designing electronic systems for emc - scitech 2011

8/20/2019 Designing Electronic Systems for EMC - Scitech 2011

http://slidepdf.com/reader/full/designing-electronic-systems-for-emc-scitech-2011 1/306

THE SCITECH SERIES ON ELECTROMAGNETIC COMPATIBILITYAlistair Duffy, PhD - Editor



DESIGNING ELECTRONIC

SYSTEM S FOR EMC



DESIGNING ELECTRONIC

SYSTEMS FOR

EMC

W illiam G. Duff

B

Scr

PUBLISUBLISHING INC



PUBLISHlNGrTINC.

Published by SciTech Publishing, Inc.

911 Paverstone Drive, Suite

B

Raleigh, NC 27615

(919) 847-2434,

fax

(919) 847-2568

scitechpublishing.com

Copyright ©2011

by

SciTech Publishing, Raleigh, NC. All rights reserved.

No part of this publication may be reproduced, stored in a retrieval system or transmitted in

any form

or by any

means, electronic, mechanical, photocopying, recording, scanning

or

other-

wise, except

as

permitted under Sections 107

or

108

of

the 1976 United Stated Copyright Act,

without either

the prior written permission of the Publisher, or authorization through payment

of the appropriate per-copy fee to the Copyright Clearance Center, 222 Rosewood Drive, Dan-

vers, MA

01923, (978) 750-8400, fax (978) 646-8600, or on the web at copyright.com. Requests to

the Publisher for permission should be addressed to the Publisher, SciTech Publishing , Inc., 911

Paverstone Drive, Suite B, Raleigh, NC 27615, (919) 847-2434, fax (919) 847-2568, or email edi-

tor@scitechpub. com.

The publisher and the author make no representations or warranties with respect to the accu-

racy

or

completeness

of the

contents

of

this work

and

specifically disclaim

all

warranties,

including without limitation w arranties

of fitness for a

particular purpose.

Editor: Dudley R. Kay

Editorial Assistant: Katie Janelle

Production Manager: Robert Lawless

Typesetting:

J. K.

Eckert Company,

Inc.

Cover Design: Brent Beckley

Printer: Sheridan Books, Inc., Chelsea,

MI

Printed in the United State s of America

10 9 8 7 6 5 4 3 2 1

ISBN: 978-1-891121-42-5

Library

of

Congress Cataloging-in-Publication Data

Duff, William G.

Designing electronic systems

for

EMC

/

William G.

Duff.

p. cm.

Includes bibliographical references.

ISBN 978-1-891121-42-5 (hardcover

:

alk. paper)

1. Electromagnetic compatibility. 2. Electromagnetic interference. I. Title.

TK7867.2.D84 2011

621.381-dc22

2011004765



This book is dedicated to a very special

lady

in my life—

my wife Sandi Her love and encouragem ent inspired and

motivated me to start writing this book Her patience and

understanding helped me to complete the task

No one knows better than an authors wife how much time

and effort is required to complete a book such as this.

Thank you,

Sandi,

for always being there when I needed

encouragement and support



Contents

Preface ix

Acknowledgment xi

Some EM C-Related and Metric Terms and Acronyms xiii

Comm on Terms and Abbreviations in EMC Literature xvii

Military EMI/EM C Standards xxiii

Chapter 1—Introduction to Electronic System Design for EM C 1

1.1 Effects of EM I 3

1.2 Sources of EM I 4

1.3 Modes of Coupling 5

1.4 Susceptible Equipm ents 6

1.5 EM C Design Consideration vs. System Life Cycle 6

1.5.1 System Definition Phase 8

1.5.2 System Design and Development 9

1.5.3 System Operation 10

1.6 Overview of Handbook 10

Suggested Readings: EM I/EMC 11

Chapter 2—B asic Terms and Definitions 13

2.1 Decibels 13

2.2 EMI Conducted Terminology 14

2.3 EMI Radiated Terminology 14

2.4 Representation of Signals in the Time and Frequency D o m a i n s . .. . 14

2.4.1 Fourier Series 15

2.4.2 Fourier Transform 16

2.4.3 Spectral Representation 16

2.5 Transients 18

2.5.1 Transient Sources 19

2.6 Narrowband Em issions 20

2.7 Broadband Emissions 21

2.7.1 Incoherent Broadband Em ission 22

2.8 Frequency and Wavelength 22

2.9 Units of Measure for EM I Signals 23

Suggested Readings: Basic Terms and Definitions 24

V l l



viii

DESIGNING ELECTRONIC SYSTEMS FOR EMC

Chapter 3—C omm unication Systems EMC 25

3.1 Comm unication System EMI Problems 26

3.2 EM I Interactions between Transm itters and Receivers 26

3.3 EMC Design of Comm unication Systems 28

3.4 Transm itter Em ission Characteristics 34

3.4.1 Fundamental Em issions 35

3.4.2 Transmitter Intermodulation 37

3.4.3 Harmonic Em ission Levels 39

3.5 Receiver Susceptibility Characteristics 40

3.5.1 Co-channel Interference 41

3.5.2 Receiver Adjacent-Signal Interference 42

3.5.3 Receiver Spurious Responses 47

3.6 Antenna Radiation Characteristics .48

3.6.1 Design Frequency and Polarization 50

3.6.2 Polarization Dependence 50

3.6.3 Nondesign Frequencies 50

3.7 Propagation Effects 51

3.8 Sample EM C Assessment 52

3.8.1 Transmitter Noise 52

3.8.2 Intermodulation 54

3.8.3 Ou t-of-Band EM I 57

3.9 Computer EMC Analysis 60

Suggested Readings: Com munication Systems EM C 60

Chapter 4—Electronic System Design for EM C 61

4.1 Basic Elements of EMI Problems 61

4.1.1 Sources of EM I 63

4.1.2 EM I Modes of Coupling 66

4.1.3 Susceptible Equipm ents 74

4.2 System-Level EM I Control 76

Suggested Readings: Electronic System Design for EM C 80

Chapter 5— Grounding for the Control of EM I 81

5.1 Definitions 82

5.2 Characteristics of Grounding Systems 83

5.2.1 Impedance Characteristics 83

5.2.2 Antenna Characteristics 90

5.3 Ground-Related Interference 91

5.4 Circuit, Equipment, and System Grounding. 94

5.4.1 Single-Point Grounding Scheme 96

5.4.2 Multipoint Grounding Scheme 97

5.4.3 Selection of a Grounding Scheme 98

5.5 Ground System Configurations 103



CONTENTS ix

5.6 EM I Control Devices and Techniques 109

Suggested Readings: Grounding 110

Chapter 6—Shielding Theory Materials and Protection

Techniques Ill

6.1 Field Theory I l l

6.2 Shielding Theory 113

6.2.1 Absorption Loss 116

6.2.2 Reflection Loss 117

6.2.3 Reflection Loss to Plane Waves 118

6.2.4 Reflection Loss to Electric and Magnetic Fields 119

6.2.5 Com posite Absorption and Reflection Loss 122

6.3 Shielding Materials 123

6.4 EM I Shield Com partments and Equipm ents 125

6.5 Shielding Integrity Protection 127

6.5.1 Integrity of Shielding Configurations 128

6.5.2 EM C Gaskets 142

6.5.3 EM C Sealants 155

6.5.4 Conductive Grease 157

Recomm ended Readings: EM I Shielding 158

Web Addresses for EM I Shielding 159

Chapter 7—B onding 161

7.1 Effects of Poor Bonds 161

7.2 Bond Equivalent Circuits, Resistance, and Impedance 162

7.3 Direct Bonds 163

7.3.1 Screws and Bolts 163

7.3.2 Soft Solder 164

7.3.3 Brazing 165

7.3.4 Welding 165

7.3.5 Cadw eld Joints 165

7.3.6 Conductive Adhesive , Caulk ing, and Grease 165

7.3.7 Bonding of Composite Materials and Conductive Plastics . . . 166

7.4 Indirect Bonds 167

7.4.1 Jumpers and Bond Straps 168

7.5 Corrosion and Its Control 169

7.5.1 Galvanic Corrosion 169

7.5.2 Electrolytic Corrosion 170

7.5.3 Finishes 170

7.5.4 Corrosion Protection 172

7.6 Equipm ent Bonding Practices 172

7.7 Summ ary of Bonding Principles 177

Suggested Readings: Bonding 178



x DESIGNING ELECTRONIC SYSTEMS FOR EMC

Chapter 8— Filters Ferrites Isolators and Transient

Suppressors 181

8.1 Filters 181

8.1.1 Pow er Line Filters 186

8.1.2 Signal Filters 196

8.2 Ferrites 199

8.3 Isolators 201

8.3.1 Isolation Transformers 204

8.3.2 Optical Isolators 211

8.4 Transient Suppressors 221

8.4.1 Crowbar Devices 224

8.4.2 Voltage-Clamping Devices 225

8.4.3 Hybrid Transient Suppressors 227

Suggested Readings: Filters, Ferrites, Isolators, and Transient

Suppressors 227

Web A ddresses for Companies that Provide EM I M itigation

Devices 227

Chapter 9—Cables and Connectors 229

9.1 Factors that Affect Shield Termination Guidelines 230

9.2 System Design for Interconnected Equipments 234

9.2.1 Cable Shield Termination Guidelines 237

9.2.2 Twisted Pairs to Reduce Magnetic Coupling 239

9.2.3 Shielded Cable Configurations 240

9.3 Connectors 240

9.3.1 Shield Termination Concepts 241

9.3.2 Connector Backshells 243

9.3.3 Termination of Individual Wire Shields 245

9.3.4 Filter-Pin Connectors 251

9.3.5 Coaxial Connectors 251

9.3.6 Summary of Connector Characteristics 252

9.3.7 Summary of EM I Control Techniques for Connectors 254

Chapter 10—Summ ary of EM I Control Techniques 257

Appendix A: Cable-to-Cable Coupling 267

Index 273

About the Author 277



Preface

Almost every aspect of modern life depends on the use of electronics.

Without electronics, the basic na ture of our society would be completely

different. The manner and efficiency in which modern life is conducted

depends on the ability to achieve and maintain electromagnetic compat-

ibility (EMC), which is a necessary condition for effective communica-

tion-electronic (CE) system performance. EMC is the ability of

electronic equipments and/or systems to function as intended in their

operational electromagnetic environment (EME) without adversely

affecting or being affected by other electronic equipments or systems.

Electromagnetic interference (EMI) is the culprit which does not

allow radio, TV, radar, navigation, and the myriad of communications-

electronic devices, apparatus and systems to operate compatibly in a

common EME. The EMI can result in a jammed radio, hear t pacemaker

failures, navigation errors and many other nuisance or catastrophic

events. In order for electronic equipments to operate compatibly, they

must share the electromagnetic spectrum without creating EMI or

reacting to EMI. The requirement for spectrum sharing has reached

international levels of concern and it must be dealt with in proportion

to the safety and economic impact involved.

The basic EMC requirement is to plan, specify and design electronic

circuits, equipments and systems tha t can be operated in their intended

EME without creating or being susceptible to EMI. To satisfy this

requirem ent, careful consideration must be given to a number of factors

that influence EMC. It is particularly necessary to consider major

sources of EMI, modes of coupling and points or conditions of suscepti-

bility.

There is much written material on EMI that is generally available in

trade journals, symposium records and other sources. In general, this

material provides a collection of miscellaneous subjects and topics that

do not interrelate very well. As a result, individuals that are seeking



XII


tutorial or how-to-do-it knowledge about EMC will find it very frustrat-

ing.

The prim ary purpo se of this book is to provide the rea de r w ith a tuto-

rial overview of the major factors that must be considered in designing

circuits, equipments and systems for EMC. This book emphasizes fun-

dam entals and provides information t ha t will help the reade r to under-

stand the rational that forms the basis for many of the EMC practices

and procedures.

—W illiam G. Duff



Acknowledgment

Electromagnetic Compatibility (EMC) is a difficult subject involving

many areas of technology. There is much written material on EMI/EMC

which is generally available in trade journals, symposium records,

reports rules, regulations and standards. However, this material repre-

sen ts a collection of miscellaneous subjects an d topics th at do not inter-

relate very well. As a result, a newcomer to the EMC discipline or

others already in the discipline who are seeking tutorial or how-to-do-it

knowledge will find it very frustrating. The primary purpose of this

book is to provide the reader with a tutorial overview of the major fac-

tors th at mu st be considered in designing circuits, equipm ents, and sys-

tems for EMC.

Over the years that electromagnetic interference (EMf) has been a

concern, many individuals have contributed to our knowledge on this

subject. I would like to acknowledge that the material presented in this

book represents the contributions of many of those individuals. One

individual that made a significant contribution to the field of EMC, in

general an d m e in particular, was Don W hite.

Don's company Interference Control Technology (ICT), was dedicated

to proving education in the field of EMC. ICT published two series of

handbooks and a magazine. ICT also provided a number of seminar

courses and computer software on EMC. For yea rs th e Don W hite

handbooks and seminar courses were the major source of information

on EMC.

Don encouraged me to write four of his handbooks and teach a num-

ber of his courses. Durin g the ye ars th at I worked w ith Don I found him

to be an enthusiastic hard worker. His enthusiasm was contagious and

I regarded him as my mentor. Before I started to write this book, I

requested Don's permission to use material from two of the books that I

wrote for ICT. Don gave me the permission that I requested along with

his blessings. I especially want to thank Don for his permission and his

blessings and I wish him well in his new endeavors.

—W illiam G. Duff

X l l l



Some EMC-Related and Metric

Terms and Acronyms

A ampere

AFCI arc-fault circuit in ter ru pt er

AM amp litude modulation

ANSA Am erican National Stan dar ds Assoc.

ASTM ASTM Int ern atio na l (formerly Am erican Society for Test-

ing and Materials)

AWG Am erican wire gage

BRH Bu reau of Radiological H ealth

C

3

I com mun ications, comm and, control, and Intelligence

CE conducted emission

cm cen time ters = 10~

2

meters

CM common mode

CMRR common-mode rejection rati o

CS conducted susceptibility (immun ity)

CSA Can adian Sta nd ard s Association

dB decibel

dB/dec dB per decade (am plitude slope)

DM differential mode

E

3

electromagn etic env ironm ental effects

EEC Eu rop ean Economic Comm unity, now EU

EED electroexplosive device

E-field electric field in Vim, ^V/m or dBV/m

EM electromagnetic

EMC electromagn etic comp atibility

EME electromagnetic environm ent

EM F electromagn etic fields

EMI electromagn etic interference

EM P electromagnetic pulse

EMV electromagn etic vuln erability

EN Europ ean norm (regulations)

ER P effective radiat ed power

ESD electrostatic discharge

X V



xvi DESIGNING ELECTRONIC SYSTEMS FOR EMC

EU

EW

FAA

FCC

FDA

FD M

FM

gauss

GFI

GHz

Gnd

HEM P

HER F

HERO

HER P

H F

H-field

HIRF

HPM

IE C

IEEE

IF

IR

ISM

ISO

IT E

kA

kHz

km

kV

LF

m

mA

mG

M F

MHz

micron

m il

m i

m m

mT

mV

European Union

electronic warfare

Fed eral Aviation Agency

Federal Communications Commission

Food and D rug Adm inistration

frequency division multiplex

frequency modulation

1(T

4

tesla

ground-fault interrupter

gigahertz = 10

9

hertz

ground

high-altitude electromagnetic pulse

hazards of EM radiation to fuels

haz ards of EM radiation to o rdnance

haz ard s of radiation to personnel

high frequency = 3-30 MHz

m agnetic field in A/m or dBA/m

high-intensity radiated fields

high power microwave (radiation)

International Electrotechnical Commission

Institute of Electrical and Electronics Engineers

intermediate frequency

infrared

industrial, scientific and medical

International Standards Organization

information technology equipment

kiloampere = 10

3

amperes

kilohertz = 10

3

hertz

kilometer = 10

3

mete rs = 0.621 m iles

kilovolt = 10

3

volts

low frequency = 30-30 0 kH z

meter = 39.37 inches

milliampere = 10-

3

amperes

milligauss = 10~

4

gauss = 10~~

7

tesla

medium frequency = 300 kH z-3 M Hz

megahertz = 10

6

hertz

10~

6

meters

10~

3

inches = 39.37 m icrons

mile = 1.609 km

millimeter = 10~

3

meters = 0.03937 inches

millitesla = 10~

3

tesla = 10 gau ss

millivolts = 10

3

volts



SOME EMC-RELATED AND METRIC TERMS AND ACRONYMS

XVII

NA RTE Na t i ona l As s oc i a t ion of R a d i o and T e l e c o m m u n i c a t io n s

E n g i n e e r s

N A S A N a t i o n a l A e r o n a u t i c s and S p a c e A d m i n i s t r a t i o n

NE C Na t i ona l E l ec t r i ca l Code

NF PA Na t i ona l F i r e P r o t ec t i on As s oc i a t ion

n F n a n o f a r a d = 1 0

9

f a r ad

n H n a n o h e n r y = 10~

9

h e n r y

NI L no i s e - i m mu ni t y l eve l of logic fam ilies)

N I S T N a t i o n a l I n s t i t u t e of S t a n d a r d s and Technology

n m n a n o m e t e r = 10~

9

m e t e r s = 10 A

n T n a n o t e s l a = 10~

9

t e s l a = 10~

5

g a u s s

P C p e r s o n a l c o m p u t e r

P C A p e r s o n a l c o m m u n i c a t io n s a s s i s t a n t

PC B pr i n t ed c i r cu it boa r d

pF p icofarad = 1CT

12

f a r ad

PLC pr o gr a m m abl e logic con t r o l le r

P L F pow er - l ine f req . = 5 0 - 4 0 0 Hz

pT picotes la = 1 0

1 2

t e s l a = 10~

8

g a u s s

R A O H A Z r a d i a t i o n h a z a r d s

R F r ad i o f r equency

R FI r ad i o -f r equency i n t e r f e r ence

S H F s upe r - h i gh f r equency = 3-3 0 GHz

SI s i gna l i n t eg r i t y

T t e s l a = 10

4

g a u s s

T C F tech nica l con s t ruc t io n file

T E M P E S T c o m p r o m i si n g e m a n a t i o n s

T H O t o t a l h a r m o n i c d i s to r t io n

TVI t e l ev i s ion in ter fere nce

T V S S t r a n s i e n t v o l ta g e s u r g e s u p p r e s s o r

| aH mi c r ohenr y = 10~

6

h e n r y

JXF mi c r o f a r ad = 10~

6

f a r ad

U H F u l t r a - h i g h f re q u e n cy = 0.3 -3 GHz

U L F u l t r a - low f r equency = 300 H z- 3 kHz

U P S u n i n t e r r u p t i b l e p o w e r s u p p l y

JO

sec mic rosecond

=

10~~

6

s e c ond

juT microtes la

= 10 ~

6

t e s l a

= 0 . 0 1

g a u s s

J V

m i c r ovo l t

=

1 0 ~

6

vol t

de voltage supply to circuits and PCBs

VHF very high frequency = 30-300 MHz

VLF very low frequency = 3-30 kHz

WGBCO waveguide beyond cutoff freq.)

ZSRG zero signal reference grid



Common Terms and Abbreviations

in EMC L iterature

Prefixes for

Multiples

10

12

10

9

10

6

10

3

10

2

10

10-

1

io -

2

IO -

3

io -

6

10-

9

i o -

1 2

te ra

giga

mega

kilo

hecto

deka

deci

centi

milli

micro

nano

pico

Decimal

T

G

M

k

h

da

d

c

m

M

n

P

Technical Terms

absolute abs

alternating current ac

Am erican wire gauge AWG

ampere A

amp ere per meter A/m

ampere-hour Ah

amp litude modu lation AM

amplitude probability

distribution APD

analo g to digita l A/D

analog-to-digital converter

ADC or

A/D

converter

anti-jamming AJ

arith m etic logic un it ALU

audio frequency AF

autom atic da ta processing ADP

auto m atic frequency control.. AFC

auto m atic gain control AGC

average avg

bandwidth BW

bin ary coded decimal BCD

bit b

bit-error ra te BER

bits per second bps

British therm al unit Btu

broadband BB

byte B

bytes per second Bps

centimeter-gram-second cgs

centra l processing un it CPU

cha ract ers per second cps

common-mode coupling CMC

common-mode rejection

ratio CMRR

complementary metal-oxide

semiconductor CMOS

continuo us wave CW

coulomb C

cubic cen time ter cm

3

X I X



DESIGNING ELECTRONIC SYSTEMS

FOR

EMC

decibel dB

decibel above 1 milliw att dBm

decibel above 1 volt dBV

decibel above 1 w at t dBW

degree Celsius °C

degree Fah renh eit °F

diameter dia

differential-mo de coupling.... DMC

digital mu ltim eter DMM

digital to analo g DA

digital voltm eter DVM

digital-to-analog converter

DAC or D/A con verter

diod e-tran sistor logic DTL

direct cur ren t de

double-pole, double-

throw DPDT

double sideb and DSB

double sideband

suppres sed carrier DSB-SC

dual in-line package DIP

electric field

E-field

electromagnetic

compatibility EMC

electromagnetic

interference EMI

electromagnetic pulse EM P

electrom otive force EM F

electron volt eV

electronic

countermeasures ECM

electrostatic discharge ESD

emitter-cou pled logic ECL

extremely high frequency EH F

extrem ely low frequency EL F

farad F

fast Fourier transform FFT

field inte nsit y FI

field inte nsit y me ter FIM

field-effect tra ns ist or FET

foot ft or

frequency freq

frequency division

multiplex FDM

frequency mo dulation FM

frequency shift keying FSK

gauss G

gram g

ground gnd

grou nd loop coupling GLC

ground support equipm ent GSE

hazards of electromagnetic

radiation to ordnance HERO

henry H

he rtz (cycles pe r second) Hz

high frequency H F

high-power transistor-to-

tran sisto r logic HTTL

high-speed complementary

metal-oxide

semiconductor HCMOS

high -thre shold logic HTL

hour hr

inch in or

inch per second ips

industrial, scientific, and

medical ISM

infrared IR

input/output I/O

inside dime nsion ID

instantaneous automat ic

gain control IAGC

insulated-gate field-effect

t rans is tor IGFET

integ rated circuit IC

interference-to-noise rat io I/N

inte rm edia te frequency IF

joule J

junction field-effect

t rans is tor JFET

kelvin K

kilogram kg

kilohertz kHz

kilovolt kV



COMMON TERMS AND ABBREVIATIONS IN EMC LITERATURE

XXI

kilowatt kW

kilowatt-hour kWh

lambert L

large-scale integration LSI

least significant bit LSB

length 1

length of cable) l

c

line impedance stabilization

network LISN

line of sight LOS

liter 1

local oscillator LO

low frequency LF

lower sideband LSB

lumen lm

lux lx

magnetic field H-field

master oscillator power

amplifier MOPA

maximum max

maxwell Mx

mean time between

failure MTBF

mean time to failure MTTF

mean time to repair MTTR

medium frequency 300 kHz

to 3 MHz) MF

metal-oxide semiconductor ...MOS

metal-oxide semiconductor

field-effect

transistor MOSFET

metal-oxide varistor MOV

meter m

microfarad JLIF

microhenry jiH

micron 10~

6

meter) \i

micro-ohm jxQ

microwave MW

mile mi

military specification...MIL-SPEC

military standard MIL-STD

milliamp mA

million instructions

per second MIPS

millisecond ms

millivolt mV

milliwatt mW

minimum min

minimum discernible

signal MDS

minute min

modulator-demodulator modem

most significant bit MSB

multi layer board MLB

multiplex, multiplexer mux

nanofarad nF

nanohenry nH

nanosecond ns

narrowband NB

negative neg

negative-positive-negative

transistor) NPN

negative-to-positive

junction) n-p

newton N

noise equivalent

power NE Por P

n

non-return to zero NRZ

N-type metal-oxide

semiconductor NMOS

nuclear electromagnetic

pulse NEMP

oersted Oe

ohm Q

ohm-centimeter Hem

ohms per square Q/sq

ounce oz

outside dimension OD

peak pk

peak-to-peak p-p

phase lock loop PLL

phase modulation PM

positive pos



XX11

DESIGNING ELECTRONIC SYSTEMS

FOR

EMC

positive-negative-positive

(transistor) pnp

positive-to-negative

(junction) p-n

poun d (sterling) £

pound per square

centimeter lb/cm

2

pound per squ are inch psi

power factor P F

printe d circuit board PCB

priva te bra nch exchange PBX

P-type metal-oxide

semiconductor PMOS

pulse per second pps

pulse repetition frequency PR F

pulse-amplitude

modulation PAM

pulse-code mo dulation PCM

pulse-duration

modulation PDM

pulse-width modu lation PWM

quasipeak QP

radiation haz ard RADHAZ

radio frequency R F

radio interference and field

intensity RI-FI

radio-frequency

interference RFI

rando m access memory RAM

receiver RX

reference ref

relative hum idity RH

resistance-inductance-

capacitance RLC

re tu rn to zero RTZ

revolutions per min ute rpm

roentgen R

root-mean-square rms

second s

sens itivity tim e control STC

shie lding effectiveness SE

sideband SB

Siemens S

signal-to-interference (ratio) S/I

signal-to-n oise (ratio) S/N

silicon contro lled rectifier SCR

single sideb and SSB

square meter m

2

standing-w ave ratio SWR

super high frequency SH F

supe r low frequency SLF

surface acou stic wave SAW

surface-m ount technology SMT

surface-mounted

component SMC

surface-m ounted device SMD

television TV

te m pe ra tu re coefficient TC

tesla T

tim e division mu ltiplex TDM

transistor-to-transistor

logic TTL

ultra high frequency

(360 MHz to 3 GHz) U H F

ultraviolet UV

very high frequency

(30 MH z to 300 MHz) V H F

very high-speed integrated

circuit VHSIC

very large-scale

integration VLSI

very low frequency

(3 to 30 kHz) VL F

volt V

volt m ete r VM

voltage standing wave

ratio VSWR

voltage-to-frequency

converter VFC

voltampere VA

volt-ohm m eter VOM

watt W

waveguide beyond

cuttoff WGBCO



COMMON TERMS AND ABBREVIATIONS IN

EMC

LITERATURE

X X l l l

weber Wb

words per min ute wpm

yard yd

Mathematical Functions

and Operators

absolute value abs

approximately equal «

argument arg

cosine cos

cosine (hyperbolic) cosh

cotangent cot

cota ngent (hyperbolic) coth

determinant det

dimension dim

exponential exp

imaginary im

inferior inf

limit lim

logarithm, common (base

10

) log

logarithm, Napierian (base

e

) In

sine sin

tangent tan

tan ge nt (hyperbolic) ta n h

Common V ariables in EMC

Equations

attenuation constant, absorption

factor a

Boltzmann's constan t K

capacitance (in farads) C

charge Q

coefficient of self-ind uctance L

conductance in mho G

conductivity, pro pagatio n

constant, leakage coefficient,

deviation a

current I

dielectric constant,

permittivity 8

frequency (in Hz) f

impedance Z

induced voltage E

indu ctance (in henry s) L

infinity ©o

length

(coil tu rn , ground

loop,

etc.) 1

length

in mill imeters l

m m

magnetic

suscep tibility %

magnetizing

force H

parasit ic

capacitance C

p

permeabil i ty

of free sp ace

JLI

0

permeabil i ty

of me dium

relat ive

to n

0

jn

r

phase

co ns tan t |3

radius

r

relat ive

perm ittivity e

r

resistance

(in ohm s) R

rise

time x

r

shield thickness d

t ime

t

t ime

constant , t ransm ission

factor

x

velocity, volume V

wavelength X



Military EMI/EMC Standards

The following military standards may be downloaded at www.jsc.mil.

• MIL-STD-449D

Radio Frequency Spectrum Characteristics, Measurement of

•

MIL-STD-461E

Requirements for the Control of Electromagnetic Interference

Emissions and Susceptibility

• MIL-STD-464

Electromagnetic Environmental Effects, Requirements for Systems

•

MIL-STD-469A

Radar Engineering Design Requirements; Electromagnetic

Compatibility

•

MIL-STD-1310G

Shipboard Bonding, Grounding and Other Techniques for EMC and

Safety

•

MIL-STD-1512

Electroexplosive Subsystems, Electrically Initiated, Design

Requirements and Test Methods

•

MIL-STD-1541A

Electromagnetic Compatibility Requirements for Space Systems

• MIL-STD-1542B

Electromagnetic Compatibility and Grounding Requirements for Space

System Facilities

•

MIL-STD-1605

Procedures for Conducting a Shipboard EMI Survey (Surface Ships)

• MIL-STD-1795A

Lig htnin g Protection of Aerospace Vehicles and H ard w are

xxv



xxvi DESIGNING ELECTRONIC SYSTEMS FOR EMC

• MIL-STD-1857

Grounding, Bonding, and Shielding Design Practices

•

MIL-A-17161D

Absorber, RF Rad iation (Microwave Absorbing Ma terial), Gen eral

Specification for

DoD Ad opted Non-Mi l itary S tanda rds

• ANSI C95.3-1979

Techniques and Instrumentation for Measurement of Potentially

Haza rdous Electromagnetic Radiation at Microwave Frequencies

•

ANSI N2.1-89

Warning Symbols—Radiation Symbol

•

IEEE 81-1

E art h Resistivity, Ground Impedance, and E art h Surface Poten tials of

a Ground System

•

IEEE C63.14

St an da rd Dictionary for Technologies of Electrom agnetic Com patibility

(EMC), Electromagnetic Pulse (EMP), and Electrostatic Discharge

(ESD)

•

IEEE C95.1-91

Safety Levels with Respect to Human Exposure to Radio Frequency

Electrom agnetic Fields, 300 kHz to 100 GHz

•

IEEE 299-1991

IEEE Standard for Measuring the Effectiveness of Electromagnetic

Shielding, Enclosures

•

SAE-ARP 1173

Test Procedures to Measure the R.F. Shielding Characteristics of EMI

Gaskets

• SAE-ARP 1972

Recommended Measu rem ent Practices and Procedures for EMC

Testing

•

SAE-J551-90

M easurem ent Practices and Procedures Recommended for EMC

Testing



Chapter 1

Introduction to Electronic

System D esign for EMC

Almost every aspect of modern life is significantly influenced by and

depends on the use of electronics. Without electronics, the basic natu re

of our society would e completely different. Electromagnetic compati-

bility (EMC) is a necessary condition for effective electronic system

performance. EMC is the ability of equipm ents and systems to function

as intended their operational electromagnetic environment without

adversely affecting the operation of, or being affected adversely by,

other equipments or systems. Thus, the manner and efficiency in

which modern life is conducted depends on the ability to achieve and

maintain EMC.

In order to permit efficient use of electronics, engineers, technicians,

and users responsible for the planning, design, development, installa-

tion, and operation of electronic systems must have a methodology for

achieving EMC. Techniques that permit them to identify, localize, and

define electromagnetic interference (EMI) problem areas before, rather

than after they waste time, effort, and dollars, must be available. More

timely and economical corrective m easures may then be taken.

The primary purpose of this book is to provide an understanding of

EMI problems and techniques for mitigating these problems. Careful

application of these techniques at appropriate stages in the system life

cycle will ensure EMC without either the wasteful expense of overengi-

neering or the uncertainties of underengineering.

EMI can occur in different levels ranging from the chip to ensem-

bles of systems, as shown in Fig. 1.1. The top level, Deployment of

Vehicles

and Plant Sites

applies to the situation where a large num-



INTRODUCTION TO ELECTRONIC SYSTEM DESIGN FOR

EM C

The Many Levels of EMI Manifestations

Large Area Deployment

Platform or Ensemble of Systems

I

Mission-Oriented System

Subsystem or

Collection of Equipments

Individual Box

or Equipment

Card Cage & Back Plane

or Mother Board

X

Printed-Circuit Board

X

Command and Control

Center, Battle Force, etc.

Aircraft, Spacecraft, Ship

Tank, Half Tread, B uilding

M I L

STD-

Fire-Control Radar,

4 6 4

Industrial Process Control

Components

Equipment Racks and

Consoles, Multiple Boxes MI L-

STD-

Receiver, Computer,

Display(s), Large Storage

4 6 1 E

Printed Wiring Distribution

+ PCBs

Cards + Traces + ICs

+ SMT + Edge Connectors

ICs, Caps, Inductors,

Resistors, F ilters, Ferrites,

FCC

RTCA

EU

IEC

Figure 1.1 The many levels of EMI manifestation.

ber of EMI sources and victims are deployed over a large area (e.g.; a

navy battle force, a military command and control center, a civilian

emergency force, etc.). The second highest level of EMI manifestation

is an Ensemble

o f

Systems

or

Vehicles

at a Site.

This level may apply to

a platform (such as a ship, an aircraft, a tank, a communications facil-

ity, etc.) containing a number of electronic equipments in a relatively

small area. The third level of EMI manifestation is labeled Mission

Oriented System. Examples of this level would be a fire control system

that includes a radar, computer, and associated missiles, or an indus-

trial control system in a process that is being monitored with sensors

th at provide information to a computer that controls the process. The

fourth level shown in Fig. 1.1 is labeled

Subsystem or

Collections

of

Equipments. Typical examples at this level are equipment racks, cabi-

nets, and/or consoles containing a number of individual equipments

connected together by signal and/or power cables. The primary

emphasis of this handbook is directed to this level. The next lower

level of EMI complexity is the

Individual Box or Equipment Level.

Examples include tran sm itters, receivers, medical instrum ents, mea-

suring instruments, etc. The sixth level from the top is the M other

Board or ackplane Assembly level. The layout and deployment of the

interconnecting wiring will affect EMI emissions and susceptibility.

Next to the bottom level is the Printed Circuit Board Level. The layout



EFFECTS

OF

EMI 3

of

components and traces on the printed circuit board will have a

major

impact on the EMI characteristics of the electronic package.

Finally,

th e

Component

level conta ins the indiv idual electronic compo-

nent s .

The EMI characteristics of the higher levels depend on the

selection of com ponen ts.

1 1 Effects of EMI

EMI may directly influence the performance of any electronic equip-

ment or system, and it can indirectly affect the overall accomplishment

of an operation or mission. Examples of direct influences of EMI on

system performance are false targets and missed targets in a radar

display system, wrong navigation data or landing system errors in an

aircraft, lost or garbled messages in a communications system, false

commands to a missile or electro-explosive device, or triggering a h eart

pacemaker demand-mode operation. Some resulting indirect effects

corresponding to the above include false alerts in an air-defense sys-

tem as a result of false targets, surprise enemy attacks as a result of

missed targets, aircraft mid-air collisions as a result of navigation

errors, aircraft crashes while landing because of altitude or glide-slope

errors, ineffective control of riots or fires because of lost or garbled

emergency fire or police communications, accidental launching of mis-

siles or detonation of explosives because of wrong electrical commands,

and the fainting, collapse, or even death of the person with a heart

pacemaker.

All of these effects, both direct and indirect, have happened as a

result of

EMI.

They can recur, and with the increase of

EMI

sources and

receptors every year, the situations will probably become more fre-

quent.

Figure 1.2 illustrates the three basic elements that must be consid-

ered in dealing with any EMI problem. These three basic elements of

EMI are discussed in the following sections.

Elements of EMI

Figure

1 2

Three basic elements of an

emitting-susceptibility

situation.



4 INTRODUCTION TO ELECTRONIC SYSTEM DESIG N FOR EMC

1 2

Sources

of

EMI

Any electrical, electromechanical, or electronic device is a potential

source of EMI.

In

general, EMI sources can be classified either as trans-

mitters (equipment whose primary function

is

to intentionally generate

and radiate electromagnetic signals) or incidental sources (equipments

that generate electromagnetic energy

as an

unintended by-product

in

the process of performing their primary function).

Transmitters generate electromagnetic energy in specific frequency

ranges. The spectrum chart shown in Fig. 1.3 illustrates various users

and identifies

the

specific frequency ranges

in

which they operate.

Fig

ure 1.3 also specifies the maximum power levels allowed for the trans-

mitter fundamental outputs.

Transmitters generate energy

not

only

in

their fundamental

or

intended frequency range, but also over a wide range of other frequen-

cies

on

both sides

of

the fundamental carrier, harmonics

of

the funda-

mental, and other undesired or spurious frequencies. These undesired

emissions result from spreading of the baseband transmitter modula-

tion spectrum, generation

of

harmonics

of

the fundamental

as a

result

of nonlinearities in the equipment output stages, and production of

broadband noise

in

the output stages.

Because of transmitter nonlinearities, signals from two or more

transmitters

can

heterodyne

in the

output stages

of one to

produce

additional signals at totally different frequencies. This is called trans-

mitter intermodulation. In designing a wireless system, all of these

10 kHz 100 kHz 1 MHz 10 MHz 100 MHz 1 GHz

Radio Frequency

10 GHz 100 GHz

Figure

1.3

US and Canadian frequency allocations and maximum effective

radiated powers.



MODES OF

COUPLING

5

t r an smi t t e r outputs must be carefully considered. They are explained

in

detail in Chapter 3 of this book.

Electrical,

electromechanical, and electronic equipments can be

potential sources of conducted and/or radiated

EMI.

Although the levels

associated

with these sources are usually relatively low compared to

th e

power output of t ransmi t t e r s , they may cause interference in sensi-

tive devices such as receiv ers. Also, because of their broadband charac-

teristics, they may represent a threat over several octaves or more of

th e frequency spectrum. Sources include computer clocks, printers,

power supplies, automobile engine ignition systems, fluorescent lamps,

electrical

motors, switches and relays, etc.

1 3 Modes of Coupling

Emissions may be coupled see Table 1.1) by one or more paths from the

interference source to the susceptible victim device s) These paths are

classified as either conduction paths or radiation paths.

Table 1 1 Emissions, Susceptibility, and Primary Modes of Coupling

EMI Sources EMI Victims

Mode of Coupling

Transmitter

Transmitter

Electronic Device

Electronic Device

Receiver

Electronic Device

Receiver

Electronic Device

Electronic Device Electronic Device

Radiated

Antenna to Antenna

Radiated

Antenna to Wires

Antenna to Case Penetration

Radiated

Wires to Antenna

Case Penetration to Antenna

Radiated

Wire to Wire

Wire to Case Penetration

Case Radiation to Wire

Case Radiation to Case Penetration

Conducted

Signal Wires

Power Cables

Common-Source Impedance

Common-Ground Impedance

Conduction paths include all forms of direct conductor, wire, and

cable coupling. Conducted interference may enter a victim receptor as a

result of directly coupled conductors or wiring leads between victim



6 INTRODUCTION TO ELECTRONIC SYSTEM DESIGN FOR EMC

receptor and a source of electromagnetic interference. Typical con-

ducted paths include interconnecting cables, power leads and control

and signal cables, common-ground impedances, and common-source

impedances.

Radiation paths involve propagation through the environment or

induction (near-field). Radiating interference includes situations in

which emissions (1)' en ter via a receiving system an ten na , if applicable,

(2) penetrate a shielded housing at the openings and couple into low-

level circuitry, or (3) couple into various signal, control, or power leads

of a receptor via radia ted pa ths .

1 4 Susc eptible Equipm ents

Any device capable of responding to electrical, electromechanical, or

electronic emissions, or to the fields associated with these emissions, is

vulnerable to EMI. Susceptibility of all such devices may be divided

into two categories: (1) devices that are frequency selective and (2)

devices susceptible to interfering emissions over a broad band of fre-

quencies. Frequency-selective devices primarily include equipments or

systems such as communication, radar, and navigation receivers. Typi-

cal devices that may be considered vulnerable to interfering emissions

over a few or many octaves include sensors, computer process control,

switches, relays, indicator lights, electro-explosive squibs, recording

devices, logic circuits, and meters.

1 5 EMC Design Consideration vs Sys tem Life Cycle

The scope of this book may be made clearer by discussing the various

phases in the life cycle of an equipment or system and the EMC design

considerations that apply to each phase. Figure 1.4 illustrates the inter-

relationship between the levels of EMC design and system life cycle

phases.

EMI is an interdisciplinary problem that can be solved by careful

consideration and attention during all phases in the life cycle of an

equipment or system. In order to achieve EMC economically and effec-

tively, it is necessary to use a combination of the following:

• Interference analy sis techn iques to identify a nd define the prob-

lems

• EMC specifications and stan dar ds to ensu re com prehensiveness

during equipment design and development stages



EMC DESIGN CONSIDERATION VS. SYSTEM LIFE CYCLE

EMC Design Consideration vs. System Life Cycle

Intersystem, Intrasystem and

Electromagnetic Environment

System

Definition

Intrasystem, Subsystems,

Equipments, Functional Stages,

Circuits and Components

System Design

and Development

Fig ure 1.4 EMC design and system life cycle phases.

V.

System

Operation

• EMI control devices and techniques during equipment or system

design, development, and production to ensure that specifications

and standards are met

• EMC system design to ensure tha t equipments and subsystems do

not have adverse EMI interactions

• Measurements to provide analysis inputs and ensure compliance

with EMC specifications and standards

• Suppression techniques during installation and operation to solve

specific problems tha t arise as a result of severe or unusual operat-

ing conditions

During each phase of the equipment or system life cycle, responsible

management and engineering personnel must give appropriate atten-

tion to the particular EMC considerations applicable to their areas of

responsibility if EMI-free operation is to be assured.

Techniques used for EMC design of system s are significantly differ-

ent from techniques used for EMC design of equipments. The system

designer is interested in determining interactions among various sys-

tems.

It is necessary to define the output characteristics of EMI sources

and the susceptibility of receiving equipments. Consequently, it is not

necessary to know detailed internal characteristics of equipments.

Thus, in system EMC design, the individual elements can be regarded

as black boxes with defined input/output characteristics. On the other

hand, in analyzing equipments to determine their EMI properties, the

designer must consider the detailed characteristics of components and

circuits that the equipment comprises. Brief discussions of the major

design considerations at each phase in the system life cycle are pre-

sented in the following sections.



8 INTRODUCTION TO ELECTRONIC SYSTEM DESIGN FOR EMC

1 5 1 System Definition Phase

The first step in the life cycle of a system is the definition phase. During

this phase, the system progresses from missions or applications to its

most basic form, which could be either an idea that originated at a

research laboratory or the operational requirement of potential users. It

then moves to the definition and specifications of the major system

characteristics such as size, weight, type of modulation, data rate, infor-

mation bandwidth, transmitter power, receiver sensitivity, antenna

gains, spurious rejection, etc. It is essential that careful consideration

be given to EMC during this definition phase, because the major charac-

teristics of equipments and systems are defined during this phase.

During the definition phase, the system planner must consider EMI

problems that are likely to he encountered (1) within or between ele-

ments of the system (intrasystem), (2) between elements of the system

and elements of other systems that are likely to be operating in the

same general area (intersystem), and (3) between elements of the sys-

tem and the electromagnetic environment in which it is to be operated.

The intrasystem EMI problem is shown in Fig. 1.5. EMI results

because noise spikes on both nearby power cables and wiring harnesses

are coupled into low-level, sensitive circuits as a result of conducted

Generator

and

Regulator

Ground

1.

Power Cable Conducted Emission

2. Power Cable Conducted Susceptibility

3.

Intercoraiecting Cable Conducted Emission

4.

Interconnecting Cable Conducted Susceptibility

5.

Antenna Lead Conducted Emission

6. Antenna Lead Conducted Susceptibility

7.

Common Ground Impedance Emission Coupling

8. Common Ground Impedance Susceptibility Coupling

9.

H Field Radiation

10. E Field Radiation

11.

H Field Susceptibility

12. E Field Susceptibility

Fig ure 1.5 The Intrasystem EMI problem.



EMC DESIGN CONSIDERATION VS SYSTEM LIFE CYCLE 9

coupling on power cables and/or signal cables, and radiated magnetic

and electric-field coupling from box to box, cable to cable, box to cables,

or cables to box.

Intersystem EMI problems may result from signals that are coupled

from the transmitting antenna of one system to the receiving antenna

of another system. This intersystem EMI problem is particu larly seri-

ous when many systems are required

to

simultaneously operate

in a

limited physical area such as

a

ship, an airpla ne,

a

vehicle, a b uilding,

a

military base, an industrial site, a hosp ital, or a city. This typ e of prob-

lem

is

il lustrated

in

Figu re 1.6, which shows

a

mobile communication

system attempting to receive signals from distant locations while oper-

ating

in

the imm ediate vicinity

of

tran sm itter s associated with other

systems.

The type of analysis that is performed at th e system definition stage

must rely on assumed or typical EMI characteristics for the individual

elements

of

the system. Concentration

is

directed

to the

manner

in

which these elements interact in th e to tal system from an EMI stand-

point. EMC design considerations during

the

definition ph as e will

include the selection of frequency bands; allocation

of

system param e-

ters such as transmitter power, antenna gains, receiver sensitivity, type

of modulation, rise time, and information bandwidth; determination of

system EMI specifications; and identification of po ten tia l deficiencies

and problem areas.

1 5 2 System Design and Development

Design and development is the second phase in the life cycle of a system.

Du ring this phase, the system progresses from the previously established

specifications

to

the final hardw are item.

In

the process

of

designing

a

Interfering

Signal

JIBlli

l l l l l

eeeee

MMM

Desired Signal

Fig ure 1.6 Intersystem EMI problems.



10 INTRODUCTION

TO

ELECTRONIC SYSTEM DESIGN

FOR

EMC

system, there are a number of decisions that must be made. In general,

an equipment may be considered to consist of combinations of functional

stages such as amplifiers, mixers, frequency converters, filters, modula-

tors, detectors, display or readout devices, power supplies, etc. For each

equipm ent, ther e are a num ber of imp ortan t factors, including EMC, th at

must be considered. For example, in the case of receivers, it is necessary

to define the number of amplifier and mixer stages that will be used and

to establish th e allocation of gain, selectivity, and sensitivity among th ese

stages. More importantly, it is necessary to develop an overall block dia-

gram for the receiver with a complete description of the gains, frequency

responses, input and output impedances, dynamic ranges, and suscepti-

bility levels for each stage.

Personnel responsible for the design and development of a system

must be concerned with EMI problems resulting from signals exter-

nally coupled between antennas of different elements of the system and

other tran sm itte rs an d receivers in the environm ent, as well as in tern al

EMI problems resulting from cable coupling, case radiation, and case

penetration.

1 5 3 Sys tem Operation

The final phase in the life cycle of the system shown in Fig. 1.4 is the

operational phase. During this phase, a system that has been designed

and developed is placed into operation. Overall, the EMI characteristics

that are considered at the operational level are similar to those per-

formed at the system definition level. Usually, personnel responsible for

compatible system operation are more concerned about the interaction

of the elements of the system, both with each other and with elements

of other systems, than they are in the interna l characteristics of the ele-

ments.

Mitigation techniques that apply to EMI between transmitters and

receivers include frequency, time, location, and direction management.

Each of these mitigation techniques results in a number of individual

EMC devices and techniques. These EMI-mitigation techniques are

also useful in the system definition stage. This especially applies for the

frequency management heading.

1 6 Overview of Handbook

Each area of technology has special terms and definitions that apply.

Chapter 2 describes the terms and definitions that are used by the



OVERVIEW OF HANDB OOK 11

EMC community, and learning these terms and definitions will help one

to understand the language used by EMC engineers.

Chapter 3 discusses the EMI considerations that apply to wireless

communication systems and presents the methods tha t may be used to

mitigate EMI problems in wireless systems. Most of the problems asso-

ciated with wireless systems result from transmitters interfering with

receivers.

Chapter 4 presents a discussion of the EMI considerations that

apply to electronic systems that are not transmitters and/or receivers.

The various EMI coupling modes that result in problems are pre-

sented, and techniques that may be used to mitigate EMI for each of

these coupling modes are identified. EMC design techniques that may

be used to mitigate EMI include Grounding (Chapter 5), Shielding

(Chapter 6), Bonding (Chapter 7), Filters, Ferrites, Isolators, and

Transient Suppressors (Chapter 8), Cables and Connectors (Chapter

9),

and Summary of EMI Control Techniques (Chapter 10).

Appendix A provides an approach to calculating crosstalk or cou-

pling between circuits, wires, or cables.

Suggested Readings: EMI/EMC

[1] Ott, Henry, Electromagnetic Compatibility

Engineering

Hoboken,

NJ: Wiley/IEEE Press, August 2009.

[2] Paul, Clayton, Electromagnetic Compatibility for

Engineers

with

Applications to Digital Systems and

Electromagnetic Interference

Hoboken, NJ: Wiley/IEEE Press, September

2003.

[3] Celozzi, Salvatore, Rodolfo Areneo, and Giampiero Lovat, Electro-

magnetic

Shielding

Hoboken, NJ: Wiley/IEEE Press, April 2008.

[4] Morrison, Ralph, Grounding and Shielding Circuits and Interfer-

ence

5th ed., Hoboken, NJ: Wiley/IEEE Press, March 2007.



Chapter 2

Basic Terms and Definitions

A number of specialized terms are applicable to the characterization,

specification, and/or measurement of electromagnetic interference

(EMI).

It is particularly important that individuals responsible for

ensuring that equipments and/or systems operate in an electromagneti-

cally compatible manner be familiar with the basic term s and definitions

that are widely used throughout the electromagnetic compatibility

(EMC) community. This chapter presents a discussion of the basic terms

and definitions that are important to the EMC engineer or technician.

2.1

Decibels

In order to characterize EMI, it is often necessary to deal with signal

and susceptibility levels tha t range over many orders of magnitude. For

example, receivers typically have sensitivities on the order of 1CT

13

watts,

whereas high-power transmitters have power outputs on the

order of kilowatts or megaw atts. Signals that range over many orders of

magnitude such as this are usually plotted on a logarithmic scale so

that the resolution may be maintained over each decade. One logarith-

mic representation that is often used in the EMC community is the

decibel, which is defined as follows:

dB = 10 logf—) (2.1)

The decibel can also be expressed in term s of a voltage or current ratio

as shown below:

d B = 10 log \-L—i = 10 log _ x -2 (2.2)

13



14 BASIC TERMS AND DEFINITIONS

= lOlogf—1 whenZj = Z

2

= 20 1og - 1 (2.3)

dB

=

10 log fil£l|

(2.4)

\

2

=

10 logl -i when

.

log(j-l) 20 log j-l) 2.5)

2.2 EMI Condu cted Term inology

The term conducted EMI refers to EMI tha t is coupled between circuits,

equipments or systems as a result of being conducted along an intercon-

necting power or signal wire or cable. The units of measure for con-

ducted EMI are usually expressed in term s of voltage or current.

2.3 EMI R ad iated Ter m inology

The term

radiated EM I

refers to EMI that is coupled between circuits,

equipments, or systems via electromagnetic fields that are radiated

from an EMI source and picked up by susceptible circuits, equipments

or systems. The units of density or measure for radiated EMI are usu-

ally expressed in terms of power density or field streng ths.

2.4 R ep rese nta t ion of S ign als in the Time and F req uen cy

D o m a i n s

In general, EMI signals can be represented in terms of their character-

istics in either the time or frequency domains, and Fourier analysis

may be used to transform signals from one domain to the other. This



REPRESENTATION OF SIGNALS IN THE TIME AND FREQUENCY DOMAINS

15

section presents the basic Fourier analysis relationships and describes

their application to some typical EMI type signals.

2.4.1 Fourier Series

A periodic function of time v(t) having a fundam ental period T

o

can be

represented

as an

infinite sum

of

sinusoidal w aveforms. T his s um ma-

tion, called a Fourier series, may be written in several forms. One such

form is the following:

4 \ A v A 27int ^ ^ . 27int

/ o

„

V

W

=

A

o

+

X

A

n

c o s

— +

S

B

n

s m

— (

2

-

6

)

n = 1

° n = 1 °

The con stant A

o

is the a verage va lue of v t) given by

1

A

o

= — f

T

°

/2

v( t)dt (2.7)

0

T

0

J-T

0

/2

while the coefficients A

n

and B

n

are given by

T

o

l

The exponential form

of

the Fou rier series finds extensive applica-

tion in communication theory. This form is given by

n

=

- 0 0

where V

n

is given by

it (2.11)

The Fourier series of a periodic function is thu s seen to consist of a

sum ma tion of harm onics of a fund am ental frequency

f

0

= 1/T

O

. The

coef-



16

BASIC

TERMS AND

DEFINITIONS

ficients V

n

are called

spectral amplitudes;

th at is, V

n

is the amplitude of

the spectral component at frequency nf

0

2.4.2 Fourier Transform

A periodic waveform may

be

expressed

as a sum of

spectral compo-

nents . These components have finite amplitudes and are separated by

finite frequency intervals f

0

= 1/T

O

. The normalized power of the wave-

form is finite, as is also th e no rmalized energy of th e sig nal in an inter-

val T

o

. Now suppose th e period T

o

of th e waveform is increased with out

limit. Then, eventually, a single-pulse nonperiodic waveform would

result .

As T

o

approaches infinity,

the

spacing

of

spectral components

becomes infinitesimal. The frequency of th e sp ectral com ponents, wh ich

in the Fou rier series w as a discontinuous variable with a one-to-one

correspondence with the integers, becomes instead a continuous vari-

able. The norm alized energy of the nonperiodic waveform rem ain s

finite, but, since

the

waveform

is not

repeated,

its

normalized power

becomes infinitesimal. The spectral amplitudes similarly become infini-

tesim al. The Fou rier series for th e periodic w aveform

v t) = I V

n

e

n = -oo

becomes

v(t)

=

f°°V(f)e

j27lft

df (2.13)

oo

The finite spectral amplitudes V

n

are analogous to the infinitesimal

spectral amplitudes V(f)df. The quantity V(f) is called the

amplitude

spectral density

or more generally the Fourier transform of v(t). The

Fourier transform is given by

V(f) = f°°V(t)e"

j27cf t

dt (2.14)

J—oo

2.4.3 Spectral Representation

As discussed in the preceding sections, the re is a direct relationship

between the time domain and frequency domain representation of a sig-



REPRESENTATION OF SIGNALS IN THE TIME AND FREQUENCY DOMAINS

17

nal.

This relation ship is illu stra ted in Fig. 2.1 for several different types

of signals. Referring to the figure, a perfect sinusoidal signal produces a

single spectral component at a frequency (f

0

), which is the reciprocal of

the period (T

o

) of th e sine wave. Signals th at are not perfect sine w aves

produce spec tral componen ts over a rang e of th e frequency s pectru m. In

general, the spectral content is related to the time domain characteris-

tics of th e signal.

For example, a periodic tria ng ula r or trapez oida l pulse will produce a

spectrum that has a fundamental frequency (f

0

), which is the reciprocal

of the pulse period (T). The spectrum will contain discrete components

at integer multiples (harmonics) of f

0

, as illustrated in Figure 2.1 . The

envelope of the spectrum of the periodic triangular or trapezoidal pulse

will be flat for frequencies less th an 1/TTC where x is the p ulse width. The

am pli tud e of th is flat po rtion of th e s pec trum will be equ al to 2A x/T,

Time Domain

(Oscilloscope View)

Frequency Domain

(Spectrum Analyzer,

EMI Receiver View)

Sine Wave

A

«—T

A --

Non Sine Wave

but Periodic

Ultra Short Pulses

Long Period

A-

2Ar

4

T

Single Pulses

2Ar

Fig ure 2.1 Time and frequency domain representation of signals.



18 BAS IC TERM S AND DEFINITIONS

where A is the pulse am plitude. The spectrum of the t rian gu lar or trap-

ezoidal pulse will start to roll off at a rate of 20 dB per decade for fre-

quencies greater than

1/rcx,

and this 20 dB per decade roll-off will

continue to a frequency equal to

l/nx

Y

where T

r

is the rise time of the

pulse. For frequencies greater than

l/nx

r

the pulse spectrum will roll

off at 40 dB per decade. Figure 2.1 illustrates the spectrum for periodic

rectangular pulses where the pulse width is much less than the period.

For this typ e pulse, the spectrum is flat out to a frequency given by

llnx,

which includes a nu m ber of harm onics of the fun dam ental frequency.

For frequencies above 1/KT, th e sp ectrum rolls off at 20 dB per decade.

Figure 2.1 also illustrates the spectrum for a single (aperiodic) trian-

gular pulse. The aperiodic pulse will produce a continuous spectrum.

The amplitude of the spectrum for the aperiodic pulse will be flat, with

spec tral den sity (amplitude) equal to 2Ax, for frequencies less t h an

1/nx.

The spectral density will start to roll off at 20 dB per decade at frequen-

cies greater th an 1/TTC, and thi s 20 dB per decade roll-off will continue to

a frequency equal to l/7CT

r

For frequencies greater than

l/ni

r

the pulse

spectrum will roll off at 40 dB per decade.

In general, signals with significant spectral energy in the higher por-

tion of the frequency spectrum will be more difficult to control or sup-

press and are more likely to create EMI problems in a system. The

po tentia l for EMI problems resu ltin g from p ulse type sign als will gener-

ally increase as

• the pulse repetition rat e increases,

• the pulse wid th decreases, and/or

• th e rise time decreases.

For this reason, special consideration must be given to pulse signals,

such as computer clocks, with high repetition rates and short rise

t imes. Also, short-duration transients (such as lightning, electromag-

netic pulses, and electrostatic discharges) with short rise times can

raise havoc in an electronic system.

2.5

Transients

Transients represent a major source of EMI. Furthermore, an under-

standing of transients and their amplitude spectrum occupancy and

phase relations are param oun t to an und erstandin g of broadband emis-

sions discussed in the next section. As presented there, broadband

emissions may be coherent (e.g., a transient or impulse) or incoherent

(e.g., bandwidth-limited white noise). The former results in a 20 dB/



TRANSIENTS 19

decade bandwidth relation, while incoherent broadband emissions

result in a 10 dB/decade bandwidth dependency.

2.5.1 Transient Sources

When either an emitting EMI source or a potential victim receptor per-

forms over a broadband of frequencies, it is likely that it develops or

responds, respectively, to transients. Transients distinguish themselves

by having a low duty cycle and fast rise and/or fall times. The duty

cycle, 8. of an emitting source is defined as:

8

= xxf

r

where,

x

= equivalent pulse or impulse width at the 50 percent height

f

r

=

pulse repetition rate, or average number of

pulses

or impulses

per second for random occurrences

Most transients from incidental emitting sources correspond to duty

cycles that

are

very small,

i.e.,

less than 10~

5

. Table

2.1

lists some

approximate duty cycles identified to the nearest order of magnitude

corresponding to

a

few transient sources. When th e duty cycle becomes

significantly greater than 10~

5

, such as 10~

3

for radar or 0.5 for a com-

puter clock, the emitting source is no longer regarded as a transient,

although it may still have fast rise times and therefore broadband emis-

sions.

Transients have become a major problem because so many emit-

ting sources are now operational and because computers, digital and

control devices, among others, are especially susceptible.

Table 2.1

Typical Transient Sources

Emitting transient source

Fluorescent lamps

Ignition systems

Idle speed

Fast speed

Relays and solenoids

Casual us e

Pinball machine

Teletype

Brush-commutator motor

On-off switches

Wall switch

Lathe

Copy machine

Repetition rate

100 pps

100 pps

10

3

pps

10"

3

pps

l p p s

10 pps

10

3

pps

10~

4

pps

10"

3

pps

10 ~

3

pps

Impulse width

10~

7

s

10

8

s

10~

8

s

10 ~

7

s

10

7

s

10~

7

s

10

8

s

10

6

s

10 ~

7

s

10 ~

7

s

Duty cycle

10~

5

10~

6

10~

5

i o

- i o

io-

7

10 ~

6

10~

5

i o

- i o

l o

- i o

l o

- i o



2

BASIC TERMS AND DEFINITIONS

The time and frequency domain representations for lightning, EMP,

and ESD are illustrated in Fig. 2.2. The parameters associated with

each of these transien t sources are also shown in the figure.

