A Practical Guide To EMC Engineering (PDFDrive)

A Practical Guide to EMC Engineering
For a complete listing of titles in the

Artech House Electromagnetics Series,
turn to the back of this book.
A Practical Guide to EMC Engineering
Levent Sevgi
Library of Congress Cataloging-in-Publication Data
A catalog record for this book is available from the U.S. Library of Congress.
British Library Cataloguing in Publication Data

A catalogue record for this book is available from the British Library.
Cover design by John Gomes
ISBN 13: 978-1-63081-383-3
© 2017 ARTECH HOUSE

685 Canton Street
Norwood, MA 02062
All rights reserved. Printed and bound in the United States of America. No part of this book
may be reproduced or utilized in any form or by any means, electronic or mechanical, including
photocopying, recording, or by any information storage and retrieval system, without permission
in writing from the publisher.
All terms mentioned in this book that are known to be trademarks or service marks have been
appropriately capitalized. Artech House cannot attest to the accuracy of this information. Use of
a term in this book should not be regarded as affecting the validity of any trademark or service
mark.
10 9 8 7 6 5 4 3 2 1
Contents
Preface xiii
Introduction 1
1.1 Electromagnetic Compatibility 2
1.2 EM Fields in Our Environment 5

1.2.1 Low-Frequency Magnetic Field Coupling 7
1.2.2 Power Absorption from EM Fields 7
1.2.3 Electromagnetic Levels in Our Environment 8
1.2.4 Epilogue 10
1.2.5 Risk Assessment and Precautionary Principle 12
1.2.6 Simple EM Calculations 14
1.3 EU EMC Directives 16
1.4 CE Marking Process 17
1.5 EMC Institutions and EMC Standards 20

1.5.1 Commercial EMC Standards 21
1.5.2 Military EMC Standards 22
1.6 EMC Limiting Values 23
1.7 EMC Tests and Measurements 26
v
vi A Practical Guide to EMC Engineering
1.8 EMC Engineering Philosophy 28
1.9 Suggested EMC Approach 29

References 31
Bibliography 32
2 Accreditation 33
2.1 Introduction 33
2.2 Accreditation 35
2.3 Accreditation Institutions 36
2.4 TÜRKAK 38
2.5 EMC Tests and Measurements 40

2.5.1 Features of EMC Tests and Measurements 40
2.5.2 Calibration 42
2.5.3 Reporting and Product Certification 43
2.6 Proficiency Testing and Interlab Comparisons 44

References 49
Bibliography 50
3 Electromagnetic Model 51
3.1 Basic Electrical Engineering Theories 51
3.2 Maxwell Equations 52
3.3 EM Scattering, Diffraction, and Propagation 53

3.3.1 EM Point and Line Sources 55
3.3.2 EM Wave Polarization 56
3.3.3 EM Plane Waves and Wave Impedance 57
3.3.4 EM Power Density 57
3.3.5 EM Problem Groups 57
3.3.6 EM Propagation Modes 59
3.4 EM Materials and Skin Depth 59
3.5 Electric and Magnetic Dipoles 61
3.6 Typical Emissions 62

Contents vii
3.7 EM Coupling Mechanisms 63

References 64
4 Circuit Model 65
4.1 Lumped Parameter Circuit Elements 65

4.1.1 Conductor Wires 66
4.1.2 Inductive Effect of a Conductor Wire 68
4.1.3 Capacitive Effect of a Conductor Wire 68
4.1.4 Realistic R/L/C Models 69
4.2 Two-Port Circuit Definitions 71
4.3 Resonance Circuits 74
4.4 Cables and Transmission Line Model 77

4.4.1 Characteristic Impedance 78
4.4.2 Propagation Constant 79
4.4.3 Voltage Reflection Coefficient 80
4.4.4 Voltage Standing Wave Ratio (VSWR) 81
4.5 Grounding 83
4.6 Common Mode and Differential Mode Currents 86
4.7 Nonlinearity Effects 87
4.8 Two-Port Circuits and S-Parameters 89
4.9 Microstipline Circuits 95

4.9.1 Characteristics of a Microstripline 95
4.9.2 Basic Microstrip Circuits 98
4.10 Crosstalk 99
References 100
Bibliography 101
5 Antennas and Antenna Calibration 103
5.1 Fundamental Antenna Terms 105
5.2 Communication Antennas 107
5.3 EMC Antennas 108

viii A Practical Guide to EMC Engineering
5.3.1 Receive Antenna and Antenna Factor 109

5.3.2 Transmit Antenna Factor 114
5.4 Antenna Calibration 116
5.5 Normalized Site Attenuation 118

5.5.1 Theoretical NSA Calculations 118
5.5.2 NSA Measurements 123
5.5.3 Performing an Antenna Calibration 125
5.5.4 Antenna Calibration with Pattern Measurements 127
5.6 Loop Antennas 132
5.7 Loop Antenna Calibration 135
5.8 Antenna Arrays 137

5.8.1 Arrays with Isotropic Radiators 138
5.8.2 A MATLAB-Based ARRAY Package 140
5.9 Antenna Types 143

5.9.1 Electric and Magnetic Dipoles 145
5.9.2 Wire Antennas 148
5.9.3 Broadband EMC Antennas 149
5.9.4 Log-Periodic Dipole Antenna 149
5.9.5 Horn Antenna 151
References 152
Bibliography 154
6 Noise and Frequency Analysis 155
6.1 Fundamental Electromagnetic Signals 155
6.2 Noise 156
6.3 Frequency Analysis and Fourier Transforms 159
6.4 Discrete Fourier Transform 163
6.5 Harmonic Generation 171
6.6 AC Mains and Power Quality 172

6.6.1 Distortions in AC Mains 173
6.6.2 Harmonic Distortions and Standards 176
6.6.3 Harmonics Measurements and Evaluation 177
Contents ix
6.6.4 Effects of Harmonic Currents and Voltages 179

References 179
Bibliography 180
7 EMC Test and Measurement Environments 181
7.1 Open Area Test Site 181
7.2 Open Area Test Site Calibration 183
7.3 Screened Room and Resonance Effects 190
7.4 Anechoic Chamber 194
7.5 TEM/GTEM Cell 197
7.6 Reverberation Chamber 198

References 199
Bibliography 200
8 EMC Test and Measurement Devices 201
8.1 Digital Multimeter 201
8.2 Noise-Limited Detection 202
8.3 The Loading Effect 204
8.4 Digital Multimeters and Specs 205
8.5 Measurement Receivers 210

8.5.1 Spectrum Analyzer 210
8.5.2 EMI Receiver 215
8.6 Other Test/Measurement Instruments 218

8.6.1 Current Probes and Current-Injection Clamps 218
8.6.2 Line Impedance Stabilization Network 219
8.7 Network Analyzer 220

References 221
9 EMC Tests and Measurements 223
9.1 Risk Analysis in EMC Tests and Measurements 224

x A Practical Guide to EMC Engineering
9.2 Emission Measurements 226

9.2.1 Detector Types 227
9.2.2 Basic Factors in EMC Tests and Measurements 227
9.2.3 Performing Emission Measurements 228
9.3 Immunity/Susceptibility Tests 232
9.4 Harmonic Measurements 236
9.5 Surge and Flicker Tests 237
9.6 Electrostatic Discharge Tests 238
9.7 Electrical Fast Transients 240
9.8 Measurements of Spurious 240
9.9 Error Analysis and Uncertainty 240

References 244
Bibliography 246
10 EMC and Protection 247
10.1 Filtering: Butterworth Filters 247

10.1.1 Butterworth LC Filters 248
10.1.2 Transmission Line Filters 254
10.1.3 Microstripline (ML) Transformation 256
10.1.4 MATLAB-Based Filter Package 257
10.2 Ferrites 260
10.3 Screening/Shielding 262

10.3.1 Shielding Effectiveness—Ideal Case 264
10.3.2 Shielding in Real Life 269
10.3.3 Practical Protection Tips 272
10.4 Cables and Connectors 272

10.4.1 Cables 274
10.4.2 Connectors 276
10.5 Switched-Mode Power Supplies 277
10.6 Practical Design Tips 277
10.7 Impedance Matching 280

Contents xi
10.7.1 Lumped Element Circuits 281

10.7.2 Scattered Parameter Circuits 282
10.7.3 Impedance Matching on the Smith Chart 283
References 286
Bibliography 287
Bibliography 289
Suggested Readings 289
Suggested Books 290
Suggested Sites 291
About the Author 293
Index 295
Preface
Every book is an accumulation of knowledge and/or experience.
The material for this book has been accumulated since 1996. At that
time, concern about cell phones, base stations, and public health was continu-
ously increasing at the same time as legislation of the necessity of Conformity
Europe (CE) marking (i.e., the Electromagnetic Compatibility Directive) on
every product entering the European Union market was taking effect. Health
concerns fall under bioelectromagnetics (BEM), while legislation is the subject
of electromagnetic compatibility (EMC) engineering. Both topics are equally
important. They cover a large number of subjects from public understanding of
science to problems of technological developments and from market control to
engineering education. In other words, they cover life itself.
During these years, while we were quietly studying in our labs or sitting in
front of our PCs, our phones started ringing. We were continuously bombarded
with questions from government offices, municipalities, chambers, and other
institutes and organizations as well as citizens, on EM pollution and possible
adverse health effects of base stations and cell phones. We did our best to an-
swer these (technical and nontechnical) questions. As time went on, our single
initial responses had turned into public seminars, panels, and radio and televi-
sion talks. We received a wide range of speaking invitations, from ministries
of health, environment, and industry to GSM companies, from military acad-
emies to medical faculties, and from schools to local associations. We wrote in
newspapers, magazines, and scientific journals. The aim was to provide people
with the right information.
In the same time slice, in September 1997, an older colleague of mine
knocked on my door and told me that in the next month, the Turkish Elec-
tronics Industry Association (TEIA) would organize two EMC training courses
at Istanbul Technical University. He said that nearly 60 attendees had already
xiii
xiv A Practical Guide to EMC Engineering
registered in each course; unfortunately, TEIA could not find a trainer for this.
I was going to ask “what is this EMC?” when he said that he gave my name as
the trainer because he thought I could teach these courses. You can imagine that
the next few weeks before the courses were a nightmare for me! Nevertheless, I
did my best and prepared myself and created course notes on time. Both courses
were highly successful and influenced many companies in Turkish industry.
Since then, I have repeated these two- to three-day EMC training courses at
many prestigious Turkish companies including Aselsan, Arçelik, Beko, Alcatel,
and Vestel, covering defense to information technology equipment (ITE) and
white goods producers. These courses enabled me to meet with engineers and
groups of people in different sectors from design to tests and measurements,
from planning to marketing. I have also organized these courses in universities
and research institutes open to personnel of small and midsize companies. There
were design/system/test engineers, quality control managers, directors, owners,
and even lawyers in these courses. There were also technical and nontechnical
representatives of ministries and municipalities, accreditation, and certification
authorities, who are responsible for market control. I have also learned much
more than I taught about the problems and challenges of different groups and
sectors. I have continuously updated and enriched my course notes. I published
three EMC-related books in Turkish.
This is what I tell the attendees at the beginning of these courses: “I know
you’re attending this course because you all face serious EMC problems. First,
I must say that that I’m not going to solve your problems; you’re going to do
that! Second, I’m not going to teach you something totally new. What I’ll do in
the following two to three days is mostly to fill in the gaps in your knowledge
and experience. I’m going to revisit everything you learned in your university
education and teach you how to look at them with an eye of an EMC engineer.
I’m quite sure that at the end of the course you will learn how to diagnose and
solve your problem; very often, you’ll suddenly realize where the problem is
during the course. That would be the success of the course.”
The title of this book is an interesting phrase that characterizes the worth
of EMC engineering. This book is unique in terms of its content. It is not a
specific EMC design book; it neither aims to be nor claims to be this! There are
several books on EMC (see the list of suggested readings) that aim to discuss
specific topics such as signal integrity, grounding, shielding, anechoic cham-
bers, and so on. The book covers essentials in EMC engineering that necessi-
tates a different look into many EM concepts and issues. As the title implies, it
contains all the procedures from design to market, including technical and non-
technical issues, which mostly have not been completely covered by any other
EMC books, such as market control, accreditation, calibration, EMC tests and
measurements, and EMC protection. There are many sections and subsections
as well as practical tips related to design. In addition, there are nearly two dozen
Preface xv
MATLAB scripts useful for EMC engineers. The purpose of giving these MAT-
LAB scripts is not to attack real-life EMC problems, but to establish the first
(most valuable) step for the beginners.
There are 10 chapters in the book. The logic of the book is as follows:
First, fundamental definitions, terms, and concepts, including processes in
achieving EMC compliance through testing and certification, are given in
Chapter 1. Accreditation is summarized in Chapter 2. Both EM and circuit
theories are revisited through the eyes of an EMC engineer in Chapters 3 and 4.
Chapter 5 is reserved for antenna and antenna calibration. Noise and frequency
analysis are given in Chapter 6. Chapter 7 focuses on EMC tests and measure-
ment environments and their calibration. Chapter 8 belongs to basic EMC
tests and measurement equipment. EMC tests and measurements methods and
implementation are presented in Chapter 9. Finally, major protection methods,
such as filtering, shielding/screening, grounding and cabling, and impedance
matching, are included in Chapter 10.
Note that modeling and numerical simulation in EMC engineering are
important topics that should be addressed in a separate book. There are many
sources on the Internet on these topics. Many commercial EM packages (e.g.,
CST, FEKO, COMSOL, EMA, REMCOM, ANSYS, SOLIDWORKS) have
continuously been modified to handle more and more complex EMC/EMI
problems. For example, several EMC/electromagnetic interference (EMI) mod-
eling and numerical simulation examples, from EM and lightning protection
to shielding, aircraft chambers to cabling for data and power lines, compos-
ite materials to EMC in radar systems, ESD to system-in-chip simulations,
which have been prepared using the commercial computer simulation technol-
ogy (CST) packages, may be found at www.cst.com/Applications/Category/
EMC. Another example is the site jointly sponsored by the IEEE EMC-S TC9
and ACES, www.ewh.ieee.org/cmte/tc9, where excellent practical examples of
EMC/EMI modeling, simulation, and validation with a focus on canonical,
benchmark, and standard problems are presented.
I have spent nearly three decades on EM and studied on topics such as
propagation through complex environments, radars and integrated surveil-
lance systems, EM scattering and diffraction, radar cross section (RCS) pre-
diction and reduction, EMC-BEM, antennas, EM modeling and simulation,
microstrip circuits, and so on. These topics certainly necessitate establishing
an intelligent balance between strong mathematical background (theory), en-
gineering experience (practice), and modeling and numerical computations
(simulation). Generally speaking, I know that EMC Society (EMC-S) people
find books written by Antennas and Propagation Society (AP-S) authors too
academic (theoretical), and that books written by EMC-S authors are found by
AP-S members to be too industrial (practical). Being a member of both IEEE
AP-S and EMC-S, I aimed, in this book, to balance theory and practice. As can
xvi A Practical Guide to EMC Engineering
be seen from the topics listed in the table of contents, they are the essentials of
EMC engineering that are mostly not covered by other published EMC books.
I believe this book will be a reference book for broad range of EMC-re-
lated technical and nontechnical people and can be used as a textbook in many
introduction and advanced-level EMC lectures. You want to have this in your
library! I hope you enjoy reading it and find it useful.
1
Introduction
The role of electromagnetic (EM) fields in our lives has been increasing [1–6].
Communication, remote sensing, integrated command/control/surveillance
systems, medicine, environment, education, marketing, and defense are only
a few areas where EM fields have critical importance. In 2015, the number of
first-time cell phone subscribers (7.5 billion) exceeded the world population
(7.4 billion) and the number of Internet users exceeded 3 billion. We are sur-
rounded by EM fields.
Figure 1.1 shows a typical recorded spectrum from DC to 1 GHz, show-
ing EM signals in our environment. These signals and levels can be recorded
anywhere on the Earth with quite similar spectra. Today, the spectrum is ex-
panded to 6 GHz.
All of these necessitate the design and production of less interfering, less
susceptible electronic devices on one hand; on the other hand, they create en-
gineering problems in terms of EM fields—human interactions. Two new en-
gineering disciplines were born: electromagnetic compatibility (EMC) and bio-
electromagnetics (BEM). EMC and BEM engineering, in general, deal with
identification of, respectively, adverse device-device and device-human inter-
actions (usually unintentional) and understanding the mechanisms occurring
there as well as their mitigation. Conventional EM engineers are concerned
with a range of product specific issues, whereas EMC engineers have to deal
with all possible external EM influences on the environment. They are both
multidisciplinary and cover electrical, electronics, communication, system en-
gineering and chemistry, physics, medicine, and biology and both have design,
test, production, quality, marketing, and legal implications.
At the same time, governments have tightened controls for the products
bought and sold in their markets. Military and aerospace systems must also
1
2 A Practical Guide to EMC Engineering
Figure 1.1 EM signals in our environment.
comply with EMC requirements. International institutions force producers, ex-

porters, importers, and sellers to trade products produced according to related
international standards. For example, the EU introduced a new EMC directive
(EMC Directive 89/336/EEC) for electrical and electronics products entering
the EU market in 1989 [7] and made CE marking mandatory. The Conformity
Europe (CE) mark is a kind of product passport that allows products to be
available freely in the EU market. There are also similar processes in the United
States and Asia-Pacific regions. For example, international standards from the
Federal Communications Commission (FCC) rules in the United States to var-
ious IEC/EN are specified for emission and immunity testing to verify EMC
compliance. On the other hand, MIL-STD-461 is applied to the U.S. Depart-
ment of Defense (DoD) procurements for equipment and subsystems.
1.1 Electromagnetic Compatibility

The International Electomechanical Committee (IEC) defines EMC as “the
ability of a device, equipment, or a system to function satisfactorily in its EM
environment without introducing intolerable EM disturbance to anything in
that environment.”
An EMC problem consists of three items: EMI source, EMI victim, and
coupling between source and victim. A special case within the EMC problem
is that the victim is a living organism. This is called BEM. EMC engineering
targets to get rid of the source, strengthen the victim, and remove the coupling.
Introduction 3
BEM engineering targets the understanding of short-/medium-/long-term ef-

fects of EM signals on living organisms and to reduce them substantially. A
typical EMC scenario is pictured in Figure 1.2, where a computer may disturb a
TV viewer. Here, the computer is the source of EMC and the TV is the victim.
Computer-generated disturbing EM signals may reach the TV via atmo-
sphere (radiated emissions) or via cables and connectors (conducted emissions).
The victim may be susceptible or immune. Radiated emissions may cause radi-
ated susceptibility. Conducted emissions may cause conducted susceptibility.
Typical protection approaches are pictured in Figure 1.3. Roughly speaking,
the source and victim may be separated (the distance between them may be
increased or a kind of EM screen may be used), better cabling and grounding may
be designed, EM disturbing signals may be filtered out, and finally, impedance
matching may be applied.
EMC deals with device-device interaction. It is concerned with the unin-
tentional generation, propagation, and reception of EM energy (emission) that
may cause unwanted effects such as EMI or even physical damage in operational
equipment. The goal of EMC is the correct operation of different equipment in
a common EM environment. Emission is the generation of EM energy, whether
deliberate or accidental, by some source and its release into the environment.
EMC studies unwanted emissions and countermeasures that may be taken in
order to reduce unwanted emissions. Susceptibility is the tendency of electrical
equipment (victim) to malfunction or break down in the presence of unwanted
emissions. Immunity is the opposite of susceptibility; it is the ability of equip-
ment to function correctly in the presence of interference, with the discipline
of hardening equipment being known equally as susceptibility or immunity.
Coupling is the mechanism by which emitted interference reaches the victim.
Figure 1.2 A typical EMC scenario.

