, the latter morerigorously.

5.4Leakage

Leakage is another fundamental issue inadequately dealt with byASME-type codes. Such codes involve the use of two gasketparameters (m and y) that together define the gasket stresses (and hencebolt load) required to "seal" the joint. In reality, leakage decreasesprogressively as gasket compressive stress is increased; there is noleak/seal divide as implied by the use of m and y. It is expected thatfuture issues of the ASME code will include an improved treatment ofthis subject, based on a procedure developed by The Pressure VesselResearch Council. The draft European code can use either approach.

6Analysis of Gasketed Flanged Joints


6.1Analytical Methods

The traditional approach is to consider a flange as an assemblage ofinteracting structural members with appropriate conditions forcontinuity of displacements, stresses, and moments imposed at theinterfaces between them (Fig. 4). The components are chosen on thebasis that analytical solutions exist for the chosen substructure andloading system. This normally means using some combination ofshells, plates, or rings. For instance, a taper-hub flange could beconsidered as a combination of

1. An annular ring that suffers rotation but not bending (this is theradial flange proper)

Figure 4Substructuring model used by Taylor Forge method.

(From Ref. 3.)


TABLE 5 Validity of Flange Component ModelsModel Validity conditionaThin circular plate w/t > 4Ring 1/4 < w/t < 4Beam w/t > 8, negligible hoop stressaSee Figure 5.Source: Ref. 3.

2. A variable-thickness cylindrical shell (the taperhub)

3. A cylindrical shell (wall of pipe or vessel)

The assembled equations for the substructurescomprise a set of simultaneous linear equationsthat can be solved by normal numerical methods.This approach can be implemented on a computerto provide a tailored model for joints of a givenshape, with dimensions and materials as variableinput data. Details of the formulation of such anapproach are given by Rodabaugh and Moore [11].

When dividing the joint into substructures, the


validity of simple models depends on the aspectratio of the component, the ratio w/t in Table 5 andFigure 5. It is also assumed that the in-plane stressvaries linearly through the thickness t. Design codesare based on this substructuring approach. Table 5summarizes key features and approximatesunderlying several codes.

6.2Finite Element Analysis

The substructuring method outlined above isessentially a finite element (FE) method with asingle element for each "component." Each of theseelements is a different type, chosen to model thatspecific component. Instead of simple polynomialinterpolation within elements, as in conventionalFE modeling, more accurate analytical expressionsare used. Interestingly, in the Taylor Forge analysis[12] the hub solution was actually obtained by anapproximate numerical method involvingminimization of energy, a method often used todevelop conventional FE elements.


However, the use of a general-purpose finiteelement stress analysis package is an alternative tothe use of a standard design code or an analyticalapproach. Advantages are

1. More accurate modeling of the geometry andmaterial properties.

2. Comprehensive output of stress and deformationdistributions.

Figure 5Notation for Table 5.

(From Ref. 3.)


On the other hand, perfection is not easilyachieved for the following reasons:

1. Setting up an FE analysis can take longer thana complete code analysis.

2. Gasket stressstrain characteristics arenonlinear and have hysteresis.

3. Gasket stress distribution and bolt load bothvary with flange rotation.

4. Elastoplastic analysis is required becauseflange stresses can exceed yield.

5. Static indeterminacy and materialnonlinearities necessitate an iterative solution.

6. The procedure still uses the "cut and try"method to optimize flange and bolt dimensions.

References


1. Flanged Joint Calculation, German DraftStandard DIN V2505, 1986.

2. Wesstrom, D. B., and S. E. Bergh, Effect ofinternal pressure on stresses and strains inboltedflanged connections, Trans. ASME 73:553568 (1951).

3. Nau, B. S., Fluid Sealing Lecture Notes:Joint Design, 1995 (unpubl.)

4. Nau, B. S., On the design of bolted gasketedjoints, Paper D1, Proc. 12th International Conf.on Fluid Sealing, BHRA, Cranfield, UK, 1989.

5. Steel Pipe Flanges and Flanged Fittings,ASME Standard B 16.5, 1988.

6. Metallic Flanges. Part 1: Steel Flanges,International Standard ISO 7005, 1990.

7. Boiler and Pressure Vessel Code for UnfiredPressure Vessels, ASME Standard, 1995.

8. Waters, E. O., D. B. Wesstrom, et al.,


Formulas for stresses in bolted flangedconnections, Trans. ASME FSP-59-4: 161169(1937).

9. Design Rules for Gasketed Circular FlangeConnections, Draft European Standard CENprEN 1591, 1994.

10. German Democratic Republic StandardTGL 32903/13, 1983.

11. Rodabaugh, E. C., and S. E. Moore,Evaluation of the Bolting and Flanges of ANSIB16.5 Flanged JointsASME Part A DesignRules, Oak Ridge Nuclear Laboratory ReportORNL/Sub/2913-3; NRC-5; P.R. 115-7b, 1976.

12. Waters, E. O., D. B. Rossheim, et al.,Development of General Formulas for BoltedFlanges, Taylor Forge & Pipe Works,Southfield, MI. Reprinted by Welding ResearchCouncil, New York, 1949.

13. Large Diameter Carbon Steel Flanges,


American Petroleum Institute Standard API605, 1980.

14. Flanges and Their JointsCircular Flanges forPipes, Valves and Fittings. Part 1. SteelFlangesPN Designated, Draft EuropeanStandard CEN prEN 10921, 1994.

15. Flanges and Their JointsCircular Flanges forPipes, Valves, Fittings and Accessories, ClassDesignated. Part 1. Steel Flanges, DN 15 to DN600 (NPS 1/2 to 24), Draft European StandardCEN prEN 1759-1, 1994.


20ASME Boiler and Pressure Vessel Code DesignRules for Bolted Flanges

George BibelUniversity of North Dakota, Grand Forks, NorthDakota

1Introduction

This chapter presents a brief discussion givinginsight into the ASME Code rules for design ofbolted flanges [1]. However, it does notsubstitute for the detailed requirements of aCode analysis.

Bolted flanged joints are a very simplemechanical component used in manyapplications. In this context a bolted flange is a


pressure vessel and piping component.

Most pressurized vessels or piping componentsfail by excessive stress. Although flanges canalso fail this way, failure is usually defined asleakage. Leakage usually occurs because ofinsufficient compressive stress on the gasket.

Bolted flange design can be broken down intothree fundamental problems:

1. Selection of an appropriate gasket for theservice conditions of pressure, temperature, andfluid compatibility.

2. Determination of required bolt loads toadequately compress the gasket and hold the jointtight under operating conditions.

3. Transmission of the bolt loading to the gasketwithout excessive flange stresses and deflections.

Gasket selection is dealt with in more detail inthe sister volume Gaskets and Gasketed Joints[2] and elsewhere.


The current ASME gasket factors m and ydetermine the required bolt loading (see Section3.3). New gasket factors Gb, Gs, and a have beendeveloped for many gasket types. The newfactors are based on extensive test data sponsoredby the Pressure Vessel Research Council(PVRC). The new gasket factors will eventuallybecome the basis of a new nonmandatoryappendix in ASME Section VIII to be used as anoption to m and y.


The discussion in this chapter is limited to ASMESection VIII, Division 1, Mandatory Appendix 2,based on gasket factors m and y. This is done forthe following reasons:

1. The use of factors m and y as the basis forASME bolted flange design is still mandatory andwill be for the foreseeable future.

2. The current design rules have been usedsuccessfully for many years in most applicationsand will remain an important benchmarkcomparison.

3. The new proposed rules with gasket factors Gb,Gs, and a are unnecessary for our purposes in thischapter. For those who are interested, they arediscussed at length in Refs. 2 and 3 and in manypapers published by the ASME and the WeldingResearch Council (WRC).

It is important to note that structural analysis of


It is important to note that structural analysis ofthe flange is essentially the same in the currentASME rules and in any proposed changes. Whatdoes change is the determination of bolt loadingrequired to compress the gasket. Therefore,statements made in this chapter will also apply tothe new proposed rules except for the calculationof required bolt loads.

2Flange Design:Background and Theory

An increase in the operating temperatures andpressures of steam generation equipment createda need for flange design rules applicable over arange of conditions. In 1935, Waters andcoworkers undertook the study of all existingflange stress formulas and the development ofnew rules over a practical range of applications.The results of their work [4, 5] were adopted bythe ASME in 1940 and are still in use today. Abrief review will give insight into the format and


complexity of these rules. Gasket factors m and y,first described by Rossheim and Markl [6], wereincluded in the ASME Code in 1942 and are stillused, with little change, in the present edition ofthe code [1].

The analysis of bolted flange structures is basedon plate-and-shell analysis of three separatecomponents as shown in Figure 1: (1) the shell,(2) the tapered hub, and (3) the ring flange. Eachcomponent is studied independently withunknown boundary conditions of displacement,rotation, shear, and moment. When the individualcomponents are assembled, it is


Figure 1The three structural components of a flange.


mathematically possible to solve all boundaryand stress conditions in terms of external loadsby equating the boundary conditions on eitherside of the junction.

This complex stress problem was reduced, withjudicious approximations, to the point wheremanual calculations could be contemplated. Theprocedure was extended into a practical designprocedure by precalculation of various functions.(Alternative design methods use different andpossibly more accurate assumptions for theboundary conditions.)

For example, the stress distribution in the shelland hub can be described in terms of Mh1, themoment on the large end of the hub. Thecalculations were done for a full range of huband shell geometries and became flange factors Fand V, given in the Code in formula and chartform. Different boundary conditions apply to


loose types of flanges. Therefore, different chartsand formulas are given for FL and VL [1, 4, 7].

The flange ring is summarized by theprecalculation of factors U, Y, Z, and T. Theflange factors F, V, U, Y, and Z, with furtheradjustments given in the Code, multiply theflange moments caused by bolt loads, gasketreaction, and pressure thrust to form the flangestresses.

For a flange with a straight hub, the criticalstresses are the radial and hoop stresses at theinside diameter of the ring and the axial hubstress at the surface of the junction with the ring.When the angle of the hub taper increases beyonda certain limit, the hub becomes more rigid andthe location of the critical axial stress in the hubmoves from the large end into the small end (seeFig. 2). A hub stress correction factor ¦ wasintroduced to obtain the "true" maximum stress.The stress at the hub ring junction multiplied by ¦is equal to the stress at


Figure 2Flange stresses.


the hub/shell junction. This factor is presented in graphical form in terms of the hubgeometric parameters.

3Flange Design:Application

3.1Scope

The rules in ASME Section VIII, Division 1, Appendix 2 apply specifically to "the designof bolted flange connections with gaskets that are entirely within the circle enclosed by thebolt holes, with no contact outside the circle." It is recommended that flanges conformingto the following standards be used when connecting to external piping:

ASME/ANSI B16.5, Pipe Flanges and Flanged Fittings [8]

ASME/ANSI B16.24, Cast Copper Alloy Pipe Flanges and Flange Fittings, Class 150,300, 400, 600, 900, 1500, and 2500 [9]

ASME/ANSI B16.42, Ductile Iron Pipe Flanges and Flanged Fittings, Class 150 and300 [10]

ASME/ANSI B16.47, Large Diameter Steel Flanges, NPS 26 Through PS 60 [11]

These standards may be used for other pressure vessel flange connections within theirspecified pressure-temperature ratings.

3.2Procedure

The complex solution method has been reduced to a "cookbook" procedure in the Code.Numerous correction factors are used to adjust for the variety of possible geometricconfigurations. Once the flange geometry and gasket type have been identified, theprocedure reduces to a four-part process.

1. Identify the required bolt loads for

a. Gasket seating.

b. Operating conditions.

2. Calculate the internal flange forces and moments for gasket seating and operatingconditions.


3. Calculate the flange factors and stresses for gasket seating and operating conditions.

4. Compare to the Code allowable stresses.

The individual steps are reviewed below.

3.3Required Bolt Load

Appendix 2 is based on two separate and parallel designs. First, the bolt loading requiredfor gasket seating, Wm2, is evaluated.

where

b =effective gasket width reduced to account for flange rotationabout the gasket

G =gasket diameter at location of gasket load reactiony =gasket seating load

Wm2=effective gasket area times a gasket seating factor y

Wm2 reflects the bolt load required to compress the gasket a sufficient amount that it flowsinto the surface imperfections of the flange to provide an adequate seal. This calculation isdone with no consideration given to design pressure.


After gasket seating, pressurization tries to separate the joint and reduce the gasket seatingstress. The operating bolt pressure Wm1 must resist the separating pressure and maintain asufficient gasket load to retain a seal.

where P is the design pressure. The first term on the right is the pressure thrust trying toseparate the joint (the pressure times the area contained by the gasket). The second term isthe residual load required on the gasket after pressurization and equals the gasket area timesthe operating pressure and a gasket factor m. Gasket factors m and y have little experimentalbasis but have proved adequate in the field for most applications. (The PVRC test programsought at first to identify "better m and y factors.")

The design bolt load W equals Wm1 for operating conditions. The flange stresses created byWm1 must be evaluated against the Code allowable stresses at the design temperature.

The flange stresses created by gasket seating are evaluated by comparison to the Codeallowable stresses for ambient conditions. Gasket seating and operating conditions must beseparately evaluated to see which one "controls" the design.

The design bolt load W for gasket seating is not Wm2 but an adjusted value. This adjustmentis made to acknowledge the reality of field conditions. If there are any leakage problems it iscommon practice to torque the bolts excessively by whatever means. Therefore, the designbolt load W for gasket seating is increased to

This is the average of the sum of the required bolt area Am, and the actual bolt area Abmultiplied by the allowable bolt stress Sa at ambient temperatures. Am is taken as the greaterof Wm1/Sb or Wm2/Sa. Sb is the Code allowable bolt stress at operating temperatures. Thisaverage value increases the design gasket seating bolt load to acknowledge the possibility ofoverbolting.

3.4Internal Flange Moments

For gasket seating the internal flange moment is given by

where

C.


C.=bolt circle diameter

G =location of the gasket loadreaction.

For operating conditions the loading forces and moment arms are as shown in Figure 2.

These three loading forces are due to internal pressure and work to separate the flanges.They are balanced by the bolt load W.

The total internal moment for operating conditions is the sum of the product of each of theabove forces and its moment arm as shown in Figure 3.

where hD, hT, and hG are the moment arms as shown in Figure 3.


Figure 3Forces and moment arms on a bolted flange.

3.5Calculation of Flange Stresses

Once the internal flange moments are determined for gasket seating and independently foroperating conditions, the flange stresses can be calculated.

Longitudinal hub stress:

Radial flange stress:

Tangential flange stress:

where t is the flange thickness and g1 is the hub thickness at the back of the flange. Factors¦, L, e, Y, and Z are correction factors given in the Code and part of the adjustmentsdescribed earlier that reduce the numerical solution of a system of differential equations toa "cookbook" procedure.

For low pressure, small diameter flanges, seating conditions are expected to control the


design, especially for gaskets with high y gasket factors. For high pressure, large diameter


flanges, operating conditions are expected to control, especially for gaskets with high mgasket factors.

3.6Flange Rotation and Rigidity

Large diameter, low pressure flanges tend to be flimsy with excessive rotation. The gasketserves as a pivot point for the rotation. It is even possible for leakage to increase withincreased bolt loading as the flange rotates off the gasket.

The original analysis developed by Waters et al. [4] included a calculation of the hub ringjunction rotation. Many designers have used this equation to place limits on flange rotationof 0.2° for loose type flanges and 0.3° on integral type flanges.

This equation, in modified form, was recently added to the ASME Section VIII,Nonmandatory Appendix S, Design Considerations for Bolted Flange Connections. Theflange rotation was changed to a rigidity index (J) because of concern that J may beinterpreted as something more physically measurable than the accuracy of the theorypermits. For integral flanges,

For loose type flanges with hubs:

For loose type flanges without hubs (and optional flanges designed as loose type flanges)the rigidity index is modified as shown below.

J must be equal to or less than 1. KI = 0.3 for integral or optional flanges, and KL = 0.2 forloose type flanges.

3.7Step-by-Step Procedure

The Code procedure identifies 61 separate terms. Most designers have access to acomputerized version of the design process. The standard for hand calculations remainssheet A from Taylor Forge's "Modern Flange Design," Bulletin 502 (Fig. 4). The advantage


of this format is that the terms are introduced as needed in the calculation process. Anupdated version of the form (Slight changes have occurred in nomenclature.) The ASMECode is still required to determine the flange factors.

4Nomenclature Used in Flange Design Formulas

The symbols defined below are used in the formulas for the design of integral flanges arereprinted with permission from the 1995 ASME Boiler and Pressure Vessel Code, SectionVIII, Division 1, Appendix 2. All references to figures, tables or paragraphs are from 1995ASME Code. (See ASME Code for loose type or lap type flanges.)

A=

outside diameter of flange or, where slotted holes extend to theoutside of the flange, the diameter to the bottom of the slots (in.).

Ab=

cross-sectional area of bolts using the root diameter of the thread orleast diameter of unthreaded position, if less (in.2).


Figure 4Welding neck flange design, integral type (see Code for loose or lap type flanges).

Adopted from ''Modern Flange Design: Bulletin 502" sheet A.(Courtesy of Taylor Forge Company.)


Am=

total required cross-sectional area of bolts, taken as the greater of Am1and Am2 (in.2).

Am1=

total cross-sectional area of bolts at root of thread or section of leastdiameter under stress, required for the operating conditions (in.2), Am1= Wm1/Sb.

Am2=

total cross-sectional area of bolts at root of thread or section of leastdiameter under stress, required for gasket seating (in.2). Am2 = Wm2/Sa.

B =inside diameter of flange (in.). When B <20g1, it will be optional forthe designer to substitute B1, for B in the formula for longitudinal hubstress SH.

B1=

B + g1 for loose type flanges and for integral type flanges that havecalculated values h/h0 and g1/g0 which would indicate an ¦ value ofless than 1.0, although the minimum value of ¦ permitted is 1.0.

B1=B + g0 for integral type flanges when ¦ > 1 (in.).

b =effective gasket or joint contact surface seating width (in.) [see Note1, 25(c)(1)].

b0=basic gasket seating width (in.). (from Table 2-5.2).

C =bolt circle diameter (in.).

c =basic dimension used for the minimum sizing of welds (in.), equal totn or tx, whichever is less.

d =factor (in.3) as follows: for integral type flanges; for loose type flanges.

e =factor (in.-1) as follows: e =F/h0 for integral type flanges; e = FI/h0for loose type flanges.

F=factor for integral type flanges (from Fig. 2-7.2).

FL=factor for loose type flanges (from Fig. 2-7.4).


¦ =hub stress correction factor for integral flanges from Fig. 2-7.6 (whengreater than 1, this is the ratio of the stress in the small end of hub tothe stress in the large end). For values below limit of figure, use ¦ = 1.

G=

diameter at location of gasket load reaction. Except as noted in sketch(1) of Fig. 2-4, G is defined as follows (see Table 2-5.2): When b0 <1/4 in., G = mean diameter of gasket contact face (in.); when b0 > 1/4in., G = outside diameter of gasket contact face less 2b (in.).

g0=thickness of hub at small end (in.).

g1=thickness of hub at back of flange (in.).

H=total hydrostatic end force (lb); H = 0.785G2P.

HD=hydrostatic end force on area inside of flange (lb); HD = 0.785 B2P.

HG=

gasket load (difference between flange design bolt load and totalhydrostatic end force) (lb); HG = W H.

Hp=total joint contact surface compression load (lb); Hp = 2b (3.14GmP).

HT=

difference between total hydrostatic end force and the hydrostatic endforce on area inside of flange (lb); HT = H HD.

h =hub length (in.).hD=

radial distance from the bolt circle to the circle on which HD acts, asprescribed in Table 2-6 (in.).

hG=

radial distance from gasket load reaction to the bolt circle (in.). hG =(C G)/2

h0=factor (in.); h0 (Bg0)1/2.

hT=

radial distance from the bolt circle to the circle on which HT acts, asprescribed in Table 2-6 (in.).

K=

ratio of outside diameter of flange to inside diameter of flange; K =A/B.


L =factor; L = (te + 1)/T + t3/d.MD

=component of moment due to HD (in.-lb); MD = HDhD.

MG=component of moment due to HG, in.-lb; MG = HGhG.

M0=

total moment acting upon the flange, for the operating conditions orgasket seating, as may apply (in.-lb) (see 2-6).


MT=component of moment due to HT (in.-lb); MT = HThT.

m =gasket factor, obtained from Table 2-5.1 [see Note 1,2-5(c)(1)].

N =width (in.) used to determine the basic gasket seating width b0 basedon the possible contact width of the gasket (see Table 2-5.2).

P =internal design pressure (psi). For flanges subject to external pressuresee 211.

R =radial distance from bolt circle to point of intersection of hub andback of flange (in.). For integral and hub flanges, R = (C B)/2 g1.

Sa =allowable bolt stress value at atmospheric temperature (psi) (see UG-23).

Sb =allowable bolt stress value at design temperature (psi) (see UG-23).

Sr =allowable design stress for material of flange at design temperature(operating condition) or atmospheric temperature (gasket seating), asmay apply (psi) (see UG-23).

Sn =allowable design stress for material of nozzle, neck, vessel, or pipewall, at design temperature (operating condition) or atmospherictemperature (gasket seating), as may apply (psi) (see UG-23).

SH=calculated longitudinal stress in hub (psi).

SR=calculated radial stress in flange (psi).

ST =calculated tangential stress in flange (psi).T =factor involving K (from Fig. 2-7.1).t =flange thickness (in.).

tn =nominal thickness of shell or nozzle wall to which flange or lap isattached (in.).

tx =

2× thickness g0 when the design is calculated as an integral flange(in.) or two times the thickness of the shell or nozzle wall requiredfor internal pressure when the design is calculated as a loose flange,


but not less than 1/4 in.U =factor involving K (from Fig. 2-7.1).V =factor for integral type flanges (from Fig. 2-7.3).VL=factor for loose type flanges (from Fig. 2-7.5).

W =flange design bolt load, for the operating conditions or gasket seating,as may apply (lb) (see 2-5(c)).

Wmi=

minimum required bolt load for the operating conditions (lb) (see 2-5(c)). For flange pairs used to contain a tubesheet for a floating headfor a U-tube type of heat exchanger, or for any other similar design,Wm1 shall be the larger of the values as individually calculated foreach flange, and that value shall be used for both flanges.

Wm2=minimum required bolt load for gasket seating (lb) (see 2-5(c)).

w =width (in.) used to determine the basic gasket seating width b0, basedon the contact width between the flange facing and the gasket (seeTable 2-5.2).

Y =factor involving K (from Fig. 2-7.1).

y =gasket or joint contact surface unit seating load (psi) [see Note 1,25(c)].

Z =factor involving K (from Fig. 2-7.1).

References

1. ASME Boiler and Pressure Vessel Code, Section VIII, Division 1,Appendixes 2 and S, 1995.

2. Bickford, J. H. (ed.), Gaskets and Gasketed Joints, Marcel Dekker,New York, 1997.

3. Bickford, J. H., An Introduction to the Design and Behavior of BoltedJoints, Marcel Dekker, New York, 1995.

4. Waters, E. O., D. B. Wesstrom, D. B. Rossheim, and F. S. G. Williams,


4. Waters, E. O., D. B. Wesstrom, D. B. Rossheim, and F. S. G. Williams,Formulas for stresses in bolted flanged connections, Trans. ASME59:161169 (1937).

5. Waters, E. O., D. B. Wesstrom, D. B. Rossheim, and F. S. G. Williams,Development of General Formulas for Bolted Flanges, Taylor Forge andPipe Works, Chicago, IL. Reprinted as a PVRC Monograph by theWelding Research Council in August 1979.


6. Gill, S. S., The Stress Analysis of PressureVessels and Pressure Components, Pergamon,New York, 1970.

7. Modern Flange Design. Taylor ForgeBulletin 502, Taylor Forge, Chicago, IL.

8. ASME/ANSI B16.5, Pipe Flanges and FlangedFittings.

9. ASME/ANSI B16.24, Cast Copper Alloy PipeFlanges and Flange Fittings, Class 150, 300,400, 600, 900, 1500, and 2500.

10. ASME/ANSI B16.42, Ductile Iron PipeFlanges and Flanged Fittings, Class 150 and300.

11. ASME/ANSI B16.47, Large Diameter SteelFlanges, NPS 26 Through NPS 60.

12. Rossheim, D. B., and A. R. C. Markl, Gasket


loading constants, Mechanical Engineering(9):647648 (1943).


21Design of Joints Loaded in Shear

Richard T. BarrettBarrett Engineering Consulting, Olmsted Falls,Ohio

1Introduction

Much more attention is paid to the design offasteners in tension than in shear, but sheardesign is just as important as tension. Inparticular, shear is the dominant fastener designfactor in the aircraft and aerospace field, wherestructural components are thin and very elastic.

The two general types of shear fasteners are hole-filling (rivets) and clearance hole (bolts andscrews). Both have important functions.


Shear loads and tension loads frequently occursimultaneously and must be combined andcompared to the total strength of the fastener.These and many other design criteria are coveredin this chapter.

2Definition of Shear Loading

Starting with the basic definition, a fastener shearjoint is one in which the load is perpendicular tothe fastener longitudinal axis as shown inFigures 1 (single shear) and 2 (double shear). Inthe single shear case, the equal and opposingforces are reacted to by shear stress on thefastener at plane A. In like manner, the doubleshear case has two reaction areas, at B and C,respectively. For riveted joints, there should beno gap between the material sheets. For a boltedjoint in shear it is imperative that the gapbetween material sheets be as close to zero aspossible to prevent fastener bending. (Note that a


single shear joint can have a small bendingmoment due to the eccentricity of the opposingforces.)

3Bolts and Screws

3.1Grip Length

The joint is weakened if the threads of thefastener lie in a shear plane between jointmembers or in the planes defined by the surfacesof the joint. To avoid this, only partially threaded


Figure 1Single Shear.

fasteners are used and a washer is placed underthe nut. Thread runout occurs above the surfaceof the joint. The nut can be tightened fully, andonly the smooth shank of the fastener lies withinthe hole.

All designs have their "weakest links." Forfasteners, the ideal shear design would fail in thesheet or plate in bearing at a load near (but still


less than) the shear yield of the fastener. Where abolt pattern is being used, the only way to loadup all of the bolts is for the bolt clearance holesto yield enough individually to close all the gapsbetween the bolts and the sheet. (See Fig. 3 forillustration.) This means that the shear strengthof the bolt must exceed the bearing yield strengthof the sheet. (The development of bearing stressallowables is covered in this chapter.)

3.2Fastener Edge Distance and Spacing

Although the actual edge distance and distancebetween fasteners can vary, a nominal value foredge distance is 2D with an absolute minimumof 1.5D, where D is the nominal diameter of thefastener. The nominal value for fastener spacingis 4D. These distances are measured from thehole centerlines as shown in Figure 4.

The thickness of the bolted sheets and whethersealing is required between sheets are major


factors in determining fastener spacing. Sheartearout calculations can be used to determinewhether edge distances are adequate. (Sheartearout is discussed in Section 12.1.)

3.3Shear Heads and Nuts

If the design loads for a fastener arepredominantly shear, it is possible to save weightand space by using fasteners with shear headsand/or nuts. The difference in thickness forcomparable heads and nuts is shown in Figures 5and 6, respectively. Note that the torque valuesmust also be reduced by approximately 50% forfasteners with shear heads, because the thin headsand nuts cannot develop the full tensile strengthsof the fastener shanks.


Figure 2Double Shear.


Figure 3Clearance hole gaps on a fastener pattern

(before loading).


Figure 4Fastener edge distance and spacing.


Figure 5Shear head vs. tension head.


Figure 6Shear nut vs. tension nut.

3.4Friction Effects

In most cases, friction forces between clamped(bolted) surfaces are not counted in determining load-


carrying capability. The reason for not includingfriction forces is the inability to determine thecoefficient of friction between clamped surfaces andaccurate clamping forces generated by each fastener.(Figure 7 illustrates the forces generated by jointfriction.) In some buildings and bridges, this contactfriction is monitored closely enough to be used as areaction force to prevent joint movement. However,the bolts are torqued to near yield levels to accomplishthis goal.

4Rivets

The analysis of rivets is not much different from boltanalysis, except that rivets are normally interference fitin holes and have a lower tensile load capacity thanbolts. The primary use of rivets is to carry shear loadsin relatively thin sheet or plate. (The maximum sheetthickness is usually less than the rivet diameter.) Thetwo main reasons for this requirement are to make therivet critical in bearing rather than shear (sheet willyield before rivet yields) and to be able to form theshop head on the rivet. (A more complete description


of rivets is given in Ref. 1.)

Figure 7Friction forces in a bolted joint. P = bolt preload; N =

friction load = Pm. Shear load on bolt = F 2N.


Figure 8Sheet inter-rivet buckling.

(From Ref. 2.)

4.1Inter-Rivet Buckling

If the distance between rivets is too large, the sheet can buckle as a column between rivets.This principle can be illustrated by pulling a sheet of paper in a three-ring binder along the11 in. border. The sheet will buckle between holes before the holes start tearing out. For apractical elastic buckling check of a riveted joint, as shown in Figure 8, use the bucklingformula [2],

where

Fbcr=buckling stress (psi)

t =thickness of sheet (in.)b =width of sheet (in.)a =length of sheet (in.)

E =modulus of elasticity(psi)

s =a = rivet spacing (in.)

and K is a factor taken from Figure 9.

Note that this formula is based on a Poisson's ratio of 0.3 and is for elastic buckling only.


Since the ratio varies from 0.28 to 0.35 for all metals, a value of 0.3 should besatisfactory for most calculations. Euler's formula can be used for more exact calculations[2]. A modification of Eq. (1) using the tangent modulus Et must be used for plasticbuckling.

4.2Countersunk and Dimpled Holes

Countersunk (flush or flat head) rivets and screws are very common, particularly forattaching aerodynamic surfaces. The most common countersunk head angles are 82° and100°, although special head angles can vary from 60° to 120°. A countersink should neverextend all the way through the first sheet, since this would create a knife edge hazard,explained below.

4.3Knife Edge Hazards

Knife edges are stress risers and are to be avoided in countersunk sheet holes (see Fig. 10).To avoid this problem, make the sheet thickness t greater than 1.5h as shown in the figure.If it is not possible to have t ³ 1.5h, then dimple the outside sheet and countersink the holein the inner (thicker) sheet as shown in Figure 11. If neither sheet is thick enough tomachine a countersink, both sheets can be dimpled to provide a flush outside surface asshown in Figure 12. However, the sheet must have sufficient ductility to dimple withoutcracking.


Figure 9K vs. a/b for sheet buckling.

(From Ref. 2.)

Figure 10Knife-edge countersunk sheet.


Figure 11Dimpled and countersunk holes.

4.4Hole-Filling Versus Non-Hole-FillingFasteners

Sometimes a designer will propose acombination of rivets and bolts or bolts anddowel pins to share a shear load. This is not anacceptable design. Rivets are hole-filling(interference fit). Dowel pins are slip or snug fit,and bolts have clearance holes. Therefore, anycombination of these three types of shearfasteners will give uneven load distribution, with


the tightest fasteners picking up most of the loadfirst.

Note that this does not preclude the use of dowelpins for alignment of a bolted joint as long as thebolts are designed to carry all of the structuralloads.

5Material Allowables

For fastener material selection, many aspectsmust be considered. Some of these are

1. Strength required

2. Operating temperature range

3. Corrosiveness of environment, including bothchemical and galvanic corrosion

4. Availability of fasteners in the desired material

5. Fatigue resistance (and general ductility)

Aerospace fasteners are not available from the


European or Asian markets. Therefore, metricaerospace fasteners are presently available onlyby special order from U.S. aerospace fastenermanufacturers to ALA/SAE specifications.