Figure 2.3 shows the spectral density for various pulse shapes. Refer-

ring to th e figure, th e high-frequency roll-off is 20 dB per decade for the

rectangular pulse, 40 dB per decade for the trapezoidal and the criti-

cally damped exponential pulses, and 60 dB per decade for the cosine

squared pulse.

2.6 Narrowband Em issions

The term narrowband

emission

means that the emission bandwidth is

narrow or less than some reference bandwidth as shown in Figure 2.4.

The reference bandwidth may be that associated with a potentially

susceptible victim receptor. Thus, if an emission source is narrowband

with respect to the victim receptor, the power received by the victim

receptor will be approximately equal to the total power present in the

emission.

0-100% points o f ideal pulse

=10-90% points o f real-world pulse

20 dB/DEC. Slope

h

-40 dB/DEC.

Frequency

^ \ P a r a m e t e r

ffe t ^ * ^ \ ^

Lightning

EM P

Electrostatic

discharge

0.5 //s ec

5 nsec

1 nsec

X

20 //sec

50

nsec

150

nsec

f

x

= Vn t

17 kHz

6.4 MHz

2 MHz

f

2

= VitTj.

640 kHz

64 MHz

300 MHz

A

100 kA

50 kV/m

5A*

2A r

4kA/kHz

5V/m/kHz

1.5

A/MHz

*Can create peak fields of several kV/m at 10 cm.

Fig ure 2.2 Time and frequency domain representa tions for lightning, EMP,

and ESD.



BROADBAND EMISSIONS

21

i/nr

20

« 4

6

m 8

| 100

55

g 120

4

°

|

16

180

200

A

=

Peak amplitude

of pulse

r

=

Average pulse

duration in //sec

Ar

=

Area under pulse

in A//sec or

A/MHz

0.1/r

1/r 10/r 100/r 10

3

/r

Frequency in

MHz

Figure 2.3 Spectral density for various pulse shapes.

10

4

/r

10

5

/r

2.7 Broadband Emissions

The term broadband emission indicates that the emission bandwidth is

broad

or

greater th an some reference b andw idth

as

shown

in

Fig. 2.4.

Here, the reference band wid th may be th at associated w ith a poten-

tially susceptible victim receptor. In th is case, th e victim receptor will

not receive all of the power pres en t in th e em ission and it will be neces-

sary to adjust th e total power to com pensate for the em ission and recep-

tor bandwidths. That is it will be necessary to reduce the power of the

emission to represent the power that will be received within the victim

bandwidth. The adju stm ent will depend on whether the emission is

coherent or incoherent.

A

broa dban d signal or emission is said to be coherent w hen neighbor-

ing frequency increments are related or well defined in both amplitude

and phase. For broadband situations, neighboring amplitudes are both

equal and in phase.

Coherent broadband voltages vary proportionally to th e ratio of th e

victim bandwidth

to the

emission bandw idth,

and it is

necessary

to

make a coheren t ban dw idth correction (CBC) to compen sate for th e dif-

ferences in bandwidth.



22

BASIC TERMS AND DEFINITIONS

Narrowband N B )

Broadband

B B )

>

Receiver

selectivity

curve

1

i

y Emission

Y bandwidth

\ <

Receiver

\ bandwidth

Frequency

Example

of NB

terms

dBV

dBmV

dBz/V

dBA

dBm

dBpT

dBV/m

dB//V/m

dB//A/m

p

u

A

Receiver '

selectivity

curve

i J Emission

^

^ bandwidth

> Receiver

bandwidth

Frequency

Example

of BB terms:

dBV/MHz

dBmV/kHz

dB//V//MHz

dB//V/nVMHz

dB//A/MHz

dB//A/m/MHz

F ig u re 2.4 Narrowband and broadband emissions relative to the mea suring

receiver bandwidth.

CBC =

Victim bandwidth

Emission bandwidth

(2.15)

2.7.1 Inco herent Broadband Em ission

A signal or emission is said to be incoherent when it is not coherent,

viz., when neighboring frequency increments are random or pseudo-

random (bandwidth limited) in either phase or both amplitude and

phase. Examples of incoherent broadband emission sources are gas

lamps (de energized), noise diodes, blackbodies including internal

receiver noise, and corona discharge from high-voltage sources.

For incoherent broadband emissions the voltage phase terms are

random from neighboring frequency increment to increment, the incre-

mental voltages do not add in phase but add in an RMS fashion. The

incoherent bandwidth correction (IBC) to compensate for incoherent

broadband emissions is

IBC =

Victim bandwidth

Emission bandwidth

(2.16)

2.8 Freq uen cy and W avelength

Sometimes the term

wavelength

instead of

frequency

is used. To convert

from frequency, f, in hertz to wavelength,

X

(length of one cycle of fre-

quency) in m eters, th e velocity of propa gation in air is use d:



UNIT S OF MEASU RE FOR EMI SIGNALS 23

wh ere C « 3 x 1 0

8

meters/second in air.

Thus,

X 3 x

10

8

/f

Hz

m ete rs (2.17)

(2.18)

where,

f

Hz

= frequency in Hz

f*MHz = frequency in MHz

2.9 U nits of M easure for EMI Sign als

The previous sections have described the different types of EMI sig-

nals and have discussed the various ways in which these signals may

be represented. This section provides a summary of the basic units

that are used for each type of signal. In general, EMI signals may be

specified in e ither linear un its [e.g., volts (V), amps (A), etc.] or in log-

arithmic units, which are usually expressed in terms of decibels (e.g.,

dBV, dBA, etc.). EMI may be present in the form of conducted signals,

in which case the units of measure will be volts or amps, or in the

form of rad iated signals that are specified in term s of power density or

field strength. It is also important to note that EMI signals may be

either narrowband or broadband, and for broadband signals, it is nec-

essary to reference the signal level to some unit of bandwidth (e.g.,

volts/hertz). Table 2.2 summarizes the units of measure that are used

for the various types of EMI signals.

Table 2.2

EMI Un its of M easure (continues)

Conducted signals

Narrowband Broadband

Power Voltage Current Power Voltage Current

W V A W/Hz V/Hz A/Hz

dBW dBV dBA dBW/Hz dBV/Hz dBA/Hz

dBm dB^iV dBmA dBm/MH z dB^V/MH z dB^iA/MHz



24 BASIC TERMS AND DEFINITIONS

Table 2.2 EMI Un its of M easure (continued)

Power

density

Narrowband

Electric

field

E)

Radiated signals

Magnetic

field H)

Power

density

Broadband

Electric field

E)

Magnetic

field H)

W/m

2

V/m A/m

dBW/m

2

dBV/m dBA/m

dBm/m

2

dB^iV/m dB^A /m

W/m

2

/Hz V/m/Hz A/m/Hz

dBW/m

2

/Hz dBV/m/Hz dBA/m/Hz

dBm/m

2

/MHz dBjuV/m/MHz dB|iA/m/MHz

Suggested Readings: Basic Terms and Definitions

[1] Hoolihan, Daniel D., "EMC and M easurem ent Uncertainty— Lab

34 and CISPR 16-4-2," Com pliance Magazine, 2010 Annual Guide,

p.

10.

[2] Hoolihan, Dan iel D., "CISPR 11: A Historical and E volutionary

Review," C ompliance M agazine, August, 2010, p. 8.

[3] Heirman, Don, and Manfred Stecher, "History of CISPR,"

Compli-

ance Magazine,

Ju ne 2010, p. 36.

[4] Jon es, B rian, "EMC S tan da rd s from a European Perspective,"

Compliance Magazine, 2010 Annual Guide, p . 54.

[5]

"List of EMC D irective Stan dard s," Compliance Magazine, 2010

Annual Guide,

p . 59.

[6]

Dash, Glen, "Why Digital Devices Ra diate," Compliance Magazine,

2010 A nn ual Guide, p. 26.

[7] Dash , Glen, "Designing for Compliance—We Put Theory to the

Test," Conformity, March 1998, p. 10.

[8] "Spectrum A nalys is B asics," from 1997 Back to Basics Seminar,

Agilent Technologies.

[9]

"EMC Narrowband and Broadband Discrimination with a Spec-

trum Analyzer or EMI Receiver,"

Conformity,

December 2007.



Chapter 3

Com m unication System s EMC

Our society relies on the ability to establish and maintain extensive,

reliable communications. In general, the requirement for use of the

electromagnetic spectrum for communication, navigation, and radar

systems has been rapidly increasing. Our military strategy is based on

the rapid deployment of dynamic forces supported by an extensive Com-

mand, Control, Communication, and Intelligence (C

3

I) network to pro-

vide the information required for battle management. In the civilian

sector, our communication requirements have increased drastically as a

result of the mobility of our society and our dependence on computers.

The cellular telephone has significantly increased the capacity of our

mobile communications, and fixed point-to-point microwave and satel-

lite communication systems provide an extensive data transmission

network for computer systems.

One of the most important considerations in the design, installation,

and operation of a communication-electronic (CE) system is that of

achieving and maintaining EMC between the system and the other CE

equipments in the immediate vicinity. EMC is the ability of equipments

or systems to function as designed, without degradation or malfunction,

in an intended operational electromagnetic environment. The equip-

ment or system should not adversely affect the operation of, or be

adversely affected by, any other equipment or system.

To succeed in achieving EMC, and to permit efficient use of the fre-

quency spectrum, it is essential that engineers, technicians, and users

responsible for the planning, design, development, installation, and/or

operation of CE equipments devote careful attention to potential EMI

problems. This will ensure EMC without either the wasteful expense of

over-engineering or uncertainties of under-engineering.

25



26 COMMUNICATION SYSTEMS EM C

3.1 Com mu nication System EMI Problems

In a typical communication situation, the receiver must be able to pick

up

its intended signal, which is probably relatively weak, while operat-

ing in the presence of

a

number of strong, potentially interfering signals

that result from other CE systems operating in close proximity. Con-

versely, the transm itter must be able to tran sm it a relatively strong sig-

nal without causing interference to sensitive receivers.

The basic EMC requirement is to plan, specify, and design systems

that can be installed in their operational environments without creat-

ing or being susceptible to interference. To help satisfy this require-

ment, careful consideration must be given to a number of factors that

influence EMC. In particular, it is necessary to consider major sources

of EMI, modes of coupling, and conditions of susceptibility. The system

designer should be familiar with the basic tools (including analysis,

measurement, control, suppression, specifications, and standards) that

are used to achieve EMC.

This chapter identifies potential EMI problems that may occur

between transmitters and receivers. The emphasis in this chapter is

specifically oriented toward EMI signals that are generated by poten-

tially interfering transmitters, propagated and received via antennas,

and that cause EMI in receivers associated with communication sys-

tems.

3.2 EMI Interaction s betw een Transmitters and

Receivers

In the planning and design of a communication system, it is important

to recognize that there are several different means by which EMI may

occur. For each situation, the appropriate types of EMI must be consid-

ered. The important types of EMI, which are shown in Fig. 3.1, may be

considered to be in one of three basic categories: co-channel, adjacent-

signal, or out-of-band. These categories are defined as follows.

Co channel

EMI

refers to interference resulting from signals that

exist within the narrowest passband of the receiver. For superhetero-

dyne receivers (which is the type used for many applications), the fre-

quency of co-channel interference must be such th at the interference is

translated to the intermediate frequency (IF) passband in the same

manner as the desired signal. This requires th at the frequency of co-

channel interfering signals equal the tuned radio frequency plus or

minus one half the narrowest IF bandwidth. Although the receiver is

most sensitive to this type of interference, it is usually easily controlled



EMI INTERACTIONS BETWEEN TRANSMITTERS AND RECEIVERS

27

Receiver Susceptibility

Transmitter Emission

a) Co-channel EMI



b) Adjacent-Channel EMI


\r~r

EMI Resulting from

Transmitter Harmonic and

Receiver Fundamental


1.

Transmitter Harmonic—Receiver Fundamental


EMI Resulting from

Transmitter Fundamental

and Receiver Spurious


2.

Transmitter Fundamental—Receiver Spurious


Transmitter Emission \ i .

E M I

Resulting from Transmitter

U Harmonic and Receiver Spurious

3. Transmitter Harmonic—Receiver Spurious

c) Out-of-Band EMI

Figure

3.1

Types of transmitter-receiver EMI.

by avoiding co-channel assignments within a relatively large control

zone

over

which

this type

of interference

may occur.

Adjacent signal

EMI

refers to potentially interfering signals that

exist

within or near the

receiver radio

frequency

RF) passband

but fall



28 COMMUNICATION SYSTEMS EMC

outside of the IF passband after conversion. The most significant adja-

cent-signal EMI effects result from modulation sidebands, intermodula-

tion, and transmitter noise.

The adjacent-signal EMI region may extend over a considerable

range of frequencies on each side of the tuned frequency. For example,

for a typical UHF communication transceiver having 25 kHz channel

spacing, the adjacent-signal EMI region may include 400 channels (i.e.,

10 MHz) on each side of the desired channel. Although the adjacent-sig-

nal EMI region includes a relatively wide range of frequencies, the

receiver is not particularly sensitive to these signals. As a result, adja-

cent-signal EMI is usually limited to co-site situations involving trans-

ceivers that are located within

1

or

2

km of each other.

Out of band EMI refers to signals having frequency components that

are significantly outside of the widest receiver passband. The most sig-

nificant out-of-band EMI effects result from transmitter harmonics

interfering with receiver fundamentals or transmitter fundamentals

interfering with receiver spurious responses. EMI between transmitter

harmonics and receiver spurious responses are also possible but

extremely unlikely. Because of the power levels involved, out-of-band

EMI is usually restricted to co-site situations .

3.3 EMC D esign of Com mu nication System s

In order to design a communication system for EMC, it is necessary to

consider the susceptibility of each receiver to both the design and spuri-

ous outputs (individually and collectively) of the potentially interfering

transmitters. The factors that must be considered for each transmitter

output (or group of transmitter outputs) include:

1. Transmitter power (PT)

2.

Transmitting antenna gain in the direction of the receiver (G^R)

3.

Free-space propagation loss between the transmitter and receiver (L)

4. Receiver antenn a gain in the direction of the transm itter (GR

T

)

5. The amount of power required to produce interference in the

receiver

(PR)

in the presence of the desired signal

Factors th at must be considered in the EMC design of communica-

tions systems include both the design (intentional) and operational

performance characteristics of equipment and the non-design (unin-

tentional) and non-operational characteristics. This chapter discusses

equipment charac teristics th at must be considered in EMC design and

describes equipment EMC characteristics of communication systems.



EMC

DESIGN

OF

COMMUNICATION SYSTEMS

29

The necessity for considering parameters such as transmitter spuri-

ous output emissions, receiver spurious responses, antenna side- and

back-lobe radiation,

and

unintentional propagation paths introduces

complications, since

it is now

necessary

to

obtain information

on

equipment non-design characteristics. Unlike equipment design char-

acteristics (which

are

usually well defined

and may be

readily

obtained from equipment specifications), equipment spurious charac-

teristics may not be identified or described in equipment specifica-

tions. Therefore, it is difficult to obtain information on the spurious

characteristics of specific equipments.

For situations in which specific detailed data are not available, sev-

eral different sources of input information may be used to derive

default data. This chapter presents equipment EMC default data that

have been derived from measured equipment characteristics, MIL-

STD-461 limits, and regulations.

The procedure tha t is used for each tran sm itter output emission can

be demonstrated by considering the interference situation that exists

between

a

particular output

of

one

of a

number

of

potentially interfer-

ing transmitters and a victim receiver. For the case of a particular

transmitter output (which may be either a fundamental or a spurious

emission), the power available

at

the receiver is given by:

P

A

(f, t, d, p) = P

T

(f, t) + C

TR

(f, t, d, p) (3.1)

where,

f, t, d, p) = power available at the receiver (in dBm) as a function

of frequency (f), time (t), distance separation (d), and

polarization (p), of both the transmitter and the

receiver and their antennas

f, t) = transm itter power (in dBm)

f, t, d, p) = transmission coupling between transm itter and

receiver in dB

In problems involving interference coupled from

a

transmitting

antenna to a receiving antenna, the transmission coupling function is

represented by:

C

TR

(f,

t, d,

p) =

G

TR

(f, t, d,

p)

-

L(f,

t, d,

p) +

G

RT

(f, t, d,

p)

(3.2)

where,

Gx

R

(f, t, d, p) = the transmitting-source antenna gain in the direction

of the receiver

in

dB



30


EMC

L(f, t, d, p) = the f ree-space pro pa gat io n los s funct ion in dB

G R T ( £

t, d, p) = the

r e c e iv i n g a n t e n n a g a i n

in the

di rec t ion

of th e

t r a n s m i t t e r in dB

B y c o m p a r i n g the power ava i l ab l e at the r e c e iv e r i n p u t t e r m i n a l s to

t he power r equ i r ed

to

p r oduce i n t e r f e r ence

in the

r ece i ve r

at t he fre-

q u e n c y in ques t i on , P

R

(f, t) , it is poss ib le to d e t e r m i n e the i n t e r f e r ence

s i t u a t i o n

for the

p a r t i c u l a r t r a n s m i t t e r o u t p u t b e i n g c o n s id e r e d .

The

r e q u i r e m e n t for EMC is t h a t the power ava i l ab l e at the r ece i ve r be les s

t h a n the p o w e r r e q u i r e d to p r oduce i n t e r f e r ence in th e rece iver. T hu s ,

the condi t ion

for

e l ec t rom agne t i c c omp a t i b i li t y

is:

P

A

( f , t , d , p ) < P

R

( f , t )

(3.3)

On the other hand, if th e power available at the receiver inpu t termi-

nals is equal to or greater th an the power req uired to produce interfer-

ence in the receiver, an electromagn etic interference problem may exist.

Therefore,

an

EMI problem will exist

if

P

A

( f , t , d , p ) > P

R

( f , t ) (3.4)

When P

A

= P

R

, EMC is m arginal, and an EMI problem may or may

not exist.

An indication of the magnitude of a potential interference problem

may be obtained by considering the difference betw een the power avail-

able

and the

susceptibility threshold. This difference

is

referred

to as

the interference margin, IM , and provides a measure of th e to tal contri-

bution to interference, i.e.:

IM(f, t, d, p) = P

A

(f, t, d, p) - P

R

(f, t) (3.5)

The interference margin is defined such that there is a potential

interference problem if the margin is po sitive, and there is little to no

chance of interference if the interference margin is negative.

The expression IM(f, t, d, p) in Eq. (3.5) can be considered to repre-

sent an equivalent on-tune interference-to-noise ratio (I/N) at the

receiver input terminals. If the expressions for P

A

(f, t, d, p) and P

R

(f, t)

are expanded,

Eq.

(3.5) becom es:

IM(f, t, d, p) = I/N = P

T

(f

E

) + G

TR

(f

E

, t, d, p) (3.6)

- L ( f

E

, t , d, p) + G

RT

(f

E

, t, d,p)

CF(B

T

,B

R

,Ai)



EMC DESIGN OF COMMUNICATION SYSTEMS 31

w h e r e ,

=

power tra ns m itte d in dBm at emission frequency (fg)

*»

d, P)

=

transmitter antenna gain in dB at emission frequency

(fg) in the direction of receiver

, t, d, p) = free-space propag ation loss in dB at em ission fre-

quency (fg) between transmitter and receiver

1, d, p) = receiver a n tenna gain in dB at em ission frequency (fg)

in direction of tran sm itte r

PR^R) = receiver susceptibility threshold in dBm at response

frequency (%)

CF(B

T

,

BR, Af) = factor in dB th at accounts for tra ns m itt er and receiver

bandw idths, B^ and BR, respectively, and the fre-

quency separation, Af, between transmitter emission

and receiver response

The final ter m in Eq. (3.6), CF(B

T

,

BR,

Af), tak es in to account t he rela-

tive bandwidths, transmitter modulation envelope, receiver selectivity

curve, and the frequency separation, if

any,

between the transm itter out-

put and the receiver response. The procedure used for determining

CF(BT,

B R ,

Af) is illustrate d by considering th e various possibilities th at

may exist between p articu lar o utpu t response pairs a s shown in Fig. 3.2.

First, if the output and response occur at the same center frequency

(i.e.,

Af = 0), there are two basic co-channel possibilities that may be

considered:

1.

Receiver band wid th is either equal to or larger th an th e tran sm itte r

bandwidth

BR

> B^). For this case, all the power associated with

the transmitter output is received, and no correction is necessary

[i.e. ,CF(B

T

,B

R

,Af) = 0].

2.

Receiver bandwidth is less than the transmitter bandwidth BR <

BT). For this case, only a portion of the power associated with the

emission output is received, and it is necessary to apply a band-

width correction, CF, to account for the bandwidth differences. This

correction for Af = 0 is depe nden t on th e ba nd wi dth ratio s a nd is of

the form:

CF(Af

=

0) = K log

10

(B

R

/B

T

) dB (3.7)

where,

BR = receiver 3 dB ban dw idth in Hz

B

T

= transm itter 3 dB bandw idth in Hz



32

COMMUNICATION SYSTEMS EMC

Power

B

R

> B

T

Power Received =

Pow er Available

Power

Received

Receiver

Susceptibility

Threshold P

R

(f )

Available

Power P

A

(f)

B

R

< B

T

Power Received < Power Available

a ) On-Tune Case Co-channel Frequency Alignment)

Maximum Receiver

Susceptibility Threshold


Threshold P

R

(f)

B

R

=

3 dB B andwidth

Available /

Power P

A

(f) 7

Minimum Transmitter

Emissions

b ) Off-Tune Case Spurious Frequency Alignm ent)

Figure 3.2 Illustra tion of frequency bandwidth relationships.



EMC DESIGN

OF

COMMUNICATION SYSTEMS 33

A co nsta nt for a pa rtic ula r em ission-respon se com bination, K, can be

represented as:

K = 0 for

B R

>

B T

an d co-channel frequency align m en t (3.8)

K = 10 for noise-like signals for which RMS levels apply and B

R

< B^

K = 20 for pulse signals for which peak levels apply and B

R

< B

T

As the transmitter and receiver center frequencies are separated, the

transmitter power can enter the receiver by either of two other possible

means (see Fig. 3.1):

1. The tran sm itte r em ission m odulation sideb ands can en ter the

receiver at the main-response frequency. For this case, the correc-

tion factor is:

CF

R

(Af) = [K log

10

(B

R

/B

T

) + M(Af)] dB (3.9)

where,

M(Af) = mo dulatio n sideban d level in dB above tr an sm itt er

power at frequency separation (Af)

K = as defined in Eq. (3.8)

2. The power at the transmitter main output frequency can enter the

receiver off-tune response. For this case, the correction factor is:

CF

T

(Af) = -S(Af) dB (3.10)

where,

S(Af) = receiver selectivity in dB above receiver fundamental

susceptibility at frequency separation Af

The final bandwidth correction factor that must be applied to the

interference margin due to non-al ignment of the transmit ter output ,

and receiver response is either CF

R

(Af) or CFT(Af), whichever is larger.

The equ ations previously prese nted are applicable to variou s types of

interference problems. In most cases, the major difficulty is to deter-

mine the parameters in the equations. Although this may appear to be

a relat ively s imple undertaking where transmit t ing and receiving

equ ipm ents a re involved, i t is not. This occurs because each tra ns m itte r

produces a number of undesired spurious emissions, and each receiver

has a number of spurious responses, and information is not usually

available on spurious characteristics.



34


Furthermore, it is necessary to consider radiation in unintended

directions via unintended propagation paths. Interactions between

transmitters and receivers having totally different operational func-

tions,

purposes, and technical characteristics also must be determined.

Hence, for the simple case of an EMI assessment involving a single

transmitter and receiver pair, information must be obtained for each

transmitter output and receiver response, and the basic EMI equation

must be applied for each output-response combination.

The following sections describe EMI characteristics for transmitters,

receivers, antennas, and propagation.

3.4 Transm itter Em ission Ch aracteristics

The primary function of a transmitter is to generate radio frequency

power containing direct or latent intelligence within a specified fre-

quency band. In addition to the desired power, transmitters produce

numerous unintentional emissions at spurious frequencies as illus-

trated in Fig.

3.3.

A spurious emission is any radiated output that is not

required for transmitting the desired information. The desired and/or

undesired radio-frequency power generated by transmitters may pro-

duce EMI in receivers or other equipments. Therefore, in evaluating

EMC,

it is necessary to consider all transmitter emissions as potential

sources of interference.

0.2 0.4 0.6 0.8 1 2 4

Frequency Relative to Fundam ental (f^ox)

8 10

Figure 3.3 Transm itter output spectrum containing broadband noise and dis-

crete emissions.



TRANSMITTER EMISSION CHARACTERISTICS 3 5

3.4.1 Fundamental Emissions

To consider th e effect of a tra ns m itt er fu nd am ental ou tpu t on EMI, it is

necessary to define the transmitter operating frequency, the fundamen-

tal power output, the band width associated with the fundamental emis-

sion, and the modulation envelope in the vicinity of the fundamental

emission.

The operating frequency is obtained from frequency assignment data

or operational information, or is defined as part of the statement of the

problem. The transmitter fundamental power output and bandwidth

are nominal data that should

be

available from

the

manufacturer 's

specifications on the transm itter . The mod ulation envelope describes

the relative power

in

the sidebands arou nd th e carrier frequency and

may be represented as described in the following paragraphs.

The trans m itte r fund amen tal outp ut is not actually confined to

a

sin-

gle frequency;

it is

distributed over

a

range

of

frequencies arou nd th e

fundamental. The characteristics of the power distribution in the vicin-

ity of the fund amen tal are determined primarily by the baseba nd mod-

ulation characteristics of the transmitter . The resulting spectral

components are termed modulation sidebands. The power distribution

in the modulation sidebands

is

represented by

a

modulation envelope

function. In general, the modulation envelopes are described by specify-

ing bandwidths or frequency ranges and functional relationships which

describe the variation of power with frequency, M(Af). The modulation

envelope model is:

M(Af) = M(Afj) + Mi log

10

(Af/Afi) (3.11)

where,

Af = m agn itud e of frequency sepa ration =

|

f

- f

or

|

Afj = m agni tud e of frequency sepa rat ion of reference p oint for appli-

cable region = | f - fj |

Mi = slope of mo dula tion envelope for applicable region (dB/decade)

broadband noise generated by the transmitter. This transmit-

ter no ise may be considered to be included in the mo dulation

envelope and may be represen ted as a noise floor th at exten ds

over

a

large portion of th e frequency spectru m.

An example of the resulting functional relationship is shown in

Fig. 3.4. The parameters that are required to specify the m odulation

envelope are th e band wi dth s of applicable regions of con stant slope and

the rate

at

which

the

envelope falls

off

over

the

frequency region

of



36 COMMUNICATION

SYSTEMS EMC

M(Afi)

S

M(Af

2

)

1

Frequency Separation Relative to Reference Frequency

Figure 3.4

Mo dulation envelope represe ntation .

inte rest . The slope, M, in dB/decade, is negative on th e up per side of the

carrier frequency and positive on the lower side of the carrier frequency.

Table 3.1 summarizes modulation envelope parameter values for

some of the more commonly used types of modulation. The off-tune

transmitter emission level is given by:

P

T

(f

0T

± Af) dBm /chan nel =

d B m

(3.12)

For adjacent-signal frequencies that are sufficiently removed from

the transmitter tuned frequency, the major source of interference may

result from the broadband noise generated by the transmitter. This

transmitter noise may be considered to be included in the modulation

envelope and may be represented as a noise floor that extends over a

large portion of th e frequency spec trum .



TRANSMITTER EMISSION CHARACTERISTICS

37

Table

3.1 Co nstan ts for Mo dulation Envelope Re presen tation

Type

of Modulation

|Afj |

M Af;)

d B above

fundamental

Mi

dB/decade)

AM communication

and CW radar

AM voice

FM

Pulse

0

1

2

0

1

2

3

0

1

2

0

-i

1

2

0.1 B

T

0.5 B

T

B

T

l H z

10 Hz

100 Hz

1000 Hz

0.1 B

T

0.5 B

T

B

T

1

lO x

1

1

0

0

- 4 0

- 2 8

- 2 8

0

- 1 1

0

0

100

0

133

67

0

- 2 8

7

60

0

3 3 3

0

2

4

3.4.2 Transm itter Interm odu lation

Intermodulation is the process by which two or more undesired signals

mix in a nonlinearity to produce additional undesired signals at fre-

quencies that are the sum or difference of

the

input frequencies or their

harmonics. In general, intermodulation may occur in both transm itters

and receivers. To determine which type intermodulation predominates

for a given EMI situation, it is necessary to assess the interference level

that results from both transmitter and receiver intermodulation and

consider the case that results in the largest potential interference.

As a rule, the most serious problems result from third-order inter-

modulation and will result from mixing products that are given by:

flM

=

2fi-f

2

(3.13)

or,

where,

f

IM

=

the resulting frequency of the intermodulation product

(3.14)



38


The transmitter third-order intermodulation problem is illustrated

in Fig. 3.5. Referring to the figure, it is seen that intermodulation will

occur in both of the two transmitters. The predominant transmitter

intermodulation situation depends on the geometry and the power lev-

els and frequencies of the two transmitters. In general, it will be neces-

sary to consider both transmitter intermodulation situations to

determine which one produces the most significant signal at the

receiver.

For cases where the frequency separation (Af) between the transmit-

ters is less than or equal to 1 percent of the transmitter frequency, the

equivalent transmitter intermodulation power (P

E

) may be approxi-

mated by Eq. (3.15).

P

E

(dBm) = P

x

(dBm) - 10 dB (3.15)

where,

Pi (dBm) = interfering power available at the transm itter where

the intermodulation occurs

For cases where the frequency separation is greater than 1 percent,

PE may be approximated by Equation (3.16).

P

E

(dBm) = P

x

(dBm) - 10 dB - 30 log

10

Af (percent)

(3.16)

It should be noted that P

E

is the intermodulation signal level at the

transmitter where the intermodulation occurs. To determine the level

at a receiver, it is necessary to include the effects of propagation loss.

Transmitter T

x

Transmitter

T2

Intermodulation

Generated in Tj

Receiver

Intermodulation

Generated in

T2

Fi gu re 3.5 Transmitter intermodulation.



TRANSMITTER EMISSION CHARACTERISTICS

39

3.4.3 Harmonic Emission Levels

Referring to Fig. 3.3, it is readily observed that transmitter emissions

are present

at

frequencies that

are

harmonically related

to the

trans-

mitter fundamental frequency. For the example illustrated in Fig. 3.3,

there are other outputs (of lesser amplitude) present at frequencies that

are harmonics of the master oscillator frequency. However, because of

their reduced amplitude, these master oscillator harmonics do not usu-

ally create EMI problems. The frequencies of harmonics of the funda-

mental output are given by:

f

NT

=

Nf

0T

(3.17)

where,

fjvjT = frequency of Nth harmonic of transmitter

N = integer associated with harmonic

fox

=

operating frequency of transmitter

The amplitude of transmitter harmonic emissions may be expressed

as follows:

P

T

(f

NT

) dBm = P

T

(fbr)

d B m

+ [(

A lo

Sio

N

)

+ B

l (

3

-

18

)

where,

A = slope of harmonic levels

in

dB/decade

B = intercept in dB relative to fundamental emission

If data on transmitter harmonic emission outputs are available from

spectrum signature measurements or other information sources, they

should be used to determine specific harmonic output levels. Con-

versely, in many instances, specific data are not available. Thus, it is

necessary to employ other techniques for determining specific harmonic

levels to be used in EMC assessment.

3.4.3.1 Harmonic Emission Levels Based on MIL-STD-461

One source

of

informat ion re garding t ran sm it ter spurious output

lev-

els

is the

specification

or

s tan dar ds associated with

the

par t i cu lar

CE

equipment. Transmitter specifications impose

a

limit on spurious out-

puts , and for

system design

it may be

desirable

to use

these levels .

If

this approach

is

used,

the

resul t ing t ran sm it te r harmonic am pli tude

levels would

be

obtained

by

set t ing

A to

zero,

and B to the

specifica-



40


EMC

t ion limit.

Thus , for example , if

t r a ns m i t t e r ha rmon i c a mpl i t ude

models were based

on MIL-STD-461, the cons tan t s for the model

would be :

A = 0

B = as indicated

in Table 3.2.

Table 3.2 Values for B Based on MIL-STD-461

Transmitter Pow er

in dBm

20

50

70

100

B in dB above

Transmitter Pow er

38

- 8 0

-100

-118

3.4.3.2

Summary

of Harmonic Amplitude Levels

In order to provide transmitter harmonic amplitude levels that may

be used in the absence of specific measured data , summaries have

been derived from available spectrum signature data. The results

obtained by summarizing data for approximately 100 different trans-

mitter nomenclatures are presented in Table 3.3. The specific values

of A and B th at correspond to the harmonic emission levels in

Table 3.3 are -70 dB/decade and -30 dB, respectively. The resulting

representation for the harmonic emission level is:

P

T

(f

NT

) dBm = P

T

(f

0T

) dBm - 70 log N - 30 (3.19)

Table 3.3 Harm onic Average Emission Levels

Harmonic 2 3 4

Average emission level -51 -64 -72

(dB above fundamental)

5

- 7 9

6

- 8 5

7

- 9 0

8

- 9 4

9

- 9 7

10

-100

3.5 Receiver Susceptibility Characteristics

Receivers are designed to respond to certain types of electromagnetic

signals within a predetermined frequency band. However, receivers

also respond to undesired signals having various modulation and fre-

quency characteristics. Thus, it is necessary to treat a receiver as

potentially susceptible to all transmitter emissions.



RECEIVER SUSCEPTIBILITY CHARACTERISTICS

41

There

are a

number of interference effects that

an

undesired signal

can produce in a receiver. In order to represent receiver composite sus-

ceptibility, it is necessary to consider these effects and to determine

which effect(s) dominate within

a

given range of frequencies.

Figure 3.6 is a functional diagram useful in discussing various

receiver EMI effects. A superheterodyne receiver generally employs

radio-frequency (RF) stages that provide frequency selectivity or ampli-

fication and one or more mixers that trans late the RF signal to interme-

diate frequencies (IF).

It

also contains

IF

stages that provide further

frequency selectivity

and

amplification,

a

detector that recovers

the

modulation, and post-detection stages that process the signal and drive

one

or

more output displays. Since tuned-radio-frequency (TRF)

and

crystal-video receivers do not use the superheterodyne principle, they

do not contain mixers and IF amplifiers.

In specifying receiver susceptibility, it is necessary to consider the

effects

of an

interfering signal

on

each

of

these stages. The resulting

susceptibility function, which

is

illustrated

in

Fig.

3.7,

represents

a

composite of the most significant effects.

3.5.1

Co-channel Interference

Co-channel interfering signals are amplified, processed, and detected in

the same manner as the desired signal. Thus, the receiver is particu-

larly vulnerable to these em issions. Co-channel EMI may either desen-

sitize

the

receiver

or

override

or

mask

the

desired signal.

It

may also

combine with the desired signal to cause serious distortion in the

detected output

or

cause

the

automatic frequency control circuitry

to

retune to the frequency of the interference, if this is applicable.

For co-channel signals, the receiver susceptibility threshold may

be

represented by the receiver (or environment) noise (i.e., signals that are

below the noise can be considered to be non-interfering). The receiver

noise level is directly related to the receiver sensitivity, which may be

obtained from nominal data on the receiver.

lstLO

1

Filters

RF

Amplifier

>

1st

Mixer

Filter

-> Amp.

2ndLO

1

2nd

Mixer

Filter

•> A m p

v

V

1st IF

F 1st IF 2nd IF

Fig ure 3.6 Representation for superheterodyne receiver.



42


EMC

A4jacent-S|gnal Region

i-Out of Band

Frequency (Log Scale)

F i g u r e 3.7

Rece ive r suscep t ib i l i ty cha rac te r i s t ic s .

3.5.2 Receiver Adjacent-Signal Interference

Adjacent-signal interference

can

produce anyone

of

several effects

in a

receiver.

The

interference

may be

translated through

the

receiver

together with the desired signal and both appear at the input to an IF

stage. In this case, the IF selectivity and the adjacent-signal emission

spectrum will both influence

the

relative level

of

the interfering signal

appearing at the input to the detector. Alternatively, one or more inter-

fering emissions may produce nonlinear effects such as desensitization,

cross modulation, or intermodulation in the RF amplifier or mixer.

Desensitization

is a reduction in the receiver gain to the desired sig-

nal as

a

result of an interfering emission producing automatic-gain con-

trol

(AGC)

action

or

causing

one or

more stages

of the

receiver

to

operate nonlinearly due to saturation.

Cross modulation is the transfer of the modulation from an undes-

ired emission to the desired signal as a result of the former causing one

or more stages of the receiver to operate nonlinearly.

Intermodulation is the generation of undesired signals from the non-

linear combination of two or more input signals that produce frequen-

cies existing

at the sum or

difference

of

the input frequencies

or

their

harmonics.

Although desensitization

and

cross modulation effects

can

occur

in

receivers, recent improvements in receiver design have significantly

reduced EMI problems due to these effects. In many cases, transm itter

noise and transmitter or receiver intermodulation are the limiting fac-



RECEIVER SUSCEPTIBILITY CHARACTERISTICS

43

tors in adjacent-signal operation. Because intermodulation is often the

most serious receiver nonlinear adjacent-signal effect, only this effect

will be discussed in th is section.

3.5.2.1 Receiver Se lectivity

The receiver selectivity determines the amount of attenuation or rejec-

tion provided to off-tuned signals by the receiver. In general, the

receiver susceptibility threshold for off-tuned signals is increased by

the receiver selectivity for the frequency separation in question. The

receiver IF selectivity, S(Af), may be expressed by a piecewise linear

function of th e logarithm of th e ma gnitu de of the frequency sep aratio n,

Af.

S(Af) = S(Afi) + ^ log(Af/Afi) (3.20)

for, Afi<Af<Af

i+ 1

where,

Sj = slope of selectivity curve for app licable reg ion

Afj = m ag nit ud e of ini tial frequency sep ara tio n of applicab le

region

Af= I fi-f()Rl

The representation can be used by specifying the frequency devia-

tions associated with the 3 dB and 60 dB selectivity levels. The result-

ing selectivity characteristics are shown in Fig. 3.8. Notice that a

BR3 =

3 dB Bandwidt

B

R20

=

2 0 d B

Bandwi

B

R 6 0

=

60 dB Bandwi

Passband

h

dth

dth

IP

/

/

/

0.5B

R 20

0.5B

R 6 0

Frequency Separation (Log Scale)

Fig ure 3.8 Receiver selectivity representation.




maximum value of 100 dB is assumed for receiver selectivity. This

implies that any emission source greater than 100 dB above receiver

sensitivity may penetra te and become a source of EMI.

One good indicator

for the

selectivity characteristics

is

given

by the

shape factor: the ratio of the 60 dB bandwidth to the 3 dB bandwidth.

Applying the shape factor (SF) concept to Eq. (3.20), the IF selectivity

relation yields:

log(Af/Af

1

)

S(Af) dB = 60 *

l

dB (3.21)

log(SF)

When Af/Af

x

is chosen to equal the sha pe factor, Eq. (3.21) yie lds 60 dB.

A typical value for receiver shape factor is four. When this value is

substituted into Eq. (3.21), the selectivity parameters can be deter-

mined. The resulting values are summ arized in Table 3.4. The receiver

susceptibility to narrowband off-tune signals is given by:

P

R

(f

0R

±

Af)

dBm =

P

R

(f

0

R)

dBm + S(Af) dB

(3.22)

Table 3.4

Summary of Receiver Selectivity Parameters

Constants for IF Selectivity Model

S fi)dB SidB/decade

0

0

2

0.1 BR

0.5 B

R

5 B

R

0

0

100

0

100

0

3.5.2 .2 Re ceiver Interm odu lat ion

For two signals to produce an intermodulation product that will cause

interference in a receiver, the two signals m ust mix in the RF amplifiers

and

the

first mixer

and

produce

an

intermodulation product that

is at

or near

the

receiver tuned frequency.

The

resulting intermodulation

product tha t

is at or

near

to the

receiver tune d frequency will

be

ampli-

fied, conv erted

to the

interm ediate frequency

and

detected

in the

same

manner as the desired signal. The frequencies of signals tha t are capa-

ble

of

producing intermodu lation interference

in a

receiver must satisfy

the following relationship:

mfx ± nf

2

- f

0

R < B

R

/2 (3.23)



RECEIVER SUSCEPTIBILITY CHARACTERISTICS 45

where,

fi and f

2

=frequencies of two interfering emissions

foR ^receiver tuned frequency

BR =IF bandwidth in which intermodulation products are sig-

nificant

m and n =integers

The only signals that

are

potentially serious sources

of

intermodula-

tion are those tha t are in the vicinity of th e rece iver frequency and pro-

duce intermodulation products that fall within the receiver oper ating or

immediately adjacent channels.

The following equations present the frequency criter ia th at two

interfering signals must meet to satisfy these constraints.

%

± fp -

foR

^

BR/2 (second order)

2f

N

- f

F

- f

0

R

< BR/2 (third order)

3f

N

- 2f

F

- f

0R

< BR/2 (fifth order)

4f

N

- 3f

F

- f

0

R < BR/2 (seventh order)

where,

foR = receiver RF tuned frequency

f = frequency of interfering emission ne are st to foR

f

F

=frequency of interfe ring em ission far the st from foR

Equation

3.23 may be

normalized

to the

receiver fundamental

fre-

quency and solved to show the relation ship betw een two culprit signals

th at will produce an intermod ulation product at the receiver fundamen-

tal frequency:

3.24

To evaluate the impact of receiver intermodulation, it is convenient

to express the effect in terms of an equivalent interference margin. This

corresponds to the margin resulting from two interfering signals that

produce an intermodulation product that falls within the receiver over-

all 3 dB passband. If the intermodulation product is off-tuned from the




r ece iver tuned f requency, the IF se lec t iv i ty can be appl i ed to de termine

th e r es ul t ing off -tune in ter fere nce m arg in . I f th e in pu t s igna l s produ c-

i ng t he i n t e r m od u l a t i o n do no t p r oduce ha r d s a t u r a t i on i n t he r ece i ve r

f r on t end , and t he des i r ed s i gna l and t he r e s u l t i ng i n t e r modu l a t i on do

not exceed the r ece iver au tomat ic ga in cont ro l (AGC) threshold , the

equ i va l en t i n t e r f e r ence m ar g i n (IM) r e s u l t i ng fr om i n t e r m od u l a t i o n i s :

IM (dB) = m P

N

+ n P

F

+ IMF - P

R

(f

0

R) (3.25)

w h e r e ,

P N a n d P

F

= power in dBm at r ece iver input r esu l t ing f rom in ter -

fer ing s ig na ls at f requ enc ies f an d f

F

m and n = cons t an t s a s s oc i a t ed w i t h i n t e r modu l a t i on o r de r ( m

cor r es pond s to t he ha r m oni c o f t he n ea r s i gna l , an d n

cor r es ponds t o t he ha r m oni c o f t he f a r s i gna l t h a t a r e

mi x i ng t o p r oduce t he i n t e r m odu l a t i on p r oduc t )

I M F = i n t e r m od u l a t i o n facto r, wh i ch depe nds on r ece i ve r non l i nea r -

i ty and RF se lec t iv i ty

From a n EMI s tand poin t , t h i rd-order in term odu la t ion i s usua l ly the

most ser ious of fender . For this case, the equivalent inter ference margin i s

IM (dB) = 2P

N

+ P

F

+ IM F - P

R

( f

0 R

) (3.26)

To use Eq . (3.26) in an EM C ass ess m en t , i t i s ne ces sary to de term ine

t he va l u e of t he i n t e r m od u l a t i o n f ac to r ( I MF) . If m ea s u r ed da t a on

r ece i ve r i n t e r modu l a t i on cha r ac t e r i s t i c s a r e ava i l ab l e , t hes e da t a may

be us ed t o eva l ua t e I MF. If m ea s u r e d d a t a a r e no t ava i l ab l e , I M F m ay

be evalua ted f rom MIL-STD-461 l imi t s to provide defaul t va lues as

descr ibed in the next sec t ion .

3 .5 .2 .3 I n t e r m o d u l a t i o n L e v e l s f r o m M I L - S T D -4 6 1

I n t e r m o d u l a t i o n m e a s u r e m e n t s m a d e i n a c c o r d a n c e w i t h M I L - S T D - 4 6 1

ar e pe r f o r med i n a manner s uch t ha t t he t wo i n t e r f e r i ng s i gna l s a r e

e q u a l i n a m p l i t u d e a n d t h e r e s u l t i n g i n t e r m o d u l a t i o n p r o d u c t p r o d u c e s

a s t andard response in the r ece iver . The MIL-STD-461 l imi t s (CS03) for

conduc t ed s us cep t i b i li t y t o i n t e r m od u l a t i o n i n t e r f e r ence s pecif y t h a t no

i n t e r modu l a t i on r e s pons es s ha l l be obs e r ved when t he i n t e r f e r i ng s i g -

na l s a r e 66 dB above t he on - t une l eve l r equ i r ed t o p r oduce a r e s pons e .

The r e s u l t i ng MI L- STD- 461 de f au l t l eve l f o r t h i r d - o r de r i n t e r modu l a -

t ion in ter ference i s :

IM (dB) = 2P

N

+ P

F

- P

F

- 3 P

R

( f

0 R

) - 19 8 (3.27)



RECEIVER SUSCEPTIBILITY CHARACTERISTICS 47

3.5.3 Receiver Spurious Responses

Strong-out-of-band interference may produce spurious responses in a

receiver. The superheterodyne receiver is most susceptible to those out-

of-band signals that mix with local oscillator harmonics to produce a

signal at the IF. Spurious responses in such a receiver usually occur at

specific frequencies, and other out-of-band frequencies are attenuated

by the receiver IF selectivity. For a tuned-RF or crystal-video receiver,

the receiver will be susceptible to those out-of-band interfering signals

th at are not adequately rejected by the RF selectivity.

There are several means by which an out-of-band emission can be

translated to one of the passband frequencies of a superheterodyne

receiver. The most significant of these occurs in the first-mixer stage.

Here, the desired signal is heterodyned with the local oscillator (LO) to

trans late the incoming signal to the interm ediate frequency. In addition

to desired signals, interfering emissions at many different frequencies

are capable of being heterodyned with the LO or other signals and

translated to the receiver IF. The amplitude of responses produced in

this manner is directly proportional to the strength of the original sig-

nals.

The level of the

LO

is typically on the order of

120

dB greater th an

desired, and interfering signals present

at the

input

to the

first mixer

stage. Therefore, heterodyne products that involve the LO are much

larger

in

amplitude than those heterodyne products that do not involve

the LO. Thus, superheterodyne receivers are most susceptible to out-of-

band signals that heterodyne with the LO to produce a product in or

near the IF passband.

In this section, the term spurious response, when applied to super-

heterodyne receivers, refers specifically

to

those undesired responses

that result from

the

mixing

of a

LO

and an

undesired emission.

The

input interfering frequencies that are capable of appearing at the IF as

a result of mixing with the LO are known as

spurious response

frequen-

cies. The am ount of power necessary to cause interference at any partic-

ular spurious-response frequency is a function of receiver susceptibility

to the response.

The frequencies for which spurious responses will occur are given by

the following expression:

f

_ P

f

LO

± f

IF /o om

f

sR —

y

d

Zb

>

where,

f

SR

= spurious response frequencies



48 COMM UNICATION SYSTEMS EMC

p and q = integers associated with the local oscillator and interfer-

ence signal

f

L0

= local oscillator frequency

fjp = intermediate frequency

Figure

3.7

(page 42) illustrates receiver spurious response suscepti-

bility. In general, the most significant responses are those for which q is

equal to 1. Higher values of

q

do not need to be considered for most EMI

situations. Receiver spurious response susceptibility for a given value

of q is given by:

PRGSR)

=

PR

(

f

OR) +

I

log P +

J

(3.29)

where,

I = slope of a spurious response susceptibility in dB/decade

J =intercept in dB relative to fundamental susceptibility

3.5.3.1 Receiver Response Levels Based on MIL-STD-461

If specific information

on the

spurious response characteristics

of a

receiver

are not

available,

it

may

be

desirable

to use

"default values"

that are based on the spurious response limits specified in CS04 of MIL-

STD-461. These limits may be used to solve for I and J of Eq. (3.29) for

the various regions of interest, and the resulting default values are pro-

vided

in

Table 3.5.

3.5.3.2 Summary

of

Receiver Spurious Response Levels

When specific receiver measured data are not available, one alternative

for obtaining

an

out-of-band susceptibility characteristic

is to

derive

summaries from data on receivers. Summaries have been evaluated

from available spectrum signature data, and the specific values for I

and

J

are 35 dB/decade and 75 dB, respectively. The corresponding spu-

rious response representation is

PR *SR) = PR *OR) + 35 log

P

+ 75 (3.30)

Table

3.6

presents

the

average spurious response susceptibility levels

obtained from measured data.

3.6 Antenna Radiation Characteristics

Antennas are designed to radiate and/or receive signals over a specific

solid angle and within a specified frequency range. For land mobile or



ANTENNA RADIATION CHARACTERISTICS 49

Table 3.5

MIL-STD-461 Levels for Spurious R esponses

• For interfering signals within t he receiver 80 dB ban dw idth (i.e., f

0

- BW/2 < f

SR

< f

0

+ BW /2):

P (f

S R

) dBm = P

R

( f

0 R

) dBm

+

i | ( f

S R

- f

o

)

• For interfering signals outside the receiver 80 dB bandwidth

but

within

the

overall

tun ing rang e of the receiver (i.e., f

L

< f

SR

< f

0

- BW/2 or f

0

+ BW/2 < f

SR

< f

H

):

p

R(

f

SR> dBm - P

R

(f

0R

) dBm + 80 dB

• For interfering signals outside the tun ing rang e of the receiver (i.e., for f

SR

< f

L

or

where,

f

0

= receiver tuned frequency

BW = receiver 80 dB bandwidth

P

R

= receiver sensitivity

fL = lowest tuned frequency of receiver

fjj = highest tuned frequency of receiver

Table 3.6 Sum ma ry of Average Spu rious Response Levels

Susceptibility q

=

1)

Local oscillator harmonic (p) 1 2 3 4 5 6 7 8 9 1 0

(image)

Average susceptibility level (dB 75 82 92 96 99 102 105 107 108 110

above fundamental sensitivity)

broadcast applications, the antenna is usually designed to radiate or

receive uniformly over all sectors surrounding the antenna. Other sys-

tems (such as fixed point-to-point communication, radar, and certain

telemetry systems) are designed to confine the functional radiated or

received signals to certain limited sectors.

In practice, however,

it

is not possible to accomplish perfect discrimi-

nation with antennas

in

either the spatial

or

frequency domain. Thus,

antennas that are intended to restrict radiation to specific regions also

radiate into or receive signals from other unintentional regions. In

addition, undesired signals

at

nondesign frequencies

are

inadvertently

radiated

or

received by antennas, and

the

spatial characteristics

of an

antenna for spurious frequencies are significantly different from char-

acteristics at the design frequency.



50


EMC

3.6.1 Design Frequency and Polarization

For the design frequency and polarization, the radiation characteristics

of an antenna may be considered

to

consist of intentional and uninten-

tional radiation regions.

In the

intentional radiation region

(for the

design frequency and polarization) the important EMC characteristics

are the main beam gain and beamwidth. These are nominal character-

istics of an antenna and they may be obtained from the manufacturers

specifications. Antennas

are

often categorized according

to

these gains

(G) as follows:

• Low-gain: G < 10 dB

• Medium-gain: 10 dB < G < 25 dB

• High-gain: G > 25 dB

For the unintentional radiation region, typical mean gain levels rela-

tive to an isotrope would be:

•

-3

dB for low-gain antenn as

• -10

for

medium- and high-gain antennas

Gain levels at a specific orientation may exhibit large variations from

these levels.

3.6.2 Polarization Dependence

If an antenna is linearly polarized, there will be a significant difference

between antenna gain,

in the

intentional radiation region,

for

vertical

and horizontal polarizations. This effect will be most pronounced at the

design frequency, and the gain will be

a

maximum

for

the predominant

mode

of

polarization.

In

general,

the

discrimination afforded

by

using

antennas that are orthogonally polarized will be on the order of 16 dB

to

20

dB,

and

this provides

one

means

of

reducing

the

probability

of

interference between different users (e.g., land mobile applications typ-

ically use vertical polarization, whereas television broadcast utilizes

horizontal polarization).

3.6.3 Nondesign Frequencies

For nondesign frequencies,

the

antenna gain

in the

intentional radia-

tion region would typically be reduced by the following:

• 13 dB for high-gain antennas

• 10 dB for medium-gain antennas

• 0 dB for low-gain antennas



PROPAGATION EFFECTS

51

The antenna gain at specific nondesign frequencies may exhibit vari-

ations from the values specified above. The overall cha racteristics of the

unintentional radiat ion region are not significantly affected by fre-

quency.

3.7

Propagation Effects

In discussing concepts regarding propagation, it is helpful to begin with

a discussion of free-space propagation between lossless isotropic anten-

nas. Once the principles governing propagation under these conditions

are understood, it is easier to follow the concept of propagation between

either omnidirectional or directional antennas in the presence of earth

and reflecting and scattering objects such as buildings, trees, etc.

Because many EMI situations involve transmitters and receivers

that are co-located or located in close proximity, free-space propagation

conditions are often assumed for the purpose of performing an EMC

assessment. If a transmitted signal is radiated from an isotropic

antenna in free space, the signal spreads uniformly in all directions.

Thus, at a distance, d, from the source, the power density is:

P

D

= P

T

/47id

2

(3.31)

where,

Pj) = power density (i.e., power per unit area)

PT

= transmitter power

d = distance from antenna to observation point

The power available at the terminals of a lossless receiving antenna

having an area, A

R

, and a gain, G, is:

R D R (3.32)

= P

T

^

2

/(47id)

2

for G = 1 (isotropic)

where,

X = wavelength in same units as d

The above relation can be expressed in terms of frequency in megahertz

(f

MHz

) and distance in statute miles (d

m

i) or kilometers (d

km

) by substi-

tuting for A,:



52

COMMU NICATION SYSTEMS

EM C

A,(km) =

O 3 k m / S

(3.33)

f

MHz

«, -i v 9 8 4

or, A,(miles) =

5280

f

MHz

1 9 y 7

P R (Q3)

2

M

Th en , ' (3 .34)

= 1.75 x l O

3

f

2

d

2

MHz km

P T

=

(4TC)

2

(5280)

2

f

or,

P R

( 984)

2

= 4560 f

2

d

2

MHz

2

k m

2

Therefore, free-space attenuation in dB between lossless isotropic

antennas for far-field conditions is:

L(f, d) = 10 lo

gl0

(P

T

/PR) (3.35)

= 32 + 201ogf

M H

z + 201ogd

k m

= 37 + 201ogf

M

Hz + 201ogd

m i

3.8 Sam ple EMC A ssessm en t

There are many analysis problems for which only a few transmitter-

receiver pairs need to be considered, and the prediction is either per-

formed manually or with the aid of a small computer program that may

be run on a personal computer or a time-share term inal. This section pre-

sents a step-by-step process for performing a manual EMC analysis

throu gh the u se of a special form. Although th e partic ular form presen ted

in this section was designed for analyzing AM and FM analog voice com-

munications systems such as those used for land mobile applications,

similar forms may be used for other types of communications systems.

3.8.1 Tr ansm itter N oise

Consider the case of a land mobile receiver operating at 150 MHz.

Determine whether the EMI will result if a land mobile transmitter,

op erat ing at 150.1 MH z, is located 122 m (400 ft) from th e receiver. The

pertinent transmitter and receiver characteristics are:



SAMPLE EM C ASSESSMENT

53

Transm itter power, P^ = 50 dBm

Transmitter antenna gain, G-p = 3 dB

Receiver antenna gain, GR = 3 dB

Receiver sensitivity = -107 dBm

Allowable degradation

=

0 dB

This is clearly a co-site, adjacent-signal situation, and the primary

cause of potential interference would be transmitter noise. The com-

pleted short form is provided for this example in Fig. 3.9. The results

indicate that a 9 dB interference margin will be obtained, and a mar-

ginal interference situation exists.

Adjacent Singnal Interference*

Transmitter Noise

1.

Transmitter Power, P

T

(dBm/Channel)

2. Noise Constant

3.

201ogAf

TO

(kHz)

4. Noise pe r Cha nnel (dBm /Channel) (1) - (2) - (3)

5. Transmitter Antenna Gain, GTR (dB)

6. Effective Radiated Noise Pow er

(dBm/Channel);

(4) + (5)

7. Propagation Constant

8 . 2 0 1 o g d

m

( k m )

9.

20 log f

R

(MHz)

10. Propagation Loss, L (dB): (7) + (8) + (9)

11.

Receive r Ante nna Gain, G

RT

(dB)

12. Noise Power Available, P

A

(dBm); (6) - (10) + (11)

13. Receiver Sensitivity Level (dBm)

14. Allowable Degr adation of Receiver Sensitivity (dB)

15. Receiver Susceptibility

Level,

P

R

(dBm ); (13) + (14)

16. Interferen ce Margin

(dB);

(12) - (15)

Third Order Intermodulation

Frequency Check

• Select Receiver to Analysis

17.

Receiver/Frequ ency, f

R

(MHz)

• Select Cosite Transm itter, T

x

, with

Frequency Nearest to f

B

18.

Transmitter Frequency, f

T

(MHz)

19. Frequency Separation

AF

TO

(MHz);

(18) - (17)

20. Frequency,

f

Tt>)

for Interm odulation; (18) + (19)

21 . Chan nel Width, (MHz)

22. Band for Intermodulation; (20) ±(2 1) \Q_^

50

56

40

- 4 6

3

- 4 3

32

- 1 8

44

58

3

- 9 8

-1 0 7

0

- 1 0 7

9

( + )

• Check Other Cosite Transmitters for Frequency

within Band Specified by (22). If one is found,

continue with

analysis.

If

none,

eliminate selected

transmitter from consideration and repeat proce ss

with another transmitter.

Interfere nce Margin < 0.10

dB ,

EMI Highly Imp robab le.

10 dB < Interferen ce Margin < 10

dB, EMI Marginal

Interfer ence Margin > 10 dB, EMI Prob able.

* Applies to co-site transmitters and receivers with frequency sep arations (Af)

less tha n 10% of operat ing frequency.

Figure 3.9 EMC ana lysis form for analo g voice system s tra ns m itte r noise.



54 COMMUN ICATION SYSTEMS EMC

3.8.2 Intermodulation

Because of nonlinearities in the preamplifier of a receiver or the final

power amplifier in a transm itter, two or more interfering signals may

mix (i.e., intermodulate) to produce new signals at other frequencies.

If the new frequencies are close to the tuned frequency of the receiver,

the signals may be amplified and detected by the same mechanism as

the desired signal. Thus possible degradation of performance may

result.