Figure 1.3 Potential solutions in the typical EMC environment.
EMI mitigation and hence EMC may be achieved by addressing any or

all of these issues (i.e., quieting the sources of interference, inhibiting coupling
paths, or hardening the potential victims). In practice, many of the engineering
techniques used, such as grounding, filtering, and shielding, apply to all three
issues.
BEM deals with device-human interaction. It is a study of (short- and
long-term) interaction between EM fields and biological entities. Areas of study
include electrical or EM fields produced by living cells, tissues, or organisms.
EMC engineering is complex but not complicated. The whole process
is controlled by international institutions and standards. The standards give
limit values for both emissions and immunity. They also contain EMC test and
measurement methods, environments, devices, and even setups. Emissions of
a product is measured and compared with the given limiting value in the stan-
dards. The product’s immunity is tested according to the related standards. If
the product passes these tests, then it enters the market.
BEM engineering, on the other hand, is totally different. EM field–hu-
man tissue interaction is complex and complicated phenomena. It is extremely
difficult to develop reliable (mathematical or not) models. Although labora-
tory studies and computer simulations have long been used in modeling and
explaining EM field–tissue interaction, it seems, unfortunately, biostatistics and
epidemiology is still the most reliable study. Biostatistics is the study of statistics
as applied to biological areas. Epidemiology, either experimental or observa-
tional, is the study of disease in human populations.
Introduction 5
1.2 EM Fields in Our Environment

EM fields in our environment fall in nonionizing radiation group. Nonion-
izing radiation is described as a series of energy waves composed of oscillating
electric and magnetic fields traveling at the speed of light. Nonionizing radia-
tion includes the spectrum of ultraviolet, visible light, infrared, microwave, RF,
and extremely low frequency. Nonionizing radiation is found in a wide range
of occupational settings and can pose a considerable health risk to potentially
exposed workers if not properly controlled.
The EM spectrum may be divided into these regions:
• 3 Hz–3 kHz range: Extremely low frequencies. (AC mains is the major
EMC problem in this region; high voltage power lines are the major
concern in BEM engineering.)
• 3 kHz–30 kHz range: Very low frequency (VLF) region.
• From a few hundred kHz up to 1012 Hz: Radiowaves.
• Around 1012 Hz: Infrared region. Everything radiates in this region.
For example, nearly 60% of our body heat is radiated out as infrared
energy. Infrared sensors (i.e., thermal cameras) are used in night vision
(detection in the dark).
• Visible light around 1014 Hz.
The EM radiation part of the nonionizing radiation covers the frequency

range from DC to 300 GHz.
• Extremely low frequency radiation (ELF): ELF radiation at 50/60 Hz

is produced by power lines, electrical wiring, and electrical equipment.
Common sources of intense exposure include ELF induction furnaces
and high-voltage power lines.
• Radio frequency (RF) and microwave (MW) radiation: MW is absorbed
near the skin, while RF radiation may be absorbed throughout the body.
At high enough intensities, both will damage tissue through heating.
Sources of RF and MW radiation include radio emitters and cell phones.
• Infrared radiation (IR): The skin and eyes absorb IR as heat. Work-
ers normally notice excessive exposure through heat sensation and pain.
Sources of IR radiation include furnaces, heat lamps, and IR lasers.
• Visible light radiation: The different visible frequencies of the EM spec-
trum are seen by our eyes as different colors. Excessive visible radiation
can damage the eyes and skin.
• Ultraviolet radiation (UV): UV has a high photon energy range and is

particularly hazardous because there are usually no immediate symp-
toms of excessive exposure. Sources of UV radiation include the sun,
black lights, welding arcs, and UV lasers.
• Laser hazards: Lasers typically emit optical (UV, visible light, IR) radia-
tions and are primarily an eye and skin hazard.
Electric and magnetic fields are defined as force lines surrounding elec-
trical devices. The unit of electric fields caused by voltage difference is [V/m]
and [A/m] for the current generated magnetic field. Electric fields are caused
by electric charges; magnetic fields are caused by moving charges (i.e., electric
currents).
The electric field is the force act on a charge. The magnetic field is the
force act on moving charges.
Electric and magnetic fields are vector quantities; they have both ampli-
tude and direction. One way of representing magnetic fields is magnetic flux
density (B); its unit is tesla [T] or gauss [G] (1T = 10000G). Magnetic fields
and magnetic flux density are related with the magnetic permeability (µ) of the
medium (B = µH). In free-space, in air and in living tissues it is µ = µ0 = 4π10–7
[Henry/m].
In general, DC magnetic fields do not induce body currents on humans.
But weak currents that can affect hearing devices and/or pacemakers may be in-
duced in industrial areas under very high DC magnetic fields. AC mains supply
electric power everywhere at 50/60 Hz frequency all around the world. Time
-varying EM fields cause EM energy absorption and in-body currents. The rela-
tion between the flowing current and electric field in a tissue is given by J=σE
[A/m2] where σ is the tissue conductivity [S/m].
We are surrounded in our environment by static (DC) and dynamic (AC)
electric and magnetic fields. Earth itself creates fields. Some typical values are:
• In air, under normal conditions, DC electric fields are around 120–150
V/m.
• On rainy days with severe thunderstorms and lightning, DC electric
field values may reach up to 10 kV/m.
• DC geomagnetic flux of Earth is around 0.5G; its AC value is around
10-5G.
• In our bodies, the DC magnetic flux is around heart 10-7G, in our ab-
domen 10-6G, and in our lungs 10-5G.
Introduction 7
• An electric shaver produces up to 25 mG magnetic flux. This value for

the fluorescent lamp is nearly 5–10 mG, and for TVs and laptops it is
around 1–5 mG.
As seen, Earth’s DC geomagnetic flux is very high (around 500 mG)

as compared to TVs (1–5 mG), but DC magnetic fields do not induce body
currents.
Time-varying EM fields cause charge movement (current flow) and charge
polarization (dipole formation) and dipole displacement in human tissues.
Their amplitudes depend on electrical parameters of the tissues. The conduc-
tivity (σ) and permittivity (ε) specify the current and polarization amplitudes,
respectively. The conductivity and permittivity depend on the type of the tissue
as well as the frequency.
1.2.1 Low-Frequency Magnetic Field Coupling

Time-varying magnetic fields produce body currents and body electric fields.
The induced electric field depends on magnetic field source (loop antenna di-
ameter), tissue conductivity, and time rate (frequency) of the signal. Basic pa-
rameters and quantities are:
• Up to 10 MHz frequency: J, surface current density [A/m2];

• Up to 110 MHz frequency: I, current [A];
• Between 100 kHz and 10 GHz: specific absorption rate (SAR) [W/kg],
• Between 300 MHz and 10 GHz: SA (for pulsed signals) specific absorp-
tion [W],
• Between 10 GHz and 300 GHz: power density [W/m2].
1.2.2 Power Absorption from EM Fields

EM fields penetrate human tissues, are absorbed, and increase body tempera-
ture. EM power absorption from electric and magnetic fields is negligible at
low frequencies and therefore do not cause any body temperature increase. The
situation changes at frequencies above 100 kHz, and EM power absorbed be-
comes significant. In general, EM absorbed power varies from tissue to tissue.
Four different frequency regions are worthwhile in terms of EM field–tissue
interaction [7]:
• 100 kHz–20 MHz range: EM power absorption is mostly absorbed in

neck and legs.
• 20 MHz–300 MHz range: Significant EM power is absorbed in the

whole body but may concentrate in regions because of the resonance
effects.
• 300 MHz–300 GHz range: Different scale local EM power absorption
may be observed.
• Above 10 GHz: EM absorbed power concentrates mostly on body sur-
face.
1.2.3 Electromagnetic Levels in Our Environment

The nongovernmental organization accepted as an independent expert body is
the International Committee on Non-Ionising Radiation Protection (ICNIRP)
[8]. It is recognized by the UN, WHO, ILO, EU, and most of the developed
countries. ICNIRP is established for the purpose of advancing nonionizing ra-
diation (NIR) protection for the benefit of people and the environment and in
particular to provide guidance and recommendation on protection from NIR
exposure. In 1998, the ICNIRP published its major guidelines for exposure
limits: General reference levels (E-field strength) at 900 MHz and 1800 MHz
are given as 42 V/m and 59 V/m, respectively.
There are two different limiting values in the ICNIRP documents [8]:
• Basic restrictions: These are related to time-varying electric and mag-

netic fields that directly cause adverse health effects and are established
from short-term behavioral disruption. Depending on the frequency
they can be:
• Surface current density (J) [A/m2];
• Specific absorption rate (SAR) [W/kg];
• EM power density (S) [W/m2].
Among these three, only power density in an environment (outside the
body) can be measured easily. The other two can only be obtained indi-
rectly; therefore, reference levels are used instead.
• Reference (derived) levels: These are the levels that are used to check
whether or not basic restrictions are satisfied. They are established in
terms of electric and magnetic fields as well as power density. For base
station measurements, it is the far field region so plane wave approxima-
tion can be made. In this case, measuring only one of three is enough to
obtain the others from H=E/377 and S=E2/377. As given in Figure 1.4,
ICNIRP reference levels are 42 V/m at 900 MHz; 59 V/m at 1800 MHz
frequency. Many countries, such as the United States, EU countries, and
Turkey, accept the ICNIRP levels. Countries that apply the precaution-
Introduction 9
Figure 1.4 Threshold E-field values of the ICNIRP.
ary principle lower their reference levels. For example, Italy uses 20 V/m
(at 900 MHz) value as a general level and further reduces it to 6 V/m
around schools, hospitals, and so on. Russia has 6 V/m and Sweden has
4 V/m at 900 MHz as reference levels.
Figure 1.4 shows ICNIRP reference levels versus frequency. These are ob-
tained as follows [8]:
• Available experimental evidence indicates that the exposure of resting

humans for approximately 30 minutes to EMF producing a whole-body
SAR of between 1 and 4 W/kg results in a body temperature increase of
less than 1°C. This value is assumed critical for the thermal effects.
• Exposure to more intense fields, producing SAR values in excess of 4
W/kg can overwhelm the thermoregulatory capacity of the body and
produce harmful levels of tissue heating.
• Occupational and general public exposure limits are then determined as
0.04 W/kg and 0.08 W/kg, using 10 and 50 as the protection factors,
respectively.
The top and bottom curves in the figure correspond to occupational and
public exposures, respectively. The general public exposure limit is 300 V/m
from DC to 1 MHz, but it is 42 V/m at 900 MHz cell phone frequency band.
Note that the ICNIRP guides do not specify EM emission levels of the
devices we use in our daily lives. They are derived under normal conditions for
average humans [8]. They are given according to available scientific findings,
are dynamic, and are revised as more and more scientific evidence appears.
1.2.4 Epilogue
EM fields and public concern is a long-term, sensitive issue [9]. EM pollution
discussion has been raised in the public all around the world parallel to expo-
nential increase in cellular phone users. Because of the lack of information and
long-term experimental and epidemiological studies, contradictory opinions
exist on EM exposure safety levels in societies. Different governmental offices,
universities, media, and nonprofit organizations may have different approaches
to this problem. Since available scientific knowledge is far from being able to
end these discussions, this is normal. Beliefs may be totally different form one
group to another. It is the obligation of the scientists not to confuse beliefs with
scientific studies.
The debate on possible adverse effects has gained increasing attention.
Therefore, one needs to be careful even when choosing the right words. For
example, effect does not necessarily mean harm. Saying “no adverse effect has
been observed” does not mean “there is no adverse effect.” In order to say “there
is no . . . ” scientifically, all possible situations must be checked, which is almost
impossible in EM wave–human interaction.
The discussion about SAR (amount of EM power in watts/kg absorbed
by human tissues) is almost over. We now know that regulations based on SAR
are used in EM field measurements around base stations, and measured values
are well below specified threshold levels. The discussion has focused recently on
the question “What if very weak EM fields, well below the thresholds specified
in the regulations, trigger adverse effects in the long term (e.g., 10 years, 20
years)?” The answer will come only after long-term, accurate epidemiological
studies.
In May 2010, the results of the Interphone study [8] coordinated by
IARC were published. It was a 10-year multinational, epidemiologic study that
included nearly 13,000 human subjects from 13 countries around the world.
The study concerned the attempt to determine whether mobile phones increase
the risk of developing cancer involving the tissues around the ears and mouth in
humans. Final results showed no association between cell phone use (less than
10 years) and cancer.
In May 2011, the World Health Organization (WHO) [3] and Interna-
tional Agency for Research on Cancer (IARC) classified mobile phone use and
other RF EM fields as possible carcinogens (group 2B). Note that, IARC clas-
sifies substances to show whether or not they are suspected to cause cancer. It
places them into one of five categories depending on the strength of the evidence
for their carcinogenicity. These categories are Group 1 (carcinogenic to humans
Introduction 11
[e.g., smoking, processed meat, solar radiation, alcoholic beverages]), Group

2A (probably carcinogenic to humans [e.g., red meat, exposures working in
hairdressing]), Group 2B (possible carcinogenic to humans [e.g., coffee, pickled
vegetables, gasoline, welding fumes)], Group 3 (carcinogenicity not classifiable
[e.g., tea, static magnetic fields, polyethene]), and Group 4 (probably not car-
cinogenic to humans).
In October 2011, a Danish study [8] published in the Journal of the Na-
tional Cancer Institute showed that there has not been a rise in brain tumors
that correlates with widespread cell phone use. The study looked at data from
60,000 people who had brain tumors—glioma and meningioma—in Denmark,
Finland, Norway, and Sweden between 1974 and 2003. This, a 30-year exami-
nation in the incidents of brain tumors, was undertaken among 20- to 79-year-
olds by the Institute of Cancer Epidemiology in Copenhagen, Denmark. The
study showed no change in the brain tumor rate. Note that although this was a
30-year study, cell phone use has escalated between 1998 and 2003; therefore,
the study should be categorized as short term (five-year) cell phone use.
In May 2016, NTP of the US found a likely link between high amounts
of cell phone radiation exposure and cancer development in male lab rats.
While older reports have shown that cell phones aren’t noticeably harmful for
humans, it’s still a potentially concerning development. On the other hand, in
the same month, a huge Australian study showed that after 30-years of use, cell
phones haven’t increased brain cancer rates [10].
The focus of the discussion presented in [11] is on the increasing concern
that most current published research findings using epidemiology and biostatis-
tics are false. Similar discussion is given in [12]. A major reason for these is that
doing research on EM wave–human tissue interaction is extremely challenging.
Note that doing research on EM wave–human tissue interaction is ex-
tremely challenging; therefore, reaching a scientific conclusion necessitates de-
cades of hard work and investigation. Moreover, results have to be presented
statistically, which, at the end, implies that the debate will continue (statistics
may only give a feeling on results).
Beside the complexity of the problem, the general public is extremely
sensitive, so experts know they have to choose every word of their statements
carefully. WHO published a handbook [13] back in 2002 just to emphasize the
importance of the dialogue, among not only experts and politicians, but also
members of the whole society. As stated in this handbook, three aspects of the
problem are (a) the dialogue, (b) the scientific evidence, and (c) risk perception
and management. WHO [7] states that providing an accurate translation of
scientific information and terminology for the media, policy makers, and the
general public is a challenge.
Just as important is the scientific content of the research in this field. For
example, the U.S. National Toxicology Program (NTP) requests a minimum of
400 animals per study (50 animals per sex per group, with three dose groups
and a control) and at least two years of chronic study with complete necropsy
and histopathology [7]. In order to study nonthermal effects of cell phone radi-
ation, NTP requires (a) an ability to expose large numbers of rodents to a high
dose; (b) animals to be unconstrained and housed in standard laboratory cages;
(c) food and water to be available on demand; (d) excellent field and specific
absorption rate (SAR) homogeneity and detailed numerical and experimental
dosimetry; (e) the ability to discern a possible dose response; and (f ) third party
verification of the correct operation of the system. A study must satisfy these
criteria first (NTP spent multimillion US dollars for the establishment of a
RF radiation reverberation chamber exposure system for rodents). Note that, a
waste number of in vitro studies were conducted with at most 20–40 animals.
This is why WHO approaches their results with caution.
1.2.5 Risk Assessment and Precautionary Principle

Risk, as commonly defined, is the effect of uncertainty on objectives. Risk man-
agement is therefore a matter of the identification, assessment, and prioritiza-
tion of risks on EM wave–human tissue interaction.
Risk assessment is the process whereby the potential adverse consequences
(hazards) associated with a technology or development are identified, and the
probability (risk) of their occurrence is estimated. Risk management is the pro-
cess by which the risks and benefits associated with a technology or develop-
ment are weighed against each other and decisions are made on whether and
how to proceed with its implementation. A common approach in risk manage-
ment is to identify a critical adverse health effect (usually that which occurs at
the lowest level of exposure). The lowest exposure at which this effect has been
shown to occur is then multiplied by an assessment factor, also known as a safety
or uncertainty factor, to derive an exposure limit or guideline.
At the international level, the precautionary principle was first recognized
in the World Charter for Nature, adopted by the UN General Assembly in
1982. It was subsequently incorporated into various international conventions
on the protection of the environment. This principle was enshrined at the 1992
Rio Conference on the Environment and Development, during which the Rio
Declaration was adopted. Principle 15 states that “in order to protect the envi-
ronment, the precautionary approach shall be widely applied by States accord-
ing to their capability. Where there are threats of serious or irreversible damage,
lack of full scientific certainty shall not be used as a reason for postponing cost-
effective measures to prevent environmental degradation.”
The precautionary principle is essentially used by decision makers in risk
management of risk. The measures based on this principle should be:
Introduction 13
• Proportional to the chosen level of protection;

• Nondiscriminatory in their application;
• Consistent with similar measures already taken;
• Based on an examination of the potential benefits and costs of action or
lack of action (including, where appropriate and feasible, an economic
cost/benefit analysis);
• Subject to review, in the light of new scientific data;
• Capable of assigning responsibility for producing the scientific evidence
necessary for a more comprehensive risk assessment.
The policy by which a precautionary approach is applied to risk manage-

ment in situations of scientific uncertainty has been termed the precautionary
principle. This principle was formally adopted by countries of the EU in the
Treaty of Maastricht (1992) and is evident in a ruling of the European Court of
Justice when it upheld the decision of the European Commission to ban beef
from the UK with a view to limiting the risk of transmission of bovine spongi-
form encephalopathy (BSE), commonly known as mad cow disease. The Court
concluded as follows:
In view of the seriousness of the risk and the urgency of the situation, and
having regard to the objective of the decision, the Commission did not
act in a manifestly inappropriate manner by adopting the decision, on a
temporary basis and pending the production of more detailed scientific
information. Where there is uncertainty as to the existence or extent of
risks to human health, the Commission may take protective measures
without having to wait until the reality or seriousness of those risks
becomes apparent.
The debate on EM waves–human interaction seems destined to continue

for a long time. In the meantime, there are several things that people can do
to lower their exposure. For example, they can get low-SAR devices, use the
speaker mode on the phone or a hands-free device, text instead of talking, for-
bid its usage by children and electromagnetically sensitive people, and, more
importantly, limit the time of use. Dose is one of the key parameters here. It
has been shown that continuous use of a cell phone for 5 min increases head
temperature around 0.1°C; 10 min use, more than 0.25°C; and 20 min use,
more than 2.0°C.
It seems that maximum continuous use of 5 min is reasonable in terms
of precautionary principle. The relation between dose and effect is not linear!
1.2.6 Simple EM Calculations

EM field measurements are performed by checking whether or not EM limit-
ing values are exceeded. Simple EM field calculations can also be done in order
to get the feeling of the amplitude of the field values. The parameters listed in
Table 1.1 are used for this purpose.
The power density S, at a distance R from a transmitter with transmit
power Pt is:
Pt
S=  W m2  (1.1)
4 πR 2  
The transmit antenna is usually directive with antenna gain Gt; therefore,
power density along the maximum radiation direction is:
Pt Gt
S=  W m2  (1.2)
4 πR 2  
The power received (Pr) by a receive antenna, having effective antenna

aperture {Ae} at a distance R is:
Pt Gt λ2
Pr = Ae [W ] Ae = Gr m 2  (1.3)
4 πR 2 4π
In other words, the received power can be calculated from
2
Pt Gt Gr λ2  λ 
Pr = = Pt Gt Gr  [W ] (1.4)
2
4 πR 4 π  4 πR 
Table 1.1
Basic Power Calculations in Communication Systems
Quantity Explanation Unit
Pt Transmit power Watt
Pr Receive power Watt
S Power density W/m2
Gt, Gr Antenna gains
R Transmit-receiver distance Meter
λ Wavelength Meter
f EM signal frequency Hz
Introduction 15
The power density at the receive point may also be found using
E2
S=  W m2  (1.5)
120 π  
if the electric field at the receive point is known. Combining (1.1)–(1.5) yields
30Pt Gt
E=
R
[ V m] (1.6)
Question 1: Calculate the power density at a distance of 100m from a base

station transmitter located on top of a 20m-high tower, having 10W transmit
power and 18 dBi transmit antenna gain.
Answer: The distance R in this case is R = 102m (R = h 2 + d 2 ). The an-
tenna gain is 10(18/10) = 63. The power density from (1.2) will then be 4.8 mW/
m2. The power density is usually expressed in terms of µW/cm2; therefore, one
only needs to divide this value by 10 in order to change mW/m2 to µW/cm2 (1
mW = 103 µW ve 1m2 = 104 cm2). As a result, 4.8 mW/m2 corresponds to 0.48
µW/cm2. The electric field value from (1.3) will then be 1.3 V/m.
Question 2: What is the transmit gain of a 1W transmitter if the power
density at 1 km distance is measured to be 1.6 µW/m2?
Answer: The power density of an isotropic transmitter may be calculated
using (1.1) as 80 nW/m2. Therefore, the transmit antenna gain will be 1.6 µW/
m2 / 80 nW/m2 = 20 (13 dB).
Figure 1.5 shows relation between time of exposure and current passing
through human body at AC mains frequencies (i.e., for 10–100 Hz AC volt-
age). In general, 30 mA is the threshold value. Current with this value passing
from a human body for a few seconds is found to be dangerous. The time of
exposure reduces to ms as the current increases.
Table 1.2 lists worst-case magnetic field emissions of typical household
appliances. Note that electric and magnetic field thresholds at low (mains) fre-
quencies are around 300 V/m and 1G, respectively. Note also that extensive
amount of house measurements show that typical magnetic field values are
around 0.5 mG–5 mG.
An important EMC problem in a house is the electricity wiring during
the construction. Very often, this is done by contractors who aren’t aware of
EMC philosophy; therefore, there may be significant levels of permanent low-
frequency magnetic fields. This is even more important in today’s intelligent
houses with wiring in every corner of every room. Remember that a circular
loop formed during the wiring results in permanent magnetic field source.
Figure 1.5 Time of exposure verssus body currents [14].
Table 1.2
Typical Magnetic Fields Produced by Household Devices [mG]
Equipment d = 10 cm d = 30 cm d > 1m
Vacuum cleener 300–400 30–50 3–5
Iron 5–10 1 0.5
Washing machine 20–30 3–5 1
Hair dryer 400 10 1
Electrical shaver 200 5 0.5
Mixer 70–200 5–10 0.5
Dishing machine 20–40 5–10 0.5-1
Electric oven 5–10 0.5 0.5
Freezer 5–10 1 0.5
Music Set 3–5 1 0.5
TV 3–5 1 0.5
Digital watch 2–3 1 0.5
Air Conditioner 100 3–5 0.5
Photocopy machine 80–150 10–30 1–3
1.3 EU EMC Directives