It is best to pick a material ultimate tensilestrength level under 200 ksi even for aerospacedesign, because of the stringent manufacturingand quality control procedures required for veryhigh strength materials. Typical aerospacefasteners are in the 140190 ksi range, whileindustrial fasteners are in the 60150 ksi range.

Figure 12Both holes dimpled.


5.1Shear Stress

Shear ultimate allowables for most aerospace materials are givenin the handbook MIL-HDBK-5 [3]. Bearing yield and ultimateallowables are given in MIL-HDBK-5 for many of the commonaerospace materials. However, commercial specifications such asthose of SAE, ASTM, and AISC, rarely give more than tensileultimate, tensile yield, and elongation.

In the absence of shear allowables, the following approximateultimate values may be used (Fsu = shear ultimate force and Ftu =tensile ultimate force):

For carbon steels, use Fsu = 0.6Ftu.For stainless steels, use Fsu = 0.55Ftu.For aluminum, use Fsu = 0.55Ftu.

Since a shear yield (Fsy) is a difficult thing to determine, my ruleof thumb is to use Fsy = 0.75Fsu for a shear yield allowable.

5.2Bearing Stress

A bearing allowable is an empirical (or fictitious) value that isdetermined from a shear test (usually per MIL-STD-1312). Thetrue bearing stress varies from 0 at the side of the hole (90° fromthe load direction) to a maximum at the center of the hole asshown in Figure 13a. This semicircular load distribution is


equated to an equivalent stress, using a rectangular area equal toDt, where D is the fastener diameter and t is the sheet thickness,as shown in Figure 13b. The bearing stress allowables such asthose given in MIL-HDBK-5 are determined (per MIL-STD-1312) by testing a sheet to failure in shear. The failure load isdivided by Dt to give the ultimate allowable bearing stress. Inlike manner, the yield load is used to calculate an allowablebearing yield stress. These allowables usually range from 1.5 to2.0 times the actual compressive yield (Fcy) and compressiveultimate (Fcu) of the material for an edge distance of 2D on thesheet. In the absence of established bearing allowables, Fbru =1.5Fcu and Fbry = 1.5Fcy may be used for design allowables forall common structural metals except magnesium (use 1.3 formagnesium).

5.3Development of Bearing Stress Distribution in a Bolt Hole

C. Anthony Hermann of NASA Lewis and I wrote a computerprogram to determine the actual stresses and deformations froma bolt in bearing in a steel plate. We used the COSMOS/M1.70Finite Element Model with plastic deformation. A summary ofthe cases run is provided in Table 1.


Figure 13Development of bearing stress allowables.


TABLE 1 COSMOS/M1.7 Shear Load SummaryBolt diam.(in.)

Shear load(lb.)

Plate thickness(in.)

Hole size(in.)

Max.displacement (in.)

0.500 10,000 0.500 0.563 0.02250.375 7,500 0.500 0.438 0.02010.250 3,500 0.350 0.281 0.0117

The bolts were assumed to be SAE Grade 8, andthe plate was mild steel with a tensile yield of37,500 psi. No fastener yield was reached inthese calculations, and all of the displacementsshown in Table 1 are in the plate material at thebottom of the clearance holes. The total contactangle was approximately 80° of the theoretical180° for each of these holes. This analysisreinforces the design concept that a fastenerpattern with clearance holes must be able toelongate the plate clearance holes to distributethe total load.

For details of the COSMOS/M program, contact


C. Anthony Hermann at Lewis Research Centerin Cleveland, Ohio.

6Countersunk Heads and Matched Drilling

A very common design error of countersunkfasteners in tapped holes is to make the matingplates on separate machines. Then an attempt ismade to assemble two separate hole patterns,both of which are self-aligning (the countersinkon one and the tapped hole on the other). Inaddition, the normal tolerances in true locationof the holes cause further mismatch of the holes.Figures 14 and 15 show the problems caused bythese various mismatches. The arrows in thecountersunk areas indicate where head bendingwill start.

Tapped holes with countersunk heads should beavoided. A better design would be to use aclearance hole and a nut. At least this allows thecountersunk head to do the alignment. If the back


side of the plate is blind (not accessible for nutinstallation), a floating nutplate can be installed.However, a floating nutplate only providesalignment adjustment parallel to the surface towhich it is riveted. It does not twist to providealignment for a bolt or screw that is notperpendicular to the mounting surface.

In summary, if countersunk holes are required,they should be line-drilled and countersunk whilethe mating sheets are clamped together to preventmisalignment problems.

Figure 14Shear loading of countersunk fasteners.

(a) Holes match but countersink not in line.


(b) Wrong countersink used.


Figure 15Shear loading of countersunk fasteners.

(a) Holes parallel but not in line.

(b) Holes not parallel.

7Counterbored Holes


If weight is not critical, it is easier to design forflush surfaces with socket-head screws incounterbored holes. However, the plate must bethick enough to have adequate shear area left onthe shank after cutting away a large portion ofthe total thickness (see Fig. 16). A hex or 12-point head can also be installed in a counterboredhole, but the counterbored diameter must belarge enough to allow a socket to go over thebolt head.

It is a little easier to get a counterbored surfaceparallel to the plate surface than to get acountersunk surface on line with the clearancehole axis.

A partial counterbore (which is also called aspotface) is sometimes used on sloped surfacesto give a top surface parallel to the bottomsurface in a joint to prevent bending on the bolthead.

8


Inserts

Inserts are normally used in relatively softmaterials to provide strong threads to engagesteel fasteners. There are two basic types ofinserts for structural applications: wire thread(helicoils) and threaded (keenserts). Each has itsadvantages and disadvantages as given below.(Helicoils are made by Emhart FasteningTechnologies of Shelton, CT; keenserts are aproduct of Fairchild Aerospace Fasteners,Torrance, CA.)

The wire thread insert is a wire spring with adiamond-shaped cross section that is installed ina tapped hole as shown in Figure 17. This insertcan also have a self-locking feature


Figure 16Counterbored hole.


Figure 17Wire thread insert installation.

as shown in Figure 18. The advantages of a wirethread insert are that it can be installed in asmaller tapped hole than a threaded insert(keensert), and it is much lighter than a keensert.On the negative side, the helicoil is harder toinstall, and it has a smaller outside diameter thana threaded insert (for the same diameter internal


thread). This smaller diameter gives less threadshear strength in the softer material.

A helicoil does not lock as well as a threadedinsert, so it will sometimes come out with thefastener when the fastener is removed.

The keensert is a solid threaded insert with bothexternal and internal threads. It can have anexternal locking feature such as the prongsshown in Figure 19, or a nylon plug or adeformed thread. The internal locking can be bydeformed thread, plug, or pellet or by a splitbeam(flat beam) locking nut on the bottom of theinsert as shown in Figure 20b.


Figure 18Wire thread insert types. (a) Free running; (b) locking.


Figure 19Keensert.

Helicoil inserts are available in lengths from 1D to 3D in order to develop the full bolttensile strength in softer materials. However, keenserts have a fixed length for a giveninternal thread size, with the outside thread made larger for heavy duty (higher load)applications.

Even if a keensert is not capable of developing the full tensile strength of a bolt, it is stillsatisfactory for shear applications if the bolt installation torque is kept low enough toprevent tensile pullout of the keensert. Remember that the edge distance (2D) must bebased on the insert's external thread diameter.

Although insert manufacturers sometimes state pullout loads, the loads frequently have tobe calculated. Pullout values for specific inserts are given in MIL-1-45914. Pulloutallowables can be calculated by using the formula

where

P=pullout load (lb)

dm=

mean diameter of threaded hole (approximately equal to pitchdiameter) (in.)

FS=

material ultimate or yield shear stress allowable (psi) (Note thataerospace design practices normally require both yield and ultimatemargins of safety.)


L=

length of full thread engagement (The last two or three threads at thebottom of the tapped hole are only partially formed.)

Note that the 1/3 factor is empirical. If the threads were perfectly mated, this factor wouldbe 1/2, since the total cylindrical shell area of the hole would be split equally between thebolt threads and the tapped hole threads. The 1/3 factor is used to allow for mismatchbetween threads.

9Nutplates

A nutplate (or anchor nut) is a blind nut that is normally riveted to the back of a sheet thatis too thin to tap. It requires the drilling of three holes: two for the installation rivets and a


Figure 20Flat beam insert manufactured by Tridair Fairchild.

(Courtesy of Fairchild Fastener Division, Torrance, CA.)

through hole for the fastener. The two basic types ofnutplates are the fixed and floating (see Figs. 21 and 22).The floating nutplate is the most popular because the nutis mounted in a crimped ''cage" to allow the nut to alignitself with the mating fastener. However, the nutplateonly moves parallel to its mounting surface, so it will notcorrect for fasteners that are not perpendicular to themounting surface. In general, nutplates can have theinternal thread-locking


Figure 21Fixed nutplates.

capabilities of a regular nut. However, the most popularone in the aerospace field is the deformed thread type.

Nutplates are frequently used for shear and access panels.Once again, a good design should have the fastenerscritical in bearing (sheet weaker than fastener) and thefastener grip length sized to have no threads in thethrough holes in the sheet.

10Lockbolts

In general, a lockbolt is a nonexpanding, high strengthfastener that has either a swaged collar or a type ofthreaded collar to lock it in place. It is installed in a


standard drilled hole with a snug fit but normally not aninterference fit. A lockbolt is similar to an ordinary rivetin that the locking collar or nut is weak in tension loadingand is difficult to remove once installed. An example isshown in Figure 23. Hence, lockbolts are normally usedfor shear designs.

The heavy truck industry uses lockbolts extensively forassembly of major structural components, because theyfit tighter than bolts and can be more quickly installed.Since these fasteners have clearances that are in betweenthose of bolts and rivets, they should not be used in ashear design with rivets and/or bolts.

11Analysis of a Pattern of Fasteners in Shear

For analyzing a pattern of fasteners, the first task is tofind the centroid of the pattern. This is done by thestandard static method of arbitrarily picking x and y axesand using a unit area


Figure 22Floating nutplates.


Figure 23Installed huckbolt fastener.

(Courtesy of Huck Manufacturing Co., Irvine, CA.)

times its distance from each axis to get the X and Y of the centroid. (See the Appendix tothis chapter.) Although it is poor design practice to have different sizes of fasteners in apattern, they can be analyzed by picking the most common one as a unit value. Then ratiothe other areas to this one. For example, if eight bolts of a 12-bolt pattern are 0.375 in. indiameter and the other four are 0.313 in., the 0.375 in. diameter shank area, 0.1104 in.2,would be the unit value of one and the value for the 0.313 in. diameter fasteners would be0.0767/0.1104 or 0.69. The shank areas can be used for shear, because it is not gooddesign practice to have threads in bearing. However, for tension loads, the thread root areashould be used for calculations.

In many cases, the fastener pattern may be symmetrical, as shown in Figure 24, so that thecentroid is obvious and calculations are unnecessary. With the centroid located, the nextstep is to divide the load R by the number of fasteners n to get the direct shear load Pc oneach fastener. To account for bending, the next step is to find for the pattern offasteners. The rn's are the radial distances from the centroid to the center of each fastener.In the symmetrical example shown in Figure 25,

Now calculate the moment of the load R about the centroid (M = Re) also from Figure 25.The shear load (pounds) for a particular fastener due to this moment is


Figure 24Symmetrical load pattern.


Figure 25Combining of shear and moment loading.

where r = distance from the centroid to the (usually) outermost fastener (in inches). Notethat the expression for Pe is analogous to the formula for torsional stress (torsional stress =Tr/J) except that Pe is in pounds. The two shear loads (Pc and Pe) can now be addedvectorially to give the resultant shear load. This load must now be combined with thefastener tension load and compared to the total strength of the fastener, using load or stressratios (see Section 13).

12Development of Fitting Stresses and Shear Tearout Stresses

12.1Shear Tearout

Earlier in this chapter we covered the development of bearing stress allowables. Theseallowables were predicted on an edge distance of 2D, where 2D is the distance from thecenterline of the fastener to the edge of the sheet (e.g., for a 0.250 in. diameter fastener, 2D= 0.500 in.). The adequacy of this assumption can be checked by taking two shear areas (seeFig. 26) out of the wall thickness and comparing their capacity to the bearing allowable of


the hole.

Since the hole is oversize, the actual length of each shear area is 2D Do/2, where D is thenominal fastener diameter and Do is the diameter of the clearance hole. Multiplying by thethickness t will give two shear areas of


Figure 26Shear tearout.

Using as an example a 0.250 in. diameter bolt in a 0.281 in. diameter clearance hole in a6061-T6 aluminum plate of 0.188 in. thickness:

There are two of these areas, so

Using yield strengths for comparisons (although ultimate strength can also be used):

For shear,

For bearing,


This shows that the 2D edge distance is more than adequate in shear for this case. An edgedistance of as little as 1.5D can be used in special cases, but it needs to be checked for sheartearout.

Shear tearout is more likely to occur along lines up to 40° from each side of the direction ofthe applied load, so taking two shear areas near the load line is slightly conservative.


12.2Development of Fitting (Lug) Stresses

Lugs are usually used for rotating connections.They are usually thicker than an ordinary shearjoint. Since lugs are thick, they normally have adouble shear connection to prevent pin bending(see Figs. 27 and 28). If the lug and clevis arerelatively soft (e.g., aluminum) compared to thepin, they may also have bushings installed forwear and lubrication. The lug end is usuallysemicircular, which is desirable from bothweight and clearance aspects.

The stress checks shown on Figure 27 must bemade to verify the lug design. Remember that thehole diameter for the bushing must be used whendetermining the net wall thickness for stresschecks. It is also imperative that the joint nothave gaps sufficient to induce pin bending.


To illustrate the necessary stress checks, we willwork through an example with the followingdimensions (from Fig. 27) and allowables:

W=1.0 in.

D =pin diameter = 0.5 in.P =3000 lbt =lug thickness = 0.5 in.

Material is 6061-T6 aluminum

From Ref. 3,

From Ref. 4,

From Ref. 3,


Figure 27Semicircular lug

(shownwithout clevis and pin for clarity).


Figure 28Semicircular lug assembly.

For tension failure:

For shear tearout failure:

with margins of safety


For bearing stress failure:


For hoop tension failure, the wall thickness is too high to use thehoop stress formula f = Pr/t, so a thick-wall formula [5] will beused:


In this case,

ro=outside radius = 0.5 in.

ri =inside radius = 0.25 in.

r =radius to the point of stressdesired

p =bearing stress = 12 ksi

Since we want the maximum stress (at the inside surface) we will let r = ri. Then


Note that safety factors (SF) have not been addressed here. They can be included in the3000 lb load or they can be added to the MS calculation:

In most aerospace applications the SF is different for yield and ultimate, so it would beadded to the MS calculation. For example, a carrier-based plane will have higher safetyfactors than a land-based plane due to the impact loads from arresting hooks. Note alsothat stress concentration factors (CF) are to be included in the safety factors.

In summary, our lug is good for the 3000 lb load, providing that no added safety factorsare large enough to give negative margins of safety. Also note that rigorous lug design


criteria are given in proprietary aircraft publications such as Lockheed's Stress MemoManual.

13Combining Fastener Shear and Tension Loads

When a fastener is subjected to both shear and tension loads simultaneously, the loadsmust be combined and compared to the total strength of the fastener. A good example ofthis type of loading is that of a fastener that has a tensile load created by torquing it (whichalso creates shear stress due to twisting) and an applied direct shear load.

The customary way to combine tensile and shear loading is to use load (stress) ratios [1].The load ratios are


Figure 29Interaction curves.

Note that the shear and tensile allowable loads can be obtained from publications such asMILHDBK-5 or Federal or SAE specifications for a given fastener size and strength.

The next step is to determine which interaction curve to use from Figure 29. The most

unconservative is . (This series of curves is from an old edition of MIL-

HDBK5.) The current MIL-HDBK-5 [3] lists only the single formula, . Ifweight is not critical, I recommend the R1 + R2 = 1. However, the accepted aerospace curve

is the . After calculating the RS and RT ratios, the margin of safety can becalculated as

where X and Y are exponents chosen to suit your own degree of conservatism. Note that thefasteners should be made of ductile material with a minimum elongation of 10%.

14Dowel Pins

Dowel pins are usually solid close tolerance pins that are used for aligning two matingcomponents. The pins are usually mounted in one half with a slight interference fit. Themating half has close tolerance clearance holes to slip over the protruding pins. Then the


two halves are bolted together and the bolts are analyzed for the total shear load. However,the joint can be designed with enough dowel pins to carry the entire shear load. In this casethe bolts would only ensure that the joint was tight enough to prevent pin bending. Notethat either the bolts or the pins (not both) carry the shear loads, owing to the differences inhole fits. The pins will load up first and yield their holes (or fail the pins) before the boltsstart to pick up shear loads.

Dowel pins installed in blind holes with interference fits should be avoided if pin removalis desired. It is much better to have a through hole, which allows the pin to be driven outfrom


Figure 30Straight (solid) dowel pin.

the back side. The through hole also allowsventing, which can be a problem on blind holeinstallations unless a longitudinal venting grooveor flat is provided on the pin.

Tapered dowel pins are available, as are pins withexternal serrations or ridges (to prevent pinrotation).

Shear allowables for dowel pins are usuallydetermined by test by the manufacturer because


the area at the shear plane may vary. (The pincross section may not be constant and/ or acommon geometric shape.)

Some common dowel pins are shown in Figures30 (straight), 31 (grooved), 32 (vented), and 33(drilled and tapped with vent).

15Roll Pins

Roll pins (also called spring pins) are usuallymade by rolling a piece of thin alloy sheet metalto a given diameter, with a chamfer on each end,and heat treating it to a high hardness. Thechamfer on the ends allows it to be started anddriven into a hole. The coiled cross sectiondecreases in diameter during driving to give aninterference fit. A typical roll pin is shown inFigure 34. However, if the cross section is notoverlapped (Fig. 35), the pin is called a slottedtubular pin.


Roll pins are used for quick-disconnect jointswith interference fit, such as for attaching ahandle to an actuating shaft. They can also beused as a fixed pin in a rotating joint where therotating component has an oversize hole to allowfree rotation around the roll pin.

Load-carrying capabilities for both roll pins andslotted tubular pins are usually determined by thepin manufacturer.

Figure 31Grooved dowel pin.


Figure 32Vented dowel pin.

Figure 33Drilled and tapped dowel pin with vent.


Figure 34Roll pin.


Figure 35Slotted tubular pin.

16Analysis of Countersunk Fasteners in Shear

Referring back to Figure 10 we see that the onlybearing area in contact with the rivet shank is (t h)Dunless the holes are match drilled and countersunk.How well the countersunk rivet head can carry shearload depends solely on its fit in the countersunk hole.The most conservative analysis would use only (t h)Dfor the shear area. However, the head will carry someload if the rivet installation is closely controlled.


Chapter 8 of MIL-HDBK-5 [3] gives shearallowables for countersunk rivets, based on test data.These values are acceptable for general design use, aslong as installation tolerances are as good as thoseused for the MILHDBK-5 test articles. Rivetstrengths must also be compared to those in the MIL-HDBK-5 table to ensure full compatibility of thetabulated values.

Since countersunk bolt heads have looser fits thancountersunk rivet heads, the best design approachwould be to use only the (t h)D bearing area forbearing stress calculations for countersunk bolts orscrews.

17Horizontal Shear Loads on Beam ConnectionFasteners

If we revisit the strength of materials section, whichdeals with beam bending and horizontal shearingstresses fS = VQ/Ib, we can determine the fastenersizes and spacing required for a composite beam. Inthis example case the uniformly loaded beam crosssection is symmetrical and consists of two flat plates


that are bolted together as shown in Figure 36 (notvery designefficient but easy to work with).

Assign specific numbers for the dimensions shown inFigure 37 as follows:


Figure 36Horizontal shear load. For this example,

total beam length = 50.

Further assume that the plates are structural steelwith Ftu = 67 ksi and Fty = 46 ksi. We will guessthat 0.50 in. diameter bolts will be sufficient. (Boltclearance hole diameter = 0.56 in.) Now fS =VQ/Ib can be calculated, where Q = staticalmoment and I is section moment of inertia.


Note: The moment of inertia I should berecalculated after the final bolt diameter isdetermined, as

where D is the bolt hole diameter. Then a final fS =VQ/Ib should be calculated. For no bolt holereduction,

This stress will be acting over the entire shadedarea of Figure 36, which is equal to 6e.


Figure 37Beam loading for horizontal shear load.


Then S(lb) = 1250 × 6e = 7500e. If we use 0.50in. diameter Grade 5 bolts, two bolts are goodfor 10,500 lb each in single shear withoutyielding. Solving for e, the maximum distancebetween rows of bolts,

which is close for our plate thickness.

Go to 0.56 diameter bolts and recalculate a newmoment of inertia with the hole areas deducted.

Now


Assume two Grade 5 (0.56 in. diameter) boltswith FSY = 13,900 lb each. Then

Of course, bearing and bending stresses must bechecked and the bolts must be critical in bearingto properly distribute their shear loads. Note alsothat bolt spacing can change as the vertical shearchanges.

18Galvanic Corrosion and Stress CorrosionCracking

Although galvanic corrosion and stress corrosioncracking should be a major consideration in allstructural design, we will touch on them here as


they apply to fasteners.

18.1Galvanic Corrosion

Galvanic corrosion is caused by contact betweentwo dissimilar metals in the presence of anelectrolyte. This electrolyte can be as simple asatmospheric air or water with other impuritiespresent. A galvanic cell is created, and the moreactive (anode) of the two materials is eroded anddeposited on the less active (cathode). Thefarther apart two materials are in the galvaniccorrosion table, the greater the galvanic actionbetween them [1].

To prevent or deter galvanic corrosion, one orboth materials are coated (or plated) with a moreinert material. For example, aerospace aluminumis anodized or iridited for protection. Thefastener holes are coated with MIL-S-8802sealant or zinc chromate paste to inhibit galvaniccorrosion. Steel fasteners are also coated with


cadmium or zinc. Aluminum rivets are usuallypainted with zinc chromate paint prior toinstallation.

It is good design practice to put hardened flatwashers under both the head and nut of a bolt tominimize damage and corrosion to the coatedsurfaces around the holes.

18.2Stress Corrosion

Stress corrosion occurs when a corrosion-susceptible material incurs tensile stress in acorrosive environment. An otherwise ductilematerial will fail at a stress much lower than itstensile yield strength due to surfaceimperfections (usually pits or cracks) caused bythe corrosive


environment. For a material that is susceptible tostress corrosion, the higher the strength (and thelower the ductility), the more sensitive it is tostress corrosion cracking.

Although no master list of stress corrosionsensitive fasteners exists, two common ones areH-11 tool steel bolts and 5056 aluminum rivets.The H-11 bolts are usually high tensile strength(200260 ksi), so they must have a protectivecoating to prevent corrosion. On the other hand,the 5056 aluminum has a low tensile strength(42 ksi) so it can be used for coldheaded rivets. Itwill sometimes crack from stress corrosionunder ''no load" from the tensile stresses inducedduring cold heading. However, 5056 aluminumis acceptable for riveting of magnesium becauseof their relative positions on the galvanic scale.

19


Helpful Hints for Fasteners Used in Shear Joints

1. Avoid feather edges on countersunk orcounterbored holes. Feather edges are prone todeveloping cracks when loaded in shear.

2. Shear allowables for many countersunk andprotruding head rivets are given in Section 9 ofMIL-HDBK-5 [3].

3. Single shear rivet joints produce some bendingand some rivet tension due to the change in loadpath at the rivet. This load condition should bechecked on critical designs, because rivets havesmall tension capabilities compared to their shearcapacities.

4. Do not combine rivets and bolts to share ashear load.

5. Do not combine dowel pins and bolts to sharea shear load.

6. Size the grip length of a shear fastener so thatno threads are in bearing (only smooth shank).


7. Do not use interference fit fasteners incomposite materials, as the compressive stressescan cause decomposition of the holes.

8. Do not subject fasteners to bending loads.

9. A pattern of fasteners must be critical inbearing (rather than shear) to distribute a shearload properly.

10. Check fasteners for combined shear andtensile stress.

11. Use washers under both the head and nut of abolt to avoid embedment in the joint material andgive more predictable torque vs. tension values.

12. Avoid the use of countersunk holes unlessthey are match-drilled with their mating parts.

13. Perpendicularity of fastener holes to theirmounting faces is critical (SAE allows ±2°tolerance) to avoid bending moments on thefastener heads.


14. Avoid the use of short "fat" fasteners intension. Since a fastener is a tension beam withan elongation of Dl = PL/AE (where P = tensionload, L = length between nut and under side ofhead, and A = cross-sectional area), we can seethat a smaller A and larger L will give ourfastener more elongation for a given load. Thejoint will then be easier to keep undercompression and will be more fatigue-resistant.

15. Avoid the use of tapped holes if possible. Athrough hole with a nut is much cheaper andmore convenient. Tapped holes have cut threadsthat are difficult to form and hard to inspect. Ifthe tapped hole is in a much softer material thanthe fastener, it is usually necessary to use aninsert. This adds further labor and cost to thedesign. Another disadvantage of tapped holes isthat the elastic working length of the fastener(from hint 14 above) is shortened when part ofthe joint becomes the "nut."


16. Use a nut with a maximum hardness that isnear the minimum hardness of the rolled externalthreads. This will allow the nut threads to yieldenough to distribute the bolt load to adjacentthreads. (My thanks to my friend DaveRichardson of Lockheed-Martin for this hint.)


Figure A1This joint is used as an example in the text to compute the location ofthe centroid (C) of the bolt group. The bolts are 1/4 in. in diameter.

The x and y axes are arbitrarily located along the left and bottom sidesof the plate.

Appendix:Finding the Centroid of a Bolt Group*

We can usually find the centroid of a bolt group by inspection. If the bolts are not in neatrows, however, as in Figure A1, we can use the following equations to locateit [8,9].

where

Ai=

cross-sectional area of the body of bolt i (in.2,mm2)xi=

distance of bolt i from an arbitrarily located y axis(in., mm)yidistance of bolt i from an arbitrarily located x axis


=(in., mm)=distance of the centroid from the y axis (in., mm)=distance of the centroid from the x axis (in., mm)

With reference to Figure A1, let us assume that the nominal diameter of the bolts is

*This appendix is reprinted from Ref. 7 with minor changes. Reprinted with permission.


1/4 in. and that we have arbitrarily located x andy axes along the bottom and left edge of thesplice plate as shown. The distance of bolt 1from the y axis is x1; its distance from the x axisis y1; etc.

Since each bolt has the same cross-sectional areain this example,

Similarly,


The centroid so located is labeled C in FigureA1.

References

1. Barrett, R. T., Fastener Design Manual, NASARP-1228, NASA Lewis Research Center,Cleveland, OH, 1990.

2. Peery, D. J., Aircraft Structures, McGraw-Hill, New York, 1950.

3. Metallic Materials and Elements forAerospace Vehicle Structures, MIL-HDBK-5G,Wright Patterson AFB, Dayton, OH, 1994.

4. Reynolds Structural Aluminum Design,Reynolds Metals Company, Richmond, VA,1962.

5. Timoshenko, S., and G. H. MacCullough,Elements of Strength of Materials, VanNostrand, New York, 1949.

6. Richardson, D. G., Unpublished data on


design of nuts, Lockheed Martin Company,Marietta, GA, 1995.

7. Bickford J. H., An Introduction to the Designand Behavior of Bolted Joints, 3rd ed., MarcelDekker, New York, 1995.

8. Shigley, J. E., and C. R. Mischke (eds.),Standard Handbook of Machine Design,McGraw-Hill, New York, 1986, Section 23,Bolted and Riveted Joints, pp. 23.1223.16.

9. Shigley, J. E., Mechanical EngineeringDesign, 3rd ed., McGraw-Hill, New York, 1977,p. 258.


22Design Rules for Structural Steel Bolting


William A. MilekConsultant, Glen Ellyn, Illinois

1Introduction

The Research Council of Structural Connections(RCSC), an organization of researchers, fastenermanufacturers and suppliers, structuralengineers, steel fabricators and erectors, andothers first authored and adopted theSpecification for Structural Joints Using ASTMA325 and A490 Bolts (hereinafter referred to asthe Specification) in 1952 [1]. Since that time


the council has taken cognizance of the researchdevelopments and periodically updated theSpecification based on the advances in the stateof the art and usage. The council's specificationsof various adoption dates are reviewed,endorsed, and published by the AmericanInstitute of Steel Construction (AISC). Thecouncil also sponsored the writing of Guide toDesign Criteria for Bolted and Riveted Joints[2], currently in its second edition (hereinafterreferred to as the Guide). The Guide is based onapproximately 800 references of experimentaland theoretical studies dating from the late1800s to recent times and is a comprehensivesource of information on the state of the art ofthe behavior and strength of riveted and boltedjoints. Additionally the AISC implements theprovisions of the council's specifications in theform of extensive design aids and tables for awide variety of typical structural applications inManual of Steel Construction, Volume II [3].Currently the RCSC Specification and the


implementing AISC Manuals Volume II areavailable in two formats, "Allowable StressDesign" (ASD) and "Load and Resistance FactorDesign" (LRFD) [4].

These documents are of interest because,although focused on structural steel applications,they are the only bolting specifications, of whichwe are aware, that not only provide authoritativecriteria for strength of joints subject to variousloading conditions but also tell the designer anduser how much tension or preload is required(when preload is required) in installed bolts. Todo this, the specifications address all significantfactors affecting the bolts, from their selection bythe designer to their inspection upon receipt atthe job site, handling and storage, verification ofthe suitability of fastener components (bolts,nuts, washers and


lubrication thereof; each probably from more than one source) plus their suitability as anassembly and when installed by the method to be used in the work to achieve the specifiedpretension. They then specify how to achieve the required preload and how to inspect for it.

Because it is difficult to maintain control over the many variables that affect bolt preload,this specification has inspired the development of special fasteners that attempt to controlor to indicate the attainment of the required preload in the fasteners during assembly. Thesedocuments address the inherent uncertainties of bolting and direct the designer's attentionto the installation and inspection procedures necessary to ensure performance of theconnections assumed for the design. Other industries could benefit from such detailedattention to the selection and use of fasteners.

Another reliable structural steel bolt that should not be overlooked by designers is theASTM A307. This bolt is a structural steel fastener for use primarily in shear-loadedapplications, where the higher shear capacity or the slip resistance of tensioned highstrength bolts is not required. It is not covered by the specifications that are the subject ofthis chapter because it is not considered high strength nor is it suitable for pretensioning toresist shear loads by friction between connected parts. On the other hand, it is a totallyreliable bolt that was the standard common bolt for shear/bearing applications prior to thedevelopment of the ASTM A325 and A490 bolts. Use of ASTM A307 bolts, whereverapplicable, provides for economy, not only through their lower cost but also becauseinstallation tensioning to a specified preload and inspection to ensure full pretensioning isnot required, which greatly reduces field labor cost.

2Two Versions of One Specification

The RCSC and the AISC provide two versions of the Specification based on differentdesign philosophies. Both versions are supported by the same criteria for the strength andserviceability of fasteners and connected material extensively documented in researchliterature. However, one version, subtitled "Allowable Stress Design" (ASD) [1] is basedon the more traditional approach of applying a global factor of safety, based on experienceand judgment, that is deemed adequate to provide for all the uncertainties in strength ofmaterial, tolerances in the manufacture of fasteners and connected material, improperstorage and handling, variability of lubrication, human error in fabrication and fieldinstallation, reliability of estimation of applied loads and load combinations, reliability ofengineering design formulas to predict actual connection performance, etc. The secondversion, subtitled "Load and Resistance Factor Design" (LRFD) [4] separates the loadconsiderations from considerations of the resistance to load of the fastener and connected


material. The resistance of bolts and the connected material to different potential modes offailure is developed by theory and test with resistance factors (f factors) selected on thebasis of statistical study of the data. The resistance, so determined, is required to be equalto or greater than the results of explicit consideration of the sum of loads in combinations(each load multiplied by a load factor g, which may vary according to the combination).The general expression may be written as

where Rn = nominal tensile strength of the bolt (kip).

