Testing 7-Wire Strand forPrestressed ConcreteThe State of the Art

H. Kent PrestonSenior ConsultantWiss, Janney, Elstner Associates, Inc.Princeton, New Jersey

Special equipment and some experi-ence in the details of its use are re-

quired to obtain the necessary degree ofaccuracy in testing 7-wire prestressedconcrete strand,

ASTM A416 is the Standard Specifi-cation for Uncoated 7-Wire Stress-Relieved Steel Strand for PrestressedConcrete. The Supplement to ASTMA416 covers Low-Relaxation 7-WireStrand.

AASHTO M203 is the same as ASTMA416.

Testing equipment and proceduresare covered by Section VII of ASTMA370. Definitions of terms common to7-wire strand and prestressed concreteare given at the end of this paper.

ASTM A416 addresses the following_properties for stress-relieved strand:

Section 3.1 Strand — The require-ment of "... a center wire tightly en-closed by six helically placed outer

wires .. ." means that the outer wiresshall be neither overpreformed nor un-derpreformed. See comments near theend of "Modulus of Elasticity" and"Definitions."

Section 5 Materials and Manufac-ture — This section does not requireaction by the testing agency. Subse-quent tests will show whether or notthe requirements of Section 5 havebeen met.

Section 6.1 Breaking Strength—Minimum breaking strengths are tabu-lated in ASTM A416. Special problemsin testing strand are discussed later inthis paper under "Gripping a 7-WireStrand."

Section 6.2 Yield Strength — If aload-elongation curve is plotted, theyield strength can be read directly fromit. If not, the yield strength can be ob-tained by the test described in Section6.2.2 using equipment of the type dis-

134

cussed under "Modulus of Elasticity."Section 6.3 Elongation — In order to

reach the required 3' percent elonga-tion, it is almost always necessary to usegrips of the type described under"Gripping a 7-Wire Strand."

It is dangerous to leave sensitive ex-tensometer equipment on the stranduntil a 3' percent elongation has beenreached. A premature failure due to awire weld or a gripping problem couldruin the extensometer. Since strandsgenerally have an elongation of 4½ per-cent or more, perfect accuracy is not re-quired for this measurement. One typeof device that can be used is shown inFig. 8.

Section 6.4 — This section calls for aretest if the specimen fails in the gripwithout meeting the required ultimatestrength and/or elongation. Somelaboratories interpret this to mean thatone retest is required and that the strandis rejected if the retest has a prematurefailure in the grip. Section 6.4 actuallymeans that all tests which fail prema-turely in the grips without meeting therequired ultimate and elongation are in-valid.

Section 7.2 — The requirement thatthe center wire be at least so muchlarger than any outside wire is very im-portant. If a proper relation is not main-tained, the outer wires can form a pipearound the center wire without bearingon it and the center wire will slip andnot carry its share of the load.

Section 7.3 Permissible Variations inDiameter — These limitations were es-tablished many years ago in collabora-tion with the producers of strand an-chors and gripping devices. A strand of agiven nominal diameter meeting thesetolerances should work in a gripspecified for the same nominal diame-ter. Section 7.4 permits any size up to0.75 in. diameter as a special case.

Section 8.3 — If one or more wires flyout of position when cut and cannot bereplaced by hand, that wire is under-preformed. This condition may be

SynopsisPresents the state-of-the-art of

testing procedures for stress-relievedand low-relaxation 7-wire strand.

Specific information is given onmaterial properties, gripping devices,determining the modulus of elasticity,testing equipment, and short-term re-laxation test.

Definitions common to 7-wire strandand prestressed concrete are in-cluded.

Discussions of some of the testingmethods are given together with sug-gested procedures.

spotty or for the full length of the pack-age of strand. If the loose wires can betwisted back into place and held bytaping at the end, the strand should haveits normal properties except for a prob-able foot at the beginning of the load-elongation curve.

Section 8.4 — Oil and grease preventproper bond to the concrete and an oilycenter wire will often slide through theouter wires and not carry its share ofload. Many oils and greases become acidicwith the passage of time and invite cor-rosion.

Light rust improves bond and is notdetrimental in typical applications. A pitlarge enough to be seen without a mag-nifying glass is a stress raiser. A strandwith a pitted wire can fail at 30,000 cy-cles of repeated load when a strand withno pits will carry 2,000,000 cycles with-out failure.

ASTM A416 Supplement -Low-Relaxation Strand

The requirements for low-relaxationstrand are the same as those for stress-relieved strand with the following twoexceptions:

PCI JOURNAIJMay-June 1985 135

Table 1. Thin Line Splice for 270K Strand.

Strandsize, Catalog Color

Spliceddiameter,

Splicelength,

Ratedstrength, Splice

Standard carton

Weight, Units,in. number code in. in, lbs sets lbs in.

3/8 TLS 1120 Yellow 0.615 52 23000 1@3, 26–b4 55 257116 TLS 1121 White 0.718 60 31000 1@3, 2@4 35 10112 TLS 1122 Red 0.854 70 41300 2@3, 1@4 55 10

Section S5 Yield Strength — Theyield strength must be at least 90 per-cent of the specified minimum breakingstrength instead of the 85 percent re-quired for stress-relieved strand.

Sections S2, S3, S4, and S6 — Low-relaxation strand must have a stress lossthat does not exceed the specifiedamount when loaded to the specifiedload for the specified time under thespecified conditions listed in these sec-tions.

Very special equipment is required toconduct the relaxation test required bythese sections. The strand being testedmust be held at exactly constant length,and the load measuring device mustread load changes that are very small incomparison with the total load on thestrand.