EU EMC directive 89/336/EEC for the inspection of products entering the
EU market was first introduced in 1989 and fully came into effect in January 1,
Introduction 17
1992, after a three-year transition period. It delivers all necessary and sufficient
conditions that must be satisfied by the products. After years of feedback from
the market, its revised version, 2004/108/EC, was introduced. This new direc-
tive was based on the assumption of a healthy market with the most applicable
legal infrastructure. The new directive was published in the EU Official Journal
on December 31, 2004. On July 20, 2007, directive 2004/108/EC repealed the
long-standing EMC directive 89/336/EEC. A major difference is the change in
definition of apparatus to equipment. The term apparatus in the new directive
is now specific to devices or product only, and excludes fixed installations. The
term equipment covers both apparatus and fixed installations. It also simplified
the CE process. Finally, both component body and notified body referred to
the same institution. The last EU EMC directive was introduced in 2014 (EU
EMC directive 2014/30/EU). It covers products that are new to the EU market
and applies to all forms of supply, including distance selling. After April 2016,
the new directive was required for all EMC compliance files, and declarations
referencing 2004/108/EC were longer valid. In addition to those for produc-
ers and importers, it also brings obligations to law makers, market controllers,
and surveillance (i.e., to ministries and national accreditation institutions). The
EMC directive differs from most other directives in that its primary require-
ment is protection of the EM spectrum rather than safety of the equipment.
The EMC directive does not address safety!
1.4 CE Marking Process

CE marking on a product is a manufacturer’s declaration that the product com-
plies with the essential requirements of the relevant European health, safety, and
environmental protection legislation, in practice by many product directives.
For example, most electrical products must comply with the Low Voltage Di-
rective (LVD) and the EMC directive; toys must comply with the Toys Safety
Directive (TSD). CE marking on a product indicates to governmental officials
that the product may be legally placed on the market in their country.
The CE marking shall be affixed only by the manufacturer or its autho-
rized representative. By affixing the CE marking, the manufacturer indicates
that it takes responsibility for the conformity of the product with all appli-
cable requirements set out in the relevant community harmonization legislation
providing for its affixing. Member states shall ensure the correct implementa-
tion of the regime governing the CE marking and take appropriate action in
the event of improper use of the marking. Member states shall also provide
for penalties for infringements, which may include criminal sanctions for seri-
ous infringements. Those penalties shall be proportionate to the seriousness of
the offence and constitute an effective deterrent against improper use. It is a
product passport for free circulation in EU market. CE marking is mandatory

for certain product groups within the EU member states’ markets plus Switzer-
land and Turkey.
CE marking is a self-certification scheme. Retailers sometimes refer to
products as “CE approved,” but the mark does not actually signify approval.
Certain categories of products require type testing by an independent body to
ensure conformity with relevant technical standards, but CE marking in itself
does not certify that this has been done.
Responsibility for CE marking lies with whoever puts the product on
the market in the EU (i.e. an EU-based manufacturer, the importer or dis-
tributor of a product made outside the EU, or an EU-based office of a non-EU
manufacturer).
The manufacturer of a product affixes the CE marking to it but has to take
certain obligatory steps before the product can bear CE marking. The manufac-
turer must carry out a conformity assessment, set up a technical file, and sign a
declaration stipulated by the leading legislation for the product. The documen-
tation has to be made available to authorities on request.
Importers of products have to verify that the manufacturer outside the
EU has undertaken the necessary steps and that the documentation is available
upon request. Importers should also make sure that contact with the manufac-
turer can always be established.
Distributors must be able to demonstrate to national authorities that they
have acted with due care, and they must have affirmation from the manufac-
turer or importer that the necessary measures have been taken.
There are certain rules underlying the procedure to affix the marking:
• Products subject to certain EU directives or EU regulations providing

for CE marking have to be affixed with the CE marking before they can
be placed on the market.
• Manufacturers have to check, as their sole responsibility, which EU leg-
islation they need to apply for their products.
• The product may be placed on the market only if it complies with the
provisions of all applicable directives and regulations and if the confor-
mity assessment procedure has been carried out accordingly.
• The manufacturer draws up an EU declaration of conformity and affixes
the CE marking on the product.
• If stipulated in the directive or regulation, an authorized third party (no-
tified body) must be involved in the conformity assessment procedure or
in setting up a production quality system.
Introduction 19
• If the CE marking is affixed on a product, it can bear additional mark-

ings only if they are of different significance, do not overlap with the
CE marking, and are not confusing and do not impair the legibility and
visibility of the CE marking.
Here are the CE marking steps for a product to be circulated in EU

market:
• Identify applicable directive(s) and standard(s).

• Identify applicable “Conformity Assessment Module.”
• Test one sample of the product, either by yourself or in an accredited
test lab.
• Compile a “Technical Documentation File (TDF).”
• Sign the EC “Declaration of Conformity (DOC).”
• Affix “CE Mark” on the product. Your product is now ready to enter
the EU market.
CE marking, like other certifications marks, is misused. CE marking is

sometimes affixed to products that do not fulfill the legal requirements and
conditions, or it is affixed to products for which it is not required. There are
mechanisms in place to ensure that the CE marking is put on products cor-
rectly. Controlling products bearing CE marking is the responsibility of public
authorities in member states in cooperation with EU authorities. Citizens may
contact national market surveillance authorities if the misuse of the CE mark-
ing is suspected or if a product’s safety is questioned.
The procedures, measures, and sanctions applying to counterfeiting of the
CE marking vary according to the member state’s national administrative and
penal legislation. Depending on the seriousness of the crime, economic opera-
tors may be liable for a fine and, in some circumstances, imprisonment. How-
ever, if the product is not regarded as an imminent safety risk, the manufacturer
may be given an opportunity to ensure that the product is in conformity with
the applicable legislation before being forced to take the product off the mar-
ket. For example, EMC directive is enforced by the communications regulator
(OFCOM) in UK. In addition to a market ban, a misuse may face up to three
months imprisonment and a £5,000 fine.
There are agreements on mutual recognition of conformity assessment
between the EU and other countries such as the US, Japan, Canada, Australia,
New Zealand, and Israel. Consequently, CE marking is now found on many
products from these countries.
The CE mark is neither a quality mark nor a guarantee that the product
meets all of the requirements of relevant EU product safety law.
1.5 EMC Institutions and EMC Standards

Technology is becoming more and more complex, and users and consumers
are depending increasingly on products whose design and construction they
may not understand. Conformity assessment (CA) provides the reassurance
they need. Consumers, through CA, can be confident that electrical and elec-
tronic products are safe to use, energy efficient and environmentally friendly,
and perform to expectations. CA refers to any activity that determines whether
a product, system, or service corresponds to the requirements contained in a
standard. A standard is a technical description of the characteristics a product,
service, or object is required to fulfill.
The world’s leading organization that prepares and publishes interna-
tional standards for all electrical, electronic, and related technologies is the In-
ternational Electrotechnical Commission (IEC) [15]. The IEC, International
Organization for Standardization (ISO), and International Telecommunication
Union (ITU) are three global sister organizations that develop international
standards for the world. All the equipment, devices, and systems that contain
electronics and use or produce electricity rely on IEC international standards
and conformity assessment systems to perform, fit, and work safely together.
IEC provides a platform to companies, industries, and governments for meet-
ing, discussing, and developing the international standards they require. The
two EMC-related organizations under IEC are Technical Committee on EMC
(TC77) and International Special Committee on Radio Interference (CISPR)
[16].
The Advisory Committee on Electromagnetic Compatibility (ACEC) is
an advisory committee of the IEC on EMC. ACEC advises and reports to the
Standardization Management Board (SMB) and guides and coordinates IEC
work relating to EMC. ACEC members, nominated by member national com-
mittees, are experts on EMC. ACEC is responsible for providing guidelines on
drafting EM publications and through which it coordinates the EMC work of
IEC’s technical committees. It maintains a list of all the main product families
covered by IEC EMC standards and organizes workshops to help improve com-
munications between the IEC and industry on EMC-related issues. It also co-
ordinates the IEC’s work related to EMC with other organizations’ committees
such as the European Committee for Electrotechnical Standardization (CEN-
ELEC) [13]. EMC engineering, from design to market, is governed these by
organizations and institutions. They work to promote cooperation on standard-
ization, including publishing various EMC standards. The US FCC, British
Introduction 21
Standards Institute (BSI), and Deutsche Kommission Elektrotechnik (DKE) in

Germany are worth mentioning as national organizations.
EMC-related bodies described in the EU EMC directive are:
• Competent authority: A mechanism established to satisfy EMC require-

ments in member states.
• Competent body: A body responsible from the approval of technical
documentation files. It is authorized by the competent body of a mem-
ber state and/or by the national accreditation body. It is not necessarily
a test center.
• Notified body: An organization that has been accredited by a member
state to assess whether a product meets certain preordained standards.
Member states do not have to establish a notified body.
• EMC test versus measurement centers/labs: Legal commercial centers/
labs established according to related EMC standards (e.g., EN45000
series).
1.5.1 Commercial EMC Standards

National and international authorized organizations and technology frontier
companies prepare standards. The members of these organizations are esteemed
experts in their areas. There are four different standards mentioned in EU EMC
directive:
• Basic standards: They define and describe the EMC problem, measure-
ment and test methods, principle measuring equipment, and test setup.
They contain no limits and state no interference criteria. Basic immu-
nity and emission EMC standards are specified by the IEC and CISPR,
respectively, (e.g., CISPR 16 and IEC 61000).
• Generic standards: These are standards that have to do with a particular
EM environment. They specify appropriate series of requirements and
tests that are used for all products and systems in this environment.
There are two major environments: (1) residential, commercial and light
industry; and (2) industry.
• Product family standards: These standards contain special limits for
emission and immunity for a specific category of devices (e.g., CISPR
22, information technologies). They contain many basic standards.
• Product-specific standards: They are for specific products (e.g.,
prEN50220 is the immunity standard for hearing aid products).
Electrical and electronics devices in commercial markets all around the

world must comply with EMC requirements. EMC standards have slight dif-
ferences from region to region and from country to country, but in general, they
are the same. Turkey, being in the middle of the Asia-Europe-Africa triangle,
adopted EN standards. A few other worldwide recognized standards are IEEE
Std. XXXX, ANSI Std. XXXX, CISPR XXXX, and IEC Std. XXXX. Tables 1.3
and 1.4, respectively, lists examples of EU and US standards (http://ec.europa.
eu/growth/single-market/european-standards/harmonised-standards/electro-
magnetic-compatibility/ index_en.htm):
FCC, IEEE, and ANSI are authorities in the US. Various IEC standards
are used for emission and immunity testing to verify EMC compliance.
1.5.2 Military EMC Standards

Over the years, national and international bodies have devised a complete series
of emission and susceptibility EMC tests. The same was made in the military
domain by various departments of defense. Military EMC is environment spe-
cific and necessitates services for the land, sea, and air sectors [17]. Defense
standards are also called military standards, MIL-STD, or MIL-SPEC. The
main difference between commercial and military EMC standards is that com-
mercial standards are legally enforced, while compliance to military standards is
a matter of contract fulfillment.
Military standards are more strict; therefore, they are considered reliable
enough that they are often used by other government organizations and even
nongovernment technical organizations or general industry. A few well-known
Table 1.3
EU EMC Standards
EN50081-1:1992— Electromagnetic compatibility—Generic emission standard—Part 1:
Residential, commercial, and light industry
EN50081-2:1994— Electromagnetic compatibility—Generic emission standard—Part 2:
Industrial environment.
EN50082-1:1998— Electromagnetic compatibility—Generic immunity standard—Part 1:
Residential, commercial, and light industry
EN50082-2:1995— Electromagnetic compatibility—Generic immunity standard—Part 2:
Industrial environment
EN55013:2001— Limits and methods of measurement of radio disturbance characteristics of
broadcast receivers and associated equipment
EN55014-1:2001— Electromagnetic compatibility—Requirements for household appliances,
electric tools, and similar apparatus—Part 1: Emission—Product family
standard
EN55014-2:1997— Electromagnetic compatibility—Requirements for household appliances,
electric tools, and similar apparatus—Part 2: Immunity—Product family
standard
Introduction 23
Table 1.4
IEEE/ANSI ASC-C63 EMC Standards Used in the US
C63.16-2016- ANS Guide for Electrostatic Discharge Test Methodologies and Acceptance
Criteria for Electronic Equipment
C63.7-2015- ANS Guide for Construction of Test Sites for Performing Radiated Emission
Measurements
C63.23-2012- ANS Guide for Electromagnetic Compatibility—Computations and Treatment
of Measurement Uncertainty
C63.7-2005- ANS Guide for Construction of Open-Area Test Sites for Performing Radiated
Emission Measurements
C63.5-1998- ANS for Calibration of Antennas Used for Radiated Emission Measurements
in Electromagnetic Interference (EMI) Control Calibration of Antennas (9 kHz
to 40 GHz)
C63.15-2010- ANS Recommended Practice for the Immunity Measurement of Electrical
and Electronic Equipment
C63.5-2006- ANS for Electromagnetic Compatibility Radiated Emission Measurements in
Electromagnetic Interference (EMI) Control Calibration of Antennas (9 kHz to
40 GHz)
military standards are DEFSTAN, GAM-EG-13, MIL-461, AECTP 500, and

DO-160. The UK’s defense EMC standards are DEFSTAN 59-41 and 59-
411. These cover all aspects of testing based on the environment in which the
product is used. GAM-EG-13 is a French national defense standard for envi-
ronmental providing of defense equipment. MIL-461 is the American military
EMC standard, widely adopted across the world for more than 40 years. Vari-
ous revisions of MIL-STD-461 have been released. Many military contracts re-
quire compliance to MIL-STD-461E. The latest revision (as of 2015) is known
as MIL-STD-461G. AECTP 500 is the North Atlantic Treaty Organization’s
(NATO) standard for electrical and EM environmental tests. Finally, DO160,
the Avionics EMC standard for all aircraft, is recognized globally.
Military standards when compared to commercial ones are more compre-
hensive and contain specific conditions.
1.6 EMC Limiting Values

An EMC standard presents requirements that a product shall meet and limiting
values; details of tests and measurement methods that are used to show how
these requirements are satisfied; and test and measurement environments and
equipment. Standards also tell what shall be done for any circumstance that is
not included or explained in the standards.
For example, EN50081-1, which was introduced in 1992, is the generic
emission standard for residential, commercial, and light industrial equipment.
On the other hand, EN50081-2 is the same standard for industrial equipment.
In Figure 1.6, CE limiting voltage values, in dBµV, versus frequency for the
range of 150 kHz–30 MHz for both EN50081-1 and EN50081-2 standards
are shown. As observed, limiting CE voltage values vary between 40–60 dBµV
in residential regions, but it is between 65–80 dBµV in industrial areas.
Figure 1.7 shows FCC and CISPR RE limits as electric field (in dBµV/m)
versus frequency in the frequency range of 30 MHz–1 GHz. Here, Class A and
B, respectively, correspond to industrial and residential cases in EN standards.
The values in the figure belong to 10m distance between EUT and the EMI
Figure 1.6 Typical CE thresholds given in EN 50081.

Introduction 25
Figure 1.7 FCC and CISPR typical RE thresholds.
receive antenna. As observed, FCC and CISPR values are similar but not ex-
actly the same.
Note that standardization is a dynamic process where EMC limiting values
and frequency ranges are continuously updated. For example, IEC/EN61000-
4-3 (1995) RI values were first given in the 80 MHz–1 GHz frequency range.
After 2001, tests were mandatory up to 2.5 GHz. Finally, in the 2006 issue of
IEC/EN61000-4-3 the maximum test frequency became 6 GHz.
1.7 EMC Tests and Measurements

A test and/or a measurement, in general, is done for a process control, design,
and/or optimization. In EMC engineering, in addition to these, tests and mea-
surements are also done for certification (i.e., to show compatibility with the
standards). The three fundamental components of an EMC test and/or mea-
surement are:
• The knowledge of product-related standard;

• The knowledge of the functions of the product under tests and measure-
ments;
• The procedure of the tests and measurements.
The characteristic feature of EMC tests and measurements is repeatability.

Every single detail of tests and measurements (from methods to environments,
setup to equipment, procedures, and steps) are determined. If anything is not
covered in the standards, then every detail is recorded and reported including
photos for the sake of repeatability.
The three factors of an EMC test and measurement are:
• Control components: The physical environment of the EMC tests and

measurements;
• Coupling components: Sensors and coupling with the EUT;
• Test/measurement device: Calibration and use of EMC tests/measure-
ment setup.
These components are critical for EMC tests and measurements. For ex-
ample, effects and/or contributions of tests and measurement environments
shall clearly be known; otherwise, the difference between tests and measure-
ments conducted in a screened room and anechoic chamber shall not be un-
derstood and evaluated. Resonance effects of screened rooms, low and high
frequency contributions, and couplings in an anechoic chamber must be well
understood. Similarly, the isolation of the EUT and EMI receiver, the isolation
of mains, and the EUT must clearly be supplied and demonstrated before the
tests and measurements.
AC mains (power supply) is the worst interference source.
Introduction 27
EMC tests and measurements can be divided into two groups: emission
and immunity (or susceptibility).
Emission is measured; immunity/susceptibility is tested.
Both emission measurements and immunity/susceptibility tests can be
performed either radiated or conducted:
• Radiated emission (RE);

• Conducted emission (CE);
• Radiated immunity (RI)/radiated susceptibility, (RS);
• Conducted immunity (CI)/conducted susceptibility (CS).
In addition to these, there are also harmonic tests and measurements,

ESD tests, voltage/current surge measurements, fast transient tests, flicker tests,
and so on.
Commercial EMC tests and measurements are conducted in the fre-
quency range of 150 kHz–6 GHz. According to the standards, the critical fre-
quency that separates conducted and radiated emission measurements is 30
MHz. Conducted emission (through cables, connectors, and groundings) are
required in the range of 150 kHz–30 MHz frequencies. Above 30 MHz, con-
ducted emissions attenuate rapidly and therefore are assumed negligible. Radi-
ated emissions (through atmosphere) are required in the frequency range of
30 MHz–6 GHz. Below 30 MHz, radiated emissions attenuate rapidly and
therefore are assumed negligible.
Emission measurements are relatively simple and easy to conduct but time
consuming. For example, for the RE measurements one needs to (1) locate the
EUT on a wooden turntable, operate as instructed, put the receive antenna at
a specified distance (e.g., 3m, 10m, 30m), scan vertically between 1m and 4m
heights (in order to find maximum value when direct and ground-reflected
emission constructively interfere) and record the E-field in dBµV/m; (2) repeat
this at different EUT locations and for both vertical and horizontal polariza-
tions until worst-case value at that frequency is recorded; (3) do this for every
frequency in 30 MHz–6 GHz range with the specified resolution bandwidth;
(4) compare the results with the threshold given in the standards. The E-field
values are in the range of several tens to hundreds dBµV/m.
Immunity/susceptibility tests are relatively difficult and necessitate spe-
cial equipment. They require generation of high and homogeneous electric and
magnetic fields and also clear definition of performance. The field values are in
the order of V/m and/or A/m.
The calibration of the EMI receiver and other test/measurement equip-
ment and test/measurement environment is also critical. Calibration answers
whether or not the measured value belongs to only the EUT, and not from
improper use of equipment, not from screened room, not from the mains, and
so on.
1.8 EMC Engineering Philosophy

EMC deals with device-device EM interaction. EMC engineers are very often
faced with problems that are not taught at universities in either EM or circuit
lectures. EMC problems are mostly hidden in details and can be solved by
trial and error. This necessitates a good understanding of physics and a well-
established knowledge of EM and circuit theories, but in the end one needs
to look at each problem in terms of EMC engineering. Here are a few sayings
about EMC engineering:
• Two EMC engineers do not agree with each other unless they are drunk.
• If an EMC engineer wants to be sure, he or she never measures twice.
• Nobody grows up wanting to be an EMC engineer; it usually just hap-
pens.
• Very often, what you predict falsifies you.
• The number of your mistakes is proportional to your experience.
Another example that summarizes EMC philosophy is the measurement

results given in Figure 1.8. In 1999, AT-type and tower-type desktop computers
and microprocessor speeds increased to 200 MHz.
The RE of these two type computers were measured at 3m distance be-
tween 30 MHz–10 GHz frequencies. The results were given in terms of shield-
ing effectiveness (SE) versus frequency. As will be discussed in Chapter 10,
simply, SE is the isolation supplied by any object. Here, it is the ratio of the
amplitudes of the fields measured without and with the shield.
The solid lines in the figure belong to AT and tower cases. The dashed
lines also belong to AT and tower cases but after modifying the isolation. The
following observations show fundamentals of EMC philosophy:
• In general, SE of the tower case is better than AT case (this may change
with brand, with design, and so on).
• EMC problems of a computer with a 200-MHz processor may be im-
portant up to 10th harmonics (i.e., up to 2–3 GHz).
• Extra screening of both cases seems to modify SE in certain frequencies
but also degrade at some other frequency regions.
• SE versus frequency behaviors are too complicated.
Introduction 29
Figure 1.8 A typical SE measurement conducted in 1999.
Two important conclusions that can be derived from this example are:
• EMC behaviors of devices and systems are extremely complicated; very

often the predictions falsify us. For example, it is almost impossible to
predict high (low) frequency behaviors by looking at low (high) frequen-
cy behaviors; therefore, every single frequency has to be measured or
tested separately.
• There is no general or generic solution, approach, or method for EMC
problems. The protection method may solve your EMC problem in a
certain frequency range but, whether or not you are aware of it, may
make the problem worse at some other frequencies.
1.9 Suggested EMC Approach