This deceptively simple formula is a truism that shrouds many problems. All it says is thatthe resistance (including all of its uncertainties) must be greater than the effects of theloads (including all of their uncertainties). The Greek letter f (phi) is a coefficient less thanunity reflecting the degree of uncertainty to which the strength may be overestimated, andthe Greek letter g (gamma) is a coefficient greater than unity reflecting the degree to whichthe various load effects may be underestimated. With traditional ASD, the factors of safetyfor various modes of failure and load combinations varied widely. One basic objective inthe development of the LRFD format was to calibrate the design concept to provide anessentially


uniform "factor of safety" for the differentfailure modes. The safety factor 1.67 that hadproven satisfactory in uncountable applicationsfor the definitive case of a simple span beamsubject to uniform loading was chosen as thetarget value for all cases of failure by yielding. Alarger safety factor target value for the limitstates of fracture was taken as approximately 2.0.A target central overall reliability factor b,encompassing both loads and resistance, wastargeted at approximately 3.0. The result is thatthere are only small differences between theconnections designed by the two procedures,with the more involved LRFD procedureproviding slightly greater economy. The moreimportant advantage of LRFD over ASD is in themore rational formulation, which is suited tological evaluation of the effects that may resultfrom advances in the state of knowledge ofconnection performance and to evaluating the


effect of changes in any of the individualparameters on the load or resistance side or toevaluating any unique design problem.

A Word of Warning: The Specifications aresubject to continuous scrutiny and updating byRCSC and AISC. While they are not legaldocuments in and of themselves, they becomelegal documents when incorporated in municipalbuilding codes and in contract documents.Generally, designers and users must follow thesespecifications if involved with steel structuressuch as buildings or bridges. If you deal withsuch structures, you should obtain the latestversion of these documents from AISC and usethem, rather than this chapter, to direct youractivities. Hopefully, the following discussionwill be helpful in imparting an understanding anduse of the Specifications; but it should not beused as a substitute for the latest version of thedocuments on which this chapter is based.(Information on how to obtain these documents


is included in the references at the end of thechapter.)

This chapter alone may be helpful in the designof nonstructural steel joints and in thedevelopment of similar specifications for otherapplications.

3The Contents of the Specifications

As mentioned earlier, the provisions of the twoversions of the Specifications are identical instating requirements that are independent of thedesign calculation procedure to be used. Thus itis possible to describe both documents with asingle outline.

Section 1. Scope: The ASD and LRFDspecifications relate to the design for strengthand slip resistance of structural steel jointsusing ASTM A325 or A490 bolts orequivalent fasteners, and to the real-world


requirements for use to ensure that reliableperformance is achieved.

Section 2. Bolts, Nuts, Washers, and Paint:In both ASD and LRFD, each component ofthe fastener assembly is prescribed in detail byreference to applicable ASTM specificationsand/or American national standards. Insofar aspaint applied to faying surfaces is concerned,the coefficient of friction of the cured paintcoating is important only for joints specifiedas slip-critical. The Specifications require thatto be used in slip-critical joints, the paint mustbe qualified. A test method to determine theslip coefficient for coating used in boltedjoints, which was developed at the request ofSteel Structures Painting Council and is basedon research done at the University of Texas, isprovided as an appendix to the Specification.

Section 3. Bolted Parts: ASD and LRFDrequire that all material within the grip besteel. No material such as gaskets or


insulation is permitted. Surface conditionsthat affect the serviceability limit state (slipresistance) are specified. Hole sizes andtypesstandard, oversized, short slotted, andlong slottedare defined.

Section 4. Design for Strength of BoltedConnections and Section 5. Design Checkfor Slip Resistance (different for ASD andLRFD) are addressed in detail subsequently.

Section 6. Loads in Combination: ASD andLRFD have different wording but the sameobjective. More liberal stresses are permittedfor the unlikely simultaneous occurrence ofmaximum loads of different types.


Section 7. Design Details: This section ofboth ASD and LRFD specifies that standardholes may be used for any application butstipulates the conditions under whichoversize, short slotted, and long slotted holesmay be used and details how they are to beused when their use is permitted. The sectionalso stipulates the conditions under whichhardened washers must be used and wherethey must be located to prevent galling ofsurface of softer connected material (whichwould cause high and misleading installationtorques). Special attention is given to the useof special 5/16 in. plate washers orcontinuous bars with standard holes tocompletely cover long slotted holes and"bridge the gap" when they occur in an outerply.

Section 8. Installation and Tightening: This


section is identical in ASD and LRFD. Itcovers the requirements for handling andstorage to maintain the bolts and nuts in theclean, lubricated condition required forreliable installation. It specifies that atensionmeasuring device must be on the job tobe used daily where high strength bolts arebeing installed and tightened. (See Chapter42.) It then goes on to establish in detail theprocedure that is to be followed for testingfastener assemblies, using the selectedinstallation procedure to ensure that thespecified installed tension requirements areachieved. Finally, it sets forth in detail theprocedure to be followed for installation andtightening of fasteners in the work using anyone of four recognized methods (turn of thenut, calibrated wrench, alternative designbolts, or direct tension indicating devices).

Section 9. Inspection: Again, ASD and LRFDare identical. The most reliable method for


ensuring the installation to a tension at leastas great as the specified minimum is strictadherence to the recommended installationrequirements of Section 8. Inspection shouldbe directed primarily toward this goal. After-the-fact determination of the installedpretension to any degree of reliability is notpossible with any known economical means.Testing by torque methods is subject to ±30%variation; thus, attempts to verify the tensionin a bolt installed in accordance with specifiedprocedures by torquedependent methods isanalogous to verifying the accuracy of amicrometer using a yardstick. For this reason,the section is primarily directed to inspectorresponsibilities to ensure proper installationin the first place. An arbitration inspectionmethod is specified for use in cases ofcontroversy.

4Design of High Strength Connections


ASD and LRFD provisions in Section 4 aredifferent in terms and formulations, each beingunique to the design philosophy being presented;however, both may be related to the same criteriafor the strength of the fastener or connectedmaterial as justified by the theories and testresults documented in the Guide. The twophilosophies of design lead to comparableconnection designs. LRFD requires the detailedevaluation of the effects of multiple loads (deadloads, occupancy loads, snow loads, wind loads,earthquake loads) in various combinations andeach multiplied by a modifying factor, g, thatvaries according to the combinations of loadsbeing considered. A discussion of the load partof LRFD is beyond the scope of this chapter.Therefore, direct comparison of ASD and LRFDmust be limited to the general statement thatLRFD generally leads to a more economicaldesign and a better understanding on the part ofthe designer of the effect of individual variableson the end result.


on the end result.

The performance of a bolt centered in the hole ofa single bolt connection (low strength, highstrength, fully or only partially tightened) subjectto applied shear loading is showndiagrammatically in Figure 1. Note that as loadis increased from zero, the deformation is smalland proportional to the applied load. In thisregion, the bolt really does not "feel" the appliedload. That is, the load is transferred by frictiondue to tension in the bolt between the boltedparts. At some level of loading, dependent on thecoefficient of friction of the faying surfaces ofthe connected material and the preload in thebolt, a small well-defined displacement will


Figure 1This graph illustrates the performance of a bolt centered in the hole of

a single bolt connection if the bolt is subjected to an applied shear load.The bolt first deforms slightly (1) until the shear load causes major slip

(2). The load transfer between joint members has been by friction up tothis point; now and at higher loads it is transferred by bearing and friction.The bolt then experiences additional elastic deformation (3) until its yield

strength is reached (4). If the shear load continues to increase,the bolt will experience plastic deformation until the bolt breaks (5).

occur, but nothing more. This displacement is due toslippage of one part of the connected materialrelative to the other as the frictional capacity isexceeded. Slippage halts abruptly when the boltcomes into bearing against the sides of the holes. As


the load is increased to higher levels, thedeformations still remain essentially proportional tothe load and on about the same slope as at low loads.In this region, load is probably transferred primarilyby friction but with a little help from small boltbearing forces against the sides of the holetherelative proportions cannot be determined. As theload is increased still further, the proportiontransferred by friction will reduce until eventuallythe connected material of the net section will beginto yield, with consequent reduction in thickness. Atthis point the bolt clamp force reduces to zero, butthe connection still does not fail. Even greater loadsmay be supported at the expense of deformation ofthe bolt and deformation due to bearing pressure onthe connected material around the hole.

This graphical representation of the performance of abolt in a hole is very instructive, and note should betaken of several points. Slip may occur early or latein the loading. The level at which it will occurdepends entirely on the condition of the fayingsurfaces of the connected parts and the clampingforce provided by bolt pretension. Slip is not a signal


of imminent failure. Slip load is far below theultimate strength of the connection and generallysignificantly below the level at which the bolt will becalled upon to resist the load by direct shear alone.The strength of the connection (maximumsupportable load) is totally independent of whetherthe bolt is finger tight or fully tightened. This typicalperformance has given rise to the classification ofjoints loaded in shear as either "slip-critical" (that is,connection strength is not the primary concern; slipcannot be tolerated) or "shear/bearing" (that is, slipwill not occur, or if it does it is notimportantstrength is the primary concern). Becausethe predictable performance of slip-criticalconnections is dependent on a known minimum boltpretension, the terminology provides convenientterms of reference to define the tighteningrequirements for connections that must be fullytightened even though slip per se at a clearly knownvalue of pretension may not be a consideration. Thedesign procedure for both ASD and LRFD is to firstdesign the connection for the particular orientationof applied loading on the basis


of strength and then check the connection for adequacy to prevent slip, if slip into bearingmust be avoided.

4.1Strength Criteria for Fasteners Subject to Applied Tension

The strength of a bolt subject to tension has been determined by theory and test to beaccurately predicted by the product of the ultimate strength of the bolt material (Fu) timesthe stress area of the threaded part of the bolt (Asa). Because different dimensionalproperties of the bolt (body area, tensile stress area, area at root of the threads) are thecritical geometrical parameters for the failure modes for different types of loading, thepertinent geometrical property has been used in the theoretically correct formula for eachfailure mode. Then, for convenience of the user, the resulting formula is multiplied by theratio of the theoretically correct dimensional parameter divided by the full body area, andalso multiplied by the body area to provide formulas for all failure modes in terms of thesingle dimensional property Ag. This procedure provides a sufficiently accurate formulabut avoids the pitfall of inadvertently using the wrong area for a given case. The followingformula is used to express the strength of a bolt subject to tension.

where

Rt =strength of bolt subject to tension (ksi)Fu=

specified minimum tensile strength of the bolt(ksi)

Asa=tension stress area of the bolt (in.2)

Ag=

area corresponding to the nominal diameter of thebolt (in.2)

Because the ratio of tensile stress area to body area is essentially constant at approximately0.75 for bolts with UNC threads of the usual diameter for structural applications (5/8 in.to about 1 1/4 in.), the tensile strength of a bolt in tension may be simply stated as

Thus, the strength of high strength bolts subject to tension reduces to the following:

For ASTM A325 bolts, Rt = 120 × 0.75 = 90 ksi (on the nominal diameter area).


For ASTM A490 bolts, Rt = 150 × 0.75 = 112.5 ksi (on the nominal diameter area).

The ASD specification establishes an allowable stress for A325 bolts in tension as 44 ksi,providing a factor of safety of approximately 90/44 = 2.05 to cover all uncertainties inload estimation, under-strength material, out-of-tolerance manufacturing, etc. Similarly,the allowable stress in tension for A490 bolts is established at 54 ksi, providing a factor ofsafety of approximately 112.5/54 = 2.08.

The LRFD specification provision for bolts subject to tension derives from the sameformula for strength. However, it is termed the nominal strength Rn, and it is required thatit be multiplied by the reduction factor f to cover all of the uncertainties (manufacturingtolerances including dimensions and heat treatment on bolts and nuts, material strength,concentricity of applied load, etc.) that bear on the accuracy with which the strengthformula for Rn predicts the strength.

By means of statistical evaluation of test data, it has been determined that a value for f of0.75 is appropriate. The LRFD specification specifies that


(as appropriate to the particular load case).Because specific requirements of loading casesare beyond the scope of this chapter, a discussionof LRFD is not pursued further for this case.

AISC has carried the process of simplificationforward even further by implementing thespecification requirements in the form ofallowable load tables in kilopounds (kip) forbolts of each size. Similarly, the nominalresistances for LRFD are presented in tabularform for each bolt size.

One important requirement applicable to boltssubject to tension loading that has not beentouched on yet must be discussed briefly.Depending on the arrangement of connected partsand the stiffness of the parts, connections thatsubject the bolt to applied tension most oftentend to deform in a manner that develops pryingforces that increase the tension in the bolt beyond


the level attributable to the externally appliedload. Such prying forces, which are illustrated inFigure 2, may be small or as great as fifty percentof the applied load. Calculations require that apreliminary selection of connected fittingmembers be made and the number and size ofbolts determined. The preliminary selection mustthen be analyzed to estimate the prying inducedin the bolts. The connection must be modifiedaccordingly to account for the increased forcesand again reanalyzed. This can be tedious. Thereader is directed to the AISC manuals of steelconstruction for reliable guidance and tominimize the effort required.

4.2Shear on Fasteners and Bearing on ConnectedMaterialGeneral

In the second introductory paragraph of Section4, the history of events that occur in the loadingof a joint subject to shear was explained tointroduce the difference between design for


strength

Figure 2A bolt subjected to prying must support the external load Lx plus a

reaction force created by the leverlike influence of that portionof the joint which lies outside of the bolt pattern. This phenomenon

can significantly increase the loads borne by the bolts. Becausethe prying force is dependent upon the relative stiffness of the

connected parts and the bolts, the magnitude of this "lever action"is difficult to estimate.

(From Ref. 5.)


and design check for resistance to slip orserviceability. This section is devoted to thediscussion of strength per se of shear/bearingconnections after the slip occurs.

Several interrelated parameters influence theshear and bearing strength of connections. Theseinclude such geometric parameters as the ratio ofnet area to gross area of the connected parts (seeFig. 3), the ratio of the net area of the connectedparts to the total shear-resisting area of thefasteners, and the ratio of transverse fastenerspacing to fastener diameter and to the connectedpart thickness. In addition, the ratio of yieldstrength to tensile strength of the steelcomprising the connected parts and the totaldistance between extreme fasteners measuredparallel to the line of applied tensile force effectconnection strength.

In the past, a balanced design concept was sought


in developing criteria for mechanically fastenedjoints to resist shear between connected parts byhaving fasteners bear against the sides of theholes. This philosophy resulted in widevariations in the factor of safety for the fasteners,because the ratio of yield to tensile strengthincreases significantly with increasingly strongergrades of steel. Also, it had no application at allin the case of very long joints used to transferdirect tension, because the end fasteners''unbutton" before the plate can attain its fullstrength or before the interior fasteners can beloaded to their rated shear capacity.

By using a mathematical model it has beenpossible to study the interrelationship of thepreviously mentioned parameters. It has beenshown that the factor of safety against shearfailure ranged from 3.3 for compact (short)joints to approximately 2.0 for joints with anoverall length in excess of 50 in. It is of interestto note that the longest (and often the most


important) joints had the lowest factor,indicating that a factor of safety of 2.0, as provensatisfactory in service, should be appropriate forall cases.

Figure 3Illustration of joint tearout failure modes and ofthe definition of net and gross cross-sectional

areas of the joint members.(a) End tear-out failures.

(b) The net cross-sectional area is thearea through a line of bolt holes;

the gross area is that of theuninterrupted joint member.


(c) Zigzag failure.(From Ref. 5.)


4.3Strength Criteria for Fasteners Subject to Shear

The determination of the strength of a bolt in shear is based on the observation that theratio of the shear strength through the body area of a single high strength bolt to the tensilestrength of the bolt is 0.62. Joints involving only a single bolt are rare, and it has beenobserved that with more than two bolts in the line of force the deformations of theconnected material in bearing against the bolt and along the length of the joint causenonuniform distribution of force to the bolts, with the bolts at the ends of the jointcarrying a higher portion of the load than those at the center (see Fig. 4). The strength ofthe connection in terms of the average force on the bolts decreases as the joint lengthincreases. At the failure load of a joint with length of about 50 in., the average force on abolt is about 0.80 times the strength of a single bolt in shear. To avoid the need for areduction formula based on joint length, a single reduction factor of 0.80 has been appliedto the 0.62 factor. This permits the development of a formula that is applicable to joints upto 50 in. long that is not overly conservative for shorter joints. It must be noted that forjoints longer than 50 in., the Specification requires that an additional 0.80 reduction beapplied. Further, it has been observed in tests that the average value of the strength of abolt loaded in shear through the threads is 0.833 times the strength of the same bolt loadedthrough the body area (see Fig. 5). In the specification, a conservative value of 0.80 hasbeen applied to account for the shear strength of a fastener loaded through the threads interms of the body area. Thus, the strength of a high strength bolt in shear/bearing inconnections up to 50 in. long may thus be reliably determined by the following formulas.

With threads not excluded from the shear plane,

With threads excluded from the shear plane,

The ASD specification establishes allowable stresses for bolts loaded in shear. They arepresented in Table 1 together with the strength as calculated above and with the apparentfactors of safety provided.

The LRFD specification uses the same formulation for the nominal strength, Rn, for boltssubject to shear but stipulates that Rn shall be multiplied by a reduction factor f = 0.75 tocover uncertainties in the determination of the resistance. For reasons explained above, thefurther treatment of the LRFD specification is beyond the scope of this chapter.


4.4Design Criteria for Combined Tension and Shear

The adequacy of bolts subject to tension combined with shear may be checked by means ofthe elliptical interaction formula

Figure 4The bolts in a long joint loaded in shear do not see equal shear loads.

Those at the outer ends of the bolt pattern are subjected to higher loadsthan those in the center of the pattern due to build up of stress in the

plates along the length from opposite ends.(From Ref. 5.)


Figure 5The shear strength of the bolt is reduced if bolt threads coincide with

one or more of the shear planes of the joint, as they do here in examplesA and B. The number of shear planes shown varies from one (A) to two

(B and C).(From Ref. 5.)

where

fv =the applied shear stress (ksi)Fv=the allowable shear stress (ksi)

ft =the applied tensile stress (ksi)

Ft =the allowable tensile stress(ksi)

This concept is illustrated in Figure 6.TABLE 1 Shear Strength Versus Allowable Stress for Bolts Loaded in ShearCase description Shear

strengthAllowable

stressFactor of

safetyA325: Threads not excluded from shearplane 48 ksi 21 ksi 2.29

A325: Threads excluded from shearplane 60 ksi 30 ksi 2.0

A490: Threads not excluded from shearplane 60 ksi 28 ksi 2.14

A490: Threads excluded from shearplane 75 ksi 40 ksi 1.87


Figure 6An elliptic curve can be used to relate the tensile capacity

of a bolt to the shear stress imposed simultaneously on it, or viceversa. ST is the ratio of the shear stress in the bolt to the

ultimate tensile strength of the bolt. TT is the ratio of tensilestress in the bolt to the same UTS. G is the ratio of shearstrength to tensile strength of the bolt material. The curveshown here is for static loads on the thread stress area of

ASTM A325 or A354 BD bolts.(Modified from Ref. 2.)

4.5


Strength Criteria for Connected MaterialSubject to Bearing Stress

Bearing stress produced by a high strength boltpressing against the side of the hole in aconnected part is important only as an index tobehavior of the connected part. It is of nosignificance to the bolt. The critical value can bederived from the case of a single bolt at the endof a tension member. (See Fig. 7.)

It has been shown, using finger-tight bolts, that aconnected plate will not fail by tearing throughthe free edge of the material if the followingrelationship is satisfied.


Figure 7Many of the design rules discussed in this

chapter are based on the assumption that theedge distance, Lc, from centerline of the boltto the nearest edge of the joint in the directionof the applied load is at least 1.5d. To tear outof the joint, such a bolt would have to shear

the joint area illustrated here (i.e., 2tLc).


where

Rp=nominal bearing pressure (ksi)

Fu=specified minimum tensile strength of the connected part (ksi)

L=

distance measured parallel to the line of applied force from centerof a bolt to the free edge of the connnected material (in.)d=nominal diameter of the bolt (in.)

Although Eq. (8) has been proved by test to accurately predict strength, practicalconsiderations must be observed, such as providing reasonable lower limits on edgedistance and spacing to rule out excessive hole ovalization with excessively large edgedistances and spacing as well as a lower limit to rule out ridiculously low allowablestresses at impractically small end distances. By rearranging Eq. (8) and providingjudgment-based practical limits, criteria with a factor of safety of 2.0 have been providedin the ASD specification.

1. Connections using high strength bolts in slotted holes with load applied in a directionother than approximately normal (between 80° and 100°) to the axis of the hole andconnections in oversize holes shall be designed for resistance against slip.

2. Tabulated values (Table 2) apply when the distance L parallel to the line of force fromthe center of the hole to the edge of the connected part is not less than 1.5d and the distancefrom the center of a bolt to the center of an adjacent bolt is not less than 3d (see Fig. 7).When either of these conditions is not satisfied, the allowable stress shall be determinedaccording to the equation.

3. For connections with more than a single bolt in the direction of force, the resistance maybe taken as the sum of the resistances of the individual bolts.

The absence of allowable stress or strength provisions specifically for the case where a boltin double shear has a nonthreaded shank in one shear plane and a threaded section in the


other shear plane (Fig. 5) is for purposes of simplification and because the impliedaccuracy would be inconsistent with practical experience and with test data relative to theperformance of such connections. It also recognizes that knowledge of bolt placement(which might leave both shear planes in the threaded section in thin material) is notordinarily available to the detailer. If threads occur in one shear plane, the conservativeassumption is made that threads are in all shear planes.TABLE 2 Allowable Working Stress on Fasterners Based on ConnectedMaterialLoading condition Allowable

stressa (ksi)Bearing on connected material with single bolt in line of force ina standard or short slotted hole 1.0 Fu

Bearing on connected material with two or more bolts in line offorce in a standard or short slotted hole 1.2 Fu

Bearing on connected material in long slotted holes 1.0 Fu


4.4Design Check for Slip Resistance

As discussed in the introductory paragraph of Section 4, a small amount of slip will occurin shear-loaded connections prior to reaching the maximum strength load. (This may not betrue for the case of failure by bearing against connected material with very small edgedistances.) Slip may also occur prior to reaching the allowable load (maximum strengthdivided by a factor of safety) as determined by strength considerations. In a greatpreponderance of connections, strength is the primary concern governing the design or slipcannot occur because in the real world joint assembly procedures generally ensure that thebolts will be in bearing with either the dead weight of the piece bearing on the looseerection bolts prior to final tightening or to the application of any other load or where, dueto numerous bolts in large connections and the in-tolerance misalignments of holes, boltsare in bearing at all stages of loading. Reports of slip at design loads have rarely if everbeen reported; however, there are situations where a high degree of assurance is requiredthat slip will not occur. Slip must be prevented to avoid even minor changes in thegeometrical relationship between members to ensure the integrity of the structure subjectunder static loads or to ensure that back-and-forth slippage of connections does not occurwhen they are subjected to significant reversal of loading or for other reasons. Suchconnections are advisedly identified in the Specification as "slip-critical" connections tocall attention to the fact that they are intended to be used where special performancecharacteristics are important and to emphasize that their installation and tensioning requirespecial control and inspection at additional cost. Joints considered to be slip-critical arethose subject to significant load reversal or to cyclic fatigue loading, those with boltsinstalled in oversize holes, those with bolts installed in slotted holes when the line of actionof the applied loads is more or less parallel to the long axis of the slots, those in whichbolts and welds share the applied loads, and those that the engineer of record thinks wouldbe adversely affected by slip.

For situations where slip into bearing must be prevented in order to ensure the connection'ssuitability for service (slip-critical connections), the Specification requires that, in additionto satisfying the requirements for strength, the slip resistance of the connection be checked.The allowable capacity to resist slip must be greater than the force tending to cause slip atworking loads. The allowable slip resistance is expressed by the formula

where

Fsallowable load unit area of bolt based on slip


=resistance (Ksi)Ab=

area corresponding to nominal body area of bolt in2).

Nb=number of bolts in the connection

Ns=number of slip planes

The allowable slip load unit area of the bolt, Fs, is given in Table 3 for A325 and A490bolts, for three classes of connected faying surface treatments and various bolt sizes. (It istrue from basic principles that slip load is solely dependent on coefficient of friction andclamping force; however, out of concern for the ability to maintain installed clamping forcewhen the high clamping force is applied to a reduced area at oversize and slotted holes, thespecificationallowable loads at such holes is reduced.) For A325 bolts, the allowable sliploads vary by 1028 ksi, with the highest allowed for bolts in standard holes with Class Bsurfaces and the lowest assigned to bolts in long slotted holes parallel to the direction ofloading with Class A surfaces. A490 bolts are treated in a similar manner, with allowableslip loads Fs ranging from 13 to 34 ksi.

In contrast to earlier versions of the Specification, which recognized a wide variety ofsurface treatments, and to prevent the chaos that could result from unlimited smalldifferences in coefficients of friction claimed for various proprietary surface treatments,only three coef-


TABLE 3 Allowable Slip Loads per Unit of Bolt Area for Slip-Critical Connections(Data in ksi)

Hole type and direction of application of loadAny direction

Standard Oversize Transverse, long slots Parallel, longslots

Surface condition of joint membersa A325A490A325A490 A325 A490 A325 A490Class A 17 21 15 18 12 15 10 13Class B 28 34 24 29 20 24 17 20Class C 22 27 19 23 16 19 14 16aClass A: Slip coefficient of 0.33 is expected if the joint surface consists of clean mill scale orblast-cleaned surface to which Class A coatings have been applied. Class A coatings are thosethat carefully defined tests show will provide a slip coefficient not less than 0.33.Class B: Slip coefficient of 0.50; blast-cleaned surfaces or blast-cleaned surfaces with Class Bcoatings (which tests show will provide a coefficient not less than 0.5).Class C: Slip coefficient of 0.40; hot dip galvanized and roughened surfaces.Note: Joint surfaces must be blast cleaned before Class A or B coatings are applied.

ficients of friction are recognized by the current Specification, eachcorresponding to a group or class of surface treatment.

Class ASlip coefficient 0.33: Appropriate to clean mill scale jointcontact surfaces or blast-cleaned surfaces to which a Class Acoating has been applied

Class BSlip coefficient 0.50: Appropriate to uncoated blast-cleanedsurfaces or blastcleaned surfaces to which a Class B coating hasbeen applied

Class CSlip coefficient 0.40: Appropriate to hot-dip galvanizedsurfaces that have been roughened

Note: Coatings for Class A and Class B surfaces must have beendemonstrated by test to provide a coefficient of 0.33 or 0.50,


respectively and must be applied only to blast-cleaned surfaces.

An extensive appendix to the Specification describes the tests to beused to determine the slip coefficient of a coating. The appendixstipulates long-term creep tests as well as shortterm tests, such thingsas thickness of coating and curing time, and complete documentationto enable reproducibility in application.

A good practice followed by experienced designers and detailers onprojects where some bolts are non-slip-critical and others are slip-critical or require full tightening for other reasons is to design all slip-critical connections using bolts of a single diameter that is differentfrom that of all other bolts used in the job. In this way all bolts thatrequire special tightening and inspection are clearly identified on thedrawings and in the material delivered to the job by their diameter. Thedrawings do not carry special notations for individual connections.The ironworkers and inspectors need not refer to drawings for specialinstructions.

4.5Combined Axial Tension and Shear in Slip-Critical Joints

Consideration must be given to the situation in which the externallyapplied load is applied in such manner that it reduces the clampingforce in the connection that is provided by the bolt pretension. Asillustrated in Figure 8, the oblique load in the brace is resisted by axialtension


Figure 8An oblique load of the sort shown

here is resisted by axial tension in thebolts as well as by a shearing force atthe faying surface. The tension in thebolts is not significantly increased bysuch a load, but the clamping force

(and therefore slip resistance)between joint members is reduced.

in the bolts as well as by a shearing force at the fayingsurface. The axial tension from the external load doesnot increase the tension in the bolt to a level higher thanthe installed pretension; rather it reduces the clampingforce at the faying surfaces, which are counted on toprovide resistance to slip. That is, a portion of theclamping force that was provided by bolt pretension is


taken over by the externally applied axial tension force.For such cases, the allowable design slip load, Ps [Eq.(10)], of the connection must be reduced to the degreethat the clamping force at the faying surface is reducedby multiplying it by the factor

where

Tu=

axial force per bolt induced by the externally appliedload (kip)

Tm=specified installed prestress per bolt (kip)

The reduction in allowable slip load does not apply to allcases where the clamping force may be reduced aroundindividual bolts by the externally applied load. Asillustrated by Figure 9, when the externally applied loadinduces a bending moment as well as shear (a tendencyto slip), the reduction in clamping force due to theapplied load at the top bolts is exactly matched by anincrease in the compression forces around the bottombolts. The slip resistance is not affected.


5Slip Resistance Per LRFD

Before discussing the LRFD requirements for slipresistance it should be repeated that it is not possible topresent a truly usable summary of LRFD here becausethe load functions required for the calculations are wellbeyond the scope of this chapter. It might also be notedthat there is considerable resistance to the LRFDspecification by practicing engineers. They claim that thework required to deal with the load aspects of thisspecification is excessive. They cannot see the merit oradvantage of spending the money required to changetheir current (often computerized) design procedures toachieve marginal economies for their customers. Theacademic community, however, believes that LRFD is asignificant step forward and resists any further use ordevelopment of the ASD specification. Since thesepeople have considerable influence and may indeedsucceed in replacing ASD with LRFD, it is felt that theintroduction


Figure 9Combined axial and tensile loads do

not always result in a loss of slipresistance. In the joint shown here,

for example, any reduction in clampingforce at the top bolts is matched by an

equal increase around the bottombolts, thanks to the bending moment

on the beam.

to LRFD contained in this chapter, even if unusable on a stand-alone basis, is necessary.With all that in mind, we now take a look at the LRFD approach to slip resistance.

As in the ASD version, the LRFD specification distinguishes between slip-critical andshear/bearing connections, with slip resistance being of concern only for the first. Onceagain the frictional slip coefficient of the faying surfaces must be controlled; this is definedby three classes: A, B, and C, with coefficients of 0.33, 0.50, and 0.40, respectively, asdefined in Section 4.4. And once again bolts must be installed and tightened to or above theminimum tensions listed in Table 1, following one of the tightening procedures recognizedby both versions of this specification (and discussed in Section 6).

5.1Slip Considerations for Nominal Load Levels

The LRFD specification allows two different design criteria for slip-critical joints. In thefirst, nominal (service) loads are the design criterion and must not exceed the design slipresistance, which is given as fRs, where

The factor D is the slip probability factor.

D=0.81 for Class A surfaces (m = 0.33)


D=

0.86 for Class B or C surfaces (m = 0.50or 0.40)

The factor m is the coefficient of friction for Class A, B, or C surfaces (namely 0.33, 0.50,or 0.40) or as determined by the experiments discussed in Section 4.4.

This time the factor f can take one of several values.

f=1.0 for standard holes

f=0.85 for oversized or short slotted holes

f=

0.70 for long slotted holes transverse to the directionof the load

f=

0.60 for long slotted holes parallel to the direction ofthe load

This is not quite the end of the nominal load story. Modifications to Eq. (11) are oftenrequired, as discussed, for example, in Section 5.3, below. First, though, let us look at thesecond approach to slip resistance outlined in the LRFD specification.