Temperature must be constant at68°F for the duration of the test. Thespecial equipment required for this testis generally available only in thelaboratories of fabricators of low-relaxation strand.

Since it is not practical to conductlong-term relaxation tests on every 20tons of strand, as required by Section 9for all other tests, Section 56.1 permitsuse of data from tests on similarly pro-cessed strand. Some fabricators of strandhave had long-term tests in their labo-ratory witnessed and certified by repre-sentatives of a commercial testing labo-ratory.

A 30-minute relaxation test having thesole purpose of separating stress-relieved strand and low-relaxation

strand is described under "Short-TermRelaxation Test."

GRIPPING A7-WIRE STRAND

The configuration and metallurgicalproperties of a 7-wire strand are suchthat premature failure of the strand inthe gripping devices occurs when thestrand is tested using grips that are nor-mally used for testing steel bars. A typi-cal grip designed for a round steel barwill bear on only a small part of theouter surface of some of the outsidewires.

The resulting load per tooth fromthe grip is excessive and the wires fail inshear at a load below their actual tensilecapacity.

The problem is increased because thewire, with a strength in excess of270,000 psi, is more notch sensitive thanmost steel bars. The occurrence of pre-mature breaks has been so prevalentthat ASTM A416 now includes Section6.4 which makes the test invalid if faiI-ure occurs in a grip before the specifiedultimate tensile strength and elongationare reached.

The true properties of a 7-wire strandcan be measured only if an efficientgripping procedure is used. ASTM re-quirements are met if the specifiedminimum ultimate load and elongationare reached prior to failure. It is prefer-able to use equipment that will developthe full ultimate strength and elongation

136

LiPITCH LENGTHϗ

•

CENTER MARK

- - -y^^-- -- - - ^- THIN LAYER CF TAPE

tҟ SUBSETI.D. TAG

Fig. 1. Installing Thin Line Splice.

Fig. 2. Installing Thin Line Splice

of the strand being tested. This meansthat a reasonable number of failuresshould occur in the clear between thegrips.

All of the reusable wedge grips cur-rently being used in casting yards causeshear failure of wires in the grip beforethe required ultimate load andlor elon-gation has been reached.

None of the post-tensioned anchorsavailable will consistently reach the re-quired 3.5 percent elongation beforefailure.

Section VII of ASTM A370 includes a

section on gripping devices. It includesa discussion of the use of cushioningmaterial such as lead foil, aluminum foil,carborundum cloth, brass shims, etc.,between the strand and the teeth in thetesting machine grip. These methodshave been successful where special longjaws with fine teeth have been used.

Four methods of gripping strandwhich have given good results are dis-cussed in the following paragraphs. TheSand Grip yields about 50 percent clearbreaks. The PLP, the Tinius Olsen, andthe aluminum insert method yield 90

PCI JOURNALMay -June 1985ҟ 137

24 PLP GRIPҟMIN. CLEAR STRAND 28

5 ., °V" GRIPS IN TESTING MACHINE

t .. +ҟI5°ҟ 1524"EXTENSOMETER

24 PLP GRIP

1 5"t .

I° ±

L-- 3--------------

MINIMU M LENGTH OF TEST SAMPLE =5'-4

USE TESTING MACHINE "v "GRIPS FOR 3/4" DIA TO I" DIA BARS.

THESE DETAILS APPLY TO 3/8" AND 7/I6" DIA. STRANDS.

Fig. 3. Placement of Thin Line Splice on test specimen.

percent or better clear breaks.Under "Specimen Preparation"

ASTM A370 says "Wire slippage may beminimized by fusing together the cutends of the specimen......" ASTMA416 requires the center wire to belarger than the outer wires by aspecified amount. This difference is re-quired to make sure each outer wirebears on the center wire and grips it asthere is no other means of applying andkeeping tension in the center wire dur-ing its service life as a tendon.

If the wires are properly dimensionedand the wires are not lubricated, itshould not be necessary to fuse the endsof the specimen. After all, the centerwire of an unbonded post-tensionedstrand is gripped only by the outer wiresof the strand and the ends are often cutoff with a carborundum wheel ratherthan with a torch.

The PLP GripThe PLP Grip is made from devices

for splicing 7-wire strand that are man-ufactured by the Preformed Line Prod-ucts Company, P.O. Box 91129, Cleve-land, Ohio 44101 and marketed as theThin Line Splice by Spancrete North-east Inc., 8 Cairn Street, Rochester, NewYork 14611.

Dimensions of the Thin Line Spliceare shown in Table 1 and details in Figs.1 and 2. The splice for 1 -in, diameter270 Grade strand is composed of twosets of three wires each plus one set offour wires and has an outside diameterof about Ifs in. Each set of wires is heli-cally preformed so that it will fit tightlyaround the strand for which it is de-signed.

The three or four wires in each set arecemented together and the inside sur-face of the set is coated with a grit whichkeeps the strand from slipping. Whenused to splice a ' -in, diameter 270Grade strand, the TLS 1122 Thin LineSplice will develop the full ultimatestrength and elongation of the strand.

Fig. 3 shows details of the assembly ofa strand and PLP Grips in a testingmachine. Best results are obtained usingthe longest strand "V" grips and finestteeth available. In the unlikely eventthat failure occurs at a grip because ofcoarse teeth or too short a grip, using apiece of 50 grit carborundum cloth be-tween the jaws and the PLP Grip withthe grit against the PLP Grip may cor-rect the problem. Fig. 4 shows a PLPGrip and strand assembly in a testingmachine and Fig. 5 shows the resultingclear break with pure tensile failures ofthe wires.