EMC engineers must be well-equipped with EM fields and waves, circuit the-
ory, transmission lines, microstrip circuits, antenna theory, EM scattering and
diffraction, frequency transforms, DFT, FFT, electronic elements, noise and
random signals, signal integrity, stochastic modeling, numerical modeling and
simulation, signal processing, material science, and system theory. They must
revisit these topics in terms of EMC engineering. EMC field is usually consid-
ered a black magic by those who do not understand EM. In reality, one can
solve some of the most complex aspects of EMC by understanding the funda-
mental or basic aspects of Maxwell’s equations.
Critical parameters/quantities of an EMC problem are frequency, signal
amplitude, time behavior, geometrical discontinuities, and boundary/edge con-
ditions. Specifically these interrelations are vital:
• Time-frequency;
• Wavelength (frequency)-geometry (size);
• Frequency-impedance.
Because:
• EMI may be temporary or permanent in the time domain. These two

create totally different EMC problems, and solution approaches are to-
tally different.
• Identifying and solving an EMC problem, in general, is easier in the
frequency domain. Unfortunately, almost all of today’s devices and sys-
tems are digital. An EMC engineer must use DFT instead of analytical
Fourier transform. Fourier transform is a mathematical definition; DFT
is a numerical approach. It brings in problems such as aliasing, spectral
leakage, maximum frequency and frequency resolution, and so on.
• Theoretically, discrete signals have infinite number of harmonics. In
practice, a good EMC approach is to pay attention up to 10th harmon-
ics. Standards for AC mains (power supply) make it mandatory up to
40th harmonics. For the 50 (60) Hz AC mains frequencies, up to 2 kHz
(2.4 kHz) will be important.
• Geometry/size/length is a critical parameter in EMC engineering. Sig-
nal frequency or size alone does not mean anything; an EMC engineer
needs to look at wavelength-size (λ -l) relation.
• EM emissions vary with current frequency, amplitude, and size.
• Equal and opposite flowing currents are usually useful signals in electri-
cal circuits, and they reduce unintentional emissions.
• Cables and connectors act like a wire antenna; apertures and holes act
like an aperture antenna. A practical tip to reduce their unintentional
emissions is to keep their sizes much less than the signal wavelength (l
<< λ, e.g., l ≤ λ/20).
Introduction 31
Yesterday, many companies’, directors’, and engineers’ approach was “Let’s

design and prototype first; we’ll then take care of EMC problems on the way.”
Today, they all say “EMC engineering from design to market.” Unfortunately,
they have learned this by experience (i.e., after exercising years of EMC prob-
lems and wasting a lot of money and time). Now, small companies hire at least
one EMC engineer and train all other engineers; big firms establish their own
EMC groups and labs. EMC engineering is not something that is being taught
in colleges. Undergraduate engineering classes on EMC are very few; therefore,
EMC engineers set up self-training programs and attend hands-on professional
EMC courses. An effective way of EMC training is to witness EMC tests and
measurements.
Choose your approach as an EMC engineer:
• Crisis approach: Disregard EMC until the design is completed. Use add-
ons to fix the problems as they arise during testing and field experience.
This is also called the band-aid approach and tends to be expensive.
• Systems approach: Consider EMC throughout the design. Anticipate
all possible problems from the start. Do thorough EMC test of final
prototypes. This is more cost effective because EMC is designed into the
product and not added on when problems later arise.
In summary, EMC is achieved when (1) products do not adversely inter-

fere with their environment (emission); (2) the environment does not upset the
operation of the products (immunity); (3) products do not adversely interfere
with themselves (signal integrity).
References
[1] Sevgi, L., Elektomanyetik Uyumluluk—Elektromanyetik Kirlilik, Elektrik Mühendisleri
Odası (EMO) Yayınları, Istanbul, December 2000.
[2] Sevgi, L., Tasarımdan Piyasaya Elektromanyetik Uyumluluk, OKAN Üniversitesi Yayınları,
Istanbul, March 2015.
[3] Sevgi, L., Tasarımdan Üretime Elektromanyetik Uyumluluk, Eksen Yayıncılık, Istanbul,
June 2006.
[4] Sevgi, L., Ruediger Vahldieck (Guest Editor), Special issue on “From Design to Market of
Electronic Systems: EMC Engineering,” TJEECS, Turkish Journal of Electrical Engineering
and Computer Sciences, Vol. 17, No. 3, Nov. 2009, http://mistug.tubitak.gov.tr.
[5] Sevgi, L., “Modeling and Simulation Strategies in EMC Engineering: Problems, Chal-
lenges and Tools,” EMC Zurich 2009, 20th International Zurich Symposium on Electro-
magnetic Compatibility, Jan 12–16, 2009, Switzerland.
[6] Sevgi, L., “EMC Testing Centers and Labs: Problems Related to CE Marking,” EMC Zu-
rich 2009, 20th International Zurich Symposium on Electromagnetic Compatibility, Jan
12–16, 2009, Switzerland.
[7] WHO, World Health Organ�ization, http://www.who.int.
[8] ICNIRP, http:// http://www.icnirp.org.
[9] Sevgi, L., “Biostatistics and Epidemiology: Hypothetical Tests on Cell Phone Users,” IEEE
Antennas and Propagation Magazine, Vol. 52, No. 1, Feb. 2010, pp. 267–273.
[10] American Cancer Society, www.cancer.org.
[11] Ioannidis, John P. A., “Why Most Published Research Findings Are False,” PLoS Medicine,
Vol. 2, 8, Aug 2005, pp. 696–701.
[12] Sevgi, L.. “Biostatistics and Epidemiology: Hypothetical Tests on Cell Phone Users,” IEEE
Antennas and Propagation Magazine, Vol. 52, 1, Feb. 2010, pp. 267–273.
[13] CENELEC, http://www.cenelec.eu.
[14] Siemens, www.siemens.com
[15] IEC, http://www.iec.ch.
[16] CISPR, http://www.iec.ch/emc/iec_emc/iec_emc_players_cispr.htm.
[17] Military Standards, www.dsp.dla.mil.
Bibliography
EU Directive, http://ec.europa.eu/enterprise/sectors/electrical/emc.
2
Accreditation
Accreditation is discussed in this chapter. Worldwide accreditation institutes
are reviewed. Procedures and standards for accreditation, certification, and cali-
bration are summarized through EMC tests and measurements. Error, uncer-
tainty and statistical evaluations, and proficiency tests with typical results are
presented.
2.1 Introduction
Performance enhancing drugs (PEDs) have long been used in professional
sports and the Olympic Games. The first documented misuse was in the 1904
marathon by the winner Thomas Hicks. The International Olympic Commit-
tee (IOC) banned PEDs in 1967 and introduced the first drug use controls at
the 1968 Winter Olympics. These controls include systematic testing of ath-
letes’ urine and blood. The World Anti-Doping Agency (WADA) became the
authoritative body on the use of PEDs in 1999. The PEDs have become more
and more sophisticated, but WADA has also innovated new ways to detect these
drugs. The first positive test was at the 1968 Summer Olympics, and the Swed-
ish pentathlete Hans-Gunnar Liljenwall lost his bronze medal for alcohol use.
A well-known case was the Canadian sprinter Ben Johnson, who won the men’s
100 meters at the 1988 Seoul Olympics but lost his gold medal when he tested
positive in doping tests.
Diana Taurasi, a star of the Women’s National Basketball Association,
tested positive and was suspended for a banned substance in December 2010
while playing for the Turkish basketball team Fenerbahce, which could have
jeopardized not only her competition at the 2012 Olympic Games in Lon-
don but also her entire career. Tests done by the Doping Control Center at
33
Hacettepe University, Turkey, were then found to be mistaken. The Disciplin-

ary Committee of the Federation of Women Basketball Sports then lifted the
suspension. In 2011, WADA indefinitely suspended the Turkish drug-testing
laboratory that reported an apparent “false positive.”
EU rapid alert system RAPEX [1] informs member states and the EC on
measures taken to prevent or restrict the marketing or use of products posing
a serious risk to the health and safety of consumers with the exception of food,
pharmaceuticals, and medical devices, which are covered by other mechanisms.
Every Friday, the EC publishes a weekly overview of the products posing a
serious risk, which gives information on the product, the identified risk, and
measures that were taken in the notifying country. A sample notification report
on a battery charger is given in Figure 2.1.
What is in common in these examples and many more? Testing— doping
testing, EMC testing, safety testing, and so on. Depending on the results of a
test, you can ban millions of products from the market, imprison a producer,
ruin the life of a champion, and so on. What if it is wrong as in Taurasi case?
Figure 2.1 A typical RAPEX notification.

Accreditation 35
How can we be sure on the results of the tests? How can we be sure on the preci-
sion of the measurement and accuracy of the result? How can we be sure that the
test/measurement devices/instruments are capable (calibrated)? How can we be
sure that the test/measurement operator is competent? How can we be sure that
everything is certified? The answer is accreditation.
2.2 Accreditation
Annual world trade (global export) approaches dozens of trillion US dollars.
Consumers buy products imported/exported all around from China to Africa,
South America to Australia. The safety and quality of the trade is controlled by
international standards and compliance programs that facilitate the movement
of goods and services across national borders. The ISO is the world’s largest
standard-setting body. Many of ISO’s standards include guidance on how to
demonstrate that a product, person, service, or system meets the requirements
contained within a standard. ISO/IEC 17011 (Conformity Assessment-Gener-
al Requirements for Accreditation Bodies Accrediting Conformity Assessment
Bodies) defines accreditation as “a third-party attestation related to a confor-
mity assessment body conveying formal demonstration of its competence to
carry out specific conformity assessment tasks.”
The philosophy of the accreditation is given in Figure 2.2. The aim is
to remove all the barriers of international trade without degrading the safety
and quality. It is an effective marketing tool that brings public confidence. It
gives confidence through ensuring consistently high standards in the quality of
products/services purchased, and minimizes failure or recalls. This is achieved
through repeatable tests/measurements carried out by a reliable independent
Figure 2.2 The philosophy of accreditation.

organization. Accreditation also prevents dangerous outbreaks such as food poi-

soning, water contamination, and so on, that can be hazardous to the public.
Accreditation is beneficial to all parties—to the accredited body itself,
to the governments, and to the users of accredited bodies. Test and calibration
laboratories can demonstrate through accreditation that their measurements
are traceable to international standards. The benefits for users include assurance
that test results are obtained using properly calibrated equipment by staff with
the right level of expertise. It enables manufacturers to produce reliable prod-
ucts, eliminate technical barriers to trade encountered by exporters, provide
added credibility to the results issued by government institutions that have been
accredited, and provide increased confidence that data generated from multiple
accredited laboratories are comparable.
Accreditation is a dynamic process and has distinguishing characteristics.
First, accreditation is independent, impartial, and operates on a not-for-profit
basis. It operates within a legal framework and is completely independent from
commercial motivations. When carried out against the recognized harmonized
standards, accreditation is regarded as a public authority activity. It competes
with neither other accreditation bodies nor conformity assessment bodies.
Note that although sometimes used interchangeably the terms accredita-
tion and certification are not synonymous. Certification is the process of veri-
fying that personnel have adequate credentials practicing certain disciplines
and/or verifying that a product meets certain requirements. Accreditation, on
the other hand, is used to verify that test/measurement centers/labs have ap-
propriate quality management systems, can properly perform certain tests, do
calibrations, and measure quantities. Roughly speaking, accreditation is one
step higher than certification. For example, universities themselves are first ac-
credited and then yield degrees (a kind of certification) recognized worldwide.
2.3 Accreditation Institutions

Worldwide accreditation bodies include International Accreditation Forum
(IAF) [2], International Laboratory Accreditation Cooperation (ILAC) [3],
European Accreditation (EA) [4], Inter-American Accreditation Cooperation
(IAAC) [5], and Asia Pacific Laboratory Accreditation Cooperation (APLAC)
[6]. A global acceptance of the services provided with these bodies is established
by signing a multilateral agreement (MLA), which constructs the essential
trusting mechanism. MLAs create confidence in, and acceptance of, accredited
certifications, inspections, and test reports, eliminating the need for suppliers
to be certified in each country where they sell their products or services. The
MLA signatories recognize and accept the equivalence and the reliability of the
accreditation systems operated by the signing members.
Accreditation 37
The activities and standards covered by most of the MLAs are listed in
Table 2.1. As listed there, members of the regional accreditation bodies provide
accreditation of labs (testing according to ISO/IEC 17025, medical examina-
tions according to ISO 15189, calibration according to ISO/IEC17025, in-
spection bodies according to ISO/IEC 17020), certification bodies (product
certification according to EN45011-ISO/IEC 17065, certification of persons
according to ISO/IEC 17024), management systems certification according to
ISO/IEC 17021.
The IAF oversees, evaluates, and formally recognizes bodies that accredit
organizations that certify management systems, products, or persons. The ILAC
oversees, assesses, and formally recognizes accreditation bodies that participate
in the accreditation of calibration and testing laboratories. It has regional co-
operations with EA, APLAC, and IAAC. The merging of IAF and ILAC in the
future is a common expectation.
Like all the other regional accreditation bodies, the EA is a nonprofit
association set up in November 1997 and registered in The Netherlands in
June 2000. EA has the overall strategic objective to safeguard the credibility of
accreditation services delivered by its member countries within the European
market.
The regional accreditation bodies’ missions include defining, harmoniz-
ing, and building consistency in accreditation in their region. They ensure com-
mon interpretation and application of the standards used by their members.
They also ensure transparency of the processes and procedures. They maintain
a multilateral agreement based on mutual recognition between accreditation
activities and reciprocal acceptance of accreditation services. They manage a
peer evaluation system consistent with international practices.
A typical structure of the EA is pictured in Figure 2.3 (structures of the
IAAC and APLAC are also the same). Some of the national institutes within
EA are the Danish (DANAK), Finnish (FINAS), French (COFRAC), German
(DAR-DKD), English (UKAS), Czech (CAI), Slovak (SNAS), Irish (INAB),
Table 2.1
MLA Accreditation Coverage
Accreditation of Scope/Activity: Standard
Laboratories: Testing and Medical examinations ISO/IEC 17025 ISO 15189
Laboratories: calibration ISO/IEC 17025
Certification bodies: product certification EN 45011 and ISO/IEC 17065
Certification bodies: certification of persons ISO/IEC 17024
Certification bodies: management systems certification ISO/IEC 17021
Inspection ISO/IEC 17020
Verification EN/ISO 14065
Figure 2.3 The structure of regional accreditation bodies.
Spanish (ENAC), and Portuguese (IPAC) accreditation bodies. The Turkish ac-
creditation body (TÜRKAK) is also a full member of EA, IAF, and ILAC.
All accreditation bodies aim to provide a high level of service for regional
economies, industries, governments, and citizens. An accredited organization’s
competence, impartiality, and sustainable performance have to be demonstrat-
ed. An accreditation organization can ensure that everyone from specifiers, pur-
chasers, and suppliers to consumers have confidence in the safety and quality of
certified goods and services throughout the supply chain.
There are numerous mutual recognitions of conformity between EU and
other countries such as the United States, Japan, Canada, Australia, New Zea-
land, and Israel. Because of this, products from these countries may not have af-
fixed CE mark. Japan has its own marking known as the Technical Conformity
Mark. Switzerland and Turkey (which are not members of the EU) also require
products to bear CE marking as an affirmation of conformity.
2.4 TÜRKAK
Being related with Ministry of Industry and Trade and subject to private law
provisions, TÜRKAK was established as a legal entity having administrative
and financial autonomy with its headquarters in Ankara. (TÜRKAK Law No.
4457 was adopted on 27/10/1999 and published in the 4/11/1999/No.23866
issue of the Official Gazette.) The objective of TÜRKAK is to facilitate the free
movement of goods in the Turkish market by assuring the reliability of reports
and certifications granted by Conformity Assessment Boards (CABs) in the
Accreditation 39
direction of international agreements and criteria. A global acceptance of the

services provided with an accreditation is also subject to some questions, as to
establish a confidence in the quality of the accreditation service itself. The trust-
ing mechanism between accreditation bodies is constructed on the MLAs at the
international and regional accreditation body organizations. TÜRKAK, started
to provide accreditation services in 2001 and became a signatory of MLA with
EA for all the available accreditation schemes in 2008. Currently, TÜRKAK is a
full member of EA, IAF, and ILAC. Its specialties are accreditations of laborato-
ries, inspection/certification/product certification/personnel certification bod-
ies, reference material producers accreditation, as well as accreditation body op-
erations training. TÜRKAK also performs other assessments and approvals of
competence of bodies carrying out conformity assessment activities. TÜRKAK
makes arrangements to encourage the use of markings and certificates issued
by the accredited organizations; establishes relationships and cooperates with
international and regional accreditation bodies and those of other countries;
ensures the confidentiality of information obtained in relation with the applica-
tion, evaluation, and accreditation of the organizations applying for accredita-
tion; carries out activities to promote the importance of accreditation and the
consciousness of quality; and performs other related duties. The logo, address,
and link of TÜRKAK in IAF’s webpage is as given in Figure 2.4.
TÜRKAK is responsible for accreditation in accordance with the relevant
national, international, and EU standards and guides, and the applicable na-
tional, international, and European requirements.
Figure 2.4 TURKAK info in IAF site (http://www.iaf.nu/articles/IAF_MEM_Turkey/110).

2.5 EMC Tests and Measurements

An interesting procedure showing the importance of accreditation is the EC
EMC directive [7]. EMC has become critical in the electronic industry/market
as well as academia for the last few decades because of the worldwide inter-
est in commercial and domestic equipment sectors. It’s a dynamic area where
evolution has accelerated in 1990s with the introduction of the EC EMC di-
rective. The directive applies to all electrical/electronic products causing or be
disturbed by EM disturbances. A large number of electrical/electronic manu-
facturers need to show that their products are compliant with the requirements
of the directive.
The EC EMC directive harmonizes regulations and applies to almost all
electro-technical equipment, and this yields EMC standards. The standards are
continuously updated by international expert bodies, such as the IEEE and
IEC, whose role is to issue international standards.
2.5.1 Features of EMC Tests and Measurements

Emission is measured; immunity is tested. Tests and measurements are con-
ducted according to related standards. Standards contain detailed information
of the test and measurement (1) environments; (2) instruments and setups; (3)
methods and procedures; (4) evaluation and reporting. EMC tests and mea-
surements toward CE marking necessitate monitoring of a product from design
to market [8–10].
EMC tests should be reliable, realizable, and repeatable; therefore, trace-
ability is a must. Traceability is the ability of a test or measurement result/value
available for a comparison within a chain of national/international references
within a specified uncertainty. A typical emission measurement setup is pic-
tured in Figure 2.5.
To satisfy traceability, the following steps should be taken according to the
related standards:
• The measurements can be conducted in an anechoic chamber or in an

OFTA. In either case, one needs to be sure that all undesired environ-
mental interference and noise source are well-isolated. This is achieved
by OFTA/anechoic chamber calibration [11].
• The EUT is located on a turntable with negligible EM scattering char-
acteristics. Standards list all critical information related to that.
• EUT is connected to AC mains and worst-case emissions are recorded
by a broadband receive antenna. This needs to find out EUT’s maxi-
mum emission direction, orientation, and so on. The staff must show
Accreditation 41
Figure 2.5 A typical emission measurement setup: (1) eliminate external interference; (2)
minimize turntable contributions; (3) mutually isolate EUT and mains; (4) mutually isolate re-
ceive antenna and mains.
that there is no mutual disturbance between the mains and EUT and
that they are well isolated.
• Emissions captured by the receive antenna are recorded by an EMI re-
ceiver. The antenna, cables, and the EMI receiver must all be calibrated
[12, 13].
Only accredited labs can maintain all these requirements, and tests and
measurements can be traceable. National accreditation bodies do the accredi-
tation. Lab accreditation covers personnel management system accreditation,
the calibration of the test/measurement environments and instruments, and
accreditation of the test/measurement methods. Note that accreditation is a
dynamic process and the continuity is maintained through proficiency tests
summarized in Section 2.6.
Emissions are typically measured from a few hundred kHz up to several
GHz frequencies. Roughly speaking, CE along cables and wiring are performed
in the frequency range of 150 kHz–30 MHz and results are presented in dBµV.
Above 30 MHz CE is negligible. On the other hand, RE are performed in the
frequency range of 30 MHz–1 GHz (up to 6 GHz) and results are presented in
terms of EM field strength dBµV/m. Below 30 MHz RE is negligible.
Like all the others, EMC tests and measurements are challenging and
time consuming. For example, EN55016-2-3:2010 (CISPR 16-2-3:2010) cov-
ers RE measurements for household appliances in OATS and anechoic chambers
in the range of 30 MHz to 1 GHz. Measurements are performed with a resolu-
tion bandwidth (RBW) of 120 kHz with a scan step much less than this value.
Usually, a scan step of 50 MHz is chosen. This results in 970 MHz / 50 kHz =
19,400 steps. Note that peak (PK), quasi-peak (QP), and average (AV) detec-
tors are used in EMC measurements. PK measurements are a few order magni-
tudes faster than QP and AV measurements, but the limits are given in terms of
QP and AV values. EMC measurements are performed in two steps in order to
speed up the work: prescan and final. Prescan is done with the PK detector; QP
detector is used in final measurements. The typical measurement durations of
PK and QP detectors are 20s/MHz and 1 ms/MHz, which mean PK measure-
ments (one scan in 30 MHz–1 GHz band lasts 324 min in QP mode, but less
than a second with PK detector).
Note that CISPR 22 (2009) for ITE (also EN55022) necessitates condi-
tional RE measurements in the range of 30 MHz–6 GHz. If the internal EUT
sources are less than 108 MHz, then RE measurements are mandatory up to
1 GHz. If it is between 108 MHz and 500 MHz, then measurements shall be
performed up to 2 GHz. For EUTs with internal sources between 500 MHz
and 1 GHz, RE measurements are requested up to 5 GHz. Finally, if the EUT
has sources above 1 GHz, RE measurements shall be conducted up to 6 GHz or
5fmax, whichever is less (e.g., if EUT has a source at 1.1 GHz, then RE measure-
ments will be done in the range of 30 MHz–5.5 GHz). For RE measurements
above 1 GHz, the RWB is 1 MHz; measurements are requested with both PK
and AV detectors.
Both prescan and final measurements can be done in a 10m anechoic
chamber. First, the maximum emission direction of the EUT and its polariza-
tion are found with at least 4 prescans. These fast prescans yield critical frequen-
cy regions where recorded emissions are close to or above the limits given in
the standards. Then final measurements in these critical regions are done with
QP detector. the 10m anechoic chambers are expensive; therefore, a significant
number of EMC labs have 3m anechoic chambers. In this case, prescan is done
in the chamber, but final measurements are performed in an OATS. Note that
the reliability of an OATS is higher than an anechoic chamber.
2.5.2 Calibration
The CE marking of a product is possible if tests and measurements reveal that
the product meets all the conditions given in related standards. Tests and/or
measurements done in an accredited lab assuming the following:
1. The personnel is professional and know how to measure;