5.2Slip Considerations at Factored Load Levels

When factored loads are the design criterion, they must not exceed the design slipresistance, which is defined as fRst, where

If you run calculations comparing the results obtained with Eq. (12) to those obtained byuse of the equivalent expressions in the ASD specification, you will find, we think, thatload and resistance factor design is somewhat less conservative than allowable stressdesign. This is why the above-mentioned "economies for their customers" can be achievedby use of LRFD.

5.3Modifications to Accommodate Tensile Loads on the Joint

If a tensile load, illustrated in Figure 10 (i.e., a load acting more or less parallel to the axesof the bolts) is applied to a joint, it tends to pull the joint apart, partially reducing theinterface clamping force the bolts had created between joint members. This phenomenon isdiscussed at length in Chapter 42. It is the clamping force, however, that generates thefriction forces that resist joint slip. As a result it is important to take tensile loads, if any,into account when estimating the design slip resistance of the joint. The rather lengthyprocedures for doing this are given in the LRFD specification.

6Assembly and Tightening

As already mentioned, both ASD and LRFD versions of the Specification incorporate thesame criteria for the installation and tightening of high strength bolts. The resulting rulesdepend on whether or not the joint is loaded in tension or shear, whether it is subject tocyclic (fatigue) loading, and whether a shear-loaded joint is considered to be slip-critical.Although different


Figure 10The upper sketch shows a joint loaded in axial tension, the lineof action of the load being parallel to the axes of the bolts. The

lower sketch shows a joint loaded in shear, with the line ofaction of the load perpendicular to the axes of the bolts.

(From Ref. 5.)


TABLE 4 Minimum Fastener Tension (Tm)Required for Slip-Critical Connections andConnections Subject to Direct Tension Loads

Minimum tension (kip)Nominal bolt size (in.) A325 bolts A490 bolts1/2 12 155/8 19 243/4 28 357/8 39 491 51 641 1/8 56 801 1/4 71 1021 3/8 85 1211 1/2 103 148

tightening requirements might have beenjustified for each type of loading or applicationthat requires tightening, the single criterion isapplied to all joints requiring tightening tominimize opportunities for errors in the field.The pretension required to be developed in thetightening procedure is that upon which theallowable loads for slip-critical connections are


based. The minimum installed fastenerpretensions for ASTM A325 and A490 bolts ofdifferent sizes are as specified in thespecification (see Table 4).

Bolts in connections not classified as slip-criticalor subject to tensile loads or cyclic fatigueloading need to be tightened only to a ''snugtight" condition. This degree of tightness isdefined as "tightened to the degree necessary tobring all joint members into firm contact." TheSpecification suggests that snug tightness cannormally be achieved by a few impacts of animpact wrench or by the full effort of a manusing an ordinary spud wrench.

In contrast to the rules for snug-tightened joints,the installation requirements to ensure that theminimum pretensions equal to those specified inTable 4 (70% of the ultimate tensile strength ofthe bolts) are detailed in the requirements forproper receipt inspection, storage, job site testingcapacity of the fastener assembly, as it will be


used in the work, to achieve a tension 5% higherthan required by the Specification and the use ofone of four recognized methods for installationand tightening of bolts in the work.

It is informative to learn why 70% of ultimatestrength was established as the requiredinstallation tension. Early versions of theSpecification were limited to A325 bolts, andtightening was required to achieve the specifiedproof load of the bolt. Proof load is defined inthe ASTM specifications basically as theminimum required yield point as determined in astandard direct tension test (generally taken as thestress at 0.2% offset).

The Specification was extended to include thehigher strength A490 bolts in 1964, and in theversion of that adoption date the installationtension requirement was established at the proofload. Proof load of the higher strength lessductile A490 bolts is specified by ASTMspecifications to be 80% of the ultimate tensile


strength and is also determined in a standarddirect tension test. In the practical applicationand use of bolts, however, the tension is inducedby torquing a nut on the threaded end of the bolt,subjecting the area at the root of the threads tocombined torsion and tension stresses. Thefailure load of bolts subject to combined torsionand tension is approximately 20% less than thefailure load in direct tension as suggested inFigure 11; thus twist-off failures were soonencountered when bolts were being tensioned tomeet the Specification's pretension requirements.Additionally, to minimize rejection rate,manufacturers must target production at, say,10% above the minimum specified acceptancelevel.


Figure 11A fastener tightened with a torque tool, and therefore exposed to simultaneoustorsion and tension, will yield at a slightly lower level of tensile stress (A) than

will a fastener loaded in pure tension. This same fastener, as a result, will supportan additional tensile load in service because the torsional stress component will,

in general, disappear rapidly after the wrench is removed. Data are for a 7/8A325 bolt with a 4 1/8 grip length.

(Adapted from Ref. 2.)

Thus the actual yield point approaches the ultimate tensilestrength. This increased the potential for failures inservice due to hydrogen embrittlement and stresscorrosion cracking. To address these problems the


specified minimum proof load in the ASTM specificationfor A490 bolts was reduced by 10 ksi, and the RCSCinstallation pretension requirement was changed fromproof load to 70% of the ultimate tensile strength, whichjust happens to be the percentage of ultimate strength forproof load with A325 bolts.

6.1Preinstallation Check Testing

Specifying a minimum installation based on requirementsto satisfy in-service fastener performance is easy. Anassembly comprising a bolt, nut, washer, and connectedmaterial is deceptively simple, and for most applicationstightening the assembly is a simple, straightforwardoperation. However, when the tightening of the assemblymust be controlled to ensure a specified minimumtension, the control becomes very difficult. It haschallenged the best and most experienced minds and stilldoes. Literally hundreds of variables affect the results ofdeveloping tension in a bolt by tightening the nut. ASTMand ANSI specifications establish geometrical, chemical,and physical property tolerances on the separatecomponents (bolts, nuts, and washers) that are presumablyright and proper for service; however, the user is notinterested in a bolt or a nut or a washer. The only thing


that is of importance is the complete assembly (includinglubrication and connected material). The cumulativeeffect of tolerances on the individual parts, whether theywere produced with new tooling or tools about to beretired; the conditions of handling and storage in thewarehouse, during shipment, and at the job site; theadequacy and condition of the installation tools; theknowledge and diligence of the workmen; and many otherfactors all bear on the problem. Even in controlledlaboratory environments, it has been demonstrated thattorquetension relationships vary by as much as ±3040%due to the effect of the numerous uncontrolled variables.Because it is not possible to adequately control thesenumerous individual variables, the Council specificationrequires that their cumulative effect be compensated forby a testing and calibration procedure prior tocommencement of work and repeated daily when theinstallation method might be dependent on tool operatingcondition, using the fastener components from each lotassembled as they will be assembled in the work and usingthe method that will be used by the crew that will do the


work. The specified demonstration testing offastener assemblies under the real conditions thatwill be used is essential to reliable results. (SeeChapter 52 for further discussion.)

One point concerning which questions have beenraised is the fact that the specified procedurerequires testing to a level of tension 5% higherthan the specified minimum pretension level. Ithas been argued that this does not constitute afair and enforceable requirement because theminimum level for acceptance of the assembly byfield test is 5% higher than the ASTM specifiedminimum proof load for A325 bolts. Theargument is not valid. Rarely does the testidentify bolts or nuts that do not comply withASTM minimum requirements, but when it does,that's good: Such fasteners can be rejected beforethey are incorporated in the work rather than laterwhen they are identified by postinstallation


inspection. The proof load for high strength boltsis comfortably lower than the ultimate tensilestrength; thus proper bolts and nuts can easilyachieve loads 5% above proof load. What thetest does identify is improper use of the selectedinstallation method, mismatch of componentsfrom different sources (although eachcomponent might be within tolerance for theindividual part, they may not work welltogether), fasteners that have becomecontaminated by dust or rust from improperstorage, dry or improperly lubricated fastenersthat need relubrication to overcome the loweringof developed tensile strength when tension isinduced by tightening the nut, overtapping ofnuts in the case of galvanized fasteners, and otherconditions. The pretesting of fastener assembliesand installation method is the critical element inensuring proper tension in the work. It must notbe given short shrift.

6.2


TighteningGeneral

Inspection testing to ensure that the fastenersmeet requirements and the components will worktogether as an assembly is required for fastenerassemblies that are to be tightened in slipcriticaljoints. Further testing of fastener assemblies isnot specified in great detail. However, forassemblies requiring full tensioning, theSpecification recognizes four tighteningmethods, any one of which will ensure that thespecified tension is achieved if it is properlyused. It can be said with equal assurance thatunsatisfactory results and unknown tensions willresult if any of the methods is misused.

6.3Turn-of-Nut Tightening

Once the job site demonstration testing hasshown that the fastener assemblies and procedureto be used will induce a tension 5% higher thanthe specified minimum tension, installation of


fasteners may proceed. The method requires thatbolts be installed in all holes of the connectionand tightened to the snug tight condition startingwith the fasteners in the most rigid area of theconnection and proceeding systematicallyoutward to the free edges. Generally one or twoimpacts of an impact wrench will signal that abolt is snug; however, with large joints in thickmaterial more than a single cycle of snugglingbolts from the center outward may be required tobring all bolts to the snug condition. Once allbolts are essentially uniformly snug, finaltightening may begin. Again starting with thebolts in the most rigid part of the connection andproceeding outward systematically to the freeedges, the nuts (or bolt heads) are given not lessthan the required number of turns (see Table 5).In the case of large joints in thick material, it maybe necessary to "touch up" the bolts that weretightened first as they may have been loosened bytightening of adjacent bolts.


The turn-of-nut method has been shownrepeatedly to be one of the most reliablemethods, when diligently applied, for attainingthe required pretension in the bolts. This is truebecause none of the variables that affect torque-dependent methods affect the turn-of-nut methodto the same degree. The tension developed in thebolt depends on the amount of bolt extensionfrom the snug tight condition. Therein, however,also lies the source of inaccuracy of resultsachieved by the method. Care must be exercisedto ensure that the connected material is in contactthroughout the area of the joint and that no boltsare loose at snug tight. Care


TABLE 5 Nut Rotation from Snug Tight Condition for Turn-of-NutProcedurea,b

Disposition of outer face of bolted partsBolt length(underside ofhead to endof bolt)

Bothfaces

normalto boltaxis

One face normal to boltaxis and other sloped notmore than 1:20 (beveled

washer not used)

Both faces sloped notmore than 1:20 from

normal to the bolt axis(beveled washer not used)

Up to andincluding 4diameters

1/3 turn 1/2 turn 2/3 turn

Over 4diametersbut notexceeding 8diam.

1/2 turn 2/3 turn 5/6 turn

Over 8diametersbut notexceeding12 diam.

2/3 turn 5/6 turn 1 turn

aNut rotation is relative to bolt regardless of the element (nut or bolt) beingturned. For bolts installed by 1/2 turn and less, the tolerance should be±30°; for bolts installed by 2/3 turn and more, the tolerance should be ±45°.bApplicable only to connections in which all material within the grip of thebolt is steel.Source: Refs. 1 and 4.


must be exercised to apply required turns from themost rigid area to the free edges. Because the turnsspecified are adequate to bring the bolts to slightlymore than the specified minimum, they will be slightlypast the knee of the turnstension curve (Fig. 12) wheresmall differences in the initial snug-tightness tension ofindividual bolts make even less difference in finaltension. The main source of abuse and inaccuracy withthe method is haste and lack of care on the part of theworkmen. A frequently occurring abuse is to put boltsin all holes and then tighten them in any convenientorder to final tightness, trying to note orientation of thechuck on the fly as it begins to impact and then addingan estimated number of turns. This procedure willmost assuredly render the first installed bolts loose tosome unknown degree when adjacent bolts are latertightened. Another disadvantage of turn-of-nut, evenwhen diligently applied, is that the inspector mustmonitor the work in progress to be able to assure theowner that the method was properly applied. He cannotcome to the work after completion at his convenienceand ascertain that the work was properly done.

6.4


Calibrated Wrench Tightening

The calibrated wrench tightening method is directlydependent on the torque-tension relationships of thefastener assemblies. Therefore, all of the numerousvariables that affect the torque required to developtension in the bolts by torquing the nut also affect thefinal result. Daily demonstration testing of the systemis essential and must involve a representative sample ofeach lot, length, diameter, and grade of fastenerassembly as they will be used on the job in a devicecapable of indicating tension in the bolt, using theactual tightening equipment that will be used. (SeeChapter 42.) Retesting and readjustment of the systemis required if any significant change in condition of thefasteners is noticed. In the testing, the requiredminimum tension must be developed withoutexceeding the bolt-head-nut rotation from snug tightspec-


Figure 12Illustration of the forceelongation curve generated when a bolt is tighteneduntil it breaks. At first the bolt deforms elastically; tension builds up in it,

but there is little axial deformation. Finally the bolt yields, at the point shownby the knee of the curve. Any attempt to tighten it further is accompanied

by significant deformation, until it breaks. In the turn-of-nut procedurethe bolt is first tightened to a minimum of 70% of its ultimate strength.This brings it to "proofload," which is just below the knee of the curve.

The following, specified, turn of the nut (typically half a turn) will bring the boltslightly farther into the inelastic region. As you can see from the curve, however,

there is a large margin of safety (equal to about three turns) against overtightening.(From Ref. 5.)

ified in Table 5. If manual torque wrenches are to be used,the torque must be measured while the nut is being turned


in the tightening direction. These are demanding butnecessary requirements that must be met on a daily basis ifreliable results are to be achieved.

It must be noted that the Council does not recognize anytable of standard torque values or any formulas for thecalculation of required torque for installation orinspection. Such tables and formulas are based onconsideration of ideal thread helix angle, bolt diameter,and thread pitch but totally ignore the many variables thataffect the results in the real world. They are rejected asunreliable and misleading.

Once the job site demonstration testing has shown that thefastener assemblies and procedure to be used will induce atension 5% higher than the specified minimum tension,installation of fasteners in the work may proceed. Boltsshall be installed in all holes of the connections withhardened washers under the element to be turned intightening, and all bolts brought to the snug tightcondition. Snug tightening is required to start at the mostrigid part of the connection and proceed systematically tothe free edges until all bolts are uniformly snug tight andthe mating parts of the joint are in firm but not necessarilycontinuous contact. Following this initial tightening,which may be accomplished using any wrench, thecalibrated wrench shall be used in a systematic manner


from the most rigid part of the connection to the freeedges. The element not turned in tightening must be held toprevent rotation during final tightening operation and theturns observed to ensure that the threads are not stripped ordamaged. More than a single cycle of tightening may berequired to ensure that all the bolts are simultaneously atthe same proven torque level.


6.5Alternative Design Fastener Installation

Alternative design fasteners are those that meetall the geometrical, chemical, and physicalproperties specified in ASTM A325 or A490specifications as appropriate but have someadditional feature such as a twist-off splined endand special installation wrenches to establish aspecific level of torque at which the splined endwill shear off and the tool will cease operation.Although such fasteners are in fact torque-controlled fasteners and therefore subject to thesame effects of the many variables that affect theperformance of conventional high strength boltsinstalled by torque-controlled methods, they havea good track record and despite high cost areproving popular in the marketplace. This isprobably because they are generally distributedthrough the producer's distribution


representatives rather than through brokers whosupply fastener components from many sources,because they are probably produced underconditions of greater quality control, becausethey are supplied as an assembly with the boltand nut under the control of the same producer,and because they are probably lubricated, stored,and shipped under better control. (Theirperformance is strongly dependent on thelubrication provided.) Even so, they are subjectto the same preinstallation demonstration testingby lot, diameter, and length, to test their ability toprovide a tension 5% higher than the specifiedinstalled tension requirements. If they fail to passthe test, the problem is usually one of failure tomaintain proper cleanliness and lubrication andcan often be corrected by cleaning andrelubrication in accordance with the producer'srequirements.

Alternative fastener installation can be abusedand the good performance characteristics


defeated by installing bolts in the holes andcarrying the tensioning forward in a singlecontinuous operation. Only if the prescribedprocedure is faithfully followed will the absenceof the splined end give witness to a properlytensioned bolt. Only if the inspector monitorswork in progress to ensure that the correctprocedure is routinely used may it be assumedthat the specified tension has been installed.

6.6Direct Tension Indicator Devices

The typical direct tension indicator (DTI) isrepresented by devices that meet the requirementof ASTM A959 although others are notprecluded. (See Chapter 33.) Although suchdevices do not provide the specified installedtension, they do constitute an integral part of thefastener assembly and are subject to the samedemonstration testing for each lot, diameter, andlength of fastener being used. It is sometimesargued that since they measure only the tensile


stress developed in the bolt, repeated testing withfasteners of different lengths is excessive. This isnot deemed to be a valid argument because theobjective is to ensure that the fastener assembliesdevelop a tension 5% higher than the minimumspecified tension. Instances have beenencountered wherein DTIs that were targeted at astrength level slightly higher than the specifiedminimum so as to minimize the reject rate wereactually as much as 1520% over strength. Inapplication such DTIs may be combined withbolts that are close to the minimum specifiedstrength by direct tension testing. Whentensioned in the joint by tightening the nut, thecombined torque-tension effects may cause boltfailure because of the over-strength DTI. In suchcases neither the bolts nor the DTIs should besummarily rejected; rather the solution is to befound in improving the fastener lubrication.

Once the job site demonstration testing hasshown that the fastener assemblies including the


DTIs, placed as they will be installed in the workand with the procedure to be used, will induce atension 5% higher than the specified minimumtension, installation of fasteners in the work mayproceed. In the work bolts with properlypositioned DTIs must be installed in all holes ofthe connections and brought to the snug tightcondition by the same systematic tightening fromthe most rigid area to the free edges but with carebeing taken to only partially compress theindicator protrusions. Final tightening mayproceed when the mating parts are in firm contactand all bolts are uniformly tight. Tighteningmust, again, be accomplished by


tightening in a systematic manner from the most rigid area to the freeedges. In large connections, tightening should not be carried forwardto final gap dimension in a single operation; connections may requiremore than a single cycle of tightening. Final compression of the DTIsto the specified gap should be accomplished after all bolts areuniformly tight.

Flat hardened washers must be placed under DTIs when the DTIs areinstalled in oversized or slotted holes in an exterior ply to avoid falseindication of protrusion compression by dishing of the DTI into thelarge hole.

Direct tension indicator installation can be abused and the propertensioning of the joints defeated by installing bolts in the holes andcarrying the tensioning forward in a single continuous operation. Onlyif the prescribed procedure is faithfully followed will the specified gapgive witness to a properly tensioned bolt. The protrusions do notrebound once compressed if bolt tension is reduced; thus thetightening must be carried out in a manner that minimizes relaxationof previously tightened bolts. If a DTI is compressed to the final gaprequirement before all bolts are uniformly tensioned, it should beremoved and replaced. Only if the inspector monitors work in progressto ensure that the correct procedure is routinely used may it beassumed that the specified tension has been created.

7Inspection Requirements

Bolts in slip-critical connections or in connections that must supportdirect tension loads are subject to inspection, and such joints must be


clearly identified on the working drawings.

Ideally the inspector will be present at the job site to witnesscalibration tests, installation, and the actual tightening of some of thebolts. If not, or if the tightening of the bolts is challenged, theinspector must conduct a test in which a representative sample of fivebolts of each diameter, length, and grade used on the job are tested inthe tension-measuring calibrator described earlier and in Chapter 42.The bolts shall be first snugged to about 15% of the value shown inTable 4, then further tightened to the Table 4 amount. The torquesused to create the final tension are recorded, and the middle threevalues are averaged to define a "job inspection torque," which is thenapplied to a small sample of the bolts that have been challenged. Ifnone of them turn, the inspection is complete. If any one of them turnsunder the job inspection torque, all bolts shall be so tested.

Bolts that have been installed some considerable time before beingchallenged cannot be accurately inspected by the methods justdescribed. Alternative test means must be defined and agreed upon. Ina few cases the ultrasonic equipment and techniques discussed inChapter 34 have been used successfully for this purpose.

Symbols and Units

AbCross-sectional area of bolt corresponding to nominal diameter(in.2)

AgCross-sectional area of bolt corresponding to nominal diameter(in.2)

AsaTensile stress area of the threaded region of a bolt (in.2)DSlip probability factor (dimensionless)DAnticipated dead load on structure (kip)dBolt diameter (in.)


FpAllowable bearing pressure on a bolt (ksi)FsAllowable slip load per unit area of bolt (ksi)FtAllowable tensile stress (ksi)ftApplied tensile lstress (ksi)FuSpecified minimum tensile strength of connected part (ksi)FvAllowable shear stress (ksi)fvApplied shear stress (ksi)


LAnticipated live load on a structure (kip)

L

End distance from center of the nearest round bolt hole (ortransverse slotted hole) to the edge of connected part, in thedirection of the applied load; also the center-to-center distancebetween standard holes or transverse slots (in.)

NbNumber of bolts in jointNsNumber of slip planes in jointPsAllowable slip resistance of a joint loaded in shear (kip)

RnNominal tensile strength of a bolt subjected to axial tension orshear; or to a combination of tension and shear (kip)

RnNominal bearing pressure exerted by bolt of joint or the bearingresistance of the connected material (ksi)

RsNominal slip resistance of the bolts in a joint for use at nominalloads (kip)

RstSlip resistance of the bolts in a joint for use at factored load levels(kip)

RtStrength of bolt subject to tension and shear (kip)RvStrength of bolt in a shear/bearing connection (kip)TmSpecified minimum fastener tension (kip)TuAxial force per bolt induced by the externally applied load (kip)WA load placed on the structure (kip)

gA coefficient less than unity reflecting the degree to which variousload effects may be underestimated (dimensionless)

fResistance factor; a coefficient less than unity reflecting the degreeof uncertainty to which the strength of the joint may beoverestimated (dimensionless)

mMean slip coefficient of the joint's faying surfaces as listed in thespecification, or as established by tests (dimensionless)


References*

1. Research Council on Structural Connections, Specification forStructural Joints Using ASTM A325 or A490 Bolts: Allowable StressDesign.

2. Kulak, G. L., J. W. Fisher, and J. H. A. Struik, Guide to DesignCriteria for Bolted and Riveted Joints, 2nd ed., Wiley, New York,1987.

3. AISC, Manual of Steel Construction, Vol. 2, Connections (ASD,9th ed.; LRFD, 1st ed.), AISC, Chicago, IL, 1992.

4. Research Council on Structural Connections, Specification forStructural Joints Using ASTM A325 or A490 Bolts: Load andResistance Factor Design.

5. Bickford, J. H., An Introduction to the Design and Behavior ofBolted Joints, 3rd ed, Marcel Dekker, New York, 1995.

*Note: The AISC specifications can be obtained for a modest price from the AmericanInstitute of Steel Construction, Inc., One East Wacker Drive, Suite 3100, Chicago, IL60601-2001. Publication sales can be arranged by telephone:1-800-644-2400.


PART VASSEMBLY OF BOLTED JOINTS

23Behavior of a Bolted Joint during Assembly


1Introduction

Many different assembly tools and proceduresare discussed in this section of the Handbook.Some of these will seem overly complicated tothose who don't appreciate the complexity ofbolted joints. In this chapter we examine thebehavior of the joint during assembly,


introducing many of the factors that make itdifficult to do a "perfect" job with even the mostsophisticated tools and procedures.

That all-important clamping force that holds thejoint togetherand without which there would beno jointis not created by a good joint designer orby high quality parts. It is created by themechanic on the assembly line or job site, usingthe tools, procedures, and working conditions heor she has been provided with. The force isbrought into being as energy in the mechanic orpower tool is converted to potential energystored in the joint and bolt members. The correctamount of force cannot be created if the design isfaulty or the parts don't fit together properly or ifthey break; but getting all this right, whilenecessary, is not enough. The final, essentialcreator of the force is the mechanic, and the timeof creation is during assembly. So it is veryimportant for us to understand the assemblyprocess.


How do the bolts and the joint members respondas we tighten the bolts? We will see that thebehavior of the parts during assembly iscomplex. This chapter takes a close look atseveral unseen, difficult to detect, difficult toquantify factors that can have a significantimpact on the results, on the amount of clampingforce developed in the joint. This will teach usthat it isn't easy to control the buildup ofclamping force in the joint and that mechanicsneed all the help we can give them, both asproduct designers and as production engineers.

In eleven of the chapters that follow, we willlook at many of the options we have for controlof the bolt-tightening process. Our knowledge ofthe behavior of the bolts and joint duringassembly will help us evaluate the merits of theseoptions.

Accurate control of the clamping force is oftennecessary; too much or too little clamp candegrade the behavior and life of the joint in


service. If the bolts and joint members do notcontain the correct amount of stored energy tocreate the correct amount of clamping force,there will be joint problems. In other words,proper assembly can be essential.


2Initial vs. Residual Preload

The clamping force a bolt exerts on the joint isusually calledor equated tothe "preload" in thebolt. This term is used in general in most of theliterature on bolting to describe the tension in thebolt at any time, but this, in my opinion, is amistake. I like to think of the preload created inan individual fastener when it is first tightened as"initial preload," even though that term may beredundant. As you'll see, the effects we're aboutto discuss will frequently modify this preload asthe fastener relaxes and/or as we tighten otherfasteners in the joint. I call the final preload inthe bolts the "residual preload."

When the joint is put into service, a variety ofthings can act to modify the preload in individualfasteners still further. This could be called in-


service tension in the bolts.

Each of these preloads or tensions is directlyproportional to the amount of potential energystored in the bolt as it is first tightened, afterrelaxation occurs, or in service. In most casesthese preloads or tensions will also be directlyproportional to the clamping force between jointmembers, but there are exceptions.

Now that we have these definitions under ourbelt, let's get on with it. What happens duringassembly?

3Starting the Assembly Process

We are going to assemble a hypothetical joint,using as our example a round, gasketed, pipeflange joint held together by sixteen 1 1/88ASTM A193 B7 bolts. (See Fig. 1.) The largediameter and the presence of a gasket make thisassembly a little more difficult than most but


allow us to look at a more complete range ofassembly problems than would a simplerexample. Most of the discussion will apply tojoints in general.

We are also going to measure the torque weapply to the nuts to control the buildup of initialpreload in these bolts. This is the most commontype of control, and one of the simplest. It is thesubject of Chapter 29, so we won't go into a lotof detail here about its pros and cons. We willjust use it for now.

3.1Assembling the Parts

We start by roughly aligning the flanges so thatwe can insert the bolts by hand. When we finishpushing and pulling on the flanges, their matingsurfaces are not exactly parallel and the


Figure 1A sketch of the large-diameter pressure vessel joint used as

an example in this chapter.


holes aren't aligned perfectly, so we have to tap afew of the bolts with a hammer to get themthrough their holes, and some of them stick out alittle farther, on the nut end, than others. Nowwe're going to apply a preliminary ''snuggingtorque" to run the nuts down and pull the flangestogether.

3.2Tightening the First Bolt

To load the joint (and gasket) evenly we applythe snugging torque in a cross or star pattern, asshown in Figure 2. We would use a similarpattern on a square or rectangular joint if thebolts were all around the edge. In a rectangularstructural joint pattern with several rows ofbolts, we would start snugging at the center ofthe bolt pattern and work our way out to the freeedges.


We use 225 lb-ft of torque for this first,snugging pass. This is about one-third of thefinal torque we're planning to use, and we followit with a second pass at two-thirds of finaltorque, and then a third and final pass at fulltorque. In a structural steel joint we wouldfollow the snugging pass with a second (last)pass at the final torque. Note that in each case weare following basically a two-step procedure:Pull the joint together and then tighten it.Because this is a learning experiment we useultrasonic equipment (Chapter 34) to measurethe preload in each bolt as we tighten it. We alsomeasure the angle through which the nut turnsafter it contacts the surface of the joint, and wemeasure the amount by which the bolt stretchesand the amount by which the joint is compressed.

We now apply the snugging torque to the firstbolt and use the resulting preload, torque, andturn data to plot the curves shown in Fig. 3. Weare doing work on this fastener as we tighten it.


The amount of work is equal to the area underthe torqueturn curve (measured in pound-feet ornewton-meters times radians). Ideally, all of thiswork would be converted to potential energy inthe bolt and in those portions of the jointmembers that surround it. If that were the case,all of the work we do on this fastener would endup contributing to the clamping force.Unfortunately and unavoidably, most of ourinput work is lost.

Typically about 90% of the work we do on a nutis converted to heat, owing to the frictionalresistance between the face of the nut and thesurface of the joint and between male and femalethreads. About 50% is lost under the nut andabout 40% within the threads, as


Figure 2We'll tighten the bolts of our example

joint in the "star pattern" sequenceshown here. We will use three

passesat one-third, two-thirds, andfinal torquefollowing the same

sequence on each pass.


Figure 3As we tighten the first bolt in our

example joint we plot(a) the buildup of initial

preload vs. applied torque

(b) applied torque vs. the angle


through which the nut turns.The area under the torqueturn

curve is equal to the amount of workwe are doing on the nut and to the energy

delivered to the fastenerjoint system.

shown in Fig. 4. Only 10% of the input worktypically ends up as potential energy in the bolt,so only 10% ends up as bolt preload or asclamping force between joint members.

We would like to apply a given torque to eachbolt and create a given amount of initialpreloadthe same amountin each bolt. But the factthat most of the work we do on the nuts isconverted to heat makes this virtuallyimpossible, because these frictional losses areextremely difficult to predict or control. Let'sassume, for example, that this first nut we'retightening is a little drier than average. As aresult, let's assume that 52% of the input work isconverted to heat at the nut/joint interface, ratherthan the "typical" 50%. A 4% increase infrictionfrom 50% to 52% of the input workis


easy to come by.

Figure 4This diagram shows the approximate way in which the energy

delivered to the fastener-joint system is absorbed by it. About 50%of the input is lost as friction-generated heat between the face ofthe nut and the surface of the joint. Another 40% is lost as heat

between male and female thread surfaces. Only about 10%, on theaverage, ends up as potential energy stored in the bolt and joint

"springs"; and only that 10%, therefore, ends up as preload in thebolt and clamping force on the joint.


Figure 5(a) As we tighten the first bolt we also plot the buildup ofpreload (Fp) in the bolt vs. the increase in length (DL) ofthe bolt, and the buildup of clamping force on the joint

(FCL) vs. the compression or change in thickness (DT) ofthe joint. At this point we assume that the preload will be

equal to the clamping force,

(b) We then combine thosetwo plots to start constructing a "joint diagram."

This 4% increase in friction lossthat extra 2% of the input work going into heatmeans that2% less of the input work will be converted to tension in the bolt. We started with theassumption that an average of only 10% of the input would be going into preload; nowwe've lost one-fifth of that. This bolt will therefore end up with only 80% as much preloadas we expected it to. A 4% swing in friction has caused a 20% change in assembly preload,a very bad "leverage" situation. There are a lot of factors that can cause this kind ofvariation in friction.

Although in our learning experiment we're measuring both torque and bolt tension, we willnot attempt to compensate for the frictional differences between bolts; we will apply thesnugging torque of 225 lb-ft to the first bolt and let the initial preload end up where it may.

We also plot the deflections in the bolt and in the joint material surrounding the bolt versusthe preload we create in the bolts and the presumably equal and opposite clamping force onthe joint. (See Fig. 5a.) We then combine these forcedeflection curves, plotting the preloadon a common axis, as shown in Fig. 5b. This creates what the bolting world calls a jointdiagram. You may complain that the preload in the bolt and the clamping force on the jointare equal and opposite action and reaction forces and that both should not be shown aspositive values, but this joint diagram is a great convenience.