In order to analyze intermodulation, it is necessary to identify pairs

of transmitters within the electromagnetic environment that can

intermodulate and cause EMI in a receiver. Next, it is necessary to

determine the interference margin that results from intermodulation

occurring in each of the transmitters and the receiver. The only sig-

nals that are considered serious sources of intermodulation interfer-

ence are those that are in the vicinity of the receiver frequency and

produce intermodulation products that fall within the receiver 60 dB

bandwidth.

Consider the case of a land mobile receiver operating at 450 MHz in

the vicinity (12 m, or 40 ft) of a land mobile transmitter at 451 MHz.

Determine whether an intermodulation problem will result if a second

transmitter operating at 452 MHz is located 30.5 m (100 ft) from the

receiver on a site that is 24.5 m (80 ft) from the first transmitter. The

pertinent transmitter and receiver characteristics are:

Transmitter power, P

1

and P

2

= 50 dBm

Transm itter an tenna gain, G^i and G^2 = 3 dB

Receiver antenna gain, G

R

= 3 dB

Receiver sensitivity = -10 7 dBm

Allowable degradation = 0 dB

Channel width = 50 kHz

This situation could result in either transmitter or receiver third-order

intermodulation. To determine whether third-order intermodulation is

possible, it is first necessary to perform the frequency check indicated

on the short form (Fig. 3.10). This has been checked, and the results

indicate that an intermodulation problem may occur.

Next, it is necessary to calculate the interference margin resulting

from both transmitter and receiver intermodulation situations to

determine the corresponding interference potential. These calcula-



SAMPLE

EMC

ASSESSMENT

55

tions, which are straightforward, have been performed on the appro-

priate forms (Fig. 3.11). The calculations indicate that the receiver

intermodulation results in a +33 dB interference margin, and trans-

mitter

intermodulation results in a +59 dB interference margin. For

this situation, transmitter intermodulation will predominate, and

EMI is probable.

Adjacent Signal

Interference*

Transmitter Noise

1.


T

(dBm Channel)

2.

Noise Constant

3.

20 log Af^ (kHz)

4. Noise per Channel (dBm Channel) (1) ~ (2)

-

(3)

5. Transmitter Antenna Gain, G

TR

(dB)

6. Effective Radiated Noise Power (dBm Channel); (4) + (5)

7.

Propagation Constant

8. 20^ ^ 1011)

9. 20 log f

R

(MHz)

10. Propagation Loss, L (dB): (7) + (8) + (9)

11.

Receiver Antenna Gain, G ^ (dB)

12.

Noise Power Available, P

A

(dBm); (6) - (10) + (11)

13.

Receiver Sensitivity Level (dBm)

14.

Allowable Degradation of Receiver Sensitivity (dB)

15.

Receiver Susceptibility Level, P

R

(dBm); (13) + (14)

16. Interference Margin (dB); (12) - (15)

Third Order Intermodulation

Frequency Check

• Select Receiver to Analysis

17.

Receiver Frequency, f

R

(MHz)

• Select Cosite Transmitter, Tj, with

Frequency Nearest to f

R

18.

Transmitter Frequency,

f

T

(MHz)

19.

Frequency Separation AF

TO

(MHz); (18 )-( 17)

20. Frequency,

F

Ts?

,

for Intermodulation; (18) + (19)

21 .

Channel width, (MHz)

22 .

Band for Intermodulation; (20) ± (21)

• Check Other Cosite Transmitters for Frequency

within Band Specified by (22). If one is found,

continue with analysis. If none, eliminate selected

transmitter from consideration and repeat process

with another transmitter.

Interference Margin < 0,10

d B,

EMI Highly Improbable

10 dB < Interference Margin < 10

d B,

EMI Marginal

Interference Margin > 10 dB, EMI Probable.

* Applies to co-site transmitters and receivers with frequency separations (Af)

less th an 10% of operating frequency.

(-) 451.95

450

451

452

.050

(+) 452.05

Figure 3.10 EMC analysis form for analog voice systems intermod ulation.



56


Adjacent Signal Interference*

Receiver Intermodulation

23. Transmitter Power, P

T

(dBm)

24.

Transm itter Antenna Gain,

GTR

(dB)

25. Effective Radiated Pow er (dBm ) (23) + (24)

26.


27. 20 log d ^ (km)

28. 20 log f

T

(MHz)

29. Propagation Loss (dB); (26) + (27) + (28)

30. Receiver Antenna Gain, (dB)

31. Pow er Available at Receiver, (dBm ); (25) - (29) + (30)

32. Multiply T

x

Pow er Available, Line (31), by Two

33, T

2

Pow er Available, Line (31)

34. Intermodulation Con stant

35.

Freq uency S eparatio n, Af (%) [(19) + (17)) x 100

36. 601ogAf(%)orO

37. Equivalent Intermodulation Pow er (dBm);

(32)+ (33)+ (34)-(36)

38.

Receive r Susceptibility Level, P

R

(dBm)

39. Interfe renc e Margin, (dB); (37) - (38)

Transmitter Intermodulation

40. Pow er of T

2

(dBm)

41. T

2

Antenna Gain (dB)

42.

T

2

Effective Radiated P ower (dBm), (40) + (41)


44. 2 0 i o g d

T i T 2

( k m )

45.

201ogf

Ti?

(MHz)

46. Propagation Loss L (dB); (43) + (44) + (45)

47. Tj Antenna Gain (dB)

48. T

2

Signal at T

x

(dBm ); (42) - (46) + (47)

49.

Intermodulation Constant

50. 30 log A f (%), (line 35), or 0; Whichever is Larger

51. Intermodulation Power at T

x

(dBm ); (48) - (49) + (50)

52.

T

t

Antenna G ain(dB )

53. Intermodulation E RP (dBm); (51) + (52)

54. Propagation Constant (dB)

55.

20 log dp.

R

(km)

56. 20 log f

H

(MHz)

57. Intermodulation Propaga tion Loss (dB); (54) + (55) + (56)

58. Receiver Antenna Gain (dB)

59. Intermodulation P owe r at Receiver (dBm); (53) - (57) + (58)

60.

Receiver Susceptibility Level (dBm)

61. Interference Margin (dB)

50

3

53

32

-3 8

53

47

3

9

18

50

3

53

32

-30

53

55

3

1

1

-9 3

0.22

0

-7 4

-107

+33

50

3

53

32

-3 2

53

53

3

3

10

0

_7

3

-4

32

-3 8

53

47

3

-4 8

-107

59

Interference Margin < .10

dB,

EMI Highly Improbable.

-10 dB < Interference Margin < 10 dB, EMI M arginal

Interference Margin > 10 dB, EMI Probable.

* Applies to co-site transmitters an d receivers with frequency s eparatio ns (Af)

less tha n 10% of ope rating frequency.

Figure 3.11 EMC ana lysis form for analog voice system s interm odu lation .



SAMPLE EMC ASSESSMENT 57

3.8.3 Ou t-of-Band EMI

Consider that an industrial user desires to operate a land mobile base

receiver a t 158.1 MHz. The receiving antenna will be located on top of

a building, and a survey of the immediate vicinity reveals th at there is

a public safety transmitter operating at 39.525 MHz and a land trans-

portation transmitter operating at 452.9 MHz. The separations

between the industrial receiver and the public safety and land trans-

portation transmitter are 100 and 20 ft (30.5 and 6.1 m), respectively.

Determine whether an EMI problem exists if the system characteris-

tics are as follows:

Industrial Receiver

Frequency =158.1 MHz

Intermediate frequency

=

10.7 MHz

Local oscillator = 147.4 MHz

Fundamental sensitivity

=

-10 7 dBm

Antenna gain = 3 dB

Public Safety Transmitter

Frequency = 39.525 MHz

Power output = 50 dBm

Antenna gain = 0 dB

Land Transportation Transmitter

Frequency

=

452.9 MHz

Power O utput = 47 dBm

Antenna gain = 6 dB

These two potential interference situations are clearly examples of

out-of-band EMI. The most probable causes of interference for these sit-

uations would be a harmonic of the public safety transmitter interfer-

ing with the industrial receiver fundamental, and a spurious response

of the industrial receiver being interfered with by the fundamental of

the land transportation transmitter. The calculations have been per-

formed on the accompanying forms (Figs. 3.12 and 3.13). The results

indicate that both of these transmitters pose a potential EMI problem

to the receiver.



58


EMC

Out of Band Interference*

Transmitter Harmonic to Receiver Fundamental;

1.

Receiver Frequency, fjj (MHz)

2.


T

(MHz)

3.

(1) + (2)

and

Round Off to Nearest Integer, N

4. Transmitter Harmonic Frequency, N%

(MHz);

(3) x (2)

5. Frequency Separation,

I (4)

- (1)

I,

(MHz)

6. Receiver Bandwidth

• If (5) > (6) No Harmonic Interference

If (5) < (6) Continue

7.


T

(dBm)

8. Harmonic Correction, (dB); from Table 3.3

9. Harmonic Power

(dBm);

(7)

+

(8)


11.

201ogd

TR

(km)

12.

20 log f

R

(MHz)

13.

Propagation

Loss,

L, (dB) (10) + (11) + (12)

14. Receiver Antenna Gain, Gg (dB)

15.

Power

Available

at Receiver

(dBm);

(9) - (13) + (14)

16. Receiver Susceptibility Level, P

R

(dBm)

17.

Interference Margin, (dB); (15) - (16)

Transmitter Fundamental to Receiver Spurious:

18.

(2) +(1) and Round Off

to

Nearest Integer, P

19.

Local Oscillator Frequency, ^(MHz)

20. Intermediate Frequency, % (MHz)

21.

Pf

L0

± % -

f

T

j

; (IS) x (19) ± (20) - (2) |

If (21 +) or (21 -) > (6) No Spurious Interface

If (21 +) or (21 -) < (6) Continue

22.


T

(dBm)

23.

Transmitter Antenna Gain,

Gj

(dB)

24.


25. 201ogd

TR

(km)

26. 20 log f

T

(MHz)

27. Propagation Loss, L (dB); (24) + (25) + (26)

28.

Power Available at Receiver, (dBm); (22) + (23) - (27)

29. Receiver Fundamental Susceptibility, P

R

(dBm)

30. Spurious Correction, from Table 3.6

31. Spurious Susceptibility, (dBm); (29) + (30)

32.


Interference Margin < -10

dB,

EMI Highly Improbable

-10

dB <

Interference Margin

< 10 dB, EMI

Marginal

Interference Margin > 10

dB,

EMI Probable.

* Applies to cosite transmitters and receivers with frequency

separations (Af) greater than

10%

of operating frequency.

t These entries are also required for transmitter fundamental to receiver

spurious.

158.lt

39.525f

_

158.1

_

0.015+'

50

-72

-22

32

-30

44

46

-65

-107

+42

32

Figure 3.12 EMI from public safety transmitter.



SAMPLE EMC ASSESSMENT

59

Out of Band Interference*

Transmitter Harmonic to Receiver Fundamental; f > f q

1.

Receiver Frequency, % (MHz)

2.


T

(MHz)

3. (1) + (2) and Round Off to Nearest Integer, N

4. Transmitter Harmonic Frequency, N%

(MHz);

(3) x (2)

5. Frequency Separation, I (4) - (1) I, (MHz)

6. Receiver Bandwidth

• If

(5)

> (6) No Harmonic Interference

If (5) < (6) Continue

7. Transmitter Power, P

T

(dBm)

8. Harmonic Correction, (dB); from Table 3.3

9. Harmonic Power

(dBm);

(7) + (8)


11.

20logd

TR

(km)

12.

20 log f

R

(MHz)

13.

Propagation

Loss,

L, (dB) (10) + (11) + (12)

14.

Receiver Antenna Gain, Gg (dB)

15. Power Available at Receiver

(dBm);

(9) - (13) + (14)

16. Receiver Susceptibility Level, P

R

(dBm)

17.

Interference

Margin,

(dB); (15) - (16)

> fit

ransmitter Fundamental to Receiver Spurious:

18.

(2) + (1) and Round Off to Nearest Integer, P

19.

Local Oscillator

Frequency,

f^ (MHz)

20. Intermediate Frequency, fjp (MHz)

21.

IPfLQ ± % - f

T

l; (18) x (19) ± (20) - (2)1

If

(21

+) or

(21

- ) > (6) No Spurious Interface

If (21 +) or (21 - ) < (6) Continue

22.


T

(dBm)

23.

Transmitter Antenna Gain, % (dB)

24.


25. 20 log d

TO

(km)

26. 20 log f

T

(MHz)

27.

Propagation

Loss, L (dB);

(24)

+

(25) + (26)

28. Power Available at Receiver, (dBm); (22) + (23) - (27)

29.

Receiver Fundamental Susceptibility, P& (dBm)

30.

Spurious Correction, from Table 3.6

31. Spurious Susceptibility, (dBm); (29) + (30)

32.


Interference Margin < -10 dB, EMI Highly Improbable

-10

dB <

Interference Margin

< 10 dB, EMI

Marginal

Interference Margin > 10

dB,

EMI Probable.

* Applies to cosite transmitters and receivers with frequency

separations (Af) greater than 10% of operating frequency.

f These entries are also required for transmitter fundamental to receiver

spurious.

158.lt

452.9t

0.015*

32

147.4

10.7

-021.4

47

32

44

53

41

12

107

92

-15

27

Figure 3.13 EMI from land transportation

transmitter.



60 COMMUN ICATION SYSTEMS EMC

3.9 Com puter EMC An alys i s

The previous section presented forms that may be used to perform a

manual EMC assessment.

All of

the operations indicated

on the

forms

may be easily programmed on a computer or calculator to assist the

system designer in evaluating EM C. If one is to be involved in the plan-

ning and design of a large system (e.g., a statewide public safety sys-

tem) it is recommended t ha t a computer be used to assist in the many

calculations that will be required

to

en sure proper EMC design. Also,

it

is suggested that a computer databa se be established on all other users

in the area.

Suggested Readings: Communication Systems EMC

[1] Case, David A., U nde rstan din g

the

Changes

to FCC

Pa rt 15.407

Regulations, ITEM interference technology,

2010 EMC

Test

Design Guide, p. 60.

[2] Radio Noise, ITU-RP.372-8, 2003.

[3] Spauld ing, A. D., an d R. T. Disney, Man-M ade Noise Esti m ate s for

Business, Residential and Rural Areas, NTIA, 1974-38.

[4] An Upd ate of CCIR Bus iness and Residen tial Noise Levels, IE EE

International Symposium on Electrom agnetic Compatibility, 1994,

pp. 348-353.

[5] The Natura l and Man-Made Noise E nvironm ents in Personal

Com mun ications Services Ban ds, NTIA Repo rt 96-330, May 1996.

[6] Acharz, R.

J.,

Y. Lo, P. Pa pa zia n, R. A. Dalke

and

G. Hufford, Man-

Made Noise in the 136-138 MHz VHF Meteorological Sa tellite

Band, NTIA Report 95-355, 1998.

[7]

Acharz, R. J., and A. Dalke, Man-Made Noise Power Measu re-

ments at VH F and UH F Frequencies, NTIA Report 02-390, 2010.

[8] Rantakko, J., F. Lofsved, and M. Alexanderson, M easurem ents of

Man-Made Noise a t

VHF,

EMC E urope W orkshop, 2005.

[9] Classification of Electromagnetic Env ironments, Basic EMC Pub-

lication,

IEC

61000-2.5, Technical Report Pa rt 2— Environm ent,

Section

5.

[10] ANSI C63.10: Procedures for Testing Compliance of a Wide Vari-

ety of Un licensed W ireless Devices, ITEM interference technology,

2009 EMC Test and Design Guide,

p. 8.



Chapter 4

Electronic System D esign for EMC

The basic

EMC

requirement

is to

plan, specify,

and

design devices,

equipments, and systems that can be installed in their operational

environments without creating

or

being susceptible

to

interference.

In

order to satisfy this requirement, careful consideration must be given to

a number of factors th at influence EMC.

In

particular,

it

is necessary to

consider major sources of electromagnetic interference (EMI), modes of

coupling, and points or conditions of susceptibility. The electronic

equipment

or

system designer should

be

familiar with

the

basic tools

(including prediction, analysis, measurement, control, suppression,

specifications, and standards) that are used to achieve EMC.

The first step in the system-level EMC design process is to define the

ambient environment. During this step,

it is

necessary

to

identify cul-

prit EMI sources and victim circuits and specify the EMI emissions

from sources

and the

susceptibility

of

victims. Information about

the

environment EMI sources and victims may be provided by applicable

regulations and stan dards (i.e., EMC, safety, etc.).

The next step in system-level EMC design is to identify major EMI

coupling mechanisms

and

determine

EMI

suppression

and

control

requirements that

are

necessary

to

achieve EMC. Trade-off consider-

ations (i.e., EMI vs. safety, shielding vs. circuit design, etc.) should be

addressed, and the applicable EMI

fixes

should be selected and incorpo-

rated. Measurements should be performed throughout the design and

development process to verify compliance.

4 1 Basic Elements of EMI Problems

Three basic elements are common to all EMI situations. These three

basic elements are a source of EMI, a transfer or coupling medium, and

a susceptible device. Figure 4.1 illustrates the three basic elements of

61



62

ELECTRONIC SYSTEM DESIGN FOR EMC

Coupling Path

Fig ure 4.1 The three basic elements of EMI.

an EMI situ ation . Figu re 4.2 identifies v arious po ssible sources of EM I,

modes of coupling, and potentially susceptible devices. In order to effec-

tively suppress and control EMI problems, it is necessary to develop an

awareness of the role that each of these basic elements plays, assess

potential EMI problems (which requires quantitative information on

EM I levels produced by sources, coupling from source to victim, a nd vic-

tim susceptibility), and understand how to minimize the resulting EMI

impact on potentially susceptible devices.

Conduction and

Radiation Emitting

Sources

Transfer or

Coupling Media

Radiated

Antenna-to-Antenna

Case Radiation

Case Penetration

Field-to-Wire

Wire-to-Field

Wire-to-Wire

Conducted

Common Ground

Impedance

Power Line

Interconnecting Cable

Receiving or

Receptor Elements

Radio T ransmitters

(Broadcast,

Communications,

Navigation, Radars)

Receiver Local

Oscillators

Motors, Sw itches,

Fluorescent Lights,

Diathermy,

Dielectric Heaters,

Arc Welders

Engine Ignition

Computers

&

Peripherals

Natural Sources:

Lightning,

Galactic Noise,

Electrostatic Discharge

F ig u re 4.2 Sources of EM I, modes of coupling, an d poten tially susceptible

devices.

Radio Receivers

Analog Sensors and

Amplifiers

Industrial Control

Systems

Computers

Ammunition and

Ordnance

Human Beings

(Biological H azards)



BASIC ELEMENTS OF EMI PROBLEMS

63

4.1.1 Sources of MI

Any electrical, electromechanical, or electronic device is a potential

source of EMI. In general, EMI sources can be classified either as trans-

mitters (i.e., equipment whose primary function is to intentionally gen-

erate or radiate electromagnetic signals) or incidental sources (i.e.,

equipments that generate electromagnetic energy as an unintended by-

product in the process of performing their primary function). Sources of

EMI may be divided into natural and man-made sources. This hand-

book is concerned with only man-made sources of EMI. Examples are

shown in Figure 4.3.

The energy generated by EMI sources can either be radiated from

the source into the surrounding environment and then picked up by

potentially susceptible devices or conducted from the source into poten-

tially susceptible devices via power leads, signal leads, or any other

interconnecting wires, cables, or other conductors. In general, it is nec-

essary to consider both radiated and conducted emissions from an EMI

source.

Although any source of EMI can produce radiated emissions, radio

transm itters are intentionally designed to generate and radiate electro-

magnetic signals, and they usually represent the most serious threat

from a radiated emission standpoint. Transmitters may cause EMI

problems in equipments that are located within several (or in some

cases many) kilometers of the source. Other equipments can cause EMI

as a result of their radia ted emissions, but they will usually cause prob-

lems only in their immediate vicinity.

Figure 4.4 displays the frequency bands allocated for various radio

and communication services and indicates the maximum effective radi-

ated power allowed for each service. The levels shown in Fig. 4.4 repre-

Man

Made

Sources

of EMI

Communications

Electronics

Electric Power

I

Tools

and

Machines

Ignition Systems

Industrial

Consumer

Broadcast

Relay Comm.

Navigation

Radar

Communications

-

Generation

- Conversion

- Transmission

- Distribution

Power Tools

Appliances

Office Business

Machines

Ind.

Machines

Transporters

- Engines

-Vehicles

-Tools

-

Welders

Heaters

-

Ultrasonic

Cleaners

-

Medical

- Ind.

Controls

Computers

L Lights

Figure 4.3 Sources of EM I.



64


ERP = P ower O utput x Antenna Gain

ITACAN/IFF

Fixed

Microwave

Link

10kHz 100kHz 1MHz 10MHz 100MHz

Radio Frequency

lGHz lOGHz lOOGHz

F ig u re 4.4 Frequency allocations and maxim um effective radia ted power.

s en t t he m ax i m um e ffective r a d i a t ed pow er t h a t w i l l be p r oduce d by t he

f u n d a m e n t a l ( i n te n t io n a l ) o u t p u t f ro m t h e t r a n s m i t t e r s . F u n d a m e n t a l

output s tha t a re r e l a t ive ly h igh power exhib i t a ser ious potent i a l for

caus i ng i n t e r f e r ence p r ob l ems t o equ i pmen t s l oca t ed w i t h i n s eve r a l

k i l om et e r s of t he t r a ns m i t t e r s . F i gu r e 4 .5 i l l u s t r a t e s fie ld s t r en g t h s a s

a funct ion of e ffec tive r ad ia t ed po wer an d d i s t a nc e from t h e sou rce .

Note that low-power sources close to a vict im can produce high f ield

s t r eng t hs . Thus , a l ow- power t r ans mi t t e r c l os e t o a v i c t i m can have t he

s ame po t en t i a l f o r caus i ng EMI as a h i gh - power t r ans mi t t e r t ha t i s f a r -

t he r away f r om t he v i c t i m . Low- power ed t r ans mi t t e r s s hou l d no t be

i gnor ed

Al l e l ec t r i ca l and e lec t ronic equipment can be potent i a l sources of

EMI . In genera l , t he EMI l evel s r ad ia ted f rom e lec t r i ca l or e l ec t ronic

equ i pm en t a r e r e l a ti ve l y low power, an d t he r e f o r e t hes e eq u i p m en t s

us ua l l y pos e an EMI t h r ea t on l y t o communi ca t i ons r ece i ve r s o r s ens i -

t i ve equ i pm en t ope r a t ed i n clos e p r ox i mi t y w i t h t h e s our ce .

F o r e l e c t r i c a l o r e l e c t r o n i c E M I s o u r c e s , o t h e r t h a n t r a n s m i t t e r s ,

s i gn i f i can t em i s s i on s ma y occupy s ev e r a l oc t av es o r m or e of t h e f re -

q u e n c y s p e c t r u m . S o m e of t h e m o r e i m p o r t a n t s o u r c e s i n c l u d e p o w e r

l i n e s ,

au t om ob i l e en g i n e i gn i t i o n s ys t em s , fluorescen t l am ps , e l ec t r i -

c a l m o t o r s , s w i t c h e s , a n d r e l a y s . I n c i d e n t a l r a d i a t i o n m a y c a u s e

E M I i n c o m m u n i c a t i o n s r e c e i v e r s o r o t h e r s e n s i t i v e e q u i p m e n t s o r

s y s t e m s .




65

Use only for Far Field Situations, i.e. R

me

ters ^ 50 /FMHZ)

Transmitter-to-Victim Distance

in

meters

10 30 100 300 1000

3000

10k 30k

100k

0.01

Figure 4.5

source.

0.03 0.1 0.3 1 3 10

Transmitter-to-Victim Distance

in

Kilometers

*ERP

=

Effective Radiated Power

=

Transmitter Power

x

Antenna Gain

Note:

Below VHF, where Non-Directional Antennas

ar e

Used,

Ground Wave

May

Cause

a

3

dB

Increase

in E

field

Field strength

vs.

maximum radiated power and distance from

Equipment-generated EMI can be conducted from a source to a

potentially susceptible device via power leads, signal leads, or any other

interconnecting conductors (e.g., metal structures, racks, equipment

housings, etc.). This conducted EMI can also cause problems in suscep-

tible devices that are connected to an EMI source, either directly or

through a shared common-ground or common-source impedance.

Although any electrical or electronic device can produce conducted

EMI, electrical power systems are often the most serious source of con-

ducted interference. As loads are switched on and off of electrical cir-

cuits,

large transients may be produced, and these transie nts can cause

EMI in systems. Maximum transients in unprotected electrical power

systems may be on the order of ten times the normal line voltage (i.e.,

1200 V transients in a 120 V electrical power system) as shown in

Fig. 4.6. In order to avoid EMI problems in susceptible equipments, it is

necessary to provide transient suppression to control the transients

resulting from surges in the electrical system.

In general, it is difficult to determine the EMI levels generated by

various sources. However, if the equipment was required to conform to

EMI rules, regulations, or standards, the EMI limits imposed by the



66


Total

Outage

Under-Voltage

90

V (or 180)

A

Over-Voltage

Up to Several

kV

Spikes

Induced Radio

Frequency Signals

Fig ure 4.6 Transients in electrical power systems.

rules, regulations, or stan dar ds may be used to provide an uppe r bound

on the em issions th at will be produced by th e device.

4.1.2 EMI Modes of Coupling

Emissions may be coupled by one or more paths from the interference

source to the susceptible receiving device(s). Basically, these paths are

classified as either (1) conducted paths, which include all forms of direct

conductor, wire, or cable coupling, or (2) radiated paths, which involve

near field effects or propagation through the environment.

The most important radiated and conducted EMI coupling paths are

listed in Table 4.1 and are ill us trate d in Fig. 4.7. While not all inclusive,

the se pa th s account for, perha ps, 95 perce nt of all intra -sys tem EMI sit-

uations. The object is to classify each potential EMI situation into one

or more of the coupling paths illustrated.

Table 4.1

Major Conducted and Radiated EMI Coupling Pa ths

• Conducted power or signal cable coupling

• Common-ground impedance common-mode coupling

• Field-to-cable or cable-to-field common-mode coupling

• Field-to-cable or cable-to-field differential-mode coupling

• Case radiation-case penetration

The mode(s) of coupling from an emitter to a receptor can become

very complicated. In general, the coupling paths are extensive and may



BAS IC ELEMENTS OF EMI PROBLEM S

67

Antenna Antenna

Filter

->

Receptor

a ) Antenna-Box-Wire Radiation Coupling Paths

Emitter Filter

Wire Conducted

Filter

Receptor

Common

Source Impedance

Regulation

Impedance

b) Conduction Coupling Paths

Figure 4.7 Illustration of major coupling paths .

not be well defined. Coupling can also result from a combination of

paths , such as conducted from an emitter to a point of radiation, then

picked up by induction and conducted to the v ictim.

Conducted EMI may enter a victim as a result of directly coupled

wiring leads between the receptor and some source of electrical distur-

bances. Typical conducted paths include interconnecting cables, power

leads, control and signal cables, and shared source or ground imped-

ances.



68


The major conducted paths are:

• Power cable coupling

• Sign al cable coupling

• Common-source imped ance coupling

• Comm on-ground impe dance coupling

There are several mechanisms by which conducted emissions can be

coupled into equipments

or

systems

and

produce

an EMI

problem.

First, conducted emissions

on

interconnecting signal, control,

or

power

leads can couple interfering emissions directly into other equ ipm ents

and cause problems. This

is the

most obvious mech anism

for

conducted

emissions to produce EMI. The conducted emissions can be either dif-

ferential mode or common mode. For differential-mode emissions, the

currents

in the two

intercon necting wires (i.e.,

the

signal wire

and the

return) flow in opposite directions as shown in Fig. 4.8. For common-,

mode

emissions,

the

currents

in the two

wires flow

in the

same direc-

tion,

as

i l lustrated

in

Fig.

4.9.

Power Source

DCM1

>

DCM2

Load

F i g u r e

4.8

Differential-mode EMI cu rren t flow.

Power Source

CMC1

CMC2

CMC

Load

Metallic Structure

F i g u r e

4.9

Common-mode EMI curre nt flow.



BASIC

ELEMENTS OF EMI PROBLEMS

69

Second, in situ ation s w here a num ber of different equ ipm ents or cir-

cuits use a common ground, emissions generated by one equipment or

circuit can couple into other equipments or circuits as a result of them

sharing a common ground impedance. This coupling mechanism, illus-

trated in Fig. 4.10, involves a ground loop and is often referred as com-

mon-mode common-ground impedance

coupling. The coupling is

"common" mode because the currents in the two interconnecting wires

will be flowing in a "common" direction. The

common-ground

imped-

ance

term is used because the coupling results from the fact that equip-

ments (or circuits) are sharing the same ground wire, bus, plane, trace,

etc. Examples where this may be a problem would be in installations

where electrical machinery, computers, and sensitive instrumentation

all use the same ground system or in equipments where analog and dig-

ital logic circuits use the sam e ground.

The common-mode voltage (Vj) shown in Fig. 4.10 resulting from

common-ground impedance coupling is equal to the product of the EMI

ground current and the shared common-ground impedance (ZQ). Char-

acteristics of ground impedances as a function of frequency are pro-

vided in Chapter 5 for various types of ground conductors. However, it

is important to recognize that the common-mode voltage, Vj, is not the

direct cause of the problem. Instead, the problem results from the dif-

ferential-mode voltage, V

o

, that appears at the input to the victim as

shown in Fig. 4.11. The ratio of V

0

/Vj is referred to as the

ground-loop

coupling, and it depends on the distribution of impedances in the

ground loop as shown in Fig. 4.12.

Radiated interference includes situations in which emissions enter

via a receiving system antenna, if applicable. Other radiated paths,

shown in Fig. 4.7, include situations in which emissions are coupled

Power

A A A

EM I

EM I

El

Cul

EM I

prit

F ig u re 4.10 Common-mode common-ground impedance coupling.



7


Box# l

Box #2

t y

Metal Ground Plane

F ig u re 4.11 Ground-loop EMI coupling.

B ox#l Box #2

Metal Ground Plane

-

Common Ground

-

Notes: Z = Ground Plane impedance between Points

A

and H.

Vi

=

Voltage Drop x Z, between Points

A

and H.

Ig = External Ambient Current Flowing through Z.

v

o

=

Differential-Mode Voltage Developed from Common-Mode Voltage,

v

i .

F ig u r e 4.12 Conversion of common-mode voltage to differential-mode.

into or out of signal, ground, or power leads or penetrate a shielded

housing at points of leakage and couple into low-level circuitry.

The major radiated paths are:

• Antenna-to-antenna

• Antenna-to-box

• Antenna-to-wire



BASIC ELEMENTS OF EMI PROBLEM S

71

• Box-to-antenna

• Box-to-wire

• Wire-to-antenna

• Wire-to-box

• Wire-to-wire

• Box-to-box

Interconnecting wires can act as antennas and "pick up" and/or "radi-

ate"

EMI.

One such "pickup" or "radiated emission" mode is illustrated in

Fig. 4.13, wh ere th e in terconn ecting wires or cables (or th e circuit

itself) act as an antenna. In this case, if the interconnecting wires or

cables are exposed to an electromagnetic field, a voltage will be induced

in the loop formed by the interconnecting wires or cables (or the cir-

cuit).

This situation, as illustrated in Fig. 4.14, is often referred to as

field-to-cable differential-mode coupling

because the curre nts in the two

wires forming the loop will be flowing in "different" directions. Alter-

nately, differential mode curre nts flowing in th e loop will rad iat e EM I.

This situation is referred to as cable-to-field differential-mode coupling.

A

third "pickup" and/or "radiated emission" mode for radiated fields

is illustra ted in Fig. 4.15. In th is situ ation , the loop formed by the inter-

connecting wires or cables and the ground acts as an antenna and picks

up the radiated field incident on the equipments or circuits. This situa-

tion, which involves a "ground loop," is referred to as field-to-cable com-

mon-mode coupling because the currents in the two interconnecting

wires will be flowing in a "common" direction. Alternatively, common-

Radiations from

ICdips

Logic families

clock rates

Large single-layer board

PCB card cage with back plane

Multi-layer board

Radiation from

ribbon cables

Fig ure 4.13 Principal radiation sources from a printed circuit board.



72


Electromagnetic Wave

Box 1

Ground Plane

I

1?

I

2

Represent Differential M ode Current

Figure 4.14

Field-to-cable differential-mod e coup ling.


, I2 Represent Common

Mode Currents

Signal

Beference

Plane

Figure 4.15 Field-to-cable common-mode coupling.

mode curre nts flowing in a ground loop will rad iat e E MI. This situa tion

is referred to as cable-to-field comm on-mode coupling.

The coupling of an electric field into or out of a loop area, as ind icate d

in F ig. 4.14 an d Fig. 4.15 is a function of th e d ime nsions of th e loop (i.e.,

length (L) of the interconnecting wires and either the spacing (s)

between them for differential-mode coupling or their height above

ground for common-mode coupling and frequency.

The equations presented in Table 4.2 may be used to calculate the

voltage induced in a loop as a res ult of exposing the loop to an exte rna l




73

Table 4.2 Voltage Indu ced in a Loop Exposed to a Field

Small loop

Large loop

L <

2

V (volts) =

2TEEA

f,

MHz

3

V

(volts)

= TCES

wh ere, E = incident electric field in V/m

A = area of

loop

in square m eters = L s

L = length of

loop

in m eters

s = spacing between w ires for differential mode or spacing between

wires and ground for common mode

fjVIHz = frequency in megahertz

A, = wavelength in m eters

field. The equations presented in Table 4.3 may be used to calculate the

electromagnetic field strength radiated from a loop or dipole.

Table 4.3 Field Ra dia ted from a Dipole or Loop

Dipo le radiation far field d > X/2n Loop radia tion far field d > A/2TT

Small

dipole

Large

dipole

L <

2

E(V/m) =

Z

o

I L f

M H z

600 d

E V/in)«JL

where,

Z

o

= plane wave impedance (120

n

ohms)

I = dipole current in amps

L = length of dipole in m eters

d = distance from source in meters

A, = wave length in m eters

=

frequency in megahertz

Small

loop

Large

loop

L <

2

E(V/m) =

JAf

2

MHz

(300)

2

d

L >

2

E(V/m) =

w Z

o

I S

f

MHz

600 d

where,

Z

o

= plane wave impedance (120

n

ohms)

I = loop current in amps

A = area of

loop

in square m eters = L s

d = distance from source in meters

— wavelength in meters

^MHz

=

frequency in megahertz

The fourth mechanism by which conducted EMI emissions can cou-

ple from a source to a victim involves coupling of EMI (or crosstalk)

betw een two p airs of wires (one pair carryin g conducted em issions from

a source and the other pair connected to a susceptible device). Coupling



74

ELECTRONIC SYSTEM DESIGN FOR EM C

between two wire pairs, between two coaxial lines, or between one wire

pair and one coaxial line involves both electric- and magnetic-field cou-

pling. The former is represented by mutual capacitive coupling between

the lines,

and the

latte r corresponds

to

m utu al inductive coupling

between the EMI source and victim lines. The procedure for calculating

cable-to-cable coupling is pre sen ted in Appendix A.

When the victim circuit impedance is high relative to the character-

istic imp edance of free space (377 fit), capacitive coupling p redom inates.

This coupling increases with frequency, the length of the wires, the

spacing between t he w ires in a pair, and th e proxim ity of th e wire p airs .

Figure 4.16 shows the network involving capacitive coupling between

culprit line an d victim circuits. A portion of th e available c ulprit source

line voltage (V

c

) is coupled into th e victim load (Zj). This type of wire to

wire coupling is often referred to as

crosstalk.

Figure 4.17 shows a similar cable network involving inductive cou-

pling between culprit and victim line circuits. As before, with capacitive

coupling, a portion of th e av ailable cu lprit source line voltage (V^ is cou-

pled into the victim load. Inductive coupling predominates when the cir-

cuit impedances are low relative to 377 Q. This coupling also increases

with frequency, the length of the wires, the spacing between the wires in

a pair and the proximity of the wire pairs. The ratio of victim-to-culprit

voltages represents the cable-to-cable coupling or crosstalk.

4 1 3 Susceptible Equipments

Any device capable of responding to electrical, electromechanical, or

electronic emissions, or to the fields associated with these emissions, is

Culprit Line

Voltage

Victim Input

—

v

,

v

—7

Voltage

F ig u re 4.16 Circuit repres entatio n of capacitive coupling between parallel

wires over a ground p lane.




75

Figure 4.17 Circuit representation of inductive coupling between parallel

wires over a ground plane.

potentially vulnerable to EMI. Susceptibility of all such devices may be

divided into two categories: (1) devices susceptible to interfering emis-

sions over a broadband of frequencies, and (2) devices that are fre-

quency selective. Typical devices that may be considered vulnerable to

interfering emissions over a few or many octaves include remote-control

switches, relays, indicator lights, electro-explosive squibs, recording

devices, logic circuits, and meters. Frequency-selective devices prima-

rily include equipments or systems such as communication, radar, and

navigation receivers.

Figure 4.18 shows such an organization and identifies typical recep-

tors for each category. Receptors of EMI can be divided into natural and

man-made. This handbook is concerned with only man-made receptors.

EMI can cause problems in susceptible equipments as a result of either

radiated or conducted emissions, and therefore it is important to con-

sider the susceptibility of equipments to both emission types.

Communication receivers are potentially very susceptible to radiated

emissions that fall within the receiver passband. In addition, receivers

will respond to strong radiated emissions at other frequencies.

Other electronic equipment used in system applications may also be

susceptible to radiated emissions, and this must be considered by the

system designer.

Electronic circuits are susceptible to conducted EMI that is coupled

into the circuits through interconnecting wires and cables. The suscep-

tibility of various electronic circuits may vary widely. Sensitive circuits

such as analog amplifiers will typically be susceptible to signals in the

microvolt to millivolt range, whereas digital logic circuits will typically

be susceptible to signals in the volt range.



76


Man-Made

Receptors of EMI

1

Communications

Electronics

Receivers

Amplifiers

- Broadcast

- Relay Link

- Navigation

-Radar

L

Communications

Industrial &

Consumer

- I F

- Video

- Audio

- Controls

- Bio Medical

- Instruments

- Audio/Hi-Fi

- Public Address

- Telephones

- Sensors

- Computers

^ Status M onitors

Ordnance

RADHAZ

hEEDs

•-Fuels

Fig ure 4.18 Receptors of electromagnetic interference.

Electronic components are also susceptible to burnout as a result of

exposure to high levels of electromag netic energy.

4.2 Sy stem -Lev el EMI Con trol

System-level EMI control techniques involve both hardware and

methods and procedures. Engineers and technicians must become

knowledgeable and accomplished in system EMI control techniques.

Chapters 5, 6, 7, 8, ad 9 present an overview of the major techniques.

Figu re 4.19 illust rate s th e basic EMI characte ristics of concern in a sys-

Test Specimen

I conducted

Antenna

Terminal

Susceptibility

Power Mains

Interconnecting

Cable

Antenna

Conducted

Emission

{Key

up and Down}

F ig u re 4.19 Basic EMI charac teristics of concern for EMI systems problem.



SYSTEM-LEVEL EM I CONTROL

77

tern EMI problem. The specimen may be a single box, an equipment, a

subsystem, or a system (an ensemble of boxes with interconnecting

cables). Problems associated with either (1) susceptibility to outside

conducted and/or radiated emissions or (2) tendency to pollute the out-

side world from its own undesired emissions come under the primary

classification of intra-system EMI. Corresponding EMI control tech-

niques address themselves to emission/susceptibility in accordance

with applicable EMI specifications.

Figure 4.20 presents an organization tree that groups system EMI

control techniques by five fundamental categories often appearing in

the literature: equipment selection, grounding, wiring, filtering, and

shielding. Bonding, connectors, gaskets, and other topics appear as sub-

categories, as shown in the figure. In general, system-level EMI control

is accomplished through the application of one or more of the following

control considerations:

• Control of EMI emissions at th e source

• Control of EMI coupling between sources and susceptible components

• Control of th e EMI susceptib ility of victim s

System

EMI Control

1

Equipment

Selection

Power

Supplies

Rotating

Devices

Arc

Suppressior

Induction

Solid S tate

i

Relays

Solenoids

^Filters

•-Clamps

Electronic

Circuits

Grounding

" Objects

-Buildings

-Rooms

-Cabinets

-Chassis

Circuits

Cable

Bonds

[-Types

[-Surfaces

•-Corrosion

1

Wiring

• Cabling

-Grouping

-Types

-Ground

-Loops

-Shielding

-

Connectors

[-Shielded

L Filter Typ<

Filtering

Power

Mains

-Filters

-Beads/Rods

-Lossy line

-Connectors

Isolation

Transformers

• Low Level

|_LP, BP,

HP

&BR Filters

Shielding

Housing

Chassis &

Cabinets

-Rooms

-Matrials

-Thickness

Packaging

-Gaskets

-Seals

-Apertures

F ig u re 4.20 Organization tree of system EMI control techniques.



78 ELECTRONIC SYSTEM DESIGN FOR EMC

Control of EMI emissions at the source and susceptibility of victims can

be accomplished by:

• Carefully selecting equipments for the ir EMI and susceptibility

characteristics

• Using shielding to control radiated EMI effects

• Using EMI suppression devices such as filters, ferrites, and isola-

tion transformers to control conducted EMI effects

One of the most important considerations in designing a system for

EMC is to give proper attention to the selection of the various equip-

ments that compose the system. Equipments should be selected on the

basis of their EMC characteristics as well as other considerations such

as operational specifications and cost. Particular attention should be

directed toward a consideration of the EMC characteristics of equip-

ments that are likely to present problems because they are either

potential sources of EMI or potentially susceptible to EMI.

Equipments that contain power supplies, rotating devices, relays,

and solenoids should be suspect as being potential sources of

EMI,

and

the equipment selection should consider the extent to which EMI emis-

sions are suppressed at the source or contained within the equipment

enclosure by filtering, shielding, etc.

Equipments that contain electronic circuits should be suspect as

being either potential sources of EMI or potentially susceptible to EMI.

Circuits such as clock circuits, switching rectifiers, oscillators, and so

forth should be regarded as sources, whereas circuits such as analog

amplifiers, digital logic, sensors and controls, and so on should be

regarded as susceptible. In either instance, EMI characteristics should

be an important consideration in selecting equipments containing these

types of circuits. In addition to the different modes of coupling EMI,

there are a number of other factors that must be considered. For exam-

ple,

consider the situation shown in Fig. 4.21, where there are two

interconnected equipm ents. In this case, there are a total of 29 question

marks indicated. Each question mark identifies a decision point that

requires a "y

es

" or "no" answer. Thus, a question mark associated with

one of the multiple grounds requires a "yes" or "no" answer signifying

whether there is a connection to ground at the indicated point. A ques-

tion mark associated with an EMI mitigation component requires a

"yes"

or "no" answer signifying whether the component is or is not used.

For the purpose of identification in the figure, FR refers to ferrites, IT to

isolation transformers, IS to isolators (optical or transformers), F to fil-

ters, C

to connectors, and PS to power supply.



L

c

n

e

d

T

a

n

m

e

s

C

=

C

o

F

F

e

F

R

F

e

I

T

I

s

o

a

o

T

a

n

o

m

V

7

I

S

=

I

s

o

a

o

O

c

o

T

a

n

o

m

P

P

w

S

y

M

a

n

M

ad

N

s

e

—

G

o

P

a

n

o

S

e

y

W

i

r

e

F

i

g

u

r

4

2

I

n

e

c

o

e

d

e

q

p

m

n

s

w

h

2

f

x

c

h

c

r

e

u

n

n

m

e

h

n

5

m

o

d

g

o

o



80 ELECTRONIC SYSTEM DESIGN FOR EM C

For this simple two-box system, there are a total of 29 question

marks, each requiring a "y

e s

" or "no" answer. This results in more than

500,000,000 combinations. Some

of the

com binations will res ult

in

EMC, others will res ult in E MI. Some of the co mbinations t h at res ult in

EMC will be bette r th an others . Also, it must be recognized that the

problem often involves more than a simple "yes" or "no" answ er. For

example, if th e decision is m ade to u se a filter, this results in a number

of new questions, such as: What

is

th e cutoff frequency? W ha t

is the

slope in the stop band ? ...a nd so on.

Suggested Readings: Electronic System Design for EMC

[1] Dash, Glen, "Minimizing Ringing an d C rosstalk,"

Compliance Mag-

azine, 2010 Annual Guide, p . 50.

[2] Montrose, Mark, Printed Circuit Board Design Techniques for EMC

Compliance, IEEE Press, 1996, p. 85.

[3] Black, Jack , "EMC an d Aerospace,"

Compliance Magazine,

July

2010, p. 18.

[4]

Tab ataba ei, Sas san, "Clocking S trateg ies

for EMI

R eduction,"

ITEM interference technology, 2010 EMC Test

Design Guide,

p.

46.

[5]

Arc ham beau lt, Bruce, "Distributed Decoupling Capacitor Effective-

n e s s / ' ITEM interference technology, 2009

EMC

Directory

and




Chapter 5

Grounding for the

Control of

EMI

There are two primary reasons for grounding devices, cables, equip-

ments, and systems. The first reason is to prevent shock and fire haz-

ards in the event that an equipment frame or housing develops a high

voltage due to lightning or an accidental breakdown of wiring or compo-

nents.

The second reason is to reduce EMI effects resulting from elec-

tromagnetic fields, common impedance, or other forms of interference

coupling.

Historically, grounding requirements arose from the need to provide

protection from electrical faults, lightning, and industrially generated

static electricity. Because most power-fault and lightning control relies

on a low-impedance path to earth, all major components of an electrical

power generation and transmission system were earth grounded to pro-

vide the required low-impedance path .

As

a result, strong emphasis was

placed on earth grounding of electrical equipment, and the overall phi-

losophy was ground, ground, ground without regard to other prob-

lems,

such as EMI, that may be created by this approach.

When electronic equipments were introduced, grounding problems

became evident. These problems resulted from the fact that the circuit

and equipment grounds often provided the mechanism for undesired

EMI coupling. Also, with electronic systems, the ground may simulta-

neously perform two or more functions, and these multiple functions

may be in conflict either in terms of operational requirements or in

terms of implementation techniques. For example, as illustrated in

Fig. 5.1, the ground network for an electronic equipment may be used

as a signal return, provide safety, provide EMI control, and also per-

form as part of an antenna system.

Therefore, in order to avoid creating EMI problems, it is essential to

recognize that an effective grounding system, like any other portion of

81



82

GROUNDING FOR THE CONTR OL OF EMI

Electronic Enclosure

6\S>

Signal Return, de Common

Signal Ground, Cabinet Ground, Safety Ground, etc.

Building Ground, Power Ground, Safety Ground, etc

Antenna Ground,

Building Ground,

Lightning Ground

etc.

Figure 5.1 The multiple functions of grounds.

an equipment or system, must be carefully designed and implemented.

Grounding is a system problem and in order for a grounding arrange-

ment to perform well it must be well conceived and accurately designed

and implemented. The grounding configurations must be weighed with

regard to dimensions and frequency, just like any functional circuit.

The objective of this chapter is to help engineers, designers, and

technicians to optimize the functionality and reliability of their equip-

ment by providing an orderly systems approach to grounding. Such an

approach is highly preferable to the em pirical and sometimes contradic-

tory approaches th at are often employed.

5.1 Definitions

The term ground is one of the most abused words in the electronic engi-

neering vocabulary. In addition, several other words are often used in

conjunction with the term ground, and these words are also often mis-

used. For the purpose of this chapter, it is important to carefully define

these terms. The definitions that follow are given in terms of the noun

rather than the verb.

Ground: Any reference conductor th at is used for a common re turn.

Earth:

The soil into which a safety conductor (rod, grid, plate) is driven

or buried to provide a low-impedance sink for fault and lightning cur-

rents.



CHARACTERISTICS OF GROUNDING SYSTEMS

83

Reference: Some object whose potential (often 0 V with respect to ea r th

or a power supply) is th e one to which analog and logic circuits,

equipments ,

and

sys tems

can be

related

or

benchmarked.

Return:

Th e low

(reference) voltage side

of a

wire pair (e.g., ne utra l),

outer jacket

of a

coax

or

conductor providing

a

pa th

for

intent ional

current to get back to the source.

Bond: The process used to join two metal surfaces via a low-impedance

pa th .

Connection: A mec hanical joint between tw o electrical condu ctors, often

including

an

interm ediary conductor such

as a

jump er, pigtail ,

or

shield braid.

Figure 5.2 i l lus t ra tes th e reason that th e term ground can be a mis-

leading, ambiguous term

if one

does

not

consider

i ts

electrical parame-

ters . Referring to Fig. 5.2, it is apparent that significant voltages m ay

exist between

two

different points

on the

ground associated with

a

platform, facility,

or

rack. This potential difference

is a

major cause

for

EMI problems resulting from grounding of circuits , equipme nts , or sys-

t ems .

5.2 Ch aracteristics of Grounding System s

Ideally, a ground system should provide a zero-impedance path to all

signals for which it serves as a reference. If this were the situation, sig-

nal currents from different circuits or equipments th at are connected to

the ground could return to their respective sources without creating

unwanted coupling between the circuits or equipments. Many interfer-

ence problems occur because designers treat the ground as ideal and

fail to give proper attention to the actual characteristics of the ground-

ing system. One of the primary reasons th at designers trea t the ground

system as ideal is that this assumption is often valid from the stand-

point of the circuit or equipment design parameters (i.e., the impedance

at power or signal frequencies is small and has little or no impact on

circuit or equipment performance). However, the non-ideal properties of

the ground must be recognized if EMI problems are to be avoided.

5.2.1 Impedance Ch aracteristics

Every element (conductor) of a grounding system, whether it be for

power grounding, signal grounding, or lightning protection, has proper-

ties of resistance, capacitance, and inductance. Shields and drain wires



84

GROUNDING FOR THE CONTRO L OF EMI

G round Means any Referen ce Conductor

that is U sed for a Common Return

Earthing is only a particular case of grounding.

On an aircraft,

10 to 100 V

differences may exist

between structural points.

DDnnnn

In a

building,

levels of several

kilovolts develop on grounds

when lightning creates e arth

gradients.

Ground?

In vehicles, differences of

several volts develop betwee n

points on the steel body.

What For?

Where?

How?

Is this ground

really equipotential?

In a ship, levels of several

hundred volts exist

between decks,

supe rstructures and rigging.

In racks, several hundred

millivolts can develop between

different drawers.

Fig ure 5.2 Ground can be a misleading, ambiguous term if one does not con-

sider its electrical parameters.

of signal cables, the green wire power safety ground, lightning down

conductors, transformer vault buses, structural steel members—all

conductors have these properties. The resistance property is exhibited

by all metals . The resistan ce of a ground pa th conductor is a function of

the material, its length, and its cross-sectional area. The capacitance

associated with a ground conductor is determined by its geometric



CHARACTERISTICS OF GROUNDING SYSTEMS 85

shape, its proximity to other conductors, and the natu re of the interven-

ing dielectric. The inductance is a function of its size, geometry, length,

and, to a limited extent, the relative permeability of the metal. The

impedance of the grounding system is a function of the resistance,

inductance, capacitance, and frequency.

Because the inductance properties of a conductor decrease with

width and increase with length, it is frequently recommended that a

length-to-width ratio of 5:1 be used for grounding straps. This 5:1

length-to-width ra tio provides a reactance th at is approximately

45

per-

cent of that of a straight circular wire.

The impedance of straight circular wires is provided as a function of

frequency in Table 5.1 for several wire gauges and lengths. Typical

ground plane impedances are provided in Table 5.2 for comparison. Note

that for typical length wires, ground plane impedances are several orders

of magnitude less than those of

a

circular

wire. Also

note tha t the imped-

ance of both circular wires and ground planes increase with increasing

frequency and become quite significant at higher frequencies.

A commonly encountered situation is that of a ground cable (power

or signal) running along in the proximity of a ground plane . This situa-

tion is illustrated in Fig. 5.3 for equipment grounding. Figure 5.4 illus-

tra tes a representative circuit of this simple ground pa th. The effects of

the resistive elements of the circuit will predominate at very low fre-

quencies. The relative influence of the reactive elements will increase

at increasing frequencies. At some frequency, the magnitude of the

inductive reactance (jcoL) equals the magnitude of the capacitive reac-

tance (1/jcoC), and the circuit becomes resonant. The frequency of the

primary (or first) resonance can be determined from:

f = — L = (5.1)

where L is the total cable inductance, and C is the net capacitance

between the cable and the ground plane. At resonance, the impedance

presented by the grounding path will either be high or low, depending

on whe ther it is para llel or series reso nan t, respectively. At paralle l res-

onance, the impedance seen looking into one end of the cable will be

much higher than expected from R + jcoL. (For good conductors, e.g.,

copper and aluminum, R « coL; th us , jooL generally provides a n accu rate

esti m ate of th e imped ance of a ground conductor a t frequencies above a

few h un dred hertz). At parallel resonance:

Z

p

= QcoL (5.2)



86

GROUNDING

FOR THE CONTROL OF

EMI

Table 5

.1 Impedance of Stra ight Circular Copper Wires

AWG# = S

Freq.

10HZ

20Hz

30Hz

50Hz

70Hz

lOOHz

200Hz

300Hz

500Hz

700Hz

1kHz

2kHz

3kHz

5kHz

7kHz

10kHz

20kHz

30kHz

50kHz

70kHz

100kHz

200kHz

300kHz

500kHz

700kHz

1MHz

2MHz

3MHz

5MHz

7MHz

10MHz

20MHz

30MHz

50MHz

70MHz

100MHz

200MHz

300MHz

500MHz

700MHz

lGHz

l=lcm

5.13*1

5.14*1

5.15*i

5.20*i

5.27*i

5.41*1

6.20*i

7.32*i

10.1*1

13.2*1

lain

35.2*1

52.5*i

87.3*i

122*i

174*1

348*1

523*i

871*i

1.22m

1.74m

3.48m

5.23m

8.71m

12.2m

17.4m

34.8m

52.3m

87.1m

122m

174m

348m

523m

871m

1.22ft

1.74ft

3.48ft

5.23ft

8.71ft

12.2ft

17.4ft

I,

D = 6.54mm

I

= 10cm

51.4*1

52.0*1

52.8*i

55.5*1

59.3*i

66.7*i

99.5*1

137*1

219*1

303*1

429*i

855*i

1.28m

2.13m

2.98m

4.26m

8.53m

12.8m

21.3m

29.8m

42.6m

85.3m

128m

213m

298m

426m

853m

1.28ft

2.13ft

2.98ft

4.26ft

8.53ft

12.8ft

21.3ft

29.8ft

42.6ft

85.3ft

128ft

213ft

298ft

426ft

l = lm

517*1

532*i

555*i

624*i

715*i

877*i

1.51m

2.19m

3.59m

5.01m

7.14m

14.2m

21.3m

35.6m

49.8m

71.2m

142m

213m

356m

498m

712m

1.42ft

2.13ft

3.56ft

4.98ft

7.12ft

14.2ft

21.3ft

35.6ft

49.8ft

71.2ft

142ft

213ft

356ft

498ft

712ft

1.42kft

2.13kft

3.56kft

4.98kft

7.12kft

UlQm

5.22m

5.50m

5.94m

7.16m

8.68m

11.2m

20.6m

30.4m

50.3m

70.2m

100m

200m

300m

500m

700m

1.00ft

2.00ft

3.00ft

5.00ft

7.00ft

10.0ft

20.0ft

30.0ft

50.0ft

70.0ft

100ft

200ft

300ft

500ft

700ft

l.OOkft

2.00kft

3.00kft

5.00kft

7.00kft

lO.Okft

20.0kft

30.0kft

50.0kft

70.0kft

*AWG = Am erican Wire Gage

D = wire diameter in mm

I =

wire length

in cm or m

M-

=

nicrohms

m = m illiohms

Q. =Dhms

AWG#=10,

1 =

lcm

32.7*1

32.7*i

32.8^1

32.8*1

32.8*i

32.9*1

33.2*1

33.7*1

35.3*1

37.7*i

42.2*i

62.5*i

86.3*i

137*1

189*i

268*1

533*i

799*1

1.33m

1.86m

2.66m

5.32m

7.98m

13.3m

18.6m

26.6m

53.2m

79.8m

133m

186m

266m

532m

798m

1.33ft

1.86ft

2.66ft

5.32ft

7.98ft

13.3ft

18.6ft

26.6ft

I = 10cm

327*1

328*1

328*i

329*1

330ji

332*1

345*1

365*1

425*1

500*i

632*i

1.13m

1.65m

2.72m

3.79m

5.41m

10.8m

16.2m

27.0m

37.8m

54.0m

108m

162m

270m

378m

540m

1.08ft

1.62ft

2.70ft

3.78ft

5.40ft

10.8ft

16.2ft

27.0ft

37.8ft

54.0ft

108ft

162ft

270ft

378ft

540ft

D

=

2.59mm

l = lm

3.28m

3.28m

3.28m

3.30m

3.33m

3.38m

3.67m

4.11m

5.28m

6.66m

8.91m

16.8m

25.0m

41.5m

58.1m

82.9m

165m

248m

414m

580m

828m

1.65ft

2.48ft

4.14ft

5.80ft

8.28ft

16.5ft

24.8ft

41.4ft

58.0ft

82.8ft

165ft

248ft

414ft

580ft

828ft

1.65kft

2.48kft

4.14kft

5.80kft

8.28kft

< = 10m

32.8m

32.8m

32.9m

33.2m

33.7m

34.6m

39.6m

46.9m

64.8m

84.8m

116m

225m

336m

559m

783m

1.11ft

2.23ft

3.35ft

5.58ft

7.82ft

11.1ft

22.3ft

33.5ft

55.8ft

78.2ft

111ft

223ft

335ft

558ft

782ft

l.llkft

2.23kft

3.35kft

5.58kft

7.82kft

ll.lkft

22.3kft

33.5kft

55.8kft

78.2kft

>

AWG# =

22

1= lcm

529*1

529*i

529*i

530*1

530*1

530*1

530*1

530*A

530*i

530*i

531*i

536*i

545*i

571*i

609p

681*1

1.00m

1.39m

2.20m

3.04m

4.31m

8.59m

12.8m

21.4m

30.0m

42.8m

85.7m

128m

214m

300m

428m

857m

1.28ft

2.14ft

3.00ft

4.28ft

8.57ft

12.8ft

21.4ft

30.0ft

42.8ft

I =

10cm

5.29m

5.29m

5.30m

5.30m

5.30m

5.30m

5.30m

5.30m

5.31m

5.32m

5.34m

5.48m

5.71m

6.39m

7.28m

8.89m

15.2m

22.0m

36.1m

50.2m

71.6m

142m

214m

357m

500m

714m

1.42ft

2.14ft

3.57ft

5.00ft

7.14ft

14.2ft

21.4ft

35.7ft

50.0ft

71.4ft

142ft

214ft

357ft

500ft

714ft

1 1 Non-Valid Region

1

J

f o r w h i c h ^ A / 4

D = .64mm

i = l m

52.9m

53.0m

53.0m

53.0m

53.0m

53.0m

53.0m

53.0m

53.2m

53.4m

53.9m

56.6m

60.9m

72.9m

87.9m

113m

207m

305m

504m

704m

1.00ft

2.00ft

3.01ft

5.01ft

7.02ft

10.0ft

20.0ft

30.1ft

50.1ft

70.2ft

100ft

200ft

301ft

501ft

702ft

l.OOkft

2.00kft

3.01kft

5.01kft

7.02kft

lO.Okft

UlOm

529m

530m

530m

530m

530m

530m

530m

531m

533m

537m

545m

589m

656m

835m

1.04ft

1.39ft

2.63ft

3.91ft

6.48ft

9.06ft

12.9ft

25.8ft

38.7ft

64.6ft

90.4ft

129ft

258ft

387ft

646ft

904ft

1.29kft

2.58kft

3.87kft

6.46kft

9.04kft

12.9kft

25.8kft

38.7kft

64.6kft

90.4kft

where

Q, the

qua lity factor,

is

defined

as:

Q =

ooL

R

(ac)

5.3)

where R(

ac

)

is the

cable resistan ce

at the

frequency

of

resonance. Then:



CHARACTERISTICS OF GROUNDING SYSTEMS

87

T a b le 5

Freq.