2. The lab can cope completely with the requirements;
3. Measurement method, setup, and instruments are all as specified in
the standards;
4. The measured value belongs to only the product under test with the
evaluated uncertainty bounds.
Accreditation 43
The key issue is the calibration [11–13]. Calibration is a comparison be-

tween measurements; one of known magnitude or correctness made or set with
one device and another measurement made in as similar a way as possible with
a second device. The device with the known or assigned correctness is called the
standard. The second device is the unit under test (UUT), oequipment under
test (EUT), device under test (DUT), test instrument, or any of several other
names for the device being calibrated.
The formal definition of calibration by the International Bureau of
Weights and Measures (IBWM) is the following: operation that, under speci-
fied conditions, in a first step, establishes a relation between the quantity val-
ues with measurement uncertainties provided by measurement standards and
corresponding indications with associated measurement uncertainties (of the
calibrated instrument or secondary standard) and, in a second step, uses this
information to establish a relation for obtaining a measurement result from an
indication.
National Metrology Institutes (NMI), which exist in many countries,
maintain primary standards of measurement (the main SI units plus a number
of derived units) that will be used to provide traceability to customers’ instru-
ments by calibration. NMIs support the metrological infrastructure in coun-
tries by establishing an unbroken chain from the top level of standards to an
instrument used for measurement.
2.5.3 Reporting and Product Certification

Reporting and product certification are achieved via statistical evaluation of the
recorded EMC tests and measurements. Evaluation is different for emission
measurements and immunity tests. Frequency by frequency, measured emis-
sions are compared with the EMC limits given for the frequency band in the
related standard. A typical emission measurement result is shown in Figure 2.6.
Here, PK values of the conducted emissions are recorded in dBµV between 150
kHz and 30 MHz. AV and QP limits are also plotted. PK value of a measure-
ment is always greater than QP and AV values; therefore, if measured PK values
are (10–15 dB) below the limits, the product passes the test. If not, measure-
ments are repeated in critical regions with QP and AV detectors. As observed,
the emissions of the EUT of this measurement are below the limit; therefore,
it passes the test.
Immunity tests are based on pass/fail. The EUT must operate in a fully
exercised mode, allowing for the easy observance of failures. The tests neces-
sitate clear definition of performance criteria against which the EUT will be
assessed. These are commonly divided into three categories. During an immu-
nity test the EUT continues to operate as intended (criterion A), degradation
of the EUT’s performance occurs, but normal operation resumes at the end of
Figure 2.6 A typical CE measurement result.
the test with no data loss (criterion B), and the EUT either stops functioning
or its performance degrades and does recover after the test without intervention
(criterion C).
Once all the tests and measurements are over, an EMC compliance report
is issued by the accredited lab. This report contains the product info, lab ad-
dress, a copy of the lab’s accreditation certificate, info about lab facilities, mea-
surement instruments and their calibration certificates, all the standards used
in the tests and measurements, test/measurement configurations, tables and/or
figures of the test/measurement data and pass/fail info, and signatures of the
technical personnel.
2.6 Proficiency Testing and Interlab Comparisons

Proficiency (comparative) testing is a scheduled part of lab quality assurance
where a group of labs compare their results of a test with many others who
have tested the same sample (without knowing its reference value) [14]. The
samples must be homogenous and stable. It’s a mechanism by which a lab can
demonstrate competence based on practical evaluation and an important way
of meeting the requirements of ISO/IEC 17025 standard in the area of quality
assurance of laboratory results. Proficiency tests certainly validate the participat-
ing labs’ measurement method, personnel and technical training, traceability of
standards, and uncertainty budgets [15].
The typical format of proficiency testing programs is that the accredi-
tation body issues a set of samples to each participant together with a set of
Accreditation 45
instructions and any necessary background information. The participating labs

then carry out the requested analyzes in their normal manner and submit their
results. The results are then statistically handled [16, 17] by the body to generate
a report. Each participant is confidentially provided with a report to allow them
to compare their performance with the other participants. The performance of
individual laboratories will only be known by that particular laboratory and a
limited number of management personnel.
Typical proficiency test results obtained from 11 independent lab mea-
surements are shown in Table 2.2 and Figure 2.7 [18]. Here, CE of a hair
dryer (220V, 50 Hz, operating at hot-air blowing mode) is recorded in the QP
detection mode at different frequencies. Table 2.2 lists all the recorded values.
Figure 2.7 shows results at 1 MHz frequency with upper and lower conducted
emission QP measurement limits.
Note that upper/lower limits are obtained by first finding the average
(Vave) and standard deviation (∆V) and then adding/subtracting the value twice
of the standard deviation (Vsup = Vave + 2∆V, Vsub = Vave –2∆V ) [17] for 95%
confidence level. According to ISO/IEC Guide 43-1, z-value of a lab is found
from
Vi − V ave
zi =
∆V (2.1)
where Vi is the measured value of lab i. In conclusion, |zi| ≤ 2 is accepted as sat-

isfactory, 2 ≤ |zi| ≤ 3 is suspicious, |zi| = 3 is unsatisfactory. Suspicious QP values
of the conducted emission measurement are shown in Table 2.3.
Table 2.2
QP Values Recorded in Different Labs (CE Measurements of a
Hair Dryer Operating at Hot-Air Blowing Mode)
Lab|Freq 160 kHz 500 kHz 1 MHz 3 MHz 5 MHz 10 MHz 20 MHz 25 MHz
101 69.6 61.5 55.1 45.3 54.2 50.4 38.5 30.3
102 70.8 64.5 55.6 45.6 54.8 52.1 41 31.6
103 70.56 58.95 48.77 40.89 48.06 50.04 40.21 30.6
104 70.8 61.5 56.2 45.9 56.1 52.9 42.7 34
106 70.5 61 55 46 55 53 41 34
107 71.2 60.1 52 46 54 50 39 30
108 70.6 60.6 55 45.3 54.1 51 38 28
109 70.1 56 55.1 46.5 51 50 36 29
110 67 58.8 51.4 38.9 42.2 53.8 37.8 30.1
111 72.2 63.7 57.7 47.9 57.3 54.4 41.4 33.2
113 72.5 61.5 55.3 45.5 54.5 51.7 38.5 32.5
Figure 2.7 Typical results of a proficiency test (CE of a hair dryer operating at hot-air blowing
mode, QP values at 1 MHz).
Table 2.3
Suspicious Values of the CE Measurements Listed in Table 2.2
Conducted Em. –QP
Lab No 0.16 MHz 0.5 MHz 1 MHz 3 MHz 5 MHz
101 — — — — —
102 — — — — —
103 — — X — —
104 — — — — —
106 — — — — —
107 — — — — —
108 — — — — —
109 — X — — —
110 X — — X X
111 — — — — —
112 — — — — —
113 — — — — —
The proficiency tests map current status of accredited labs and measure-
ment centers. It increases the confidence level of the personnel of the participat-
ing labs in their abilities and knowledge of their capabilities. They can detect
Accreditation 47
any difficulties they may have with analyses, identify training needs, and have a
mechanism where these needs can be met.
The ISO/IEC 17025 standard used for laboratory accreditation necessi-
tates a laboratory to produce an estimation of the uncertainty of the conducted
measurements using accepted methods of analysis, through the production and
application of suitable uncertainty of measurement procedures [15]. Uncertain-
ty is a range that is likely to contain the true value of a quantity being measured
or calculated.
In general, no measurement or test is perfect, and the imperfections give
rise to error of measurement in the result. Consequently, the result of a mea-
surement is only an approximation to the value of the measurand and is only
complete when accompanied by a statement of the uncertainty of that approxi-
mation. It is related with the accuracy as well as the precision [16].
Accuracy is the closeness of the measured or calculated quantity to its ex-
act value and is expressed in terms of error. Error is the difference between a
measured or calculated value of a quantity and its exact value. The error may
be systematic or random. Systematic error is error that plagues experiments or
calculations caused by negative factors (e.g., a dc voltage component that is
unintentionally present, a drift because of the failure of a component, and so
on). Systematic errors can be complex but can be removed once understood or
discovered via careful controls and calibration. Random error is an error that is
always present but varies unpredictably in size and direction. Random errors are
related to the scatter in the data obtained under fixed conditions that determine
the repeatability (precision) of the measurement. Fortunately, random errors
usually follow well-behaved statistical rules. Their effects can be reduced by
repeating the measurement as often as possible.
Note that in addition to systematic and random errors, measurement er-
ror also includes drift error due to system performance changing after the cali-
bration, mostly because of climatic (e.g., temperature) variations.
Precision is a measure of the closeness of the value obtained via multiple
measurements to the true value. It is the total amount of random error present.
A very precise measurement means a small random error. Precision is given as
the percentage of the ratio of the value region to the true value of the quantity
being measured. The value region is the difference between the maximum and
minimum values in multiple measurements. As illustrated in Figure 2.8, high
accuracy means high precision, but high precision does not necessarily mean
high accuracy.
Confidence level [17] is the probability that the true value of the measure-
ment or calculation falls within a given range of uncertainty caused by the
inherent random nature of the measurement or calculation. Confidence levels
can be defined through a good understanding of the nature (probabilistic dis-
tributions) of the errors.
Figure 2.8 Accuracy versus precision.
Unless otherwise noted, certification measurement results are stated with

95% confidence level. This is calculated as follows. Suppose, X is a Gaussian
distributed measurand with mean µ and standard deviation σ then Z = (X –
µ)/σ is normally distributed with zero mean and unit standard deviation. The
probability of z1 ≤ z ≤ z2 (which corresponds to the confidence level of s) can
be calculated from:
z2
1  v2 
s = P ( z1 ≤ z ≤ z 2 ) = ∫ exp  −  dv (2.2)
2π z1  2
Note that P(–∞ < z < ∞). The probability of getting z between z1,2 =
±1.96 is 0.95. In other words, we’re 95% confident that z would be in the range
of –1.96 ≤ z ≤ 1.96 (i.e., the measured value is in the range of –1.96∆x ≤ x ≤
1.96∆x or approximately –2∆x ≤ x ≤ 2∆x).
The success of the world trade is directly related to the elimination of
technical/bureaucratic barriers. This can be achieved only with worldwide ac-
creditation and multilateral agreements. Accreditation is essential for the cor-
rect operation of a transparent and quality-oriented market.
The goal is to find the most effective way of achieving a single system that
will allow companies with an accredited conformity assessment certificate in
Accreditation 49
one part of the world to have that certificate recognized elsewhere. International
institutes and standards aim to ensure that accreditation bodies only accredit
bodies that are competent to do the work they undertake and reduce risks to
business and customers by ensuring that an accredited certificate may be relied
upon anywhere in the world.
References
[1] EU rapid alert system RAPEX, http://ec.europa.eu/consumers-/safety/rapex/alerts.
[2] IAF, International Accreditation Forum, http://www.iaf.nu.
[3] ILAC, International Laboratory Accreditation Cooperation, http://www.ilac.org.
[4] EU, The European co-operation for Accreditation, http://www.european-accreditation.
org.
[5] IAAC, Inter-American Accreditation Cooperation, http://www.iaac.org.mx.
[6] APLAC, Asia Pacific Laboratory Accreditation Cooperation, https://aplac.org.
[7] http://ec.europa.eu/enterprise/sectors/electrical/emc.
[8] Sevgi, L., and R. Vahldieck, special issue on “From Design to Market of Electronic Sys-
tems: EMC Engineering,” Turkish Journal of Electrical Engineering and Computer Sciences,
Vol. 17, No. 3, Nov. 2009, http://mistug.tubitak.gov.tr.
[9] Sevgi, L., “EMC Testing Centers and Labs: Problems Related to CE Marking,” EMC Zur-
ich 2009, 20th International Zurich Symposium on EMC, Jan. 12–16, 2009, Switzerland.
[10] Sevgi, L., “From Design to Market EMC,” IBCAST 2011, 8th International Bhurban
Conference on Applied Sciences and Technology, Jan. 13–16, 2011, Islamabad, Pakistan.
[11] Eser, S., and L. Sevgi, “Open Area Test Site (OATS) Calibration,” IEEE Antennas and
Propagation Magazine, Vol. 52, No. 3, Jun. 2010, pp. 204–212.
[12] Sevgi, L., S. Çakır, and G. Çakır, “Antenna Calibration for EMC Tests and Measurements,”
IEEE Antennas and Propagation Magazine, Vol. 50, No. 3, Jun. 2008, pp. 215–224.
[13] Çakır, S., R. Hamid, and L. Sevgi, “Loop Antenna Calibration,” IEEE Antennas and
Propagation Magazine, Vol. 53, No. 4, Oct. 2011, pp. 243–254.
[14] ISO/IEC Guide 43-1:1997 Proficiency Testing by Interlab Comparisons—Part 1:
Development and Operation of Proficiency Testing Schemes (revised by 17043:2010).
[15] The Expression of Uncertainty and Confidence in Measurement, UKAS Publication M3003,
First Edition, Dec. 1997.
[16] Sevgi, L., “Innumeracy: The Meaning of the Numbers We Use,” IEEE Antennas and
Propagation Magazine, Vol. 49, No. 2, Apr. 2007, pp. 195–190.
[17] Sevgi, L., “Hypothesis Testing and Decision Making: Constant-False-Alarm Rate,” IEEE
Antennas and Propagation Magazine, Vol. 51, No. 3, Jun. 2009, pp. 218–224.
[18] TSE Proficiency Test, http://global.tse.org.tr.
Bibliography
http://ec.europa.eu/enterprise/policies/single-market-goods/cemarking.
http://ec.europa.eu/enterprise/sectors/electrical/lvd.
http://ec.europa.eu/enterprise/sectors/toys.
3
Electromagnetic Model
EMC engineering necessitates a different look at two classical lectures, EM fields
and waves, taught to undergraduates. EM theory is briefly revisited through
they eyes of an EMC engineer. For further reading, a few suggested classical
books on EM are [1–5].
3.1 Basic Electrical Engineering Theories

Two basic theories/models are used in electrical engineering: the EM model and
circuit model [1–5]. The EM model deals with electric and magnetic fields as
independent quantities/parameters; voltages and currents are used instead in cir-
cuit model. The two models are dependent. Any problem modeled and solved
in one model can also be handled with the other. A problem may be easier to
handle in one model by nature, but still can be solved by the other. For example,
the EM model fits better with an antenna problem if its radiation pattern is of
interest, but certainly the circuit model is better if the input power or impedance
matching is desired. Similarly, the circuit model is better for a microstrip struc-
ture if its input/output characteristics are of interest; but EM theory is easier if
the unintentional emissions of this structure are going to be modeled.
Fundamental physical explanations/definitions behind integro-differen-
tial Maxwell equations are:
• Stationary electric charges generate stationary electric fields around.

• Static magnetic fields are formed around constant moving electric
charges (i.e., DC currents).
• Time-varying moving charges (i.e., AC currents) generate both electric
and magnetic fields (i.e., EM waves).
51
• EM waves interact with objects. Surface currents are generated during
this interaction, and objects radiate like secondary EM sources.
• Materials with nonzero total magnetic moments of their atoms (like a
magnet) generate magnetic fields around.
EM fields are related to the movement of electric charges.
3.2 Maxwell Equations

EM theory is well established with Maxwell equations (see Table 3.1). Eigh-
teenth century physicist James Clerk Maxwell wrote these fundamental equa-
tions using Gauss, Ampere, and Faraday laws. Gauss's law is related to the total
flux of electric and magnetic fields. Ampere's law explains magnetic field circu-
lation around currents. Faraday's law deals with electric fields around charges.
Maxwell combined all of these and delivers his integro-differential equations
which models electrostatics, magnetostatics, and electrodynamics.
Maxwell equations, together with the Lorentz force law, well-establish
classical electrodynamics, classical optics, and circuit theory. As shown in Table
3.1 they have both integral and differential forms. The two operators, Rot (∇×)
and Div (∇⋅), correspond to spatial differential operators in vector and scalar
forms, respectively.
 Geometric meanings of these operators are pictured in Fig-
ure 3.1. If A is a vector function, then ∇ × A = 0 and ∇ ⋅ A = 0 correspond to
rotation-free and divergence-free vector functions, respectively.
Maxwell equations tell that spatial variation of electric field is related to
the time variation of magnetic field or vice versa. They are dependent. EM
properties of a medium (e.g., air, a conducting wire, wood, glass, and human
tissue) can be characterized in terms of three parameters: permittivity (dielectric
Table 3.1
Maxwell Equations and EM Model
Differential Form Integral Form
  
∇ ⋅D = ρ

∫ Dds = ∫ pdv
S V
∇⋅B = 0  


∂B 
∫ Bds = 0
S
∇×E = − +J 
∂t ∂  
    
D = εE , B = µH , J = σE
∫ Edl ∂t ∫S
=− Bds
C
     
∫ Hdl = ∫ Dds + ∫ Jds
C S S
Electromagnetic Model 53
Figure 3.1 Geometric meaning of Rot and Div operators.
constant) (ε) in [F/m], permeability (magnetic constant) (µ) in [H/m], and

conductivity (σ) in [S/m]. The medium parameters of the free-space are ε0=10–
9/36π [F/m], µ =4π10–7 [H/m], and σ=0 [S/m]. Usually, ε is used as a refer-
0 0
ence permittivity value; any other medium is referred to this value in terms of
relative permittivity ε=εrε0. A medium with εr=2, for example, has a permit-
tivity of ε=2ε0 [F/m]. The permeability and conductivity of copper (µcu=1.0
[H/m] and σcu=5.8107 [S/m], respectively) are used as the reference values
therefore a medium with  µr=2 and σr=3 and correspond to µ=2µcu and σ=3σcu
real values. The vector J and scalar ρ represent EM sources. The quantities and
their units are listed in Table 3.2.
 In summary, in EM theory, E and H represent independent variables,
J and ρ represent EM sources that generate electric and magnetic fields, and
ε/µ/σ represent electrical parameters of the medium.
Electric field is the force act upon a unit charge. The direction of this
force is shown by an arrow from a positive charge to a negative charge. In other
words, electric field lines are shown with arrows from a high potential point to
a low potential point.
 Magnetic
 
field is the field that acts as a force on a moving charge given
by F = q (v × B ) in this field. Magnetic field lines are loops enclosing a current.
3.3 EM Scattering, Diffraction, and Propagation

EM waves interact with objects and scatter [4, 5]. The word scattering is used
for all wave phenomena except the initial wave (i.e., total field = initial field
Table 3.2
EM Quantities and Their Units
Symbol Quantity Unit
E Electric field Volt/m
H Magnetic field Amper/m
J Surface current density Amper/m2
ρ Volume charge density Coulomb/m3
ε Medium permittivity Farad/m
σ Medium conductivity Siemens/m
µ Medium permeability Henry/m
+ scattered field). EM reflection and refraction are called primary scattering

components. EM edge and tip diffraction are called secondary scattering com-
ponents. Figure 3.2 shows two-dimensional (2D) variations of various wave
components generated by a sinusoidal line source. As observed, a line source
generates cylindrical waves that interact with nonpenetrable ground (bottom)
as well as with the nonpenetrable rectangular object. Since the objects are non-
penetrable, there is no refraction in this scenario [5]. In region 1, only outgo-
ing cylindrical waves are observed. The interaction of the direct and ground-
reflected waves is observed in region 2. Region 3 shows multiple reflections.
Finally, only tip diffracted waves appear in region 4 (i.e., in shadow region).
The right corner of the object acts like a secondary line source in this scenario.
Although there are two corners on top of the object, diffractions only from the
right corner is visible. Diffracted waves from the left corner are masked by the
incident and reflected waves in this region. Primary waves are much stronger
than secondary waves.
Figure 3.2 EM reflection and diffraction.