Since the diagram records the forces developed in bolt and joint and the deflection of eachpart, it also gives us a visual indication of the stiffness of the bolt and the stiffness of thejoint. These are proportional to the slopes of the two straight lines and, as we will see inChapter 36, can be computed as follows:


where

FP=preload in the bolt and joint (Ib, N)

DL=deflection (stretch) of the bolt (in., mm)

DT=deflection (compression) of the joint (in., mm)

KB=stiffness of the bolt (Ib/in., N/mm)

KJ =stiffness of the joint material being loaded by this bolt(Ib/in., N/mm)

Note that the areas under the bolt and joint curves also equal the amount of energy stored inthese parts. So this simple diagram (Fig. 5b) contains a lot of useful information. Thisdiagram is extended in Chapter 36 to add the effects of external loads on the joint. Jointdiagrams are also used when we design joints. Now, however, we are merely interested inusing the joint diagram to illustrate the preloading of bolts.

Our supervisor has used the torquepreload equation [Eq. (1), Chap. 29)] to compute theaverage preload he expected us to get in this first bolt when we applied 225 lb-ft of torqueto it. He tells us that we should have created 12,000 lb of tension in the bolt. Because ofthe slightly higher than average friction loss described earlier, however, this bolt has endedup with only 80% of that preload, or 9600 lb.

Has this really created 9600 lb of clamping force between joint members? Our jointdiagram assumes it has, but the correct answer is "probably not"at least as far as this firstbolt is concerned. Remember that we had to tap some of those bolts into their holes? Thisimplies that there was contact between the sides of those bolts and their holesbolt-holeinterference. Furthermore, the flange surfaces were not pulled into full contact when wefirst assembled the parts. They were slightly misaligned, as we could tell by the fact thatsome bolts stuck out farther on the threaded end than others. Before we go on to snugtighten the remaining 15 bolts in our example joint, let us take a look at how holeinterference and/or nonparallel flanges might affect the buildup of clamping force duringthe assembly process.


4Bolt Preload vs. Clamping Force on the Joint

The main purpose of the bolts is to clamp the joint members together. A commonmisconception is that there is always an equal and opposite actionreaction relationshipbetween the tension in the bolts in a joint and the clamping force between the jointmembers. If there are eight bolts in the joint and an average tension of 10,000 lb in each ofthe eight bolts, then, simplistically, the joint is clamped together with an interface forceequal to eight times 10,000, or 80,000 lb. That's usually true, but there can be somesignificant exceptions.

4.1Effects of Hole Interference

Consider the situation shown in Figure 6. We are tightening this stud by turning the uppernut. Our goal is to clamp the joint members together. The hole in the upper joint member isundersized, so the stud is a press fit in this member.

Because of frictional and/or embedment constraints between the sides of the bolt and thewalls of the hole, it will take some positive force to pull the bolt through the hole and thento stretch it within the hole. Where does this force come from?

The force is created, obviously, when we turn the nut, creating tension in the bolt. Thanks tohole interference some of this tension will not end up as clamping force between jointmembers. Part of it will be lost as the bolt is forced past the walls of the hole.


Figure 6If there is interference between

the bolt and the hole, the clampingforce between joint members maybe less than the tension in the bolt,

changing the torque-clampingforce relationship.

Envision the extreme case in which the holes aregrossly undersized and the torque normally


specified for a bolt of this diameter insufficientto even bring the joint members into contact,much less provide any real clamping force. Notethat misalignment between the holes of upperand lower joint members could create a similarhole/bolt interface problem. This is obviouslynot a good situation, but hole misalignment,undersized holes, press-fit fasteners, etc. arerelatively common in the bolting world.

Hole interference is used on purpose by theairframe industry, for example, to reduce thepossibility of fatigue failure of the shear-loadedjoints used in airframe structures. (Compressivestress built up in the walls of the bolt holesopposes the formation or growth of fatiguecracks.)

The holes are purposely drilled smaller than thediameter of the bolts to create this interference.There is, of course, a manufacturing tolerance onthe diameters of both holes and bolts, so theamount of interference varies. The greater the


interference, the greater the force required to pullthe bolt through its hole. It also requires moreforce to pull a bolt through a thick plate thanthrough a thinner one, for a given amount ofinterference.

One airframe manufacturer measured the amountof force required to pull bolts of a given nominaldiameter through holes drilled in plates ofvarying thickness. Some of the results are shownin Figures 7 and 8. The bolt and hole diametersused in the experiment varied through the fullrange of the manufacturing tolerance. It wasfound that the force required to pull some ofthese bolts through their holes exceeded theaverage preload that would be developed by thespecified torque in a bolt of that diameter even ifthe bolts were used in regular holes with normalclearance. In other words, the specified torquecould not be counted on to pull all of the boltsthrough their holes, much less go on to developany clamping force between joint members. In


spite of this, the same torque is specified by theairframe manufacturer for all bolts of a givendiameter, without regard to the amount of holeinterference seen by a given bolt or the thicknessof the plates in which the bolt is used.

As a result, torque is not used in this applicationto pull the bolts through the holes. A bolt pullerdoes that job, and temporary clamps are used tohold the joint members together until the boltshave been installed and tightened. But in somejoints the act of tightening the


Figure 7Chart showing the force required to push 5/16-in. fastenersthrough interference fit holes in aluminum plates varying in

thickness from 1/4 to 3/4 in. Bolt diameters are larger than the holediameters by the amounts shown on the horizontal axis. Note the

wide scatter inthe results, which are summarized in Figure 8.


Figure 8The three vertical bars at the right in this chart show the range

of force required to thrust 5/16 in. fasteners throughinterference fit holes in aluminum plates of various thicknesses,

summarizing some of the data shown in Figure 7. The three barsat the left show the nominal preload that would be generated

by 5/16 in. bolts tightened to 50% of yield. Three bolt materialsare shown. Note that only the Inconel bolts would haveenough preload to overcome the worst-case interference

forces and still provide some interface clamping force on thejoint if 1/2 or 3/4 in. plates were involved.


bolts is supposed to create some clamping forcebetween upper and lower joint members, and theamount of this clamping force must vary widely.

4.2Resistance from Joint Members

Another factor that can rob from the clampingforce between joint members is shown in Figure9. A heavy cover is being lifted up against aflange on a pressure vessel. As shown in thefigure, the joint members have not yet beenbrought into contact, but the nut is being turnedon the stud to bring them into contact. At themoment shown in the figure, there is alreadytension in the study equal to the weight of thecover. As a result, it will take torque to advancethe cover up against the flange, due to the normalfrictional constraints between male and femalethreads and between the nut and the cover.


So, at the point shown, there is torque, there istension, there is friction loss, and there probablywill be torsion in the stud, but there is zeroclamping force between joint members.

Eventually the two joint members will bebrought into contact. Further torque will berequired at this point to create a clamping forcebetween joint members to load the gasket. Thattorque should presumably be added to the torquerequired to pull the joint members together in thefirst place, but in practice it rarely is.

There aren't many applications in which a heavyweight is raised against a mating joint member bytightening the bolts, of course. It is, however,common for large joint members to bemisaligned or nonparallel as the assemblyprocess starts. Getting a pipe flange, for example,to mate with the flange of a pump or valve oftenrequires a lot of motion in the flange members.The forces required to align such systems willhave the same effect that the force created by the


weight in Figure 9 has on that joint.

I wish I could tell you how much extra torque toadd for misalignment or the like, but I can't. Ihave never seen anything in the literature on thissubject, either. Yet I'm convinced that this factorcan seriously degrade the relationship betweentension in the bolts and clamping

Figure 9Heavy and/or misaligned or warped joint members can also


affect the relationship between the torque applied to thefastener, the tension in the fastener, and the clamping force

between joint members.


Figure 10Puget Sound Naval Shipyard used a hydraulic tensioner as

shown in this sketch to learn how much force would berequired to pull misaligned flanges together. The data wererecorded as a function of the size of the "gap" shown here.

force between joint members, especially in largejoints, as one example. Warped or nonflat (e.g.,wavy) joint members, incidentally, could createthe same sort of problem.


I once met a maintenance supervisor in a largepetrochemical plant who took this problemseriously. He insisted that his crews aligngasketed flanges within 12 mils before boltingthem together. Bolting nonparallel flanges, hesaid, was a waste of time; "they'll always leak."

The Puget Sound Naval Shipyard also studiedthis problem [18]. They made theoreticalcalculations and conducted experiments todetermine the forces, stresses, and moments inpipes and flanges when the flanges aremisaligned. They used a hydraulic tensioner, forexample (see Chapter 26), to measure the forcerequired to pull misaligned flanges together, asshown in Figure 10. The chart in Figure 11shows some of the results of their work: thestresses in the pipe adjacent to the flange as afunction of nominal pipe diameter and withflange gaps of 0.015, 0.020, and 0.025 in. Theforces (and torques) required to pull the flangestogether would be proportional to these stresses.


In each case documented in Figure 11 it isassumed that there is a length of pipe equal to100 times the nominal pipe diameter between theflange and the first rigid pipe support (a rodhanger or intersection with a larger pipe).

As a result of these studies, the shipyarddeveloped some flange parallelism criteria,designed, as I understand it, to keep pipe stressbelow 3 ksi near any flange that is connected toturbines or other rotating equipment. Pipestresses beside flanges located 50 or more pipediameters away from rotating equipment wereallowed to go slightly higher. The resultingcriteria are shown in Table 1.

As a final note, if misalignments and stresses aretoo high they are sometimes reduced by the useof bellows-type expansion joints at some point inthe pipeline.

5Continuing the Snugging Pass


We are now going to continue tightening the 16-bolt gasketed flange joint we are taking as anexample in this chapter. We apply the snuggingtorque of 225 lb-ft to each of those bolts,following the cross-bolting pattern shown inFigure 2. We measure the preload, turn, and


Figure 11A plot of the stress in the pipe adjacent to a misaligned

flange vs. the nominal diameter of the pipe, for misalignmentgaps of 0.015, 0.020, and 0.025 in. It is assumed that the

nearest rigid support for the pipe is located 100 pipediameters away, along the pipe, from the misaligned flange.The moment on the flange and the force required to pull the

two halves into full contact would be proportional to thepipe stress.

deflection in and around each bolt as we tighten it, andwe record these data. We find that, on average, we have


created the anticipated preload in these bolts but thatindividual bolt results vary from the averageby ±30%.That, our supervisor assures us, is a typical result oftorquetightening a group of unlubricated, as-receivedsteel bolts and nuts against steel joint surfaces.Everyone's happy, so we now go out and have lunch.

When we come back from lunch we find our boss'sboss on the job site, reviewing our data. He wants tosee how we managed to measure the tension or preloadin these bolts, soTABLE 1 Flange Parallelism Criteriaa

Flanges adjacent to rotating equipment Line flanges

Pipe size 1 in. orless 1 1/48 in. All sizes

Wallthickness All Sch 40 or

belowOver 40 throughSch 80 All thicknesses

Raisedface 0.10 in. 0.025 in. 0.020 in. 0.005 in./in. of contact

diameter

Flat face 0.020in. 0.035 in. 0.030 in. 0.005 in./in. of contact

diameteraIf the pipe diameter is over 10 in. or wall thickness greater than Schedule80, the specifications say that a special analysis is required.


we plug in our ultrasonic instrument andremeasure the preload in bolt number 1, the firstbolt we tightened to a preload of 9600 lb. To ourembarrassment we find only 900 lb of tension inthat bolt. And we find a wide range of residualpreloads in the other 15 bolts. We think that ourinstrument is misbehaving, but the bossdisagrees. He gives us a lecture on boltrelaxation, which some people call "torque loss."Here's what he says.

6Short-Term Relaxation of Individual Bolts

Whether or not there's a one-to-one relationshipbetween bolt tension and interface clampingforce, there will often be some initial loss oftension in individual bolts after they are initiallytightened. Call this short-term relaxation todistinguish it from other effects such as stress


relaxation [22] that will cause further loss oftension over a longer period of time.

In general, short-term relaxation occurs in abolted joint because something has been loadedpast its yield point and will creep and flow to getout from under the excessive load. This can be acomponent such as a soft bolt or a gasket; morecommonly it is only a portion of a component,such as the first threads in a nut. Let's look atsome examples.

6.1Sources of Short-Term Relaxation

Here are some things that can cause relativelyshort term relaxation, starting with the mostcommon of all, embedment.

The surfaces of the threads in the nut, the bolt,and the faying surfaces of the structuralmembers, washers, etc. are never perfectly flat,even if such parts are given a high polishwhich israrely the case with industrial structures and


fasteners. Under a microscope they are a series ofhills and valleys.

When such parts are first loaded, they contacteach other only through high spots on the metalsurfaces. Even a small bolt, however, is able toexert extremely high surface pressures onstructural members or on its own threads. Threaddimensions have been selected to support thesehigh loads, but only if a significant percentage ofthe total thread surface shares that load.

Since initial contact areas are relatively small, themetal at the contact points cannot stand thepressures. Plastic deformation occurs untilenough of the total thread surface has beenbrought into play to stabilize the situation andsupport the load without further deformation.The same thing happens in the faying surfaces ofthe structures, though perhaps to a lesser extentbecause larger surfaces are involved and initialcontact areas are larger . Embedment isillustrated in Figure 12.


Many of these surface high spots are smoothedaway during the tightening process [1]; at leastthey will be if the fasteners are torqued.Hydraulic tensioners don't load the activethreads, or even the joint surfaces underneath thenut or washer, until they let go of the bolt. As aresult, there is often more embedment relaxationafter tensioning than there is after torquing.

Embedment is worse on new parts than on reusedones. In critical applications it can be minimizedby tightening, loosening, and retightening thefasteners several times. This is done on cameramounts in space satellites, for example [2].

6.1.1Poor Thread Engagement

If the bolt is undersized, or the nut oversized,thread contact areas will be less than thoseplanned by the designer, and substantial plasticdeformation may occur, as shown in Figure 13[3].


6.1.2Thread Engagement Too Short

The length of thread engagement for steelfasteners should be at least 0.8 times the nominaldiameter of the fastener. If the engagement lengthis too short (too few threads support the


Figure 12High spots on thread and other contact surfaces will yield and

creep under initial contact forces. As a result, the surfaceswill settle into each other until enough surface area has been

brought into contact to stabilize the joint. The process iscalled embedment.

load), thread contact areas are again smaller thanthose intended by the fastener manufacturer andexcessive relaxation can result. One authorclaims that if thread engagement length is greaterthan 1.25 times the nominal diameter, then''permanent set is negligibly small" [1].


6.1.3Soft Parts

If parts are softer than intended by thedesignerperhaps because of improper heattreatment or incorrect materialthey may creepand relax substantially even if the geometry iscorrect and loads are normal.

6.1.4Bending

If the fastener is bent as it is tightened, it will seehigher stresses along one side than along theopposite side, as we have seen in Chapter 10.These higher stresses mean more plastic flow andtherefore greater than normal embedment orthread relaxation.

6.1.5Nonperpendicular Nuts or Bolt Heads

The contact faces of nuts and bolt heads arenever exactly perpendicular to the axis of the


threads or to the axis of the bolt hole. This meansthat only a portion of the contact surface of thenut or bolt head is loaded when we first tightenthe fastener. These abnormally loaded

Figure 13Poor thread engagement may be a major

source of plastic deformation andtherefore joint relaxation.


Figure 14Oversized fillets and/or

undersized holes may result intotal relaxation of a preloaded

fastener.

surfaces will creep until enough additionalcontact area has been involved to reduce contactpressures and stabilize the joint.

6.1.6Fillets or Undersized Holes

If the head-to-body fillet contacts the edge of thebolt hole as shown in Figure 14, the edge of the


hole will break down under initial contactpressures. This may result in a complete loss ofpreload, since such effects are usually largecompared to the amount by which the bolt wasstretched when it was initially tightened [4].

6.1.7Oversized Holes

Oversized holes can also be a problem. Nowthere is too little contact between nut and jointsurface or between bolt head and joint surface.Unless a washer or something is used todistribute contact pressures and limit contactstresses, the head and/or nut will embed itself inthe joint surfaces, as suggested in Figure 15 [3].The amount of relaxation will, of course, dependon the strength of the surface supporting the nutor washer. The oversized or slotted holes used toaid the assembly of structural joints, forexample, do not increase relaxation appreciably[5].


6.1.8Conical Makeups

Surface irregularities exist on conical jointsurfaces as well as flat ones. The effect on axialtension in the fastener, however, is magnified ifthe embedment occurs on a conical surface.

Figure 15Oversized holes may also increasecontact stress levels and thereforeincrease embedment relaxation.


Figure 16Conical or tapered joints usually relax more than flat ones for

reasons given in the text.

A given amount of relaxation perpendicular to the surface may meansubstantially greater relaxation in the axial direction, as suggested inFigure 16 and Eq. (3).

where

e=

embedment relaxation perpendicular to the surface of a conical jointmember (in., mm)

r=resulting relaxation parallel to the axis of the fastener (in., mm)

q=half-angle of the cone (deg)

6.2


Factors Affecting Short-Term Relaxation

Overstressed parts relax. Overstressing may be created in a number ofdifferent ways, as we have seen. The amount of relaxation thatoverstressing causes in a given bolt and joint, however, can depend ona number of secondary factors.

6.2.1Bolt Length

Long, thin bolts will relax by a smaller percentage than short, stubbyones. The total embedment relaxation or the like will be the same for agiven initial preload, but that embedment will be a different percentageof the total length of the bolt and therefore account for a differentpercentage loss in length. Preload loss will be proportional to thechange in length (see p. 61 of Ref. 5).

Many people take advantage of this fact. They add bushings above andbelow flange surfaces, for example, as shown in Figure 17a. Thismakes it possible for them to use longer bolts on a given joint.

6.2.2Belleville Washers

Another common way to reduce the change in clamping force producedby a given amount of embedment is to use Belleville washers (Fig.17b). These springs have a very flat rate compared


Figure 17Since long, thin bolts will relax bya smaller percentage than short,stubby ones, many people use

bushings as shown in(a) to reduce percentage relaxation

in a given joint. Stacks ofBelleville washers


(b) are also effective.

to the stiffness of either bolt or joint members.They will therefore determine (limit) the preloadand clamping force in the system. Because oftheir flat rate, a small deformation in the bolt orjoint will not make an appreciable difference inforce levels. For the same reasonand morecommonlyBellevilles are used to compensate forthe effects of differential temperature expansion.Spring manufacturers have trouble controllingthe stiffness of Bellevilles, however, so thissolution to a temperature or relaxation problem


can increase the basic scatter in preload.

6.2.3Number of Joint Members

Increasing the number of surfaces in a joint mayincrease relaxation effects because there are nowmore high spots to embed and settle in together.Doubling the number of contact surfaces, forexample, will almost double relaxation in manycases [6].

6.2.4Tightening Speed

Creep and flow take time. Fasteners that aretightened very rapidly do not have time to settlein together during the tightening process andrelax more after tightening [7]. This is one of theadvantages of hesitation tightening or torque-recovery tightening. Tighten with high speedtools, but pause to give the parts time to relax.(See Chapter 28.)


Tightening bolts in a series of passes rather thanapplying full torque on the first pass allows timefor relaxation. This procedure also pulls the jointtogether uniformly. For both of these reasons,progressive tightening is a virtual necessity onlarge gasketed joints.

6.2.5Simultaneous Tightening of Many Fasteners

Some experiments [8] have suggested thattightening a group of fasteners one at a timeresults in more relaxation in a given fastener thandoes tightening several or all of them at once.Presumably a fastener tightened before itsneighbor is subjected to higher stressconcentrations than if it is tightenedsimultaneously with the rest and all share thedeveloping load.

6.2.6Bent Joint Members

If joint members are soft or warped or bent,


tightening one fastener can cause relaxation (oradditional stress) in other fasteners. This sort of"cross-talk" between fasteners is very common,although it is not usually seen or recognized.More about this in Chapter 24.


6.3Amount of Relaxation to Expect

The factors that cause and contribute torelaxation are many and difficult to predict.Although attempts have been made to writeequations for the amount of relaxation to expect[6,8,9], in most cases the amount must bedetermined experimentally. And, as usual whendealing with bolted joints, it won't be "an"amount but rather a distribution of values aroundsome anticipated mean.

In general, fasteners relax rapidly followinginitial tightening, then relax at a slower rate,following the pattern shown in Figure 18. Theamount of relaxation varies greatly, dependingon the condition of the parts, finishes, initial andlocal tension levels, fit of parts, and all of theother factors discussed earlier. Here are some ofthe relaxation amounts and times found in the


references. Fisher and Struik report [5, p. 61]that tests of A325 and A354 Grade BD bolts inA7 structural steel showed a loss of 211% ofpreload immediately after tightening, followedby another 3.6% in the next 21 days, followed byanother 2% over the next 11.4 years. BethlehemSteel reports [10] that only 5% of the initialtension will be lost in structural bolts set by turn-of-nut techniques5% over the total life of thestructure. Chesson and Munse report [11] avariety of results on a variety of structural boltswith different types of bolt heads, different nuts,with and without washers, etc. As one example,an A325 bolt with a regular (not heavy) head, aflanged nut, and no washer relaxed 2.6% in thefirst minute after tightening (most of this in thefirst 1520 s). It had relaxed by 6.5% after 5 days.Hardiman [13] reports that most relaxationoccurs in the first few seconds but thatrelaxation, usually, never stops. SouthwestResearch Institute [14] suggests that fastenerslose an average of 5% of preload right after


tightening, "because of elastic recovery." The 23/4-8 × 12 3/4 Nitronic 50 top guide studs in aboiling water reactor (BWR) relaxed an averageof 43% after tensioning. Grip length was 4.75in.; studs were hydraulically tensioned to160,000 lb; nuts were run down with a measuredtorque of 500 in.-lb. This is not all embedment,as we will see in Section 7.

Gasketed joints will relax substantially whetherthe bolts are torqued or tensioned. This isespecially true during preliminary passes, whenloss of as much as 80100% of initial tension isnot at all uncommon for reasons to be discussedsoon. Gaskets will eventually stabilize, however,and will retain the tension introduced in finaltightening operations.

6.4Torsional Relaxation

Preload or tension relaxation is of primeimportance to us because of the general


importance of correct preload. We must notforget, however, that torsional stress is also builtup in a fastener as it is tightened and that thisstress is also subject to varying amounts ofrelaxation. Many people, in fact, insist thattorsional stress disappears immediately andcompletely when the wrench is removed from thefastener. Others find that it doesn't disappearuntil a breakaway torque is applied [15]. Ourexperience indicates that, like tension relaxation,torsional relaxation depends on many factors; theamount and rate of torsional relaxation will varysubstantially from bolt to bolt as well as fromapplication to application.

Figure 19 shows the tension and torsionrelaxation measured in an experiment with a 21/4-8 × 12 B16 stud that had been lubricatedwith molybdenum disulfide (Moly-lube).Torsion relaxed 50% when the wrench wasremoved; tension actually increased 12% duringthis period. I have since seen this phenomenon on


many other types and sizes of bolts. My guess isthat, as embedment allows relaxation of bothtension and torsion stress to occur, some of thetorsional stress is turned into a little moretension stress. The twisted bolt screws itselffarther into its own nut. The exchange isencouraged if you lubricate the threads but donot lubricate the face of the nut.

Sizable relaxation of tension occurring as thetorsional stress disappears can, of course, maskthis exchange of torsion for tension. Thus I haveobserved this exchange only on hard joints. But itis relatively easy to reproduce, and I think it iscommon. Many bolts in large


Figure 18Most short-term relaxation occurs in the first few seconds

or minutes following initial tightening but continues at alesser rate for a long period of time.


Figure 19Relaxation of torsional stress in a bolt canbe accompanied by an actual increase intension. The bolt screws itself into its nut.

Data shown were taken in tests on a2 1/4-8 × 12 B16 stud.


joints mysteriously "grow" a little betweenpasses, for example, even when there is notemperature change or the like to explain thegrowth and even when neighboring bolts remainat constant length. Presumably elasticinteractions play a role here as well.

A torque wrench appears to respond to torsionalstress levels in the bolt as well as to preloadlevels. The torque required to restart a nut can beless than that required to tighten it in the firstplace, even if there has been no loss in preload inthe meantime, if torsional stress has disappearedor been reduced. Repeated "torque recovery" can,as a result, gradually increase the tension in afastener until it is substantially above initialanticipated levels. In one set of measurements ontank tread end connector bolts, for example, wefirst tightened the 5/8-18 × 1 3/4 Grade 8 boltswith a torque of 150 lb-ft. This stretched the bolt


0.0015 in. This was a tapered joint, so the boltnow relaxed to a stretch of only 0.001 in. Wereapplied 150 lb-ft of torque, and the stretchreturned to 0.0014 in. Incidentally, it took only100 lb-ft to restart the nut.

The bolt now relaxed again, was tightened again,relaxed some more, etc., as shown in Figure 20.Final preload (stretch) was 33% greater than thatachieved in the first pass, even though we neverapplied more than the initial 150 lb-ft of torque.The final restarting torque was still only 87% ofthe rated torque. Many mechanics wouldconclude from this that preload was still 13%below the initial value. This possible interactionbetween torsional and tension stress furthercomplicates the task of predicting how much agiven fastener will relax, of course. This isanother complex situation.

7Elastic Interactions between Bolts


Even if we can avoid the problems cited aboveand can count on achieving a certain amount ofpreload in the bolts we tighten at assembly one byone, there are going to be times when that preloadwill be significantly modified as we tighten otherbolts in the same joint, owing to "elasticinteractions" between bolts. Let's look at anexample.

Assume that we are planning to tighten a circularflanged joint that contains eight bolts. We aregoing to use ganged hydraulic wrenches orhydraulic tensioners to tighten these bolts


Figure 20Torquestretchrelaxation history of a 5/8-18 × 1 3/4 Grade 5 bolt. A

torque of 150 lb-ft was applied repeatedly to this fastener, with apause for relaxation between passes. Final preload was 33% greater

than that achieved on the first pass.


Figure 21A simulated model of a bolted joint in which the joint membersare represented by a large spring, here "loaded" by the first two

bolts to be tightened.

two at a time. The bolts we tightensimultaneously, of course, will be opposite eachother on the flange, 180° apart.

To explain the process of elastic interactions,think of the joint as a large spring connected byrigid top and bottom plates to the bolts (smaller


springs) that are going to be used to clamp it. Thisarrangement is suggested in Figure 21, whichshows the first two bolts to be tightened.

Now, let us assume that we have tightened bolts 1and 2 in this joint and that magically we haveachieved exactly the initial preload we wanted ineach bolt. Let's say that this preload is 10,000 lbof tension in each bolt.

The 20,000 lb of force that these two bolts arecreating on the joint partially compresses thejoint. We now go on to tighten bolts 3 and 4located 90° away from bolts 1 and 2. Again, ourtools work magic for us, and we create exactly10,000 lb of initial preload in bolts 3 and 4 whenthey are tightened (Fig. 22).

We now have four small springs (bolts)compressing the joint spring rather than the twosmall springs we had a moment ago. If bolts 1 and2 had retained their full preload, we would nowhave 40,000 lb of force on this joint instead of


20,000. Doubling the compressive force on thejoint spring would, of course, double the amountby which it is compressed. But what

Figure 22The joint model of Figure 21, but now with four bolts to be tightened.


happens in bolts 1 and 2 when we tighten 3 and 4?Bolts 1 and 2 are allowed to relax a little as the jointis compressed by bolts 3 and 4.

At this point in the process, therefore, bolts 1 and 2have a slightly lower preload in them than bolts 3and 4even though each of the four bolts started withthe same initial tension of 10,000 lb. When we nowgo on and tighten bolts 5 and 6 in a third step, bolts3 and 4 will relax a little, and bolts 1 and 2 will relaxfurther. Tightening bolts 7 and 8 to complete theassembly will create relaxation in each of the sixbolts tightened earlier.

The result is four different levels of residual preloadin the eight bolts when they are tightened two at atime, even though the initial preload in each one wasidentical. And this is not just a theoreticalpossibility; it is a very common occurrence. Mostpeople are not aware of this interaction, however,which is visible only if you use ultrasonics or straingages or something to monitor the tension in the


bolts. This process, called elastic interaction, is thesubject of Chapter 24 [16,17,1921].

We can now modify the joint diagram of Figure 5bto show the loss in preload created by embedmentand by the average elastic interaction effect, as inFigure 23. The dashed lines show the situationboltand joint forces and deflectionsbefore relaxation.The solid line shows the residual preload situation atthe end of the assembly operation.

8The Assembly Process Reviewed

After learning about elastic interactions, we makeone more pass around our 16-bolt joint, in thereverse order, to reduce the scatter in residualpreload. We remeasure the preloads in each boltafter that pass and find that each has changed. Weaccept the final results and then draw the blockdiagram shown in Figure 24 to summarize the thingswe now know can affect the amount of clampingforce created on a multibolt joint when it istightened. We include a few things learned


elsewhere, the fact that some bolts will be bentslightly if tightened against nonparallel jointsurfaces, and some minor factors not discussed, suchas the very small heat loss in the tools and drive barswe're using. We also include the possibility of heatloss through

Figure 23The joint diagram of Figure 5 modified to show the effects of embedmentand elastic interaction on initial preload. The largest triangle reflects theinitial, "just tightened" situation. Embedment reduces the initial preloadby an amount shown as A. The average loss in a group of bolts caused

by elastic interactions is shown as loss B. We use the average lossinstead of the individual loss because joint behavior will more likely bedetermined by average loss than by worst case loss. The final average

residual preload in this group of bolts is represented by the solid triangle.


Figure 24Block diagram showing most of the factors that affect the relationshipbetween the input work done by the tool on the bolt or nut and the

subsequent inservice clamping force between joint members.


''prevailing torque"when interference of somesort between male and female threads makes itnecessary to use a wrench simply to run a nutdown before the nut contacts the joint.

Because of our interest in stored energy, we alsolist the many ways the input work we do on thenuts is absorbed by the bolts and joint members.The work

Heats the parts and toolsEnlarges interference fit bolesBends and twists the boltsDilates the nutsPulls the joint members togetherOvercomes prevailing torqueDeforms the parts plastically (embedment andgasket creep)

It also deforms the bolts and joint memberselastically. It stretches the bolts and compresses


the joint. Only the work done deforming the partselastically ends up as stored energy, creatingclamping force, but then some of this work islost as the parts embed and relax.

We now realize why accurate control of theassembly process is so difficult. We would haveto predict or control such unpredictable things asthe friction forces between parts, the effortrequired to pull the joint together, the amount ofhole interference, and the elastic interactionsbetween bolts to achieve exactly the sameamount of residual preload in each bolt. I thinkthat it's safe to say that we'll never be able toafford the effort required to predict or controlthe many variables involved. This does not mean,however, that all is lost.

9Optimizing Assembly Results

A great deal of work has been and is still beingdone on assembly tools and techniques. There are


ways to get better results than we achieved withour example joint in this chapter. Most of thesetechniques are discussed in this section of thehandbook. This information will allow you topick the most appropriate assembly methods foryour own applications.

Regardless of the type of control you adopt,furthermore, there are a number of relativelysimple things you can do to improve yourchances of success during assembly operations.In fact, the things listed below can often makemore improvement than the use of elaborate orexpensive preload control equipment. As you cansee, most of the items on the list would tend tomake your assembly practices and proceduresmore consistent. Although you cannot predict orcontrol the many variables affecting results, youcan reduce their variation by being consistent.

1. Be as consistent as possible in your choice anduse of tools, procedures, calibration frequency,etc.


2. Train and supervise the bolting crews. This canhelp far more than a "better" lubricant or moreexpensive tool. Let the people know howimportant good results are and how difficult theyare to obtain. Enlist their help.

3. Make sure the fasteners are in reasonableshape. Wire-brush the threads if they're dirty andrusted. Use stainless steel bristles on alloy steelmaterials. Chase threads with a tap or die ifthey're damaged.