10HZ

20Hz

30Hz

50Hz

70Hz

lOOHz

200Hz

300Hz

500Hz

700Hz

1kHz

2kHz

3kHz

5kHz

7kHz

10kHz

20kHz

30kHz

50kHz

70kHz

100kHz

200kHz

300kHz

500kHz

700kHz

1MHz

2MHz

3MHz

5MHz

7MHz

10MHz

20MHz

30MHz

50MHz

70MHz

100MHz

200MHz

300MHz

500MHz

700MHz

lGHz

2GHz

3GHz

5GHz

7GHz

lOGHz

2 Metal Ground Plane Impedance i

COPPER,

COND-1,

t = .O3

5 7 4 M

5 7 4 M

574|i

5 7 4 M

574|n

574^1

5 7 4 M

5 7 4 M

574|i

5 7 4 M

574ji

5 7 4 M

5 7 4 M

5 7 4 | L I

5 7 4 M

5 7 4 M

574^

5 7 4 M

5 7 4 M

5 7 4 M

5 7 5 M

5 7 5 M

576jx

5 7 8 M

582jx

604M-

6 3 8 M

7 3 6 M

8 5 5 M

1.04m

1.61m

2.03m

2.62m

3.09m

3.69m

5.22m

6.39m

8.26m

9.77m

11.6m

16.5m

20.2m

26.1m

30.9m

36.9m

t = .i

1 7 2 M

172JJ.

172|i

172M.

1 7 2 M

172fx

1 7 2 M

172M

1 7 2 M

1 7 2 M

172^

172M

172M

172M

172M

172M

172M

172M

173M

173M

175M

1 8 3 M

195M

2 3 0 M

2 7 1 M

3 3 5 M

5 1 6 M

6 4 3 M

8 2 7 M

9 7 7 M

1.16m

1.65m

2.02m

2.61m

3.09m

3.69m

5.22m

6.39m

8.26m

9.77m

11.6m

16.5m

20.2m

26.1m

30.9m

36.9m

t = .3

5 7 . 4 M

5 7 . 4 M

5 7 . 4 M

5 7 . 4 M

5 7 . 4 M

5 7 . 4 M

5 7 . 4 M

5 7 . 4 M

5 7 . 4 M

5 7 . 4 M

5 7 . 4 M

5 7 . 5 M

5 7 . 5 M

5 7 . 6 M

5 7 . 8 M

5 8 . 2 M

6 0 . 4 M

6 3 . 8 M

7 3 . 6 M

8 5 . 5 M

140M

1 6 1 M

2 0 3 M

2 6 2 M

3 0 9 M

3 6 9 M

5 2 2 M

6 3 9 M

8 2 6 M

9 7 7 M

1.16m

1.65m

2.02m

2.61m

3.09m

3.69m

5.22m

6.39m

8.26m

9.77m

11.6m

16.5m

20.2m

26.1m

30.9m

36.9m

* t is in units of mm

M<

m

=

m icrohms

=

milliohms

= ohms

PERM-

t = i

17. 2M

17. 2M

17. 2M

17. 2M

17. 2M

17. 2M

17. 2M

17. 2M

17. 3M

17. 3M

1 7 . 5 M

18. 3M

19. 5M

2 3 . 0 M

2 7 . 1 M

3 3 . 5 M

51. 6M

6 4 . 3 M

8 2 . 7 M

9 7 . 7 M

116M

1 6 5 M

2 0 2 M

2 6 1 M

3 0 9 M

3 6 9 M

5 2 2 M

6 3 9 M

8 2 6 M

9 7 7 M

1.16m

1.65m

2.02m

2.61m

3.09m

3.69m

5.22m

6.39m

8.26m

9.77m

11.6m

16.5m

20.2m

26.1m

30.9m

36.9m

1

t = 3

5 . 7 4 M

5 . 7 5 M

5 . 7 5 M

5 . 7 6 M

5 . 7 8 M

5. 82M

6 . 0 4 M

6 . 3 8 M

7. 36M

8 . 5 5 M

10. 4M

1 6 . 1 M

2 0 . 3 M

2 6 . 2 M

3 0 . 9 M

3 6 . 9 M

52. 2M

6 3 . 9 M

8 2 . 6 M

9 7 . 7 M

116M

1 6 5 M

2 0 2 M

2 6 1 M

3 0 9 M

3 6 9 M

5 2 2 M

6 3 9 M

8 2 6 M

9 7 7 M

1.16m

1.65m

2.02m

2.61m

3.09m

3.69m

5.22m

6.39m

8.26m

9.77m

11.6m

16.5m

20.2m

26.1m

30.9m

36.9m

t = 1 0

1.75M

1.83M

1.95M

2 . 3 0 M

2 . 7 1 M

3 . 3 5 M

5. 16M

6 . 4 3 M

8 . 2 7 M

9 . 7 7 M

11. 6M

16. 5M

2 0 . 2 M

2 6 . 1 M

3 0 . 9 M

3 6 . 9 M

5 2 . 2 M

6 3 . 9 M

8 2 . 6 M

9 7 . 7 M

116M

1 6 5 M

2 0 2 M

2 6 1 M

3 0 9 M

3 6 9 M

5 2 2 M

6 3 9 M

8 2 6 M

9 7 7 M

1.16m

1.65m

2.02m

2.61m

3.09m

3.69m

5.22m

6.39m

8.26m

9.77m

11.6m

16.5m

20.2m

26.1m

30.9m

36.9m

n Ohms/Square

t = .O3

3.38m

3.38m

3.38m

3.38m

3.38m

3.38m

3.38m

3.38m

3.38m

3.38m

3.38m

3.38m

3.38m

3.38m

3.38m

3.38m

3.40m

3.42m

3.50m

3.62m

3.85m

4.95m

6.23m

8.62m

10.5m

12.7m

17.9m

21.9m

28.3m

33.5m

40.0m

56.6m

69.4m

89.6m

106m

126m

179m

219m

283m

335m

400m

566m

694m

896m

1.06ft

1.26ft

STEEL, COND-17, PERM-20C

t = .i

1.01m

1.01m

1.01m

1.01m

1.01m

1.01m

1.01m

1.01m

1.01m

1.01m

1.01m

1.02m

1.03m

1.06m

1.10m

1.18m

1.57m

1.99m

2.75m

3.35m

4.03m

5.66m

6.93m

8.96m

10.6m

12.6m

17.9m

21.9m

28.3m

33.5m

40.0m

56.6m

69.4m

89.6m

106m

126m

179m

219m

283m

335m

400m

566m

694m

896m

1.06ft

1.26ft

t = .3

3 3 8 M

3 3 8 M

3 3 8 M

3 3 8 M

3 3 8 M

3 3 8 M

3 4 0 M

3 4 2 M

3 5 0 M

3 6 2 M

3 8 5 M

4 9 5 M

6 2 3 M

8 6 2 M

1.05m

1.27m

1.79m

2.19m

2.83m

3.35m

4.00m

5.66m

6.94m

8.96m

10.6m

12.6m

17.9m

21.9m

28.3m

33.5m

40.0m

56.6m

69.4m

89.6m

106m

126m

179m

219m

283m

335m

400m

566m

694m

896m

1.06ft

1.26ft

t = l

1 0 1 M

102M

1 0 3 M

106M

110M

118M

157M

199M

2 7 5 M

3 3 5 M

4 0 3 M

5 6 6 M

6 9 3 M

8 9 6 M

1.06m

1.26m

1.79m

2.19m

2.83m

3.35m

4.00m

5.66m

6.94m

8.96m

10.6m

12.6m

17.9m

21.9m

28.3m

33.5m

40.0m

56.6m

69.4m

89.6m

106m

126m

179m

219m

283m

335m

400m

566m

694m

896m

1.06ft

1.26ft

t = 3

3 8 . 5 M

4 9 . 5 M

6 2 . 3 M

8 6 . 2 M

105M

127M

179M

2 1 9 M

2 8 3 M

3 3 5 M

4 0 0 M

5 6 6 M

6 9 4 M

8 9 6 M

1.06m

1.26m

1.79m

2.19m

2.83m

3.35m

4.00m

5.66m

6.94m

8.96m

10.6m

12.6m

17.9m

21.9m

28.3m

33.5m

40.0m

56.6m

69.4m

89.6m

106m

126m

179m

219m

283m

335m

400m

566m

694m

896m

1.06ft

1.26ft

t = 1

4 0 . 3 M

5 6 . 6 M

6 9 . 3 M

8 9 . 6 M

106M

126M

179M

2 1 9 M

2 8 3 M

3 3 5 M

4 0 0 M

5 6 6 M

6 9 4 M

896M

1.06m

1.26m

1.79m

2.19m

2.83m

3.35m

4.00m

5.66m

6.94m

8.96m

10.6m

12.6m

17.9m

21.9m

28.3m

33.5m

40.0m

56.6m

69.4m

89.6m

106m

126m

179m

219m

283m

335m

400m

566m

694m

896m

1.06ft

1.26ft

NOTE: Do

not

use table

at

frequencies

in MHz

above

5/l

m

since

the

separation distance

in

meters,

l

m

,

of two

grounded equ ipments will exceed

0.05A,

where error

becomes significant.

Z

p

=

QcoL

=

(5.4)

v

(ac)

v

(ac)



GROUNDING

FOR THE

CONTROL

OF EMI

Equipment

Grounding C onductor

z

0

= S/L/C

Z

L

= 0

y

Ground Plane

Fig ure 5.3 Idealized equipment grounding.

Ground

Cable

Ground

Plane

Fig ure 5.4 Equivalent circuit of a ground cable parallel to a ground plane.

Above the primary resonance, subsequent resonances (both parallel

and series) will occur between the various possible combinations of

inductances and capacitances (including parasitics) in the pa th.

Series resonances in the grounding circuit will also occur between

the inductances of wire segments and one or more of the shunt capaci-

tances. The impedance (Z

s

) of a series resonant path is:

_

COL

5.5)

Therefore,

R

(ac)

5.6)

The series resonant impedance is thus determined by, and is equal to,

the series ac resistance of the particular inductance and capacitance in

resonance. (At the higher ordered resonances, where the resonant fre-

quency is established by wire segments and not the total path, the

series impedan ce of th e pa th to ground may be less tha n pred icted from

a consideration of the entire ground conductor length).

An und ers tan di ng of th e high-frequency b ehavior of a ground ing con-

ductor is simplified by viewing it as a transmission line. If the ground



CHARACTERISTICS OF GROUNDING SYSTEMS 89

pa t h i s cons i de r ed un i f o r m a l ong i t s r un , t he vo l t ages and cu r r en t s

a long the l ine can be descr ibed as a funct ion of t im e an d d i s t anc e . If th e

r es i s t an ce e l em en t s i n F ig . 5 .4 a r e s ma l l r e l a t i ve to t h e i ndu c t an ces an d

capac i t ances , t he g r ound i ng pa t h has a cha r ac t e r i s t i c i mpedance , Z

o

,

equal to JLTC wh er e L an d C a r e t he pe r - u n i t l eng t h va l u es of i nduc -

t anc e an d cap ac i t ance . Th e s i t ua t i o n i l l u s t r a t ed i n F i g . 5 .3 i s of pa r t i cu -

l a r i n t e r e s t i n equ i pmen t g r ound i ng . The i npu t i mpedance o f t he

g r ound i ng pa t h , i . e . , t he i mpedance t o g r ound s een by t he equ i pmen t

case, i s :

5.7)

where,

P = CGVLC = the pha se constant for the transm ission line

= the len gth of th e pa th from the box to the short

where

(3%

is less tha n

n/2

radians, i.e., when the electrical path length is

less th an a qu art er w avelen gth (A/4), th e inp ut im pedan ce of th e s hort-

circuited line is inductive w ith a value rang ing from 0

(p%

= 0) to °°

(P%

=

n/2 radian s). As

p%

= increases beyond n/2 radians in value, the imped-

ance of the grounding path cycles alternately between its open- and

short-circuit values.

Thus, from the vantage point of the device or component that is

grounded, the impedance is analogous to that offered by a short-cir-

cuited trans missio n line. Where px =

n/2,

the impedance offered by the

ground conductor behaves like a lossless parallel LC resonant circuit.

Ju st below resonance, the impedance is inductive; jus t above resonance,

it is capacitive; while at resonance, the impedance is real and quite

high (infinite in the perfectly lossless case). Resonance occurs at values

of

equal to integer multiples of quarter wavelengths, such as a half

wavelength, three-quarter wavelength, etc.

Typical ground networks are complex circuits of Rs, Ls, and Cs with

frequency-dependent properties including both parallel and series reso-

nances. These resonances are im porta nt to the performance of a ground

network. Resonance effects in a grounding path are illustrated in

Fig. 5.5. The relative effectiveness of a grounding conductor as a func-

tion of frequency is directly related to its impedance behavior (Fig. 5.6).

It is eviden t from Figs . 5.5 and 5.6 that , for max im um efficiency, grou nd

conductor lengths should be a small portion of the wavelength at the

frequency of the signal of concern. The most effective performance is

obtained at frequencies well below the first resonan ce.



90


Parallel

Resonances, f

p

=

R

a c

Series

Resonances,

f

s

F ig u re 5.5 Typical impedance vs. frequency behavior of a grounding conduc-

tor.

.3

200

150

100

50

0

f

\

•

- -

50 100 150

Frequency

in

MHz

F ig u re 5.6 Photog raph of the swept frequency behavior of a grounding strap .

5.2.2 Antenna Characteristics

Antenna effects are also related to circuit resonance behavior. Ground

conductors will act as antennas to radiate or pick up potential interfer-

ence energy, depending on their lengths relative to a wavelength, i.e.,



GROUND-RELATED INTERFERENCE 91

their efficiency. This fact permits a waveleng th-to-physical-length ratio

to be derived for ground conductors. The efficiency of

a

conductor a s

an

antenna is related to its radiatio n resistan ce. Radiation resistanc e is a

direct measure

of

th e energy ra dia ted from the an ten na . A good mea-

sure of performance for a wire is a quarter-wave monopole, which has a

radiation resistance of 36.5 Q. An ant en na th at tran sm its or receives 10

percent or less th an a monopole can be considered to be inefficient. To

be effective, a ground wire should be an inefficient an te nna . A conve-

nient criterion

for a

poor antenna, i.e.,

a

good ground wire,

is

tha t

its

length be A/10 or less. Thu s, a recommended goal in the design of an

effective grounding system is to m ainta in ground wires exposed to

potentially interfering signals at leng ths less th an 1/10 of a w avelen gth

of th e interfering signal.

5.3 Ground-Related Interference

Interference is any extraneou s electrical or electromagnetic disturbance

th at tend s to disrupt t he reception of desired signals or produces un desir-

able responses in a circuit or system . Interference can be produced by

both natural and man-made sources, either external

or

internal

to the

circuit. The correct operation of complex electronic equipment and facili-

ties is inhe rently depen dent upon th e frequencies a nd a mp litudes of both

the signals utilized in the system and the pote ntial interference emis-

sions that are present.

If

th e frequency of an und esired signal

is

within

the operating frequency range of a circuit, the circuit may respond to the

undesired signal (it may even happen out of band). The severity of the

interference is a function of the amp litude and frequency of the undes-

ired signal relative to that of the desired signal at the point of detection.

Ground-related interference often involves one of two basic coupling

mechanisms. The first mechanism results from the fact that the signal

circuits of electronic equipments share the ground with other circuits or

equipments. This mechanism

is

called com mon-ground imped ance cou-

pling. Any shared impedance can provide a mechanism for interference

coupling. Figure 5.7 illustrates the mechanism by which interference is

coupled between culprit and victim circuits via the common-ground

impedance. In this case, the interference current, I, flowing throu gh the

common-ground impedance, Z, will produce

an

interfering signal volt-

age, V

c

, in the victim circuit. I t should be emphasized that the interfer-

ence current flowing in the common impedance may be either a current

that is related to the normal operation of the culprit circuit or an inter-

mittent current that occurs due to abn orm al eve nts (lightning, power

faults, load changes, power line transients, etc.).



92

GROUND ING FOR THE CONTR OL OF EMI

Culprit

/

Source

^

Rg2

Victim

Receptor

Finite Common Impedance

in Ground

F ig u re 5.7 Common-mode impedance coupling between circuits.

Even if the equipment pairs do not use the signal ground as the sig-

nal retu rn, the signal ground can still be the cause of coupling between

them. Figure 5.8 illustrates the effect of a stray current,

IR,

flowing

n

the signal ground. The current IR may be the result of the direct cou-

pling of another equipment pair to the signal ground. It may be the

result of external coupling to the signal ground, or induced in the

ground by an incident field. In either case, IR produces a voltage Vjsj in

the ground impedance Z

R

. This voltage produces a current in the inter-

connecting loop, which in turn develops a voltage across ZL in Equip-

ment B. Thus, it is evident that interference can conductively couple

through the signal ground to all circuits and equipment connected

across the non-zero impedance elements of tha t ground.

Equipment A

Equipment B

Ground

IR

F ig u re 5.8 Conductive coupling of extraneo us noise into equipm ent intercon-

necting cables.



GROUND-RELATED INTERFERENCE

93

The second EMI coupling mechanism involving ground is a radiation

mechanism whereby the ground loop, as shown in Fig. 5.9, acts as a

receiving or transmitting antenna. For this EMI coupling mechanism,

the characteristics of the ground (resistance or impedance) do not play

an im po rtan t role, because th e induced E MI voltage (for th e susceptibil-

ity case) or the emitted EMI field (for the emission case) is mainly a

function of the EMI driving function (field strength, voltage, or cur-

rent), the geometry and dimensions of the ground loop, and the fre-

quency of th e EMI sig nal.

It should be noted that both the conducted and radiated EMI cou-

pling m ech ani sm s identified above involve a ground loop. However, it


Signal

Reference

Plane

Signal

Reference

Plane

Ij_,

I2 Represent Common

Mode Currents

WftWflMSZL*

(a) Susceptibility Case


Plane

(b) Emission Case

Figure 5.9

Comm on-mode rad iation into and from ground loops.



94 GROUNDING FOR THE CONTRO L OF EMI

should be recognized that ground loop EMI problems can exist without

a physical connection to ground. In particular, at RF frequencies, dis-

tributed capacitance to ground can create a ground loop condition even

though circuits or equipments are floated with respect to ground.

Also,

it should be noted that, for both of th e EMI coupling mecha-

nisms involving the ground loop, the EMI currents in the signal lead

and the return are flowing in the sam e direction. This EM I condition

(where the currents

in

the signal lead and the re tu rn are

in

phase)

is

referred to as common-mode EMI. The EMI control technique s th at will

be effective

for

ground loop problems a re those th at eithe r reduce

the

coupling of EMI into the ground loop or provide suppression of the com-

mon-mode EMI that is coupled into the ground loop.

5.4 Circuit Equipment and System Grounding

In the previous section, EMI coupling mechanisms resulting from cir-

cuit, equipment, and system grounding were identified and discussed.

At this point,

it

should

be

obvious th at grounding

is

very impo rtant

from

the

standpoint

of

minimizing

and

contro lling EM I. H owever,

groun ding is one of th e least und erstood and most significant cu lprits in

many system-level EMI problems. The grounding scheme

of a

system

must perform the following functions:

• Analog, low-level, and low-frequency c ircuits m us t hav e noise-free

dedicated re tu rn s. Due to th e low frequencies involved, wires are

generally used (more or less dictating a single-point or sta r ground

system).

• Analog high-frequency circu its {radio, video, etc.} m us t hav e low-

impe dance, noise-free r et ur n c ircuits, generally in form of plan es or

coaxial cables.

• Re turns of logic circuits, especially high-speed logic, must have low

impedances over the whole bandw idth (dictated by the fastest rise

times), since power and signal returns share the same paths.

• R et ur ns of powerful loads (solenoids, mo tors, lam ps , etc.) shou ld be

distinct from any of th e above, even thou gh th ey m ay end up in th e

same terminal of the power supply regulator.

• Re tur n pa th s to chass is of cable shields, trans form er shields, filters,

etc.

mus t not interfere with functional re tu rn s.

• W hen the electrical reference is distinct from th e chassis ground,

provision a nd accessibility m ust exist to connect and disconnect one

from the other.



CIRCUIT, EQUIPMENT, AND SYSTEM GROUNDING

95

• More generally, for signals tha t communicate within the equipment

or between pa rts of

a

system, the grounding scheme must provide a

common reference w ith minimum ground shift (unless these links

are balanced, optically isolated, etc.). Minimum ground shift means

that the common-mode voltage must stay below the sensitivity

threshold of th e most susceptible device in the link.

All the above constraints can be accommodated if their functional

returns and protective grounds are integrated into a grounding system

hierarchy as shown in Fig. 5.10. The application of this concept is the

subject of the following discussion.

Modern electronic systems seldom have only one ground. To miti-

gate interference, such as due to common-mode impedance coupling, as

many separate grounds as possible are used. Separa te grounds in each

subsystem for structural grounds, signal grounds, shield grounds, and

primary and secondary power grounds are desirable if economically

and logistically practical. These individual grounds from each sub-

system are finally connected by the shortest route back to the system

ground point, where they form an overall system potential reference.

Low-Level,

Low-Frequency

Ground

(//VtomV

de to a few

100

kHz)

V W A

AWfty

Relays, etc ,

Signaling Groun ds

( 5 V t o 5 0 V

de to a few kHz)

Low-Level

High-Frequency

Ground.

Radio Com munication

jN

to mV, kHz to GHz)

Digital Levels,

High-Frequency

Ground.

(Volts, de to 100 MHz)

Lightning, EMP

Ground

(Tens of kA,

de to a few ten s of MHz

DC Power Ground

(Returns for Loads > 1 A)

AC Pow er Safety G round

(50 Hz/60 Hz o r 400 Hz)

Fig ure 5.10 Grounding hierarchy.



96

GROUNDING FOR THE CONTROL OF

EMI

This method is known as a single-point ground and is illustrated in

Fig. 5.11.

5.4.1 Single-Point Grounding Scheme

The single-point or star type of grounding scheme shown in the figure

avoids problems of common-mode impedance coupling discussed in the

previous section. The only common path is in the earth ground (for

earth-based structures), but this usually consists of a substantial con-

ductor of very-low impedance. Thus, as long as no or low ground cur-

rents flow in any low-impedance common paths, all subsystems or

equipments are maintained at essentially the same reference potential.

The problem

of

implementing

the

above single-point grounding

scheme comes about when (1) interconnecting cables are used, espe-

cially ones having cable shields that have lengths on the order of 1/20 of

a wavelength or greater, and (2) parasitic capacitance exists between

subsystem or equipment housings or between subsystems and the

grounds of other subsystems. This situation is illustrated in Fig. 5.12.

Here, cable shields connect some of the subsystems together so that

more than one grounding path from a particular subsystem to the

ground point exists. Unless precautions are taken, common-impedance

ground currents could

flow.

At high frequencies, the parasitic capacitive

reactance represents low-impedance paths, and the bond inductance of

a subsystem-to-ground point results in higher impedances. Thus, again,

common-mode currents may flow or unequal potentials may develop

between subsystems.

Subsystem

(or Equipment)

#1

Subsystem

(or Equipment)

#2

System (or

Groun

J

Subsystem)

d Point

L

§§

/ Earth Jy

I Ground

V

Subsystem

(or Equipment)

#4

Prime Power

Generator

Subsystem

(or Equipment)

#3

Subsystem

(or Equipment)

#N

Fig ure 5.11 Single-point or star grounding arrangem ent.




97

Interconnecting Cable

Subsystem

(or Equipment)

# 1

Subsystem

(or Equipment)

# 2

Parasitic Capacitance

Subsystem

(or Equipment)

# 3

System (or Subsystem)

i

Ground Point

Earth

Ground

Jl

Parasitic

*7 Capacitance *

r

Subsystem

(or Equipment)

# 4

Prime Power

Generator

Subsystem

(or Equipment)

#N

Figure 5.12 Degeneration of single-point ground by interconnecting cables

and parasitic capacitance.

5.4.2 Multipoint Grounding Scheme

Rather than have

an

uncontrolled situation

as

shown

in

Fig. 5.12,

the

other grounding alternative is multipoint grounding as illustrated in

Fig. 5.13. For the example shown in Fig. 5.13, each equipment or sub-

system is bonded as directly as possible to a common low-impedance

Ground Plane

Subsystem

(or Equipment)

# 1

Subsystem

(or Equipment)

# 2

y

Ground Lugs or

Bonds on Unit Frame

Grounds

Subsystem

(or Equipment)

# 4

Subsystem

(or Equipment)

# 3

Interconnecting

^ Cables

O ~ n Earth

V

Ground

Prime Power

Generator

Subsystem

(or Equipment)

#N

Ground Plane

Figu re 5.13 Multipoint grounding system.



98 GROUNDING FOR THE CONTROL OF EMI

gr ound p l ane t o f o r m a homogeneous , l ow- i mpedance pa t h . Thus , com-

mon- m ode cu r r en t s and o t he r EMI p r ob l em s wi l l be mi n i m i zed . Th e

gr ound p l ane t hen i s ea r t hed f o r s a f e t y pu r pos es .

5.4.3 Se lect ion of a Grou nding Schem e

The facts are that a single-point grounding scheme operates better at

low frequencies, and a multipoint ground behaves best at high frequen-

cies. If the overall system, for example, is a network of audio equip-

ment, with many low-level sensors and control circuits behaving as

broadband transient noise sources, then the high-frequency perfor-

mance is irrelevant, since no receptor responds above audio frequency.

For this situation, a single-point ground would be effective. Conversely,

if the overall system were a receiver complex of 30 to 1,000 MHz tuners,

amplifiers, and displays, then low-level, low-frequency performance is

irrelevant. Here, multipoint grounding applies, and interconnecting

coaxial cables should be used .

The above comparison of audio versus VHF/UHF systems makes

clear the selection of the correct approach. The problem then narrows

down to one of defining where low- and high-frequency crossover exists

for any given subsystem or equipment. The answer here in part

involves the highest significant operating frequency of low-level circuits

relative to the physical distance between the farthest located equip-

ments. The determination of the crossover frequency region involves

consideration of (1) magnetic versus electric field coupling problems

and (2) ground-plane impedance problems due to separation. Hybrid

single and multipoint grounding systems are often the best approach

for crossover region applications.

When printed circuits and ICs are used, network proximity is consid-

erably closer. Thus, multipoint grounding is more economical and prac-

tical to produce per card, wafer, or chip. Interconnection of these

components through wafer risers, motherboards, etc. should use a

grounding scheme following the illustrations of previous paragraphs.

This will likely still represent a multipoint or hybrid grounding

approach in which any single-point grounding (for hybrid grounds), if

used, would be to avoid low-frequency ground current loops and/or com-

mon-mode impedance coupling.

In summary, many system-level EMI problems can be avoided by

paying careful attention to the grounding scheme used. Common-mode,

common-ground impedance problems may be reduced by application of

one or more of the following techniques.




99

Elim inate common impedance by using a single point ground

(Fig. 5.11) if possible. This configuration is usually op timal for

power frequencies and signal frequencies below 300 kHz.

Separate

and

isolate g round s

on the

basis

of

signal type, level,

and

frequency

as

il lustrated

in

Fig.

5.10.

Minimize ground impedance

as

il lustrated

in

Fig. 5.14

by

using

ground bus, ground plane,

or

ground grid.

Float circuits

or

equipments

if

practica l from

a

safety standpoint

as

i l lustrated

in

Fig. 5.15. The effectiveness

of

floating circu its

or

equipments depends on their physical isolation from other conduc-

tors. In

large facilities,

it is

difficult

to

achieve

a

floating system.

Use an inductor or capacitor in the ground connection to provide

high-

or

low-frequency isolation, respectively,

as

il lustrated

in

Figs. 5.16 and 5.17.

Daisy Chaining (Poor)

Heavier Ground Path (Better) or: Parallel Ground Wires (Better)

T \ \

Ground Plane (Better Still)

Ground Grid (Better Still)

T T T

Figure 5.14 Means of decreasing common-impedance coupling by decreasing

ground path impedance. From the bad practice of daisy-chain (top), the

improvement evolves toward a plane (left) or a grid (right).



100


Box 1 Box 2

Safety Bus

y

(a ) Float Equipment Enclosures

Box

1

Box 2

(b) Float Circuits and Boards

Fig ure 5.15 Float circuits or equipments.

F ig ure 5.16 Capacitive grounding.

• Use filters or ferrites in ground loops to lim it comm on-mode cur-

ren ts or provide a common-mode voltage drop.

• Use a common-mode choke as ill us tra ted in Fig. 5.18 or a common-

mode isolation transformer as illustrated in Fig. 5.19 to suppress

ground-loop EMI. These devices may provide on the order of 60 dB




101

F ig u re 5.17 Inductive grounding.

RF Choke

O

Figure 5.18

Comm on-mode chokes.

Primary

B

Case

C

Victim

Secondary

-D

Or Green

Wire

Ordinary

Isolation

Transformer

A Parasi t ic

> 1 Cap

Victim

D

F ig u re 5.19 Common-mode isolation transformer.



102

GROUND ING FOR THE CONTR OL OF EMI

of common-mode rejection a t frequencies u p to several h un dr ed

kilohertz.

• Use optical isolators and /or fiber optics to block common-mode EM I

effects as illustrated in Fig. 5.20. Optical isolators provide a high

degree of common-mode rejection at frequencies up to and including

th e H F ban d (i.e., 3 to 30 MH z). Optical isolators a re us ually lim-

ited to digital app lications (they a re not applicable to low-level an a-

log circuits ).

• Use bala nce d circuits to min imize effects of common-m ode EM I in

th e grou nd loop as illu stra ted in Fig. 5.21. W ith a perfectly b al-

anced circuit, the c urr en ts flowing in the two pa rts of th e circuit

will produce eq ual an d opposite voltages across the load, so th e

resulting voltage across the load is zero. Balanced circuits can pro-

vide significant (greater than 20 dB) common-mode reduction for

low-frequency conditions. However, at higher frequencies (above

30 MHz), other effects start to predominate, and the effectiveness of

balanced circuits dim inishes.

Common-mode radiated EMI effects resulting from emissions that

are ra dia ted or picked up by a ground loop may be reduced by the appli-

cation of one or more of th e following tech niq ues:

• Minimize the common-mode ground loop are a by rou ting intercon-

necting w ires or cable close to the ground.

• Reduce the common-mode ground loop cu rre nts by floating circuits

or equipments; using optical isolators; or inserting common-mode

filters, chokes, or isolation tran sform ers.

• Use balan ced circuits or balan ced drivers and receivers.

-A/W

LED

Photo

Detector

O

F ig u re 5.20 Use of optical isolation to combat common-mode impe dance.



GROUND SYSTEM CONFIGURATIONS

103

-

I

Balanced ~

Signal Source

Common-Mode

Noise Source

F ig u re 5.21 Balanced configuration with respect to common-mode voltage.

5.5 Ground System Configurations

The ground system for a collection of circuits within a system or facility

can assume any one of several different configurations. Each of these

configurations tends to be optimal under certain conditions and may

contribute to EMI problems under other conditions. In general, the

ground configurations will involve either a floating ground, a single-

point ground, a multipoint ground, or some hybrid combination of

these.

A floating ground configuration is illus tra ted in Fig. 5.22. Th is type of

signal ground system is electrically isolated from the ground and other

conductive objects. Hence, noise currents present in the ground system

will not be conductively coupled to the signal circuits. The floating

ground system concept is also employed in equipment design to isolate

signal retu rns from equipm ent cabinets and thu s prevent unw anted cur-

ren ts in ca binets from coupling directly to signal circuits.

Effectiveness of floating ground systems depends on their true isola-

tion from other nearby conductors; floating ground systems must really

float. In large facilities, it is often difficult to achieve and maintain an

effective floating system. Such a floating system is most practical if a

few circuits or a few pieces of equipment are involved and power is

applied from eithe r batt erie s or dc-to-dc converters.



104

GROUN DING FOR THE CONTR OL OF EMI

Equipment

T

Structure or Other Grounded Objects

F ig u re 5.22 Floating Signal Ground.

A single-point ground for an equipment complex is illustrated in

Fig. 5.23. With this configuration, the signal circuits are referenced to a

single point, and this single point is then connected to the facility

ground. The ideal single-point signal ground network is one in which

separate ground conductors extend from one point on the facility

ground to the return side of each of the numerous circuits located

throughout a facility. This type of ground network requires an

extremely large number of conductors and is not generally economically

feasible. In lieu of the ideal, various degrees of approximation to single-

point grounding are employed.

Equipment

Structure or other Grounded Objects

F ig u re 5.23 Single-point signal ground.




105

The configuration illustrated in Fig. 5.24 represents a ground bus

arrangement that is often used to provide an approximation to the sin-

gle-point grounding concept. The ground bus system illustrated in

Fig. 5.24 assumes the form of a tree. Within each system, the individual

subsystems are single-point grounded. Each of the system ground

points is then connected to the tree ground bus with a single insulated

conductor.

The single-point ground accomplishes each of the three functions of

signal circuit grounding. That is, a signal reference is established in

each unit or piece of equipment, and these individual references are

connected together. These, in turn, are connected to the facility ground

at least at one point, which provides fault protection for the circuits and

provides control over static charge buildup.

An important advantage of the single-point configuration is that it

helps control conductively coupled interference. As illustrated in

Fig. 5.23, closed paths for noise currents in the signal ground network

are avoided, and the interference currents, or voltages in the facility

ground system, are not conductively coupled into the signal circuits via

the signal ground network. Therefore, the single-point signal ground

network minimizes the effects of any noise currents th at may be flowing

in the facility ground.

In a large installation, a major d isadvantage of a single-point ground

configuration is the requirement for long conductors. In addition to

System

C

,**' ,' System

A

/ System B

Subsystem

LO

A

Subsystem

B

Subsystem

C

Earth Ground

Fig ure 5.24 Single-point ground bus system using a common bus.



106

GROUNDING FOR THE CONTR OL OF EMI

being expensive, long conductors prevent realization of a satisfactory

reference for higher frequencies because of large self-impedances. Fur-

thermore, because of stray capacitance between conductors, single-

point grounding essentially ceases to exist as the signal frequency is

increased. In general, for typical equipments, systems, or facilities, sin-

gle-point grounds tend to be optimum for frequencies below approxi-

mately 300 kHz.

The multiple-point ground illustrated in Fig. 5.25 is the third config-

uration frequently used for signal ground networks. This configuration

establishes many conductive path s to various electronic systems or sub-

systems within a facility. Within each subsystem, circuits and networks

have multiple connections to this ground network. Thus, in a facility,

numerous parallel paths exist between any two points in the m ultiple-

point ground network.

Multiple-point grounding frequently simplifies circuit construction

inside complex equipment. It permits equipment employing coaxial

cables to be interfaced more easily, since the outer conductor of the

coaxial cable does not have to be floated relative to the equipment cabi-

net or enclosure.

However, multiple-point grounding suffers from an important disad-

vantage. Power currents and other high-amplitude, low-frequency cur-

rents flowing through the facility ground system can conductively

couple into signal circuits to create intolerable interference in suscepti-

ble low-frequency circuits. Also, multiple ground loops are created, and

this makes it more difficult to control radiated emission or susceptibil-

ity resulting from the common-mode ground loop effects. In addition,

Equipment

Facility Ground

Fig ure 5.25 Multiple-point ground configuration.




107

for multiple-point grounding to be effective, all ground conductors

between the separate points must be less than 0.1 wavelength of the

interference signal. Otherwise, common-ground impedance and ground-

radiated effects will become significant. In general, multiple-point

groun ding configurations tend to be optim um at high er frequencies (i.e.,

above 30 MHz).

To illustrate one form of a hybrid-ground system, Fig. 5.26 shows a

19-in cabinet rack containing five separate sliding drawers. Each

drawer contains a portion of the system (top to bottom): (1) RF and IF

preamp circuitry for reception of microwave signals, (2) IF and video

Single-Point Pow er Line

Gnd Return (SPPL&GR)

Earth

K

—Sk

IF

Ampl.

1

Log

IF

Ampl.

Demod-

ulator

Video

Ampl.

t

Multipoint Ground Plane

Display

Drawer

Recorders

Audio

Driver

Singlepoint Ground Plane

Ground Distribution Block

Gnd

Fig ure 5.26 Grounding arrangement used in cabinet racks.



108

GROUNDING FOR THE CONTROL OF EMI

signal amplifiers, (3) display drivers, displays, and control circuits, (4)

low-level audio circuits and recorders

for

documenting sensitive multi-

channel, hard-line telemetry sensor outputs, and (5) secondary and reg-

ulated power supplies. The hybrid aspect results from:

• The RF and IF video drawers are similar. Here, unit-level boxes or

stages (interconnecting coaxial cables are grounded

at

both ends)

are multipoint grounded to the drawer-chassis ground plane. The

chassis is then grounded to the dagger pin, chassis ground bus

as

suggested in Fig. 5.27. The power ground to these drawers, on the

other hand, is using a single-point ground from its bus in a manner

identical to the audio drawer.

Insulator

Antenna Jack

Computer [~

Clock Inpu t L

Low-Level RF Circuits

&

IF

^ _ Preamp

To Power Gnd

To Signal Gnd

RF-IF Coax C ables

IF Amplifiers, BP F ilters

Demodulators, Video Ampl

,1 Video Cables, Coax or Twisted

^T~ Shielded P air

Display Drivers and

Readout Circuitry

Multiplex Input [~~

Sensor Jack L

Low-Level Audio

Sensor Circuits

Display

Secondary

Regulated

Power Supplies

Ground Distribution Block

Figure 5.27 Block diagram detail of hybrid grounding arrangement.



EMI

CONTR OL DEVICES AND TECHNIQUES

109

• The chassis or signal ground and power ground busses each consti-

tute

a

multipoint grounding scheme to the drawer level. The indi-

vidual ground busses are single-point grounded at the bottom

ground distribution block. This avoids circulating common-mode

current between chassis or signal ground and power grounds, since

power ground current can vary due to trans ient surges in certain

modes of equipment operation.

• Interconnecting cables between different drawer levels are run sep-

arately, and their shields, when used, are treated in the same

grounding manner as at the drawer level.

• The audio and display drawers shown in Fig. 5.27 use single-point

grounding throughout for both their unit-level boxes (interconnect-

ing twisted cable is grounded at one end to its unit) and power

leads.

Cable and unit shields are all grounded together

at

the com-

mon dagger pin bus. Similarly, the outgoing power leads and

twisted retu rns are separately bonded on their dagger pin busses.

To review the above scheme, the following is observed:

• The audio and display drawers have ground runs of about 0.6 m

and

an

upper frequency of operation of about 1 MHz (driver and

sweep circuits). Thus, single-point grounding to the strike pins

is

indicated.

• The RF and IF drawers process UHF and 30 MHz signals over

a

distance of a meter so that multipoint grounding is indicated.

• The regulated power supplies furnish equipment units having tran-

sient surge demands. The longest length is about 1.5 m, and signifi-

cant transient frequency components may extend up in the HF

region. Here, hybrid grounding is indicated: single-point within

a

drawer and multipoint from the power bus to all drawers.

5.6 EMI Control Devices and Techniques

The performance of

some

EMI control techniques or devices may be sig-

nificantly influenced by grounding. In particular, cable shields; isola-

tion transformers; EMI filters; ESD, lightning, and EMF protection

techniques;

and

Faraday shields must

be

properly grounded

so as to

provide maximum

EMI

protection.

A

detailed discussion

of

specific

grounding considerations associated with these EMI control techniques

or devices is beyond the scope of this book. However, it is important to

emphasize the importance of grounding on the performance of these

techniques or devices, and details may be found in the references.



11 o GROUND ING FOR THE CONTRO L OF

EMI

Suggested Readings: Grounding

[1] Morrison, Ralph, an d W.

H.

Lewis,

Grounding

and

Shielding

in

Facilities, Hoboken, NJ : Jo hn W iley & Sons, 1990.

[1] Morrison, Ralph,

Grounding

and

Shielding Techniques

in

Instru-

mentation, 3rd ed., Hoboken, N J: Joh n Wiley & Sons, 1990.

[1] Denny, Hugh W.,

Grounding for

the

Control

of

EMI,

Gainesville,

VA, Interfe rence Con trol Technologies, Inc.

[1]

Groun ding, Bonding an d Shielding for Electronic Equ ipmen t and

Facilities, MIL-HDBK-419.



Chapter 6

Shielding Theory, Materials,

and Protection Techniques

Shielding

is a

major means

of

EMI control

at all

levels

of

EMC, viz.,

component; chassis or black box; equipment; subsystem; system; and

entire vehicular

or

housing structures, such

as

ships, aircraft,

and

buildings. This chapter presents shielding theory, shielding materials,

and some mathematical models of shielding effectiveness.

The performance of shields is a function of whether the source

appears as an electric or magnetic field in the near-in induction region

or an electromagnetic field in the far-field region. These considerations

are a function of both the source and receptor geometry separation and

frequency of operation. Consequently, it is pertinent to first establish

criteria for near and far fields as a function of these param eters.

6.1 Field Theory

The purpose of this section is to present some pragmatic relations about

magnetic, electric, and electromagnetic fields

as

pertinent background

to understanding and applying shielding criteria. The literature con-

tains excellent discussions of Maxwell's equations and field theory.

Therefore, only a few aspects are presented here .

The electric (E

e

, E

r

) and magnetic

H^)

fields existing about an oscil-

lating doublet (or circuit), exhibiting high impedance and oriented as in

Fig. 6.1, are obtained from applying Maxwell's equations:

E, . - ^ ^ ( A j

c o s v

- ( _ L )

s i l l v +

(_L)

cosv

] (6. )

111



112 SHIELDING THEORY, MATERIALS, AND PROTECTION TECHNIQUES

F ig u re 6.1 Fields from a vertical doublet.

2Z

0

ID7i cos6

-—] simp

2nr)

Y

J

(6.2)

H

A

=

2mJ

2n r

(6.3)

where,

Z

o

= free-space impedance (for r »

X/2n

= J\i/e =

120TC

= 377 Q)

I = current in short wire (doublet)

D = leng th of sho rt wire (doublet) in which D « X

0 = zenith angle to rad ial distance r

X = wav elength co rresponding to frequency, f

= c/X

r = distance from sh ort wire doublet to m easu ring or observation

point

\|/ = 2nr/X - cot

co = ra d ia l frequency = 2rcf

t = tim e = 1/f

c = 1 J\LE = 3 x 10

8

m/sec

1. The electric and m agnetic field comp onents con tain term s th at

involve

X/2nr.

For m ost conditions, the E

r

term will be small rela tive

to the E

0

term, and it is usually considered to be insignificant.

Thus, E

r

will not be considered further.

2. When the multiplier, X/2nr, equ als 1 in th e electric-field an d mag-

netic-field terms, all coefficients of either the sin or cos are unity

and eq ual. Thus, r =

X/2n

(about one sixth wav elength) corresponds



SHIELDING THEORY

113

to the transition-field condition or boundary between the near field

(first term of both equations) and far field (last term).

3. When r » X/2n (far-field conditions), only the last term of each

equa tion is significant. For this condition, th e wave impeda nce Z

o

=

EQ/H^ = 377 Q. This

is

called the radiation field (plane waves), and

both E

e

and H^ are in time phase , although in directional quadra-

tu re .

4. When r «

A/2TE

(near-field conditions), only the first term of each

equation is significant. For this condition, the wave impedance, E

e

/

H^ = Z

o

A/27ir. Note that the wave impedance is now » Z

o

. This is

sometimes called simply an electric field or a high-impedance field

i.e., high relative

to a

plane-wave impedance.

It is

also the induc-

tion field, and E

e

and H^ are in both tim e p hase and directional

quadra ture .

5.

If th e oscillating source had been low impe dance, th e electric and

magnetic field equations would be similar to the ones given above

except that the first term in Eq. 6.1 would vanish, and a similar

first ter m would have to be added to Eq. 6.3. For this condition, t he

wave impedance in th e ne ar field

EQ/H^

= Z

o

2nr/X. This is some-

times called

a

magnetic field

or a

low-impedance field

(i.e.,

low

impedance relative to Z

o

, the p lane wave radiation impedance.

Figure 6.2 illustrates conceptually the fourth and fifth conditions in

the ne ar or induction field. Situatio n (a) is a monopole, strai gh t w ire, or

circuit in which the RF current is low. Consequently, the source imped-

ance = V/I is a high impedance. The wave impedance near in is also

high, being made up predom inantly of th e electric field. The electric

field attenuates more rapidly (1/r

3

) with an increase in distance tha n

th e m agnetic field (1/r

2

) in the inductio n reg ion [cf. Eqs. (6.1) and

(6.2)].

Thus, the w ave impedance decreases with distance w here it asymp toti-

cally approaches Z

o

= 377 Q in the far or radia tion field. The converse

applies for s ituatio n (b), wh erein a low-impedance source cre ates a low-

impedance wave of predominantly the magnetic-field component. This

impedance increases with distance where it asymptotically approaches

377 Q in the far field. Figure 6.3 illustrates these impedances of both

fields as a function of distanc e, r.

6.2 Shielding Theory

Shielding provided by a metallic barrier can be analyzed from either of

two viewpoints: (1) tha t of field or wave theo ry or (2) th a t of circuit th e-

ory. In the circuit-theory approach , c urr en ts from the interference



114

SHIELDING THEORY, MATERIALS, AND PROTECTION TECHNIQUES

Monopole

o

Low Current Corresponds

to High Impedence

High Current C orresponds

to Low Impedance

y

Loop

H ,

(a ) H igh-Impedance, Electric-

Field Source and Wave

(b ) Low-Impedance, M agnetic-

Field Source and Wave

F ig u re 6.2 Conceptual illustration of field intensities vs. source type and dis-

tance.

10K

3K

IK

300

i

l

100

30

10

»»

—

1

I

\ l

^*

i

e;

k

?

***

'ft

5a

jr

cCJ

7*"

Held

i0r ii Pi

>

K -

A'

1

^.j

.

fi

(

>r

'If

•^

i-r

» EX-

I

^ ^

, s

'A

i

_

;

1

Btrt

--

/

h~

y

-»

E and

am

Kac

ir

¥

t i

16

01

1

i

• \ \

0

el

3

r

rx

t

1

—

.2

.3 .4 .5 .7 1 2

Distance from Source in units of r =

4 5

F ig u re 6.3 Wave impedance as a function of source distance.

source induce currents in the shield such that the associated external

fields due to both curr en ts a re out of pha se a nd t end to cancel. Since the

field-theory approa ch is more widely adopted in th e lite rat ure , however,

it will be used in the re m ain der of thi s discussion.

Figu re 6.4 depicts the p heno me na of both reflection and tran sm issio n

that are utilized in removing energy from an incident wave (plane-wave

example shown). If an incident plane wave encounters a barrier to its



SHIELDING THEORY

115

Inside of Enclosure

Transmitting Wave

B ) E

y

Outside World

Barrier of Finite

Thickness

Fig ure 6.4 Representation of shielding phenomena for plane waves.

passage, at region A of the interface, both reflection and transmission

occur. The amplitudes of these two portions of the original wave depend

on the surface impedance of the barrier material with respect to the

impedance of the wave. Since the reflected wave is not proceeding in a

direction that contributes to the surviving wave on the far side of the

barrier, this is considered a loss mechanism.

The transmitted portion of the incident wave, continuing on in

approximately the same direction after penetrating the interface, expe-

riences absorption while traversing the finite thickness of the barrier.

At the second barrier interface B of Fig. 6.4, reflection and transmission

phenomena again occur. The transmitted portion is the amount of

energy that traversed the first interface less the energy absorbed in tra-

versing the barrier and that reflected at B. The second reflection con-

tribu tes an insignificant amount in the removal of energy and is usually

neglected.

At plane-wave (far-field) frequencies, the shielding effectiveness of a

barrier in reducing the energy of an electromagnetic field can be readily

computed. Each of the contributing factors discussed above is computed

separately, and then their total contribution is summarized. This is

accomplished in the following manner for expressing shielding effec-

tiveness in dB, S Q:

S

d B ~

(6.4)



116 SHIELDING THEORY, MATERIALS,

AND

PROTECTION TECHNIQUES

where,

R

d B

= reflection loss in dB

i = transm ission or absorption loss in dB

i

= internal reflection loss a t exiting interface in dB (usu ally

neglected)

The shielding effectiveness

to

electric

or

electrom agnetic fields

may

also be measured in te rms of the fraction of the impinging field th at

exists a t th e othe r side of th e bar rier:

S

d B

=

20 Iog

10

[ — )

(6.5)

where,

E = impin ging field inten sity in V/m

E

2

= exiting field inte nsi ty in V/m

The individual contributing factors to the shielding effectiveness in

Eq. (6.4) are separately computed in the next sections.

6.2.1 Absorption Loss

The absorption loss, A Q, is ind epe nde nt of th e type of wave impin ging

on the shield and is expressed as follows:

A

d B

= 3.34xlO~

3

tVfGJI = 3.34t^f

MH z

G|Li dB (6.6)

where,

A= atten uation in dB

t = thickne ss of ba rrie r in m ils (unit of 0.001 in)

f = frequency in Hz

fMHz

=

frequency in MHz

G = conductivity rela tive to copper

\i = permeab ility relativ e to copper

Equation (6.6)

is

plotted in Fig. 6.5 for the p ara m ete rs copper (G =

1,

ja = 1), iron (G = 0.17, (I = 1000), an d hy pern ick (G = 0.6, \x = 80,000).

Absorption loss is the dep end ent v ariable , and frequency is the inde-

pendent variable, with thickness in mils as a second parameter. It is

noted that the brute-force appro ach of using a thick sheet (1/8 in) of

iron at low frequencies (e.g., at 60 Hz) results in a significant absorp-



SHIELDING THEORY

117

30Hz 300Hz 3kHz 30kHz 300kHz 3MHz 30MHz 300MHz

U se B-Factor f or this Region:

= 3dB, B = - 2 d B

= 4dB,B=

OdB

lOHz lOOHz

1kHz

10kHz 100kHz

1MHz

10MHz 100MHz lGHz

Radio Frequency

Fig ure 6.5 Shielding absorption (penetration/attenuation) loss vs. radio fre-

quency, material, and thickness (independent of wave impedance).

tion loss (approx. 45 dB). On the othe r han d, a th in she et (e.g., 1 mil) of

copper at 1 GHz yields significant (>100 dB) absorption loss. Th is illus-

trates the difficulty of achieving a significant absorption loss at E LF in

contrast to UHF.

The internal reflection loss, B, in Eq. (6.4) is negligible wh en A

d b

is

greater t ha n about 4 dB. When A

df

is not greater th an 4 dB, B

d B

is neg-

ative, since

it

is

a

coherent term, which would have made E

2

in Eq. (6.5)

larger. The valu e of B^g is shown in th e lower righ t corner of Fig. 6.5.

6.2.2 Reflection Loss

Reflection loss, R^g, is represented by forming the ratio of the wave

impedance, Z

w

to the surface impedance of the barrier material, Z^.

R

dB

= 20 log,

K+ir

3

4K

= 20

*•• £ •

for K > 10

(6.7)

Equation (6.7) indicates that if either the wave impedance is high

(e.g., electric field) and/or the bar rier surface impedance is low (e.g.,



118


copper), the loss will be substantial. Conversely, if the wave impedance

is

low

(e.g., magnetic field) and/or

the

barrier impedance

is

relatively

high (e.g., iron),

the

reflection loss will

be

significantly less. Each

of

these situations is now discussed

in

further detail.

6.2.3 Reflection Loss to Plane Waves

The reflection loss of a plane wave, R^B*

m a

Y also be calcu lated from:

R

dB

=

101og

10

(G/^f

M H z

)dB

(6.8)

Equation

(6.8) is

plotted

in

Fig.

6.6 for

copper, iron,

and

hypernick.

Compared with absorption loss,

the

figure ndicates that the reflection

loss of plane waves at low frequencies is the major attenuation mecha-

nism. High-conductivity (G), low-permeability (\x) m aterial is more

effective

in

establishing reflection loss, since

the

barrier surface

impedance is lower with regard to that of a plane wave where Z

w

=

377 £2, and the ratio of the latte r to the former (the loss mechanism)

is

greater

[cf. Eq.

(6.7)].

At

UHF,

the

reflection loss becomes less effec-

tive,

since

the

barrier skin depth decreases (surface resistivity

increases),

and the

barrier impedance increases, resulting

in a

smaller ratio of plane wave to barrier impedance. In comparing Figs.

6.5 and 6.6, note tha t, at UHF, the absorption loss becomes the more

significant loss mechanism of the two.

200

« 150


100

50

• —

- — ^

• —

1

—«

1

— . .

1

— .

• * • * —

^ " - HI

• * •

' *

^

.

- .

Jrpn

|

er |

—f^ <

i

—-^

— «»^

• —

— -

-

- - -

— .

• — ^

•—«.

• —

* * —

— - * .

10Hz lOOHz 1kHz 10kHz 100kHz 1MHz 10MHz 100MHz lGHz

Radio Frequency

Valid for Thickness > 3 £

S=

Skin Depth

F i g u r e 6.6 Reflection loss of pla ne wav es vs. rad io frequency.



SHIELDING THEORY 119

6.2.4 Reflection Loss to Electr ic and M agnetic Fields

When there is a substantial difference in the impedance of the incident

wave and the shielding barrier, reflection at the boundary is significant

and good shielding is obtained. The high impedance wave in the near

field is known as a n electric-field wave, and its reflection loss is:

(6.9)

where r = the distance from source to barrier in inches; the other terms

are as defined under Eq. (6.6).

Equation (6.9) is plotted in Fig. 6.7 for the parameters of separation

distances, r, of 1 in, 1 m (3.3 ft), and 30 m (100 ft) and for copper and

iron materials. As before, frequency is the independent variable, and

reflection loss, R^b, is the dependent variable. The above distance

parameter covers a range of 1200 or about 62 dB difference in reflection

loss,

whereas the G/|i range for copper to iron is about -38 dB.

Figure 6.7 shows that the reflection loss of an electric field decreases

with frequency until the separation distance becomes

XI2n,

whence far-

field conditions prevail. Thus, Eq. (6.9) applies until the losses meet

th at of Eq. (6.8), th e plane-wav e losses. Thereafter, th e two me rge. For

300

100Hz

10kHz 1MHz 100MHz

lOHz 1kHz

100kHz

Frequency

10MHz

lGHz

F i g u r e 6.7 Reflection loss of electric fields vs. rad io frequency.



120

SHIELDING

THEORY,

MATERIALS,

AND PROTECTION TECHNIQUES

this reason,

the

plane wave reflection losses

are

also shown

as a

refer-

ence

in

Fig.

6.7 and are

identical

to

those previously shown

in

Fig.

6.5.

For low-impedance or magnetic-field waves, the reflection loss is:

R

dB

= 20

Iog

10

[(0.462/r)

+ 0.136rV(Gf)/|i + 0.354]

dB

(6.10)

Equation (6.10) is plotted in Fig. 6.8 for the parameters of separation

distance,

r, of

1

in,

1

m

(3.3

ft), and 30 m

(100

ft) and for

copper

and

iron

materials.

The

reflection loss

to

iron

(1 in

separation) approaches

0 dB

at about

30 kHz,

when

the

magnetic-field wave impedance approxi-

mates that

of

th e barr ier im pedanc e [loss

= 0 dB

from

Eq.

(6.7)]. Below

30

kHz, the

wave impedance

is

less than

the

barrier impedance,

and

the loss again increases.

The

reflection loss

of a

m agnetic field shown

in

the figure increases with frequency until the source-to-barrier separa-

tion distance

is

about

1/2,

whence

the

plane-wave losses

of Fig. 6.6

again prevail.

In comparing Figs.

6.7 and

6.8,

it is

noted that reflection-loss shield-

ing

for

providing

a

reduction

in

absolute field intensity

to

magnetic

fields at

low

frequencies

is

distinctly different from that

for

electric

fields. Magnetic fields

are

shielded

at de and ELF

only

by

providing

a

low-reluctance path

as an

alternative

for the

incident mag netic field.

100Hz

10kHz 1MHz

100MHz

lOHz

1kHz

100kHz

Frequency

10MHz lGHz

F i g u r e 6.8 Reflection loss of magnetic fields vs. radio frequency.



SHIELDING THEORY

121

Figure 6.9 depicts a simple representation of a uniform magnetic

field existing in free space. The vertical lines show the direction of the

orientation of the magnetic-field vector throughout the two dimensions.

Figure 6.10 shows the effect on the field lines by including

a

hollow per-

meable object in this uniform magnetic field. The field-intensity ines

enter the object at an angle of 90° to its surface. In the interior of this

hollow object, the field ntensity lines are less intense than in the sur-

rounding free-space medium. However, these magnetic field lines

in

the solid barrier are much more intense th an in either the hollow center

or the exterior of the barrier. This effect is due to the relative higher

reluctances of free space, both surrounding the barrier and in the inte-

rior, versus that of the barrier itself. The lower reluctance of this bar-

rier divides the field-intensity ines, thus reducing the intensity of the

Air

Figure 6.9 Uniform magnetic field.

High Permeability ju

>>

1)

Material Offering Low

3

Reluctance Path ' . .

Air

A i r I

I

Magnetic Field G reatly

Reduced Inside to jut/s

of Outside

Figure 6.10 Cross section of a hollow rectangular solid of high permeability

in uniform field.

* The magnetic field in the inside is about \xtfs of the value on the outside, where \i is the

relative permeability,

t is the

thickness,

and s is the

dimension of one side.




absolute magnetic field

in

the interior of the enclosure to yield

a

shield-

ing effect. This effect is quite pronounced at de, where shielding effec-

tiveness values

in

excess

of 50 dB

have been achieved through

the

utilization

of

extremely high-permeability materials configured

on a

double-barrier enclosure.

6.2.5 Composite Absorption and Reflection Loss

When either

Eqs. (6.6)

through (6.10)

or

Figs.

6.5

through

6.8 are

combined,

the

overall attenuation

or

shielding effectiveness given

in

Eq. (6.4) results. These relationships are plotted in Fig. 6.11. Since

there

are

many variables,

the

composite curves represent

the

param-

eters of copper and iron materials having a thickness of one mil and

1/32 in;

electric

and

magnetic fields

and

plane-wave sources;

and a

source-to-barrier distance

of 1 in

and 1 m (3.3 ft). Except

for

L-F mag-

netic fields, the figure shows that reflection loss is the principal

attenuation mechanism

at low

frequencies, whereas absorption loss

is

the

main mechanism

at

H-F. Figure 6.11

is but one of a

family

of

mathematical models that define shielding attenuation. Other mod-


lOHz lOOHz 1kHz 10kHz 100kHz 1MHz 10MHz 100MHz lGHz

Frequency

F ig u r e 6.11 Total shielding effectiveness vs. frequency for electric and mag-

netic fields and plane waves.