138

Fig. 4. Test assembly with Thin Line Splice in testing machine.

As shown in Table 1, the Thin LineSplice for ½-in. diameter 270 Gradestrand is 70 in. plus or minus 3/a in. long.Using a carhorundum disc, one suchsplice can he cut into three lengths tomake three PLP Grips. When used fortesting strand, PLP Grips have been re-moved and reused but the cost of thelabor involved more or less offsets thesaving achieved.

Table 1 does not include a Thin LineSplice for 0.60-in, diameter strand. Thecost of making one as a special itemwould probably be high.

Preformed Line Products Companysnakes dead end anchors for all sizes ofstrand. Their Number BG 2111 is for a%-in, diameter galvanized strand. The

straight portion of this is approximately54 in. long and could he used to maketwo PLP Grips each about 27 in. long.There are no reports of testing with thisbut, unless the inside diameter is toolarge, it should work for 0.60-in. strand.

The Sand Grip

Details of a Sand Grip are shown inFig. 6. These are U grips with no teethin the gripping faces. In the area wherethe strand is gripped it is covered with agritty mixture which prevents slippage.The grips are not patented and can befabricated by any reasonably wellequipped machine shop. Details can beobtained from major strand producers.

PCI JOURNALJMay-June 1985ҟ 139

Fig. 5. Pure tension failure between ThinLine Splice grips.

Two types of grit have been used:1. Ordinary sharp concrete sand

screened to remove any oversize parti-cles. The sand is mixed with enoughSAE-10 or SAE-20 oil to make a cohe-sive but not wet mixture. Ifwater is usedin place of oil, the mixture will alsofunction satisfactorily but it dries if al-lowed to stand and must be remixed.

2. 80 grit aluminum oxide and water.Use enough water to form a cohesive butnot wet mix.Sand Grip Testing Procedure:

(1) Install testing machine grips withflat file faces.

(2) Apply sand or grit mixture to en-tire surface of U grooves in SandGrips.

(3) Install Sand Grips and strand intesting machine.

(4) Apply hydraulic pressure or othermeans to cause U grips to squeezestrand.

(5) Apply tension to strand.Note that a little of the grit mixture

may fall from the grips during the test.Therefore, consider the need to protectany measuring devices attached to thestrand.

The Tinius Olsen Grip

The Tinius Olsen Testing MachineCompany, Inc., P.O. Box 429, WillowGrove, PA 19090 has developed and fab-ricates the SLS grip for testing 7-wirestrand (see Figs. 7 and 16).

The SLS grips are toothless U gripsthat have a 10 in. length of contact withthe strand. They are designed so thatpressure on the strand is very lightwhere it enters the grip and becomesheavier further into the grip after someof the tension has been transferred fromthe strand to the grip.

Since the gripping faces of the U areinclined to wear with the repeated usethey get in a production laboratory, theU's are 10 in. long relatively small steelinserts that can be removed and re-placed as necessary, Although the strand

140

3" 3"

IҟIҟIҟ̂ 7/32ҟ1

-SEEҟ ILD STEELTABLE

CLAMP PLATES TOGETHERWITH SHIM BETWEEN THEMAND BORE HOLE SHOWN.

STRANDD1A. BORE SHIM

5/32 9/64 .',020

3/I6 ., 11/64" ".020

1/4" I%4..

3/8 23,,..

7/16 27/64.. ".025

1/2 , 3%64., .025

9/I6 3%4'•

.600 19/32" .025

2 I/2'

7"

Fig. 6. Details of Sand Grips.

may be gripped directly in the U's with-out any special preparation, it has beenfound that the number of clear breaksdecreases as the grips wear and the useof aluminum oxide grit improves theperformance of worn grips.

The Aluminum Insert

Aluminum insert angles have beenused successfully in several ways. Twoprocedures have been used in machinesthat have tapered mechanical or wedgegrips:

1. A thin layer of grease is applied tothe ends of the test piece followed by

the application of silicon carbide grit.2. A mixture of epoxy compound with

sand or grit is applied to the inner face ofthe aluminum insert.

A more recently installed procedureuses a machine with hydraulic gripswhich press the serrated teeth of the90-deg V groove machine grips into theouter faces of the aluminum angle. Thepressure also forces the strand config-uration into the inner faces of thealuminum angles. With this procedureclear breaks are experienced all thetime.

The testing machine being used toobtain these results is a 750 KN Avery

PCI JOURNAL Way-June 1985ҟ 141

Denison 7155. The aluminum angles are12 x 12 x 1.4 mm to 1.6 mm thick and±120 mm long (% x ' x 0.055 in. to0.063 x 43/1 in.). They are not reusable.

MODULUS OFELASTICITY

ASTM A416 does not require a mod-ulus of elasticity (MOE) test, but data onits load vs elongation properties are re-

quired almost every time a 7-wire strandis installed in a prestressed concretemember. Most specifications governingthe fabrication of prestressed concretemembers require the load indicated bythe tensioning jack and the computedstrand elongation to agree within 5 per-cent or work must be halted until thatdegree of agreement is achieved.

Although a possible error of 5 percentin initial tension may seem rather largefrom the designer's viewpoint, it re-

Fig. 7. Tinius Olsen SLS Grips, new liners for grips and strand that broke in clearbetween grips.