Cylindrical waves far from a line source may be assumed planar and are
called plane waves. The plane wave is the simplest wave object that can easily
be modeled mathematically. Both reflection and refraction occur when an EM
wave hits a penetrable surface. Reflection and refraction depend on the electri-
cal parameters of the surface as well as the polarization of the source.
In free space, EM waves propagate with the speed of light (c=3108 m/s).
The wavelength (λ) of the wave in free space is obtained from λ=c/f, where f is
the frequency of the wave. For example, at 10 GHz the wavelength is λ=3 cm.
At 1 GHz it is λ=30 cm. At 100 MHz it is λ=3m. A practical equation for the
wavelength is λ=300/fMHz [m].
3.3.1 EM Point and Line Sources

The first two Maxwell equations are first-order differential equations that in-
clude electric and magnetic fields at the same time. These forms are called first
-order coupled equations. These two can be combined and either of electric/
magnetic fields can be eliminated. This yields a second-order decoupled equa-
tion called wave equation. In the frequency domain (∂/∂t → jω) source-free
wave equation for both field components (ψ=E or H) is then given as
 ∂2 ∂2 ∂2 2
 ∂x 2 ∂y 2 ∂z 2 + k  ψ = 0
+ + (3.1)
 
The source term in the wave equation is given on the right-hand side.
Source-free means either the sources are very far, or that they are first applied,
the fields are excited, and then are removed. Source-free wave equation (3.1) is
used to investigate the environment. The solutions are called eigenfunctions, and
characteristic values that satisfy eigenfunctions are called eigenvalues. Therefore,
together with boundary/radiation conditions, (3.1) is called an eigenvalue
problem [6].
If a point source is located at the origin of Cartesian coordinate system,
(3.1) becomes
 ∂2 ∂2 ∂2 2
 ∂x 2 ∂y 2 ∂z 2 + k  g = − δ ( x ) δ ( y ) δ ( z )
+ + (3.2)
 
Together with the boundary/radiation conditions, (3.2) represent Green’s

function problem. Any source function f(x, y, z) may be used on the right-hand
side of (3.2), but, whatever it is, this function may be represented as a super-
position of point sources. Therefore, if (3.2) is solved for the point source (i.e.,
if the Green’s function is obtained), then the solution of any function would
be known. The Green’s function is the solution of (3.2) with the point source.
The Green’s function in EM theory corresponds to the impulse response in the
circuit theory.
Due to angular symmetry of the point source at the origin, it would be
better to deal with (3.2) in spherical coordinates. The full symmetry in both
horizontal and vertical angles the solution (i.e., the Green’s function would be
only radial distance dependent):
e − jkr e jkr
g= or g = (3.3)
4 πr 4 πr
These two, respectively, represent outgoing waves from the origin and
incoming waves toward the origin. The Green’s function (i.e., the solution of
the wave equation under the point source) decays with distance as 1/r and the
phase changes as (–jkr). Here, k = ω/c is the wave number).
If there is symmetry along one direction (e.g., along z), the wave equation
can be reduced to 2D (by assuming, ∂/∂z ≡ 0). In this case, the point source
reduces to a line source and can be given as δ(x)δ(y). The Green’s functions in
2D will then be:
e − jk ρ e jk ρ
g= or g = (3.4)
2π ρ 2π ρ
Here, ρ is the distance from the origin on the xy plane. As observed, waves
decay with one over square root of the distance (1 ρ) and the phase change
will be in the same form (exp(–jkρ)).
The Green’s function problem and the eigenvalue problem are not in-
dependent. One can be obtained from the other. This means EM behavior of
the structure at hand can be used to obtain the source effects, or vice versa.
The characteristic problem which connects source-driven and source-free solu-
tions (i.e., the Green’s function and the eigenfunction) is called Sturm-Liouville
problem [7, 8].
3.3.2 EM Wave Polarization

EM polarization is defined by the geometrical locus of the electric field vector
as it varies during one wave period at a fixed point in space. In the most general
case, EM waves are elliptically polarized. There are two limiting cases:
• Linear polarization: If the tip of the electric field vector moves along a
straight line, then the polarization is said to be linear. The two cases of
linear polarization are the vertical and horizontal polarizations. This is

best understood when propagation over the Earth’s surface is the case.
The polarization is said to be vertical (horizontal) when electric field vec-
tor is perpendicular (parallel) to the Earth’s surface.
• Circular polarization: When the locus of the tip of the electric field vec-
tor is a circle, then the polarization is said to be circular. Or, when verti-
cally and horizontally polarized waves are combined with a 90° phase
difference, the resultant wave will be circularly polarized. Circularly
polarized waves are mostly used in satellite communication, since they
can be generated and received using antennas that are oriented in any
direction.
3.3.3 EM Plane Waves and Wave Impedance

EM waves become plane waves in the far field. A plane wave is a nonphysical
wave where electric field, magnetic field, and direction of propagation are per-
pendicular to each other, and wave impedance (Z) is the ratio of the amplitudes
of electric and magnetic fields. In the free space, this value is Z0 = E/H = 377Ω.
Figure 3.3 shows a plane wave traveling along the z-direction.
3.3.4 EM Power Density

The energy
 of EM waves is carried with electric and magnetic fields. The power
density
   (in [W/m2]) is the vector product of electric and magnetic fields
P
P = E × H and called Poynting vector. The direction of the Poynting vector is
the direction of wave energy transport (see Figure 3.3).
3.3.5 EM Problem Groups

Figure 3.4 shows some different areas where EM modeling and simulation is
used. Different EM problems necessitate different modeling approaches. There
Figure 3.3 An EM Plane wave traveling along the z-direction.

Figure 3.4 Challenging EM problem groups.
are three characteristic problem groups in EM, and every area in Figure 3.4 falls
in one of these three groups.
• Antenna and radiation problems: The conventional way is to start with

Maxwell’s equations plus boundary and radiation conditions written in
terms of total EM fields; then, vector and scalar potentials are defined.
The wave equation is written in terms of these auxiliary functions and
then solved. Finally, electric and magnetic fields are obtained [5]. This is
because the vector and scalar potential functions are directly related with

the two source functions, J and ρ, respectively. The equations they need
to satisfy are in the form of (3.1)-(3.2).
• Scattering problems: The best way to approach these problems is to di-
vide total EM fields into incident and scattered fields, write down Max-
well’s equations plus boundary and radiation conditions, and solve for
the scattered fields.
• Guided wave problems: Cables (TL, coaxial lines, microstrip lines) and
waveguides guide EM waves. For these types of problems, it is best to
start with Maxwell’s equations plus boundary and radiation conditions
and then decompose total fields into transverse and longitudinal compo-
nents. Transverse and longitudinal equations will then yield eigenfunc-
tions and eigenvalues, respectively.
3.3.6 EM Propagation Modes

EM guided waves can be investigated under three different modes. All other
conditions can be represented with these forms. These modes are named ac-
cording to direction of propagation. Transverse electric (TE) mode has no lon-
gitudinal electric field component. Transverse magnetic (TM) mode has no lon-
gitudinal magnetic field component. Transverse electromagnetic (TEM) mode
has neither longitudinal electric nor longitudinal magnetic field components.
Plane waves are TEM modes. Note that there are slight difference in these
definitions in propagation and optics societies [6–8]. In groundwave propaga-
tion terminology, in addition to these, vertical and horizontal polarizations are
also used for TE and TM modes. Vertical (horizontal) wire antennas above the
ground produce vertically (horizontally) polarized waves.
It is essential to identify the direction of propagation (longitudinal direc-
tion) first in order to define TE/TM/TEM modes.
3.4 EM Materials and Skin Depth

Any object, a medium, or a material is electromagnetically represented in terms
of these three (electrical) parameters:
• Conductivity, σ, is the ratio of surface current density and electric field,

and is given as σ = J/E [S/m].
• Permittivity, ε, is the ratio of electric displacement and electric field
(e = D/E [F/m]).
• Permeability, µ, is the ratio of magnetic flux density and magnetic field
(µ = B/H [H/m]).
Depending on these parameters, there may be three different material

groups. Insulators (σ=0) are materials that contain free electrons and are as-
sumed as electromagnetically passive structures. Lossy materials (σ≠0, σ≠∞)
are materials that contain a finite number of free electrons. Magnetic materi-
als (µ≠µ0) are materials that have the ability to guide magnetic fields. Table
3.3 lists relative permittivity, relative permeability, and relative conductivity
of some well-known insulators. In Table 3.4, relative permeability and relative
conductivity of some conductive materials are given.
Communication engineers assume conductive materials like metals, such
as iron, copper, and steel, are nonpenetrable (σ→∞). This assumption always
works because they deal with EM scattering characteristics of these materials.
They don’t need to know how much EM fields penetrate into these materials.
On the other hand, this assumption is not true for EMC engineers. Metals are
Table 3.3
EM Materials and Their Electrical Parameters (σCu=5.8 107
S/m, µCu= 1.0 H/m )
Relative
permittivity Relative Relative
(wrt free- Permeability Conductivity
Material space) (wrt copper) (wrt copper)
Air 1.0005 0.0000 0.0000
Glass 5.0000 0.0000 0.0000
Nylon 3.5000 0.0000 0.0000
Teflon 2.1000 0.0000 0.0000
Bakelite 4.9000 0.0000 0.0000
Porcelain 6.0000 0.0000 0.0000
Silicon 12.0000 0.0000 0.0000
Table 3.4
EM Materials and Their Electrical Parameters
(σCu=5.8 107 S/m, µCu= 1.0 H/m)
Relative Relative
Conductivity Permeability
Material (wrt copper) (wrt copper)
Copper 1.0000 1.0000
Brass 0.260 1.0000
Gold 0.700 1.0000
Silver 1.0800 1.0000
Nickel 0.2200 1.0000
Aluminum 0.6100 1.0000
Zinc 0.2900 1.0000
Steel 0.1000 500.000
Stainless Steel 0.0200 1000.0000
Iron 0.1700 1000.0000
Permaloy 0.0300 80,000.0000
Mu-metal 0.0300 30,000.0000
not nonpenetrable. EM waves penetrate all materials. The penetration distance

depends on the conductivity and permeability of the material, and the wave fre-
quency. Skin/penetration depth (δ) is defined as the penetration distance inside
a material where the wave amplitude decays to 1/e (ω = 2πf ):
2
δ= [m] (3.5)
ωµσ
Skin depth is a critical parameter in EMC engineering, especially for EM

screening. It is also a critical parameter that determines AC/DC resistances of
cables.
3.5 Electric and Magnetic Dipoles

Any object may act as an antenna when interact with EM fields and/or when
AC current flows through. In order to mitigate EMI, one needs to identify the
type of the unintentional source. The source may look like an electric dipole or
a magnetic dipole. Figure 3.5 shows these two sources. Any AC current flowing
on a piece of conductor or an open-ended TL acts like electric dipole. Any AC
current flowing through a loop of wire or a short-circuited TL act as a magnetic
dipole. These two sources are totally different in the near field. Electric (mag-
netic) field is dominant around electric (magnetic) dipole in the near field. The
wave impedance of electric (magnetic) dipole is high (low) in the near field.
Electric (magnetic) dipole is a high voltage/low current (low voltage/high cur-
rent) element. In the far field, they both yield plane wave with Z0 = 377 Ω.
Figure 3.6 shows radiation characteristics of basic radiators. A point
source is a theoretical radiator with isotropic radiation characteristics. It gener-
ates spherical waves. In 2D, the theoretical radiator is the line source, and it
generates cylindrical waves. In practice, electric and magnetic dipoles are basic
radiators. They are also the basic emission sources in EMC engineering.
Figure 3.5 Electric and magnetic dipoles.

Figure 3.6 Basic radiators and their radiation patterns: (a) line source; (b) Poynting vector;
(c) vertical dipole; (d) horizontal dipole.
In EMC engineering it is critical to identify the type of the interference

source first. This is because the protection approaches are totally different in
the near field. In the far field, the type of interfering source is not important.
Emission from electric dipole type interfering source can be mitigates by
using electrical screening. For example, 15–25 dB isolation can be supplied by
using a thin aluminum folio. Unfortunately, this does not work if the interfer-
ing source looks like a magnetic dipole (e.g., electrical screening does not work
in the vicinity of a high power transformer; ferrite type filters are required).
EMC protection against electric and magnetic dipole type interference
sources is different.
3.6 Typical Emissions

As presented in Chapter 1, EMC thresholds in emission standards are in the
order of 40–80 dBµV for conducted emissions and 40–80 dBV/m for radiated
emissions. To get the feeling of the sources that can create such values, the ex-
ample shown in Figure 3.7 is given.
Here, radiated emissions from two simple sources at 3m are predicted.
One of the source is a 10-cm wire with AC current flowing as low as 1 µA. The
Figure 3.7 E/H dipoles, their emissions, and FCC limits.
other is a 0.5 cm × 5 cm rectangular wire loop with 1-mA AC current. Even

these small dipoles with highly low AC currents can create radiated emissions
close to the limiting values given in the standards.
A good EMC engineer needs to identify the type of interfering sources
first. This requires a good understanding of antennas, electric/magnetic dipoles,
TLs, microstrip lines, common/differential mode radiations, filtering, screen-
ing, grounding, gasketing, and so on.
3.7 EM Coupling Mechanisms

EM fields can penetrate solids, liquids, and gases and propagate inside. They
penetrate everywhere air, water, light, and sound penetrate. They also penetrate
into materials that these cannot penetrate. One reason for these penetrations/
emissions is apertures and/or holes. In order to reduce this, the sizes of the ap-
ertures and holes must be much smaller than EM signal wavelength (a rule of
thumb: they must be less than λ/20). The other reason is the skip depth, which
is especially important for mono-block materials. These two will be investigated
in detail in the Chapter 10 (Section 10.3) discussion of shielding effectiveness
(SE) modeling.
The coupling mechanisms that can create EMC problems are illustrated
in Figure 3.8. Basic coupling mechanisms are atmospheric coupling, induc-
tive coupling, capacitive coupling, and conductive coupling. Any AC current-
flowing material may act as an antenna and radiate. It creates RE problems.
Cable wires and conductive plates show resistance; therefore, they can create
conductive coupling problems. Surface irregularities, twists, and so on cause
Figure 3.8 Major EM coupling mechanisms.
leakage inductance and can create inductive coupling problems. Conductive

wires, plates, and so on close to each other cause leakage capacitor and can cre-
ate capacitive coupling problems.
In order to get rid of EMC problems (1) identify the type of the EMC
source, and (2) understand the mechanism of the coupling.
References
[1] Collin, R. E., Field Theory of Guided Waves, New York: McGraw-Hill, 1960.
[2] Jordan, E. C., and K. G. Balmain, Electromagnetic Waves and Radiating Systems, Engle-
wood Cliffs, NJ: Prentice-Hall, 1968.
[3] Dudley, D. G., Mathematical Foundations for Electromagnetic Theory, New York: IEEE
Press (Series on Electromagnetic Waves), 1994.
[4] Felsen, L. B., and N. Marcuvitz, Radiation and Scattering of Waves, Englewood Cliffs, NJ:
Prentice-Hall, 1973, Reissued IEEE Press, 1994.
[5] Sevgi, L., Electromagnetic Modeling and Simulation, New York: IEEE Press–John Wiley &
Sons, 2014.
[6] Sevgi, L., “Guided Waves and Transverse Fields: Transverse to What?” IEEE Antennas and
Propagation Magazine, Vol. 50, No. 6, Dec. 2008, pp. 221–225.
[7] Sevgi, L., “Guided Waves and Eigenvalue Extraction from Propagation Characteristics,”
IEEE Antennas and Propagation Magazine, Vol. 50, No. 2, Apr. 2008, pp. 222–234.
[8] Sevgi, L., “Sturm-Liouville Equation: The Bridge Between Eigenvalue and Green’s Func-
tion Problems,” ELEKTRIK, Turkish J. of Electrical Engineering and Computer Sciences,
Vol. 14, No. 2, 2006, pp. 293–311.
4
Circuit Model
One of the first technical lectures in electric, electronics, control, computer, and
communication engineering is circuit theory. Three important circuit elements
are resistor, inductor, and capacitor. Students learn definitions, their serial and
parallel connections, and functions in simple few-loop circuits, and they calcu-
late element currents (voltages) for given voltage (current) sources. For example,
resistance is defined as the ratio of voltage between its ends and current flowing
through and is the same for both DC and AC circuits. The impedance concept
is used to define inductor and capacitor in an AC circuit, which act as short-
circuited (SC) and open-circuited (OC) in DC circuits, respectively. Everything
works perfectly in circuit theory when definitions are used in ideal conditions.
However, EMC engineers have to go further and need to understand and model
these three elements and their circuits under realistic conditions. How does a
resistor/inductor/capacitor act at high frequencies? Is a resistor only a resistor
in practical circuits? How do EMC engineers define these elements in practice?
What is the difference between “lumped” and “scattered” element definitions?
4.1 Lumped Parameter Circuit Elements

Symbols and definitions of resistor, inductor, and capacitor are given in Figure
4.1 (ω=2πf is the angular frequency). Their impedances are also shown.
Typical ranges of practical resistances/inductances/capacitances are 100–
10 Ω (from kΩ to MΩ), 100–10–3 Henry (around mH), and 10-6–10–12 Farad
6
(from µF to pF). A resistor in DC is defined with Ohm’s Law (V = RI). It states

that a resistor is an element, current flowing on it is proportional with the ap-
plied voltage. In AC circuits, it becomes V = ZI, all of which are complex.
65
Figure 4.1 Basic circuit elements (V=ZI).
4.1.1 Conductor Wires

Although contributions of conductor wires are neglected in circuit theory, their
physical properties and effects are also important in EMC engineering. In the
RF range (i.e., typically, from several hundred kHz up to several GHz frequen-
cies), circuit behavior of a conductor wire is significant and depends on wire
cross section (e.g., its radius) and its length. Table 4.1 shows wire radii accord-
ing to American wire gauge (AWG) standards. As seen, the radius of a wire
doubles for every 6 AWG increase.
Electric current is the flow of charges inside conductors. In a DC circuit
(and for low frequencies), the charges flow through the whole cross section area.
As the frequency increases, the charges concentrate near the surface of the wire
just because of the skin depth (discussed in Chapter 3). As given there, skin/
penetration depth is given as δ = 2 ωµσ [m] and is inversely proportional
with the frequency. The skin/penetration depth becomes zero when the fre-
quency goes to infinity. For example, the skin depth of the copper (σ = 5.8107
[S/m], µ = 1.0 [A/m]) is 1 cm at 50 Hz; but becomes 7 µm at 1 MHz.
The decrease of the skin depth of a conductor wire increases its resistance.
This is illustrated in Figure 4.2. Here, charges flow through the cross-section
Table 4.1
AWG Standards (1 mil = 10-3 in = 2.54
10-3 cm)
Standard Diameter Diameter
Wire [mil] [mm]
AWG 50 1.0 0.0254
AWG 44 2.0 0.0508
AWG 38 4.0 0.1016
AWG 32 8.0 0.2032
AWG 26 16.0 0.4064
AWG 20 32.0 0.8128
AWG 14 64.0 1.6256
Circuit Model 67
Figure 4.2 DC and AC resistances of a circular cross-section wire.
in gray areas. The unit resistances of the wire for DC and AC circuits are also
given in the figure. As shown there, unit resistance is inversely proportional
with conductivity and the cross section. Since the area A1 is much less than A,
RAC will be much greater than RDC.
Figure 4.3 shows RAC/RDC versus frequency of copper and iron wires for
different wire cross sections. As observed, this ratio is nearly 20 for a circular
cross-section copper wire with 20-mm radius at 1 kHz, but it increases to 700
at 1 MHz.
EMC engineers must pay attention to skin depth even in very simple
circuits in order to prevent circuit failures.
Figure 4.3 Wire resistance versus frequency.

4.1.2 Inductive Effect of a Conductor Wire

A magnetic field is generated around a current flowing wire. A reaction voltage
will be induced between the terminals if this current is AC. This reaction is
called inductance. The inductance of a wire having length l and diameter d can
be approximated by
  4l 
L = 0.002 l 2.3 Log 10  − 0.75  (4.1)
 d 
For example, the inductance of 1.2-cm-long AWG 14 wire is approxi-

mately 8.6 nH. For a two-wire TL with diameter d and interwire distance (as
shown in Figure 4.4) the inductance is approximated by
bl
L=µ (4.2)
d
A resistor is not a resistor at high frequencies; it is a combination of resis-

tor and leakage inductor.
4.1.3 Capacitive Effect of a Conductor Wire

A potential difference occurs between charged parallel conductor wires/plates,
such as cables, connectors, and so on. This potential difference depends on the
amount of charges, the distance between conductors, and permittivity of the
material between conductors (see Figure 4.4). This element is called a capacitor,
and its capacitance is given by C = Q/V. The capacitor between two AWG wires
with diameter d and distance b can be calculated from
Figure 4.4 A two-wire TL model.