4. Use hardened washers between the turnedelement (nut or bolt head) and joint members.

5. If lubricants are to be used, make sure they areclean and fresh. Apply them consistently,applying the same amount to the same surfacesby using the same procedure. Preload scatter willbe minimized if lubricants are used on boththread and nut (or head) contact surfaces.

6. Run the nuts down by hand. If you can't, the


threads may need to be cleaned or chased.


7. Hold wrenches perpendicular to the axes ofthe bolts.

8. Apply torque at a smooth and uniform rate.Avoid a "stick-slip" situation as you approach thefinal torque. If necessary, back the nut off a littleso the specified torque can be reached with thewrench in motion.

9. If hydraulically powered wrenches are used, besure that adequate reaction points are used andthat the tools aren't twisting or cramping as aresult of cocked or yielding reaction surfaces.

10. Snug the joint first, with a modest torque;then tighten it. Try to align the joint membersbefore tightening the bolts.

11. Tighten from the center of bolt patternstoward the free edges if the bolt pattern isrectangular (as in a structural steel joint). Workin a cross-bolting pattern on circular or oval


in a cross-bolting pattern on circular or ovaljoints.

12. Keep good records of the tools, operators,procedures, torques, and lubricants used. Boltingis an empirical "art" at present. If you recorddetails of your "experiment," you'll have theinformation you need to make a controlledmodification in the procedure the next time if thefirst procedure doesn't give the results you want.

13. If possible, develop your own nut factorsexperimentally rather than relying on a table.Perform the experiment under actual jobconditions if possible, though a lab test on yourjoint (or a simulation of your joint) is better thannothing.

References

1. Landt, R., Preload Loss and VibrationLoosening, SPS Technologies, Jenkintown, PA,1979.

2. Conversation with Carl Osgood, consultant,


Cranbury, NJ, 1974.

3. Friesth, E. R., Performance characteristics ofbolted joints in design and assembly, SMETechnical Paper AD77-715, 1977.

4. High strength bolted joints, SPS FastenerFacts, SPS Technologies, Jenkintown, PA, SecIV-C-4.

5. Fisher, J. W., and J. H. A. Struik, Guide toDesign Criteria for Bolted and Riveted Joints,Wiley, New York, 1974, p. 179.

6. Meyer, G., and D. Strelow, How to calculatepreload loss due to permanent set in boltedjoints, Assembly Eng., February and March1972.

7. Ehrhart, K. F., Preloadthe answer to fastenersecurity, Assembly and Fastener Eng. (London),9(6):1923 (1971).

8. Prodan, V. D., More precise load-relief factorfor tightened screw connections. Vest. Mashin.,


54(1):2728 (1974).

9. Junker, G., Principles of the Calculation ofHigh Duty Bolted ConnectionsInterpretation ofGuideline VDI 2230, VDI Berichte, No. 220,1974, published as an Unbrako technical thesisby SPS, Jenkintown, PA.

10. High Strength Bolting for Structural Joints,Booklet 2867, Bethlehem Steel Co., Bethlehem,PA, p. 9.

11. Chesson, E., Jr., and W. H. Munse, Studiesof the Behavior of High Strength Bolts andBolted Joints, University of Illinois College ofEngineering, Engineering Experimental StationBulletin 469, 1964.

12. Mayer, K. H., Relaxation tests on hightemperature screw fastenings, Wire, 23:15(1973).

13. Hardiman, R., Vibrational looseningcausesand cures, Presented at the Using Threaded


Fasteners Seminar at the University ofWisconsinExtension, Madison, May 8, 1978.

14. Investigation of Threaded Fastener StructuralIntegrity, Report prepared by SouthwestResearch Institute, San Antonio, TX, underContract NAS9-15140, DRL Tii90, CLIN 3,DRD MA 129 TA, SWRI Project 15-4665,October 1977.

15. Blake, T. C., and H. J. Kurtz, Theuncertainties of measuring fastener preload,Machine Design, Sept. 30, 1965.

16. Van Campen, D. H., A systematic bolt-tightening procedure for reactor pressure vesselflanges, ASME First International Conference onPressure Vessel Technology, Delft, Netherlands,Sept. 29-Oct. 2, 1969, Part 1, pp. 131141.


17. Rumyantsev, D. V., V. D. Prodan, and A. F.Pershin, A tightening procedure for fasteningbolts applicable to high-pressure apparatus,Khim. Neftyanoe Mashinostroeine, 2:3-5(1973).

18. Tripp, O. S., and E. D. Stevens, Flangeparallelism criteria, Presented to the PVRCSubcommittee on Bolted Flanged Connections,San Francisco, CA, January 1983.

19. Bibel, G., and R. Ezell, An improved flangebolt-up procedure using experimentallydetermined elastic interaction coefficients, J.Pressure Vessel Technol., 114: 439 (1992).

20. Bibel, G., and R. Ezell, Bolted flangeassembly: Preliminary elastic interaction dataand improved bolt-up procedures, Draft of papersubmitted to the Pressure Vessel ResearchCommittee, May 18, 1993.


21. Bibel, G. D., and D. L. Goddard, Preloadvariation of torqued fasteners, a comparison offrictional and elastic interaction effects,Fastener Technol. Inter., Feb. 17, 1994.

22. Bickford, J. H., Introduction to the Designand Behavior of Bolted Joints, 3rd ed., MarcelDekker, New York, 1995.


24Tightening Groups of Fasteners in a Structureand the Resulting Elastic Interaction

George BibelUniversity of North Dakota, Grand Forks, NorthDakota

1Introduction

The relationship between fastener torque andresulting preload is a mathematical equationbased on summation of forces on the thread. Thisequation is described in detail elsewhere in thishandbook. Other than statistical variation ofdimensions due to manufacturing, all the termsin the equation are precisely known except forfriction.


Many factors affect the coefficient of friction.(Again this is a topic discussed in detailelsewhere in this handbook.) These factors turnthe mathematical relationship between torqueand preload into a statistical phenomenon bestdescribed with experimental data. The purpose ofthis chapter is to describe another phenomenonthat results in significant preload reduction. Thisreduction occurs when groups of fastenersinteract within a structure, and the phenomenonis known as elastic interaction or bolt cross talk.

In the sections that follow, elastic interaction isdefined, data is presented that demonstratesgreater preload variation than caused by friction,and elastic interaction is shown to depend on theratio of joint stiffness to fastener stiffness.Finally an improved flange bolt-up procedurethat compensates for elastic interaction isdescribed.

2


Elastic Interaction Defined

When a group of fasteners are tightened in ajoint, the elongation of each individual boltcauses it to structurally interact with the boltedjoint being compressed. (The tightened bolt actslike springs in parallel with the joint beingcompressed.) Subsequent tightening compressesthe joint further and reduces the preload infasteners previously tightened. This structuralinteraction is best explained with a series ofsimple illustrations.


Figure 1a shows bolt 2 initially tightened. In thiscase the joint consists of a bolted flange andgasket. Local compression of the joint occursunder bolt 2. When the adjacent bolt 1 istightened, additional compression of the joint isshown in Figure 1b. The preload is reduced inbolt 2 as shown. When bolt 3 is tightened,additional compression of the joint occurs, withadditional loss in preload of bolts 1 and 2 asshown in Figure 1c.

Elastic interaction always results in a loss of boltpreload. This differs from preload variation dueto friction. Frictional variation results in arandom preload, higher or lower, about somemean value. If the fasteners are torqued,frictional variation will be superimposed on theelastic interaction preload reduction. If thefasteners are loaded to a precise value withhydraulic tensioners or torqued to a precise


preload value by monitoring bolt strains withultrasonics, micrometers, strain gages, etc., thenonly preload reduction due to elastic interactionoccurs.

3Elastic Interaction Data

3.1Procedure for Obtaining Data

The following procedure can be used to obtainelastic interaction data.

1. Define a method to accurately measure bolttension. This can be done with strain gagesmounted on the fasteners, ultrasonics, digitalmicrometers, or strain-gaged washers.

2. Define the final preload desired in the bolts, interms of microstrain, stress, or tension,whichever is appropriate for the means used tomeasure tension.


3. Tighten the bolts one at a time to this preload,following the normal sequence. The bolts in acircular pressure vessel flange, for example, areusually tightened in a cross or star patternsequence. Alternatively, the bolts can betightened in several passes, each at a highertorque than the last, to probe the interactions thatoccur during the normal assembly operations.

4. Measure the tension in each bolt when it isfirst tightened to the full or preliminary preloadand after you have tightened each of the otherbolts. Record the data to a data file. This data canlater be displayed, perhaps following the formatof Table 1, to reveal the effects of the elasticinteractions between bolts.

This procedure results in many more bolt tensionreadings than are normally taken. A joint with sixbolts would have a total of 21 readings. The firstbolt would be measured once when it is firsttorqued to its target load and five additionaltimes, once after each of the other five bolts is


torqued. The second bolt would be measuredonce as it reaches its target and four additionaltimes, once for each bolt tightened afterwards.The subsequent bolts would be measured fourtimes, three times, twice, and once, for a total of21 readings for the six bolts.

There may be obvious patterns to the interactionsthat may reduce the number of readings required.For example, for the data presented here,significant interactions occurred only in theadjacent bolts.

3.2An Example

Elastic interactions were measured on severalstandard ANSI pipe flanges (24 in. Class 150 andtwo 16 in. Class 300 [9]). Testing was done withincreased joint flexibility (by varying the stiffnessof the gasket) and increased flexibility of thebolts (by using longer studs with spacers.)

The target bolt strain was either 850 or 1010


microstrain. This converts to about 25,000 psi or30,000 psi bolt stress by multiplying stud strain,measured with strain gages, by the modulus,29,300,000 psi [1]. The studs were tightened in a"standard" star pattern as shown in Figure 2.


Figure 1(a) Tighten bolt 2.


(b) Tighten bolt 1; preload in bolt 2 is reduced.

(c) Tighten bolt 3; preload is reduced in bolts 1 and 2.


Figure 2Bolt number identification. Bolts weretightened in the following sequence: 1,

11, 16, 6, 2, 12, 17, 7, 3, 13, 18, 8, 4, 14,19, 9, 5, 15, 20, 10

(four-point "star" pattern).

Typical elastic interaction data are shown in Figures 3 and 4. The final bolt strain isplotted mafter all of the studs are tightened to an initial value of approximately 850mstrain. Figures 3 and 4 represent extremes in joint flexibility: Figure 3 data are fora stiff joint with no gasket and little interaction, whereas Figure 4 is for a flexiblejoint (same flange with a soft gasket) with large interaction.

Another way to describe interaction is with a defined bolting efficiency, which isequal to the average bolt stress (or strain) divided by the design bolt stress. ForFigure 3, even without a gasket, the bolting efficiency is calculated to be 84%. Forthe flexible joint elastic


Figure 3Elastic interaction data for a stiff joint. A 24 in. Class 150 ANSI pipe flange with no gasket. Bolting efficiency is 84%.


Figure 4Elastic interaction data for a flexible joint. Same flange as in Figure 3 except that a flexible spiral wound gasket

was used. Bolting efficiency is 50% after the first pass and only 60% after tightening all of the bolts twice to 100%of target load.


Figure 5Elastic interaction for a modified four-point star sequence. Bolts were tightened as follows.

Pass 1: One-third of 30,000 psi target stress with four-point star pattern starting at bolt 1. Pass 2: Two-thirds of30,000 psi target with four-point star pattern starting at bolt 3. Pass 3: 30,000 psi target starting at bolt 6.

Pass 4: 37,500 psi on consecutive bolts starting at bolt 1, moving clockwise.


interaction data graphed in Figure 4, the bolting efficiencyis seen to be only 50% after the first pass (i.e., tighteningall of the bolts once) and 69% after the second pass.

Preload reduction due to elastic interaction occurs inpatterns based on bolt tightening sequence. The first boltstightened have the most preload reduction. The last boltstightened have none. Bolts tightened between first and lastsuffer an intermediate amount of elastic interaction.Therefore, changing the sequence of the bolt tighteningpattern or the number of bolts tightened simultaneouslywill change the amount of preload reduction. Figure 5shows the results of a "scrambled" bolting sequence. Thecyclic pattern of interaction can also be expected to bedisturbed whenever the star pattern is applied to a groupof bolts that is not a multiple of four. (Of course, otherpatterns can be envisioned. The star pattern can be appliedin groups of three with bolts being tightened every 120°.)

Another way to present the data is to tabulate the historyof each stud throughout the bolt-up procedure, as shownin Table 1. The data presented this way show theproximity of interaction. Another mechanism, describedlater as rocking, can also be observed.


Every time a bolt is tightened, the strain in all 20 bolts ismeasured and recorded to a data file. A one-passprocedure would create 20 × 20 or 400 strain readings,19 of which are zero on the first pass only. (Tighteningthe first bolt results in one actual reading and 19 zeroTABLE 1 Elastic Interaction Data for Passes 1 and 2 of a 20-Bolt FlangeTightened in a Star Pattern (see Figure 2) as Follows: Bolts 1, 11, 16, 6, 2, 12,17, 7, 3, 13, 18, 8, 4, 14, 19, 9, 5, 15, 20, 10

Pass 1Bolts 16 strain data493.00 0.00 0.00 0.00 0.00 501.00243.00 485.00 0.00 0.00 0.00 469.00251.00 492.00 0.00 0.00 0.00 485.00219.00 508.00 0.00 0.00 0.00 501.00235.00 532.00 0.00 0.00 0.00 282.00172.00 305.00 493.00 0.00 0.00 219.00180.00 305.00 493.00 0.00 0.00 219.00110.00 289.00 516.00 0.00 0.00 235.00110.00 305.00 532.00 0.00 0.00 195.00118.00 250.00 360.00 501.00 0.00 117.00125.00 250.00 352.00 501.00 0.00 117.00

71.00 211.00 360.00 516.00 0.00 133.0071.00 219.00 367.00 532.00 0.00 133.0071.00 211.00 297.00 376.00 492.00 39.0079.00 219.00 305.00 383.00 492.00 47.0024.00 140.00 266.00 383.00 508.00 47.0032.00 140.00 273.00 383.00 524.00 47.00

Bolts 712 strain data8.00 0.00 0.00 0.00 492.00 0.00


8.00 0.00 0.00 0.00 508.00 0.008.00 0.00 0.00 0.00 297.00 493.000.00 0.00 0.00 0.00 305.00 501.00

509.00 0.00 0.00 0.00 281.00 517.00509.00 0.00 0.00 0.00 297.00 524.00




524.00 0.00 0.00 0.00 258.00 337.00532.00 0.00 0.00 0.00 258.00 337.00329.00 493.00 0.00 0.00 211.00 329.00297.00 501.00 0.00 0.00 211.00 337.00297.00 517.00 10.00 0.00 211.00 282.00305.00 517.00 0.00 0.00 219.00 282.00251.00 329.00 501.00 0.00 133.00 258.00180.00 313.00 501.00 0.00 140.00 258.00188.00 313.00 509.00 0.00 156.00 251.00188.00 321.00 517.00 0.00 148.00 251.00180.00 274.00 321.00 516.00 54.00 188.00Bolts 1318 strain data

0.00 0.00 0.00 501.00 0.00 0.000.00 0.00 0.00 524.00 0.00 0.000.00 0.00 0.00 516.00 0.00 0.000.00 0.00 0.00 297.00 508.00 0.000.00 0.00 0.00 305.00 508.00 0.000.00 0.00 0.00 313.00 524.00 0.00

492.00 0.00 0.00 266.00 532.00 0.00508.00 0.00 0.00 211.00 289.00 501.00516.00 0.00 0.00 219.00 297.00 509.00524.00 0.00 0.00 227.00 297.00 516.00320.00 477.00 0.00 157.00 266.00 532.00328.00 493.00 0.00 157.00 211.00 313.00328.00 508.00 0.00 164.00 211.00 313.00336.00 508.00 0.00 164.00 219.00 321.00


281.00 305.00 493.00 94.00 156.00 297.00281.00 313.00 501.00 102.00 141.00 235.00258.00 313.00 509.00 102.00 141.00 235.00Bolts 1920 strain data

0.00 0.000.00 0.000.00 0.000.00 0.000.00 0.000.00 0.000.00 0.000.00 0.000.00 0.000.00 0.000.00 0.00

486.00 0.00486.00 0.00509.00 0.00509.00 0.00290.00 483.00290.00 485.00



ContinuedTABLE 1

Pass 2Bolts 16 strain data994.00 39.00 86.00 329.00 516.00 55.001002.00 39.00 86.00 337.00 532.00 62.001017.00 47.00 94.00 337.00 563.00 70.001033.00 39.00 78.00 266.00 164.00 1001.00532.00 1001.00 0.00 172.00 93.00 970.00540.00 1009.00 0.00 180.00 93.00 1001.00517.00 1048.00 0.00 180.00 101.00 1009.00524.00 1087.00 0.00 180.00 70.00 563.00423.00 618.00 993.00 86.00 31.00 422.00431.00 618.00 1001.00 86.00 23.00 438.00321.00 610.00 1032.00 86.00 31.00 446.00321.00 626.00 1072.00 94.00 31.00 360.00345.00 539.00 696.00 1002.00 15.00 188.00352.00 539.00 704.00 994.00 7.00 188.00165.00 500.00 719.00 1025.00 7.00 195.00165.00 500.00 727.00 1056.00 15.00 211.00188.00 500.00 626.00 704.00 1017.00 39.00204.00 516.00 626.00 689.00 1017.00 47.00

47.00 344.00 594.00 704.00 1048.00 47.0047.00 344.00 602.00 712.00 1071.00 55.00

Bolts 712 strain data196.00 274.00 321.00 516.00 54.00 196.00211.00 274.00 235.00 188.00 985.00 24.00


219.00 274.00 235.00 195.00 1001.00 24.0039.00 157.00 188.00 195.00 1017.00 24.0039.00 165.00 204.00 219.00 1024.00 24.0047.00 180.00 204.00 133.00 547.00 1002.0039.00 180.00 204.00 141.00 555.00 1017.00

994.00 71.00 94.00 102.00 555.00 1041.00986.00 79.00 110.00 109.00 555.00 1041.001009.00 79.00 126.00 117.00 461.00 603.001017.00 79.00 118.00 117.00 469.00 618.00626.00 994.00 32.00 39.00 391.00 618.00579.00 1010.00 32.00 47.00 399.00 618.00579.00 1025.00 40.00 55.00 414.00 524.00587.00 1033.00 40.00 55.00 414.00 524.00501.00 650.00 994.00 8.00 250.00 493.00352.00 611.00 1010.00 8.00 258.00 493.00344.00 611.00 1025.00 8.00 281.00 493.00360.00 618.00 1025.00 8.00 289.00 493.00360.00 525.00 673.00 1001.00 78.00 384.00Bolts 1318 strain data266.00 320.00 524.00 110.00 149.00 211.00156.00 266.00 516.00 110.00 164.00 211.00




133.00 156.00 172.00 1002.00 31.00 70.00140.00 164.00 180.00 1009.00 23.00 78.00140.00 172.00 188.00 1041.00 31.00 102.00

39.00 70.00 125.00 1017.00 31.00 102.0039.00 62.00 86.00 485.00 1001.00 8.0039.00 70.00 94.00 493.00 1009.00 8.0039.00 62.00 94.00 501.00 1041.00 8.00

1001.00 7.00 39.00 399.00 1033.00 8.001025.00 7.00 47.00 305.00 563.00 1009.001048.00 15.00 47.00 313.00 563.00 1017.001056.00 7.00 47.00 321.00 571.00 1033.00618.00 985.00 8.00 172.00 508.00 1064.00625.00 1017.00 8.00 172.00 415.00 642.00641.00 1048.00 8.00 180.00 422.00 642.00649.00 1048.00 8.00 180.00 422.00 649.00547.00 626.00 1002.00 63.00 282.00 610.00555.00 633.00 1025.00 70.00 282.00 524.00532.00 649.00 1041.00 78.00 289.00 524.00Bolts 1920 strain data180.00 141.00188.00 141.00149.00 141.00149.00 149.00141.00 102.00149.00 102.00

47.00 71.00


47.00 63.0063.00 78.0071.00 78.008.00 31.000.00 31.008.00 39.000.00 39.00

1002.00 16.001002.00 16.001018.00 8.001049.00 8.00642.00 1017.00634.00 1017.00

entries. Tightening the second bolt gives a strain gagereading for the second bolt, an interaction reading for thefirst bolt and 18 zero readings, etc.)

Pass 2 for Table 1 is read as follows: The first 20 rows ofcolumn 1 are the strain history of bolt 1 while tighteningall other bolts according to the star pattern 1, 11, 16, 6, 2,12, 17, 3, 13, 18, 8, 4, 14, 19, 9, 5, 15, 20, 10. Row 5shows the first significant interaction when adjacent bolt2 is tightened and bolt 1 strain drops from 1033 to 532.Additional interaction is seen in bolt 1 when the otheradjacent bolt (#20) is tightened (row 19) and the strain isseen to drop in bolt 1 from 204 to 47 mstrain.


3.3What the Data Reveals

Presenting the strain history for each boltdemonstrates that most of the interaction is local(atleast for the pipe flanges tested) in that onlythe two adjacent bolts have an appreciable effect.It is suggested that this is true in general, but oneshould be careful to extrapolate data of this typeto other structures and bolting configurations.

In addition to frictional variation and elasticinteraction, a third effect can be noticed duringthe bolt-up of pipe flanges. A rigid-body motionor rocking of the flanges can occur. Although thefollowing bolting pattern was never used, it bestillustrates the condition.

Consider 12 bolts located like numbers on aclock. If the 12 o'clock bolt is tightened first,followed by positions 5, 6, and 7, then the 12


o'clock bolt will in fact increase in preload. Thisoccurs because bolts 5, 6, and 7 try to open theflange as shown in Figure 6.

Using the star pattern on a 20-bolt flange withtightening sequence 1, 11, 16 and 6, the tensionin bolt 1 did, in fact, increase by up to 48% insome cases. This worst-case condition occurredwhen the bolts were tightened to the target loadon the first pass (no incremental tightening). Thisshows that the potential exists for a bolttightened near yield to yield. This increase in boltload also decreases the reduction of preload inbolt 1 during testing. Bolt 1 did consistently haveless reduction than the other bolts. Other boltingpatterns as extreme as the case described abovecould be expected to create a significant increasein preload in the first bolt tightened.

The rocking disappears rapidly or after enoughbolts are tightened to make the incrementalloading associated with a single boltinsignificant. Other ways to deal with rocking are


to

1. Use incremental tightening. This seemed toeliminate the problem.

2. Use symmetrical tightening. For example, usefour hydraulic wrenches, one every 90°.

3. Bring the flange faces into contact by anyavailable means before tightening the bolts. Thesimplest way may be to apply a small amount oftorque, by hand, to each of the bolts.

Bringing the flanges flush may create a differentbolting problem. In a stiff piping system withmisalignment, a significant portion of the boltload may in fact be used up to correct anyflange/piping misalignment. This may leave theactual bolt load available for clamping the jointor sealing a gasket far below design.


Figure 6Flange opens up if several bolts are

tightened on the bottom of the flange.This can increase the preload by as much

as 48% on the top bolt.


Figure 7Single pass to 25,000 psi stress on a 16 in. Class 300 ANSI flange without a gasket.

3.4The Effect of Flange Surface Gaps and Misalignment

Two ''identical" pairs of 16 in. Class 300 flanges were tested, with results asshown in Figures 7 and 8. Some preliminary conclusions can be drawnabout flange surface gaps and misalignment from comparing test data for theflange pairs.

Data from bolts 1, 11, 16, and 6 on each pair of 16 in. flanges show "soft"spots. Stress reductions of up to 67% can be seen in bolts 6 and 16 forflange 1 (see Fig. 7). A reduction of 87% is present in bolt 11 of flange 2


(see Fig. 8). Differences among apparently identical flange pairs arepresumably due to surface gaps and misalignments.

Flange surface misalignment apparently results in gaps of unknown size andlocation. The flange is not as stiff when the gap is open but becomes stifferas the gap is closed. As the joint becomes more rigid with increased boltloading and gap closure, the interaction is reduced. An infinitely stiff flangewithout a gasket would have no interaction, conversely, very flexible bolts,made for example out of rubber, would also have no interaction. (Therubber bolts would be incapable of compressing the joint.)

The grip length of the bolts used in this example was 4.86 in. This results ina bolt stretch of about 0.005 in. It is easy to imagine flange distortion due towelding and/or machining resulting in gaps of this order or a significantfraction thereof.

3.5Finite-Element Analysis of Elastic Interaction

Discrepancies between finite-element (FE) modeling of interaction andexperimental data can also be explained with surface misalignment. Anaxisymmetric model with nonaxisymmetric loading was used for the finite-element analysis. The method used is described in Ref. 2. The model doesnot account for any flange misalignment.


Figure 8Identical to Figure 7 except that data are for a second pair of 16 in. flanges. Differences between Figures 7 and 8 are

presumably due to nonuniform contact in the flange faces.

The model predicted elastic interaction much less than that observed experimentallywhen only four bolts were tightened. However, with 19 of the 20 bolts tightened, theFE model predicted elastic interaction very close to the experimental value. With 19bolts tightened, any gaps that exist are closed and the flange is much stiffer.

In Ref. 2, the finite-element model gave very good results with experimentallymeasured interaction data for a gasketed joint. In a gasketed joint, the gap isimmediately filled by the relatively soft gasket material and is therefore less of afactor.

Reportedly FE analysis has been used with modest success to predict elasticinteraction in a nuclear reactor. The flange of a nuclear reactor is essentially metal-to-metal contact. (Although there is a gasket, it is structurally insignificant.) Ibelieve that FE analysis was of limited success because of the inability to accurately


believe that FE analysis was of limited success because of the inability to accuratelymodel any misalignments. Later in this chapter an algorithm is presented tocompensate for elastic interaction. Since the algorithm is based on experimental(instead of analytical) determination of interaction coefficients, there is automaticadjustment for any misalignments that occur.

3.6More Data Are Available

Detailed elastic interaction data are presented in Ref. 3 for 24 tests on pipe flangesdescribed above.

Data for a more complex flange, a heat exchanger with 48 1 3/4 in. diameter studs(effective length of 22.7 in.) and a spiral-wound gasket are given in Ref. 4, which isa case study of a problem flange. The required torque was initially evaluated withwell-known torque-preload equations. Measurements with ultrasonics found theactual bolt stretch to be much less than expected. This decrease in bolt preload wasattributed to elastic interaction. Retorquing to


compensate for elastic interaction increased boltstretch by 44%. Subsequently the exchangeroperated for four years without leading. Beforethe change in bolt-up procedure, leakage wasexpected every few weeks.

Interaction data for an automotive engine blockare given in Ref. 5. (This test is discussed inmore detail in the final section of this chapter.)

3.7Design to Minimize Elastic Interaction

By varying the joint stiffness and the stiffness ofthe bolts during testing, a pattern in the effect ofthe ratio of the stiffness of the joint to thestiffness of the fasteners was determined. Elasticinteraction depends on the relative stiffness ofthe joint compared to the stiffness of thefastener. For very stiff joints or very flexiblebolts there is little interaction (i.e., rubber bolts


have no interaction because no joint compressionoccurs). For very flexible joints or stiff boltsthere is a great deal of interaction. This givessome insight into the effect of joint design onelastic interaction and assembly.

In this context the joint refers to the structure andthe gasket, if any. Although any sort of gasket isexpected to be significantly more flexible than asteel pipe flange, the data show much interactioneven without a gasket. The flange will dimpleand bend under the contact load. Although it ishard to imagine much deflection of the flangecompared to that of a gasket, the flangedeflections are relative to the equally smalldeflections of the bolt.

The stiffness of the gasket is not necessarilyinsignificant compared to that of the bolts. Theaxial stiffness of an elastic rod is AE/L, where Ais the cross-sectional area, E is the modulus ofelasticity, and L is the length of the rod.Although this is a very simplistic model to


describe the axial stiffness of a nonlinear gasket,it does correctly illustrate the effect of gasketthickness and area. L for the gasket is muchsmaller than that of the bolts. Also, the cross-sectional area of the gasket is typically 23 timesgreater than that of the bolts. This somewhatcompensates for the fact that the modulus of theusually nonmetallic gasket material is much lessthan the modulus of the steel bolts. Gasketstiffness data are given in Ref. 6.

4Comparison of Frictional Variation of Preloadwith Elastic Interaction

After testing with strain-controlled loading toisolate elastic interactions, follow-up tests wereconducted to focus on frictional scatter in thesame set of bolts.

In the first test the preload scatter created byrepeatedly torquing an individual stud wasobserved. For this type of test the same


studwashernut configuration was tightened,loosened, tightened, loosened, etc. This was donewithout any relubrication or any other changes.Torque was measured with a torsional load cellusing internal strain gages. This type of testingeliminates the effects of elastic interaction andproved to be quite repeatable. Preload scatterwas ±5%. Typical scatter bands for two bolts arepresented in Figure 9. Although each boltexhibited about the same low amount of preloadscatter, the slope of the torquestrain curve isnoticeably different for each bolt.

The test was then repeated for 14 different studs.The scatter bands are shown in Figure 10. The 14bolts when plotted together had scatter bands ofabout ±13%. This result is considered consistentwith work done by others who have studied thestatistical scatter of preload vs. torque. Note thatcare must be taken to separate out the elasticinteraction from the frictional scatter.

It is hypothesized that minute misalignments of


all of the surfaces (i.e., threads, nut face, washer,flange, etc.) due to manufacturing variationcause greater scatter in a group of bolts than in asingle bolt. A single bolt tested repeatedly alwayssees the same misalignments. For groups ofbolts, the misalignments are random andnonrepeatable.


Figure 9Typical scatter bands for two individual bolts torqued repeatedly. This testing was done to

compare the preload variation due to friction to that due to elastic interaction. Althougheach bolt is quite repeatable, note how the strain vs. torque line has a different slope for

each bolt.


Figure 10Scatter bands for a group of 14 bolts thatwere torqued. This testing was done toevaluate the repeatability of an individualbolt compared to a group of bolts and tocompare frictional variation in preload to

elastic interaction variation.


TABLE 2 Preload VariationCasea Preload variationI ±5%II ±13%III +035%IV +059%V +095%VI +075%aCase I: Preload variation in a single stud torquedrepeatedly (no interaction).Case II: Preload variation in a group of 14 studs torquedrepeatedly one at a time (no interaction).Case III: Preload variation of 20 studs in a stiff structurewithout a gasket (24 in. ANSI Class 150 pipeflange). Torquing is done to precise strain valuesby using strain gages mounted on the studs (nofrictional variation).Case IV: Same as above except joint is made moreflexibleby adding a 0.060 in. thick sheet gasket.Case V: Same as above except joint is made moreflexibleby adding a 0.180 in. thick spiral wound gasket.Case VI: Same as above except studs are made moreflexible


by using 4 in. spacers and longer studs. Thisapproximately doubled the grip length.Note: All testing in this data set was done with the samestuds and same structure (i.e., 24 in. pipe flange).

Table 2 summarizes a comparison of preloadvariation for friction only and elastic interactiononly for various degrees of joint flexibility. Jointflexibility is varied by (1) using no gasket in thepipe flange; (2) using a thin, stiff sheet gasket;(3) using a thicker, flexible spiralwound gasket;and (4) using a spiral-wound gasket and longerstuds with 4 in. spacers. The addition of thegasket makes the joint more flexible. Addingspacers to the studs makes the studs moreflexible.

A few trends are seen that may affect bolt-upstrategy.

1. Frictional variation is a random phenomenonwith equal probability of plus or minus variation.

2. Elastic interaction can result only in the


reduction of individual bolt preloads. Care wasalways taken to separate frictional preloadvariation from elastic interaction variation. Inmost real-world applications, the two effectswould, of course, superimpose.