142

quires a considerable degree of accuracyat the job site and in the laboratory. Sev-eral things, including variations in pis-ton friction and gage accuracy, can causean error in the hydraulic jack's mea-surement of load in the strand. If thisamounts to 2.5 percent of the load, thenthe measured elongation must be within2.5 percent of the computed elongation.

Extremely good laboratory equipmentand procedures are needed if the MOEestablished in the laboratory is to alwaysbe within 2.5 percent of the actual MOEof the strand being tensioned at the jobsite. As indicated in the following dis-cussion, strand fabricating tolerancesare such that the actual MOE of twolengths of strand fabricated on the sameequipment to the same specification candiffer by as much as 2.4 percent. In ad-dition, it is difficult to attach a straingage to a member composed of sevenindividual wires so that it will measurethe MOE with complete accuracy acrossa relatively small gage length.

In order to reduce the effect of shoptolerances to a minimum, some strandfabricators furnish a typical curve whichis the average of the last 30 to 50 tests onthe specific size and grade being fur-nished. If the MOE test of the strandbeing shipped is within 2.5 percent ofthe typical, it is within the accuracy thatcan be expected and is considered ade-quate. Tensioning to the typical curveeliminates the large differences that cancome from testing a strand that happensto have a maximum MOE and then get-ting a strand with a minimum MOE atthe job site.

ASTM A416 requires a minimumyield strength of 85 percent of the ulti-mate for stress-relieved and 90 percentof ultimate for low-relaxation strand.This is defined as the load at which thestrand reaches 1.0 percent elongationand the straight line MOE does not ex-tend to these values. The load-elongation curve furnished by the strandfabricator should be consulted for ten-sions within 10 percent of the specified

yield strength.The MOE of a strand is affected by the

total area of the seven wires and, to alesser degree, by the pitch or lay of thesix outside wires.

ASTM A416 permits a pitch of any-thing from 12 to 16 diameters but re-quires that whatever pitch is used beconstant for the full length of the strand.A strand with a long pitch has a largerMOE than a strand with a short pitch.

ASTM A416 gives a nominal area foreach strand but permits a considerabletolerance with respect to actual area aslong as the specified minimum ultimateload is met, Because of different rodproperties and wire drawing proce-dures, there is a very appreciable differ-ence between the actual area of thestrand with the smallest area on themarket in the United States and that ofthe strand with the largest area.

When wire is drawn for a prestressedconcrete strand, the shop practice of thestrand manufacturer will specify a di-ameter, but it must also permit some de-gree of tolerance. The wires of a viz-in.diameter strand have a diameter ofabout 0.167 in. The minimum reason-able tolerance is ±0.001 in. This varia-tion in diameter of 0.001 in, represents achange in area, and MOE, of 1.2 per-cent.

The diameter of the bole in the lastdie of the wire drawing operation de-termines the diameter of the finishedwire. Since the diameter of this hole in-creases with wear as wire is drawnthrough it, it is standard practice to starta new die with a hole slightly under sizeand continue to use the die until thehole is producing wire that meets themaximum permissible oversize toler-ance. The result is a deliberately createdvariation in area, and MOE, of-_* 1.2 per-cent from the nominal.

It is essential that a standard proce-dure for computing MOE be establishedand adhered to. A complete series oftests on a strand includes measuring thediameter of each wire to he sure that the

PCI JOURNALIMay-June 1985ҟ 143

required difference between center andoutside wires is met. Some laboratoriescompute the actual strand area fromthese diameters and use this area inconjunction with the load-elongationcurve to compute MOE. Otherlaboratories use the nominal area fromASTM A416 in conjunction with theload-elongation curve to compute MOE.

In a number of instances, an excessivediscrepancy between jack load reading

— —FRAM CLAMPED-ҟE cL ҟTOt*ҟSTRAND AT TOP ONLY

TECHNICIAN READINGSCALE SHOULD BELOOKING THROUGHSHATTER PROOFGLASS

NN

y^SCALE CLAMPEDTO STRAND

READ CHANGE INELONGATION ASBOTTOM EDGE OF

aeҟFRAME MOVESACROSS SCALE

Fig. 8. Device for measuring total elonga-tion in 24 in. at failure.

and computed elongation has caused awork stoppage finally traced to an errorin MOE. The laboratory reported aMOE which they computed using actualarea. The field engineer, not knowingthis, used the laboratory's MOE and thenominal area of the strand with a resul-tant error in computed elongation.

In order to eliminate the type of errorillustrated in the previous paragraph, itis strongly recommended that the areaused in its computation be given at thesame time the MOE of a strand is given.

Although the MOE of most 7-wirestress-relieved and low-relaxationstrands is a little above 28,500,000 psi, itis common practice for the design en-gineer to use 28,000,000 psi and thenominal area of the strand for structuraldesign calculations. The 28,000,000 psivalue should not be used when com-puting the elongation required to pullthe strand to specified tension at the jobsite.

Testing Equipment andProcedure

ASTM EI11 covers standard details oftesting for Young's modulus or MOE.The chief difference between conduct-ing a load-elongation or stress-strain teston a 7-wire strand and on a single wireor a steel bar is that of attaching the ex-tensometer to the six outside wires ofthe strand so that it does not slip or ro-tate and so that any effect of bending asload is applied is cancelled out.

Experience has shown that the exten-someters which are normally attached towires or steel bars for making auto-graphic curves, generally do not performproperly when attached to a 7-wirestrand. One laboratory reports improvedresults when masking tape is placed onthe strand at the points where the gagepoints contact the strand.

A procedure which gives excellent re-sults using parts that can be fabricatedby a good machine shop is the DoubleDial method devised by Howard J.