SHIELDING MATERIALS

123

els would reflect different materials, thickness, and emission source

distances.

6.3

Shielding Materials

Good shielding efficiency for electric (high-impedance) fields is obtained

by use of materials of high conductivity, such as copper and aluminum.

As shown in Eq. (6.9) and Fig. 6.7, the shielding effectiveness for elec-

tric fields is infinite at de and decreases with an increase in frequency.

However, magnetic fields [Eq. (6.10)] are more difficult to shield, since

the reflection loss may approach zero for certain combinations of mate-

rial and frequency. With decreasing frequency, the magnetic field reflec-

tion and absorption losses of nonm agnetic m aterials such as a lum inum

decrease. Consequently, it is difficult to shield against magnetic fields

using nonmagnetic materials. At high frequencies, the shielding effi-

ciency is good due to both reflection and abso rption losses, so th e choice

of m aterials becomes less imp ortant.

Regarding plane waves, magnetic materials provide better absorp-

tion loss (Fig. 6.5), whereas good conductors provide better reflection

loss (Fig. 6.6). These and the above relations are summarized qualita-

tively in T able 6.1.

Table 6.1

Materials

Summary of Shielding Effectiveness of Permeable and Nonpermeable

Permeable

mater ia ls

Magnetic

(H >

1000)

Nonmagnetic

(n = i)

Frequency

Low:

< l k H z

Medium:

1-100 kHz

High:

> 100 kHz

Low:

< l k H z

Medium:

1-100 kHz

High:

> 100 kHz

Absorption

loss A

d B

,

all fields

B ad

Good

Excellent

Fail

B ad

Good

Assumptions:

Material

thickness:

1/32 in

Source distance: 10 ft (3 m)

Radio frequency: as shown

Reflection loss, R^ B

Electric

fields

Excellent

Good

Fair

Excellent

Excellent

Good

Magnetic

fields

Fail

B a d

Poor

B ad

Poor

Fair

Plane

w a v e s

Good

Fair

Fair

Good

Good

Fair

Attenuation scores:

Excellent: > 150 dB Poor: 30-50 dB

Good: 100-150 dB Bad: 10-30 dB

Fair: 50-100

dB Fail: <10 dB



124


Table 6.2 summarizes the absorption loss of a number of different

materials that, in one form or another, may be used for shielding. The

loss is given in decibels per mil thickness of the metal. The high-perme-

ability

(JJ,

> 80,000) materials shown are especially interesting for their

low-frequency, magnetic-field shielding properties. However, they are

prone to saturation at lower field densities, and they require careful

handling procedures.

Table 6.2

Ch aracteristic s of M etals Used for Shielding

Metal

Silver

Copper, annealed

Copper, hard drawn

Gold

Aluminum

Magnesium

Zinc

Brass

Cadmium

Nickel

Bronze

Iron

T in

Steel (SAE 1045)

Beryllium

Lead

Hypernom®

Monel

Mumetall®

Permalloy

Stainless steel

Conductivity

relat ive

to copper

1.05

1.00

0.97

0.70

0.61

0.38

0.29

0.26

0.23

0.20

0.18

0.17

0.15

0.10

0.10

0.08

0.06

0.04

0.03

0.03

0.02

Relative

permeabil ity

(100 kHz)

1

1

1

1

1

1

1

1

1

1

1

1,000

1

1,000

1

1

80,000

1

80,000

80,000

« 1

Absorption loss in

dB per mil (0.0001 in)

100 Hz

0.03

0.03

0.03

0.03

0.03

0.02

0.02

0.02

0.02

0.01

0.01

0.44

0.01

0.33

0.01

0.01

2.28

0.01

1.63

1.63

0.15

10

kHz

0.34

0.33

0.32

0.28

0.26

0.20

0.17

0.17

0.16

0.15

0.14

4.36

0.13

3.32

0.11

0.09

22.8

0.07

16.3

16.3

1.47

1MH z

3.40

3.33

3.25

2.78

2.60

2.04

1.70

1.70

1.60

1.49

1.42

43.60

1.29

33.20

1.06

0.93

228.00

0.67

163.00

163.00

14.70

It is often assumed that most materials that have adequate struc-

tural rigidity will also possess sufficient thickness to provide satisfac-

tory shielding efficiency. This is not generally true for equipments

operated in the audio-frequency region. At these low frequencies, it is

necessary to use a high-permeability material such as Hypernom,

MuMetal®, or Netic® or Co-Netic® foil to provide satisfactory shielding

efficiency to magnetic fields.

While the above equations and figures show a theoretical value of

shielding efficiency from magnetic materials that is quite high, in prac-

tice,

such levels are seldom achieved, particularly at low frequencies

where the required thickness is substantial. Some of the best results

have been obtained by the use of multiple permalloy sheets or the Netic



EMI

SHIELD COMPARTMENTS AND EQUIPMENTS

125

and Co-Netic sandwich foils. These latter products

are

available

in a

variety of ready mad e forms a nd sizes to

fit

diverse ap plications.

Illustrative Example 6.1

A sensitive parallel-T amplifier tuned

to 120 H z is to be located about

1 m away from a 60-Hz am plidyne. By meas urem ent, the magnetic flux

density, B, from the amplidyne

at a

1 m distance

at

its second harm onic

is 180 dBpT or 10 gauss (10~

3

weber/m

2

). The cable feeding the tuned

amplifier is 16 in (0.4 m) long and is equivalent to a conductor separa-

tion of 0.1

in

(0.0025 m). De term ine the induced voltage a nd specify th e

magnetic shield required to protect the 1 JLIV amplifier sensitivity, if

nec-

essary.

The cable loop are a is A = lw = 0.4 m x 0.0025 m = 10"

3

m

2

. The mag-

ne tic flux, <>, cro ssing t he cable loop is BA = 10~

3

weber/m

2

x 10~

3

m

2

=

10~

6

we bers. The induced voltage, V, is:

V = - - $ = - —(10~ webers x coscot)

at at

=

|colO"

6

sin cot| volts =

In X

120 Hz x 10"

6

= 750 ^V (58 dBjuV)

Since the induced voltage is 58 dB above the 1 ^V amplifier sensitiv-

ity, about 60 dB of m agn etic sh ielding of th e cable is req uire d at 120 Hz.

At this frequency, from Fig. 6.10,

a

1/32-in iron sheet offers about 15 dB

attenuation, and copper of any th ickn ess offers about 40 dB. Neither

will provide the shielding required. Table 6.2 indicates that Hypernom

offers 2.3 dB per mil thickness at 100 Hz. Thus, about 26 mils of Hyper-

nom (60 dB attenuation) should adequately shield the twin-T amplifier

cable.

The attenuation offered by materials

to

electric, ma gne tic, a nd elec-

tromagnetic waves described

in

th e previous sections

is

achieved theo-

retically. In practice, however, this att en ua tio n is not often achieved,

because a

shielded enclosure

or

housing

is not

completely sealed.

In

other words, nearly any practical application of shielding has necessary

penetrations of one kind or another. The next section discusses the loss

of such shielding integrity

and the

practices t ha t may

be

followed

to

reclaim the integrity.

6.4 EMI Shield Compartments and Equipments

The preceding sections covered the subject of shielding, theory, and

materials. It was shown that, for othe r th an low-frequency mag netic

fields, it

is

easy

to

obtain more tha n

100 dB

shield ing effectiveness



126


across the spectrum for nearly any metal. The shielding problem then

develops from the fact that practical enclosures have apertures and

penetrations that compromise the effectiveness of the basic shield

m ater ial. T hu s, shielding effectiveness of a hous ing could be reduced to

60 dB or less because of th e loss of enclosu re integ rity.

It now remains to bring the foregoing material together in the form

of practical shielded-housing applications. Consequently, this section

reviews the subjects of shielded compartments, chassis and equip-

ments, and cabinets. Typical examples of the chassis of equipment-level

shielded housing include electronic test instruments, biomedical equip-

ment, mobile transceivers, hi-fi amplifiers, and microcomputers.

Figure 6.12 illustrates a typical equipment case with a number of

represe ntative shielding compromises such as:

• Cover pl ate for access

• Holes or slots for cooling

• Power and signal cable entr y

• Displays, inst rum ents , and switches

The designer of an equipment case must give careful consideration to

these shielding compromises and should incorporate various protective

measures to minimize the compromise in shielding integrity.

Holes or Slots

for Convection Cooling

Cover Plate

for Access

Screw Spacing for Slot Radiation

Forced Air

/~~ Cooling

Panel M eter

F ig u re 6.12 Some principal box shielding compromises.



SHIELDING INTEGRITY PROTECTION

127

Many of the shielding integrity compromises in an equipment case

(such as openings or seams) can be regarded as apertures, and the leak-

age will be a function of the aperture size relative to wavelength.

Figure 6.13 illustrates the principal of leakage through an aperture.

Referring to Fig. 6.13, it can be observed that as the aperture size

approaches one-half wavelength, the leakage increases and, a t one-half

wavelength, the aperture does not provide any shielding. Therefore, in

designing equipment cases, it is particularly important to keep the size

of any apertures much less tha n one-half wavelength at th e highest fre-

quency for which shielding is required. Shielding integrity protection

techniques are described in the following section.

6.5 Sh ielding Integrity Protec tion

The previous sections discussed the subjects of shielding theory and

materials. With the exception of low-frequency magnetic-field shielding,

it was shown that it is quite simple to obtain more than 100 dB of

shielding effectiveness across the entire spectrum from de to light for

electric and electromagnetic waves. However, since any practical enclo-

sure has apertures, the theoretical shielding is never obtained, due to

• Worst Case, Simplified Model:

• Vertical Polarization:

S E

d B

>

20\og(A/2l),

for

I <

1/2

• Horizontal Polarization:

S E

d B

> 201og(l /2h), for h < 1 /2

•

Best

Case,

Simplified

Model:

• S E

d B

< 201og

10

(A . A ) for I h <

A 12

\2l 21y

where:

X

= wavelength in same units

as slot dimensions

I &

h

SE^g < Shielding Effectiveness of

Base Shield Metal

• Default Model (Diagonal POL)

• S E

d B

= Lesser of Worse Case + 3dB

1 = 1/2

log Frequency

Fig ure 6.13 Slot and aperture leakage.




loss of integrity . This section disc usse s t h e res ul t an t loss of shie ld ing

integrity, how t h e in tegr i ty can be rec la imed, a n d pract ica l ap pl icat ion s

to shielded boxes, chassis , equipment , a n d cabinets .

6.5.1 Integr ity of Sh ieldin g Configurations

The attenuation offered by materials to electric, magnetic, and electro-

magnetic waves described in the previous section is achieved theoreti-

cally. In practice, however, this attenuation is not often achieved,

because a shielded enclosure or housing is not completely sealed. In

other words, nearly any practical application of shielding has necessary

pen etrations and ap ertu res of one kind or another.

Thus, it is not uncommon to find the plane-wave attenuation of a

basic shield material to be 120 dB, for example, while the actual enclo-

sure will exhibit 50 dB in the VHF/UHF portion of the spectrum. Here,

leak age compromises the integrity of the basic shielding m aterial. Pro-

tective measures that may be used to reduce leakage are described

below.

6.5.1.1 Bon ding of Seam s and Joints

Loss of RF shielding integrity across the interface of clean mating

material members is a main reason why shielding effectiveness is com-

promised. Here, the conductivity of the interface may be much higher,

and/or the permeability may be much lower, because of the type of

interface bond used. Thus, resulting material interfaces may be classi-

fied into two types: physically inhomogeneous and physically homoge-

nous.

A physically inhomogeneo us interface bond res ult s when shielding

members are directly connected by screws, rivets, spot welds, and the

like. The interface connection is not continuous, and there results a

bowing or waviness effect between connected members. This in turn

develops slits or gaps, which leads to radiation or penetration at fre-

quencies approaching 0.01. The attenuation, A, in dB at such a gap fol-

lows the waveguide-beyond-cutoff criteria:

A

d B

= 0.0046 l

d

f

MHz

J ( f

c

/ f

M H z

)

2

-

1

dB (6.11)

where,

Id

=

&

a

P depth in inches for overlapping members or the thickness

of the m aterial for bu tting mem bers

fMHz

=

op era ting frequency in MHz (6.12)




f

c

= cutoff frequency of gap in MHz

= 5900/g for a rect an gu lar g ap

= 6920/g for a circu lar gap

g = largest gap tran sve rse dimension in inches

When f

c

»

f

MHz

,

Eq. (6.11) becomes:

A

d B

« 0.0046tf

c

= 27 1/g dB for rectangular gap

= 32 1/g dB for circular gap

129

(6.13)

(6.14)

(6.15)

Figure 6.14 is a plot of Eq. (6.11) representing attenuation through a

rectangular gap versus frequency as a function of gap dimensions. The

figure shows th at more th an 100 dB atte nu ati on exists over th e de to 10

GHz spectrum for both git ratios greater th an about 4 and the largest

gap dimension less than 0.2 inches (cutoff frequency of about 30 GHz).

A num ber of techn ique s are av ailable for reducing electromagn etic

emission leakage or receptor penetration of a shielded specimen. If

members are joined by screws or rivets, Eq. (6.15) shows that

A^B

m ay

be significantly increased by using more screws or rivets per linear

140

Largest Gap Dim ensions, g, in Inches

20 15 10 6 4 3 2 1.5 1 .6 .4 .3 .2 .15

300MHz 500 lGHz 2 3 5 7 lOGHz 20 30 60

Radio Frequency

F ig u re 6.14 Atten uation through a metallic gap vs. frequency.



130


dimension of the interface, creating a reduction in the gap, g.

Figure 6.15 shows a joint shielding effectiveness

as a

function

of

screw

spacing

for the

indicated p arameters. Also note

the

improvement

due

to

the

application

of a

typical EMI mesh gasket.

Other techniques available for reducing the leakage in a physically

inhomogeneous mating member bond involve attempting to eliminate

or reduce the inhomogeneity. Figure 6.16 illustrates some of these

approaches. Where members do not have to be disengaged or separated,

a continuous seam weld around the periphery of the mating surfaces is

preferred. This type of weld is not critical provided it is continuous and

has no weld pin holes. One exception involves the departure of the weld

filler m aterial from the basic shield member material. Hence, either the

conductivity or permeability of the weld filler may be much lower,

resulting in degradation of shielding effectiveness. The seam weld tech-

nique is of questionable value when used with the more exotic magnetic

materials (jn > 1000; see Table 6.1), which must be annealed before

assembly. Here, welding will destroy

the

specific properties that

the

annealing produced.

An alternative technique shown in Fig. 6.16 is the overlap seam. All

nonconductive material (e.g., paint, rust, coatings, etc.) must be

removed from

the

mating surfaces before they

are

crimped. Crimping

must be performed under sufficient pressure to ensure positive contact

between all mating surfaces.

Shield members, such

as

cover and access plates, may have to be sep-

arated from time to time for equipment alignment or maintenance.

120

100

g

80

60

? 40

1

|

20

0

—

•

**

u.

1

^ «

1

h

i

~~T~

]

For 1/2" m

0.090 Alun

_ [ „

j

L

eft

lir

1

— r~

1

4-

tr

il-to-metal joint

mm at 200 MHz

.

p .

o

I

1

•.Tf-

11

s

' N

— •

ztz

_

i-

•«

—

—

.2 .3 .5 .7 1 2 3 4 5 7 10 20 30 50

Screw Spacing in Inches

F i g u r e

6.15

Sh ielding effectiveness

for

screw-secured joints.




131

Weld Material

Non-Step /

/

Continuous Butt Weld

^

Fused Material

Formation of Permanent

Overlap Seam

-^

Note: Soldering or W elding is

Desirable for M aximum

Protection

,1 m c

Spot Weld

Courtesy of USAFSC DH 1-4

Figure 6.16

Perm anent and semipermanent shield seam configurations.

Therefore, no ne of th e above techniqu es is acceptable. A tem po rary bu t

good bond is required , an d th is is the role of RF ga sketin g m ate rial such

as fingerstock and re silien t m esh. The subject of gas kets is discussed in

a later section.

6.5.1.2 Ve ntilatio n O pen ing s

Most shielding housings or enclosures require either convection or

forced-air cooling. Since associated openings will compromise the integ-

rity of the basic shield material, a suitable electromagnetic mask must

be sought tha t will provide sub stan tial a ttenu ation at RF while not sig-

nificantly imped ing th e mech anical flow of air. Two approaches are pos-

sible: screened covers and honeycomb aperture covers. As explained in

the next section, screens are inexpensive approaches to this problem

but are limited in shielding effectiveness and tend to block the flow of

air due to turbulence. Thus, a honeycomb material is generally used,

because it provides higher shielding effectiveness and maintains a

strea m line flow of air.

In typical honeycomb construction, illustrated in Fig. 6.17, the hex-

agonal elements use the waveguide-beyond-cutoff technique to accom-

plish the desired shielding effectiveness. One representative

honeycomb configuration is shown in Fig. 6.18. Equation (6.11) previ-

ously indicated the expected attenuation. However, for honeycomb, the



132

SHIELDING THEORY, MATERIALS,

AND

PROTECTION TECHNIQUES

Foil Direction of

Upper Honeycomb

Foil Direction of

Lower Honeycomb

Figure 6.17

Typical honeycomb construction.

Figure 6.18 Representative honeycomb configurations.

shielding effectiveness

at

frequencies well below cutoff

is

reduced

by

the number of waveguide elem ents, N, in the panel, since the emerging

field from each hex cell coherently combines with its neighbor. Thus,

there results for honeycomb ven tilation covers:

A

d B

« 2 7 1 / g - 2 0 1 o g

1 0

N

(6.16)



SHIELDING

INTEGRITY PROTECTION

133

Figure

6.19

illustrates typical performance

of

different honeycom b

configurations.

The L-F

magnetic field performance, however, does

not

follow

Eq.

(6.16). Rather,

the

applicable relatio n

is Eq.

(6.6).

Sometimes,

it is

necessary

to

provide redu ction

or

removal

of

dust

in

the ventilation process. Honeycomb construction will

not

remove dust.

Thus,

a

shield screen

is

fabricated

of a

woven-wire mesh.

The

shielding

mesh medium

can be

either

dry (see Fig. 6.20) or wet (to

accommodate

an

oil

coating

for

more dust removal;

see

Fig. 6.21). Figure

6.22

shows

typical attenuation of shielding m esh covers vers us frequency.

When ventilation cover panels

are

used

for

convection cooling,

it is

often common practice

to

employ

a

number

of

perforations

in the

panel

ra ther than

to use

honeycomb

or

screen. Holes

are

punched

out

with

a

0

10kHz

1MHz 10 100

Radio Frequency

lGHz

10

F i g u r e

6.19

Typical shield ing effectiveness

of

honeycomb v ent covers.

F i g u r e 6.20 Rep resentative shield screen mesh ventilation covers for air fil-

ter ing.



134


F ig u r e 6 .21 S h ie ld s c r e e n me sh ve n t i l a t i on pe r m i t t i ng dus t r e mov a l by o il

i m p r e g n a t i o n .

10kHz

1MHz 10 100

Radio Frequency

lGHz 10

F i g u r e 6 .22 Typica l sh ie ld in g e f fec t iveness of sh ie ld sc reen m esh ve nt covers .

die, which also cuts the cover panel. For this situation, the shielding

effectiveness, A^b, is:

(6.17)

where,

k = 27 for squa re perforations (opening holes)

= 32 for circu lar p erforatio ns

1 = thi ckness of cover pa ne l in inche s (or cm)

g = wid th of squ are perforations or diame ter of circular perfora-

tions in inches (or cm)



SHIELDING

INTEGRITY PROTECTION 13 5

C = cente r-to-cen ter spacing of perfo rations in inche s (or cm)

D = leng th of ap ert ur e for squ ares or diam eter for circular aper-

tu res in inches (or cm)

If the cover plate perforations are not equally spaced, then C

2

in

Eq. (6.17) may be replaced by C

2

= A/N, where A = area of ape rture = D

2

and N = number of perforations or holes. For this situation, Eq. 6.17

becomes:

(6.18)

(6.19)

Both the honeycomb and mesh covers are mounted over the ventilation

opening with gasketing material.

6.5.1 .3 Vie win g Ap ertures

Another req uirem ent th at compromises the integrity of the basic shield

material is the need for viewing panel meters, digital displays, scopes,

and other types of status monitors or readout presentations contained

inside the shielded housing or enclosure. This is accomplished by either

a laminated-screen window or a conductive-optical substrate.

Screen Windows

A shield screen window may be used to block RF penetrations in which

fine knitted wire is laminated between two layers of acrylic or glass.

Figu re 6.23 illu stra tes th is. The wire may be monel with typical sizes of

0.002 in. diameter (20-25 openings per inch) or 0.0045 in. diameter

(10-1 3 openings pe r inch). This correspond s to a low-shadow a rea (15 to

20 percent blockage, giving good visibility). Typical shielding effective-

ness is shown in Fig. 6.24. This app roach is becoming less pop ular th an

that of the conductive-optical substrate described below because of the

less-esthetic aspects of the former. Furthermore, under some condi-

tions,

a screen window exhibits undesired diffraction-grating viewing

problems.

Conductive Optical Substrate Windows

Another approach is available for providing shielding across apertures

through which either optical viewing or the transmission of light is also

necessary. This approach involves the use of a conductive window, a

technique in which a thin film of metal is vacuum deposited on an opti-



136


F ig u re 6.23 Re presentative shield screen windows for viewing.

100

r

10kHz

1MHz 10 100

Radio Frequency

IGHz

10

Fig ure 6.24 Shielding effectiveness of shield screen windows.

cal substrate. These conductive window designs, such as shown in

Fig. 6.25, ar e evolved by estab lish ing some or all six basic design

parameters, as applicable:

• Window m ateria l

• Reticle requ irem ents

• Conductive coating

• EMI gasketing

• Optical coating and finishes

• Fram ing and moun ting




137

F ig u re 6.25 Typical conductive optical viewing pan els.

Most plastic and glass panel materials are suitable as subs trates for

the application of conductive coating. The commonly accepted, more

standard materials are glass, acrylic, polycarbonate, and fluorocarbon

plastics. The substrates may be clear or colored, as required by the

application. There are no restrictions on substrate thickness. Curved or

three-dimensional parts can generally be coated.

Most thermosetting and thermoplastic substrates have minute sur-

face scratches produced in their normal manufacture. The application

of the coating will inherently make these more apparent, although

actual user experience indicates that no functional problem will arise.

The following list illustrates a sample of the large selection of substrate

materials suitable for conductive coating.

• Glass, plate

• Plexiglas, thermoplastic acrylic

3

• Glass, single strength

• Plexiglas, transparent, colorless

• Glass, float

• Plexiglas, frosted, colorless

• Glass, tempered

• Plexiglas, colored: yellow, amber, grey, bronze, green, red, blue

• Glass laminated, PVB

• Homalite, thermosetting plastic

4

film , safety



13 8 SHIELDING THEORY , MATER IALS, AND PROTECTION TECHNIQUES

•

Glass, quartz

•

Kapton

5

• Crystals, ruby

• Mylar

5

• Crystals, quartz

• Abcite, coated acrylic

5

• Vycor

1

• Polycarbonate

• Pyrex

1

• Self extinguishing Plexiglas

• Lexan

2

•

Fluororocarbons

Trademarks of: 1. Corning, 2. General Electric, 3. Rohm & Hass, 4. Homalite, and 5. DuPont.

In the plastic substrate g roup, the mos t scratch-resistant mate rials are Abcite followed by

Homalite.

Polarized filter laminate finishes are available for contract improve-

ment. Coatings are unaffected by application of laminated circular

polarizers. Translucent or frosted finishes, rough in surface nature, are

available. They are best employed on the side opposite the conductive

face.

They can be used only for display of rear projections or where the

object is extremely close to the window surface. Antireflective, vacuum-

deposited coatings may be applied to windows before coating.

Figure 6.26 illustrates typical shielding effectiveness versus fre-

quency for different film coating thicknesses on glass measured in sur-

face resistance units of ohms/square. Since the film thickness is

deposited in microns, little contribution to attenuation comes from

absorption loss. Accordingly, reflection loss, as previously shown in

Figs.

6.6 and 6.7, is the medium of attenuation. Above about 1 MHz, the

loss decreases with an increase in frequency at the rate of approxi-

mately 20 dB per decade and becomes negligible above about 1 GHz.

Light transmission versus surface resistance for the above conduc-

tive glass is shown in Fig. 6.27. Transmission values of 60 to 80 percent

correspond to resistances of about 10 to 100 Q/square. Thus, these val-

ues shown in Fig. 6.26 may now be compared with the attenuation data

of the shield screen depicted in Fig. 6.24 for comparable area size speci-

mens.

The shield screen is seen to be everywhere superior in shielding




139

120

100

1

60

40

20

0

-H

•

1

1

„_ __

i

-4 —

__4—

~4-~

h+-

70

oh

a

\

i

i n s

/sqi

1 Ctl

1

• s

la

?**

ir e

j

JL

:

i

— I—

t

— T ~ 1

;

t

\

s

s«

Hi

s >

11 ti

t

j

1(

/

s

)or

ft

s

,_.

S ,

]

I

,

im s

/

-v

\

s

or

S

s.

S,

%

J (

J l

—

-

l

-

i r e

40

/

sr~

C

q

»h

i t

s

s ,

Is

li<

m

s,

—

lie

s/s

s

qua

r

s

fs.

i N

i

__

li

S

sr

-f-

id

-

~v

s

S^

•x

s

*s

S i

100kHz 300 1MHz 3 10MHz 30 100MHz 300 lGHz

Radio Frequency

F i g u r e

6.26 Sh ie ld ing effec t iveness of conduc t ive g lass .

1UU

r

n

t

s

g

8 0

|

70

2 60

-

<•*

I ,

>•

•

I

•

;

i

•

I

s

10 20 30 50 70 100 200 300 500 1000

Surface Resistance in Ohms/Square

Fig ure 6.27 Light transmission of conductive glass.

effectiveness, as shown in Table 6.3, in which the difference becomes

greater with increasing frequency. Thus, it is concluded that if signifi-

cant VHF and UHF attenuation is required for viewing apertures,

shield-screen windows should be used. If the esthetics or other consid-

erations do not permit this, conductive glass cannot be relied upon to

provide significant RF attenuation to E-fields much above 30 MHz.

6.5.1.4 Control-Shaft Apertures

Another aperture class that compromises the shielding integrity of

an equipment housing or instrument panel is that resulting from

shafts of potentiometers, tunin g dials, and control devices. Generally,




Table 6.3 Comparison of Shielding Effectiveness of Screen and Conductive Glass

Windows

Superiority of

Frequency Shield screen Conductive glass shield screen

lMHz

10 MHz

100 MHz

lGHz

98 dB

93 dB

82 dB

60 dB

74-95 dB

52-72 dB

28-46 dB

4-21 dB

3-24 dB

21-41 dB

36-54 dB

39-56 dB

an external metallic front panel or housing is either drilled or

punched with sufficient clearing tolerance through which the control

shaft extends

to

result

in a

leaky aperture.

The

inside wall

of the

panel hole forms an outer conductor to a coaxially situated internal

control shaft (i.e.,

the

inner conductor).

In

other words, poten tial

EMI

can enter or exit through this effective short-length coaxial line, and

the extended shaft beyond the panel acts as a pickup or radiating

antenna.

To preserve the shielding integrity of otherwise leaky control-shaft

situations,

one

method

of

minimizing the degradation

of

shielding effec-

tiveness is to design a supporting bushing extender to act as a circular

waveguide-beyond-cutoff attenuator [cf. Eq. (6.11)]. For 100-dB a ttenu-

ation

in a

circular waveguide,

the

length

of the

waveguide must

be

somewhat more than three times its diameter [1/g > 3 in Eq. (6.15)]. Fig-

ure 6.28 shows an acceptable use of a metal tube bonded to the wall

containing the clearance aperture

for

control shafts.

If the preceding situation were implemented without regard to the

control shaft properties

and

relations

to the

added metal tube, little

improvement could result for typical metal shafts. This situation corre-

sponds

to a

low-impedance coaxial line

in

which

an

intervening dielec-

tric may result from contaminants such as oil films or oxides. To

preclude this , one of

two

techniques is followed: (1) replace the metallic

control shaft with

a

non-conductive shaft

as

shown

in Fig. 6.28, or (2)

use a cylindrical-shim EMI gasket between the shaft and tube. The lat-

ter method does not require modification of existing control shafts.

6.5.1.5 Indicator Bu ttons and Lamps

Some instruments or equipments require the use of pushbuttons, sta-

tus indicator buttons, and/or indicator lamps. These devices also pro-

vide another compromise of shielding integrity by virtue of the required

apertures in a front panel or housing. Two techniques are available to

mitigate the EMI leakage through such devices:




141

Panel

Weld or Braze

Metal Tube

Nut

Lock Washer

Mounting Bracket

( a )

Nut

Lock Washer

Panel

RF Gasket

Control or Switch

7

Enclosure

Panel

Metal Tube Acting as

Circular Wave Guide

Non-Conductive Shaft

and Knob

(c )

Courtesy of

USAFSC

DH 1-4

Fig ure 6.28 Use of circular waveguide in a permanent aperture for control-

shaft EMI leakage control.

1.

Encase them in a shielded comp artment behind th e front panel

when they are mounted, as shown in Fig. 6.29. Feed-through capac-

itors or filter-pin conductors are used for hard wiring from outside

the com partmen t to the butto ns or indicator lamp s, since conducted

EMI could exist on eithe r side of the barrier.

2.

Use special EMC-designed ha rd w are wh ere such devices are

mounted directly to a front panel.



142


EMI Gasket

Shielded Compartment

Shielded

Compartment

Front Panel

Feed-Through Capacitors

(Filtered Leads)

Multi Filter-Pin

Connector

-Option to Fill

With Lossy Dielectric

For Additional

Energy A bsorption

Fig ure 6.29 Shielded and filtered compartment technique to restore shielding

integrity of button and lamp apertures.

6.5.2 EMC G as ke ts

This section discusses a very im portant class of techniques used to rein-

state loss of shielding integrity at seams and joints where nonperma-

nent fastening methods are permitted.

6.5.2.1 Gasketing Theory

Gaskets are employed for either temporary or semipermanent sealing

applications between joints or structures, such as:

Temporary

RF

Sealing Applications

• Securing access doors to enclosures, cabinets, or equipments

• Mounting cover plates or removal panels for equipment mainte-

nance, alignment, or other purposes

Semipermanent RF Sealing Applications

• Mounting either screen or conducted glass windows to housings

containing electrical or electronic test equipment

• Mounting honeycomb and other ventilation covers to enclosures,

cabinets, or equipment

• Securing parallel members of an equipment housing to a frame

structure using machine screws

All gaskets of the non-spring fingerstock type (whether they seal

EMI, contain higher-pressure fluid, make a container dunk proof, or



SHIELDING INTEGRITY PROTECTION 143

simply keep forced ventilating air from escaping at a door-to-cabinet

joint) conform to the unavoidable irregularities of the mating surfaces

of a joint. Some examples are:

• The joint between a garden hose and wate r faucet

• Hou sing for an eme rgency radio or beacon to be dropped into th e

sea

• The joint betw een the cover an d enclosure of a ra da r pulse m odula-

tor

In each example, the joint has two relatively rigid mating surfaces,

and neither surface is perfectly flat. When the surfaces are mated

without a gasket, even high closing forces will not cause the two sur-

faces to mu tuall y seal. Res ulta nt gaps will allow leak s to exist. A gas-

ket resilient enough to comply with both surfaces under reasonable

force, however, will eliminate these leaks. In the garden-hose exam-

ple,

try to prevent a leak by force alone without a gasket. With a gas-

ket placed in the hose fitting against a faucet, even hand torque

res ults in a w ater -tigh t joint. To try to get the s am e wat er tig htn ess by

accurate machining of both surfaces would be prohibitively expensive.

Thus ,

in most cases, the least expensive way to obtain a tight joint

(watertight, oil-tight, or EMI-tight) is to make the mating surfaces to

normal tolerances on flatness, rigidity, and tolerance buildup, and

then to add a gasket to compensate for the resulting misfits between

the two surfaces.

6.5.2.2

Joint Uneven ness

The degree of misalignment or misfit of the mating surfaces is com-

monly called

joint unevenness

and is designated H in Fig.

6.30a.

It is

the maximum separation between the two surfaces when they are just

touching and in the limit becomes the sum of the peak irregularities of

both surfaces. If the surfaces are not rigid, then the joint unevenness

also includes any additional separation between the two surfaces due to

joint distortion when pressure is applied.

Figure

6.30b

shows the same joint with a gasket installed. The

dashed lines indicate the gasket height, H

g

, before compression. The

compressed minimum gasket height, H

m i n

, occurs at the point where

the surfaces would touch without a gasket. Compressed maximum gas-

ket height, H

m a x

, is at the point of maximum joint separation. Thus,

joint unev enne ss of the m ating surface is:

Joint unevenness = AH = H

m a x

-

H

m

j

n



144 SHIELDING THEOR Y, MAT ERIALS, AND PROTECTION TECHNIQUES

I

Joint Unevenness

= AH

Jus t Touching

( a )

H

m

Gasket ^ LJU J ^ ^ ^ / U n c o J n p r e s s e d

Gasket

.1.

AH H

m a x

- l

(b ) Compressed Gasket

in

Place

F i g u r e

6.30

Description of joint unev enn ess.

6.5.2.3 Required Compression Pressure

Three factors determine the required compression press ure on a gasket:

its resiliency,

the

minimum pressure required

for a

seal,

and the

total

joint unevenness.

(A) Resiliency

Resiliency

is the

amount

by

which

a

gasket compresses

per

uni t

of a

percentage

of

original (uncompressed) gask et heig ht, divided

by

pres-

sure in psi. A soft gask et w ould compress more th an a hard gasket with

the same applied pressure. Stated another way,

a

soft gasket requires

less pressure than a hard gasket to compress the same percentage of

gasket height.

For

example,

a

sponge neoprene gasket might compress

10 percent under

an applied compression pressure of 6 psi, but a solid

neoprene g asket would require 40 psi for the sam e 10 pe rcen t deflection

as shown

in

Fig. 6.31.

(B) Minimum Pressure

for

Seal

A gasket must

at

least make contact

at the

point

of

maximum separa-

tion between ma ting surfaces,

i.e.,

H

m a x

<

H

g

in

Fig. 6.30. Actually,

the

pressure

at

this point must

be a

stated minimum amount

in

order

to

assure

an EMI

seal. T his

is

easy

to

understand

in the

case

of a

high-

pressure lubricating system. If there is not some required minimum

pressure

at the

point

of

H

m a x

,

oil

will blow

by

between

the

flanges

and

the gasketing material . Thus,

the

pressure

at the

H

m a x

point must

be

high enough

to

prevent blow-by.

For EMI

gaskets, this m inimum pres-

sure, P

m

i

n

, is determined by the pressure required to break through




145

100

9 0

80

| 70

o 60

50

0

i l U -

s

s:

—I-H 1—5k

: • • :

A h - . : • : .

: ; ; - ; ; -

^GI]

' I i

•"• . . J . J 4 J -T

T

. ~ •::.:. ..:. i .

^

T

ip

m^

^ | - ^ - H

>

a i l Gas tel;

j

•

s = =

-

• • : . • • : : • . . - . " . • • . . - - • :

- - - -

X ——

0 10 20 30 40 50 60 70

Gasket Pressure in PSI

90 100

Fig ure 6.31 Typical hard and soft EMI gasket height vs. pressure relations.

corrosion films and to make a suitable low-resistance contact. P

m

i

n

is

typically abou t 20 psi bu t can be as low as 5 psi.

C) Average Pressure

The average pressure applied to the gasket must also be large enough

to compress the overall gasket so that the difference between the mini-

mum height and the maximum gasket height (determined by P

m

i

n

from

the previous paragraph) is equal to the joint unevenness, i.e., H = H

m a x

- H

m i n

, as previously presented in Eq. 6.19. In general, the average

pressure should equal or exceed that corresponding to the average com-

pressed gasket height, H

a v g

:

H

m in

)/2

(6.20)

(6.21)

The required compression force, F, in units of points, may be calcu-

lated from P

a v g

by determining the surface area of the gasket to be

sandwiched between the mating members:

F =

P

a v g

x A pounds

where A = gasket are a in square inches

Required Gasket Height

To obtain the required EMI seal from a gasketed joint, the gasket

height must meet these criteria:




• The press ure a t t he point of ma xim um joint sepa rat ion (H

m a x

) must

correspond

to th e

min imum pressu re

to

obta in

t h e

requi red

E M I

seal .

• The

difference betw een m ax im um

a n d

mini mum compressed

heigh ts of the gask et mu st equal the joint une ven nes s of the ma ting

surfaces.

If the average pressure , avai lable to compress t h e gaske t is P

a v g

, t h e

maximum pressure , P

m a x

,

is

obtained from

Eq. 6.22:

Pmax

=

2 P

a v g

- P

m i n

(6.22)

The percentage

of

uncom presse d height corresponding

to

P

m

i

n

a n d

P

m a x

in Fig. 6 .31 are

H

m a x

a n d H

m

j

n

,

respectively.

To

calculate

t h e

required uncompressed gasket height ,

EL, as a

dimension:

o

A T T

H =

(in or cm)

( 6 2 3 )

AH

decimal

Thus, the required height is the actual joint unevenness in inches

divided by the joint unevenness expressed in decimal equivalent of per-

cent gasket compression (See Fig. 6.31).

Compression Set

Some gaske ts do not retu rn to their original uncompressed height after

release of compression. This is called compression set. It may be visual-

ized by assuming that the lower curve shown in Figure 6.31 applies for

a particular soft gasket. When compression pressure is reversed, the

gasket re tu rn s to a lesser height whose properties might look somew hat

like the upper curve in Fig. 6.31 (this is exaggerated for illustrative

purposes). The importance of compression set depends on how the gas-

ket is to be used. The classes of use are defined below:

• Class A, permanently closed. Compression set is unimportant, since

the gask eted component, in all probability, will never be removed.

• Class B, repeated identical open-close cycles (e.g., hinged door or

symmetrical covers). Here, compression set problems are marginal;

further exam ination of details, however, is indicated.

• Class C, completely interchangeable (complete freedom to reposi-

tion g asket on repea t cycles; e.g., round gask et in w aveguide). Since

the com pression-set height at a point of max imu m com pression may

end up being less th an minim um compressed height, no contact at

all would result between gasket and mating surfaces at this point.




147

For class

C

applications, do not reuse gaskets with compression set

limits; instead, use a new gasket.

6.5.2.4

Gasket Types and M aterials

There exists a plethora of

EMI

gasket

types,

shapes, binders, and mate-

rials. In fact, the profusion of gaskets is so great that it is likely to be

confusing to all but those who specify or use them with some degree of

regularity. This is recognized by the suppliers to the extent that they

have produced creditable application notes and design and order

guides.

For convenience of discussion here, EMI gaskets are divided into four

types:

(1) knitted wire mesh, (2) oriented immersed wires, (3) conduc-

tive plastics and elastom ers, and (4) spring fingerstock. The las t type is

different from the first three types and operates on a significantly dif-

ferent principle. A brief sum mary of each is presented below followed by

a comparison of all four types.

Knitted-Wire Mesh Gaskets

Figure 6.32 shows some examples of knitted-wire mesh gaskets. They

are m ade from resilient, conductive, knitted wire and somewhat resem-

ble the outer jacket of a coaxial cable. Nearly any metal that can be pro-

duced in a fine-wire form can be fabricated into these EMI gaskets.

Typical materials used are monel; aluminum; silver-plated brass; and

tin-plated, copper-clad steel. These gaskets may employ either an air

Fig ure 6.32 Typical knitted wire mesh gaskets.




core or, for maximum resiliency, they may use a spongy neoprene or sil-

icone core. Cross sections m ay be round, rec tang ular, or roun d w ith fins

for mounting. They are generally applied to shielding joints having a

periphery of greater than 4 in. (10.2 mm) and cross sections between

0.063 in. (1.6 mm) a nd 0.75 in. (19 mm).

Oriented Immersed-W ire Gaskets

Figure 6.33 shows some examples of oriented immersed-wire gaskets.

They are made with a myriad of fine parallel, transverse-conductive

wires whose parallel impedance across the gasket interface is very low.

Each convoluted wire is insulated from its neighbor. They represent a

density of about 1000 wires per square inch. Typical materials used are

monel or aluminum embedded in either a solid silicone (hard gasket) or

a sponge silicone (soft gasket) elastomer. As such, this gasket provides a

simultaneous EMI and pressure seal. The embedded wires protrude a

few mills on each side to assist in piercing any residual grease/oil film

and oxide on the surface of the mating numbers. This characteristic is

especially good where aging and subsequ ent main tenance may result in

a panel number being no longer clean and degreased. Available cross

sections range from 3.175 mm sq. (0.125 in. sq.) to 15.875 x 12.7 mm

(0.626 x 0.500 in.) and come in any length.

Conductive Plastics and Elastomer Gaskets

Figure 6.34 shows some examples of conductive plastic and elastomer

gaskets. They are made with a myriad of tiny silver balls immersed in a

silicone rubber or vinyl elastomer binder and carrier. As such, this gas-

Fig ure 6.33 Typical oriented immersed-wire gaskets.




149

F ig u re 6.34 Typical conductive elastomer gaske ts.

ket provides a simultaneous EMI and hermetic seal. Offering volume

resis tivities from 0.001 to

0.01

Q-m and useful over a wide range of tem-

perature, these gaskets are provided in sheets, die cuts, molded parts,

and extruded shapes. Some versions are operable down to cryogenic

temperatures. They offer low closing pressures, low compression set

and maintenance, and long life.

Spring Fingerstock Gaskets

Figure 6.35 shows some examples of beryllium copper, spring-finger

gaskets stamped into different configurations. Basically, gaskets simi-

lar to these were introduced over 30 years ago and were the firs t type of

EMI gasket appearing on the market. Since there existed little elas-

tomer technology in the 1940s, it is na tural th at joint unevenness could

be accommodated by a series of individual fingers, each capable of flex-

ing a different amount. Thus, shielded enclosures, cover plates , and

Fig ure 6.35 Typical spring fingerstock gaskets.



150 SHIELDING THEOR Y, MATE RIALS, AND PROTECTION TECHNIQUES

other heavy-duty applications used, and still use, this type of gasket.

Recent design changes, shown in Fig. 6.35, make this type of gasket

more competitive with the other gaskets. The spring-finger contact

strips offer self-adhesive backing to eliminate older mechanical fasten-

ing methods. They are available in a wide variety of sizes and shapes.

The principal disadvantages are tendency of the fingers to oxidize and

to break off.

Pressure-Sensitive, Foam-Backed Foil Gaskets

Another type of gasket differing from the above is a beryllium-copper

foil backed by a highly compressible neop rene foam. Th e foam side, con-

taining a synthetic rubber pressure-sensitive adhesive, is applied to

cover plates. When placed over an electronics package containing

shielded compartments, the foam-backed foil assumes the irregularities

of the compartment heights, including outside plates to result in a con-

tinuous EMI seal. This 1/16-in. gasket is available in sheet widths to

6 in. or may be die cut. EMI shielding effectiveness of 90 dB to electric

fields is claimed over the 1 kHz to 10 GHz frequency spectru m.

Com parison of Gasket Types and Ma terials

With the profusion of different gasket types and materials (over 1000

variatio ns), it is confusing to th e design or specification engin eer task ed

with the responsibility of selecting one or more best candidates for his

pa rtic ular app lication . Accordingly, Table 6.4 is a com parison of some of

the principal characteristics of EMI gaskets. No one type is the best for

all applications. For example, those gaskets having relatively low cost

tend to have relatively higher volume resistivity, resulting in a less-

impressive shielding effectiveness. Some gaskets are designed to oper-

ate down to cryogenic temperatures or up to 500°F, but not both. Since

there exist several different methods of mounting, gaskets are available

in sheets and strips, die cuts, molded shapes, and extruded forms. At

the risk of generalizing, conductive plastics and elastomers seem to

offer the widest range of applications and price.

6.5.2 .5 Ga sket Sele ct io n and M ountin g

EMI gasket selection involves making suitable matches and tradeoffs

between (1) available EMI gasket materials and their characteristics

(see Table 6.4) and (2) performance requirements of equipment and

design constraints of mating surfaces. Gasket mounting (and hence

selection) involves a number of alternatives.

Gasket Selection

In selecting one or more suitable EMI gaskets for sealing mating sur-

faces, gasket characteristics, application requirement and constraints,




Table 6.4

Comparison of Gasket Types and Materials

151

Comparison

factors

Available forms

Size

Type of seal

EMI only

EMI + hermetic

Conductive

material

Binder or core

material

Temp, range

Available gasket

heights

Joint unevenness

accommodations

Compression

height range

Compression

pressure

EMI shielding

performance

10

kHz (H)

10 MHz

lG H z

10 GHz

Gasket types

Knitted wire

mesh

Strips, jointless

rings

Periphery

Min. cross section

Max. cross section

Good-excellent

NA

Silver plate,

monel, alumi-

num, steel Sn/Cu/

Fe

Rubber, air core,

neoprene, silicone

sponge

Limited to core

0.062 to 0.500"

(1.57 to 12.7 mm)

0.020 to 0 .160"

(0.5 to 4.1 mm)

5 to 100 psi (14.5

to 290 kg/cm

2

)

25-30 dB

>100 dB

>90dB

Oriented

immersed wires

Strips

sheets,

jointless rings,

die-cut shapes

>

4" (102 mm)

0.063"

(1.6 mm)

0.750" (19 mm)

Good

Fair-excellent

Monel, aluminum

Solid sponge-

silicone

-70 to 500°F

(-57 to 260°C)

0.062 to 1.000"

(1.57

to 25.4 mm)

0.010 to 0.100"

(0.25

to 2.5 mm)

20 to 100 psi (58

to 290 kg/cm

2

)

>45dB

>100 dB

>90dB

Conductive

plastics &

elastomers

Strips sheets,

die-cut, molded,

extruded shapes

Also seals her-

metically

Many tiny silver

balls

Silicone or plastic

-100 to 400°F

(-73 to 204°C)

0.020 to 0.160 "

(0.5

to 4.1 mm)

0.003 to 0.030"

(0.076 to 0.76 mm)

20 to 100 psi (58

to 290 kg/cm

2

)

>35dB

>100 dB

>95dB

>70dB

Spring

fingerstock

Strips

Any

Good-excellent

Beryllium-copper

NA

-65 to 100°F

(-57 to 38°C)

0.062 to 0.400"

(1.57 to 10.2 mm)

0.035 to 0.250"

(0.89 to 6.4 mm)

7:1

>10dB

>120 dB

>100 dB

>100 dB

and price are the major considerations. These topics are summarized as

follows:

Application Requirements. This is usu ally sta ted in the form of

equipment performance specifications. They include amount of shield-

ing, pressure sealing, and environmental exposure (e.g., temperature,

salt spray, ambient pressure, and corrosive material).

Application Constraints. This is usually imposed by equipment hous-

ing design. They include space available, com pression force, joint uneven-

ness, contact surface characteristics, and attac hm ent possibilities.



152 SHIELDING THEOR Y, MATE RIALS, AND PROTECTION TECHNIQUES

The important matches and tradeoffs between application require-

ments and constraints on one hand and gasket characteristics and price

on the other are:

• Gasket height and compressibility must be large enough to compen-

sate for joint unevenness under the available force.

• The gasket m ust be capable of providing the required EMI sealing

and hermetic sealing (when applicable) when compressed by the

available force.

• There must be sufficient space for the gasket within the design lim-

itations of the application.

• The gasket m ust be attached or positioned by a means that fits in

with the joint design.

• The metal portion of the EMI gasket m ust be sufficiently corrosion

resistant and compatible with the mating surfaces.

• The EMI gasket must meet the temperature and other environmen-

tal needs of the equipment specifications.

Gasket manufacturers and suppliers provide design guide tables to

assist the user to select the gasket most nearly meeting the application

requirements and constraints.

Gasket Mounting

A

number of methods are available to position the gasket to a metal

mating surface: (1) hold in slot, (2) pressure-sensitive adhesive, (3) bond

non-EMI portion of gasket, (4) conductive adhesive, (5) bolt through

bolt holes, and (6) special attachments situations. Each of these meth-

ods is summarized below.

HOLD IN SLOT.

This method is recommended if the slot can be provided

at relatively low cost, such as in a die casting. All solid elastomer mate-

rials,

which embody the gasket material, are essentially incompress-

ible. These products appear to compress because the material flows

while it maintains a constant volume. Therefore, when these products

are used in a slot, extra cross-sectional area must be allowed for the

material to flow axially. At least 10 percent extra volume, and more if

possible, is recommended such as shown in Fig. 6.36.

PRESSURE-SENSITIVE ADHESIVE.

This method of mounting is often the

least expensive for attaching EMI gasket materials. Installation costs

are substantially reduced, with only a slight increase in gasket cost

over a material without adhesive backing. Most sponge-elastomer

materials are used for applications that do not require any hermetic

sealing. The adhesive-backed rubber portion of this material serves

only as an inexpensive a ttachm ent method for the EMI portion.




153

a. Making allowance

for solid elastomer

gasket flow

T7

MIJ

rJT//

Poor Design Good Design

b. Areas where non-

conductive or dry-

back adhesive can be

used

(a)

( c )

( d )

c. Bolt-through holes

Cover

rrv

Rivet

or Spot

weld

[U Box

strip

over

fin EMI Mesh Strips Gasketing

}

Cover

Cabinet

Door

Rivet

or Spot

weld

Aluminum

Extrusion

to cover

Box

rr

Cabinet

Door

Metalastic Gasketing

d. Special mounting

methods

Bailor

machine screw

E Z M Z Z 2 -

Gasket

Fastener

Figure 6.36

Different methods

of

mounting gaskets.




B O N D N O N - E M I P O R T I O N O F G A S K E T .

M an y good nonco nductiv e adh e-

sives are now available to bond an EMI gasket in position by applying

the adhesive to the non-EMI portion of the gasket. This can be insu-

lated from the mating surfaces by a nonconductive material and is often

a good way of mounting EMI gaskets. This method is shown in Fig.

6.36b.

The designer specifying nonconductive adhesive at tachment must

include adequate warnings in applicable drawings and standard proce-

dures for production personnel . These cautions sta te that adhesive is to

be applied only to the portion of the gasket material not involved with

the EMI gasket ing funct ion. Experience indicates that insta l la t ion

workers ,

either through carelessness or a misguided desire to do a bet-

ter job, will apply the nonconductive adhesive to the entire gasket,

including the EMI gasket portion. It is not uncommon to hear, "This

gasket would hold better if I glued all of it rather than half of it." This

occurrence completely degrades the EMI performance.

CONDUCTIVE ADHESIVE.

Since good conductive adhesives can provide an

adequate e lectr ical contact between the EMI gasket and the mounting

surfaces, they can also be used to mount the gaskets. However, the fol-

lowing cautions should be observed:

• Mos t conductive adh esiv es ar e ha rd and incom press ible. Th us , if

too much adhesive is applied, and it is allowed to soak too far into

the EMI gasket materia l , the compressibi l i ty wil l be destroyed.

Irregularly applied adhesive also has the effect of increasing joint

unevenness .

• Th e volu me resisti vity of th e adhes ive shou ld be 0.01 Q-cm or less,

preferably 0.001 Q-cm.

• Mos t conductive adhe sive s do not bond well to eith er neop ren e or

silicone. This is why all products that have conductive paths in elas-

tomer are rated "poor" for conductive adhesive bonding by the man-

ufac turers .

• Ap ply ing a 1/8 to 1/4 in. di am et er spot of cond uctiv e ad hes iv e ever y

1 to 2 in. is preferred over a continuous bead.

• Conductive epoxies wil l a t t ach the gasket perman ently. Th us,

removal of EMI gasket without destroying it is almost impossible.

B O L T - T H R O U G H B O L T H O L E S .

Th is is a very common an d i nexp ensi ve

way to hold gaskets in position, as shown in Fig. 6.36c. For most prod-

ucts , providing bolt holes involves only a small initial tooling charge.

There is generally no extra cost for bolt holes in the piece price of the

gasket. Bolt holes can be provided in the fin portion of EMI strips or in




155

rectangular cross section EMI strips if they are sufficiently wide, such

as >3/8 in.

SPECIAL ATTACHMENT MEANS PROV IDED. The knitted-mesh fins provided

on some versions of EMI strips a nd th e alum inum extrusions in alum i-

num gasketing were designed to attach these products as shown in

Fig. 6.36d. The mesh fins could be clamped under a strip of metal that

is held down by riveting or spot welding, or th e m esh fins can be bonded

with an adhesive or epoxy. The aluminum extrusions of aluminum gas-

ketin g can also be held in position by riveting or bolting.

EMI gaskets should be positioned so they receive little or no sliding

motion when being compressed. This is illustrated in Fig. 6.37. The

EMI gasket shown in Fig. 6.37a is subject to sliding motion when the

door is closed. This may cause it to tear loose or to wear out quickly. In

Fig. 6.37b, the gasket is subject to almost pure compression-only forces.

Th is is th e preferred position.

6.5.3 EMC Se ala nts

This section discusses another form of EMC shield integrity protection

in th e form of conductive epoxies and caulk ing.

6 .5 .3 .1 Co nd uc t iv e Ep oxi es

Conductive epoxies are used to join, bond, and seal two or more metallic

mating surfaces. The silver-epoxy resins replace soldering and other

bonding techniques and cure at room temperatures. The conductive

epoxy adhesive and solder families are used in the following applica-

tions:

• Electrical connections to heat-s ensitiv e components, capacitor

slugs,

ferrites, and integrated circuits

• Connect electrolum inescent pan els

(a ) Poor design, door

slides on EMI gasket

(b ) Good design, door

compresses on EMI gasket

Figure 6.37 Prop er me thod of mo unting gask et in cabin et door wall.




• Fo r bu s b a r s o r s t r i p s on con duc t i ve g l a s s

• B on din g flanges to w av eg uid es

• B o n d i n g w a v e g u i d e s e c t i o n s

• Bo l t ho l e s a nd f a s t e ne r s on e l ec t r on i c en c l o s u r e s

• J o i n i n g d i s s i m i l a r m e t a l s

• S e a l i n g I C p a c k a g e s a g a i n s t m o i s t u r e a n d E M I

• R ep a i r of p r i n t e d c i r cu i t s

• I n t e r c o n n e c t i n g c o n d u c t i v e - m e t a l g a s k e t s

• F i e l d r e pa i r s t o c i r cu i t s

• P e r m a n e n t s e a m s h i e ld i n g

• S e a l i n g E M I s h i e l d s

Preparation and Curing

T h e con duc t i ve epox i es a r e eas i l y mi xe d on a vo l u m et r i c ba s i s , e l i m i -

n a t i n g m u c h t i m e a n d e q u i p m e n t t h a t w o u l d o t h e r w i s e b e n e c e s s a r y fo r

w e i g h i n g . M o s t e p o x i e s c a n b e p r e p a r e d w i t h e i t h e r e q u a l v o l u m e s o r

w e i g h t s of t h e c o m p o n e n t s . T h e y a r e f o r m u l a t e d w i t h m i x e d v i s c o s it i e s

t h a t p r o d u c e a l ig h t , c r e a m y p a s t e to m a k e a p p l i c a t i o n w i t h s t a n d a r d

d i s p e n s i n g

e q u i p m e n t r e a s o n a b l e e a s y a n d f oo lp ro of. T y p i c a l c u r e t i m e s

a r e o n e d a y a t r o o m t e m p e r a t u r e o r 3 0 m i n u t e s a t 2 0 0 ° F .

Typical Properties

Depending upon the type of silver-epoxy resin used, typical volume

resistivity will range from 0.001 to 0.02 Q-cm. Operating temperature

range is about -80 to +250°F. Shear strength is about 1200 psi, and ten-

sile strength varies with type but averages about 2500 psi. It exhibits

excellent moisture resistance. The cured specific gravity is about two,

suggesting its relative light weight for many pay-load-limited applica-

tions.

6.5.3.2 Cond uct ive Caulk ing

Conductive caulking is used to EMI shield and seal two or more metal-

lic mating members mechanically held by other means. Silver particles

are suspended in resin to provide conductive sealing. Conductive caulk-

ing is used in th e following application s:

• Caulking EMI-shielded shelter panels

• Caulking EMI-tight cabinets and enclosures

• Improv ing joint and seam inte grity of electronic enclosures

• Protec ting m atin g me mb ers of shielded conduits



SHIELDING INTEGRITY PROTECTION 157

• EMI sealing and ground ing bul khe ad pan el fittings

• Mo isture sealing of ma ting mem bers

• Ad hering metal-foil tap e to shielded room joints

• Rep airing dam aged conductive gas kets

• Caulking fasteners, panels, and handle s

6.5 .3 .3 Pr ep ar ati on an d U se

The conductive caulking compounds, as with any EMI sealant and

bond, require that the surfaces be thoroughly degreased and cleaned of

oxide coatings. The ca ulking m ay be applied with con ventional caulking

guns and dispensing equipment such as small bead-orifice syringes.

Hand application with spatula or putty knife may be used. The caulk-

ing is free of any corrosive bind ers. It is used at room tem pe ra tur es , an d

most caulking will not cure (i.e., are permanently non-setting). This

feature permits easy disassembly of caulked parts for movement or

maintenance.

6.5 .3 .4 Typical Pr op er tie s

De pendin g upon the ty pe of silver resin used, typical volume resistivity

will range from 0.005 to 0.02 Q-cm. Operating temperature is -80 to

+400°F (-62 to 204°C). Moisture resistance is excellent. The final spe-

cific gravity is about 1.8, suggesting its relative light weight for many

payload-limited applications.

6.5.4 Conductive Grease

Conductive grease is not a member of EMI gaskets and sealants collec-

tion discussed in this chapter. However, it is related in that one of its

functions is to provide a low-resistivity contact to mating members.

Here, mating members may engage and disengage more often than in

most EMI gasketing applications, excepting finger stock used in

shielded enclosures.

Conductive grease is a low-resistivity, silver-silicone grease that con-

tains no carbon or graphite fillers. The material will maintain its elec-

trical and lubricating properties over a broad environmental range.

These conditions include high and low temperatures, resistance to

moisture and humidity, and inertness to many chemicals, ozone, and

radiation. Most conductive greases are viscous pastes that can be

applied at elevated operating temperatures to vertical or overhead sur-

faces without dripping or running.




Conductive grease is used on power substation switches and in sus-

pension insulators to reduce EMI noise. It also reduces make-break arc-

ing an d pitt ing of th e sliding m etal contact surfaces of switches and fills

in pitted areas with silver/silicone. In addition, normally-closed

switches are prevented from sticking due to corrosion or icing. The

grease is effective in maintaining a continuous electrical path between

contact surfaces that must be free to move. These include ball-and-

socket connections of power insulators, which, if allowed to arc, can

generate EMI. Conductive grease is designed to maintain low-resis-

tance electrical contact and thereby m ainta in equ ipment operating over

extended environmental conditions, helping to deliver continuous elec-

trical service.