3. Adding a gasket makes the joint more flexibleand results in more elastic interaction.

4. Making the studs longer (i.e., making the studsmore flexible) reduces elastic interaction.

5Improved Flange Bolt-Up Procedures

5.1The Procedure Summarized

The data given in Table 1 can be used in a matrixequation to compute elastic interactioncoefficients. These coefficients define therelationship between initial preload in a givenbolt,


the final load in that bolt (after interaction), and the initial loads of the other bolts thatcause interaction. This relationship can be put into equation form and expanded to includeall the bolts of any size flange. Each bolt is represented by a linear equation that defines itsrelation to the other bolts in the flange. These equations can be easily solved with readilyavailable engineering software.

Any bolted connection will have a set of equations that can be written in matrix form as

where

{Sf}=is an n × 1 matrix correspovding to the final bolt load

{Si}=is an n × 1 column matrix describing the initial bolt load

[A]=

is an n × 1 matrix of interaction coefficients thatcorrectly transforms Si in to Sf; n equal the number ofboth in the flange

Coefficients of the [A] matrix can be found experimentally by tightening the bolts once andmeasuring the interactions. Once the coefficient matrix is found, the matrix equation can besolved for the required initial bolt loads that result in any desired final bolt loads as shown.

Typically each bolt must be tightened to a different initial load to create uniform final loadsin each bolt.

5.2Determining Interaction Coefficients

Consider the hypothetical elastic interaction data shown in Table 3 for a three-bolt system.Each bolt is initially tightened to 40,000 psi. The history of bolt 1 is as follows: Initiallytighten to 40,000 psi. Bolt 1 drops to 25,000 psi after bolt 2 is tightened. Bolt 1 then dropsto 22,000 psi due to interaction when bolt 3 is tightened.

Bolt 1:


TABLE 3 Hypothetical Interaction Data for a Three-Bolt SystemaResulting load (psi)

Tighten bolt Bolt 1 Bolt 2 Bolt 31 40,0002 25,000 40,0003 22,000 27,000 40,000aThe data are used in the text to demonstrate how the interactioncoefficients are calculated.


where a11 describes the interaction of bolt 1 on itself.

where a12 is the interaction of bolt 2 on bolt 1.


Bolt 2:


Since bolt 1 is already tightened, it does not affect bolt 2.

where a22 is the interaction of bolt 2 on itself.



Bolt 3:

Equation (1) becomes

5.3Procedures for Large Flanges

The procedure described above can be extended for use onlarge conventional flanges. A 20-bolt flange, for example,would be represented by a system of 20 equations with 20unknowns. The method was first described by Van Campen[7], who developed interaction equations for reactor vesselsusing hydraulic jack pressures and bolt strain. He alsoassumed flange symmetry, which reduced the number of


equations.

Van Campen assumed the interaction coefficients remainconstant and can be determined by a single test. I havedeveloped a more accurate procedure that involves twosingle-pass tests


and results in interaction coefficients that areprecise enough to allow the flange to besubsequently tightened with a final preloadscatter of ±2%. This procedure is as follows.

1. Test 1: Tighten all bolts to initial preload orstress equal to that ultimately desired for thefinal preload. Let us use, by way of example, astress goal of 25,000 psi. The final average boltstress during this test is, of course, much lessthan 25,000 psi.

2. Use the data from Test 1 to computepreliminary interaction coefficients (i.e.,establishing a preliminary [A] matrix).

3. Using Eq. (2), predict the initial bolt loadspresumably required to obtain a uniform finalbolt load of 25,000 psi.

4. Test 2: Loosen all the bolts and then retightenthem to the initial stress values predicted in step


them to the initial stress values predicted in step3.

5. Use the data obtained in Test 2 to update the[A] matrix and then recompute the initial stressvalues required to create a final stress of 25,000psi in this set of bolts with Eq. (2).

The initial stress values can now be used onsubsequent tightening of these bolts, with aresulting scatter in stress of only ±2% of 25,000psi. Note that if the interaction coefficients wereconstant, as assumed by Van Campen, or ifgreater scatter could be tolerated, then only onetest would be needed to determine the correctinitial loads. Because the coefficients changesome, a second test is required to improve theaccuracy of the coefficients and ensure ±2%uniformity in future bolt-ups.

5.4An Example

Experimental testing was completed on a 16 in.


Class 300 RFWN (raised face welding neck)flange with twenty 1 1/4 in. (31.75 mm)diameter bolts. The flanges and instrumentationare as described earlier.

Initial testing was done without a gasket(simulating a nuclear reactor head where thegasket is structurally insignificant). Subsequenttesting was completed with a flat sheet gasket ona second pair of 16 in. flanges and on a pair of24 in. flanges. Each test consisted of onetightening pass using the four-point startightening sequence (see Fig. 2). The first fourbolts (1, 11, 16, 6) were torqued together,simulating simultaneous tensioning. This wasdone to eliminate flange rocking and excessivestress on the first bolt. Other bolting patternscould be used to eliminate this problem (e.g.,hand tightening of all bolts to about 25 ft-lb).Tightening four bolts simultaneously eliminatedthe rocking problem by bringing the two flangesurfaces rigidly against each other. Technically


this resulted in what might be called a two-passprocedure; however, this first pass could be donewithout torquing equipment, without anytimeconsuming torque or bolt stretchmonitoring, and without a special pattern. If theflange is brought flush with the first four boltstightened simultaneously as described above,then perhaps the process should more accuratelybe described as a 1.2-pass procedure. Usingmultiple tensioners would result in a true one-pass procedure.

The procedure used was the one describedearliera series of two tests designed toexperimentally obtain and improve thecoefficients in the [A] matrix, which were thenconfirmed in a third test. The initial bolt stressfor each of the three tests is shown in Table 4.

5.5Discussion of the Results

A three-test sequence completed with metal-to-


metal contact (no gasket installed) is shown inFigure 11. The graph shows the final bolt stressfor each one-pass procedure. In test 1 all boltswere tightened to approximately 25,500 psi(172.4 MPa). In tests 2 and 3, the bolts weretightened as shown in Table 4.

The first test shows large variations in bolt stresswith an average value of only 18,500 psi (127.6MPa). Stress load reductions of up to 67%occurred. The second test in the series


TABLE 4 Initial Bolt Stress in a Three-Test SeriesaTest results (psi)

Bolt tightening sequence Test1 Test2b Test 3c1 25,500 42,090 42,21011 25,500 41,190 40,50016 25,500 47,190 41,8206 25,500 46,620 43,1102 25,500 34,170 34,62012 25,500 31,110 32,28017 25,500 34,170 34,0807 25,500 34,260 34,2303 25,500 31,650 32,31013 25,500 30,210 30,81018 25,500 30,570 32,0708 25,500 32,040 32,8204 25,500 30,630 30,69014 25,500 30,630 31,23019 25,500 30,630 30,6309 25,500 29,790 30,5105 25,500 25,500 25,50015 25,500 25,500 25,50020 25,500 25,500 25,50010 25,500 25,500 25,500aFinal stress values are plotted in Figure 11.bCalculated from Test 1.cCalculated from Test 2.


Source: Reprinted with permission, from Welding Rsearch CouncilBulletin 408, Jan. 1996.

has an average stress load of 25,500 psi (172.4MPa) with variations of +17 to 6.5%. Theaverage bolt load for the final test is at the targetstress of 25,500 psi (172.4 MPa) with scatter ofonly ±2%. This small amount of scatter can beattributed to instrumentation noise (±0.9%) andthe inability to hand tighten to precise values(±1.8%).

Figure 12 shows similar results for the secondpair of 16 in. flanges. Figure 13 shows resultsfor a three-test sequence conducted with a sheetgasket 0.060 in. (1.52 mm) thick.

Analysis of the coefficient matrix shows that themajority of interaction is caused by the adjacenttwo bolts. However, the other coefficients werenot negligible. Collectively they also contributeto the final bolt load. A complete 20 × 20interaction coefficient matrix is shown in Table 5


as an example. Table 5 is the [A] matrix for test 3of a test series done on a 16 in. flange without agasket.

The amount of interaction changes for differentloading conditions. High clamping force createsless interaction than bolting with low clampingforce.

(Note: The coefficients used for the procedureare ordered by tightening sequence number ratherthan bolt number. A1,5 is the interactioncoefficient for bolt 1 when the fifth bolt of thesequence is tightened. The fifth bolt is theadjacent bolt 2. Ordering in this manner makesthe matrix inversion easier. The matrix becomesupper triangular with all 1s on the diagonal.)

As seen from Table 5, primary coefficientsaverage 0.334 for the first four bolts tightened.The averages for the second, third, and fourthgroups are 0.290, 0.254, and 0.215, respectively.This decreasing pattern is also consistent with the


theory that interaction decreases with increasedclamping force.


Figure 11Final stresses for a three-test sequence (single-pass tests) on a set of 16 in. flanges without a gasket.


Figure 12Final stresses for a three-test sequence (single-

pass tests) on a second set of 16 in. flanges without a gasket.


Figure 13Final stress for a three-test sequence (single-pass tests) on a 16 in. flange with a sheet gasket.

Additional testing was done to illustrate the stability of the coefficients.For example, if bolt A is tightened to 30,000 psi (206.9 MPa) and adjacentbolt B is tightened to 36,000 psi (248.2 MPa), the coefficient is

If bolt A is increased to 36,000 psi (248.2 MPa) or decreased to 24,000 psi(165.5 MPa), the coefficient remains almost constant:


Even though bolt A was varied by ±20%, the coefficients varied only from0.0 to 5.6%. A 5.6% variation in the interaction coefficient results in onlya 4.4% error in the predicted stress of bolt A after interaction. The methodworks because the coefficients become stable as the correct solution isapproached.

5.6Tests on a Noncircular Joint

The question often asked is: Is a circular symmetrical flange needed for thematrix method to work? The matrix method was applied to an automotivecylinder head and engine block [5], a highly irregular and noncircularshape. (I do not believe the circular shape of a pipe flange is of anysignificance to the matrix method. There is nothing symmetrical abouttightening one bolt at a time in a circular flange.)


TABLE 5 20 × 20 Coefficient Matrix for Test 3 of the Three-Test Series Without a Gasket1 11 16 6 2 12 17 7 3 13

1 1.0000 0.0000 0.0000 0.0000 0.3292 0.0075 0.0191 0.0000 0.0000 0.000011 0.0000 1.0000 0.0000 0.0000 0.0000 0.3657 0.0066 0.0262 0.0000 0.007916 0.0000 0.0000 1.0000 0.0000 0.0140 0.0075 0.3120 0.0000 0.0000 0.05316 0.0000 0.0000 0.0000 1.0000 0.0140 0.0149 0.0000 0.3316 0.0436 0.00692 0.0000 0.0000 0.0000 0.0000 1.0000 0.0224 0.0066 0.0203 0.2670 0.000012 0.0000 0.0000 0.0000 0.0000 0.0000 1.0000 0.0191 0.0271 0.0076 0.261617 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 1.0000 0.0000 0.0076 0.00267 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 1.0000 0.0218 0.01573 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 1.0000 0.007913 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 1.000018 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.00008 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.00004 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.000014 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.000019 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.00009 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.00005 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.000015 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.000020 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.000010 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000Source: Reprinted with permission from Welding Research Council Bulletin 408, January 1996.


The bolts were threaded into the engine. This made it very difficult to mount strain gages onthe fasteners. An alternative method was chosen to apply the matrix method. A plastic sensorabout 0.005 in. thick (described in Ref. 8) was placed on top of the gasket to monitor gasketcontact stresses. Fifteen sensors were paired up with 15 bolts. Instead of strain gagereadings on the bolt before and after elastic interaction, gasket contact stress before and afterinteraction was used in the matrix equations.

One sensor consistently gave a very low value. The matrix method predicted an excessivetorqueenough to yield, if not break, the bolt. The assumed cause was a lack of flatness of theengine block and/or head. Instead of aiming to achieve uniformity, the goal was changed toobtaining a calculated distribution of sensor readings in a single pass by using the matrixalgorithm. This test was considered successful and is described in detail in Ref. 5. All of thissuggests that one application of the matrix method is to predict or ''dial in" any favorablehead gasket load distribution by varying bolt load, size, torque, or location.

6Two-Pass Matrix Algorithm

The one-pass procedure worked very well for rigid flanges (metal-to-metal contact and flat


sheet gaskets). It did not work as well with flexible systems such as spiral-wound gaskets. Atwo-pass matrix procedure was developed next.

where

[A] =interaction coefficient matrix for the first pass{Si}

=the initial value to which each bolt is tightened in thefirst pass

[B] =interaction coefficient matrix for the second pass


(table continued from previous page)18 8 4 14 19 9 5 15 20 10

1 0.0617 0.0071 0.0016 0.0079 0.1156 0.0079 0.0095 0.0095 0.1669 0.000011 0.0000 0.0769 0.0080 0.0149 0.0000 0.1308 0.0095 0.0095 0.0000 0.215116 0.0000 0.0063 0.0070 0.0782 0.0313 0.0000 0.0000 0.1858 0.0094 0.00006 0.0000 0.0134 0.0858 0.0000 0.0000 0.0226 0.2130 0.0095 0.0000 0.00962 0.0139 0.0071 0.0240 0.0000 0.0000 0.0000 0.0095 0.0095 0.0271 0.009612 0.0000 0.0134 0.0000 0.0228 0.0000 0.0079 0.0000 0.0035 0.0000 0.038217 0.3128 0.0071 0.0080 0.0079 0.0382 0.0079 0.0095 0.0284 0.0094 0.00007 0.0000 0.3217 0.0000 0.0079 0.0078 0.0268 0.0367 0.0095 0.0000 0.00963 0.0139 0.0143 0.2345 0.0079 0.0157 0.0000 0.0284 0.0000 0.0094 0.000013 0.0139 0.0206 0.0000 0.2010 0.0000 0.0147 0.0000 0.0272 0.0000 0.009618 1.0000 0.0071 0.0080 0.0149 0.3066 0.0000 0.0000 0.0095 0.0271 0.00848 0.0000 1.0000 0.0230 0.0000 0.0000 0.2763 0.0178 0.0083 0.0000 0.02874 0.0000 0.0000 1.0000 0.0000 0.0157 0.0069 0.2130 0.0083 0.0082 0.000014 0.0000 0.0000 0.0000 1.0000 0.0069 0.0079 0.0000 0.2308 0.0094 0.000019 0.0000 0.0000 0.0000 0.0000 1.0000 0.0000 0.0000 0.0000 0.2012 0.00009 0.0000 0.0000 0.0000 0.0000 0.0000 1.0000 0.0000 0.0000 0.0000 0.21515 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 1.0000 0.0000 0.0082 0.008415 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 1.0000 0.0000 0.000020 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 1.0000 0.000010 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 1.0000

{Sinc}=

the incremental load added to each bolt in the secondpass

[Sf] =the final bolt load after interaction in the first andsecond passes

However, since [A]{Si} equals the bolt load in each bolt at the end of the first pass, thisequation can be simplified to

where {SP1} is the bolt load in each bolt at the end of the first pass.

Once the [B] matrix is found experimentally, the following matrix equation can be solvedfor the required increment in the second pass to give any final bolt load.

What follows is an example of how the coefficients are calculated and a description ofexperimental results.


6.1Determining Two-Pass Interaction Coefficients

Hypothetical interaction data for a two-pass procedure are given in Table 6. In the first passall the bolts are tightened to 850 mstrain, but interaction results in final values of 450, 550,700, and 850 mstrain in bolts 1, 2, 3, and 4, respectively, as shown. There are no blankentries in the second-pass data matrix because all the bolts are tight at the end of the firstpass and receive interaction in the second pass.

The two-pass matrix procedure is based on increments. Therefore the increments used inthe second pass must be calculated. Bolt 1 ends the first pass at 450 mstrain. At the start ofthe second pass bolt 1 is incremented by 400 to 850. This results in bolt 2 dropping to 400.Bolt 2 is incremented by 450 to 850 mstrain during the second pass. Similarly, bolt 3 isincremented 275 and bolt 3 is incremented by 375.


TABLE 6 Hypothetical Interaction Data for a Two-Pass,Four-Bolt SystemaFinal values at end of first pass

Bolt 1 Bolt 2 Bolt 3 Bolt 4450 550 700 850

Second pass interaction dataBolt 1 Bolt 2 Bolt 3 Bolt 4

Bolt 1 850 400 650 600Bolt 2 700 850 575 550Bolt 3 650 725 850 475Bolt 4 550 650 700 850aUsed in text to illustrate calculation of two-passinteraction coefficients.

The change in bolt load is divided by the initialload of the bolt being tightened. For the [B]matrix coefficients, the denominator becomes theincrement added to the bolt instead of the initialload used to determine the [A] matrix.


Using this information the [B] matrix can becalculated as follows:

For bolt 1:

For bolt 2:

And so on for bolts 3 and 4.

Correct calculation of the [B] matrix can beverified by substituting the values into Eq. (4).The final values obtained in the first pass and the


increments used in the second pass should, alongwith the correct [B] matrix, correctly predict thefinal bolt stress values. The completesubstitution of the hypothetical data into Eq. (4)results in the matrix equation


Figure 14Data for a two-

pass procedure (16 in. flange, no gasket). First pass to 15,000 psi stress; second pass to 25,000 psi.

6.2Experimental Data

Figure 14 shows interaction data for a two-pass procedure. In the first passeach bolt is tightened to 500 mstrain. In the second pass each bolt istightened to 850 mstrain. The bolting efficiencies are 0.39 and 0.81,respectively. This data is for a 16 in. ANSI pipe flange without a gasket.

Figure 15 shows data for a two-pass procedure using the two-pass matrixtechnique. The first-pass target was 500 mstrain and was obtained with good


accuracy by using Eqs. (1) and (2). The second-pass target of 1000 mstrainwas obtained by using Eq. (5). This test was done on a 16 in. flange withouta gasket. The bolting efficiencies are seen to be nearly perfect.

Figure 16 shows the results of using the two-pass matrix technique on aspiral-wound gasket. The target strain is 500 and 850 mstrain for the firstand second passes, respectively. The bolting efficiencies are 0.60 and 0.88,respectively. Although not as good as the results shown in Figure 15, thesevalues compare favorably with those of a non-two-pass matrix procedure ona spiral-wound gasket with much lower bolting efficiencies of 0.26 and0.53.

6.3Discussion of Results

The two-pass matrix procedure worked very well for a rigid system withouta gasket. It did not work as well with a flexible spiral-wound gasket. Thethicker, more flexible spiral-wound gasket has interaction coefficients thatare not as stable as those for a rigid system. However, the boltingefficiencies for a two-pass matrix procedure were much higher than for anyother spiral-wound bolt-up procedure after two passes.

The nonlinear bolted flangegasket interactions are approximated with asystem of linear equations. Performing a series of tests and updating thecoefficients after each test is equivalent to experimentally iterating to thecorrect nonlinear solution with a series of linear approximations of theelastic interactions. If the spiral-wound gasket has less constant interaction


Figure 15Final bolt stress for a 16 in. flange (without a gasket) using two-pass matrix technique. Second pass target is 30,000 psi stress.


Figure 16Final bolt stress for a 16 in. flange with a spiral-wound gasket using two-pass matrix technique.

Second pass target is 25,000 psi stress.


coefficients, improvements can be expected withmore iterations (i.e., more tests). The predictionshould become better as the correct solution isapproached and the coefficients become moreaccurate.

7Applications

Any critical joint is a good candidate for a matrixbolt-up procedure. A bolted connection with aspecific bolt-up procedure or one that requiresbolt load monitoring or stud tensioningequipment and/or receives additional engineeringattention for whatever reason (i.e., history ofleakage and/or critical nature of the joint) is agood candidate for a one-pass matrix bolt-upprocedure.

For new equipment, it makes sense to requirethat a series of tests be done in the shop


immediately after fabrication to determine the[A] matrix. Critical flanges should be purchasedwith their interaction data supplied.

It may be difficult to run a series of tests onoperating equipment. However, most of the datacan be taken during the normal bolt-upprocedure. The [A] matrix can be determinedwithout disrupting the existing procedure. Thealgorithm can then be used during the nextboltup.

A modified two-pass matrix procedure could beused as an alternative approach in the field. Datafrom the first pass could be used to predictoverbolting required for the second pass.

Data specific to each application must bedeveloped. However, this does not prevent amatrix procedure from being used on ordinaryflanges that are torqued. The accuracy fortorqued flanges will never by ±2%. However, byvirtue of torquing, the operator is accepting


tremendous scatter in bolt-to-bolt loading. Ageneric torquing procedure could be developed,for example, for all 10 in. Class 300 flanges. Thematrix method would give less scatter in fewerpasses.

References

1. 1995 ASME Boiler and Pressure VesselCode, Section II, Materials, Part DProperties,The American Society of Mechanical Engineers,New York, 1995.

2. Bibel, G. D., Experimental and analyticalstudy of elastic interaction in a pipe flange,International Conference on Pressure VesselTechnology, Duesseldorf, Germany, 1992.

3. Bibel, G. D., and R. Ezell, Bolted FlangeAssembly: Preliminary Elastic Interaction Dataand Improved Bolt-up Procedures, WeldingResearch Council Bulletin 408, May 1996.

4. Bickford, J. H., K. Hayashi, A. T. Chang, and


J. R. Winter, A Preliminary Evaluation of theElevated Behavior of a Bolted FlangedConnection, Welding Research Council Bulletin341, February 1989.

5. Goddard, D. L., and G. D. Bibel, Achieving aselected load distribution in the bolted joint of acylinder head of highly variable stiffness andcontact geometry, SAE Tech. Paper Series940693.

6. Bazergui, A., and L. Marchand, PVRCMilestone Gasket TestsFirst Results, WeldingResearch Council Bulletin 292, February 1984.

7. Van Campen, D. H., A systematic bolt-tightening procedure for reactor vessel flanges,University of Technology, Delft, TheNetherlands, 1969.

8. Czernik, D. E., and F. L. Misczk, A newtechnique to measure real time static anddynamic gasket stresses, SAE InternationalCongress and Exposition, Detroit, MI, Feb. 25,


1991.

9. Ezell, R., The analytical and experimentalstudy of elastic interaction in a bolted flangedgasket joint, Master's Thesis, University ofAkron, December 1992.


25Hydraulically Powered Wrenches

George A. SturdevantFastorq Bolting Systems, Inc., Houston, Texas

1Introduction

Hydraulic wrenches are relatively new in thebusiness of power tools. The first commerciallyavailable models came to market in 1971. Theneed for greater power and control to tighten andloosen threaded fasteners of ever greater size andstrength has driven the development of thesetools. Worker safety and the advent of stricterenvironmental regulation of fugitive emissionsare also of great concern. The replacement ofsledgehammers and striking wrenches, hand


wrenches, or impact wrenches with hydraulicallypowered wrenches reduces the hazards of boltingwork and improves the integrity of bolted joints.A wide variety of hydraulically poweredwrenches has evolved over the last 25 years. Inthis chapter I discuss the major types of wrenchesand the power units required to drive them.

2Hydraulic Wrench Types

The four basic types of hydraulic wrenches arelow profile, square drive ratchet, ratcheting boxwrench, and bar tooling. All of these typesconsist of the same basic components, as shownin Figures 1ad: hydraulic cylinder (1), wrencharm (2), and a reaction device (3). A power unitto provide hydraulic fluid flow within acontrolled range of pressure is required tooperate all four types.

All four types of wrenches are often referred toas hydraulic torque wrenches. This is due to the


fact that the torque output of the wrench can beadjusted and controlled by controlling the fluidpressure delivered by the pump.

2.1Low Profile Wrenches

Low profile wrenches were the first type to bedeveloped. They are also the most durable andleast complicated. As illustrated in Figure 2, therod end of a hydraulic cylinder (1) is connectedto the arm or handle of a box end wrench (2), andthe rear clevis is anchored to a reaction device(3) that reacts against adjacent bolts in the samepattern. The centerline of the hydraulic


Figure 1Illustration of the basic components of a hydraulic

wrench configured for each of the four types:(a) low profile

(b) square drive


(c) ratcheting box

(d) bar tooling.


Figure 2Low profile wrench with cylinder at midstroke and forming

a 90° angle with the wrench.

cylinder forms a 90° angle with the centerline ofthe wrench arm with the cylinder at midstroke.This provides for maximum torque at cylindermidstroke for each hydraulic pressure setting.

A 12-point box end wrench design is normallyused on low profile hydraulic wrenches. Thisallows the wrench to be positioned on a hexagon


nut with a rotation of no more than 15°. In orderfor the hex nut to be turned one flat (60°), thewrench must be rotated an equal amount.

A typical low profile wrench is designed withtwo to four positions on the wrench head so thatthe nut will be rotated 30° per cylinder stroke inthe case of two positions, 20° per stroke withthree positions (Fig. 3), and 15° per stroke withfour positions. The nut must be rotated a total of60° so that the wrench can be rotated back to thestarting position.

The power of a hydraulic wrench is determinedby three factors: hydraulic fluid pressure in thecylinder, cylinder piston area, and wrench armlength. For instance, 100 psi (7 bar) fluidpressure in a cylinder with a piston area of 2 in.2(0.13 mm2) will deliver a force of 200 lb (890N) at the cylinder rod. When this force is appliedperpendicular to a point on the wrench arm 1 ft(30.48 cm) from the center of the bolt, 200 ft-lb(271 N.m) of torque is applied to the nut.


As illustrated in Figure 4,

Figure 3Typical 12-point low profile wrench with

three-position head.


Figure 4Torque equals force times distance.

where

T=torque in ft-lb (or N.m)F=force in lb (or N)D=distance in ft (or m)

A=angle between cylinder and wrench arm (when A = 90°,sin A = 1)

The speed of a hydraulic wrench is determined by the flowrate of hydraulic fluid provided by the power unit at the


fluid pressure required for each specific torque valuedivided by the volume of flow required to extend andretract the cylinder. I discuss this in more detail in Section3.

The advantages of the low profile style of wrench are

1. Durability. There are a few moving parts, and thecomponents are generally forged or cast in high strengthalloy steels.

2. Ability to fit into restricted spaces. The 12-pointwrench and the reaction devices are thinner than the nutheight in each size, and the outside diameter of the wrenchhead is the same as or smaller than standard box endstriking wrenches.

3. Safety. Reaction forces are contained in line with thecenterline of the cylinder. There is no tendency to twist orpull the wrench off the nut under load.

Disadvantages are

1. Limited applications. This style is generally used onlyon regular bolt patterns such as pipe flanges.

2. Manual handling requirements. Low profile wrenchesmust be manually removed from the nut and reset aftereach 60° turn of the nut.


Low profile wrenches are adjustable to fit on a number ofbolt patterns. For instance, one wrench may be used on aflange with twelve 1 7/8 in. bolts or one with forty-eight1 7/8 in. bolts simply by adjusting the position of thereaction device so the hydraulic cylinder at midstrokeremains in the same 90° angle orientation with respect tothe wrench head. (See Fig. 5.)


Figure 5One low profile wrench shown on two very different bolt patterns.


Figure 6Low profile wrench in (a) tightening (clockwise) position


(b) loosening (counterclockwise) position.

Low profile wrenches may be changed from tightening toloosening or vice versa by reversing the direction of thewrench on the bolt. (See Fig. 6.)

2.2Square Drive Wrenches

Square drive ratchet wrenches are the most common ofthe four types of hydraulic wrenches. The basic elements


of this style (Fig. 7) are a drive gear, square drive element,drive pawl(s), body, hydraulic cylinder, and reactiondevice. The first four elements form the wrench armcomponent. The centerline of the hydraulic cylinder mustbe positioned perpendicular to the drive pawl mechanismat midstroke.

Standard impact sockets are used with square driveratchets with compatible square drives. The wrenches aredesigned with a square drive size capable of transmittingthe maximum torque that the wrench will produce. Table1 lists some typical square drive sizes and their maximumtorque.

Ratchet wrenches all use drive wheels that have ratchetteeth. The number of teeth on the wheel divided by thenumber of drive pawl positions determines the minimumdegrees of


Figure 7Components of a typical square drive mechanism.

rotation of the drive wheel required for theratchet to function. For instance, a wrench with adrive wheel with 12 teeth and a single drive pawlposition (Fig. 8) must rotate at least 30° percylinder stroke in order for the drive pawl toengage a new tooth when the cylinder retracts.


(360° divided by 12 equals 30°)

In practice, ratchet wrench cylinders must have astroke that rotates the ratchet wheel more thanthe minimum number of degrees in order tocompensate for wind-up or flexing of the boltand/or wrench. It is important to minimize thedegrees of turn of the wheel per cylinder strokein order to maintain accuracy of torquedelivered. The use of multiple pawl positions hasthe effect of reducing the degrees of turn of thewheel per cylinder stroke without increasing thenumber of ratchet teeth. A ratchet with 12 teethdriven by two pawls in different positions (Fig.9) will require only 15° of rotation per cylinderstroke.

Fewer teeth on a ratchet wheel of a given sizewill allow the teeth to be larger and stronger.However, these larger teeth will require moredegrees of rotation per cylinder stroke. Thedesign compromise must be made between sizeand quantity of teeth, number of pawl positions,


and cylinder stroke.TABLE 1 Typical Square Drive Sizes forMaximum Torque Ranges

Maximum torqueSquare drive (in.) ft-lb N · m3/4 1,500 2,0341 3,500 4,7461 1/2 11,500 15,6002 1/2 40,000 54,2403 1/2 80,000 108,480


Figure 8Ratchet with 12 teeth and single pawl position.


Figure 9Ratchet with 12 teeth and two pawl positions.


Figure 10Multiple pawls engage simultaneously.

The number of drive pawls engaged with the


ratchet wheel on each cylinder stroke is also animportant consideration. Multiple pawls spacedequally around the wheel will share the load andbalance the forces (Fig. 10).

Anti-wind-up or anti-reverse pawls (Fig. 11) areoften used in hydraulic torque wrenches toprevent the ratchet wheel from rotating backwardas the cylinder is retracted. If the wheel doesrotate backward, the drive pawl(s) will notengage the next tooth and the ratchet will notfunction.

The square drive may be changed from one sideof the wrench to the other, as illustrated inFigure 12, to change from tightening toloosening position and vice versa.

Square drive ratchet wrenches also require areaction device in order to transmit torque. Insome cases, the cylinder is an integral part of thereaction device. The most common practice is tofix the reaction device to the rear of the hydraulic


cylinder or wrench body by means of a femalespline or polygonal reaction component attachedto a matching male shape on the cylinder or body,as shown in Figure 13.

Reaction devices may also be attached to awrench frame or body so that the distance fromthe center of the square drive element to thereaction point can be adjusted. This allows forthe use of a ''slave" socket on an adjacent bolt asthe reaction point (Fig. 14). This configuration issometimes called a blind flange adapter.

The advantages of a square drive ratchet wrenchare


Figure 11Anti-reverse pawl in position when cylinder is

extended.


Figure 12Typical square drive mechanism.


Figure 13Typical reaction device with spline.

1. Standard impact sockets allow one wrench tocover a wide range of bolt sizes.

2. The nut can be continuously rotated withoutresetting the wrench.

3. It will fit on a wide range of bolt patterns.

Disadvantages are

1. Space is required above the nut for socket andwrench height.

2. Reaction forces tend to twist the wrench and


2. Reaction forces tend to twist the wrench andlift the socket off the nut.

2.3Ratcheting Box Wrenches

Ratcheting box wrenches (Fig. 15) are the newestdevelopment in this field. This style is an effortto combine the advantages of the low profile boxwrench with those of the square drive ratchetwrench. The elements of this style are the sameas for a square drive ratchet wrench except thatthere is no square drive. The ratchet wheel of aratcheting box wrench has a hexagonal hole cutto the size of the nut it is intended to fit.

The ratchet wheel, drive pawl, and wrench armare integral parts of the same component in aratcheting box wrench. Therefore, to change nutsizes, the wrench must be partially taken apartand the component replaced.