144

DIAMETERI.2 OF STRANDIr^l

IϗO.I" I 1Iϗ ,

DBO" SPLIT COLLET

[DIMENSIONS BEFORE SLITTING)DRILL FOR I/4"

^^CONE POINT25'^

I.0" I DRILL ANOP FOR I/423•' "OLLOW FLATT POINT 2.5

SET SCREWD.eD"

COLLET HOLDER

20"1 DRILL AND TAP FOR {/4"HOLLOW CONE POIN < DRILL FOR 1/4"

0.5"r > I 0TRj SET SCREW CONE POINT

L I'S PIVOT RING

^J DRILL ANU IAP FOR2 1/4" HOLLOW CONE 2 7••POINT SET

SCREW

I--1 0.75 ^ 1

LFl

L0" 1 10..

I 4 / I25" A / A

I/4" BOLT HOLES1 4/2 .. Y.D.ALUMINUM TUBE

SECTION A-A BASE RING

50 INCH EXTENSOMETER UPPER FIXED GAGE POINT

Fig. 9. Details of extensometer for Double Dial method of measuring elongation.

Godfrey who is now a consultant inPennington, New Jersey. The followingdiscussion and illustrations have beenfurnished by Mr. Godfrey.

THE 50-IN. EXTENSOMETERFOR PRESTRESSEDCONCRETE STRAND

The 50-in. extensometer consists of analuminum tube somewhat less than 50in. in length. A flexible fitting is at-tached to the upper end of the tube so

that the extensometer will hang verti-cally. The flexible fitting is fastened tothe strand by split collets which have aninternal diameter equal to the strand di-ameter.

The dimensions of the collets shownin Fig. 9 are those of the solid sleevebefore slitting with a milling cutter. Thecut allows the two halves of the splitcollet to be pressed against the strand bymeans of the 1'a in. hollow flat pointsetscrew which is held in a 1.3 in. out-side diameter (OD) steel ring.

The 1.3 in. OD ring is pivoted inside a

PCI JOURNAL/May-June 1985ҟ 145

2 in. OD steel ring by' in. hollow conepoint setscrews.

The 2.0 in. OD ring is pivoted in a 2.7in. OD steel bracket which is attached tothe top end of the aluminum tube. Thepivot axis of the 2.0 in. OD ring is at 90deg to the pivot axis of the 2.7 in. ODbracket. Only a limited amount of mo-tion is required of the upper fitting be-cause the extensometer is not attachedto the strand until an initial load hasbeen applied. However, the availablemovement allows the extensometer toadjust for any further straightening ofthe strand during testing.

The various parts of the upper flexiblefitting are shown in Fig. 9 and a photo-graph of the assembled fitting is pre-sented in Fig, 9a. The dimensions havebeen estimated and may be revised asrequired.

The parts for the lower gage point fit-

Fig. 9a. Top clamp for Double Dial method

ting are shown in Fig. 10 and consist of a6 in. OD aluminum disc and two steelbars which are machined to hold thesplit collets. One of the steel bars ispermanently fastened to the bottom ofthe aluminum disc and the second bar isattached to the permanent bar after thestrand is under an initial load.

Before placing the strand in the test-ing machine grips, it is threaded throughthe aluminum tube and the top fittingwithout the collets in place. The top endof the strand is then fixed in the testingmachine grips, after which the lowergage point fitting is threaded on to thestrand. The bottom end of the strand isthen placed in the testing machine gripsand the initial load applied.

The split collets are then placed in theupper fitting and attached to the strandby means of the setscrew. The lower fit-ting is then attached to the strand by asecond set of collets held in the two barclamp under the aluminum disc. The50-in, gage length is measured by a steeltape from the top edge of the upper col-lets to the top edge of the lower collets.

The dial gages reading to 0.001 in.should then be attached to the lowerend of the aluminum tube by an adjust-able aluminum clamp as shown in Fig.11. The aluminum clamp should be sodesigned that the dial stems areequidistant from the centerline of thestrand and are on the same axis throughthe center of the strand. Under theseconditions any rotation of the aluminumdisc or tube, or any change in the levelof the disc during testing will not resultin an incorrect measurement of theelongation.

In order to eliminate excessive lateralmovement of the bottom end of thealuminum tube due to accidental con-tact, a slit cork stopper should be in-serted into the lower end of the tube.Before slitting, a hole should be drilledon the vertical axis of the cork stopperslightly larger than the diameter of thestrand. A view of the stopper in place isshown in Fig. 11. This guide system

146

ALUMINUMDISC

r"0.4,.ҟ

4-DRILL 0.4"a TAP

14'DRILLa TA P

---5 ud^6n

I,Z„ ^O

3/$3 ../4

I'ҟ3'

3/4 ' 6"

516"DRILL

y4

1/4 DRILL B TAP rite

3/ҟ21/4„

50 INCH EXTENSOMETER

LOWER FIXED GAGE POINT

Fig. 10, Details of extensometer for Double Dial method of measuring elongation.

does not result in any friction betweenthe strand and the stopper.

Either 1 in. or 2 in. travel dial gagesmay be used for measuring the elonga-tion of the strand as the load is applied.The full range of 1 in. travel dial gages isequivalent to an elongation of 2.0 per-cent. In general, load-elongation testsare discontinued when the elongationhas exceeded 1.0 percent which is theextension for determining the yieldstrength according to ASTM A421.