Conductive grease is used on the co ntacting surfaces of circuit break -

ers and knife-blade switches. It reduces localized overheating or hot

spots in turn maintaining the blades spring properties and current rat-

ing of the switch or breaker at original equipment level. Lubricating

conductively prevents freezeup in operating equipment and permits

restoration of marginal or discarded breakers to rated capacity.

Typical volume resistivity is about 0.02 Q-cm. Operating tempera-

ture range is -650 to +450°F (-650 to 232°C). Conductive grease pro-

vides excellent moisture resistance and has no corrosion effect on

metals. Its pot life is unlimited, and unused portions can be returned to

the container.

Recommended Readings: EMI Shielding

[1]A Dash of Maxwell's Equations—A Maxwell's Equation Primer,

Part 4. Glen Dash, Ampyx, LLC,

Compliance Magazine,

April,

2010, p . 28.

[2]

A Da sh of Maxwell's Eq uations—A M axwell's E quatio n Primer,

Part 6, The Method of Moments, Glen Dash, Ampyx, LLC, Compli-

ance M agazine,

Ju ne , 2010, p. 20.

[3] The Basic Principle s of Shield ing, G ary F enical, L aird Technolo-

gies, Compliance Magazine,

Ju ne 2010, p. 12.

[4] Circu it Models M ake Shield D esign Sim ple, Glen D ash , Am pyx,

LLC,

Com pliance Magazine, 2010 Annual Guide,

p . 46.

[5] Antennas, 2nd ed., J. D. K rau s, New York, McG raw-Hill.

[6] Design and Selection of Shielding Gaskets for Medical devices and

the Effect of Cleaning Solutions on Material Performance, Anjali

Khosla, Claydine Lumibao-Arm, and Douglas S. McBain, Laird

Technologies, Compliance Magazine, July, 2010, p. 52.



SHIELDING INTEGRITY PROTEC TION 159

[7]

Differential Transfer Impedance of Shielded Twisted Pairs, Michel

Mardiguian, private EMC consultant, 2010 ITEM interference tech-

nology, EM C Test & Design Guide, p . 62.

[8] Low Frequen cy M agnetic Shielding: An Int eg rate d Solution, Rich

Emrich and Andrew Wang, Integran Technologies, Inc., Tronto,

Canada, ITEM interference technology, 2009 EMC Directory &


[9]

RF Shielding Materials: An Update on Selection and Cost Consid-

erations, Gary Fenical, Laird Technologies, St. Louis, MO.

ITEM

interference technology, 2009 EMC Directory & Design Guide, p.

134.)

[10]

Architectural Electromagnetic Shielding Handbook, A Design and

Specification Guide, Hemming, L.H., IEEE Press, 1992, ISBN 0-

87942-287-4.

[11] Cable Shielding for Electromagnetic Com patibility, Anatoly Tsalio-

vich, Hoboken, NJ: Jo hn Wiley and Sons, 1995.

[12] Coupling to Shielded Cables, E. F. Vance, Hoboken, N J: Jo hn Wiley

& Sons, 1978.

[13]

Design of Shielded Enclosures: C ost-Effective Me thods to Prevent

EMI,

Louis T. Gnecco, Newn es, 2000.

[14]

Electromagnetic Shielding,

Vol. 3, EMC Handbook Series, Don

W hite & M. Mardig uian, DWCI Pres s, 1988, 616 pp., 178 illus.

[15]

Electromagnetic Shielding Handbook for W ired and Wireless EM C

Applications, Anatoly Tsaliovich, New York: Kluwer Academic Pub-

lishers, 1999.

[16] Grounding and Shielding in Facilities, R. Morrison and WHo

Lewis, Hoboken, N J: Joh n W iley and Son s, 1990.

[17] Grounding and Shielding Techniques in Instrumentation, R. M orri-

son, 3rd ed., Hoboken, NJ: John Wiley and Sons, 1986.

[18] The Shielded Enclosure Handbook, Louis T. Gnecco, Tempest Incor-

porated, 1999.

[19]

Shielding Design Methodology and Procedures,

Don White, DWCI

Press, 150 pp., 65 illus .

Web Ad dresses for EMI Shield ing Sou rces

Spira Mfg. Corp. www.Spira-emi.com

MAJR Pro du cts www.MAJR.com

Leade r Tech, Inc. www .LeaderTechinc.com

Tech-Etch, Inc. ww w.tech-etch.com/shield

Arc Technologies, Inc . www.arc-tech.com



160


Fotofab

Braden Shielding Systems

Spectrum Advanced Specialty Products

Parker Hannifin Corp.

W. L. G ore

Assoc, Inc.

Intermark USA, Inc.

MuShield Co.

A-Jin Electron

www.fotofab.com/RF

www.bradenshielding.com

www.SpecEMC.com

www.parker.com

www.gore.com/emi

www.intermark-usa.com

www.mushield.com

www.ajinelectron.co.kr



Chapter 7

Bonding

Electrical bonding refers

to

th e process by which pa rts of an assembly,

equipments, or subsystem s are joined tog ether in a ma nner such tha t

they provide low contact impe dance. Th e objective is to make the joined

structures homogenous with respect

to the

flow

of

RF currents. This

mitigates electrical potential differences that can produce EMI among

metallic parts.

7 1 Effects of Poor Bonds

Poor bonds lead

to a

variety

of

hazard ous and interference-producing

situations. For example, loose connections in ac power lines may cause

hea t to be generated in the joint and dam age the insu lation of the wires

or loosen the co ntact pre ssu re. Loose or high-imp edance join ts in sig nal

lines are particularly annoying because of intermittent signal behavior

such as decreases

in

signal am plitude, increases

in

noise level, or b oth .

Degradations in system performan ce from high noise levels are fre-

quently traceable to poorly bonded joints in circuit retu rn s an d signal

referencing n etwork s.

Bonding is also important to the performance of interference control

measures. For example, adequate bonding of connector shells to equip-

ment enclosures is essential to the maintenance of the integrity of cable

shields and

to

the retention

of

th e low-loss trans m issio n prop erties of

the cables. The careful bonding of seams and joints in enclosures and

covers

is

essential

to the

achievement

of a

high degree

of

shielding

effectiveness. Interference-reduction components and devices (such as

filters and isolation transformers) also may require proper bonding for

optimum performance. Poorly bonded joints can behave

as

nonlinear

junctions and produce audio rectification, cross modulation, and inter-

modulation effects.

161



162

BONDING

7.2 Bond Eq uivalent Circuits Resistan ce and

Impedance

A primary requirement for effective bonding is th at a low bonding resis-

tance path must be established between two joined objects.

A

bonding

resistance of 1 mQ indicates a high-quality junction. Experience shows

that 1 mQ can be achieved if surfaces are properly cleaned and ade-

quate pressure is maintained between the mating surfaces. There is lit-

tle need to strive for a junction resistance that is appreciably less than

the intrinsic resistance of the conductors being joined.

A similarly low value of resistance between widely separated points

on a ground reference plane or network ensures that all junctions are

well made and that adequate quantities of conductors are provided

throughout the plane or network. In this way, resistive voltage drops

are minimized, which enhances noise control.

It should be recognized th at a low de bond resistance is not a reliable

indicator of the performance of the bond at high frequencies. Inherent

conductor inductance and stray capacitance, plus associated standing-

wave effects and path resonances, will determine the impedance of the

bond. Thus, in RF bonds, these factors must be considered along with

the de resistance.

A low-impedance path is possible only when the separation of the

bonded members is small compared to a wavelength of the EMI being

considered, and the bond is a good conductor. This was discussed in

Chapter 5. At high frequencies, structu ral members behave as trans-

mission lines whose impedances can be inductive or capacitive in vary-

ing magnitudes (depending upon geometrical shape and frequency).

Figure 7.1 shows the equivalent electrical circuit of

a

bond s trap . The

circuit contains resistance due to the finite conductance of the strap in

series with the self-inductance of the bond. Shunt capacitance exists

due to the residual capacity of the strap and its mounting. This capaci-

p

a

Anti-

Resonance

Increasing RF

Figure 7.1

Equ ivalent circuit of bond strap and its impedance.



DIREC T BON DS 163

tance and self-inductance form a parallel antiresonant circuit, resulting

in the adverse impedance response shown in the figure.

7.3 Direct Bon ds

Direct bonding is where specific portions of the surface areas of the

members are placed in direct contract. Electrical continuity is obtained

by establishing a fused metal bridge across the junction by welding,

brazing, or soldering or by maintaining a high-pressure contact

between the mating surfaces w ith bolts, rivets, or clamps. Properly con-

structed direct bonds exhibit a low de resistance and provide an RF

impedance as low as the configuration of the bond members will permit.

Direct bonding is always preferred, but it can be used only when the

two members can be connected together w ithout an intervening conduc-

tor and can remain so without relative movement.

Direct bonds may be either permanent or semipermanent in nature.

Permanent bonds may be defined as those intended to remain in place

for the expected life of the installation and not required to be disassem-

bled for inspection, maintenance, or system modifications. Joints that

are inaccessible by virtue of their locations should be permanently

bonded, and appropriate steps should be taken to protect the bonds

against deterioration.

Many bonded junctions must retain the capability of being discon-

nected without destroying or significantly altering the bonded mem-

bers. Junctions that should not be permanently bonded include those

that may be broken for system modifications, network noise measure-

ments, resistance measurements, and other related reasons. In addi-

tion, m any joints cannot be permanently bonded for reasons of cost.

All such connections not permanently joined are defined as semiper-

manent bonds. Semipermanent bonds include those that use bolts,

screws, rivets, clamps, or other auxiliary fastening devices.

7.3.1 Screws and Bo lts

In many applications, permanent bonds are not desired. The most com-

mon semipermanent bond is the bolted connection (or one held in place

with machine screws, lag bolts, or other threaded fasteners), because

this type of bond provides the flexibility and accessibility. The bolt or

screw should serve only as a fastener to provide the necessary force to

maintain the 85 to 110 kg/cm

2

(1200 to 1500 psi) pressure required



164

BONDING

between

the

contact surfaces for satisfactory bonding. Except for

the

fact that metals

are

generally required

to

provide tensile strength,

the

fastener does not have to be conductive.

Star

or

lock washers

or

lock nuts should be used

to

ensure

the

con-

tinuing tightness of a semipermanent bond, but preferably not

directly

on the

mating surfaces. Figure

7.2

shows

one

recommended

arrangement. Star washers

are

sometimes relied

on for

cutting

through protective and insulating coatings

on

metal such

as

anodized

aluminum

and

unintentional oxides

and

grease films developed

dur-

ing periods between maintenance. But this can cause long-term corro-

sion under the washer teeth .

7 3 2 Soft Solder

Soft solder

is

attractive because

of

the ease with which

it can be

applied. Properly applied to compatible m aterials, the bond provided by

solder is nearly as low in resistance as one formed by welding or braz-

ing. Because of its low melting point, however, soft solder should not be

used

as the

primary bonding material where high currents

may

be

present, as in power fault or lightning discharge paths.

Bonding

or Current

Return Jumper -^

J

Plated Steel, —

/

or CR Steel /

or T itanium

/

Steel Locknut -L^-

or Plate _ / ^ ^ 5

/—

Screw or Bolt

/

/

/ r- Steel Lockwasher

^L/^~

Steel Washer

•3

r- Clean to Base Metal

^ ^ / ^ Area 1-1/2 Dia.

of

l^fet/ Term

Figure 7.2

EMC

-Refinish after Instl.

1-1/2 Dia. of Cleaned

Area

Bonding connections (courtesy AFSC Design Handbook DH1-4



DIRECT BON DS 165

7.3.3 Brazing

Brazing (also including silver soldering) is another metal flow process

for pe rm an en t bond ing. As with w elds, th e resista nce of th e brazed joint

is essentially zero. Since brazing frequently involves the u se of a m etal

different from the primary bond members, precaution must be taken to

protect the bond from deterioration through corrosion.

7.3.4 Welding

In ter m s of electrical performance, welding is th e ideal bonding method .

The inte nse he at involved is sufficient to boil away con tam inatin g films

and foreign sub stanc es. A continuo us m etallic bridge is formed across

the joint; the conductivity of this bridge approximates that of the pri-

mary members. The net resistance of the bond is essentially zero,

because the bridge is very short relative to the length of the bond mem-

bers.

The mechanical strength of the bond is high; the strength of a

welded bond can approach or exceed the strength of the bond members

themselves. Since no moisture or con tam inants can pene trate the weld,

bond corrosion is m inimized.

7.3.5 Cadweld Jo ints

A cadweld joint is obtained by bring ing th e two surfaces togeth er a t a

high temperature and fusing them with a metallic powder, which is

ignited by a special cartridge. The process is extremely dependable and

not subject to corrosion. It is especially recommended for bonds sub-

jected to harsh climatic or corrosive elements.

7.3.6 Conductive Ad hesive Caulking and Grease

Conductive adhesiv e is usu ally in th e form of a silver-filled, two-compo-

nent, thermosetting epoxy resin that, when cured, produces an electri-

cally conductive material. It can be used between mating surfaces to

provide low-resistance bonds. It offers the advantage of providing a

direct bond without the application of heat. When used in conjunction

with bolts, conductive adhesive provides an effective metal-like bridge

with high mechanical strength. It should be used with care, however,

for ther e are indications tha t its properties may deteriorate w ith time.

Conductive grease is used to provide electrical bonding between two

parts that have relative motion such as sliding, rotation, etc. It is usu-

ally a low-resistivity, silver-silicone grease. Applications include:



166 BONDING

•

Sw itche s, bla de s (knife type)

a n d

insula tor suspensions

in

power

substat ions. This reduces arcing, pi t t ing,

a n d E M I

noise

a n d

pre-

vents st icking

by

corrosion

or

arcing.

•

Ball bea rin gs used with noncond uctive pulleys, belts, t ire s, etc. Thi s

reduces

t h e

con sta nt microscopic arcs caus ed

b y

sta t ic charging.

•

Potent iometers

a n d

rot ary sw itches' shafts. Thi s rest ores shield

in tegr i ty

a t

shaft penetrat ion through

t h e

enclosure wall.

Typically, greases have

a

volume resist ivi ty

of

0.02 Q-cm. Th eir ti m e

and tempera ture s tab i l i ty

is

excellent. However, when using them

in

equipment containing printed circui t boards, connectors,

an d so

forth

i t

is necessary

to be

extre mely careful ab out cleanline ss, since even

a

minuscule film

of

grease

c an

crea te

a

short between traces, pins,

etc .

7.3.7 Bonding of Com posite M aterials and Cond uctive

Plastics

Composite materials such as carbon or boron fibers used in aeronautics

pose serious problems for electrical bonding. The first problem lies with

the material

itself.

Carbon fiber composite (CFC), for instance, is made

of layers of carbon fibers em bedded in nonconductive lay ers, at different

angles. The media is both nonisotropic an d nonhom ogeneous. R esistivity

of CFC, depending on the number of plies and their weaving angle,

ranges from 3 mfl-cm to more than 100 mQ-cm. This is three or four

orders of magnitude larger than copper or aluminum. Therefore, it is

pointless to try to achieve de bonding resistan ces much below 1

£1,

since

they will be overridden by the ma terial's poor conductivity anyway.

An effective method for bonding composite materials is to coat the

material with a thin layer of conductive film such as zinc spray, copper

or silver paint, etc. This will not add much weight penalty and can cre-

ate surface resistances of 5 to 100 mQ-cm. This is far superior to the

composite material itself as far as RF bonding and shielding effective-

ness are concerned.

Conductive coatings are widely used, too, in commercial and con-

sumer equipment since the enforcement of national and international

RFI limits. They, as well, pose the problem of making simple inexpen-

sive RF bonds. Making a low-resistance and long-term reliable electri-

cal contact between a ground lug, a filer case, etc., with a sufficient

pre ssu re, is not so easy on a thin film, especially if the u nde rlying m ate-

rial is simply plastic. Figure 7.3 shows some alternative solutions to

this problem.



INDIRECT BONDS

167

Washer

Conductive Plastic

or Coating

Component

to be

Bonded

Threaded °.\

Metallic *g

t

Insert « '

?

©

Preferred, Especially for

Replaceable Items

Acceptable for

One-Time Mounting Only

Avoid

Figure 7.3

Direct bond ing over me tallized plastics or com posites.

7.4 Indirect Bon ds

Operational requirements or equipment locations often preclude direct

bonding. Many times, the metal-to-metal contact provided by the

mechanical fixture is not dependable electrically, such as in the case of

parts that have relative motion, are exposed to corrosion, or are

removed frequently. In such cases, it becomes necessary to dissociate

the electrical function from the mechanical one. When physical separa-

tion is necessary between the elements of an equipment complex or

between the complex and its reference, auxiliary conductors such as

bonding straps or jumpers must be incorporated. Such straps are com-

monly used for bonding of shock-mounted equipment to the structural

ground reference. They are also used for bypassing structural elements

such as the hinges on distribution box covers and equipment covers to



168

BONDING

eliminate th e wideband noise generated by those elements when illumi-

nated by intense radiated fields or when carrying high-level currents.

Bond strap s or cables are also used to prevent static charge buildu p and

to connect metal objects to lightning down conductors to prevent flash-

over.

7.4.1 Jum pers and Bond Straps

Bonding jumpers are short, round, braided, or stranded conductors

used in applications where EMI currents exist at frequencies below

about 10 MHz. They are frequently used in low-frequency devices to

prevent the development of static charges. They are also used to pro-

vide good electrical continuity across tubing members and associated

clamps such as shown in Fig. 7.4. The clamp itself should not be relied

on for continuity, because it is affected by tubing finishes, grease films

and oxides.

To provide a low-impedance p ath at radio frequencies, one mu st min-

imize both the self-inductance and residual capacitance of a bond to

maximize the parasitic resonant frequency. Since it is difficult to

change the residual capacitance of the strap and mounting, self-induc-

tance becomes the main controllable variable. Thus, flat straps are pref-

erable to round wires of equivalent cross-sectional areas.

Bond straps consist of either solid, flat metallic conductors or a

woven braid configuration where many conductors are effectively in

parallel. Solid metal straps are generally preferred for the majority of

applications. Braided or stranded bond straps are not generally recom-

Tab W elded

to Tubing

Clean Tab to

Basic Metal and

Seal After Installation

F ig u re 7.4 Bonding of tubing across clamps.



CORROSION AND ITS CONTROL 169

mended because of several undesirable characteristics. Oxides may

form on each strand of unprotected wire and cause corrosion. Because

such corrosion is not uniform, the cross-sectional area of each strand of

wire will vary throughout its length.

The nonuniform cross-sectional areas (and possible broken strands of

wire) may lead to generation of EMI within the cable or strap. Broken

strands may act as efficient antennas at high frequencies, and interfer-

ence may be generated by intermittent contact between strands. Solid

bond straps are also preferable to stranded types because of lower self-

inductance. The RF impedance of conductors was discussed in

Chapter 5. Because of the increase of impedance with frequency, there

is no subs titute for direct metal-to-metal contact. A rule of thumb for

achieving minimum bond strap inductance is that the length-to-width

ratio of the strap should be a low value, such as 5:1 or less. This ratio

determines inductance, the major factor in the high-frequency imped-

ance of the strap.

7 5

Corrosion

and Its

Control

Corrosion

can

occur between metal parts,

and it

results

in a

nonlinear

junction that may cause undesirable EMI effects. Corrosion can occur

as

a

result

of

either

of

two chem ical processes.

7.5.1 Galvanic Corrosion

The first process, galvanic corrosion, results from the formation of a

voltaic cell between metallic parts with moisture acting as an electro-

lyte.

The degree of the resultant corrosion depends on the relative

positions of the metals in the electrochemical series. This series is

shown in Table 7.1, with the metals listed at the top of the table cor-

roding more rapidly than those at the bottom. If the metals differ

appreciably, such as aluminum and copper (2.0 V difference), the

resulting electromotive force will cause a continuous ion stream with

a significant decomposition of the more active metal as it gradually

goes into solution.

Corrosion caused

by the

electrochemical action between dissimilar

metals is minimized if the combined potential does not exceed approxi-

mately 0.6 V. Us ing 0.6 V as a maximum , Table 7.2 shows the allowable

combinations of mating metal parts. Combinations above the dividing

line (in the shaded area) should be avoided.



170

Table 7 1 Electrochemical Series

Metal

Magnesium

Magnesium alloys

Beryllium

Aluminum

Zinc

Chromium

Iron or steel

Cast iron

Cadmium

Nickel

T in

Lead-tin solders

•Reliable values N/A.

EMF (volts)

+2.37

+0.95

+1.85

+1.66

+0.76

+0.74

+0.44

*

+0.40

+0.25

+0.14

*

BONDING

Metal

Stainless steel (10-18)

Lead

Brass

Copper

Bronze

Copper-nickel alloys

Monel

Silver solder

Silver

Graphite

Platinum

Gold

EMF (volts)

*

+0.13

*

-0.34

*

-0.35

*

-0.45

-0.80

-0.50

-1.20

-1.50

7 5 2 Electrolytic Corrosion

The second process, electrolytic corrosion, also results when two metals

are

in

contact with

an

electrolyte present. However,

the

metals do

not

have to be different; i.e., they can be the same material. In this case,

decomposition

is

attributed

to the

presence

of

local electric currents,

which may be flowing as a result of using a structure as a power system

ground return.

7 5 3 Finishes

Since mating bare metal

to

bare metal

is

essential

for a

good bond,

a

conflict arises between bonding

and

finishing specifications. Oxides

that form on metal are, as a rule, nonconductors. For this reason, it is

desirable that they be softer than the base metal

and as

thin

as

possi-

ble. Oxides of common structure materials like aluminum are much

harder tha n the base metal. So, an ideal contact for bonding would con-

sist

of

plating

the

contact area with

a

metal (such

as

copper) whose

oxide melts at a lower temperature than the metal. However, corrosive

or salt-spray environments

may

exist,

so

this factor usually prevails,

and exposed surfaces must be coated with a protective finish to avoid

corrosion.

For EMI control,

it

is preferable to remove th e finish where bonding

effectiveness would be otherwise compromised. Conductive coatings

generally

do not

need

to be

removed. Most other coatings, however,

are nonconductive and destroy the concept of a bond offering a low-

impedance RF path. For example, anodizing appears to the eye to be

a



T

b

e

7

2

E

e

r

o

h

m

i

c

C

o

o

T

b

e

MM

0

Z

0

0

0

8

Z

Z

0

5

0

0

0

0

0

2

0

1

0

C

0

8

0

3

0

2

0

1

0

AM

0

8

0

3

0

3

0

1

0

0

0

A

=

s

v

A

=

a

u

m

i

n

m

C

=

c

o

m

i

u

m

C

c

m

i

u

m

C

=

c

M

g

=

m

a

u

m

N

=

n

c

R

=

r

h

u

m

S

=

s

a

n

e

s

e

Z

=

z

n

M

0

4

0

3

0

2

0

1

0

0

0

D

0

5

0

4

0

3

0

2

0

1

0

1

0

L

1

0

0

5

0

5

0

3

0

2

0

2

0

1

0

0

0

C

S

1

1

0

6

0

5

0

4

0

3

0

2

0

2

0

1

0

0

0

CN

T

1

C

i

o

0

6

0

4

0

3

0

3

0

2

0

1

0

1

0

5

0

H

CS

•

H

»

J

J

i

S

0

5

0

4

0

4

0

3

0

2

0

2

0

1

0

1

0

C

i

•

i

f

0

5

0

5

0

4

0

3

0

3

0

2

0

2

0

1

0

S

A

•

i

n

i

1

0

6

0

5

0

5

0

4

0

3

0

3

0

2

0

1

0

0

0

N

•

H

I

• i

I i

0

6

0

5

0

4

0

4

0

3

0

3

0

2

0

0

0

0

0

S

i

•

• i i s 1

0

6

0

5

0

5

0

4

0

3

0

2

0

2

0

1

0

RAC

S

i

I

m

i

1t

i

1

l

0

6

0

5

0

5

0

4

0

3

0

2

0

2

0

0

0

C

•

•

•

H

I

•

1

i

• i i

i

0

6

0

5

0

4

0

3

0

3

0

2

0

1

0

0

0

G

P

i

• n

i

I

j

•

1

l

l

l

l

I

•

•

0

6

0

5

0

4

0

3

0

3

0

1

0

1

0

0

0

M

g

M

g

a

o

Z

Z

a

o

8

t

n

2

Z

o

s

e

Z

o

r

o

s

e

A

C

o

s

e

A

M

g

a

o

M

i

d

s

e

D

u

m

i

n

L

C

o

s

e

S

s

o

d

C

o

i

o

s

e

T

n

o

s

e

1

C

S

H

g

C

S

C

c

a

o

S

v

s

o

d

A

n

c

S

N

o

s

e

S

v

R

o

A

n

C

S

v

g

d

a

o

C

b

G

d

p

a

n

m



172 BONDING

good c o n d u c t i v e s u r f a c e

for

b o n d i n g ,

bu t in

r e a l i t y

i t is a n

i n s u l a t e d

c o a t i n g .

7 5 4 Corrosion Protection

The m ost effective way to avoid the a dve rse effects of corrosion is to use

metals (such as tin, lead, or copper) that are low in the electrochemical

activity table. In ma ny stru ctu res (e.g., aircraft) this is not generally

practical due to weight consid erations. Consequently, the more active,

lighter metals such as magn esium and alum inum are employed. How-

ever, stainless steel has been used in many missile programs.

Joined metals should be close together in the activity series

if

exces-

sive corrosion is to be avoided. If dissim ilar m etals m us t be used, select

replaceable components

for

th e object

of

corrosion, such

as

grounding

jumpers, washers, bolts, or clamps, rath er th an structural members.

Thus, the smaller mass should be of the high er p oten tial (cathode),

such as steel washers for use with brass structures. For instance, bond-

ing a steel box with a copper strap will result in min imal corrosion d ue

to reduced cathode surface. Also, the part that deteriorates will be the

replaceable one.

When joined members are widely separa ted in the activity table ,

plating may be used to help reduce the dissimilarity. Som etimes it is

possible to electrically insulate metals with organic and electrolytic fin-

ishes and seal the joint against moisture to avoid corrosion. However,

this is an unacceptable practice for EMI control. One solution for elec-

trolytic corrosion is to avoid th e u se of stru ctu re or equipm ent housing

for power ground ret ur n. Any anticipate d corrosion should occur in eas-

ily replaceable items, as previously mentioned.

A galvanic cell requires the presence of an electrolyte to function.

Therefore, joints should be kept tight and well coated after bonding to

prevent the entrance of liquids or gases that can act as an electrolyte. If

a joint involves dissimilar metal contact, coating just one of th e elec-

trodes is not sufficient. Com plete coating, or at least sealing the edges,

is required.

7.6 Equipm ent Bon ding Prac tices

This section presents design and construction guidelines for effective

bonding of equ ipm ent circuits, enclosures, and cabling. T hese guide-

lines are not intended as step-by-step procedures for meeting EMC

specifications and standards. Instead, they are aimed at focusing atten-



EQUIPMENT BONDING PRACTICES 173

tion on the principles and techniques, listed below, that lead to

increased EMC between circuits, assemblies, and equipments.

1.

Welded seam s should be used where ver possible because they are

permanent and they provide a low-impedance bond.

2. Spot welds may be used where RF tightness is not necessary. Spot

welding is less desirable than continuous welding because of the

tendency for buckling and the possibility of corrosion occurring

between welds.

3. Soldering should not be used whe re mecha nical s tren gth is

required. If mechanical strength is required, the solder should be

supplem ented w ith screws or bolts.

4.

Fasteners such as rivets or self-tapping sheet metal screws should

not be relied upon to provide the primary current path through a

joint.

5.

Rivets may be used to provide mechanical stre ng th to soldered

bonds.

6. Sheet metal screws should not be used to secure an electrical bond.

The following precautions should be observed when employing bond-

ing straps or jum pers.

1.

Ju m pe rs should be bonded directly to the basic struc tur e ra th er

than through an adjacent part .

2. Ju m pe rs should not be installed w ith two or more in series.

3.

Jumpers should be as short as possible.

4. Jumpers should be installed so that vibrations or motion will not

affect the impedance of the bonding path.

If electrical continuity is required across shock mounts, bonding

jumpers should be installed across each shock mount. Jumpers for this

application should have a maximum thickness of 0.063 cm (0.025 in) so

the damping efficiency of the mount is not impaired. In severe shock

and vibration environments, solid straps may be corrugated, or flexible

wire braid may be used.

W here RF shielding is require d a nd w elded joints cannot be used, th e

bond surfaces must be machined smooth to establish a high degree of

surface contact throughout the joint area. Fasteners must be positioned

to maintain uniform pressure throughout the bond area. Chassis

mounted subassemblies should utilize the full mounting area for the

bond as illustrate d in Figs. 7.5 and 7.6. Se par ate jump ers should not be

used for this purpose. Equipment racks provide a convenient means of

maintaining electrical continuity between rack mounted chassis, pan-



174

BONDING

Direct

Bonding Method

(Preferred)

Bond Area

(Clean both members over entire

mating surface.)

Figure 7.5

Bonding of subassemb lies to equipm ent chassis.

Clean Faying

Surfaces at

all

Four Comers

Figure 7.6 Bonding of equipm ent to mo unting mem bers.

els,

and ground planes. They also provide an electrical interconnection

for cable trays. Typical equipment cabinets with the necessary modifi-

cations to provide such bonding are shown in Figs. 7.7 through 7.10.

Bonding between equipment chassis and rack is achieved through

equipment front panel and rack right angle brackets. These brackets

are grounded to the u nis trut horizontal slide tha t is welded to the rack

frame. The lower surfaces of the rack are trea ted with a conductive pro-

tective finish to facilitate bonding to a ground plane. The ground stud at

the top of the rack is used to bond a cable tray, if used, to the rack struc-

ture, which is of welded construction.



EQUIPMENT BONDING PRACTICES

175

Clean to

Base M etal

Rack

Clean Flange

to Base Metal

F ig u re 7.7 Typical method of bonding equipment flanges to frame or rack.

Clean each mating surface 3.2 mm (1/8 )

around the bushing periphery.

Rear of

Electronic

Equipment

v?

Dagger Pins

F ig u re 7.8 Bonding of rack-moun ted equipm ent employing dagger pins.

Figure 7.10 illustrates a typical bonding scheme of a whole cabinet

intended for very severe EMI requirements. Cable trays are bonded

together, and the cable tray is bonded to the cable chute. The cable

chute is bonded to the top of the rack or cabinet; the cabinet is bonded

to the flush-mounted grounding insert (which is welded to the ground

grid; and the front panel of the equipment is bonded to the rack or cabi-



176

BONDING

Cadmium Plated

Surface

Welded to

Cabinet

Horizontal

Slide

Cadmium

Plated

Front Panel

Mounting

Surface

Grounding

Stud

Figure 7.9

Recommended practices for effective bonding in cabinets.

net front-panel mounting surface. Nonconductive finishes are removed

from the equipment front panel before bonding. The joint between

equipment and cabinet may serve a dual purpose—that of achieving a

bond and that of preventing interference leakage from the cabinet if the

joint is designed to provide sh ielding.

If shielding is a requirement, conductive gaskets should be used

around the joint to ensure that the required metal- to-metal contact is

obtained. If equipment is located in a shock-mounted tray, the tray

should be bonded across its shock mounts to the rack structure. Con-

nector mounting plates should use conductive gasketing to improve

chassis bonding. If chassis removal from the rack structure is

required, a 25.4 mm (1 in) wide braid with a vinyl sleeve should be

used to bond the back of the chassis to the rack. The braid should be



SUMMARY OF BONDING PRINCIPLES

177

Cable Chute

Cable Tray

Rack-to-Grounding

Insert Bond

Flush Mount Insert

ith Floor

Grounding

Insert

Weld to Grid

Ground

Grid

Welded or Explosive )

Fused Interconnections

Fig ure 7.10 Typical bonding scheme for severe EMI requirements.

long enough to permit partial withdrawal of the chassis from the

rack.

7.7 Su m m ary of Bo nding Prin ciple s

1. Bonds m us t be designed as com ponents of th e grou nding system,

because th ey affect the system's overall p erformance.

2.

Electrical continuity an d m echanical fastening are two different

functions, and they should be considered separately. Fasteners,

spring washers, threads, etc. are strictly to apply mechanical pres-

sure; then the curre nt can flow throug h base m etal ma ting surfaces.

3. Bonding must achieve and m ainta in in tim ate contact between

metal surfaces. The surfaces must be smooth, clean, and free of

nonconductive finishes. Fasteners must exert enough pressure to

hold the surfaces in contact in the presence of the deforming



178 BONDING

stresses, shock, and vibration associated with the equipment and

its environment.

4. The effectiveness of the bond depends on its construction, the fre-

quency and magnitude of the currents flowing through it, and the

environmental conditions to which it is subjected.

5.

Bonding jumpers are only a substitute for direct bonds. If the jump-

ers are kept as short as possible, have a low resistance, have low

length-to-width ratio, and are not higher in the electrochemical

series than the bonded members, they can be reasonable substi-

tu tes .

6. Bonds are always best when similar metals are joined. If this is not

possible, attention must be paid to selecting metals that will mini-

mize corrosion, using supp lemen tary components, such as was hers,

to ensure that corrosion will affect replaceable components only,

and the use of protective finishes.

7. Even if th e m etals are similar, a protective coating m ust be pro-

vided if moisture or con tam inants are presen t.

8. Finally, throughout the lifetime of the equipment, system, or facil-

ity, the bonds m ust be inspected, tested, an d ma intained .

Suggested Readings: Bonding

[1] M ardiguian, Michel,

Grounding and Bonding,

Vol. 2, A Handbook

Series on Electromagnetic Interference and Compatibility, Gaines-

ville, VA: Inte rferen ce Con trol Technologies, 1988.

[2] W hite, Donald R. J. and M ardiguian , M ichel, EMI Control Meth-

odologies and Procedures,

Vol. 8, A Handbook Series on Electro-

ma gnetic Interfe rence and Com patibility, Gaine sville, VA:

Interference Control Technologies, Gainesville, 1988.

[3]

Duff,

William G., EMC Design of Electronic Sys tem s,

EMC EXPO

88 Symposium

Record, Gainesville, VA: Interference Control Tech-

nologies, 1988.

[4] MIL-HNBK-419,

Grounding, Bonding and Shielding for Electronic

Equipment and Facilities,

[5]

MIL-B-5087B, Bonding, Electrical and Lightning Protection for

Aerospace Systems,

October, 1964.

[6] M IL-STD-188-124,

Grounding, Bonding and Shielding.

[7] Denny, Hugh, et al.,

Grounding and Bonding,

Vol. 2, Gainesville,

VA: DWCI P res s, 1988.



SUMMARY OF BONDING PRINCIPLES 179

[8]

Morrison, R., Grounding and Shielding Techniques in Instrumenta-

tion, 3rd ed., Hoboken, N J: Joh n W iley & Sons, 1986.

[9] Grou nding of Ind us tria l an d C omm ercial Power System s, ANSI/

IEEE Std. 142-1992, Piscataway, NJ: IEEE, 1992.

[10] Um an, M. A., Lightning, M ineola, NY, Dover Publica tions, 1984.

[11]

H art, W. C , and E. W. Malone, Lightning and Lightning Protection,

Gainesville, VA: Don W hite C ons ultan ts, 1985.

[12] Golde, R. H., Lightning Protection, Gloucester, MA: Chemical Pub-

lishing Co., 1973.

[13] Fisher,

F.

A., R. A. Pe ral a, an d J . A. Plum er,

Lightning Protection of

Aircraft, Pittsfield, MA: Lightning Technologies, Inc.

[14] Natio nal Electrical Code (NEC ), Quincy, MA: Natio nal Fire Protec-

tion Association, 2 002.

[15] Kervill, Gregg,

The Practical Guide to Electrical Product Safety,

UK: M&M Business Comm unications, Ltd.

[16] Shipbo ard Bonding, G rounding, and O ther Techniques for EMC

and Safety, MIL-STD-1310.



Chapter 8

Filters,

Ferrites, Isolators, and

Transient Suppressors

There are several different types

of

EMI control devices t h at may

be

placed in

a

conducted p ath (either sign al or power lines)

to

selectively

pass intended signals and reject unintended EMI signals. The rejection

is provided on the ba sis of some charac teristics of the EMI sign al, which

differs from the intended signal. Thus, these EMI control devices pro-

vide a means of suppressing conducted interfering signals tha t have

certain characteristics. Filters, which are discussed in Section 8.1, dis-

criminate between desired and interfering signals on the basis of fre-

quency. Ferrites may also be used to provide frequency selectivity, and

these devices

are

discussed

in

Section 8.2. Isolators, which

are

dis-

cussed in Section 8.3, discrim ina te betw een com mon-mode and differ-

ential-mode signals existing

in the

conducted path . Transien t

suppressors, which are discussed in Section 8.4, discrim inate betwee n

signals on the basis of signal level. All four of thes e device types are

very important

in

system applications, because th ey can usu ally

be

used at equipment inputs or outputs to control EMI problems th at

occur as

a

resu lt of integ rating the equipm ent into

a

system.

8.1 Filters

An electrical filter is a network of lumped or distributed resistors,

inductors, and capacitors that exhibit signal selectivity as a function of

frequency. T hu s, an EM I filter is one th at pas ses signals whose frequen-

cies are in certain ranges or bands, called the

passbands,

and blocks, or

attenuates, signals whose frequencies are in othe r ran ges , called th e

stopbands.

The nature

of

the am plitud e function

or

th e loss function

may be used to classify the various types of filters according to the loca-

tion of the ir p ass- and stopb ands . An ideal filter is one th at ha s a linear

181



182 FILTE RS, FERRITE S, ISOLATORS, AND TRANSIENT SUPPRESSORS

p h a se r e sp o n se in i t s p a s sb a n d , z e ro lo s s i n i t s p a s sb a n d , a n d in f in i t e

loss in i t s s top ban d . A l th ou gh a n idea l filter r esp on se i s phy s ica l ly

unreal izable , i t is possible to design f i l ters that have low loss in the

p a s s b a n d a n d s ig n if i ca n t a t t e n u a t i o n i n t h e s t o p b a n d .

EMI f i l t e r s a re inse r ted be tween the source o f EMI and the load . The

filter a t te nu a t es th e leve l of no ise re ac h in g th e load e i t he r by d is s ipa t -

ing the RF energy as hea t o r by re f lec t ing the energy back to the source .

D iss ip a t iv e filte rs a r e tho se w i th res i s t iv e e le m en ts . Lossy fe r r i te s a r e

us ed in som e f ilters to p rov id e e le m en ts th a t ap pe ar res i s t iv e above

50 MHz or so .

Th e m os t of ten en co un te re d ty pe s of f req uen cy selec t ive fi lters ar e

defined as follows:

• A low -pa ss f ilter is one w ith a s ingle p as sb a nd below a cutoff f re-

q u e n c y {Q w ith a l l f requen c ies h ig he r th a n f c on s t i tu t ing th e s top-

b a n d .

• A hi gh -p as s f ilter is one w it h s top ba nd below a cutoff f req uen cy f

and a passband for a l l f requencies above f(>

• A ba nd -p as s filter i s one w i th a pa ss ba nd be tw ee n two cu to f f f re -

que nc ies fL an d fu an d s to pb an ds over th e re m ai nd er o f th e f re-

q u e n c y s p e c t r u m .

• A ba nd -re jec t f ilter is one w it h a s t op ba nd bet w ee n tw o cutoff f re-

que nc ies f]^ an d f an d pa ss ba nd s over th e re m ai nd er of th e f re -

q u e n c y s p e c t r u m . ( O t h e r t e r m s u s e d a r e

band-elimination

a n d

band-stop.)

A t te n u a t i o n o v e r a p r e sc r ib e d f r e q u e n c y r a n g e i s p e rh a p s t h e m o s t

com mo n w ay of spec i fy ing filte r s pe c t ru m per f o rm anc e a nd i s a l so one

of th e mos t ab us ed te rm s in EM I f ilters . Filter attenuation re fe rs to th e

ra t i o of ou tp u t vo l tag es , before an d af ter f ilter in se r t i on , as a funct ion of

f r eq u e n cy . F ig u re 8.1 i l l u s t r a t e s t h e a t t e n u a t io n c h a ra c t e r i s t i c s a s a

func t ion of f requ ency for ea ch of th e fi lter ty pe s des cr ib ed ab ove .

F i l te r s a re us ed in sys t em EM I con t ro l in one o r m ore of th e fo l lowing

w a y s :

• R F su p p re s s i o n of u n w a n te d s ig n a l s o th e rwi s e e n t e r in g o r e x i t i n g

f rom the power l ines o f ac power mains .

• R F i so la t io n of com m on- im ped anc e coup led c i rcu i t ry , suc h as sev-

e ra l ne tw or ks fed f rom co mm on power supp l ies , v ia low -pass filters.

• R F su p p re s s io n of u n w a n te d E M I a t t h e s ig n a l i n p u t of su sc e p t ib l e

d e v ic e s su c h a s a n a lo g a mp l i f i e r s , a n a lo g c o n t ro l c i r c u i t s , c o m mu n i -

c a t i o n e q u i p m e n t s , e t c .



FILTERS

183

S

I

Frequency

(a ) Low Pass Filter

Frequency

(b ) High Pass Filter

73

1

I

Frequency

(c )

Band Pass Filter

Frequency

(d ) Band Reject F ilter

Figure 8.1 Characteristics of

various

filter types.

• Conducted bro adb and noise supp ression from power tools, appli-

ances, industrial machinery, office equipment, and other devices

developing transients due to arc discharge at the brush-commuta-

tor interface of motors.

• Conducted broa dba nd noise supp ression from non-motor, tran sie nt-

developing devices such as fluorescent lamps, electric ignition sys-

tems, industrial controls, relays and solenoids, and other switching-

action devices.

• Protection of susceptible devices such as tran sd uc ers , compu ters,

and electro-explosive devices.

With some exceptions, EMI filters are characterized by having

unequal input and output impedances when installed in their opera-

tional environments. For example, impedance sources of power mains

are frequently less than 1 Q at low frequencies, while their loads rep-



184 FILTE RS, FERR ITES, ISOLATO RS, AND TRANSIENT SUPPRESSORS

resent high impedances. Furthermore, both source and load imped-

ances are frequency dependent. Emphasis for system-level filtering is

to suppress the source when feasible rather than protecting suscepti-

ble circuits.

A num ber of different types of filters are commercially available for

use in system-level applications. The systems engineer should give

careful consideration to the various factors that influence filter perfor-

mance so that the proper choice may be made. The major factors that

should be considered in selecting a filte r are given below.

• Identify the filter type required (i.e., low-pass, high-pass, band-

pass, or band-reject). The frequency characteristics of the intended

signal and the interference will influence this decision.

• Define the cutoff frequency or frequencies required. This will be pri-

marily determined by the characteristics of the intended signal.

• Define the attenuation required in the stop band as a function of

frequency. This will determine the number of elements that will be

required for the filter. In general, each element will contribute

20 dB/decade of attenuation in the stop band as shown in Fig. 8.2.

Note that there is, in general, a maximum attenuation that may be

expected from a filter, and th is maximum attenuation will be a

function of the num ber of elements and the manner in which the fil-

ter is installed.

• Define the installation configuration for the filter. For example, if

the filter is to be installed at the input or output of an equipment,

does the equipment have a shielded enclosure? As mentioned above,

this will have a major impact on the maximum attenuation tha t the

filter may be expected to provide. Table 8.1 provides information on

the maximum values of attenuation that a filter may be expected to

provide in various frequency ranges relative to the cutoff frequency

and for various installation practices involving shielded equipment

compartments and connectors, shielded com partments only, and no

shield.

• Define the inpu t and output impedances th at the filters will

encounter in the operational configuration. This is very important,

because the terminating impedances can have a major impact on

the filter performance. For example, Fig. 8.3 compares the attenua-

tion provided by a single shunt capacitor (C = 0.63 pF) or a single

series capacitor (L = 1.6 mH) when installed in applications involv-

ing different values for the source and load impedances. Referring

to the

figure,

t is obvious that the filter attenuation is dependent on

the terminating impedances.



FILTERS

185

5

5

1

15

2

25

35

4

5

55

6

65

7

Pass Band

ii

It

\

1

\

\

\

\

i

n =

4;

80dB/Decade

-

1

\

\ i

>

\

S

\

V

\

s

\

V

\

1

\

\

2,

r

\

s

\

\

V

\

D

s

\

\

n

=

Number of Stages

\

V

1

s

Stop Band

\

s

50dB Default Limit

60dB Maximum Default

L

in

B LSt III

eu

70dB Default Limit

.2

1 2

3 5 10 20 30 50 100 200 500 Ik 2k

Relative Frequency in U nits of

Figure 8.2 Filter response vs. number of stages and frequency.

In summary, it is emphasized that caution should be exercised in the

selection and installation

of

filters

for

systems applications.

In

particu-

lar, it should be recognized that the performance of

filters

will be depen-

dent

on the

terminating impedances

(and

therefore

may be

quite

different from the filter specifications), and the maximum attenuation

will be dependent on installation. In general, attenuations of more than

100 dB

are

difficult

to

achieve

due to

input-output crosstalk coupling,

and the filter may completely degenerate in performance a few decades

above cutoff due to parasitics. Where open circuitry

is

used, not involv-

ing either connectors or filter shields, it is not uncommon to have direct

input-output coupling of the order of 40 to 60 dB, especially in minia-

ture

and

integrated circuits. Regarding parasitics, unless special

pre-

cautions are taken in the filter design and fabrication, a filter may offer

little

to

no attenuation

at

two

or

more decades above cutoff. This

is the

very essence of most EMI filters. They will continue to offer a pre-

scribed amount of attenuation up to 1 GHz, 10 GHz, or whatever the

rated value may be.



186

FILTERS, FERRITES, ISOLATORS, AND TRANSIENT SUPPRESSORS

Table 8.1 Typical Maximum Attenuation of Electrical Filters outside their Passb ands

Reject ion-Band Shield and

Frequency Range Connectors

f

co

<f<10f

co

1 0 f

c o

< f < 1 0 0 f

c o

f>100f

co

Shield Only

Microminiature or IC Filters

NA 60 dB

NA 40 dB

NA 20 dB

Communications Filters (No Special EMI Precautions)

f

co

<f<10f

co

80 dB 70 dB

10 f

co

<

f < 100 f

co

60 dB 50 dB

f>100f

co

40 dB 30 dB

f

co

<f<10f

co

1 0 f

c o

< f < 1 0 0 f

c o

f>100f

co

f

co

<f<10f

co

1 0 f

c o

< f < 1 0 0 f

c o

f>100f

co

f

co

<f<10f

co

1 0 f

c o

< f < 1 0 0 f

c o

f>100f

co

Communications Filters (EMI Hardened)

90 dB NA

80 dB NA

70 dB NA

Power Line Filters

<

80 dB

80 dB

70 dB

Power Line F ilters >

100

dB

100

dB

90 dB

10

Amps (EMI

Type

NA

NA

NA

10 Amps (EMI Type)

NA

NA

NA

No Shield/

Connectors

50 dB

30 dB

20 dB

60 dB

40 dB

20 dB

NA

NA

NA

NA

NA

NA

NA

NA

NA

The following sections provide discussions of some of the consider-

ations that apply specifically to power line or signal filters.

8.1.1 Pow er Line Filters

Most conducted forms of system EMI result from equipment or systems

sharing the same source of ac power mains. Here, an electrical noisy

source may pollute the power distribution wiring by injecting broad-

band emissions into wires also feeding other potentially susceptible

equipm ents. Ano ther m echanism involves common impedance coupling

in which two or more circuits are fed from a common regulated or

unregulated power supply. On the other hand, it frequently develops

that a potentially susceptible equipment sharing a common power bus

with an EMI source may not be affected thereby. Rather, th e power line

may have been electromagnetically contaminated to begin with, and

the m utu al connection thereto is academic.

Power lines feeding a given area can act as pickup antennas for

broadcast, shortwave, HF, FM, TV, communication emissions, radar,



FILTERS

187

Attenuation (Insertion Loss) of a Single Element Filter in a

50 Ohm & in a Low

Impedance Source

&

Load System

A

d B

= 20 Log

10

= 20Log

1(

1 +

jcoL

coL

f

c

=10kHz; C = 0.63p£

For coL » Z

g

+ Z

L

g

Z

L

= 50O;

coL=100

f

c

= 10kHz; L = 1.6mh

1kHz 3

10kHz

30

100kHz

300 1MHz 3

10MHz

30

100MHz

Frequency

Figure 8.3 Attenuation (insertion

loss)

of a single-element filter in a

50-£2

and

a low-impedance source and load system.

etc.

across the frequency spectrum. Furthermore, these lines can con-

duct wideband ignition and overhead fluorescent-lamp noise, harmon-

ics from the ac power mains, nearby office and machine noise, and

virtually any electrical noise that couples to the input power lines by

electric, magnetic, or electromagnetic means. Since these potentially

disturbing noises can cause EMI in sensitive equipment, it is para-

mount to filter them out—preferably before they get to user areas. This

is accomplished by the use of power line filters. They must pass the de,

60 Hz, and/or 400 Hz power-mains frequencies with very little attenua-



188


tion (e.g., 0.2 dB or less) and provide per ha ps 60 dB or more atte nu ati on

from a low frequency such as 10 kHz to 10 GHz, or another frequency

depending on the EMI bounds of potential susceptibility.

Any discussion of power line filters requires an understanding of

common and differential modes. Single-phase power is generally pro-

vided via three terminals: line, neutral, and ground. This makes the ac

power port a two-port network (line to ground and neutral to ground).

These ports could be treated as independent noise sources. However, it

is much more convenient to deal with the two-port network in terms of

its common-mode and differential-mode equivalent circuits, each of

which is a one-port (two-terminal) network. The reasons are described

below.

Noise that is conducted through the power mains into an equipment

generally appears on the line and neutral leads at the same potential

with respect to ground. Noise with thi s chara cteristic is defined as com-

mon-mode (CM) noise. The noise emissions of electronic equipment

with linear power supplies often are primarily common-mode as well.

This type of noise behaves as if the line and neutral leads were con-

nected in parallel. CM noise circulates between this pair and the

ground terminal. Thus, the CM equivalent circuit of a noise source or a

filter treats the line and neutral terminals as though they were a single

term ina l, referenced to ground (G). Figu re 8.4 illu stra tes power line sit-

uations involving common-mode EMI.

Switch-mode power supplies produce noise that includes both com-

mon- and differential-mode components. Differential-mode (DM) noise

is that which appears on the line and neutral leads at the same magni-

tud e bu t 180° out of ph ase . Othe r sources of DM noise pre sen t on the ac

line include locally generated switching transients, motor noise, etc.

Differential-mode noise current circulates only between the line and

neutral leads, as shown in Fig. 8.5. Thus, the DM equivalent circuit

includes only the line (L) and neutral (N) terminals.

Any power line noise characteristic can be completely represented by

its C M and DH components. Thus, in the power line environment, noise

Equipment Equipment

F ig u re 8.4 Common-mode EMI.



FILTERS

Equipment

189

o—

G

Figure 8.5 Differential-mode EMI.

behaves as if it were derived from independent CM and DM sources.

Power line filters also behave as if they were composed of CM and DM

filters, where some components and characteristics function only for

CM, wh ile others function only for DM.

Common-Mode Inductors

Common-mode inductors are those for which high values of common-

mode inductance and operating current are achieved at the expense of

differential-mode indu ctanc e. This is accomplished by providing identi-

cal windings on a common core for all current carrying lines to be fil-

tered. A single-phase example is shown in Fig. 8.6; however, the

technique may also be applied on multiphase filters.

Identical line and neutral side windings are arranged on a core such

that the flux developed in the core cancels when currents in the wind-

ings are equ al bu t of opposite ph ase . Such cu rren ts are differential

mode and includ e th e ac power cu rren t. Th us, ideally, th e ac power cur-

rent does not generate flux that could saturate the core. CM currents

th at are in phase on the L and N windings generate flux tha t ad ds.

This technique permits the design of compact large-value CM induc-

tors that tolerate much larger values of ac line current than would a

comparable inductor with only one winding.

Because the common flux and thus the mutual inductance cancel for

DM currents, the full inductance is realized only for CM currents.

Core Saturation Effects

The permeability of materials used in inductor cores is a function of

mag netic field. Above low levels of excitation, in sta nta ne ou s permeab il-

ity (and thus inductance) decreases with instantaneous magnetic field;

i.e., the core sat ur ate s. The m axim um field in the core is propo rtional to

the peak current in the winding. Thus, the current that a filter may

handle while providing its design performance is limited.

Fortunately, the ability to separate noise and filter performance into

CM and DM components permits a convenient solution to the design of



190

FILTERS,

FERRITES, ISOLATORS, AND TRANSIENT SUPPRESSORS

H

I 1/2 i

CM

1/2 l

C M

Common Mode Noise Current

( NCL I

MAINS)

ONCL

AC Line and Differential Mode Noise Currents

F ig u re 8.6 Common-mode inductor.

h i gh - i ndu c t anc e com pon en t s t h a t a r e capab l e of ca r r y i n g s i gn if ican t

l evel s of power cur rent . The so lu t ion i s the common-mode inductor .

Leakage Inductance

In prac t i ce , cancel l a t ion i s not per fec t . Some f lux genera ted by one

winding l eaks out of the core before i t can cancel l eakage of the o ther

winding . Thi s i s ca l l ed leakage flux. Th e i ndu c t an ce co r r e s pon d i ng t o

this leakage f lux i s cal led

leakage inductance.

I t i s th e va lu e of ind uc -

tance achieved for DM cur rent s and i s genera l ly l es s than a f ew percent

of the CM value .

Again, because the f lux cancel lat ion i s not perfect , the core does sup-

po r t some flux as a r es u l t of th e ope ra t i ng cu r re nt . T hu s , avoidan ce of

core sa tura t ion s t i l l p l aces a l imi t on the usable opera t ing cur rent of a

g iven CM inductor .

A ba s ic po we r l ine EM I f ilter i s i l l us t ra te d in Fig. 8.7. Ha vi ng

descr ibe d the op era t io n of th e CM inductor , i t i s now ap pr op r ia t e to

present the equivalent ci rcui t models for power l ine f i l ters . These mod-



FILTERS

191

N a

rrrr

c

lg

-O L

x

O G

O N

Figure 8.7 Typical power line EMI filter.

els illustrate the effects of components in the filter upon the CM and

DM performance and provide the basis for quantitative analysis. A rep-

resentative filter schematic is shown in Fig. 8.8. This filter employs

both independent and CM inductors.

Common-Mode Equivalent Circuit for Filter

The common-mode equivalent circuit for this filter is shown in Fig. 8.9.

The CM inductor appears in the model as its total self-inductance.

Because the line and neutral leads are essentially in parallel, the line-

to-ground capacitors (C

lg

) and the independent inductors appear in the

model in parallel, while line-to-line capacitors (C^) do not appear in the

model. Parasitic elements (Rp, Cp, 1, and r) are also shown. Each repre-

sents values for individual components (windings in parallel for the CM

coil).

Parasitics

These parasitic elements degrade filter attenuation from the ideal. Cp

causes the inductors to appear capacitive at moderate to high frequen-

cies.

Also,

1

causes C

l g

to series resonate and turn inductive at higher

frequencies. Both have the effect of reversing the attenuation slope of

the filter at higher frequencies (i.e., the low-pass filter is turned into a

high-pass filter).

CM

Independent

Inductor Inductors

Fig ure 8.8 Representative filter schematic.



192

FILTERS, FERRITES, ISOLATORS,

AND TRA NSIENT SUPPRESSORS

Rp 1/2 Rp

Lp

Cp

CM Coil

1/2 Lp

(Clgl + Clg2)

-o

2Cp

Independent

Inductor

-1/21

-1/2 r

Figure

8.9

Common-mode equivalent circuit.

High-Frequency Performance

It is important to maintain the high-frequency performance of

the

filter.

Conducted emissions performance requirements extend

to 30 or

50

MHz.

Also, radiated emissions problems at higher frequencies

may

be due to high-frequency noise th at is conducted either into or out of the

electronic device via the power

cable.

These problems may be addressed

through filter attenuation at the point where the cable enters the

device.

Controlling Parasitics

The effects that degrade high-frequency performance may be controlled

by winding techniques th at minimize Cp and by assembly and construc-

tion techniques th at limit 1. One technique for significantly reducing

1

is

to provide separate input

and

output leads

on the

capacitor, thus turn-

ing it into a three-terminal device

(Fig.

8.10).

The

ground connection

must, of course, be kept extremely short. This technique removes the

inductance

of

each long lead from

the

shunt branch. It thereby greatly

increases the series resonant frequency of the C

lg

structure.

While these techniques for extending high-frequency performance

are quite effective, they alone generally will not permit compliance

to

Tempest requirements, which extend to 1

GHz.

These filters require

additional elements

to

maintain high performance up to

1

GHz.

Differential-Mode Equivalent Circuit

for

Filter

The differential-mode equivalent circuit for the filter of

Fig.

8.8 is

shown

in

Fig. 8.11. Here the line and neutral leads are

in

series, and

current flows hrough these leads in opposite directions. Thus the inde-



FILTERS

193

3-Terminal

Capacitor

2-Terminal Capacito r

Fig ure 8.10 Limiting parasitic inductance by providing separate input and

output leads on capacitor.

Rp

Lp

2Rp

2Lp

Cp

CM

Coil

1/2

Cp

_Ji

Independent

Inductor

~ 2 r

Fig ure 8.11 Differential-mode equivalent circuit.

pendent inductors are in series for the DM model, and the CM coil

appears as its leakage inductance and the associated parasitics. These

elements, I/p, C'p, and R'p, are those measured on a CM coil with the

windings connected series opposed (i.e., in differential mode).

Capacitors

for

this model

are

only those tha t appear between line

and neutral; i.e., the line-to-line capacitors and the series combination

of the line-to-ground capacitors. Due to their relative magnitudes, C

lg

is often ignored when it is in parallel with C;Q. However, this simplifica-

tion ignores the higher-frequency effects of series resonance of the C^

arm and the parallel resonance of the combination of C^ and C

lg

.



194 FILTERS, FERRITES, ISOLATORS, AND TRANSIENT SUPPRESSORS

Basis for Filter Selection

This section deals with the issue of determining which filter will solve a

problem in a given application. I t is very tempting to use published val-

ues of insertion loss as a tool for selecting a filter that will provide a

given level of attenuation in a circuit. Unfortunately, insertion loss is

only a measure of a filter's RF attenuation when measured in a 50-Q

system. This impedance bears little relation to the ac power line port on

an electronic device. Since a large part of filter attenuation comes from

mismatching the terminations, performance with nonrepresentative

terminations will not predict in-circuit performance.

Insertion loss in useful only as a means of verifying the uniformity of

a product over time and as a qualitative means of evaluating filters

with identical schematics.

Nevertheless, a curve of attenuation versus frequency could be used

as a basis for filter selection. It would, however have to represent per-

formance achieved when the filtered is terminated by the device to be

filtered on one end and by the ac line (or equivalent) on the other.

Filters for Linear Power Supplies

Conducted immunity and emission problems on devices with linear

power supplies generally involve moderately high-frequency noise that

becomes stray coupled between the ac line and the digital electronics.

Simple wideband models do not exist either for the coupling mode or for

the input circuit of the supply. Thus, an analytical approach is not

available to solve these problems. The selection process for filters in

these applications is usually empirical. Filters for these applications

generally provide performance in the range of

1

MHz and above and are

primarily CM filters. Emphasis is placed on controlling element para-

sitics and placement so that high-frequency performance is preserved.