Circuit Model 69
dl (4.3)
C =ε
b
Equations (4.2) and (4.3) yield L = µ and C = ε when the length, diam-
eter, and distance are assumed 1m. This helps to bring different interpretations
for EM parameters permittivity and permeability. The inductive and capacitive
effects are related with magnetic and electric energy storage capacities, respec-
tively. It should be noted that models given in Figure 4.1 belong to ideal case;
in practice element models are quite different.
Here are few circuit tips in terms of EMC engineering:
• In addition to their resistive effects, conductors begin to act as an induc-

tor at frequencies above 10–15 kHz.
• As a rule of thumb, 7–8 nH/cm inductance effect should be taken into
account for conductor wires and microstrip lines. This corresponds to an
additional reactance of Z = j12 Ω at 1-MHz frequency.
• Wire-wounded inductors can be used up to 50-MHz frequencies. Reso-
nances maybe observed because of leakage capacitors, which can roughly
be calculated from fo = 200/L [MHz] where L is in µH.
• EMI ferrite inductors show resistive properties at high frequencies but
are useful in the 10 MHz–1 GHz frequency range.
• Typical unit inductances for surface mounted and leaded capacitors can
be taken, respectively, as L = 0.4 nH/cm and L = 8 nH/cm.
• Aluminium electrolytic, paper, mica, and ceramic capacitors can be used
up to 100 kHz, 5 MHz, 500 MHz, and 1 GHz frequencies, respectively.
4.1.4 Realistic R/L/C Models

None of the circuit elements acts as theoretically taught at high frequencies.
A resistor is no longer a resistor only; it is also the same for an inductor or a
capacitor.
A realistic resistor model for its high frequency behaviors is given in Figure
4.5. As observed, the model includes a serial inductor and a parallel capacitor.
Similarly, a realistic capacitor model for its high frequency behaviors is
given in Figure 4.6. Finally, Figure 4.7 presents a realistic model for the high fre-
quency behaviors of an inductor. Typical impedances versus frequency of these
three basic circuit elements are pictured in Figure 4.8. As observed, they all are
linear at low frequencies but certainly not linear at high frequencies.
Figure 4.5 A high frequency equivalent of a resistor (R).
Figure 4.6 A high frequency equivalent of a capacitor (C).
Figure 4.7 A high frequency equivalent of an inductance (L).
Figure 4.8 R/L/C elements and their realistic frequency behaviors.

Circuit Model 71
4.2 Two-Port Circuit Definitions

Most of the practical EMC problems belong to two-port circuits. Usually, the
input port is connected to the AC mains or a power supply. The output port
is connected to a refrigerator, washing machine, a TV receiver, a radio, an an-
tenna, an EMI receiver, and so on, which can simply be called load. The source
and load may directly be connected or an auxiliary circuit, like an attenuator,
filter, or an impedance matcher, may be used in between. Figure 4.9 pictures
this generic scenario. The generic source is modeled with a voltage source (Vs)
and a source resistance (Rs). The load is represented with a load impedance (ZL)
or resistance (RL). The two-port circuit is used in between.
Low and high frequency treatments in circuit theory are different. At low
frequencies (i.e., for frequencies where wavelength is very large compared with
the circuit elements’ sizes) lumped element parameters are used. Here, resis-
tor, inductor, and capacitor are the basic circuit elements. All other connecting
wires, cables, connectors, and so on, are neglected. At high frequencies when
the wavelength and sizes are comparable, scattered parameters are used. Every
element, every geometry and discontinuity in the circuit may be modeled as a
TL and acts as an antenna.
Impedance (Z = V/I; V1 = Z11I1 + Z12I2; V2 = Z21I1 + Z22I2) and admit-
tance (Y = I/V; I1 = Y11V1 +Y12V2; I2 = Y21V1 + Y22V2) parameters are used in
lumped element model. These parameters are called OC and SC parameters,
respectively (e.g., Z11 = V1/I1 when I2 = 0, the output port is OC). Scattered
parameters (S) are used at high frequencies.
The following two quantities are widely used in two-port circuit analysis:
insertion loss (IL) and reflection loss (RL). IL is defined at the output port
in terms of output (load) voltages VL1 and VL2, where VL1 is the load voltage
when the two-port circuit under investigation is inserted between the source
Figure 4.9 Two-port circuit definitions.

and load, and VL2 is the load voltage when the source is directly connected to
the load. IL in dB is then given as ILdB=20log10|VL2/VL1|. IL is negative with
the maximum value of 0 dB, which corresponds to VL2 = VL1. It should be
noted that the definition of IL may also be given as ILdB=20log10|VL1/VL2|. In
this case, IL would be positive with the minimum value of 0 dB. RL is defined
at the input port in terms of incident (V +) and reflected (V –) voltage waves as
RLdB=20log10| V –/V +|. The two-port circuit’s performance is then given as IL
versus frequency or RL versus frequency curves. Note also that voltage transfer
function (VTFdB=20log10|V2/V1|) is also used to show the performance of a
two-port circuit. Actually, the VTF curve is simply a 6-dB shifted (below) IL
curve.
Calculation of both IL and RL necessitates input impedance (Zin) and
input voltage (Vin). As shown in Figure 4.10, the input impedance (Zin = ZC in
this case) may be obtained, from load to source, by serial/parallel reduction of
the circuit impedances. Then, input voltage can directly be calculated from a
single loop circuit with the equivalent input impedance. Once the source cur-
rent is obtained, the currents through all branches, from source to load, can eas-
ily be obtained. This will yield all required voltages for IL and RL calculations.
In Figure 4.10, a π-circuit with three impedances is used as a two-port
example. The impedance Za is the parallel combination of RL and Z3; Zb is the
serial combination of Za and Z2; the input impedance Zc is the parallel com-
bination of Zb and Z1. Once they are calculated, currents flowing through all
branches will easily be obtained as shown in the figure. Table 4.2 lists a short
MATLAB code for the calculation of IL versus frequency of the π-circuit shown
in Figure 4.10. A typical result obtained with this code is shown in Figure 4.11.
In Figure 4.11, both IL and VTF versus frequency variations are plotted.
Z1 is a parallel resonance circuit with L1 = 0.949 nH and C1=11.867 pF. Z3 is
also a parallel resonance circuit with L3 = 0.964 nH and C3 =11.676 pF. Z2 is a
Figure 4.10 Current distribution on a two-port circuit and equivalent impedance.

Circuit Model 73
Table 4.2
Short MATLAB Code for the IL Calculations of a π-Circuit
%--------------------------------------------------------------------------
% Program: IL versus Frequency calculations for a PI circuit
%--------------------------------------------------------------------------
clc; clear all; rs=50; rl=50; vs=100; nn=500;
fmin = 1; fmax = 3e3; df=(fmax-fmin)/nn; % Frequency range
% Impedance values for L/C elements)
l1 = 0.949e-9; c1 = 11.867e-12; % Z1 parallel resonance circuit
l2 = 20.372e-9; c2 = 0.552e-12; % Z2 serial resonance circuit
l3 = 0.964e-9; c3 = 11.676e-12; % Z3 parallel resonance circuit
for k=1:nn
f(k) = (fmin+(k-1)*df); omega = 2*pi*f(k)*1e6;
zl1 = 1i*omega*l1; zc1 =-1i/(omega*c1); Z1=(zl1*zc1)/(zl1+zc1);
zl2 = 1i*omega*l2; zc2 =-1i/(omega*c2); Z2=zl2+zc2;
zl3 = 1i*omega*l3; zc3 =-1i/(omega*c3); Z3=(zl3*zc3)/(zl3+zc3);
za=(rl*Z3)/(rl+Z3); zb=Z2+za; zc=(zb*Z1)/(zb+Z1);
I1=vs/(rs+zc); I1B=I1/(1+zb/Z1); IY=I1B/(1+rl/Z3);
vl=IY*rl; vtf(k)=vl/vs; IL(k)=2.*vl/vs;
end
plot(f,20*log10(abs(vtf)),’b’,f,20*log10(abs(IL)),’r--’); ylim([-50 5]);
legend(‘GTF’, ‘AGK’); title(‘2-port behavior’)
ylabel(‘IL [dB]’); xlabel(‘Frequency [MHz]’) % Program End
Figure 4.11 IL versus frequency of the sample π-circuit.
serial resonance circuit with L2 = 20.372 nH and C2 = 0.552 pF. These values
belong to a three-element band pass filter. As observed, IL is 0 dB in the pass
band (i.e., between 1.3 and 1.8 GHz).
4.3 Resonance Circuits

Resonance circuits are used for frequency (channel) selectivity in communi-
cation engineering. In EMC engineering, they are used for both design and
protection. They filter out undesired emission. Figure 4.12 shows serial and
parallel resonance circuits. The important properties of the resonance circuits
are as follows:
• Resonance frequencies of both serial and parallel resonance circuits are
1
f0 = (4.4)
2 π LC
• Impedance characteristics at their terminals are just the opposite.

• In an ideal case, the impedance of a serial resonance circuit at the reso-
nance frequency is zero, and it acts as a SC circuit. In practice, because
of losses and leakage effects, the impedance is minima around the reso-
nance frequency.
• In an ideal case, the impedance of a parallel resonance circuit at the
resonance frequency is infinite and it acts as OC circuit. In practice,
because of losses and leakage effects, the impedance is maxima around
the resonance frequency.
• In a two-port circuit, a serial resonance circuit inserted serially between
input and output show minimum impedance (SC) and therefore allows
maximum power transfer at its resonance frequency. On the other hand,
if it is connected in parallel, it totally blocks the transfer between input
and output ports.
Figure 4.12 Theoretical serial and parallel resonance circuits.

Circuit Model 75
• Similarly, in a two-port circuit, a parallel resonance circuit inserted seri-

ally between input and output show maximum impedance (OC) and
therefore blocks power transfer at its resonance frequency. On the other
hand, if it is connected in parallel, it allows maximum power transfer
between input and output ports.
• Serial and parallel resonance circuits are used for bandpass and bandstop
filters.
Transfer characteristics of a simple inductor and a capacitor, and serial/

parallel resonance circuits are illustrated in Figure 4.13.
Some remarks are as follows:
• Lowpass and highpass filters can be realized using only an inductor and/
or a capacitor. Bandpass and bandstop filters necessitate serial and/or
parallel resonance circuits.
• The inductor is SC for DC signals but acts OC at high frequencies.
On the contrary, the capacitor is SC for DC signals and acts SC at high
frequencies.
• An inductor (capacitor), connected in serial between input and output,
acts as a low (high) pass filter.
• An inductor (capacitor), connected in parallel between input and out-
put, acts as a high (low) pass filter.
Figure 4.13 Transfer characteristics of the L/C elements.

• A serial (parallel) resonance circuit, connected in serial between input

and output, acts as a bandpass (bandstop) filter.
• A serial (parallel) resonance circuit, connected in parallel between input
and output, acts as a bandstop (bandpass) filter.
Worth mentioning are t = 0 and t → ∞ behaviors of an inductor and ca-

pacitor. An inductor acts as OC (SC) when t = 0 (t → ∞). In contrast, a capaci-
tor acts as SC (OC) when t = 0 (t → ∞).
In practice, resonance circuits have losses. The two other parameters im-
portant in resonance circuits are as follows:
• Bandwidth: The 3-dB frequency band of a resonance circuit around its

resonance frequency.
• Quality factor (Q): The ratio of the center (resonance) frequency and
bandwidth as
fm
Q= ; B = fU − f L (4.5)
B
is the quality factor and show the selectivity of a resonance circuit. High-
Q means better frequency selectivity.
• Shape factor: The ratio of 60-dB and 3-dB bandwidths is the shape fac-
tor of a resonance circuit. It shows the degree of the amplitude decay be-
tween pass and stop bands. In ideal filters, the shape factor is equal to 1.
Example: A selective filter is requested for a TV broadcasting at long wave

(198 kHz). Design a parallel resonance circuit if the antenna radiation resis-
tance is 300 Ω and bandwidth is 5 kHz.
Solution: The quality factor Q is found 198 × 10 3/(2 × 510 3) = 19.8. The
inductance L = R/(ω0 Q) is then equal to 12.1 µH. Finally, the capacitor will be
C = 1/(ω02L )=53 nF.
Note that the quality factor of a resonance circuit may be defined in two
different ways: loaded and unloaded. Definition (4.5) belongs to a loaded case.
The quality factor of a resonance circuit depends on source and load imped-
ances as well as element quality factors. The quality factor of an inductor is the
ratio of its inductance value and its series loss resistance. The same definition
also holds for a capacitor.
Circuit Model 77
4.4 Cables and Transmission Line Model

Communication/data lines, cables, connectors, power lines, and AC mains
cause important EMC problems. They are the coupling means between EMI
sources and victims. In order to solve such problems and remove coupling
paths, an EMC engineer needs to understand circuit theory and transmission
line (TL) theory.
A TL is more than a set of long, parallel lines; it is a distributed-parameter
physical system. TLs are used to transmit electric energy and communication
signals from one point to the other. The TL connects a source to a load. This
may be a transmitter and an antenna, a television or a radio antenna and a receiv-
er, one port of a coupler and a power meter, and so on. A two-wire TL model,
together with characteristic parameters and quantities, is pictured in Figure 4.14.
In this model, source and load may be a power supply and a broadcast
antenna, or AC mains and a refrigerator, or a receive antenna and an EMI
receiver. Whatever they are doesn’t matter; they can all be simplified and be
modeled with TL theory. The connection may be a two-wire line, a coaxial line,
or a waveguide. Two-wire lines act as low pass filters with cut-off frequencies
around several hundred MHz. Coaxial cables also act as low pass filters with
cut-off frequencies up to several GHz. Waveguides are different; they act as high
pass filters. Microstripline family and optical waveguides may also be modeled
with TL theory.
At high frequencies voltage and current propagate as waves along the TL.
In other words, voltage waves and current waves are of interest.
Both electric and magnetic fields are present around a TL. These fields are
perpendicular to each other and to the direction of wave propagation for TEM
mode waves, which is the simplest mode, and assumed for most simulators
Figure 4.14 A TL model and its mathematical relations.

(except for microstrip lines which assume quasi-TEM, which is an approxi-

mated equivalent for transient response calculations).
Electric field is established by a potential difference between two conduc-
tors (implies equivalent circuit model must contain capacitor). Magnetic field is
induced by current flowing on the line (implies equivalent circuit model must
contain inductor).
There is an exact analogy between TL theory and plane wave representa-
tion. They are both modeled with the two first-order, coupled differential equa-
tions given in the figure. These are the characteristic equations of TL theory.
The independent variable in this model are v(z, t) and i(z, t). The voltage and
current along a TL vary with time and position. The R/L/C/G shown in these
two equations are called primary TL parameters:
• R: Unit-length resistance [Ω/m];

• L: Unit-length inductance [H/m];
• C: Unit-length capacitance [F/m];
• G: Unit-length Admittance [S/m].
Here, R and G represent TL losses; L and C represent energy storage. In

practice, it is difficult to use these parameters; they are either producer supplied
or measured with special equipment. Therefore, easily measurable secondary pa-
rameters, characteristic impedance and propagation constant, are used:
• Z0: TL characteristic impedance [Ω];

• Γ = α + jβ: TL complex propagation constant.
4.4.1 Characteristic Impedance

Characteristic impedance is an important parameter in TL theory. It is a pa-
rameter related to the primary line constants and is real for an ideal (loss-free)
TL. Its definition can be stated as (a) the ratio of voltage divided by current at
any point along an infinite-length homogeneous TL; (b) the ratio of voltage
divided by current at any point along a reflection-free homogeneous TL. It can
be calculated from the primary parameters:
R + j ωL
Z0 = [Ω] (4.6)
G + j ωC
Circuit Model 79
Optimum characteristic impedance of a coaxial cable for the best power

transfer seems to be around 30Ω. On the other hand, the lowest attenuation
along the TL is around 77Ω. Therefore, characteristic impedance of a coaxial
TL is chosen as 50Ω.
4.4.2 Propagation Constant

Complex propagation constant g of the TL represents attenuation and phase
variation along the line. The real part of the propagation constant α [dB/km]
shows attenuation per meter or km. The imaginary part of the propagation
constant β [rad/km] show phase variation per meter or km. It can also directly
be calculated from the primary parameters:
g = α + j β = (R + j ωL )(G + j ωC )

1/2
(4.7)
Characteristic impedance and propagation constant of a lossless TL (i.e.,

R = G = 0, and α = 0) reduce to Z 0 = L / C and g = j β = j ω LC . In prac-
tice, TLs (except very long ones such as power lines, data links, and so on) are
assumed lossless; they only cause reflections and phase shifts.
Field distributions around TLs are critical in EMC engineering. Figure
4.15 pictures field distributions around various TLs. Two-wire lines and mi-
crostriplines are EMI susceptible. Coaxial cables and waveguides are naturally
screened; one only needs to pay attention to their connections and grounding.
Figure 4.15 Various TLs and field distributions.

4.4.3 Voltage Reflection Coefficient

Voltage and current waves propagating along a TL toward a load that is dif-
ferent than the characteristic impedance of the line partially reflect backward.
Voltage reflection coefficient depends on the difference between line impedance
and the load impedance. In other words, any discontinuity along a TL causes
voltage reflections. Incident voltage wave (V +) interferes with the reflected volt-
age wave (V –), and voltage maxima (if they have equal phases) and minima (if
they have opposite phases) occur along the line. The current at voltage maxima
(minima) will then be minima (maxima). Voltage reflection coefficient (Γ or ρ)
is defined as the ratio of reflected and incident voltage waves:
−
V Z − Z0
Γ = ρ= = (4.8)
V + Z + Z0
Reflection is proportional with the degree of the discontinuities. Voltage

reflection coefficient is a complex number with a modulus 0 ≤ |Γ| ≤ 1. The
extreme cases are as follows:
• When ZL = Z0, reflection coefficient is zero (Γ = 0). The power along the
line totally transferred to the load. This is called matched TL.
• When ZL = 0 (SC), reflection coefficient is minus one (Γ = –1). In this
case, the power along the line totally reflects back (with a phase shift of
180°).
• When ZL →∞ (OC), reflection coefficient is one (Γ = 1). In this case,
the power along the line totally reflects back (without phase shift).
The voltage is infinite and current is zero at the OC end of a TL. On

contrary, the current is infinite and voltage is zero at the SC end of a TL. Every-
thing (i.e., voltage, current, impedance) along a TL is periodic with the period
of half wavelength (λ/2).
A typical two-port is pictured in Figure 4.16. Here, circuit A is connected
to circuit B over a piece of TL. There are two discontinuities/junctions in this
circuit. The first discontinuity is at the output of circuit A, between output
impedance (Zout) of circuit A and equivalent impedance (Zeq) seen at that junc-
tion. Note that the equivalent impedance at this junction depends on every-
thing on the right (including the cable and circuit B). The second discontinuity
is at the input of circuit B, between the characteristic impedance of the cable
(Z0) and input impedance (Zin) of circuit B.
Circuit Model 81
Figure 4.16 Impedance mismatch and reflections on a two-port circuit.
Impedance mismatch means voltage reflections. It also means part of the

power delivered by the source will be reflected. This explains why impedance
matching is critical in circuits.
Power delivered by the source is totally transferred to the load in a matched
circuit.
The reflection coefficient is defined in terms of voltages. Power is pro-
portional with the square of the voltage; therefore, the square of the reflection
coefficient is related to the power. If Pin is the input power, then reflected and
transmitted powers will be:
• Reflected power: Pref = |ρ|2 Pin [W].

• Transmitted power: Pload = (1-|ρ|2) Pin [W].
4.4.4 Voltage Standing Wave Ratio (VSWR)

Impedance mismatch generates incident and reflected voltage waves along a
TL, and this in turn causes voltage standing waves. VSWR can be expressed in
terms of reflection coefficient:
V max 1+ | ρ |
s= =
V min 1− | ρ | (4.9)
VWSR is a real number between 1 and ∞:
• VSWR = 1 when there is no impedance discontinuity along the line

(matched line).
• VSWR → ∞ when the line is terminated by OC or SC. Voltage fluctua-
tion along the line is maxima and the power is totally reflected back to
the source.
Figure 4.17 presents input impedance of a piece of TL with characteristic

impedance Z0, under both OC and SC terminations. The input impedance
equation with ZL termination is also given. As observed, the input impedance
is complex, but becomes pure imaginary (inductive or capacitive, depending on
TL length) under OC/SC terminations (i.e., when ZL→∞).
The horizontal axis is the frequency for a constant length piece of TL. It
also represents the length (l ) of the TL if the frequency is fixed. The following
observations are important for EMC engineers:
• The input impedance of any length OC/SC TL is pure imaginary—ei-

ther capacitive or inductive.
• Any inductor or capacitor can be realized with a piece of OC/SC TL by
adjusting its length.
• OC and SC terminated TLs show the same electrical properties with
quarter wavelength (λ/4) shift. That is, a capacitor or an inductor real-
Figure 4.17 Effects of short- and open-circuit terminations along TLs.