The 50-in, extensometer is not de-signed to be left on the strand whenfracture takes place so it is necessary toremove the extensometer before testingthe strand to failure. A view of the ex-tensometer attached to a test specimen

is presented in Fig. 12 and a typicalload-elongation curve is shown in Fig.13.

Curves produced by the Double Dialmethod and a testing machine with aload scale that permits the reading in-crements to be determined with accu-racy have given very good results. If it isnot possible to get a 50-in, gage lengthin a machine, the equipment can be de-signed for a shorter length but some ac-curacy will be sacrificed.

The Double Dial test is run at a con-stant speed. As each increment of load isreached the machine operator calls"Mark" and the two assistants readingthe dials record the dial readings at thatinstant. At the conclusion of the test, a

PCI JOURNAUMay-June 1985 147

Fig. 11. Dial gages and base plate forҟFig. 12. 50-in, Double Dial extensometerDouble Dial method.ҟ in use.

curve is plotted using the load incre-ments and the average of the two dialreadings. Experience has shown thatthis method gives more consistent andaccurate results than any other methodthat requires attaching the gage points tothe strand.

Experience has shown that the exten-someters which are normally attached towires or steel bars for making auto-graphic curves do not perform properlywhen attached to a 7-wire strand. After aconsiderable development effort TiniusOlsen Testing Machine Company, Inc.,succeeded in developing and now mar-kets an extensometer that can be at-tached to a strand to produce good andconsistent autographic curves (see Figs.14, 15, and 16). Fig. 17 shows a curvemade with Olsen equipment.

The Wallace No-Contact Extensome-ter eliminates the problem of slippage ofthe contact points by using opticalreading devices which do not touch the

strand. There are several models fromwhich to choose and they can be used onall types of specimens in addition tostrand. Users report good results.

In the United States, Wallace Exten-someters are available through:

Testing Machines Inc.400 Bayview AvenueAmityville, New York 11701(516) 842-5400In 1984, for an extensometer suitable

for use with 7-wire strand, TestingMachines Inc. quoted a price of$16,565.00 plus freight and installation.

Since the test for yield strength beginswith a 10 percent load on the strand,some laboratories begin their load-elongation curve at this point and thenproject it downward to zero load as aprojection of the line from 10 percentupward. If the strand is overpreformed,the bottom part of the curve is not astraight line and this procedure fails todetect the condition. When load is first

148

TENSILE TESTON 3/8"DIAMETER

LOK STRESS STRANDREEL NO. 639K-5572L

0.002 0.004 0.006 0008 0.010 0.012 0.014 0.016

ELONGATION, INCH PER INCH

8,000

6,000

4,000

2,000

00

24,000

22,000

20,000

8,000

16,000

% EXTENSION

14,000

InJ

4 12,000

OJ

ro,000

Fig. 13. Load-elongation curve plotted by Double Dial method.

applied to an overpreformed strand, asmall increase in load causes a large in-crease in elongation creating a nearlyflat line called a foot at the beginning ofthe curve. As the outer wires pull downagainst the center wire, the slope of the

curve increases until the wires are alltight and the remaining curve is astraight line to the point where yieldingbegins.

In almost all tendon installations, bothpretensioned and post-tensioned, each

PCI JOURNAL/May-June 1985ҟ 149

Fig. 14. Gage point on Tinius Olsen extensometer prior to clamping.

Fig. 15. Gage point on Tinius Olsen extensometer after clamping.

150

strand is pulled to a tension of 1000 lbsor more before elongation mea-surements are begun. Since this loadwill pull down all but very severelyoverpreformed outside wires, failure todetect the foot when plotting the load-elongation curve is not critical. It shouldbe noted that the low-relaxation processeliminates both the overpreformed andunderpreformed condition as a result ofthe permanent stretching of the wires.

Under "Specimen Preparation"ASTM A370 states "Wire slippage maybe minimized by fusing together the cutends of the specimen ..." ASTM A416requires the center wire to be largerthan the outer wires by a specifiedamount. This difference is required tomake sure each outer wire bears on thecenter wire and grips it as there is noother means of applying and keepingtension on the center wire during itsservice life as a tendon. If the wires areproperly dimensioned and are not lubri-cated it should not be necessary to fusethe ends of the specimen.

If the center wire slips during testingof an unfused strand, the cause shouldbe investigated. It could be oily wires,undersize center wire or over or under-preforming. The author does not ap-prove of fusing the cut ends of thespecimen. If one of the wires is in anyway out of place at the time of fusing,the resultant load-elongation curve willbe inaccurate.

SHORT-TERMRELAXATION TEST

The purpose of the 30-minute short-term relaxation test is to determinewhether the sample of 7-wire strandbeing tested is of stress-relieved grade(SR) or low-relaxation grade (LR). It isnot intended to provide data that couldbe used to project long-term lossesunder service life conditions.

Since the difference in relaxationlosses between SR and LR strand, mea-

Fig. 16. Tinius Olsen testing machine withSLS grips and extensometer for strand inplace.

sured as a percent of initial tension, in-creases as the initial tension is in-creased, the short-term test is conductedat a high initial tension (80 percent ofGUTS) to create as large a difference aspossible during the short term of thetest.