Options for enhanced performance include increased element values

that extend performance to lower frequencies and varieties in configu-

ration that provide steeper attenuation slopes and that may more effec-

tively mismatch the terminations.

Once a filter is selected, it is importan t to note th at it should not be

considered to be interchangeable with another filter of comparable

element values. The component parasitics shown above can make

similar-looking filters perform quite differently. Saturation effects in

the inductors can also cause filters with similar small signal element

values to perform quite differently. Finally, stray coupling between

components within a filter can seriously degrade high-frequency per-

formance. These are not indicated by the element values. Inter-

changeability is only assured by testing in the equipment to be

filtered.



FILTERS

195

Filtering for Switch-Mode Power Supplies

The performance of a filter required to limit emissions on a switched-

mode power supply (SMPS) powered equipment is generally dictated by

the noise that the SMPS itself generates. These supplies generally

require significant CM and DM filtering in order to comply with FCC,

VDE, and MIL requirements. This performance can be expensive.

8.1.1.1

Other Considerations

The discussions above addressed methods of analyzing and selecting fil-

ter attenuation levels. There are additional issues to consider when

selecting a filter.

Filters, as well as other components connected across the ac line,

must be safe to avoid potential fire and shock hazard. Various safety

agencies, including UL, CSA, VDE, etc., have set standards that pro-

vide guidelines for the designer of ac components. Components that

carry the compliance symbols from these agencies have been designed

and manufactured to comply with these standards. In addition to issues

of construction and design, agencies also specify parameters such as

line-to-ground leakage current, ac or de hipot, insulation resistance,

temperature rise, creepage distance, material temperatu re and flam-

mability ratings, temperature coefficients, environmental stress, pulse

withstanding, etc. Compliance with these safety agencies is always ben-

eficial, and in some countries (Switzerland, Denmark, Norway, Sweden)

mandatory.

Filters m ust also be capable of supporting the large inrush currents

drawn by many types of equipment upon turn on. Overcurrent due to a

fault in the equipment should not cause the filter to become a fire haz-

ard during the time it takes to open the fuse.

8.1.1.2

Filter Installation

Well designed filters provide effective isolation between their line- and

load-side terminations. However, this isolation can be degraded by the

way in which the filter is installed.

Noise on either the line- or load-side wiring to the filter can radiate

and be picked up by leads on the opposite side. Shielded equipment cab-

inets do not prevent radiation w ithin the enclosure. To prevent this deg-

radation, the line- and load-side wiring must be kept separate. Long

line-side leads inside the equipment can also pick up noise radiated by

power or logic components. The reverse is also possible. This effect is

controlled by minimizing the length of line-side wiring inside the equip-

ment and dressing it away from noisy areas. The optimum solution is to



196


employ a filter with an ac connector and mount it directly on a metal

panel

so

that

the

power line exits

the

equipment directly

at the

filter.

Examples are shown in Fig. 8.12.

8.1.2 Signal Filters

Filters may be very effective in suppressing conducted EMI in signal

circuits. For filters to be effective, the intended signal and the EMI

Metal Panel

Poor

Plastic PC Board

Filter

Filter

Better

Filter with integral

IEC connector

Metal Panel

Plastic PC Board

Metal Panel

Best

Plastic PC Board

Fig ure 8.12 Power-line filter installation.



FILTERS

197

must occupy different portions of the frequency spectrum . Various types

of signal filters are available. These include discrete filter components

that can be incorporated into the signal line at the input or output of

equipments, filter pin connectors that serve as input or output connec-

tors on signal lines and also act as filters, and printed circuit board

mountable EMI filters that may be incorporated into equipments to

provide frequency selective rejection of EMI in signal circuits.

Consideration should be given to the use of signal filters for the fol-

lowing situations. First, low-pass filters should be used at the input to

equipments that operate with low-level, low-frequency analog signals to

avoid EMI problems resulting from radio transm itters that are present

in the environment. Examples of these types of low-level, low-frequency

analog signals would be low-level audio signals at the input to an audio

amplifier or low-level control signals from an analog sensor. In these

instances, failure to properly filter the input signal lines may result in

high-level RF signals saturating the low-level analog amplifiers and

producing EMI as a result of the nonlinear "audio rectification" effect.

Second, high-pass filters should be used at the input to equipments tha t

operate with relatively high-frequency signals in the presence of high-

level, low-frequency EMI. An example would be digital computing

equipments th at may experience EMI problems as a result of power line

EMI coupling into the digital signal lines. Third, band-pass filters

should be used at the input to equipments that operate with low-level

narrowband carrier modulated signals in the presence of EMI. An

example would be a communications receiver that is operating in the

presence of a number of transmitters.

There are several different signal filter configurations that may be

selected, depending on the specific conditions in which the filter is

applied. Figures 8.13 through 8.16 illustrate different filter configura-

tions from which one may be selected to work into or out of either a high

Source Optional Load

F ig u re 8.13 Filter network for low-source and low-load imped ances.



198



F ig u re 8.14 Filter network for high-source and high-load impedan ces.


F ig u re 8.15 Filter network for low-source and high-load imp edances.


F ig u re 8.16 Filter network for high-source and low-load impedan ces.

or low, source or load impedance relative to 50 Q. All filters shown are of

the low-pass type. They use series inductors and shunt capacitors. The

philosophy then is to connect either (1) a filter series inductor to a low-

impedance source or (2) a shunt capacitor to a high-impedance source

such that the impedance source and filter element are about equal at the

desired cutoff frequency. Similarly, a series inductor should face a low-

impedance load, and a shunt capacitor should face a high-impedance



FERRITES 199

load. This ensures optimum use of filter elements and in part compen-

sates for some source and/or load impedances of typical power mains

varying over wide ranges starting about

100

times the power frequencies.

For a high-pass

filter,

nductors and capacitors would be interchanged.

Crosstalk between the filter input and output term inals may be sig-

nificant (60 dB or even lower, i.e., greater coupling) unless an infinite

baffle is used between the two terminal pair. Consequently, when the

manufacturers rate the filter attenuation with frequency, it is under-

stood that the filter is mounted in a suitable bulkhead. For shielded

enclosure filters, where attenuation is rated up to 100 dB or more, the

entire assembly is shielded in a permeable case. Thus, after filtering,

there is reasonable assurance that there will be no cross-coupling of

magnetic, electric, or electromagnetic fields to the filter output leads,

which are located inside of the enclosure.

8.2 Ferrites

Ferrite materials are available in the form of hollow core beads or tor-

oids that can be slipped over a wire and behave as a lossy inductance

with one or a few turns. They are popular among EMC specialists as

quick "last-chance" fixes and provide remarkable EMI reduction if they

are used properly. Basically, ferrites act as low-pass filters and, as such,

they may be used to provide significant attenuation of troublesome EMI

under certain limited conditions. EMI ferrites are made of lossy materi-

als having a good magnetic permeability (preferably with jx

r

being flat

over a wide frequency span and with typical values of 300 to 3,000).

They also display a resistance of ten to a few hundred ohms. Although

ferrite beads are generally thought of as inductors, they are in fact

transformers, where the wire being filtered is the primary (one or a few

turns), and the secondary is a result of the eddy currents in the bead

creating losses by Joule effect.

Due to the generally small size of ferrite beads, they can easily satu-

ra te for the normal current and become inefficient against the EMI cur-

rent. The amount of current a bead can handle without significant

decrease of ju

r

is given by the manufacturer. It is related to:

In(r

2

/r

x

)

where r

2

and ^ are the bead's outside and inside diameters. Therefore,

beads with proportionally small holes will behave better.

The permeability is also affected by frequency. Some beads are opti-

mized to work below 10 MHz, and others are suitable from 10 to 100 or

even 1000 MHz.




To effectively use ferrite beads, it must be understood that they work

by series inse rtio n loss. As a re su lt of th is fact, it is im po rta nt to note

tha t :

• Fe rri te s will be mos t efficient in low-im pedanc e circuits like power

distribution, power supplies, or radio-type circuits where imped-

ances are 75

Q

or less.

• Fe rri te s will not work efficiently in high -im ped anc e circu its.

Although significant progress has been made by ma nufactu rers, the

best ferrites today achieve values of Z

B

in the 300- to 600-Q range,

above 50 MHz.

• If th e wire impe dan ce itself is significant, th e ferrite m ay not

exhibit effective performance.

• Fe rri te s are mo st effective for rejecting E MI in th e H F an d VH F

bands .

If the attenuation with a ferrite bead is not sufficient, this can be

improved in several ways. One method is to make more than one turn

of the wire in the bead hole, using two or three turns. However, this

may rapidly bring the ferr i te into saturation. Also, the turn-to-turn

capaci tance may ru in the inductance improvement . Put t ing several

beads back to back is another method. However, this will create a cas-

cade of parasitic capacitances. If a few beads do not work, it is doubt-

ful th at mu ltiplying the ir num ber to reach se veral tens of bea ds will

ever work.

An extremely useful application of ferrite is in the blockage of com-

mon-mode currents. If the two wires of a signal pair are threaded in the

bead, the ferrite will affect only the undesired EMI currents and will

have no effect on the intentional differential-mode current. The same is

true when a ferrite is slipped over a coaxial cable.

The limitations of ferrites, besides their limited impedance are:

1. W hen be ad len gth app roa che s A/4, th e be ad becomes inefficient.

2. The end-to-en d pa ras iti c cap acit anc e of th e ferrite (typically 1 to

3 pF) may bypas s its resista nce above a certain frequency and

cause its attenuation to collapse.

3. Beyond about 1,500 to 2,000 gauss, saturation occurs and efficiency

decreases.

4. W hen slipped over mu lti-pair cables, the y may increa se inductive

crosstalk between adjacent pairs .

Figure 8.17 shows the shapes and performances of the principal types

of lossy ferrites available.



ISOLATORS

201

F ig u re 8.17 Typical available ferrite bead s.

Insp ired from the previous considerations, the ferrite-loaded wire (also

called

lossy

wire) and tubing is an interesting concept. In these wires, a

conductor is coated with a flexible compound made of ferrite plus a

binder, such that the lumped elements are replaced by a continuously

distributed insertion loss (Figs. 8.18 and 8.19). Because of the flexibility

requirement, the ferrite content of the jacket provides a permeability of a

few tens. In a 50 Q, system, little to no atte nu atio n exists below about

5 MHz (Fig. 8.20). By elim inating the impedance d iscontinu ities th at a

cascade of beads w ould create, lossy wires ha ve less of a tenden cy to rad i-

ate. They share with the beads the enormous advan tage of not depending

on grounding or bonding techniques. Also, their distributed impedance

and lossy nature mean that they can work with extremely mismatched

source and load resistance without exhibiting ringing and other mis-

match problems. Furthermore, the ferrite grains and their binder have

an £j. th at can be ra th er large, such th at a lossy line with a predete r-

mined ch aracteristic impedance can be constructed since:

Zo(lossy line)

~ 37 7 / —

W

fc

r

8.3 Isolators

Isolators are used in systems applications to control conducted EMI in

situations where the mode of the EMI signal and the desired signal are




(Shown for

1

bead or

1

turn)

z

g

+ z

L

=

Curves

10 Q 50 Q 100 Q 300

Q

A

B C D

0.1

0.3 1 3 10 30

Frequency in MHz

100

300 1,000

35

30

25

20

15

10

II-Thick Ferrite Bead

6 m m

1 mm

Ex: "Fair-Rite" Material

43

or 64

Z-

ffifc

ZZL

.._:__.

it::

~i .

\

0.1 0.3 3 10 30 100 300 1,000

Frequency in MHz

Fig ure 8.18 Insertion loss of small ferrite beads: (I) small ferrite and (II)

thick multi-hole ferrite.



ISOLATORS

2 3

100

90

80

70

.S 60

5 50

4

°

30

20

10

0

L

0.1

Rg =

R

L

* =

Tubing 1

Block 2

-

1 Q

F

A

io

a

G

B

so a

H

c

LOOQ 1,

I

D

Fair-Rite Material #4 3 I

RF Suppressant ^

Tubing ^ y ] \

1 meter (S^A *

«<

y

C

- - -

-64n

j

LU

ij

.. ^

: ^

^

m

1 3 m m

-r

300 Q

J

E

-I

1

(F

i /

/\

/ /

' (E )

i

>

(

_J _ i

1

/

fn

I

0.3

3 10 30

Freque ncy in MHz

100

300 1,000

*R

G

=

Source (Gen erator) Impedance in Q

*R

L

= Load Impedance in Q

Figure 8.19 Inse rtion loss of large ferrite suppre ssors.

Measured

p er

MIL-STD-220A

in 50 Q System

20 30 50 70 100 200 300 500 700 1,000

Freque ncy in MHz

Figure 8.20 Atten uation of ferrite-loaded tu bing vs. single bead s. (The bead s

shown were not optimized for high-frequency resistance.)



204


different. Isolators

are

generally used

to

reject common-mode

EMI

without affecting th e differential-mode desired signa ls. T he two types of

isolation devices that are most widely Used are isolation transform ers

(which

are

generally used

in

power

and

audio signal applications)

and

optical isolators (which are generally used in digital signa l applica-

tions). The following sections discuss the application of isolation tra ns-

formers an d optical isolators to solve sy stem s EM I problems.

Isolation transformers should be used at equipment power or audio

signal inputs or outputs where it is suspected that EM I problems may

result from common-mode interference conditions.

To

effectively uti lize

isolation transformers, extreme caution must be observed in installing

these devices. In particular, it is essential tha t th e shields of th e isolation

transformer

be

properly t erm inated with

a

low impedance

to

ground

at

any EM I frequencies of interest. Also, it is impor tant to install th e isola-

tion transformer such that input-to-output coupling around th e isolation

transformer is avoided. This may be accomplished by installing th e isola-

tion transformer directly at the input to a shielded equipment compart-

ment ,

as

discussed

in

th e previous section on filters.

8.3.1 Isolation Transformers

Isolation transformers offer an effective and reliable solution to many

electromagnetic interference problems from the ac mains and audio sig-

nal lines. Their simplicity belies their outstanding performance

in the

elimination of conducted EMI. Conducted EMI is distinguished accord-

ing to its relation with respect to signal or power wiring and a common

reference (ground). Two categories are identified: differential-mode and

common-mode.

8.3.1.1

Differential-Mode Noise

This interference appears differentially between two leads of the mains;

as such, it is transmitted like the normal power voltage or current

(Fig. 8.21). Consequently, devices (such as rectifiers, switching transis-

tors, regulators, etc.) located at the ac input of equipment would be sus-

ceptible to damage from high-energy transients and surges.

Furthermore, circuit malfunction could result from propagation of high-

frequency noise through the power leads.

8.3.1.2

Common-Mode No ise

This interference appears simultaneously from two leads to a common

reference; since it is equally present at both points, there exists no dif-



ISOLATORS

205

Equipment

L O

N O

G O

F ig u re 8.21 Normal-mode noise.

ferential component between the leads (Fig. 8.22). This form of interfer-

ence is the most troublesome. It could couple energy directly into an

electronic circuit through the distributed capacitance that exists

between the circuit and ground. In single-ended circuits, it could gener-

ate noise signals within their common returns. It could also generate

circulating currents within enclosure shield and grounds that, in turn,

could couple noise into the equipment. Any of these occurrences could

result in equipm ent malfunction.

8.3.1 .3 Comm on-Mode No ise A ttenu at io n

To control common-mode interference, a barri er m ust be produced th at

will prevent ingress of noise from the mains into the equipment. The

most effective metho d of realizing su ch a bar rie r is the sh ielded iso lation

transformer. This device eliminates the conductive path through which

noise could be transmitted; only the coupling capacitance between pri-

mary and secondary windings allows transfer of energy in the common-

mode. However, m eans are av ailable to control thi s param eter.

8.3.1.4 Ca pac it ive Co uplin g in Tra nsform ers

In typical transformers, an electric field is generated between the con-

ductors th at com prise th e windings because of the po tentia l differences

L O -

N O -

G O -

Equipment

Circuit

(

C M

J :: ^ :i Stray

F ig u re 8.22 Common-mode noise.




that are present; associated with this field is capacitance between the

individual turns . Although this capacitance is a distributed parameter,

it is modeled as a single element connecting the primary and secondary

windings (Fig. 8.23). There exists a substantial amount of capacitance

due to the large surface area of the coils and their close proximity to

each other and the core. If a change were to occur in the po tentials, dis-

placement currents (which are proportional to the coupling capaci-

tance) would flow between the turns. These currents could couple

significant common-mode energy through the transformer into the

equipment.

8.3.1.5 Isolation Transformers

An isolation transformer eliminates the shortcomings of an ordinary dou-

ble-wound power transformer through shielding that electrostatically

isolates the primary windings from the secondary. The shield, interposed

between the windings and suitably terminated, divides the capacitance

into two components: one from both the primary and secondary to the

shield. Any displacement currents that occur

flow

nto the shield and not

between windings. The shield is constructed of nonferrous materials and

does not hinder the magnetic coupling of

the

windings.

The shielding could be introduced in various configurations depen-

dent on the cost/performance trade-off desired. The possibilities range

from a simple, single shield between the windings to elaborate, multi-

ple box shields as used in ultra-isolation transform ers.

Single Shield

A single shield (called a Faraday shield) consists of a layer of conduc-

tive material (copper or aluminum foil) that is wound as one layer

between the primary and secondary; this layer is insulated to prevent

formation of a "shorted turn." When the shield is grounded, any com-

mon-mode interference that occurs between the primary and that

ground would return to its source through the shield and not couple

across the windings (Fig. 8.24). However, limitations exist when multi-

c

Coupling

(Core is Not Shown)

F ig u re 8.23 Capacitive coupling between windings.



ISOLATORS 207

pie grounds occur within the equipment and the ac mains. Interference

could be transferred from the shield through either winding capaci-

tance since both are common to it.

Dual Shield

The difficulty of properly grounding a single shield is eliminated with a

dual shield—one for the primary and another for the secondary. The

winding capacitances are terminated on individual, isolated shields.

With the shields connected to ground references associated with their

windings, interference would flow only in separate loops limited to

those windings.

Triple

Shield

A

further improvement is possible by adding a third shield located

between the primary and secondary shield. This allows a more favor-

able grounding arrangement with the third shield connected to the

equipment enclosure ground. It also compensates for practical deficien-

cies produced by parasitic inductance of the winding shields and their

connections. The impedance of the inductance allows circulating cur-

rents to develop voltage drops that subsequently could couple currents

into the secondary. The third shield intercepts these currents, prevent-

ing their transference.

Ultra-Isolation Transformer

The ultimate extension of

the

shielding process is found in the ultra-iso-

lation transformer. A triple shield is used with the primary and second-

ary shields totally enclosing their windings in a configuration called box

shielding

to reduce electrostatic coupling to a minimum (Fig. 8.25).

Also, the windings are physically separated by orienting them side by

side on the core instead of concentrically; a reduction in coupling is

realized through the decreased surface area and increased distance

between the coils. The third shield is located to provide total separation

of the other

two.

The resu ltan t coupling capacitance is extremely low—

less than 0.0005 pF—and the common-mode noise attenuation could

exceed 140 dB at 10 MHz.

i o L

Noise

~ Source

Fig ure 8.24 Transformer with shield.



208


Mains

Ground ~

/77

Equipment

Ground

Circuit

Common

Fig ure 8.25 Ultra-isolation transformer.

Neutral

Grounding

Apart for directly attenuating interference, an isolation transformer

could reduce equipment susceptibility by allowing a more favorable con-

nection to the ac mains. In applications requiring a grounded neutral,

the secondary could be grounded in the vicinity of the equipment. This

proves beneficial because the remotely grounded neutral is a prominent

source of common-mode interference. The neutral wire, together with

the safety ground wire and the stray capacitance within the equipment

between circuits and the enclosure, form a loop through which noise

and damaging transients could circulate (as during faults or lightning

strikes) (Fig. 8.26a). The shield or the isolation transformer would

divert those currents back to their source and prevent entry into the

equipm ent (Fig. 8.26b).

Furthermore, with a short connection between ground and neutral,

the impedance would be reduced; any interference currents would gen-

erate lower voltage drops, reducing the equipment susceptibility.

Differential-Mode

Attenuation

The shielding that is responsible for the exceptional common-mode

attenuation is ineffective for differential-mode interference. In the dif-

ferential mode, interference appearing across the primary would pro-

duce current flow through it and, by magnetic induction, would be

transferred to the secondary. However, differential-mode attenuation

could be improved through several measures.

In general, attenuation is enhanced by increasing the series imped-

ance and decreasing the shunt impedance presented to the interfer-

ence.

A

divider action is evident between the series impedance and the

shunt (essentially, a low-pass filter is produced) with the interference

being dropped across the series element.



ISOLATORS

209

Equipment

— Noise Source —

Circuit

1

c

II II Stray

(a) Remote Neutral Grounding

C

Coupling Equipment

L

— Noise Source —

(b ) Local Neutral Grounding

Figure 8.26 Neutral grounding.

Leakage Inductance

The series impedance is dependent on the leakage inductance of the

transformer (determined by the degree of magnetic coupling between

coils) (Fig. 8.27). This could be controlled through the physical location

of the coils and their geometry. The greater the separation or the taller

or narrower the coils, the greater the inductance. However, increased

leakage inductance could impact regulation at the power frequencies;

optimization is required to attain performance that is satisfactory for

both considerations.

Shunt Capacitance

The addition of a shunt capacitance across the secondary of the isola-

tion transformer markedly increases the filtering afforded by the leak-

Fig ure 8.27 Transformer equivalent circuit.



210 FILTERS, FERRITES, ISOLATORS,

AND TRANSIENT SUPPRESSORS

age

inductance.

A second-order LC filter is produced with a first-order

LR (considering

a

resistive

load). The

a t t enua t ion

of a

second-order

fil-

ter after i ts cutoff frequency increases a t 40/decade compared to 20 dB/

decade

for th e

first-order.

Also,

th e LC

filter allows

for

less loss

at th e ac

mains frequency while stil l providing good attenuation—and therefore

would improve th e regulat ion.

EMU RFI Filters

When augmented with a multistage EMI/RFI filter, an isolation trans-

former's a tte nu ati on could be extended in both common and differential

modes. The two devices would work in tand em , each complem enting the

qualities of th e other.

In the common mode, the performance of EMI/RFI filters is compro-

mised below 100 kHz by restrictions imposed by safety agencies on the

maximum leakage current flowing in the ground lead (0.5 to 5 mA,

depending on the product). This effectively limits the value of common-

mode filter capacitors and, as a result, attenuation. However, an isola-

tion transformer provides exceptional loss at lower frequencies and

would compensate to maintain an overall high attenuation across a

wide frequency range.

In the differential mode, althou gh no safety agency limitatio ns exist,

economic and size considerations would constrain the maximum value

of filter components. The leakage inductance of the transformer would

supplement the series impedance of the filter and, again, extend the

frequency rang e of useful a tten ua tio n.

Transformer Response Limitations

There also exists an inherent limitation in the ability of a transformer

to transfer high frequencies through magnetic induction. As the fre-

quency increases, the permeability of the iron core decreases until the

core no longer aids magnetic induction and the coils are coupled only

through the air. Correspondingly, the leakage inductance and the effec-

tive series impedance become very high. Furth erm ore, th e transformer

has distributed, parallel capacitance across the windings that shunts

high frequencies, reducing coupling to the secondary. Both characteris-

tics inhib it response to high-frequency interference.

Common-Mode Conversion

When one side of the secondary is grounded (to produce a neutral), the

common-mode interference transferred from primary to secondary

would appear as differential-mode interference because of the unbal-

anced impedances from the two secondary leads to ground (Fig. 8.28).

Attenuation would be enhanced because greater capacitance exists at

the output of the transformer in the differential-mode, consisting of

transform er-distribute d capacitance, cable capacitance, EMI/RFI capaci-



ISOLATORS

211

^Coupling

7

CM

~zz~

Noise

~z r

Source

F ig u re 8.28 Common-mode conversion.

tors within equipment, etc. In comparison, the coupling capacitance is

extremely small (the attendant voltage divider action would be

improved).

Common-mode conversion could also occur at the primary if an

unbalance exists in the distributed capacitance between the winding

and the shield (caused by a physical dissymmetry between them). Cur-

rent flow due to interference would not be evenly distributed along the

winding; a component of current would circulate through it (differen-

tial-mode) and, through magnetic induction, would produce a differen-

tial-mode secondary current.

Conclusions

Isolation transformers provide protection from EMI in both the com-

mon mode and differential mode. In the common mode, they have an

impressive attenuation capability resulting from their shielded con-

struction. In the differential mode, good performance is possible

through the filtering action of their inherent series impedance in con-

junction with circuit capacitance. Overall, they provide unique perfor-

mance characteristics that are essential for the protection of sensitive

equipment. Figure 8.29 illustrates the typical isolation transformer and

differential-mode rejection as a function of frequency.

8.3.2 Op tical Isola tors

Optical isolators (also called optocouplers) play a major role as isolation

elements in digital data equipment, control systems, and telephone

communications. The optical isolator consists of a photon-emitting

device and a photosensitive detector. In the optical isolator, or photon-

coupled pair, the coupling is achieved by light being generated on one

side of a transparent insulating gap and being detected on the other



212


2

8

6

4

2

60

40

20

C M R e j ec t io n = 2 0 L o g - ^ 2

V

CM

DM

Rejection

= 20 Log -

V

D

1

A = CM Rejection

B = DM Rejection

C = CM Rejection of an ordinary

unshielded tran sforme r (P = 300VA,

Prim-Sec,

capacitance

= 1-

3nF)

0

lOOHz

300 1kHz 3

10kHz

30

100kHz300

1MHz 3

10MHz

30

100MHz

Frequency

F ig u re 8.29 Typical me asured values of CM and DM rejection for Faraday-

shielded isolation transformers.

side of the gap without an electrical connection between the two sides

(except for a coupling capacitance of approximately 1 pF). In a typical

optical isolator, the light is gen erate d by an infrared light-em itting

diode (LED), and the photo-detector is a silicon diode, transistor, SCR,

or Darlington devices, as shown in Fig. 8.30.

Optical isolators have a host of applications where one or several of

the following objectives are important:

• Isola te different voltage levels betw een circu its.

• Prev ent interference between control and power circuits usin g the

unidirectional feature.

• Ins ula te people or low-voltage circuits from t he ha za rd s of high-

voltage shock.

• Elim inate de ground loops.

• Reduce common-mode EMI effects in signa l line s.

• Amplify or at te nu at e signals.

• Perform on/off sw itch ing .



ISOLATORS

213

-o o

-o a

( a )

( b )

(c )

( d )

Figure 8.30

Basic types of optical isolators: (a) LED-photodiode, (b) pho-

totransis tor with or without base terminal, (c) LED-photo-SCR, and (d) LED-

photo-Darlington.

Figu re 8.31 shows thes e common ap plications for optical isolators. As

illustrated in Fig. 8.31a, the triggering of triacs with optoisolators is

accomplished relatively easily. The diode bridge converts the ac to the

de required by the SCR. Figure 8.31b depicts how excellent isolation

between a patient and monitoring equipment can be achieved by

optoisolators when used in medical electronics. Figure 8.31c shows a

digital line receiver application.

The dynamic and EMI characteristics of optical isolators are as fol-

lows:

rise and fall time delimits the maximum useful bandwidth of the

optical isolator. The simplest and cheapest optocouplers with only a

diode/phototransistor pair typically have transition times in the 10 to

100 JIS range for a digital-type signal. By adding Schmitt triggers and

positive feedback amplifiers, the switching speed can be improved

greatly. As early as 1986, modern optocouplers with a 30 ns transition

time corresponding to a 10 MHz bandwidth were available.

Input/output de isolation is defined by at least two parameters: the

isolation resistance Ri

so

and the maximum withstanding de voltage

Vi

so

.

The former is sometimes defined by the maximum input-output

leaka ge c urr en t u nd er a given voltage like 1 or 3 dV. Typical valu es are:

R

iso

= i o

9

to 10

1 1

Q

V

is o

= 500 to 5000 V

A level of 15 kV isolation is obtainab le. This vo ltage should norm ally

be guaranteed between input and output pins or between any pin and

the device's can, whichever is less.



214

FILTERS,

FERRITES, ISOLATORS, A ND TRANSIENT SUPPRESSORS

Optoisolator

Single Con tact Circuit

(a) Triggering of Solid-State Relay

Patient

- -4

Measurement

Circuit

10 kV

Isolation

(b) Medical Electronic Sensor

O+ 5 V

3

V

minimum

0.6

y,s

< t

o n

< 1.3 ps > 1.2 kft,

0.6 s < t

o

ff < 1.3 ^s > 10%

(c) Ground-Loop Isolation of a Digital Line Receiver

OGND

Fig ure 8.31 Examples of optoisolator applications.

Input/output capacitance (Fig.

8.32)

consists

of

two capac itances:

the

internal LED-to-phototransistor

(or

othe r detector) capacitance, which

typically ranges from 0.1

to 2

pF,

and the

input-pin-to-output-pin capac-

itance, which depends greatly

on the

package style

and

ran ge s from

0.3

to

3

pF.

The

combination

of

these two capacitances dictates

the ac

isola-

tion, since they byp ass

the

Ri

so

above

a

certain frequency. Each one

par-



ISOLATORS

215

Figure 8.32

Para sitic capacitances in optoisolators. Inpu t/outpu t parasitic

capacitance consists of intrinsic LED-to-photodetector capacitance C

2

plus

package an d leads capacitance C^. It can be reduced by interna l Fa rada y thin

mesh. It can be aggravated by careless I/O wiring isolation.

ticipates in its own way to the common-mode rejection of the optical

isolator, as explained in the next paragraph.

Common-mode rejection (CMR) or transient immunity is important

to describe the device's imm un ity to EM I, bu t it is th e one criterion th a t

is usually the most poorly documented (if documented at all) by manu-

facturers' technical data sheets. Some of them label it as

maximum

input transient voltage without specifying the rise time. Some of them

define a slew rate in

V/JUS.

Some of them do not specify it at all. In the

few well-documented da ta s heets available, the CMR is described as th e

maximum slew rate in V/jus of CM voltage that can be sustained with

th e ou tpu t voltage sta yin g in eithe r a "high" (>2 V) or "low" (<0.8 V) sta-

tus .

Figure 8.33 shows that the transient immunity is indeed a two-

mechanism phenomenon.

1.

The input-lead -to-outpu t-lead capacitance (C

p l

) is a capacitance

that bypasses the Rj

so

of th e device; i.e., th e device becomes a leaky

barrier, and some percentage of the EMI voltage appears across the

load ZL without the optical isolator playing any active role in this

transfer. For instan ce, if we tak e Z L as 1 kQ (a typical dynam ic

resistance of a TTL gate input in the low-to-high transition region),

this value is shunted by the optical isolator output resistance (typi-

cally 50 to 150 Q), so the optical isolator output virtually looks like



216


_ , , _ —

(a)

1,500 V/^s

0.8

V

10

ns

( b )

Output Response:

(c )

140

dB

30

dB

:201og(F

CM

/V

L

)

•

Low Freq. Bound

2

x

Rj

so

x Cp |

( lOOHzi ins t )

F

2

= Bandwidth of

driven load

Z^

(30MHzf. inst)

Fig ure 8.33 Optical coupler CM transient response.



ISOLATORS 217

a 50 to 150

Q

load. The low-frequency bound of CM rejection is

given by:

R

max = 7^>

9

Assuming that Ri

so

= 10 and ZL = 100, this gives:

CM R

m ax

=

1 Q 0

= 10"

7

or 140 dB

100+10

This CMR starts to deteriorate as soon as the reactance of C

p l

bypasses the 10

9

Q. Assum ing C

p l

= 1.5 pF (another typical value

for optoisolators), this will occur for:

Beyond this frequency, the CMR will degrade at 20 dB/dec. There-

fore, it would theoretically reach 0 dB for:

4

F

2

= 100 Hz x 10 =

1

GHz

However, when the frequency reaches the cutoff frequency of the

load (typically the input of a digital gate, or a line receiver, compar-

ator, etc.), th e load itself, by its inpu t capacitance, star ts to have an

ac noise rejection that improves at the same rate that the CMR

degrades. For instance, with a TTL-type load beyond 30 MHz, the

CMR will stay flat to a value computed by:

CMR

(30MHz) = 140dB-201og

3 0

^

1

0

0

H z

H 2

= 30dB

2. The int ern al LED -to-detector capacitance (C

p2

) exists because of

the physical proximity of the LED and photodetector (on the order

of one to few millimeters) and is aggravated by a resin lens used to

channel the light and improve the overall efficiency. This resin has

an e

r

> 1, which ag grav ates th e capacitance C

p 2

'




Thi s capac i t ance can m a k e the opto i so la tor e l ec t r i ca l ly t r igge red if a

high enough dV/dt ex i s t s across the opt i ca l bar r i e r , l i ke a t r a n s i e n t

b e t w e e n an y or b o t h i n p u t l e a d s and the loca l grou nd. Th i s noi se cur-

r e n t ,

Ip = C

p 2

dV/dt , becom es

a

b a s e c u r r e n t i n t o

the

p h o t o t r a n s i s to r .

A s s u m i n g t h a t t h i s

one ha s a

ga i n

of 100 and a

co l l ec tor cur rent

of

1 mA w he n conduc t i ng , a b a s e c u r r e n t of 0 .01 m A wi ll tu rn on th e t r a n -

sis tor .

F r o m the p r ev i ous equa t i on , t he vo l t age s tep t h a t ca n caus e t h i s b as e

c u r r e n t is:

dV/dt = 0 .01 mA / C

p 2

I fC

p 2

= l p F ,

dV/dt = 10-

5

/10

12

= 10

7

V/s or 10 V/jis

A simple electromechanical switch can cause spikes faster than this.

Thyristors and other semiconductor switches cause transients in

excess of 100 V/ns. Static discharges induce transients in the range of

100

V/JLIS;

therefore, many real-life transients can upset an optoisolator

even though its immunity based on de data might seem impressive.

Quality optoisolators are characterized against this parasitic turn-on,

where the isolator is becoming an active device in the transmission of

EMI. A good brand of modern isolator can resist up to 500 V/|is or even

3 kV/j^s. However, these values are generally given for 25°C and

degrade rapidly with an increase in temperature.

For instance, assume there is an optoisolator specified for 1,000 V/jis

of CM transient immunity and a TTL-type output. This means that when

the output is at a low status, it takes at least a 1,000

V/JIS

spike to cause

the output to exceed 0.8 V for more than 10 ns (the typical TTL minimum

transition time). Therefore, the shortest pulses to cause an undesired

response are a 10 V pulse with 10 ns transition time or a 100 V pulse

with 100 ns transition time. For pulses having less than 10 ns transition

(i.e., a bandwidth exceeding 30 MHz), the ac noise rejection of TTL as

well as the time constant of the phototransistor will naturally improve

the rejection by the same rate (20 dB/dec) as the capacitively injected

base current increases, giving an overall flat CM rejection.

If we calculate the rejection of the above example for the worst-case

pulse of 10 V/10 ns and consider that of the 0.8 V output, half (0.4 V) is

due to the VCE saturation of the output transistor, we come up with:

10

V

Rejection = 201og * = 28 dB (for TTL)

U.O J.T

1



ISOLATORS

219

This is broadly in the same range as the CMR due to C

p l

, although one

should compute each of them separately and reta in the lower figure .

Transient immunity can be improved in several ways:

1. Decoupling of the phototransistor base. This is efficient but has a

corresponding adverse effect on bandwidth.

2. Increasing the separation distance between the LED and the opti-

cal detector. This can go as far as installing a short piece of fiber

optic to act as a light guide. This is an effective solution, but it

increases the size of the device.

3.

Inserting a thin metal mesh (optically transparent) between the

LED and phototransistor. The principle is the same as Faraday-

shielded transformers. The shield must be tied with a low-imped-

ance conductor to the common ground on the detector side (see

Fig. 8.34).

Input ac and de impedances should be considered because, being

nonlinear, the LED input cannot be assimilated to a simple RC net-

work. However, to avoid the LED overrun if the input signal exceeds

the forward or reverse break voltage (Fig. 8.35), the input of an optoiso-

lator is always driven in current mode, i.e., an input resistance is used.

The useful range of Ip current is between I mA and 100 mA, corre-

sponding to a Vp of 1.2 to 1.3 V. Thus for a given range of input signal,

the designer selects a series resistance RS such as:

1. For the minimum input signal amplitude, V

m

i

n

, the upper limit is:

T

being the minimum current to drive the optoisolator with the

desired output response, i.e., amplitude and transition time

/////////A

Fig ure 8.34 Equivalent circuit of optical coupler

CM

transient response.



22


Fig ure 8.35 Optoisolator with Faraday shield.

(response time degrades when I

F

decreases), considering also the

worst-case temperature drift

for

Vp.

2. For the

maximum input signal amplitude V

m a x

,

the

lower lim it

is:

R

V

m a x -

V

F

Smax

I

m a x

is the

curre nt compatible with th e diode safe o perating a rea

(or

maximum power dissipation Vy x I

F

) and the diode aging (for

instance, 10 percent brightn ess for I

F

= 60 mA after 10,000 hr).

Figure 8.35 also shows the LED being shunted by its own junction

capacitance , typically in the 30 to 100 pF range.

If, for

instance,

a 220 Q

series resistan ce

is

used

to

provide

a 10 mA

forward current for a 3.5 V inp ut signal and a V

F

of 1.3 V, a 50 pF para-

sitic capacitance will start

to

shunt

the

diode

VF at

around

15 MHz,

which gives

the

practical bandw idth

of

this LED input

for a

differential

signal. As seen

in

Fig. 8.35,

a

reverse voltage protection can be provided

by

a

silicon diode across

the

LED.

In some cases,

it is

desirable

to set a

definite threshold

for the LED

voltage. This

is

done

by

shunt ing

the LED by a

resistor,

the

value

of

which

is

determined

by the

applied voltage,

the

series resistance ,

and

the desired VF.

All these considerations

are

necessary

to

predict

the

behavior

of the

LED input

in the

presence

of

differential EM I. Altho ugh groun d loop

interference generally prevails in the EMI problems dealt with by

optoisolators, EMI coupled differen tially into the two wires of th e cable

pairs should not be overlooked. For instance, if comp uter cables are run-

ning

in the

same conduit with power cables over

few

m eters,

a

1 kV/jas

transient

on

these cables

can

induce several volts

per

m ete r differen-



TRANSIENT SUPPRESSORS

221

tially in the signal pairs, which is enough to drive the optoisolator into

an erroneous trigg ering. In th is case, the ground-loop isolation provided

by the optical barrier is without effect on the interference, and the solu-

tion resides in more conventional shielding and sep aration of th e cables.

Finally, another EMC aspect of optical isolators is the way they are

mounted. A few picofarads of input-output coupling are very easy to

aggravate by careless wiring practices. Two signal wire pairs spaced by

3 mm in the same cable way already represent 5 pF/m of coupling

capacitance. Inp ut w ires or traces m ust be kept away from th eir outp ut

counterparts. Figure 8.36 provides a comparison of pin-to-pin isolation

of DIP optoisolators to that of a standard logic gate. In both cases, the

devices were shut off, so the coupling is mainly due to the lead arrange-

ment .

An optoisolator must be mounted as close as possible to the I/O con-

nector. Return conductors (even if called "ground") for the input signal

should be floated and distinct from th e ground conductor of the detector

side.

An optoisolator with an external base connection for the phototrans-

istor (or SCR) should be treated carefully. Since this base lead can be

very susceptible, it should be filtered and kept away from possible noise

paths .

8.4 Transient Suppressors

Electronic systems are often subjected to high-level transient voltage

and/or current surges. Because of the increased use of electronic equip-

ment containing integrated circuits and microprocessors and the severe

negative consequences of equipm ent down time, the th re at of tran sie nts

to equipment is a major system problem. The sources of these tran-

sients can in general be classified as inductive switching, electrostatic

discharge, nuclear EMP, and lightning. The requirement for providing

n

Fig ure 8.36 Characteristic of a typical LED.



222


transient protection is a major consideration in system design and inte-

gration.

Transients may be coupled into equipments through power lines,

analog or digital lines, or ground paths. To avoid problems resulting

from transient surges, consideration must be given to each of these

paths of entry, and surge suppressors should be applied to paths that

present potential problems.

Various devices have been developed for the protection of electrical

and electronic equipment against transient overvoltages. They are

often called "transient suppressors" although, for accuracy, they should

be called "transient limiters," "clamps," or "diverters," because they

cannot really suppress the transients; rather, they limit the transients

to acceptable levels or make them harmless by diverting them to

ground. The IEEE dictionary has selected the more generic but lengthy

term of "surge protective device."

There are two categories of transient suppressors: those that block

transients, preventing their propagation toward sensitive circuits, and

those that divert transients, limiting residual voltages. Since many of

these transients originate from a current source, blocking them may

not always be possible, because the current forced into the high-imped-

ance blocking path would only result in higher voltages and breakdown.

Therefore, diverting of the t ransient is more likely to find general appli-

cation. A combination of diverting and blocking can be a very effective

approach. This approach generally takes the form of a multistage cir-

cuit, where a first device diverts the transient current to ground, a sec-

ond device offers a restricted path for transient propagation but an

acceptable path for the signal or power, and a third device clamps the

residual transient (Fig. 8.37). Thus, we are primarily interested in

diverting devices.

The diverting device can be one of two kinds: voltage-clamping or

short-circuiting devices (the latter called "crowbars"). Both of them

involve some nonlinearity, either frequency nonlinearity (as in filters)

Restrict

Divert

Clamp

Protected

Circuit

Fig ure 8.37 Multistage protection.



TRANSIENT SUPPRESSORS 223

or, more usually, voltage nonlinearity. This voltage nonlinearity is the

result of two different mechanisms—a continuous change in the device

conductivity as the current increases or an abrupt switching action as

the voltage increases.

Table 8.2 summarizes some of the major characteristics of constant

voltage and crowbar surge suppressors and presents examples of

each. Table 8.3 summarizes some of the major advantages and disad-

van tages of the basic types of trans ient protection devices. The follow-

ing sections discuss the basic types of transient protection devices in

more detail.

Table 8.2

Two Basic Tran sient Protection Devices

Constant Voltage or Solid State

Characteristics:

Little Power at Steady State

Conducts Heavily Above Clamp Voltage

Usually Reversible

Non-Destructive to C omponent under

Typical Conditions

Extremely Fast C lamp

Examples:

Zeners

Avalanche Diodes

Some Varistors

Transzorb

Crowbar or O ver Voltage

Characteristics:

Shorts Input Power for Duration of Transients

Automatic Recovery to Operating Conditions

Non-Destructive to Component und er

Typical Conditions

Gas Breakdown

Devices

Spark Gaps

Table 8.3 Lim itations of Tran sient Protection Devices

Type Advantages Disadvantages

Gas Breakdown

Devices

Solid S tate

Devices

Hybrids —

Combination of

Gas

Solid State

• Handle Large

Currents

• Fast Response

•B es t of Both Worlds

• P uts Short on

Power Line

When Fires

• Slow to Respond

• Cannot Handle

Large Currents

• Avoids Weakness

of Both



224


8.4.1 Crowbar Devices

The principle of crowbar devices is quite simple: upon occurrence of an

overvoltage, the device changes from a high-impedance state to a low-

impedance state, offering a low-impedance path to divert the surge to

ground. For example, in the case of spark gaps, the breakdown of a gas

causes the device impedance to change from

a

high sta te to

a

low state.

The major advantage of the crowbar device is that its low impedance

allows the flow of substantial surge currents without the development

of high energies within the device itself; the energy has to be spent else-

where in the circuit. This "reflection" of the impinging surge can also be

a disadvantage

in

some circuits

if

the transient disturbance associated

with

the

gap firing

is

being considered. Where there

is

no power-follow

(discussed below), the spark gap has the advantage of very simple con-

struction with potentially low cost.

The crowbar device has three limitations. One is the volt-time sensi-

tivity of the breakdown process

in air

gaps

or

gas tubes. As

the

voltage

increases across gap, significant conduction of current—and therefore

voltage limitation for the surge—cannot occur until the transition to

the arc mode of

conduction,

by avalanche breakdown of

the

gas between

the

two

electrodes.

The

load

is

unprotected during

the

initial rise

because of this delay time (typically

in

microseconds). Large variations

exist in sparkover voltage attained in successive operations, since the

process is statistical in nature.

This sparkover voltage can also be substantially higher after a long

period of res t than after successive discharges. Because of

the

physics of

the process, it is difficult to produce consistent sparkover voltage for

low voltage ratings. This difficulty is increased by the effect of manufac-

turing tolerances on very small gap distances, but it can be alleviated

by

filling he

tube with

a

gas having lower breakdown voltage than

air.

The technology developed by manufacturers of gas tube has minimized

these effects.

The second limitation is associated with the speed of the sparkover,

which produces fast current rise in the circuits. The gap does a very

nice job of diverting impinging high-energy surges, but the magnetic

field associated with the high di/dt induces a voltage in the loop adja-

cent to the surge suppressor, adding a substantial spike to what was

expected to be

a

low clamping voltage.

A third limitation occurs

if

power current from

the

steady-state

voltage source can follow the surge discharge (hence the term "power-

follow"). In ac circuits, this power follow current may or may not be

cleared at a natural current

zero.

Additional means, therefore, must be

provided to open the power circuit if the crowbar device is not designed



TRANSIENT SUPPRESSORS

225

to provide self-clearing action within specified limits of surge energy,

system voltage, and power-follow current. This combination of a gap

with

a

current-limiting, nonlinear varistor

has

been very successful

in

the utility industry

as

high-voltage surge arresters.

8.4.2 Voltage Clamping Devices

Voltage clamping occurs

on the

current flowing through

the

device

or

the voltage across its terminal. Impedance variation is monotonic and

does not contain discontinuities, in contrast to crowbar devices, which

show a turn-on action. As far as their volt-ampere characteristics are

concerned, these devices

are

time dependent

to a

certain degree. How-

ever, unlike the sparkover of a gap, time delay is not involved.

When a voltage-clamping device is applied, the circuit remains

essentially unaffected by the device before and after the transient for

any steady-state voltage below clamping level. Increased current drawn

through the device as the surge voltage attempts to increase results in

voltage-clamping action. Nonlinear impedance is the result if this cur-

rent rise is greater than the voltage increase. The increased voltage

drop

in the

source impedance

due to

higher current results

in the

apparent clamping of the voltage. It must be emphasized that the

device depends on the source impedance to produce this clamping. The

circuit behaves as a voltage divider where the source impedance (high

side of the divider) is constant, but the clamping device impedance (low

side of the divider) is changing. If the impedance of the source is very

low, the ratio is low, and eventually the suppressor could not work at all

with a zero source impedance. In contrast, a crowbar-type device effec-

tively short-circuits

the

transient toward ground but, once established,

this short circuit will remain until

the

current

(the

surge current

as

well as any power-follow current supplied by the power system) is

brought to a low level.

The action of voltage clamping can be performed by any device exhib-

iting a nonlinear impedance. Two categories of such devices, having the

same effect but operating quite different physical processes, have found

an acceptance in the industry: polycrystalline varistors and single-junc-

tion avalanche diodes. Another technology, using selenium rectifiers,

has been practically eliminated because of the improved characteristics

of modern varis tors.

8.4.2.1

Avalanche Diodes

Avalanche diodes were initially applied as voltage clamps in the form

of zener diodes,

a

natural outgrowth

of

their application

as

voltage




regulators. Improved construction, specif ically aimed at surge absorp-

tion, has made them very effective suppressors. Large diameter junc-

tion and low thermal impedance connections are used to deal with the

problem of dissipating the heat of the surge, inherent in a thin single-

layer junction.

The advantage of the avalanche diode is the possibility of obtaining

quite low clamping voltages and nearly flat volt-ampere characteristics

over its useful power range. Therefore, these diodes are widely used in

low-voltage electronic circuits to protect 5 or 15 V logic circuits, as an

example. At higher voltages, the problem of the heat generation associ-

ated with single junctions can be overcome by stacking several lower-

voltage junctions.

Silicon avalanche diodes are available with characteristics especially

tailored to providing transient suppression. These special diodes must

not be confused with regulator-type zener diodes, although many engi-

neers tend to use the generic term "zener diode."

Since the junction is very thin , the cap acitance of an a vala nch e diode

is appreciable. This capacitance can be a concern in high-frequency cir-

cuits whe re it would produce an un des irab le ins ertio n loss. It is possible

to minimize this effect by using a combination with a low-capacitance

ordinary diode in series with the avalanche diode.

Properly packages and installed avalanche diodes exhibit a quick

response to steep-front pulses and have been used for NEMP protec-

tion. However, this quick response can be completely obliterated by

improper wiring. The effect of lead length is applicable to any transient

suppressor.

8.4.2.2 Varistors

The term varistor is derived from its function as a variable resistor. It is

also called a

voltage-dependent resistor,

but that description implies

that the voltage is the independent parameter in surge protection, an

incorre ct perce ption. Two very different devices ha ve been successfully

developed as varistors: (1) silicon carbide discs have been used for years

in the surge arrester industry, and (2) metal oxide varistors are now

widely used.

8.4 .2 .3 Av alan che Di od e vs . Varistor

The

basic performance characteristics

of

these

two

devices

are simi-

lar, and

therefore

the

choice

may be

dictated

by

clamping voltage

requirements (avalanche diodes

are

available

at

lower clamping

voltages), by

energy-handling capabilities (avalanche diodes

are gen-

erally lower

in

capability

per

unit

of

cost),

or by

packaging

require-



TRANSIENT SUPPRESSORS 227

m en t s ( va r i s t o r ma t e r i a l i s m or e flexib le a nd does no t r eq u i r e

h e r m e t i c p a c k a g i n g ) .

8.4.3 Hybrid Transient Supp ressor s

From the previous discussion, it is obvious that crowbar and constant

voltage each have certain advantages and disadvantages that must be

considered in the selection of a device for a particular application. Thus,

if power follow-on is a problem, a simple spark gap device may not suf-

fice.

Also,

when very steep-front transients occur, the gap alone may let

an excessive voltage go by the "protected" circuit until the voltage is

limited by sparkover. Where the capacitance of a varistor is objection-

able, the low inherent capacitance of a gap will seem attractive. If very

high transient levels are encountered, a spark gap device has advan-

tages over a constant-voltage device. In applications such as these,

where a single device is not adequate, hybrid combinations of crowbar

and con stant v oltage devices are often utilized

Sug gested Readings: Filters, Ferrites, Isolators, and

Transient Suppressors

[1] Burket, Chris, "All Ferrite Beads Are Not Created Equal,"

Compli-

ance Magazine,

August 2010, p. 18.

[2] Muccioli, James P., and Dale Sanders, "Test Methodology for Dual-

Line EMI Filter Evaluations,"

ITEM interference technology, 2009

EMC Directory and Design Guide, p. 11.

[3] Venugopal, N aren da r "Buddy", "More Effective EM I R eduction

Techniques for High Demand Consumer Applications," ITEM inter-

ference technology, 2009 EMC Directory and Design Guide, p. 96.

[4] Keebler, Philip E , an d K ermit O. Phip ps, "Case Stud ies of EMI

Elimination and Ground Noise Reduction Using Ground Noise Fil-

ters," ITEM interference technology, 2009 EMC D irectory and

Design Guide,

p . 102.

Web Ad dresses for Com panies that Provide EMI

Mitigation Devices

Fair-Rite Prod ucts www.fair-rite.com Fe rrite s

Rad ius Power www .radiuspower.com Filters and mag netic

products

Cap tor Corp. www .captorcorp.com Fil ters




EM I Filt er Co. www .emifiltercompany.com Fi lter s

RF I Corp. www.rficorp.com Fi lte rs

Schaffner Group www.schaffner.com Fil ters , ferrites

M ur ata Mfg. Co. ww w.murata.com Filters , ferrites

MAJR Prod ucts www.majr.com Fe rrite s

Chom erics www.chomerics.com Fe rrite s

Em erson & Cum ing www.eccosorb.com Fe rrite s



Chapter 9

Cables and Connectors

Several basic electromagnetic interference (EMI) principles determine

whether an equipment will experience EMI or electromagnetic compat-

ibility (EMC) as

a

result of exposure to the electromagnetic (EM) fields

that are present in the equipment environment. These basic EMI prin-

ciples are influenced by the way that the system is configured, includ-

ing the cables. To achieve EMC, it is necessary to be careful to configure

the cables

in a

way tha t does not create EMI problems. The basic EMI

considerations and the resulting impact of the configuration of the

cables are discussed in this chapter.

It

is

important

to

realize that the cable configuration

is a

systems

problem. In order to determine the optimum configuration for any

given situation,

it is

necessary to consider the total system and iden-

tify

the

trade-offs th at result from

the

configuration used

for the

cables. One important cable decision that must be made during sys-

tem design

is

whether

a

shielded cable

is

necessary. Another impor-

tant decision is whether twisted pairs or shielded twisted pairs should

be used.

If a

shielded cable

is

used,

it is

necessary

to

define how

the

shield should be term inated . For example, there are usually trade-offs

between differential- and common-mode EMI tha t depend on the

shield terminations. A discussion of shielding of electromagnetic fields

is presented in Chapter 6.

In order

to

determine

the

optimum method

for

interconnecting

equipments,

it is

first necessary

to

identify the potential EMI source

and the victim

to

be protected. Next,

it is

necessary

to

determine

the

purpose of

the

cable shield (i.e., control radiated emissions, control radi-

ated susceptibility, or prevent crosstalk). Basically, this requires defin-

ing the susceptibility

of

the potential victim

to

threa ts resulting from

all possible types

of

EMI that are present

in

the victim's electromag-

netic environment.

It is

necessary

to

define the potential threats

and

229



230 CABLES AND CONNECTORS

identify the primary EMI coupling mechanism for each threat/victim

combination. For each EMI threa t, it is necessary to define the emission

levels vs. frequency.

The optimum method

for

interconnecting equipments depends

on a

number of factors that include cable param eters , type of EMI (i.e., radi-

ated emissions, radiated susceptibility

or

crosstalk, common-mode,

dif-

ferential-mode), EMI sources, frequencies of intended and EMI signals,

installation,

etc.

Some

of the

more important parameters

are

shown

below.

Cable Physical Parameters

• Approximate range of lengths

• Approximate range of diameters

• Number of conductors

• Size of conductors (wire gauge)

• Configuration of conductors

• Configuration of shields

Cable Signal Parameters

• Signal frequencies

• Signal levels

• Radiated emission requirements

• Radiated susceptibility requirements

• Crosstalk

• Cable runs

RF Environment

• Major sources of EMI

• Transmitters

• Shielding

• Transients

• Electric fields

• Magnetic fields

• Plane waves

9 1 Factors that Affect Shield Termination Guidelines

Some general principles may be applied to help decide how to best ter-

minate the shield. For example, consider a shielded wire pair as shown



FACTORS

THAT AFFE CT SHIELD TERMINATION GUIDELINES

231

in Fig. 9.1. Exposing the shield to an internal or external electromag-

netic field will result in a displacement current flowing in the shield.

This displacement curren t is due to the capacitance of the shield, which

is acting like a high-impedance antenna (e.g., a whip antenna). In order

to provide effective shielding, the EMI currents induced on the shield

must be terminated in a manner tha t provides these induced currents a

low-impedance path to the termination, which will be a function of the

length of the shield relative to the wavelength of the EMI signal.

It is particularly important that the EMI voltage between the shield

and the wire pair be minimized. If a significant voltage develops

between the shield and the wire pair, the EMI voltage will capacitively

couple from the shield to the wire pair. Referring to Fig. 9.1, the cable

shield and metal s tructure act like a transmission line with a short cir-

cuit on the load end. The shield must provide a low-impedance path to

the termination for the RF currents induced on the shield. For low fre-

quencies such that the path to the termination is electrically short (for

example, less than one twentieth of a wavelength), the shield will pro-

vide a low-impedance path to the termination. (If the shield is termi-

nated on both ends and is one tenth of a wavelength long, the

maximum distance to the termination is one twentieth of a wave-

length.)

However, as frequency increases, the impedance of the shield

increases. When the length of the shield is one quarter wavelength,

with the shield terminated on one end only, the shield impedance

approaches infinity, and an EMI voltage can develop between the shield

and the wire pair. For this condition, the shield is not effective, and the

EMI will couple into or out of the wire pair. If the length of the shield is

greater than a quarter wavelength, the impedance of the shield to the

termination exhibits large variations with alternating parallel and

series resonances every quarter wavelength as shown in Fig. 9.2.

r

H h

Z of Shield

above Ground

= 200Q

Fig ure 9.1 Development of shield termination.



232

CABLES AND CONNECTORS

log I

Parallel Resonances, I

p

.

log I

F ig u re 9.2 Shield impedance to termination.

In order to provide effective cable shielding at high frequencies

(>300 kHz), it may be necessary to term in ate th e shield at bo th e nds

and at intermediate points separated by less than one tenth of a wave-

length.

Control of the EMI coupling between a source and a susceptible

device is essential if EMC is to be achieved in a complex electronic sys-

tem. Cables have a major impact on the resulting EMI coupling

between elements of the system. Thus, in order to achieve EMC, it is

essential that extreme care be given to the ground system and the

interconnecting cables.

If the shield material is a good conductor (e.g., copper or aluminum),

and it is braided or foil, it will provide effective shielding at low fre-

quencies (<30 kHz) against plane waves and high-impedance EM fields

(i.e., E-fields). However, the shield will have very little impact on low-

frequency magnetic fields. In order to provide protection against low-

frequency mag netic fields, it will be necessary to use tw isted wire p airs

and/or relatively thick shields such as conduit or heavy braided shields

made out of a permeable ma terial.

In gene ral, balance d lines are preferred for interfac ing low-frequency

equ ipm ents . Balanced conditions will provide approx imately 20 dB of

common-mode rejection. All signal inputs and outputs should be bal-

anced with respect to the system common. Interconnects between

equipments should be shielded, twisted wire pairs or triads. The shields

of low-frequency signal lines should be terminated at only one end to

the system common internal to the equipment. If unbalanced signal

lines must be used, the signal return should be terminated at one end

only.



FACTORS THAT AFFEC T SHIELD TERMINATION GUIDELINES 233

It is important to consider the trade-offs between common-mode and

differential-mode effects. Term inating a shield at both ends will provide

a 6-dB improvement in the differential-mode shielding effectiveness,

but it may create common-mode problems. Therefore, for the low-fre-

quency case, where the shield is electrically short, the shield should be

terminated at one end only. Guidelines for terminating the shield for

this case are presented below:

• Shields of sensitive data lines should be terminated at the load end.

• The shields of high-level signal lines should be term inated at th e

EMI source end.

Individual shields on low-frequency signal wire pairs within a cable

bundle must be insulated from each other to minimize crosstalk. Fur-

thermore, these shields must be isolated from overall cable bundle

shields, equipment chassis, conduit, and all other elements of the sys-

tem. At terminating equipments, the shields on low-frequency signal

wire pairs may be allowed to enter the case individually, on separate

pins, or they may be connected together and carried into or out of the

case on a common connector pin. If a common pin is used, it must not

compromise the floating or single-point termination. I t is recommended

that one pin be used for low-level signal shields and a separate pin be

used for high-level signal shields. These individual shields should be

term inated to the low-frequency signal reference. Pigtails should be as

short as possible. Several options for terminating the shield are illus-

trated in Fig. 9.3. As frequency increases, the shield impedance will

increase and exhibit multiple resonances. When this happens, the sin-

gle-point termination becomes ineffective, and it is desirable to termi-

nate at m ultiple points.