The advantages of a ratcheting box wrench are

1. Reaction forces are contained "in line" as for a


low profile wrench.

2. Continuous rotation of the nut is possiblewithout resetting the wrench.

3. Low height allows use in restricted spaces.

Figure 14Slave socket used as a reaction device.


Figure 15Illustration of the components of a ratcheting

box wrench.

Disadvantages are

1. The wrench element must be changed for eachsize nut.

2. Component costs are much higher than forsquare drive sockets or low profile wrenches.

3. Reaction devices are limited to one stylereacting on an adjacent nut (Fig. 16).

4. Diameter and thickness of the head requiregreater clearance than the low profile.

2.4


Bar Tooling

Bar tooling consists of a solid or fabricated baror tubular steel bar through which a hydraulicwrench drives a nut-turning device on one endand reaction forces are absorbed on the otherend. The components of a bar tool are a bar(reaction device), hydraulic cylinder, and awrench arm. Typical bar tools are shown inFigures 17 and 18.

Figure 16Ratcheting box wrench acting on an adjacent nut.


Figure 17Typical bar tool for a socket head cap screw application:

side view.


Figure 18Typical bar tool on a turbine joint:

top view.

Bar tooling is very effective when a longextension is required to reach a fastener andavoid an obstruction. Irregular bolt patterns orchanges in elevation from nut to nut such asthose found on turbine and pump horizontaljoints are also good applications for bar tooling.

The wrench arm used on a bar tool may be asquare drive ratchet, a low profile wrench, or aratcheting box wrench. In all instances, thewrench arm drives a rotating shaft perpendicular


to the bar (reaction device).

The advantages of a bar tool are

1. The bar absorbs reaction forces and resists thetendency to twist when the torsional force isapplied an extended distance from the nut or bolthead.

2. Customized sockets and reaction devices areeasily fitted to a basic bar tool.

3. This style of wrench accommodates irregularbolt patterns and changes in elevation from boltto bolt.

Disadvantages are:

1. A wrench is usually designed for a limitedrange of applications.

2. Size and weight are greater than for the otherthree styles for a given power range.

3


Hydraulic Power and Control Units

Electric, pneumatic, and combustion enginedriven hydraulic pumps are commonly used toprovide fluid flow to hydraulically poweredwrenches. The type of power used is determinedby the job conditions and the power sourcesavailable. All of these types consist of a motor,hydraulic pump, fluid reservoir, directionalcontrol valve, pressure control valve, and remotecontrol pendant (Fig. 19). The pumps are usuallyrated at either 6000 psi (420 bar) or 10,000 psi(700 bar) maximum pressure.

3.1Electric Motor Driven Units

Electric motor driven hydraulic pumping andcontrol units are most frequently used indoors inlocations such as manufacturing plants andpower generating stations. Electric power outletsare readily available as either single-phase orthree-phase in most of these plants. The most


commonly used electrically driven power unitsare single-phase, 110 V machines withhorsepower ratings from 1 to 1.5. The controlcircuit for the directional control valve is usually12 or 24 V DC. Electrically driven pumps ofmore than 2 hp require three-phase power.


Figure 19Typical schematic of a hydraulic

power and control unit.

The advantages of electrically powered units are

1. The noise level is lower than with pneumaticor combustion engines.

2. Maintenance costs are lower than for othertypes.

3. Efficiency is higher than that of pneumaticunits.

Disadvantages are

1. Standard units do not meet explosion-proofrequirements in process plants.

2. There is greater risk of worker injury due toshock, especially in wet conditions.

3. Fluid flow rate is limited on single-phasemodels due to limited horsepower.


3.2Pneumatically Driven Units

Pneumatically driven pumps are used in chemicalplants, refineries, and other process industrieswhere electric sparks are not allowed. The twomost common types of pneumatic systems arethe rotating vane air motor and the reciprocatingor intensifier type pump.


Figure 20Typical pressure vs. flow curve for a vane air motor driven pump.

Rotating vane air motors are vulnerable to damage fromwet and dirty air. They can be protected to some degreewith the use of an air dryer, filter, and lubricator. The


maintenance cost will be greater than for other types in anycase. Vane air motors are not efficient, and therefore theair usage per horsepower output is very high (Fig. 20). Forinstance, a 3 hp vane air motor will require 6065 cfm(1.71.8 m3/min) air supply at 100 psi (7 bar) to operate atfull capacity. The hydraulic fluid flow rate at pressuresgreater than 1000 psi (70 bar) will be about 65 in.3/min (1L/min) for the most commonly used pumps of this type.

Reciprocating (intensifier) type pumps are less vulnerableto damage and considerably more efficient (Fig. 21).Efficiency can be improved by using a double-acting pump.A double-acting pump requiring 150 cfm (4.2 m3/min) airsupply at 100 psi (7 bar) will provide 700 in.3/min (11.5L/min) hydraulic fluid flow at 1000 psi (70 bar) and 500in.3/min (8.2 L/min)


Figure 21Typical pressure vs. flow curve for an intensifier type pump.


at 5000 psi (350 bar). By comparison, thereciprocating pump provides 10 times the flowrate as the vane air motor pump with an increasein air consumption of about 2.5 times.

The advantages of pneumatically driven pumpsare

1. They are spark-free in operation and thereforeexplosion-proof.

2. They are safe for operators, even in wetconditions.

Disadvantages are

1. They are not as efficient as electrically drivenpumps.

2. Their noise level is much higher than that ofelectrically driven pumps.

3.3


Combustion Engine Driven Units

Combustion engine driven pumps may beprovided in any horsepower range that is requiredbecause they require no energy input at the site.Maintenance requirements are much the same asfor a pickup truck. The flow rate at any givenpressure is limited only by the size andhorsepower of the engine.

The advantages of combustion engine pumps are

1. A wide range of horsepower and sizes areavailable.

2. They are independent of the availability ofcompressed air or electric sources.

Disadvantages are

1. Their cost of acquisition is high.

2. They are larger and heavier than pneumatic orelectric units.

3. They are not explosion-proof for use in


process plants.

4Selecting a Wrench and Power Unit

The criteria for choosing the right hydraulicwrench system concern torque, wrench type, andpower unit.

4.1Torque Requirements

Determine the torque requirements to tightenand/or loosen the bolts on the job. Select awrench with approximately 50% greater torquecapacity than the job requires. This will avoidrunning out of power at a critical point in thejob. Torque requirement may be estimated bybolt size, as shown in Table 2. This is only anestimate, as other factors such as bolt material,Kfactor, and required bolt load must also beconsidered.

4.2


Wrench Type

Choose the type of wrench that fits the job. Asquare drive ratchet wrench may work well on apipe flange with no obstructions beyond the endof the bolt. If obstructions do exist, a low profileor ratcheting box wrench will be a better choice.

Check to see if the dimensions of the wrench willfit in the space available. Use a sketch of thework area such as Figure 22.

4.3Power Unit Type

Choose a power unit that matches yourrequirements. If the job is in an enclosed roomwith 110 V power outlets, the 11 1/2 hp electricpump is the apparent selection. If onlycompressed air is available and the unit must beexplosion-proof, a pneumatic pump will berequired.


TABLE 2 Estimated Torque Requirements by Bolt SizeTorque

Stud diam (in.) Nut A/F (in.) ft-lb N·m1/2 7/8 56 765/8 1 1/16 111 1513/4 1 1/4 197 2677/8 1 7/16 318 4311 1 5/8 477 6471 1/8 1 13/16 700 9491 1/4 2 984 1,3341 3/8 2 3/16 1,335 1,8101 1/2 2 3/8 1,762 2,3891 5/8 2 9/16 2,278 3,0891 3/4 2 3/4 2,867 3,8881 7/8 2 15/16 3,559 4,8262 3 1/8 4,090 5,5462 1/4 3 1/2 6,308 8,5542 1/2 3 7/8 7,909 10,7252 3/4 4 1/4 10,639 14,4263 4 5/8 13,915 18,8693 1/4 5 17,807 24,1463 1/2 5 3/8 22,344 30,2983 3/4 5 3/4 27,627 37,4624 6 1/8 33,659 45,642


4 1/4 6 1/2 31,987 43,3744 1/2 6 7/8 38,222 51,8294 3/4 7 1/4 44,888 60,8685 7 5/8 52,594 71,3175 1/4 8 61,130 82,892Source: Based on ASTM A193, B7 Studs withASTM A194, 2H nuts, lubricated on all matingsurfaces with moly paste lubricant, and loaded to 50%of the minimum yield strength.

5Estimating Productivity

Hydraulically powered torque wrench systemproductivity is dependent on

1. The fluid flow rate of the pumping unit

2. The volume required to extend and retract thehydraulic cylinder in each specific wrench

3. The degrees of turn of the nut per stroke of thecylinder

For example, a system using a pump with a flow


rate of 65 in.3/min (1 L/min) and a wrench with acylinder of 2 in. (50.8 mm) bore and 2 in. (50.8mm) stroke will operate at a speed calculated asfollows.

The area of the cylinder piston (Fig. 23) is pR2 or3.14 × 1 [or 3.14 × (25.4 mm)2]. The stroke ofthe cylinder is 2 in. (50.8 mm); therefore thevolume of fluid required to extend the cylinder is2 × 3.14 = 6.28 in.3 [or 50.8 × 3.14 × (25.4)2 =0.103 L].


Figure 22


Work area sketch for a bolted joint.

Figure 23Schematic of hydraulic cylinder.


The volume required to extend the cylinder, 6.28in.3 (0.013 L), divided by the flow rate of thepump, 65 in.3/min (1 L/min), equals the timerequired to extend the cylinder, 0.0966 min or5.79 s. If the cylinder rod is 1 1/2 in. (38.1 mm)in diameter, then the area of the return side of thepiston is 3.14 1.76 = 1.38 in.2 (2025 1135 =890 mm2). The volume of fluid flow required toretract the cylinder is 2 × 1.38 = 2.76 in.3 (50.8× 890 = 0.045 L); 2.76/65 = 0.0425 min or 2.55s. The total cycle time for the hydraulic wrench is5.79 s plus 2.55 s = 8.34 s.

A wrench that rotates the nut 15° per stroke ofthe cylinder and requires 8.34 s per stroke willrequire 33.36 s to turn a nut one flat or 60°(8.34 × 60/15).

The speed of operation of a given hydraulicallypowered wrench can be increased only by using apump with a greater flow rate. Doubling the flow


rate at a specific pressure doubles the speed atthat pressure. It is important to examine thepump performance curve to determine the flowrate at the pressure required to achieve thedesired torque when estimating wrench speed.Determine the speed at which you want yourhydraulic wrench to operate before selecting apower unit.

6Accuracy of Torque Application

The primary factors affecting the accuracy of ahydraulically powered torque wrench are

1. Accuracy of the hydraulic pressure gauge. A10,000 psi (700 bar) gage with accuracy of ±1%of full scale is accurate to ±100 psi (7 bar) at anypoint on the scale. The gage is accurate to ±10%when reading 1000 psi (70 bar).

2. Degrees of turn of the nut per cylinder stroke.A wrench that turns the nut 30° per stroke is less


accurate than one that turns the nut 15° perstroke.

For example, a wrench arm of 12 in. (304.8 mm)rotating through an angle of 30° (Fig. 24) willhave an effective wrench arm length of 10.4 in.(264 mm) at the beginning and end of the stroke.This difference reduces accuracy by 14%.

3. Matching the wrench to the job. Select awrench with a torque range capacity to match thejob. For example, choose a wrench with1501500 ft-lb (2002000 N.m) capacity for a jobthat requires 1200 ft-lb (1600 N · m) of torque.The wrench and the pressure gage will beworking at 80% of capacity in this instance. A10,000 psi (700 bar) gage with ±1% of full-scale accuracy will be accurate to within 11/4%at this point on the scale. A 3000 ft-lb (4000 N ·m) capacity wrench operating with the same gageat the same 1200 ft-lb (1600 N · m) torque willhave gage accuracy of only ±2 1/2%. The gage inthis case will be indicating 4000 psi (280 bar).


Figure 24Illustration of effective wrench arm length.


The use of a hydraulic torque wrench is one of manymethods of tightening threaded fasteners. Accuracy oftorque application of ±35% is common for most modernhydraulic wrench designs. This is a great improvementover sledgehammers and striking wrenches, where thereis no control, or even pneumatic impact wrenches, wheretorque control is limited at best. Control of torque isonly part of the story. Threaded fasteners do the job forwhich they are intended when clamping force or preloadin the fastener clamps mating parts together. Therelationship of torque applied to a bolt and the resultingtension or preload is described by the mechanicalengineer's short formula,

where

T=

torque in inch-pounds (Newton-meters)K=nut factor

L=bolt load in pounds (newtons)

D


D=bolt diameter in inches (meters)

The nut factor (K) takes into consideration a largenumber of variables including the condition of themating threads, lubricant used, flange surface under thenut, and many others. A nut factor (K) can be establishedfor a specific set of operating conditions on a specificbolted joint by measuring torque input and resulting boltload. Bolt load can be measured by using (1) directtension indicators, (2) ultrasonic extensometers, or (3)micrometers. These devices and methods of inspectionare discussed in other chapters in this book. For thepurpose of relating torque to tension when usinghydraulic wrenches, we will consider K factor estimatesbased on data collected under specific conditions. Table3 includes a sampling of K factors.

K factors are sometimes published for specific lubricantswithout regard to other conditions of the bolted joint.These data are valid only when the assumption is madethat all other variables are constant. An example of therisks involved in this assumption is drawn from a projectI recently supervised.

The project consisted of replacing 1 3/8 in. studs of


ASTM A193 B7 material one at a time in a tower coverwith 144 studs. The purpose of the project was to seal aleak that hadTABLE 3 K Factors for Common Materials and ConditionsBolt material and conditions Estimated

K factor

1.New steel bolts and nuts, hardened steel washers under nuts, allmating surfaces coated with moly paste containing 70% solids; boltand nut temperature of 200250°F (93121°C)

0.080.10

2.New steel bolts and nuts, hardened steel washers under nuts, allmating surfaces coated with moly paste containing 70% solids; boltand nut temperature of 7090°F (2132°C)

0.160.18

3.New Zylan 1052 coated nuts and bolts with no washers and nolubrication 0.180.20

4.Used steel bolts and nuts with cleaned threads and no washers butlubricated on all mating surfaces with moly paste containing 70%solids

0.200.30

5.Used steel bolts and nuts with cleaned threads but with no washersand no lubrication 0.400.50


TABLE 4 Accuracy of Bolting MethodsBolting tool Torque accuracy (%)Preload accuracy (%)Hydraulic torque wrench ±35 ±2328Hand wrench with torque multiplier Unknown ±70150Pneumatic impact wrench Unknown 300 to +150Manual slugging wrench Unknown 48 to +50Hydraulic stud tensioners N/A ±20Dial or Click torque wrench ±5 ±6080

been detected during a fugitive emission inspection. Newstuds, nuts, and hardened steel washers were supplied toreplace the existing studs. Moly paste lubricant with 70%solids was applied to all surfaces. Direct tensionindicators (DTIs) were used to inspect for bolt tension.

A K factor of 0.18 was used to establish an estimatedtorque value of 1068 ft-lb (1448 N · m) to tighten thefirst stud to an estimated 40% of yield or 42,000 psi(2900 bar) stress. The direct tension indicator showedthat more than 100,000 lb (448,800 N) of tension orabout 80% of yield strength had been exerted on the studby using this torque setting. The difference between theestimate and actual results was due to the K factor. The250°F (121°C) temperature of the studs on this ''hotbolting" project had improved the lubricity of the


lubricant and lowered the K factor to 0.09. The boltedjoint would have been overtensioned if direct tensionindicators were not in use. In this case, the torque appliedwas lowered to 550 ft-lb (746 N · m) and DTIs confirmedthat the studs were tensioned to 40% of yield. Theseresults remind us to double-check our assumptions.

Accuracy is relative. Hydraulically powered wrenchesprovide much better control than hand tools or impacttools. Critical bolted joints may require inspection ofresidual bolt load rather than the measurement of torqueinput. Table 4 shows the relative accuracy of a number ofcommon bolt-tightening methods in terms of preloadscatter.

The accuracy of all bolting systems can be improved withthe use of better inspection methods. Direct tensionindicators and ultrasonic extensometers will both providepreload accuracy within ±2% even when very crudemethods of tightening are employed.

References

1. Shigley, J. E. and C. R. Mischke, Standard Handbookof Machine Design, McGraw-Hill, New York, 1977, p.23.23.


2. Bickford, J. H., An Introduction to the Design andBehavior of Bolted Joints, 2nd ed., Marcel Dekker, NewYork, 1990, p. 610.


26Hydraulic Stud Tensioning

William L. BiachBiach Industries, Inc., Cranford, New Jersey

1Introduction

There are two requirements that need to beserved when using a threaded fastener: (1)stretching the fastener and (2) maintaining thestretch. For most applications, the nut is used forboth purposes. Rotating the nut against thefastener thread uses the inclined plane of thethread for mechanical advantage to stretch thefastener. It is then left in place to maintain thestretch. In hydraulic stud tensioning, a hydrauliccylinder stretches the fastener; then a nut is used


to maintain the stretch. The standard process is asfollows (refer to Fig. 1):

1. The tensioner is attached to the fastener abovethe nut (usually 5075% of a fastener diameter).

2. Hydraulic pressure (600025,000 psi) isapplied to the tensioner, which pulls on thefastener and reacts on the flange.

3. This stretches the fastener, loads the flange,and compresses the gasket (if a gasket is beingused), leaving a space between the flange and thebottom of the nut.

4. The nut is then rotated under a no-loadcondition to take up the space.

5. The hydraulic pressure is released, and the nutholds the load. There is usually a small drop inload as the previously unloaded fastener and nutthreads deflects as they assume the load.

6. The tensioner is then removed.


Generally, more than one tensioner is used at atime, with a common hydraulic line. This ensuresequal, concurrent, symmetric loading of theflange. Tensioners have been used for studs fromas small as 1/4 in. with a capacity of 2500 lb upto 20 in. with a capacity of 15,000,000 lb.

2Appropriate Applications

2.1Large Fasteners

Tensioners are most commonly used on fasteners2 in. in diameter and up. In the range of 24 in.,they are an alternative to torquing. In this rangetensioners are smaller, less cumber-


Figure 1Tensioner cutaway.

some, and safer. Where fasteners are greater than4 in. in diameter, hydraulic tensioning is themethod of choice, as tensioning provides an eventightening plane that cannot be achieved withother methods.

Examples: Nuclear pressure vessels, large


autoclaves, large press tie rods

2.2Gasketed Flanges

The key to effective closure of a gasket flange isuniformity of gasket pressure. Multipletensioners provide simultaneous loading atseveral point around a flange. It is common totighten half the studs at one time in refineryapplications, which provides extremely evengasket loading.

Examples: Pipe flanges, pressure vessel covers,manway covers

2.3Controlled Bolt Group Centroid

When a structure is being anchored with a boltgroup, the structure must be designed for themost conservative case for the loading of thebolts. If the bolts are torqued, the structure mustbe designed with the assumption that the bolts. It


the bolts are torqued, the structure must bedesigned with the assumption that the bolts willbe might tighter on one side than on the other.This puts the centroid near the edge and increasesthe structural requirements. Hydraulic tensioningensures bolting uniformity, which permits theassumption that the centroid will be centrallylocated. This reduces the demands on thestructure.

Example: Water tower designed for seismicconditions

2.4Materials Subject to Galling

Since tensioning doesn't require sliding surfacesagainst each other under high load, galling isvirtually eliminated.

Examples: Stainless steel, Inconel

2.5Fasteners Operating Close to Yield


Axial stress and torsional stress are the primarycomponents to the total stress in a fastenerduring tightening. Torsional stress becomes asignificant component at high loads because the


frictional force between the threads is high. Therefore a studmay yield while being torqued to an axial stress of only 75%of yield. As there is no torsional stress with hydraulictensioning, this problem is eliminated. Although tensioningintroduces the residual loss of tensioners, the loss may bereduced to 5% with the proper procedures and equipment.

Examples: Aerospace fasteners

2.6Space Constraints

When fasteners are placed where access is limited, readilyavailable tensioning tools may be used.

Examples: Box joints, bearings

2.7Limited Reaction Surfaces

When there are no reaction surfaces or the available surfacescannot be used, tensioning tools can be used as they require notorque.

Examples: Bridge suspension cables, underwater or outerspace applications

3Tensioner Sizing


Characteristic of the tensioning process is a partial loss ofload in the fastener when the hydraulic pressure is released andthe nut retains the load. For example, if the fastener is stressedto 50 ksi while the tensioner is pressurized, the residual stressin the fastener may be 40 ksi when the nut is tightened and thetensioner is removed. The following are contributing elementsto residual loss:

Nut and stud thread deformation

Change in effective grip length from where the tensioner isconnected to the nut

Nut expansion under load

Compression of the flange material under the nut when itassumes load

Residual loss can be anywhere from 5 to 50% of the fastenertension. However, under controlled circumstances, the loss isrepeatable and predictable. Therefore, even when thecharacteristic loss of the application is high, the loading is stillvery uniform.

Since thread deformation accounts for at least 90% of thetension drop in the fastener, addressing this issue is the mosteffective way of reducing the loss. While the hydraulictensioner has the fastener under load, the threads of thefastener above the nut are carrying the load, whereas the


threads of the nut and the complementary threads of thefastener are carrying no load. When the nut takes over fromthe tensioner, the threads deform, permitting the stud tocontract, which reduces the elongation strain. The amount ofthread deflection is dependent on the design of the fastener(thread form, material, etc.) and the load. The loss in strain isessentially this deflection divided by the length of the portionof the fastener being loaded (effective length).

Example: Assume that the threads of a 3 in. diameter studunder a tension load of 325,000 lb would deflect 0.005 in.The joint length is 20 in., and we're seeking 285,000 lb ofload in the stud. The calculation is as follows.


Therefore, in order to achieve 285,000 lbresidual load under these circumstances, wewould need a tensioner capable of producing14% more load to account for residual loss.

Factors affecting residual loss include.

Nut seating torque. By torquing the nutsufficiently while the hydraulic tensioner ispulling on the stud, the threads can be


deformed before the tensioner is released,reducing the load loss.

Coatings. Thick coatings such as paint andgalvanized bolts crush when the nut takes overthe load. This permits further loss of studload.

Lubricants. Lubricants permit more "pre"deformation of the nut threads during nutrundown, thereby reducing the residual loss.This also makes the residual load moresensitive to the amount of nut seating torqueapplied. This can cause increased variation inthe final loading.


Figure 2Tensioner design factor vs. joint length

ratio for different nut seat torques.


Figure 2 can be used as a guideline to determinethe tensioning factor. The x axis shows the ratioof the effective length of the joint to the studdiameter. Each curve is for a different nut seatingtorque. The y axis is the tensioner design factor.

Example: A 3 in. stud on a 20 in. joint length isrequired to maintain 285,000 lb load. We canput 150 ft-lb of nut seating torque on the nut.What capacity tensioner is required?

To determine the factor,

1. Find 7.92 on the x axis of Figure 2. Go up tothe curve for 50 ft-lb/in.

2. Go across to the y axis to determine the


tensioner design factor.

Factor = 1.15

Tensioner capacity = 285,000 lb × 1.15 =327,750 lb

Caution: This figure should only be used as aguideline. Accurate determination at the factorcan only be made through testing.


27The Multi-Jackbolt Tensioning System

Rolf H. SteinbockSuperbolt Inc., Carnegie, Pennsylvania

1Introduction to Multi-Jackbolt Tensioners

Multi-jackbolt tensioners (MJTs) wereintroduced in 1984. The basic idea was to createlarge tightening forces with little effort. Sincetheir introduction, multi-jackbolt tensioners havebeen developed for diverse applications. Thereare over 30 design variations of the basicconcept, covering the requirements of a numberof industries. The various designs, in turn, can bemade in thread sizes ranging from 3/4 through32 in. and in load capacities from 10,000 lb to


over 20 million lb. Two samples are shown inFigures 1 and 2.

2How Multi-Jackbolt Tensioners Work

The tensioner is threaded onto a new or existingbolt, stud, threaded rod, or threaded shaft. Themain thread positions the tensioner on the bolt orstud against a hardened washer and the load-bearing surface. Actual tensioning is done withhand wrenches or light and inexpensive air toolsto torque a series of jackbolts that encircle themain thread, as shown in Figure 3.

Note: Jackbolts are in compression. Incompression some steel alloys have a yieldstrength of 300,000 psi and are ductile up to500,000 psi compressive stress. Therefore it issafe to stress jackbolts manufactured fromselected materials that have been treated towithstand stresses up to 200,000 psi for use inmulti-jackbolt tensioners.


A total load (tension) of over 10,000 tons can becreated using a 3/4 in. torque wrench. Highmechanical advantage is attained by dividing therequired total preload among many jackbolts.Although small in diameter, the jackbolts aremultiple in number; the result is a high,combined thrust force for relatively littleindividual torque.

3Basics of the MJT System

The multi-jackbolt tensioning system depends onthe relationship of load vs. torque of itsjackbolts. A low friction factor is generallydesirable to obtain a large force from littletorque.


Figure 1Flange with many small MJTs.

The general formula to calculate the torque to obtain a certain amount offorce for jackbolts is

where

F =preload (lb)

T =torque (lb-in.)


Figure 2An 18 in. MJT on a 5000 ton hydraulic press.


Figure 3The MJT system.

r =effective end of jackbolt bearing radius (in.)


R =thread pitch radius (in.)P =pitch (in.)

cos b=0.866 for 60° threads

m =friction factor (the same for jackbolt end andthread)

The "effective end" of jackbolt bearing radius r isassumed to be 0.7 times the nominal radius of thejackbolt end.

For many material combinations the friction factor µon jackbolts for graphite-based lubricants isapproximately 0.13.

The friction factor µ for molybdenum disulfidebased lubricants is approximately 0.055.

Note: To obtain high accuracy, friction factors forMJTs must be established for each combination ofjackbolt material, tensioner body material, andlubricant.

4Mechanical Advantage of the MJT System


The mechanical advantage of MJT torquing overdirect torquing is the ratio of the direct torquedivided by the jackbolt torque. The overallmechanical advantage of torquing MJTs is dependenton several factors:


1. Number of jackbolts

2. Main thread pitch diameter

3. Jackbolt thread pitch diameter

4. Torque advantage of jackbolts over tension bolts due to differentbearing radii (»37% with graphite lube)

The basic formula for the mechanical advantage of MJTs over directtorquing is

where

MA=overall mechanical advantage

Tj =jackbolt torque (see Section 3)T =main thread torquen =number of jackboltsD =main thread pitch dia.d =jackbolt pitch dia.

A =torque advantage of jackbolts over tension bolts (with graphitelube A = 1.37)

Note: A varies with lubricants of different friction factors.

A=

1.37 for hex nut graphite lube, jackboltgraphite lube

A2.75 for hex nut graphite lube, jackbolt "moly"


=lube

Example 1:

1"-8 main thread with pitch diameter D = 0.9188in.

Eight 5/16 in. jackbolts with pitchdiameter

d = 0.2854in.

Torque advantage with graphitelube: A = 1.37

Mechanical advantage »35.

Example 2:


Twenty-four 3/4 in. jackbolts with pitchdiameter

d = 0.7094in.

Torque advantage with graphite lube: A = 1.37


Example 3:



Forty-eight 1 in. jackbolts with pitchdiameter

d = 0.9459in.

Torque advantage with molybdenumdisulfide lube: A = 2.75



TABLE 1 Load-Torque Relationships for Hex Nuts and MJTs Lubricated with Graphite Oil PasteaAt 60,000 psi main bolt stress At max MJT capacity

Thread size (in.)Main

bolt load(lb)

Hexnut

torque(lb-ft)

MJT torque (lb-ft)Main

bolt load(lb)

Hexnut

torque(lb-ft)

MJTtorque(lb-ft)

Main boltstress(psi)

Mechanical advantage

1-8 33,060 480 18 64,829 941 36 117,656 262-8 159,000 4,379 69 349,458 9,624 152 131,870 633-8 379,200 15,374 154 764,255 31,002 310 120,926 1004-8 694,200 37,203 226 955,318 51,197 510 82,355 1655-8 1,104,000 73,545 427 1,343,185 89,479 520 74,621 1726-8 1,608,000 128,065 519 1,611,820 128,369 520 60,142 247All values calculated with Friction Factor = .130 (Graphite Paste).aFigure 4 illustrates these data in graph form.


5Types of Multi-Jackbolt Tensioners

5.1Bolt-Type MJTs

Assuming equal load capacity, bolt-type MJTscan generally be produced with a smallerfootprint than nut-type MJTs. Most bolt-typeMJT designs (Fig. 5) have a ring of jackbolts thatare almost touching the bolt shank. This makes itpossible to obtain preloads up to yield strength inthe minor diameter area of high strength boltshaving an outside diameter of the bolt head notlarger than 1 1/2 times the nominal boltdiameter. On some ultrahigh strength bolts(280,000 psi yield strength), the outer diameterof the bolt head may be up to 1.7 times thenominal bolt diameter, as shown in Figure 6.(This is less than the over-the-corner dimensions


of a hex head cap screw.)

Bolt-type MJTs are used where the bolt headmust fit into a counterbore hole as shown inFigures 7 and 8 or where the bolt sits close to awall as shown in Figure 9. Extremely closespacing even at ultrahigh preloads can beaccomplished with bolt-type MJTs.

5.2Nut-Type MJTs

Most standard nut-type MJTs have an outsidediameter approximately equal to the over-the-corner dimension of a heavy hex nut. Exceptionsare tensioners for steam turbines and otherapplications where studs are used and where thestud spacing is just above 1 1/2 times the studdiameter. Using alloy steel tensioner bodies andsuperalloy jackbolts, nut-type MJTs can be madeto fit into the limited space available,maintaining the tension requirements of 1000°Fsteam turbine service.


Nut-type MJTs can be designed for almost anytype of service requirements such as hightemperature, cryogenic, corrosive, abrasive, or acombination of requirements. Examples areshown in Figures 1012.

Caution! Properties of bolting materials canchange very rapidly at high temperatures. Somematerials can lose 50% of their strength withtemperature increases of only 50°F.

5.3Thrust Collar Type MJTs

Multi-jackbolt thrust collars were developed forapplications where it is impractical to use threadsto position the tensioners. Instead of the mainbolt thread, a retainer ring is used to transfer thejacking force from the MJT to the main bolt orstud. On small diameter applications the retainerring is usually a commercial snap ring as shownin Figures 13 and 14. On large diameterapplications the retainer ring is a machined split


ring as shown in Figure 15. The split ring isoften hinged and may have two or more splits.

The most common application for thrust collarsis in the steel industry. For large rolling mills,the rolls have a neck diameter of 2035 in. withbearings installed on a tapered seat on the rollneck. The bearings have to be tightened withlarge nuts that can be turned only by cranewrenching as shown in Figure 16.

Because the roll lathes that machine the large 60in. diameter, 50 ton heavy rolls have no ''leadscrew," they cannot produce a thread. It has beenthe practice to machine a rectangular groove intothe roll neck and insert a split-threaded ring intothe groove. The split-threaded ring then takes upthe thrust generated by the big nut when it isturned by the crane. A few steel mills havemultimillion dollar tightening machines to turnthe big nuts, but most mills tighten laboriouslyby crane wrenching.


Multi-jackbolt thrust collars do not need athreaded ring. The thrust can be taken up by asimple split ring. The thrust is not generated byturning the big nut but by a number of jackbolts,as shown in Figure 17. The jackbolts can betorqued by an impact wrench or a torque wrench.


handbook of bolts and bolted joints

Documents

joints bickford

joint behavior

gasketed joints

present handbook of

introduction volume

new york

joint design procedures

joint members