In the 1000 hour relaxation test re-quired by ASTM A416, the specimen isloaded to a prescribed tension and thenheld at constant length for the duration

PCI JOURNAUMay-June 1985ҟ 151

V)az0aza0

of the test. Loss due to relaxation is thedifference in load in the tendon be-tween the start of the test and its conclu-sion. The degree of accuracy with whichthe constant length must be maintainedand with which the small changes in a

v0ҟ1ҟ2ҟ3ELONGATION

TEST NO-: 1234 DATE: 1-19-84SIZE a GRADE : 1/2 f DIA. LOW-RELAXATIONCOIL NO: 54321

BREAKING LOAD IN POUNDS : 43, 3001.o % EXTENSION IN POUNDS ' 40,250ULTIMATE ELONGATION IN 24" : 4.17 %a

ELONGATION IN EO FEETAT 30,970 POUNDS=0.870°

ELONGATION IN INCHES PERINCH AT 30,970 POUNDS = 0.00725

Fig. 17. Load-elongation curve made withTinius Olsen extensometer.

large tensile force must be measured re-quire special equipment that is notavailable in otherwise fully equippedtesting laboratories. For example, eachof the relaxation units in the laboratoryof one strand producer uses a doubledial extensometer with a 192-in, gagelength and dial gages that read to 0.001in.

Basic Principles of Short-TermTest

The purpose of this test is to deter-mine within a reasonable degree of ac-curacy the loss due to relaxation of astrand loaded to a high tension for ashort period of time.

It will be necessary to record withconsiderable accuracy the:

1. Change in tension in the specimen.2. Change in length of the portion

being measured by the extensometer.3. Change in temperature.Also, it will be necessary to adjust the

results accordingly. Let the:Effective change in load = PEMeasured change in load = PMChange in load due to the measuredchange in length = PrThe effective change in load is:

PE _ ± PM + Pr.(1)

and the percentage of loss, P, is:

P = (Ps/P,) 100 (2)

where P, = initial tension.The change in load due to change in

length resulting from slippage in thegrips is:

P L = (A. E L 1 )/L (3)

in whichA = cross section area of strand (in.2)E = modulus of elasticity of strandL = gage length of extensometerL I = change in extensometer readingIt is not feasible to adjust the reading

for an appreciable change in air tem-perature because the effect on the strand

152

and other items of small mass will not hethe same as the effect on the testing ma-chine and other items of large mass. Amaximum permissible temperaturechange, say ±3°F, should be establishedand maintained. The specimen beingtested as well as the environment of thetesting machine must be maintainedwithin the limited temperature range fora considerable period of time prior tothe start of the test as well as for theduration of the test.

Procedure for Short-TermRelaxation Test

1. Make sure that strand, testing ma-chine, extensometer, etc., have been inthe same constant temperature en-vironment long enough for all parts of allitems to reach the same temperature.

2. It is not permitted to perform plasticbending or straightening on the sampleto be tested for relaxation behavior.

3. Place strand in testing machine andload to approximately 5 percent of spec-ified minimum ultimate strength. Usegripping devices with an efficiency thatwill provide a comfortable factor ofsafety above the load to be applied andthat will not allow the strand to slipduring the period of the relaxation test.Slippage and/or seating of wedges dur-ing the application of P, is not detri-mental to the test.

4. Attach extensometer to strand usinga long gage length, preferably 24 in. ormore.

5. Load at a uniform rate to P,. Loadingtime should take not less than 3 minutesand not more than 5 minutes and shouldbe approximately the same for all tests.

6. Maintain the load atP, for 1 minute.During this period make any desirableadjustments to extensometer equipmentsuch as setting dial gages to zero, etc.

7. At end of I minute hold, read andrecord extensometer reading, load instrand and temperature.

8. Maintain length constant for 30minutes.

9. At end of 30 minutes read and re-cord extensometer reading, load instrand and temperature.

10. Compute percentage of loss usingforegoing Egs.(1), (2), and (3),

Notes on Short-Term TestIt must be remembered that the func-

tion of the 30-minute test is solely todetermine whether or not the strandbeing tested is the low-relaxation grade.Test data used to make any sort of pre-diction with regard to long-term lossesmust be made in equipment capable ofmaking the 1000-hour test and should beof at least 100 hours duration.

The P, used in the short-term testshould be 80 percent of the GUTS of thespecimen being tested.

Since the change in load being re-corded is a very small percentage of thetotal load in the strand, it is importantthat the machine being used be able torecord small load increments with accu-racy.

Maintaining a constant length for the30-minute duration of the test is of ex-treme importance. This means pre-venting movement of the head of thetesting machine. By the time the speci-men has been loaded to P, and held atthat load for 1 minute, all seating ofgripping wedges or slippage within thegrips should have taken place.

The old style mechanical testing ma-chine is best suited for this purpose; aslong as the driving gears are not oper-ated, the head remains motionless andthe operator simply balances the beamto weigh the load as relaxation takesplace. Depending on their design, hy-draulic testing machines can hold a con-stant length with anything from a fair toa very poor degree of accuracy.

Theoretically the results obtainedusing Eqs. (1), (2), and (3) should be thesame whether the value of P L, is rela-tively small and the value of PM is rela-tively large or the reverse is true. Thedata from some tests suggest that the re-

PCI JOURNAL/May-June 1985ҟ 153

Metric (SI) ConversionFactors

1 ft=0.305m1 in. = 25.4 mmIin2=645.2mm'1 lb = 4.448 N1 ksi = 6.895 MPa1 psi = 0.006895 MPa1°C = (519) (°F-32)

suits are appreciably more accuratewhen P, is small indicating a very smallor zero change in length during the test.