If a cable bundle contains twisted, shielded pairs/triads, the primary

purpose of the shields is to prevent crosstalk between the wire pairs/tri-

ads. The twisted, shielded pairs/triads will also help to reduce radiated

emission and susceptibility problems. However, if the only purpose of

the shield was to reduce radiated emissions or susceptibility, an overall

shield would be more appropriate than individual shields on each wire

pair.

If the cable bundle has an overall shield, the primary purpose of the

overall shield is to protect against radiated emissions from or radiated

susceptibility to the equipment via the wire pairs. To shield against

high-frequency EM fields, the overall cable shield should be bonded to

the equipment case at both ends, as shown in Fig. 9.4, to provide a con-

tinuous RF shield barrier. For low-frequency EM

fields,

t may be better

to terminate the shield at one end only to avoid common-mode EMI.



234


To Chassis

To Chassis

Jumper and Lug

Connector

* Designated Ground Pin

Jumper and Connector Pin

Bonding Halo

g

. ^

Insulation

O v e r a l l

orFerrule

Shield

Size:

No.

16 AWG

or Larger

Length:

50

mm or Less

To Equipment

Case

Terminal Strip

Pigtails

Fig ure 9.3 Methods for terminating the shields on wire pairs.

9 2 System Design for Interconnected Equipments

A

number of options

for

terminating the shield are available to the sys-

tem designer,

and the

choice

of

these options

can

have

a

significant



SYSTEM DESIGN FOR INTERCONNECTED EQUIPMENTS

235

Fig ure 9.4 Shield is bonded to the equipment cases to form continuous enclo-

sure.

impact on the resulting EMI/EMC performance of a system. For exam-

ple, consider the instrument and control system shown in Fig. 9.5 that

consists of two interconnected equipments. The equipment on the left

may consist of a sensor that is monitoring a process and is providing

information to the equipment on the right, which is controlling the pro-

cess. Power is provided to both equ ipm ents. The two equipm ents mu st

operate in an electromagnetic environment that contains a number of

potential sources of both conducted and radiated EMI. Some of the

sources of EMI include licensed transmitters and man-made noise from

the power supply and other sources.

Figure 9.5 shows that there are four basic options for terminating

the shield at each end. They are:

• Float the shield,

• Connect th e shield to th e signal reference,

• Connect th e shield to the equip me nt enclosure, or

• Connect th e shield to ground .

float? float

Two Ends, Four O ptions

=

2

4

or

16

Com binations

The Typical Dilemma of Shield G rounding

Fig ure 9.5 Options for terminating the shield.



236


The result is that there are 16 basic options for terminating the

shield. For any given application, some of these options would result in

EMC, and some of the options would result in EMI. In general, there

may be more than one option that would work from an EMC perspec-

tive, bu t some of them ma y be be tter t h an oth ers. Also, it is necessa ry to

factor in other considerations such as cost, size, weight, reliability,

maintainability, and so forth.

The specific system being considered is a floating system as shown in

Fig. 9.6. That is, the system common is free to float with respect to the

metal enclosure.

The objective is to determine the optimum configuration for termi-

nating cable shields. In general, there are four options at each end of

the cable. One option, shown in Fig. 9.6, is to allow the system common

to float with respect to the metal enclosure.

A second option is to connect the shield to the m etal enclosure as shown

in Fig. 9.7. Because th e system common is floating, ther e m ay be a voltage

developed between the system common and the metal enclosure, and this

could result in EMI. For this case, EMI currents induced on the shield (as

Sensor

System common is not

tied to metal enclosure and

can float with respect to enclosure.

System Common

Return

V

N2

Metal Enclosure

Figure 9.6 Inst rum en t and control system floats with respect to m etal enclo-

sure.

Sensor

I/O Card

Option 2—Connect shield

to metal enclosure ground.

I

System Common

^Return

7

Metal Enclosure

Figure 9.7

Cable shield connected to m etal enclosure ground.



SYSTEM DESIGN FOR INTERCONNECTED EQUIPMENTS 237

a result of exposure to the EM environment) can result in a voltage being

developed between the shield and the wire pair inside the shield. This is

not the recommended configuration for terminating the shield if the sys-

tem common is floating. It would be appropriate to use this termination if

the system common was connected to the metal enclosure.

The third option is to connect the shield to the system common as

shown in Fig. 9.8. In this configuration, currents induced on the shield

will be diverted to the system common return. If the shield provides a

low-impedance path to the system common, there will not be a signifi-

cant EMI voltage developed between the shield and the wire pair. This

is the recommended configuration for terminating the shield. Because

the system common is floating, there may be a voltage developed

between the system common and the metal enclosure, and this could

result in EMI.

The existence of a large number of options makes the EMC design

issue difficult to deal with. The designer must have a detailed under-

standing of the impact of the various options on the resulting EMC of

the total system.

9 2 1

Cable Shield Termination Guidelines

There

are

several general guidelines that

may be

used

to

determine

how to best term inate the cable shields for an instrument and control

system. These general guidelines depend on the length of the shield rel-

ative to wavelength. Table 9.1 provides the relationship between fre-

quency and wavelength.

• For optimum protection, use a cable with each wire pair/triad

twisted and shielded to prevent crosstalk and an overall shield

around

the

bundle

to

prevent effects

to or

from

the

EMI environ-

ment.

Q

I/O

Card

Sensor

System Common

Option

3

— Connect shield

to / |

Return

floating system com mon retu rn.

Metal Enclosure

Figure 9 8

Cable shield connected to floating system common return.




Table

9.1 Relationship between

Wavelength and Frequency

Frequency

10 kH z

30 kHz

100 kH z

300 kHz

1 M H z

3 MHz

10 MH z

30 MHz

100 MH z

Wavelength m)

30,000

10,000

3,000

1,000

300

100

30

10

3

• To prote ct again st high-frequency (>300 kHz) EMI effects invo lving

the EM environment, the overall shield should be bonded to the

equipment enclosure at both ends to form a continuous enclosure.

• To prote ct ag ain st low-frequency (<30 kHz) EM I effects involving

the EM environment, the overall shield should be bonded to the

enclosure at the victim end to protect against susceptibility or at

the source end to protect against emissions.

• To protect aga inst low-frequency (<30 kHz) cross talk, ba lanc ed

lines are preferred, and the shields on each wire pair/triad should

be term ina ted at one end only as follows:

D

To reduce susceptibility, th e shield should be term ina ted to the

signal return at the victim end.

D

To reduce em issions, the shield should be term ina ted to th e sig-

nal return at the source end.

• High frequencies sho uld not pre sen t a problem, because all of the

sensor signals are at low frequencies.

• Pigtails may be used to ter m in ate low-frequency shields, bu t they

should be made as short as possible.

• To shield again st E-fields an d plan e waves, th e shield m ate rial

should be a good conductor.

• To shield ag ain st H-fields, the shield should be ma de from a high-

perm eability m ater ial. Twisted pairs a re also effective for reducin g

H-fields.

Table 9.2 Provides guidelines for terminating shields.



SYSTEM D ESIGN FOR INTERCONNECTED EQUIPMENTS

Table 9.2

Guidelines for Term inating Cable Shields

239

Purpose of

Shield

L <-

2

L >

2

To reduce suscep-

tibility? (EMI is

outside the

shield.)

To reduce emis-

sion? (EMI source

is inside th e

shield.)

Electrically short shield.

System

should be a single-point ground

(STAR) wi th prefe rably 0 volt-to-

chassis connection at receiver side.

Ground cable shield at this point.

Ordinary braid and short pigtail

are acceptable.

Electrically short shield. System

should be a single-point ground

(STAR) w ith p referably 0 volt-to-

chassis node at transmitting side.

Ground cable shield at this point.

Ordinary braid and short pigtail

are acceptable.

Electrically long shield.

Ground both ends of shield to

chassis. Use low-Z

t

shield and

integral clamp. No pigtails .

Electrically long shield.

Ground both en ds of shield to

chassis. Use low-Z

t

shield an d

integral clamp.

No pigtails .

9.2.2 Tw isted Pairs to Red uce M agnetic Coupling

Magnetic coupling into or out of interconnecting cables can be reduced

by using a dedicated ground return with each signal wire and twisting

the wire pairs (i.e., the signal wire and the corresponding dedicated

ground return). The twist tends to make the EMI contributions from

the adjacent loops cancel, since the induced affect in each incremental

twist area is approximately equal in amplitude and out of phase as

shown in Fig. 9.9. Referring to Fig. 9.9, twisting the wire pairs reverses

the direction of current flow n the adjacent

loops,

and this will result in

cancellation of the magnetic field in adjacent loops. The coupling rejec-

tion provided by twisting the wire pairs is a function on the number of

twists per wavelength and the total number of twists over the wire

length as shown in Fig. 9.10.

Magnetic Fields from a Ttoisted Pair of

Conductors Transposition)

Area 2

T

r A r e a l

Fig ure 9.9 Twisted wire pairs.



24


Total Twists, n€ , Over Wire Length

S

§

4

5

60

n - Twist per Meter

€ -

Wire Length in meters

X - W avelength in mete rs

5 10 30 50 100 300 500

Total

Twists,

n€ , Over Wire Length

F igure

9.10

Coupling rejection

for

twisted wire

pairs.

Ik

3k

9.2.3 Sh ield ed Cable Configurations

Figure 9.11 shows an assortment of different cable configurations that

include overall shields that protect against external environmental

effects and unshielded and shielded bundles of wires that include

twisted wire pairs, shielded wire pairs, and twisted shielded wire pairs

to prevent crosstalk.

9.3 Connectors

A connector is an assembly of mating contacts that permits quick link-

ing and separation of a cable with another cable or equipment. The

number of wire pins and/or coaxial sheaths making simultaneous con-

tacts may range from two to several hundred. Individual pin contacts

are embedded in insulating material to mutually isolate them and to

prevent contact with bare hands. In a link or engaged position, the con-

nector should provide a low-impedance path for all internal wires and a

low-impedance bond when an outer shell is used.

This section surveys the connector component with emphasis on EMI

control. The connector backshell, one of the major points of EMI pene-



CONNECTORS

241

F ig u re 9.11 Assortm ent of shielded cables.

tration or leakage, is discussed. Other forms of connector problems and

EMI control are reviewed along with filter-pin and unbalanced connec-

tors and adaptors.

9.3.1 Sh ield Term ination Concepts

If the shield is not properly terminated, induced RF currents that are

conducted along the cable shield may be coupled to the system wiring.

When the shield is properly terminated, the entire periphery is bonded

to a low-impedance reference. This minimizes the RF potentials at the

termination. The proper concept for terminating the shield is shown in

Fig. 9.12. The shield term ina tion concept shown in Fig. 9.13 is an exam-

ple of a bad practice. The shield should not be allowed to penetrate the

wall of the equipment enclosure.

Figu re 9.14a shows a bad exam ple wh ere th e shield is not bonded to

the o uter wall of the enclosure. In this exam ple, the shield pe net rates the

enclosure, and EMI on the shield is coupled into the enclosure. In 9.14b,

the outer sh ield is bonded to the outer w all of th e enclosure, bu t the inn er

shield p ene trate s th e enclosure and EMI on the inn er shield will be cou-

pled into the enclosure. Figure 9.14c is better because the inner shield is

terminated with a pigtail (which is not the best method of terminating

the shield) at the wall of the enclosure, and a twisted wire pair is used to




Strain Relief

Clamp o r Solder

Connection

Metal

Circuit

Enclosure

Bushing

ToPCB

Proper Connection

Fig ure 9.12 Shield termination concept.

Reradiation from

Long Cable Shield

Termination

Improper Connection

Fig ure 9.13 Bad practice for shield termination.

conduct the signal inside the enclosure. Figure 9.14d is the best method

for terminating the shield. The inner shield is terminated with a shorting

sleeve. If the signal is noisy, continue the inner shield inside the enclo-

sure and float he shield at the amplifier.

Figure 9.15a illustrates a permanent termination of the cable shield

to a connector. Here, the outer shield is made continuous with the con-

nector backshell by a soldering or metal-forming bond. Spring fingers

are used to carry the shell continuity to the mating connector. The illus-



CONNECTORS 243

I

Module Wall

C o m m o n - m o d e c u r r e n t |

Inner shield

(a)

Outer shield

j

U

\ l O p F

x

= 160

ohms

atlOOMhz

CMC-

Better

(c )

Module Wall

I

Module Wall

CMC

CMC

Note:

Continue inner

Shorting shield

if

inside

sleeve

\

s

noisy. Float

•*

shield

at

amplifier

fr

(

b

) Poor jj^Bond

( d ) B

est

jj

Module Wall

F ig ure 9.14 Comparison of shield termination practices.

tration also shows the through path for unshielded individual connec-

tors . When more than one shielded inner conductor must be routed

through a single cable and connector, the technique suggested in

Fig.

9.15b is employed to preserve individual internal-wiring shielding.

The internal coaxial shields should never be pulled back, twisted, and

then bonded to the outer connector sheath; i.e., no portion of the coaxial

shield should be broken before it is bonded to the connector shell. Indi-

vidual shields for connections tha t a re routed th roug h m ulti-pin coaxial

connectors should be terminated individually in the manner illustrated

in the figure.

Figure 9.15 indicated tha t the cable shield is perm ane ntly secured to

the connector shell. While offering the best bond, this practice is not

particularly cost effective in manufacturing time. Methods of quick

mechanical compression bonding of the cable braid to the shell have

been developed by EMC connector manufacturers. Many such connec-

tor varieties permit rapid assembly, require no special tools, are field

repairable, and permit environmental sealing. They are available in

both permeable and nonpermeable-base materials to shield against

both H- and E-fields or E-fields only.

9.3.2 Connector Backsh ells

Connector ty pes m ay be divided into three classes: (1) low-frequency

single- an d twin-conducto r co nnec tors, (2) low-frequency multi-pin con-

nectors , and (3) high-frequency unb alanced-line (coaxial, triax ial, an d



CABLES AND

CONNECTORS

Preferred: Fillet Weld Around Entire

Periphery of Female Connector Housing •

Alternative: Bolt and Tooth Type

Lack-Washer Connection as Shown

by Dotted Outline

Male Contacts

Shield

Continuous Shield-to-Shell Bond

By Solder or Metal Forming

(Never Pig-Tail the Shield)

Spring Contacts (Shield Makes Before

and Breakes After Enclosed Conductors)

a) Individual Conductors Are Unshielded

Spring Fingers

Shielded Conductor

Recessed Contacts

Connector Shell

(Male Section)

Female Portion

of Connector

b) Individual Conductors Are Shielded

igur 9 15 Shield termination

for

electrical connectors courtesy AFSC

Design Handbook DH 1-4 EMC .



CONNECTORS

245

quadrax cable) connectors. The dichotomy here between low and high

frequency may exist anywhere between 100 kHz and 10 MHz. Coaxial-

cable connectors

are

discussed

in a

later section. This section discusses

the multi-pin connector and shielding of its outer shell to mitigate radi-

ation leakage and penetration.

Multi-pin connectors generally have an external shield that slips

over the harness

at

the connector and secures to the conductor termina-

tion or mating shell. The backshell, as it is called, serves as a form of

strain relief and provides a 360° peripheral shielded configuration

around the harness assembly

at

the wire-connector interface. The back-

shell also serves to term inate (i.e., bond) the shield to either a connect-

ing housing or another mating connector shell assembly. Thus, a good

multi-pin connector is one in which the shielding effectiveness of the

mated connector equals or exceeds th at of an equal length of the inter-

connecting cable shield.

Figure 9.16 provides an illustration of the cable as a system. The

cable has a backshell, and the shields are terminated at the bulkhead.

The ferrite slug is optional and would be used for common-mode sup-

pression if required. The cable has an external shield that is used for

protecting against EMI effects associated with the external electromag-

netic environment. The connector may also have filter pins

and

planar

capacitor arrays.

Figure

9.17

shows

an

example

of a

shielded sub-D connector,

and

Fig. 9.18 shows

an

example of a CFC/CFD/CFX connector.

9 3 3 Termination of Individual Wire Shields

When

a

cable harness contains many individual shielded wires,

in

which each shield acts

as a

Faraday cage, continuity through

the mat-

Bulkhead

Common-Mode [

*p

Current

Internal

Lossy lines

And/Or

Absorptive

Jacket?

Fig ure 9.16 The cable as a system.



246


\

Metal Bulkhead or

Metalized Plastic Case

Subminiature

HDorT^peD

Connectors

(Metal or Metalized

Plastic)

Foil-Wrap in Contact with

Inside of Backshell or Boot

Cigarette Foil-Wrap

Shielded Ribbon Cable

PVC or Other Protective Jacket

(Strain Relief Not Shown)

Optional Braid Shield

Exposed Cigarette

Foil-Wrap Shield

\l) Metal Backshelt or

(2) Inside Metalized Molded Plastic or

(3) Metalized Heat-Shrinkable Boot

*

To Float Ungrownd) Foil Wrap-to-Backshell at lo w

and Mild F requen cies Interpose 40 M icron Mylar

Layer to Produce Approximately 1000 pF of RF CAP

Fig ure 9.17 Shielded sub-D connector.

Equivalent C ircuit

1000 pF

Metric Jackscrews

Flexible Outer Cover

Outer Braid Shield

Drain Wires

Cast Aluminum Nickel-Plated

Cast Connector Guard

4 Twisted P airs

8 Twisted Pairs

Inner Mylar Shield

Overlapped S eams

Reliable Strain Relief

Fig ure 9.18 CFC/CFD/CFX connector.

ing connector interface is obtained via a n indiv idual p in for each shield.

This suggests that, for non-coaxial pin connectors, twice as many pin-

receptacle contacts are ne cessary for individual shielded-wire cables. To

cut down on th e nu m ber of extra pin contacts ne cessary for shield conti-

nuity, a technique of

daisy-chaining

is sometimes employed where th e

harness contains many wires carrying signals from de up to about

1 MHz. In this technique, a single dedicated pin is not used for the con-

tinuity of each individual wire shield in the assembly. Rather, one pin

may ca rry up to five individ ual w ire shields connections.



CONNECTORS

247

The daisy-chain practice is to peel back the outer braid of each

shielded wire an d connect thes e braid s in groups of five by an insu lated

wire looping from outer shield to outer shield in a daisy-like manner.

The final wire from the shield group goes to a separate dedicated feed-

through pin. While this practice compromises Faraday shielding

between short lengths of resulting unshielded wires at high frequen-

cies,

reliance is m ade up on th e ou ter cable connector backs hell for over-

all shielding at the cable-connector interface. Cables carrying signals

above 1 MHz should n ot employ this practice; in fact, daisy-chaining is

regarded as obsolete, since multi-coaxial pin connectors are now avail-

able to give be tter performance.

One alternative to the daisy-chain technique is the halo-ring tech-

nique, in which individual shielded wires in a harness must have a

common shield ground at the connector. Here a cylindrical conductor

(the halo) is used as shown in Fig. 9.19 to connect all applicable shields

to ground through one or more connector pins. Where final termination

is to exist at an equipment housing, shield halos should be bonded to

the ground plane by 1.5 in (3.8 cm) or less of 0.25 to 0.5 in (6.35 to

12.7 mm) wide, tin-plated , copper st ra p.

The halo technique is acceptable only when a relative few shielded

wires a re involved. A preferred method wh ere cost implications become

important is to use a collectively crimped peripheral ring as illustrated

in Fig. 9.20 for all wire shields exclusive of those inten tiona lly ope rated

as eith er in dividu al coaxial cables or low-level audio shielded lea ds. Th e

collective crimping ring uses two ground wires. Connect one wire from

the ring to the connector shell where connector design permits. The

other wire is carried through the connector. Fig. 9.20 shows what the

resulting outer shield grounding configuration would look like.

Bond strap

Pins

Receptacle

N

This Halo is the

same as halo on

other side of

panel

Bond Soldered sleeving >

strap joint \y

-Panel

Figure 9.19

harness.

Bonding ring or halo at connector for terminating shields in a



248


Shield carried under

connector shell

Connector shell

Strain reliever

Crimping ring

Crimping ring

ground wire

through connector

Crimping ring ground

wire to strain reliever

Shield pigtail returned

to crimping ring

Fig ure 9.20 Crimping-ring technique for terminating shields.

The best performing method to use, but relatively expensive to man-

ufacture, is called the interlacing-strap

method,

shown in Fig. 9.21. It is

used for a common shield ground in multi-shielded wires in harnesses

that have a large num ber of individual intern al shields. The interlacing

strap should be a t least 0.25 in (6.35 mm) wide by 10 mils thick and be

bonded securely to the connector as shown in Fig. 9.22.

Where multi-shielded wires are to protect audio-susceptible circuits,

they should be grounded at one end only as shown in Fig. 9.23. Individ-

ual twisted-wire pair shields should each be insulated from other pairs

Connector

body

Soldered

connection

Strap as w ide as possible-

Fig ure 9.21 Interlacing technique for terminating shields.



C

e

v

c

m

n

r

n

8

C

m

n

r

n

1

g

o

o

s

r

a

n

r

e

e

v

o

c

o

o

B

k

d

c

o

o

§

S

e

d

c

e

d

t

h

o

c

o

o

•

S

e

d

g

o

d

a

b

h

e

n

F

g

e

9

2

W

i

r

n

s

y

e

m

e

m

n

o

o

s

h

e

d

d

w

r

e



B

k

d

\

C

o

S

e

d

G

o

d

a

O

E

O

y

S

e

d

C

e

d

T

o

C

o

o o

F

g

e

9

2

T

m

n

o

o

s

h

e

d

d

a

u

o

s

u

c

p

b

e

w

r

e



CONNECTORS

251

to prevent undesired grounding; the shield should never be used as a

signal return.

9.3.4 Filter -Pin Conn ectors

Filters offer significant possibilities for controlling conducted interfer-

ence. Generally, EMI filters are employed as lumped elements in vari-

ous portions of circuits and input-output wiring of equipments.

However, filters have been miniaturized to such small sizes that some

can now be bu ilt into the cable-pin assembly. Figu re 9.24 illu stra tes one

type of miniature multi-pin connector employing 7i-type filters in each

pin. Because of the limitation of the obtainable shunt capacitance and

series inductance that can be constructed in the pin, filters of this small

type, typically about 1/8 x 3/8 in (3.2 x 9.5 mm) in size, exhibit little or

no atten ua tio n below 1 MH z. Typical atten ua tio n offered by the se filter

pins in a 50-Q system is about 20 dB at 10 MHz and up to 80 dB at

100 MH z.

Another filter-pin connector of a somewhat larger body dimension is

designed to carry 5 A. Th us, for low de working voltages, capacitanc es

up to about 1 ^F are achievable in the larger pins. Figure 9.25 shows

insertion loss vs. frequency, Many of these filters exhibit cutoff fre-

quencies of the order of 100 kHz when measured in a 50-Q system per

MIL-STD-220A.

9.3.5 Coaxial Conn ectors

For applications above 10 kHz, and more typically above 10 MHz,

employed connectors are of the unbalanced-line, coaxial type so as to

F ig u re 9.24 Typical m inia tur e mu lti-pin filter connector.



252


m ate w ith coaxial cables.* Connectors of thi s type m ain tain

a

360°, low-

impedance integrity of the outer shield through the connector interface

to the mating connector assembly. A low-impedance shield is extremely

important, since this impedance exists

in

the return-wire path

of

the

associated coaxial cable. Thus, the outer-cable shield impedance of the

termination

is

of param ou nt consideration

in

th e perform ance of coax-

ial connectors.

One significant EMI problem with coaxial connectors is the imped-

ance mismatch in

a

50-Q, 72-Q, or other cable chara cteristic imp edanc e.

Impedance mismatch is ra ted in terms of maxim um voltage standing-

wave ratio (VSWR) vs. frequency. The maximum signal amplitude vari-

ation as a function of VSWR is the amplitude of the VSWR, per se. For

example, depending upon the length of cable, a connector rated with a

VSWR

of

2:1

at a

particular frequency could exhibit

a

6-dB peak-to-

peak variation in signal or EMI a mp litude . Th us, connector VSWR

becomes very important, especially at frequencies for which an associ-

ated cable length approaches or exceeds

X/8.

9 3 6 Summary of Connector Characteristics

Ideally, connectors shou ld hav e the following char acteristic s:

• Negligible resis tanc e

• Chemically ine rt surfaces

• Resistan ce to gouging

Rang e of

Insertion Loss

- Frequency (mega hertz) ~~

1 10 100

F ig u re 9.25 EMI filtering connector insertion loss.

* Crosstalk between wires a t high frequencies is due to electric-field coupling. To provide

both a return-wire path and a shield at and above HF, coaxial lines an d connectors are

used notw ithstanding some balanced, parallel lines used at HF/VHF.



CONNECTORS

253

• Foolproof alignm ent to minimize contact dam age

• Ad equa te force betw een contacts

• Little friction to minim ize increas e in resis tanc e with use

• Contamination-free design

• Provisions for proper connections, including sh ielding of backs hell

• Prop er dielectric prop erties

• Moisture-proofing as required

• Resistan ce to degra datio n due to age, wear, m ainte nan ce, and

repair

• Filter pins incorpo rated if necessa ry

• Comp atibility regardless of varying intersystem contractors

There should always be a proper installation, including a good bond

between the cable shield(s) and connector shell, as shown in Fig. 9.26.

Shields should be bonded completely around the periphery of the con-

nector body. All connectors used as conducting pa th s for E MI sho uld be

bonded to the static ground with de bonding resistance of the order of

1 mQ or better. Air gaps at th e con nector-chassis interface should be

eliminated by the use of woven-mesh EMI gasketing. Other desirable

features are protective coverings that extend over the male pins to

reduce pin damage, the use of caps on unused connectors, the use of

clamps to hold wires steady, contact m ate rial s designed for long life and

Shield Grounded

Around Periphery

Moisture Seal

Sheath

to Chassis-

•Chassis

Signal Ground Bus

* ^o r Bonding Strap

Shield

Termination

R < l m a

Monel RFI Gasket

Fig ure 9.26 Connector shielding and grounding.



254


proper pressure,

and no

loose

or

faulty con tacts th at might generate

EMI.

Filter pins may be used where interference is in the VHF and UH F

rang e. Most filter p ins a re not effective below 1 MH z. Use feed-through

capacitors or filters mounted in a connector box where cond ucted inter-

ference is below 1 MH z.

9 3 7 Summary of EMI Control Techniques for

Connectors

Table 9.3 summarizes EMI fixes for field-to-cable differential mode cou-

pling. Figure 9.27 illustrates

the

application

of a

ferrite

for

common-

mode cancellation. Table

9.4

summ arizes techniques

for

controlling

wire and cable crosstalk.

Table 9.3 EMI Fixes for Field-to-Cab le Differential-Mode Coupling

Balanced W iring

• Reduce wire pair insulatio n or increase

AWG

• Twist wire pa irs or increase twis t pitch

• Use twisted, shielded pairs

• Do not use pigtails, drain wires, or term ina l blocks

• Use braid-foil shields with shielded backshe lls

• Route cables in cable tray s or raceway s

• Route cables in conduit

Coaxial Cable

• Use braid-foil shields with low Z

t

connectors

• Use quad (braid-foil-braid-foil) shields

• Route cables in cable tray s or raceways

• Route cables in conduit

Ferrite Common-Mode Cancellation

Common-mode Insertion loss:

lL

dB

- 20 log

for z

cm

= 100 ohms (default)

Z

femte

50

100

150

200

300

500

IL

dB

1.9

3.5

4 9

6.0

8.0

10.9

F ig u r e 9.27 App lication of a ferrite for common-mode cancellation.



CONNECTORS 255

Table 9.4 Techniques for Controlling Wire and Cable Crosstalk

• Do not mix different cable types in bundle s.

• Route cables close to chassis or ground p lane.

• Place cable types into different cable tray s or sep arate d with in the same tr ay

• Use coax for RF signals.

• Isolate digital signal cables.

• Twist and shield low-level analog cables.

• Do not expose cable bun dles to ap ertu res or openings.

• Cross bundles at right angles.



Chapter 10

Summary of MI Control

Techniques

Control of EMI between sources and susceptible devices is essential if

EMC is to be achieved in a complex electronic system. The electronic

system designer must give careful consideration to the components that

are used in the system and must define their EMI interactions with

each other and with the system operational environment. The system

designer must define the EMI suppression and control requirements

that are necessary to achieve EMC. Also, the system designer must

define the various EMI/EMC regulations and standards that apply to

the system and must design the system so it satisfies these regulations

and standards.

Chapters 1 through 4 provided background information to address

system design for EMC. Chapter 5 discussed the considerations that

must be applied to the selection and design of a ground system.

Shielding is a major means of controlling radia ted EMI effects. Chap-

ter 6 addressed the shielding effectiveness of various materials for elec-

tromagnetic fields. Chapter 6 also addressed the design of metal

equipment enclosures with various openings such as seams, cooling

apertures, instrument displays, etc., which tend to compromise the

shielding integrity. Techniques that may be used to protect the shield-

ing integrity of these openings were described in Chapter 6.

Chapter 7 presented methods for bonding conductors such that a

low-resistance path is established between the two joined objects.

Filters, ferrites, and isolators were the major control devices for con-

ducted EMI. These devices were discussed in detail in Chapter 8. The

material presented in Chapter 8 described where to use filters, ferrites,

and isolators; how to select the proper device for a specific requirement;

how to install the device so that optimum performance is realized; etc.

7



258

SUMMARY OF EMI CONTR OL TECHNIQUES

The interconnecting cables and the system ground scheme have a

major impact on the resulting EMI coupling between elements of the

system. Therefore, in order to achieve EMC, it is essential that extreme

care be given to the ground system and the interconnecting cables.

Chapter 9 addresses cable and connector problems.

This chapter summarizes the EMI control techniques that may be

used to design a system for EMC.

As mentioned above, there are many options for controlling EMI in a

system. Table 10.1 lists some of the many EMI

fixes

ha t may be used to

fix EMI problems. Table 10.2 provides guidelines for EMI control for

components, printed circuit boards and interconnects. Table 10.3 is a

matrix th at shows various EMI

fixes

vs. applicable EMI coupling paths.

This matrix can be used to select fixes hat apply to particular EMI cou-

pling paths.

Table 10.1 Some of the Many Available EMI Fixes

Aerial terminals

Air and RF niters,

honeycomb

Air and RF niters, mesh

Aperture-leakage control

Backshells, shielded

Balanced circuits

Balun transformers

Bonding techniques

Bonds

Cable, absorption ferrites

Cable shields

Cable trays

Cable tray covers

Caps,

bypass

Caps, feed through

Caps, R F foil le ads

Choke, for isolation

Component shields

Conductive caulking

Conductive coatings

Conductive com posites

Conductive epoxy

Conductive grease

Conductive tape

Earthing techniques

Ferrites

Ferrite connectors

Ferrite-loaded cables

Fiber optics

Filter-pin connectors

Filters, power line

Filters, signal line

Floating techniques

Gas tubes

Gaskets, electrical

Grounding hardw are

Grounding methods

Grounds, instrumentation

Grounds, safety

Grounds, signal

Guard shields

Knitted-wire m esh

Inductor in safety ground

Isolation transformers

Metal conduits

Metal foils

Metal tapes

Metallized textiles

Motor-generator sets

MOVs (metal-oxide

varistors)

Optical isolators

Planar capacitor arrays

Snubbers

Shielded buildings

Shielded coax, quad shields

Shielded components

Shielded conduit

Shielded enclosures

Shielded isolation transformers

Shields, overbraid

Shielded racks and cabinets

Shield terminations

Shielded isolation transformers

Shielded rooms

Surface-mount EMC

components

Surge suppressors

Transient plates

Transient Snubbers

Transorbs

Tri-shields for coax

Twisting wires

Uninterruptible power supplies

Wire and mesh screens

Zener diodes

Zippertubing



SUMMARY OF EMI CONTR OL TECHNIQUES 259

Table 10.2

EMI Control in Components, PCBs, and Interconnects

Guidelines for C omponents

1. E nsu re th at the self-resonant frequency (SRF) of capacitors is above the high est fre-

quency to be bypassed.

2. Ensure that the SRF of inductors is above the highest frequency to be engaged.

3.

If you can't ensur e no. 1 and/or 2, add RF or microwave component(s) to tak e over at

higher frequencies.

4. Bypass EMI noise (-Ldi/dt) with caps or surge suppressors across the brushes of

motors and generators, or bypass each brush to frame ground.

5. Bypass EMI noise (-Ldi/dt) with caps or surge suppressors across the contacts of

relays and solenoids.

6. Select one of the low est-speed logic families con sistent w ith requ irem ents for digital

computation.

7.

Select one of the highest noise immunity levels (NILs) for logic devices.

8. Attempt to optimize no. 6 and 7.

9. Use ground ing st rap s or foils with length-to-width ratios not to exceed 5.

10.

Protect active devices again st R F demo dulation (i.e., audio rectification).

11.

Wrap excess cable into a serpentine (back-and-forth) pattern and tie. Do not wrap

into a helix or coil.

12. When shielding cables, use foil-braid combinations.

13. Where lightning surge protection is needed for active devices, use hybrid gas tubes

and solid-state surge su ppressors.

14.

For large capacitor filters, such as for shielded enclosures, protect with inp ut induc-

tors and surge suppressors.

Guidelines for Printed Circuit Boards

1. Make board trace height as low as possible.

2. Do not route trace s closer to the PC B edge than 3 x heigh t above the ir image p lane

(or 3 x board thick ness).

3.

Bury clock traces below deck, defined as being bound by two image ground pla nes ...

4.

...and/or use guard tr aces on noisy clock lines. Ground these tra ces to image plane at

least every tenth wavelength, d, at the clock frequency, f]yiHz> or d < ?>Q/(Ji meters)

where e = PCB dielectric ma terial (use = 4 as d efault).

5.

Use m inimu m num ber of vias (say, 2) for clock lines an d noisy high-speed lines.

6. If clock and oscillator sources are noisy, consider SMT shielding the comp onents.

7. Do not stack more than one trace height with no signal trace layers in between.

8. Pu t analog an d digital circuits on different lay ers, or sep arat e a layer into two iso-

lated analog and digital areas, or separate digital with analog moat and bridge.



260 SUMMARY OF EMI CONTR OL TECHNIQUES

Table 10.2 EMI Control in Com ponents, PCB s, and Intercon nects (Continued)

9. Separa te parallel-run traces by not less than 2 x trace widths .

10. Select and mount decoupling capacitors having SRF above logic bandwidth.

11. Remove 20h of foil from edge of VCC return plane to reduce edge radiation (h =

board layer height or thickness).

12. To reduce cable radia tion from common-mode curren t, use segm ented g round

plane(s) in the PCB.

13.

Use top and bottom ground planes to further reduce radiation from mu ltilayer

boards by 10-15 dB or more.

14. For single-layer boards, use phantom ground image plane to reduce radiation by 20 -

35 dB.

Guidelines for Interconnecting E quipments and S ystems

1. Ensure that EMI control is carried out at lower levels before attempting to control

EMI at the interconnected equ ipment and system levels.

2. To reduce ground-loop are as, bring system component equ ipm ents closer togeth er if

possible.

3. To

reduce ground-loop are as, rou te open cables close to ground p lanes or large m etal

area masses.

4. Route interconnected cables inside metal conduit, cable tray s, or raceways when ever

possible.

5. Ensure that cable types are separated into electric power, analog, and digital/RF

before placing them into dedicated cable trays, raceways, or hangers.

6. If cable shields are required, use a combination foil-braid shield (for high- and low-

frequency threa ts), which can then b e grounded or connected to equipm ent housing

at entry/exit points.

7. Add external EMI filters or filter pin connectors (FPCs) or planar capacitor arrays

(PCAs) at ho using I/O connectors.

8. Do not bring in raw cables directly to equipment's intern al term inal strips without

first removing their common-mode currents at the equipment's m etal housing.

9. Fold back-and-forth excess cable into a serpen tine p att ern and tie . Do not fold cable

into a helix. Avoid the rat's nest syndrom e.

10.

Add snap-on ferrites at eq uipm ent I/O connectors if a small amou nt of common-mode

rejection (<10 dB) is needed.

11.

Add extern al cable filters or connector filters (FPC s and PCAs) at equ ipm ent connec-

tors if significant common-mode rejection (>10 dB) is needed.

12. Any technique that causes excessive bulk at the rear of a connector is undesirable,

such as the buildup of crimp ferrules.

13. 360° circumferen tial sh ielding at th e backshell of a harn ess con nector is achieved by

a m etal cover with a strain-relief clamp or conductive epoxy potting.



T

b

e

1

3

M

a

rx

o

E

M

I

F

x

v

E

M

I

C

n

P

h

F

x

=

C

o

u

n

P

h

4

1

C

m

g

o

m

d

n

2

R

d

a

e

d

f

e

d

o

n

e

c

o

n

c

b

e

C

M

3

R

d

a

e

d

f

e

d

o

n

e

c

o

n

c

b

e

D

M

4

I

n

e

c

o

n

c

b

e

o

r

a

d

a

e

d

f

e

d

C

M

5

I

n

e

c

o

n

c

b

e

o

r

a

d

a

e

d

f

e

d

D

M

6

C

b

e

o

c

b

e

c

o

a

k

7

R

d

a

e

d

f

e

d

o

b

8

B

o

r

a

d

a

e

d

f

e

d

9

B

o

b

ra

d

a

o

1

B

o

b

co

o

1

P

w

m

n

o

b

c

o

o

1

B

o

p

w

m

n

c

o

o

1w

2

w

3

4

m

5

6

7

8

9m

1

1

1

1m

1

1

1m

1

1

1

2

oo

£



T

a

b

e

1

3

M

a

rx

o

E

M

I

F

x

v

E

M

I

C

n

P

h

C

n

d

F

i

x

•

=

C

o

u

n

P

a

h

s

£

1

C

m

m

o

g

o

m

p

d

n

2

R

d

a

e

d

f

e

d

o

n

e

c

o

n

c

b

e

C

M

)

3

R

d

a

e

d

f

e

d

o

n

e

c

o

n

c

b

e

D

M

)

4

I

n

e

c

o

n

c

b

e

o

r

a

d

a

e

d

f

e

d

C

M

)

5

I

n

e

c

o

n

c

b

e

o

r

a

d

a

e

d

f

e

d

D

M

)

6

C

b

e

o

c

b

e

c

o

a

k

7

R

d

a

e

d

f

e

d

o

b

8

B

o

r

a

d

a

e

d

f

e

d

9

B

o

b

ra

d

a

o

1

B

o

b

co

o

1

P

w

m

a

n

o

b

c

o

o

1

B

o

p

w

m

a

n

c

o

o

2

2

m

2

2

2

2

2

2m

2W

3m

3

3

3

3

3

3

3w

3

3

4

4



SUMMARY OF EMI

CONTROL

TECHNIQUES

263

Figure 10.1 identifies techniques that may be used to control EMI

resulting from conducted coupling paths. Figure 10.2 and 10.3 summa-

rize considerations associated with grounding for EMC. Figure 10.4

summarizes

techniques that may be used to control

radiation,

pick-up

or crosstalk in cables. Figure 10.5 provides techniques for controlling

Source

Power supplies

Motors

Inductive loads

High

level

analog

Digital signals

Transmitters

EM

environment

Victim

Analog equipment

Digital equipment

Video

display

Recorders

Instruments

Sensors

Control systems

Receivers

Applicable EMI Control Techniques

Differential Mode

ower

• Filters

•

Ferrites

• Isolation

transformers

• Transient

suppressors

ignal

• Filters «

•

Ferrites «

• Isolation *

transformers

•

Transient «

suppressors *

Common

Mode ground loop)

ower

> Filters

•

Ferrites

•

Isolation

transformers

• Balanced system

> Moat

•

Inductor in

ground

ignal

• Filters

•

Ferrites

•

Isolation

transformers

• Balanced

circuit

• Float

• Inductor in

ground

•

Optical isolator

Figure 10.1

paths.

Summary of EMI control techniques for conducted coupling

Grounding for

EMC

• Use a

single

point ground if applicable ( ie . low frequency)

• Separate and isolate grounds for

AC

and

DC

power, analog

signals, digital signals, chasis, etc.

• Use a dedicated return for each critical circuit

• Use large ground conductors to minimize impedance

•

Do

not daisy chain

1

'

F igur e 10.2 EMI control techniques for common-ground impedance coupling.



264


Methods of Ungrounding by Increasing Ground-Loop

Impedance or O ther Actio n to Divert CM Currents Off

Victim's In put

• Float circuits, boa rds, boxes, and equipments

• R-F float equipment, cabinets, and consoles with &*F chokes

• Float box shields inside equipment enclosures

• Use balanced circuits

• Use isolation transformers

• Use faraday shielded isolation transform ers

• Add ferrite beads and rods

• Use feed-thru CM capacitors

• U se optical isolators

• Use fiber optics

F i g u r e 1 0 . 3 EM I contro l for gro un d loop couplin g.

EMI Control of R adiation o r Pick-up by C ables

( Common Mode )j

Field to able

able to Field

• Minimize loop area

(route cables close

to ground)

• Reduce operating

frequencies

• Shield entire system

• Fiber optics

(Differential Mode)

Field to able

able to Field

• Minimize loop area


frequencies

• Twisted wire pairs

• Twisted/shielded pairs

• Coaxial cable

• Fiber optics

able to able

• Reduce length of

common runs

• Level separation


frequencies

• Twisted wire pairs

• Shielded wire pairs

• Twisted/shielded pairs

• Coaxial cables

• Fiber optics

Fig ure 10.4 EMI control for radiation or pickup by cables.

ground loop radiation or pick-up. Techniques for controlling wire and

cable crosstalk are presented in Fig. 10.6. Shielding techniques are pre-

sented in Fig. 10.7. Guidelines for good bonds are presented in

Fig. 10.8.



SUMMARY OF EMI CONTRO L TECHNIQUES

265

Controlling Ground Loop Ead iatioii or Pick-Up

• Minimize the loop area by placing wires close

to the ground

• Keep wire lengths as short as possible

• Float Equipment and/or circuits on one or both

ends if possible

• Shield entire ground-loop area

Figure 10.5 Con trolling ground loop rad iation or pickup .

Control o f Wire and Cable C rosstalk

• Do no t mix different cable types in bundles

• Route cables close to chassis

• Use coax for analog signals

• Isolate digital signal cables

• Shield low level analog

• Do not expose cable bundles to ape rtures

• Cross bundles at right angles

Figure 10.6

Control of wire and cable crosstalk.

Shielding to C ontrol EMI

• Shield componen ts, circuits, equipment, wire s

and cables to control radiated EMI

• Use me tal or metalized plastics to provide

shielding

• Protect apertures to prevent leakage

• Provide good metal-to-metal contact or EMI

Gaskets to prevent leakage from seams

Figure 10.7

Shielding to control EM I.



266


General Guidelines for Good Bonds

• Good bonding is intimate contact between metal

surfaces

• Surfaces smoo th and clean

• No non-eonductive finishes

• Fastening method must exert enough pressure

to hold surf aces in contact

• Join similar metals or

• Choose washers (replaceable)

• Use protective finishes

• Do not use solder for mechanical strength

• Protect bond from moisture other corrosion causes

• Jumpers are only a substitute for direct bonds

• Keep short for low R, low L

• Avoid jum pers lower in electro-chemical seriers then

bonded members

• Keep length/width ratio less than 5

• Bond directly to basic structure rather than through

an adjacent part

• Use no self-tapping screw s

F i g u r e 1 0 . 8 Gu ide l ines for good bon ds .



Appendix A

Cable to Cable Coupling

This appendix provides the basic approach to predicting capacitive and

inductive cable-to-cable crosstalk based on cable lengths, wire separa-

tion, wire heights above ground, load impedances, and frequency. Fig-

ures A.I and A.2 show the circuit representation of capacitive and

inductive coupling between parallel wires or circuits.

The overall procedure for calculating cable-to-cable coupling is

shown in Table A .I.

Table A.2 presents the capacitive cable-to-cable coupling in decibels

norm alized to a 1 m leng th of 22 AWG ter m ina ted with 100 Q of imped-

ance.

Table A.3 presents the inductive cable-to-cable coupling in decibels

norm alized to a 1 m leng th of 22 AWG term ina ted with 100 Q. of imped -

ance.

2 x h

o

Victim

Voltage

Figu re A.1

Circuit rep rese nta tion of capacitive coupling betw een wires or cir-

cuits.

7



268 APPENDIX A

Figu re A.2

Circuit rep rese nta tion of induc tive coupling betw een wires or cir-

cuits.

Table A.1 Overall Procedu re for Calcu lating Cable-to-Cable Coupling

• Select frequency or l/nx

r

(x

r

= pulse rise time)

• Select applicable h increment:

h = /h

c

hy he = culprit and h y = victim w ire height

• Select nearest (or interpolate) wire separation, S, in mm

• Look up (or interpo late) applicable CCC in dB for:

CCCc = capacitive coupling

CCCj = inductive coupling

• Correct for impedance and common wire length

ty .

CCC

C

'

= CCC

C

+ 20 log

10

(Zy^/100)*

^ = CCCj + 20 log

10

(100^Z

c

)t

• Select larger of CC C

C

' and CCCj' (i.e., less negative in dB)

• In any case, clamp to 0 dB maximum—crosstalk cannot be positive

*Z

V

= emitter (culprit) circuit load impedance (Z

C2

).

*ZC = receptor (victim) circuit source (Z

V1

) or load (Z

V2

) if identical. If dissimilar, use

(

yA

I, where 50 Q. in the denominator is to ac-

=

Z

V 1

X Z

V 2

Z

V1

+ Z

V2

count for the parallel combination of the two 100 Q references.

Figu re A. 3 shows an illus trativ e exam ple of coupling between back-

plane wiring.



T

b

e

A

2

C

p

v

C

b

e

o

C

b

e

C

n

n

d

B

(

N

m

z

d

o

1

m

L

n

h

1

Q

a

n

A

W

G

2

C

p

V

c

m

s

m

3

H

S

O

d

H

K

M

l

z

3

z

S

O

d

H

7

M

z

1

M

H

2

M

H

3

M

H

5

M

H

7

M

H

1

M

H

2

M

H

3

M

H

5

M

H

7

M

H

l

O

M

H

l

O

2

H

3

H

5

H

7

H

l

O

2

H

3

H

S

O

7

H

I

d

H

i

d

H

•

1

-

1

•

1

-

1

-

1

-

1

-

1

-

1

-

1

-

1

-

1

-

9

-

9

-

9

-

8

-

1

•

1

-

1

-

1

-

1

-

1

-

1

-

1

-

1

-

1

-

1

-

1

-

1

-

1

-

1

X

X

X

X

X

X

X

X

X

O

O

C

p

L

d

=

1

O

m

V

c

m

Z

=

Z

=

1

O

m

-

2

-

2

•

2

-

2

-

2

-

2

.

-

_

9

9 8 8

8

7

7

-

2

-

2

-

2

-

2

-

2

-

2

-

2

-

2

-

2

-

2

-

2

-

1

-

1

-

1

-

1

h

1

m

m

3

3

S

=

S

=

3

S

-

1

-

1

-

1

-

1

1

5

1

1

1

1

1

1

3

1

1

1

2

2

1

1

1

4

l

1

0

2

-

9

7

-

9

4

.

_

_

_

_

_

5

7

-

1

5

1

-

1

4

7

4

3

4

0

3

7

3

1

-

1

2

7

-

1

2

3

-

1

2

0

-

1

7

_

.

1

1

-

0

7

-

1

0

3

-

1

-

1

0

0

~

3

-

1

-

1

-

U

6

-

1

S

1

K

-

2



T

b

e

A

.

3

I

n

v

C

b

e

o

C

b

e

C

n

n

d

B

(

N

m

z

d

o

1

m

L

n

h

1

Q

a

n

A

W

2

C

p

V

c

m

C

p

L

d

=

1

O

m

V

c

m

Z

=

Z

=

1

O

m

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

l

O

2

H

a

a

7

m

l

O

2 3

H

5

H

T

l

d

k

2

d

H

3

k

5

d

H

7

M

H

W

W

H

2

H

S

O

d

z

5

M

H

7

d

H

2

H

3

H

5

d

H

7

d

H

1

M

2

M

3

M

5

M

7

M

l

O

M

l

2

M

3

M

5

M

7

M

.

,

„

_

_

.

.

™ _

4 4 1 3

3

1 1 1 1 1

-

1

-

1

-9 -9 -S -8 -8 -7 -7 -7

-6

-

5 -5 -5 -4

4

-3 -3 -3 -2 -2

-2

-1

-

S

•

1

7

9

1

7

0

1

6

5

M

B

2

1

3

2

0

9

2

0

5

2

0

2

2

4

9

2

4

5

2

4

2

2

3

9

2

3

3

2

2

5

2

2

2

2

3

9

2

1

3

2

0

9

2

0

5

2

0

2

1

3

7

3

3

3

1

3

0

1

2

1

3

1

7

1

1

3

1

1

0

1

0

7

3

0

1

1

2 2

1

1

1

1

0

-

9

-

9

-

9

-

8

-

1

-

1

-

1

-

1

-

1

-

1

-

1

-

1

-

1

-

9

-

9

-

8

•

1

-

-

4

-

1

-

1

-

1

-

1

-

1

-

1

-

1

.

U

-

n

-

1

-

1

-

1

-

9

1

2

7

1

1

5



CABLE-TO-CABLE COUPLING 271

• Problem: Backplane Interconnect Wiring

• Two Parallel Wires

• H = 1 mm S = 1 mm * culprit

=

l

Wire Seperation • 1 = 40 cm • Schottky Logic, T

r

= 3ns ec

t S s l m m »v =3 .5 Volt Swing =11 dBV

Average Height

# N o i s e I m m U n i < y : 3 0 0 m V

above Ground

h = lmm

•Solution (at 106

MHz):

A50x04<A

•

C Coupling = - 1 4 dB + 20 Log

10

( / ,QQ J = - 18dB

•

L Coupling =

13

dB + 20 Log

10

(

1 0

°

1

^ *

4

° )

= ~ 24 dB

•

Larger is C Coupling = -18 dB

• Coupled Voltage: 11 dBV ~ 18 dB = -7 dBV = 450 mV

•

EMI Because 450 mV > 300 mV Noise Immunity Level

•

Discuss Different EMI Fixes

Figure A 3

Illustrative example: backplane wiring.



Index

absorption loss 116

adhesive, conductive 165

adjacent-signal frequencies 36

antenna

radiation characteristics 48

antennas

characteristics 90

pow er density 51

propagation effects 51

avalanche diodes 225

silicon 226

B

backplane wiring 271

balanced circuits 102

bonding jump ers 168

bonds

composites and conductive plastics

166

direct 163

effects of poor 161

finishes 170

guidelines 266

indirect 167

lock washers 164

resistance and impedance 162

soft solder 164

brazing 165

broadband em issions 21

broadband noise 4

C

cable

important parameters 230

shielded 229

termination 2 30, 237

twisted pairs 229

cadweld joints 165

capacitive coupling 74

co-channel interference 41

common-ground impedance 91

common-mode inductors 189

common-mode rejection 215

composite absorption 122

conductor inductance 162

connector backshells 243

connectors

backshells 243

coaxial 251

EMI control summary 254

filter-pin 251

summary of characteristics 252

core saturation effects 189

corrosion control 169

corrosion protection 172

coupling

cable-to-cable 267

cable-to-field 71

calculating 268

capacitive 74, 267


field-to-cable 71

ground-loop 69

inductive 74, 268

coupling mod es 5, 66

crimping-ring termination 248

cross modulation 42

crosstalk 74, 199

crowbar devices 224

7



74


D

daisy chains 247

decibel 23

definition 13

desensitization 42

E

electric fields 113

electrolytic corrosion 170

EMC

assessment 52

gaskets, see gaskets

sealants 155

EMC design

comm unication systems 28

frequency and polarization 50

interconnected equipments 79

intersystem 9

intrasystem 8

system definition 8

system design and development 9

system operation 10

EMF 109

EMI

adjacent-signal 27

basic elements 61

characteristics of concern 76

co-channel 26

common-mode 68

conducted 6, 14, 67

control techniques sum mary 257

coupling mod es 5, 66

differential-mode 68

effects 3

fundamental outputs 64

out-of-band 57

out-of-channel 28

radiated 14

radiation paths 6

sources 4, 5, 63

system-level control 76

transmitter/receiver 26

transmitters 4

units of measure 23

victims 5

EMI control

available fixes 258


conducted paths 263

coupling paths 261

ground loop coupling 264

interconnections 260

printed circuit boards 259

radiation 264

wire and cable crosstalk 265

emissions

broadband 21

coherent broadband 21

frequency vs. wavelength 22

fundamental 35

harmonic levels 39

incoherent broadband 22

narrowband 20

spurious 34

transmitter 34

ESD 20, 109,221

F

Faraday shield 206

ferrite-loaded wire 201

ferrites 199

fiber optics 102

field strength 73

field theory 111

far-field conditions 113

near-field conditions 113

ilt rs

common and differential modes 188

comm on-mode equivalent circuit

191

control of parasitics 192

cutoff frequency 184

EMI/RFI210

for switch-mode power supplies 195

high-frequency performance 192

maximum attenuation values 184

passbands 181

power line 186

selection 194

signal 196

stopbands 181

types of 182

Fourier series 15

Fourier transform 16

free-space propagation 51

frequency



INDEX

75

adjacent-signal 36

fundamental 4

nondesign 50

separation 31

G

galvanic corrosion 169

gaskets

application requirements and con-

straints 151

comparison of types and materials

150

compression set 146

conductive plastic and elastomer 148

joint unevenness 143

knitted-wire mesh 147

mounting 152

oriented immersed-wire 148

pressure-sensitive, foam-backed foil

150

required compression pressure 144

selection 150

spring-finger gaskets 149

theory 142

ground loops 69, 93 , 94, 220

multiple 106

grounding

definitions 82

hybrid 98

in subsystems 95

influence on EMI control 109

multipoint 97

scheme selection 98

single-point 96

system configurations 103

grounding systems

high-frequency behavior 88

impedance 85

impedance characteristics 83

inductance 85

resistance property 84

H

halo rings 247

harmonic amp litude, summ ary 40

harmonic emission levels 39

hybrid transient suppressors 227

I

IF selectivity 42

impedance

common-ground 65, 91

common-source 65

input/output capacitance 214

insertion loss 194

interference

co-channel 41

ground-related 91

receiver adjacent-signal 42

interference margin 30

intermodulation 37, 42, 54

receiver 44

isolators 201

leakage inductance 190, 209

lightning 20, 81, 109, 168,221

lossy ma terials 199

M

magnetic fields 113

magnetic permeability 199

MIL-STD-220A251

MIL-STD-461 29, 39, 46

modulation sidebands 35

multi-pin connectors 245

N

narrowband em issions 20

neutral grounding 208

noise

common- and differential-mode 204

transmitter 52

nondesign frequencies 50

O

optical isolators 102, 211

with Faraday shield 220

P

parasitics 88, 185, 191

control 192

plane waves 114

polarization dependence 50

power density 51

propagation

directional 51

effects 51



276 DESIGNING ELECTRONIC SYSTEMS FOR EMC

omnidirectional 51

unintentional 29

R

radiated field strength 73

radiation

antenna 29, 48, 50

incidental 64

paths 5

printed circuit board 71

unintended 34

radiation field 113

radiation resistance 91

receiver

intermodulation 44

selectivity 3 1 , 43

spurious responses 29, 47

ventilation openings 131

viewing apertures 135

shock and fire hazards 81

shunt capacitance 209

signals

Fourier transform 16

spectral representation 16

time and frequency domains 14

single-point ground 96

spectral amplitude 16

spurious frequencies 4

stray capacitance 162

susceptibility

equipments 6, 74

receiver 40

system life cycle 6

susceptibility 40

reflection 114

reflection loss 117, 122

to electric and mag netic fields 119

to plane waves 118

relative bandwidth 31

CJ

sealants

conductive caulking 156

conductive epoxies 155

conductive grease 157

shielding

audio wire termination 250

compromises 126

effectiveness 123

materials 123

properties of metals 124

termination 23 0, 24 1, 245

wire termination 249

shielding integrity

conductive glass 139

control-shaft apertures 139

EMI gaskets 142

gap dimensions 129

honeycomb 131

indicator buttons and lamps 140

screen mesh 133

seams and j oints 128

T

transformers

capacitive coupling in 205

ultra-isolation 207

transformers, isolation 204

transient suppressors 221

hybrid 2 27

transients 18, 65

sources 19

transmission 114

transmitter

emission characteristics 34

intermodulation 4, 37

modulation envelope 31

noise 36, 52

nonlinearities 4

spurious emissions 29

transmitters 4

twisted wire pairs 239

V

varistors 226

voltage clamping devices 225

voltage standing-wave ratio 252

W

welding 165

Z

zener diodes 226



About the Author

Dr William G Duff

President

SEMTAS Corporation

E d u c a t io n

George W ashington University, B.S. in Electrical Engineering, J un e

1959

Syracuse University, M.S. in Electrical Engineering, Ja nu ar y 1969

Clayton University, D.Sc. in Electrical Engin eering, Au gust 1977

Summary

of

xperience

Dr. Duff is in tern atio nall y recognized as a leader in th e developm ent of

engineering technology

for

achieving electromagnetic compatibility

(EMC) in communication and electronic systems. He has 42 years of

experience in electromagn etic interference/electromagnetic vulnerab il-

ity (EMI/EMV) analysis, test, design, and problem solving for a wide

variety of comm unication an d electronic system s. He has applied EM I/

EMV test, analysis, modeling

and

simulation techniques

to

evaluate

EMC within and between communication and electronic systems oper-

ating in severe electromagnetic environm ents.

Dr. Duff developed and applied an analysis and test methodology for

assessing the electromagnetic susceptibility/vulnerability of communi-

cation-electronic circuits and equipm ents res ulting from both inten-

tional Electronic Counter Measures (e.g. jamming) and u nintent ional

EMI. The assessment involves applying

a

combination

of

analysis

and

tests to evaluate the overall vu lnerab ility of electronic devices to EMI

and electronic warfare (EW), determine the EMI/EW mechan isms and

develop fixes for the identified prob lem s. Dr Duff has applied the

77



278 DESIGNING ELECTRONIC SYSTEMS FOR EMC

methodology (totally or partially) to a nu m ber of poten tially susceptible

devices including:

• variou s legacy military receivers;

• line-of-sight m icrowave relay system with a n adaptiv e electronic

counter-counter measures (ECCM) antenna; and

• troposph eric scatt er comm unication system s with an ECCM modem

• electronically guided missiles and sm art bombs ;

• Black Haw k helicopter flight controls;

• M l tan k control and guidance electronics;

• shipbo ard electronic control equip men t;

• truc k anti-lock bra kes ;

• automobile emission control modu les; and

• med ical electronic devices;

This work included analysis of potential vulnerabilities, testing to

obtain quantitative data on the factors that contribute to EMV, deter-

mining the degradation mechanisms as a result of EMV, and recom-

men ding fixes to reduce the EMV for the item s of inte rest .

Dr. Duff has written more than 40 technical papers and four books

on EMC. He also regularly teaches seminar courses on EMC. He is an

IEEE Fellow, Past President of the IEEE EMC Society, and a NARTE

Certified EMC Engineer.

designing electronic systems for emc - scitech 2011

Documents