Accuracy in measuring any change inlength is imperative. The maximum al-lowable relaxation for the 30-minuteperiod will be small; probably not morethan 1.5 percent of initial tension. For anextensometer with a gage length of 24in., an error of 0.003 in. in measurementof change in length represents '/a of Ipercent of the initial tension or about 17percent of the total allowable loss_

This could easily result in the rejec-tion of a specimen that actually had ac-ceptable properties. An accuracy of0.005 percent in measuring length isrecommended. Comments on the designand use of extensometers are presentedunder the heading of "Modulus of Elas-ticity."

If a mechanical machine is used andthe gears are not engaged during the30-minute test, there should be no mo-tion of the head and an extensometermay not he needed.

When a testing machine is used thatmay have even the slightest adjustmentof the head position during the test, theaccuracy of the extensometer is criticaland the 50-in. Double Dial system isrecommended. Experience with othertypes, even those that give good resultswhen plotting load-elongation curves,has not been satisfactory. This may be

because of a slight hysteresis at theclamps when the direction of motion isreversed or just stopped. There is nohysteresis in the spring loaded dialgages of the Double Dial system.

Several factors must be considered inestablishing the maximum allowableloss for the 30-minute test:

1. TemperatureThe 1000-hour test specified by

ASTM is conducted at a constant tem-perature of 68°F ± 3.5°F. Test datashow that a low-relaxation strand willhave more relaxation at 78°F than at68° F.

2. Consistency of Test Results(A) Although 1000-hour tests show

reasonably good agreement at 1000hours, the spread between differenttests is greater for the points plottedearly in the test.

(B) The load and length measuringdevices being used for this test by acommercial laboratory may be less accu-rate than the equipment used by astrand producer for the 1000-hour test.

3. Relaxation Properties of Strand(A) Tests run on 1000-hour equip-

ment, at 68°F at a tension of 80 percentof minimum specified ultimate show a30-minute relaxation of 0.50 percent forlow-relaxation strand and 2.6 percent forstress-relieved strand. There was notension on the stress-relieved strandsduring the stress-relieving process.

(B) Some strand fabricators makestress-relieved and low-relaxationstrand on the same equipment. Thus,there may be some tension on thestress-relieved strand during the stress-relieving process. The resulting productmay have relaxation properties betterthan those of stress-relieved but not asgood as those of low-relaxation strand.

It is suggested that low-relaxationstrand subjected to the foregoing 30-minute test be required to show a re-laxation loss of not more than 1.20 per-cent of the initial tension with an ad-justment for test temperatures over72°F. If the initial 30-minute test ex-

154

seeds 1.20 percent, two additional testsshould be made of strand from the samepack and both tests must he equal to orless than 1.20 percent for acceptance.

It is further suggested that any lab-oratory planning to conduct 30-minutetests obtain a length of low-relaxationstrand with known relaxation propertiesand conduct several trial runs to makesure that its equipment and procedureproduce correct results.

DEFINITIONSWire — A wire of the type used in

7-wire strand is a single, round, hightensile strength, high carbon steel unitof considerable length.

7-Wire Strand — A 7-wire strand iscomposed of six wires wrapped helicallyaround a center wire.

Cable — A cable is a tension carryingtendon composed of two or more paral-lel 7-wire strands (or wires) installed to-gether and functioning as one unit.

Nominal Area — The nominal area ofa strand as shown in ASTM A416 Table1 is the theoretical cross-sectional areacomputed by dividing the minimumspecified ultimate strength by the unitstrength of the grade specified. Thus,the nominal area of a ½-in. diameter 270Grade strand is 41,300/270,000 = 0.153sq in. Since the diameter tolerancesspecified in Section 7.3 of ASTM A416are the only items which limit actualarea, the actual area of the strandsupplied may differ appreciably fromthe nominal area.

Pitch or Lay — The pitch, specified inSection 3.1 of ASTM A416, is the dis-tance along the strand in which an out-side wire makes one complete turnaround the center wire.

Underpreformed — During thestanding operation the outside wires arepreformed so they will fit snugly aroundthe center wire. If the amount of pre-forming is insufficient they do not re-main in place when the strand is cut. Astrand that fails to meet the requirement

of ASTM A416 Section 8.3 is underpre-formed.

Overpreformed — If the outside wiresare too heavily preformed they do not flyopen when the strand is cut, but there isopen space between some of the outsidewires and the center wire. This causesan undesirable foot on the bottom of theload-elongation curve as discussedunder "Modulus of Elasticity."

Stress-Relieved — When a 7-wirestrand emerges from the stranding ma-chine which twists the six outer wiresaround the center wire, all of the wireshave high internal stresses as a result ofthe wire drawing and stranding process.This is called a "green" strand. Thegreen strand is passed through a heatingdevice which raises its temperature toabout 725°F for a few seconds before itis water cooled. While it is at the ele-vated temperature its yield strength isappreciably reduced, the high internalstresses cause local yielding and thestress-relieved strand emerges with farless internal stress than was present inthe green strand.

Low-Relaxation — Low-relaxationstrand is made from the same greenstrand as stress-relieved and subjectedto the same stress-relieving elevatedtemperature. While at the elevated tem-perature, it is tensioned to a high stresswhich causes a permanent elongation ofapproximately 1 percent. The resultingstrand has very little remaining capacityfor creep or relaxation. The 1.0 percentelongation also eliminates any overpre-formed or underpreformed condition.

ACKNOWLEDGMENTThe information presented herein is

based upon a study on "Testing 7-WireStrand for Prestressed Concrete," spon-sored by the Post-Tensioning Institute,Phoenix, Arizona. This paper has beenreviewed and recommended for publi-cation by the PT! Ad Hoc Committee onStrand Identification and Testing.

PCI JOURNALJMay-June 1985 155