Circuit Model 83
ized with a SC TL with length l can also be realized with OC line with
length l + λ/4.
• Electrical characteristics on a TL are periodic with half wavelength (λ/2).
OC and SC terminated TLs are used (1) in the realization of inductors

and capacitor on PCBs, and (2) as stubs in impedance matching circuits.
4.5 Grounding
The voltage (potential difference) is the work capability between any two
points. In electrical engineering, voltage is the potential difference between any
two points. For example, VAB is the difference between potentials of the points
A and B. But, the potential of a single point (e.g., VA) is meaningful if a refer-
ence (zero potential) point is specified. This is called ground.
The classical definition of a ground is “an equipotential point or plane
which serves as a reference for a circuit or system” [12]. Unfortunately, this
definition is meaningless in the presence of ground current flow. Even where
signal currents are negligible, induced ground currents due to environmental
magnetic or electric fields will cause shifts in ground potential. An alternative
definition for a ground is “a low impedance path by which current can return
to its source” [12]. This emphasizes current flow and the consequent need for
low impedance and is more appropriate when high frequencies are involved. It
is important to remember that two physically separate “ground” points are not
at the same potential unless no current is flowing between them.
According to IEEE Std 142-2007, the definition of “ground” is “a con-
ducting connection, whether intentional or accidental, between an electrical
circuit or equipment and Earth, or to some other body that serves in place of
Earth.”
Grounding is one of the primary ways of minimizing unwanted noise and
of producing a safe system. That said, a noise-free system is not necessarily a safe
system, and, conversely, a safe system is not necessarily a noise-free system. It is
the responsibility of the designer to provide both a safe and a noise-free system.
One advantage of a well-designed ground system is that it can often provide
protection against unwanted interference and emission, without any additional
per-unit cost to the product.
Grounds fall into two categories: (1) safety grounds and (2) signal grounds.
The second category probably should not be called grounds at all, but rather
returns, and it could be further subdivided into either signal or power returns.
If a ground is connected to the enclosure or chassis of the equipment, then it
is often called a chassis ground. If a ground is connected to Earth through a

low-impedance path, then it may be called an Earth ground.
A typical EMC problem showing the importance of the ground is pic-
tured in Figure 4.18. Suppose, there are two installations (circuits A and B)
15m apart and a single point grounding is applied. The grounding conductor
of circuit B would be short, but 15m–20m grounding conductor would be re-
quired for circuit A. This will not create a serious problem for the power supply
(i.e., connection to the AC mains).
Suppose lightning hits. It may discharge tens of kA currents withing mi-
croseconds (di/dt ∼kA/µs). Remember, the typical leakage inductance of a con-
ducting wire is 8–10 nH/cm. Because of this, it may induce up to 150–200
kV potential difference between circuits A and B, which may destroy many
components of both circuits.
A major grounding problem is that Earth’s impedance is not equal every-
where! Because of this, potential differences may occur for different grounding
points. Figure 4.19 depicts a scenario showing problems occuring because of
nonideal grounding. In ideal case, points A and B on the ground are equipo-
tential and equal to zero. The impedance between these two points is zero. In
practice, there is a potential difference between these points. This potential
difference is usually too small to create a problem, but when a lightning hits
imperfect grounding may create a severe damage.
The input impedance of a piece of grounding wire is different at low,
medium, and high frequencies. For a ground wire with length l, cross section
Figure 4.18 Typical effects of a lightning.

Circuit Model 85
Figure 4.19 Grounding effects.
radius r, area A, distance to ground h, these impedances are Zin = l /Aσ at low
frequencies, Zin = jωL at medium frequencies, and Zin = jZ0tan(2πl /λ) at high
frequencies (the wire and ground form a kind of TL and its characteristic im-
pedance is Z 0 = 60 µr εr ln (2h r )). This input impedance becomes infinite
(zero) at frequencies when l = λ/4 (l = λ/2).
At low frequencies, a DC ground resistance gives a clue about ground im-
pedance, and single point grounding may be applied. The length of the ground-
ing conductor should be much less than a wavelength (as a rule of thumb, l<
λ/10). On the other hand, when the frequency is high, scattered inductance and
leakage capacitance effects create grounding problems.
Don’t forget: A nonzero (even as small as 0.3 Ω–0.5 Ω), imperfect ground
impedance may be negligible under normal operations but can create several
tens of kV potential differences under a 200 kA lightning. This in turn causes
high discharge and damage equipment. Grounding is also important in protect-
ing personnel, animals, structures, and buildings from harm because of electri-
cal shock or fire.
Grounding is a major problem in PCB circuits. Here are a few tips for a
good grounding:
• Reduce inductance effects. In order to do this, either the leakage induc-

tance (L) is reduced or time variation of the current (di/dt) is minimized.
Board design and circuit/element locations are critical in reducing EMC
problems.
• Avoid closed loops and minimize their area. Closed loops act as mag-
netic dipoles and create strong near-magnetic fields.
• Keep connecting lines as short as possible. They act as electric dipoles
and create strong near-electric fields.
• Keep outgoing and return lines as near as possible and main signal lines
as close to reference as possible.
• Use surface-mounted elements when possible.
A few important grounding myths in EMC engineering [1, 2]:
• The Earth is usually thought of as a low-impedance path for ground

currents, but in reality the impedance of the Earth is orders of magni-
tude greater than the impedance of a copper conductor.
• The Earth is thought of as equipotential, which is not true.
• The impedance of a conductor is thought to be determined by its resis-
tance only. In reality, leakage inductance may be critical.
• It is also thought that in operating with low noise, a circuit or system
must be connected to an Earth ground. This is also not true, because air-
planes, satellites, cars, and battery-powered laptop computers all operate
fine without a ground connection. As a matter of fact, an Earth ground
is more likely to be the cause of a noise problem. More electronic system
noise problems are resolved by removing (or isolating) a circuit from
Earth ground than by connecting it to Earth ground.
• To reduce noise, an electronic system is thought to be connected to
a separate quiet ground by use of a separate, isolated ground rod. In
addition to being false, this approach is dangerous and violates safety
requirements.
• An Earth ground is not unidirectional; any current that flows into the
ground must also flow out of the ground somewhere else.
4.6 Common Mode and Differential Mode Currents

Grounding wire or plate is a major EMI path, which may create EMC prob-
lems. A good ground design prevents the occurrence of many EMC problems.
Remember, EMI sources, electric and magnetic dipoles, differ in terms of their
near-field EMC effects. Similarly, currents flowing on circuits also differ in
terms of their EMC effects. These currents are pictured in Figure 4.20.
Electric and magnetic dipoles in EM theory are equivalent to common
mode (CM) and differential mode (DM) currents in circuit theory.
In any circuit, both CM and DM currents are present. Both types of cur-
rent determine the amount of RF energy propagated between circuits or radi-
ated into free space. There is a significant difference between the two. Given a
pair of TLs and a return path (see Figure 4.20), one or the other mode will exist
(usually both).
DM currents belong to useful signals. They are the components of RF en-
ergy present in both signal and return paths, equal and opposite to each other,
Circuit Model 87
Figure 4.20 Differential (DM) and common mode (CM) currents.
carrying data or information. DM currents do not flow on ground paths. RF

DM currents will be canceled if a 180° phase shift is established precisely. They
convey the desired information and cause minimal interference as the fields
generated oppose each other and cancel out if properly set up.
CM current is the component of RF energy that is present in both signal
and return paths, often in common phase to each other. They are generated by
any imbalance (imperfect ground) in the circuit. Radiated emissions are the
result of such imbalance. CM signals are the major sources of cable and inter-
connect EMI and contain no useful information.
DM and CM currents, respectively, act as loop and wire antennas and
they emit. The emissions caused by DM and CM currents are proportional
with the frequency and the square of the frequency, respectively. This is pictured
in Figure 4.21.
Remember, a 1m-long two-wire cable with wires 1.2 mm apart, having
20-mA DM current, may generate (worst-case) 100-µV/m electric field at 30-
MHz frequency at 3m distance, and this exceeds the FCC class-B limits. The
same electric field, at the same frequency and distance, may be generated with
only 8-µA CM current.
4.7 Nonlinearity Effects

Nothing in nature is linear, but electrical engineers love to make this assump-
tion. Assuming that an element, device, or a circuit is linear makes it simple to
model and calculate analytically as well as numerically. This assumption holds
in many engineering problems. Unfortunately, EMC engineers do not have this
Figure 4.21 Worst-case emissions of DM and CM currents.
luxury. EMC deals with real-life problems, and real-life problems are not linear.
Moreover, most of the systems have nonlinear circuit elements like a mixer,
amplifier, attenuator, and so on. Even, for a single-frequency input, they create
undesired signals at several frequencies. The signals with integer multiples of
the input signal are called harmonics. If there are multiple frequency inputs,
nonlinear circuits also create internal products (IPs) at their outputs. Genera-
tions of harmonics and internal products are pictured in Figure 4.22. Nonlinear
circuits have transfer functions with high-order terms; therefore, for an input
signal f1, they create signals f2 = 2f1, f3 = 3f1, f4 = 4f1, ..., and so on. These har-
monics cause serious EMC problems.
Similarly, a mixer, for example, with f1 and f2 signals at the input, creates
both harmonic signals (i.e., 2f1, 3f1, 4f1, …, 2f2, 3f2, 4f2, …) and internal prod-
ucts (f1 ± f2, 2f1 ± f2, f1 ± 2f2, 3f1 ± f2, ..., mf1 ± nf2) where m and n are integers.
The order of the internal product is important and would be (m + n + k) for
the internal product (mf1 ± nf2 ± kf3). Third order internal products (TOIP) are
usually critical because they fall in the signal band.
Figure 4.23 shows harmonic and internal products of two signals A and B
with close frequencies. As observed, the two third order internal products 2A-B
and 2B-A are very close to the signals A and B with significant amplitudes. The
other harmonics and internal products can be filtered out easily, but these two
TOIPs necessitate sharp filters.
Circuit Model 89
Figure 4.22 Harmonics and internal products.
Figure 4.23 Third-order internal product (TOIP) formation.
4.8 Two-Port Circuits and S-Parameters

EMC engineers are aware of circuit analysis and synthesis methods taught dur-
ing their undergraduate education. They are well-known impedance and ad-
mittance parameters. As shown in Figure 4.24, impedance parameters relate

port voltages to port currents. The impedance parameters of this two-port cir-
cuit satisfy:
V1 = Z 11I 1 + Z 12 I 2
(4.10)
V 2 = Z 21I 1 + Z 22 I 2
Here, for example, Z11 is the input impedance extracted under a specific
condition (i.e., when the output port is OC). It is obtained from
V1
Z 11 = (I 2 = 0) (4.11)
I1
Impedance (Z) parameters are OC parameters.

Admittance (Y) parameters also relate port currents and voltages. Admit-
tance parameters are SC parameters. Note that port voltages and currents are
defined according to the references shown in the figure.
At low frequencies, it is easy to make a port OC or SC. Unfortunately,
everything radiates and acts as an antenna when the frequency is high. OC or
SC terminations are difficult to realize; therefore, scattering (S) parameters are
used. S-parameters relate port incident and reflected voltage waves. Total volt-
age at a port is the addition of incident (V +) and reflected (V –) voltage waves
(V = V + + V –). These two waves can be normalized with the characteristic im-
pedance of the port and be generalized. For an n-port circuit, normalized port
voltages are defined as:
Vi +
ai = , i = 1,2,3,, N (normalized incident voltage wave ) (4.12)
2 Z 0i
Vi −
bi = i = 1,2,3,, N (normalized reflected voltage wave ) (4.13)
2 Z 0i
Figure 4.24 Impedance (OC) parameters for a two-port circuit.

Circuit Model 91
where Z0 is the characteristic impedance of i-th port. Figure 4.25 pictures a

two-port circuit and S-parameter references and definitions. A 2 × 2 system
([b] = [S] × [a]) for the two-port circuit’s S-parameters can then be written as:
b1 = S11a1 + S12a 2 (4.14)
b2 = S 21a1 + S 22a 2 (4.15)
Here, S-parameters, derived under matched terminations, are as follows:
• S11: Voltage reflection coefficient at port 1.

• S22: Voltage reflection coefficient at port 2.
• S21: Forward voltage transfer function (from port 1 to port 2).
• S12: Backward voltage transfer function (from port 2 to port 1).
Input voltage reflection coefficient is calculated from S11 = b1/a1 when a2

= 0. This means the second port is matched terminated (ZL = Z02). The term
matched termination is used when a port is terminated with its output imped-
ance. Matched termination means there is no reflection at a port.
S-parameters of two-port circuits are derived using circuit theory knowl-
edge (i.e., Kirchoff ’s voltage and current laws). Figure 4.26 depicts a few two-
port circuits for exercises of the derivation of S-parameters.
Example: Find S-parameters of the circuits in Figure 4.26(a) and 26(b).
Solution: The circuit in Figure 4.26(a) is a single-impedance inserted seri-
ally between input and output. Characteristic impedances of all ports are Z0.
In order to calculate S11, the output port is terminated with Z0. The total load
impedance is then equal to Z1 + Z0. The I1 = – I2 is equal to input voltage reflec-
tion coefficient (S11 = ρ = Z1/(Z1 + 2Z0)). This two-port circuit has a structural
Figure 4.25 S-parameter definitions of a two-port circuit.

Figure 4.26 Sample two-port circuits for S-parameter calculations.
symmetry; therefore, S11 = S22. Using current continuity (i.e., I1 = –I2), the
other two parameters are found as S21 = S12 = 1 – ρ.
The circuit in Figure 4.26(b) is impedance connected in parallel between
input and output. The S-parameters of this circuit are S11 = ρ, S12 = 1 + ρ, S21
= 1 +, S22 = ρ, where the voltage reflection coefficient is ρ = ((Z1//Z0) – Z0)/
((Z1//Z0) + Z0).
Question: For the π–circuit given in the figure, SdB = 20log10S, show that
S11 = S22 = –53dB and S12 = S21 = –6dB. Here, Z1 = 150Ω, Z2 = 150Ω, Z3 =
37Ω, and Z0 = 50Ω.
Table 4.3 lists a short MATLAB code for the S-parameter calculations of
this two-port PI-circuit. The output of this code is given in Figure 4.27. The
dB values of the calculation results match with the values given in the question.
Another MATLAB code for the TEE circuit given in Figure 4.26(d) is
listed in Table 4.4. A sample output of this code is given in Figure 4.28.
Some important properties of S-parameters are as follows:
• In circuits with passive R/L/C elements, S-parameters are complex

numbers and their amplitudes are less than 1 (0 ≤ |S | ≤ 1). If |S11| > 1,
then the reflected voltage is greater than the incident voltage. This oc-
curs only in circuits with active amplifications.
• In passive circuits, incident and reflected powers at any port are given in
terms of the square of normalized voltage waves: Pin = |a|2 and Pref = |b|2.
• S-matrix is symmetric in reciprocal circuits. Note that structural sym-
metry always yields electrical symmetry; electrical symmetry may or may
not yield structural symmetry.
Circuit Model 93
Table 4.3
MATLAB Code for Calculating S-Parameters of the PI-Circuit
%-----------------------------------
clear all; close all; clc;
z0 = 50; % Characteristic impedance of the TL;
z1 = 150; z2 = 150; z3 = 37; %z1 = 100-50j;
% Calculate Y-parameters
y1=1/z1; y2=1/z2; y3=1/z3; y11=y1+y3; y12=-y3; y21=-y3; y22=y2+y3;
% Calculate S-parameters
s11=((1/z0-(y1+y3))*(1/z0+(y2+y3))+y3^2)/((1/z0+(y1+y3))*(1/z0+(y2+y3))-y3^2);
s12=(2*y3/z0)/((1/z0+(y1+y3))*(1/z0+(y1+y3))-y3^2);
s21=(2*y3/z0)/((1/z0+(y1+y3))*(1/z0+(y1+y3))-y3^2);
s22=((1/z0+(y1+y3))*(1/z0-(y2+y3))+y3^2)/((1/z0+(y1+y3))*(1/z0+(y2+y3))-y3^2);
% Display Results
fprintf(‘ \n’); fprintf(‘ \n’);
fprintf(‘-------------Z3----------- \n’); fprintf(‘ | | \n’)
fprintf(‘ Z0 Z1 Z2 Z0 \n’); fprintf(‘ | | \n’)
fprintf(‘---------------------------- \n’); fprintf(‘ \n’);
fprintf(‘ Z1 = %5.1f Ohm; Z2 = %5.1f Ohm; Z3 = %5.1f Ohm \n’,z1,z2,z3);
fprintf(‘ \n’); fprintf(‘---S-Parameters of the above PI Circuit--- \n’);
fprintf(‘---------------------------------------------------- \n’); fprintf(‘ \n’);
fprintf(‘ S11 = ‘); disp(s11); fprintf(‘ S22 = ‘); disp(s22);
fprintf(‘ S12 = ‘); disp(s12); fprintf(‘ S21 = ‘); disp(s21); % Program END
Figure 4.27 The output of the MATLAB code in Table 4.3.

Table 4.4
Short MATLAB Code for the TEE-Circuit in Figure 4.26(d)
%-----------------------------------
z0 = 50; z1 = 150; z2 = 150; z3 = 37; %z1 = 100-50j;
% Calculate S-parameters
s11=((z1+z2-z0)*(z2+z3+z0)-z2^2)/((z1+z2+z0)*(z2+z3+z0)-z2^2);
s22=((z1+z2+z0)*(z2+z3-z0)-z2^2)/((z1+z2+z0)*(z2+z3+z0)-z2^2);
s12=(2*z2*z0)/((z1+z2+z0)*(z2+z3+z0)-z2^2);
s21=(2*z2*z0)/((z1+z2+z0)*(z2+z3+z0)-z2^2);
fprintf(‘ \n’); fprintf(‘ \n’); % Display Results
fprintf(‘--------Z1---------Z3--------- \n’);
fprintf(‘ | \n’);
fprintf(‘ Z0 Z2 Z0 \n’);
fprintf(‘ | \n’);
fprintf(‘----------------------------------- \n’); fprintf(‘ \n’);
fprintf(‘ Z1 = %5.1f Ohm; Z2 = %5.1f Ohm; Z3 = %5.1f Ohm \n’,z1,z2,z3);
fprintf(‘ \n’);
fprintf(‘---S-Parameters of the above TEE Circuit--- \n’);
fprintf(‘--------------------------------------------- \n’); fprintf(‘ \n’);
fprintf(‘ S11 = ‘); disp(s11); fprintf(‘ S22 = ‘); disp(s22);
fprintf(‘ S12 = ‘); disp(s12); fprintf(‘ S21 = ‘); disp(s21); % Program END
Figure 4.28 The output of the MATLAB code in Table 4.4.

Circuit Model 95
• S-matrix is unity in lossless circuits; that is, total power entering from
the ports is equal to the total reflected power (|a1|2 + |a2|2 = |b1|2 + |b2|2).
Why S-Parameters?
Because they are relatively easy to obtain at high frequencies; voltage traveling
waves can be measured with a vector network analyzer; they don’t need SC/OC
terminations, which can cause active devices to oscillate or self-destruct; they
are related to familiar measurements (gain, loss, reflection coefficient, and so
on); S-parameters of multiple devices can be cascaded to predict system perfor-
mance; H, Y, or Z parameters can be computed from S-parameters if desired;
and S-parameter files can easily be imported and used in numerical simulation
tools.
4.9 Microstipline Circuits

Most electronic products are based on PCB technology; therefore, understand-
ing microstrip TL is a must for engineers in this area. Microstrip antennas,
multiport microstrip networks (e.g., impedance matchers, filters, couplers, and
so on) based on multilevel PCB technology are very cheap, they are small, and
production in large quantities is easy.
Planar TLs are inhomogeneous, since they consist of two propagation
media, like the dielectric substrate and air. Inhomogeneous lines do not sup-
port TEM modes. This is because the propagation constant of a TEM wave
depends only upon frequency and the material properties of the medium. The
dominant mode of many inhomogeneous TLs, such as insulated two-wire pair,
loaded coaxial, and microstrip lines, behaves like a TEM mode at low frequen-
cies, called quasi-TEM. The longitudinal components of the quasi-TEM mode,
while nonzero, have negligible amplitudes. The propagation of this mode can
be determined approximately from electrostatics. The inhomogeneous struc-
ture is replaced by an equivalent homogeneous one, which is characterized by
the effective permittivity obtained from electrostatic study of the structure.
4.9.1 Characteristics of a Microstripline

As pictured in Figure 4.29, a microstripline is a structure having a narrow PEC
ribbon on one face of the dielectric substrate; the other face is entirely PEC. The
characteristic physical parameters of microstrip line are as follows:
Figure 4.29 A basic microstripline circuit.
• Relative permittivity er of the substrate: The higher the permittivity, the

more concentrated fields within the substrate, and the less (undesired)
radiated power.
• The substrate thickness h: Generally, this is a fraction of a millimeter.
• The width W of the PEC ribbon: This width is of the same order as the
substrate thickness, h (0.1 ≤ W/h ≤ 3.0). The characteristic impedance
of the microstrip line is controlled by varying W.
• The thickness b of the PEC ribbon: Generally, this is negligible (b/h <<
1).
As shown in the figure, electric field lines go from the ribbon to the
ground plane, most of which are concentrated underneath the ribbon. Some
lines partially extended into the free space above the substrate. The magnetic
field lines encircle the microstrip ribbon and extend above the substrate. The
presence of electric field lines in the air reduces the effective dielectric constant
seen by the propagating wave. If all the fields existed between the ribbon and
ground plane, the dielectric constant would be that of the substrate. Instead, it
is somewhat less than that constant, and it depends on ribbon width, substrate
dielectric constant, and height.
EM energy is trapped under the top ribbon inside the dielectric substrate.
The trapped energy is proportional with the permittivity of the substrate. In
terms of EMC engineering, this reduces undesired emission. Characteristic im-
pedance of the microstripline is determined by the ratio W/h (0.1 ≤ W/h ≤
3.0).
In Figure 4.30, characteristic impedance versus frequency for different
W/h ratios is plotted. When the thickness of the strip is negligible (compared
to its width), characteristic impedances between 10–20Ω up to150–250Ω can
be obtained by changing the width, the substrate height, or substrate relative
permittivity.
Circuit Model 97
Figure 4.30 Z0 versus W/h ratio for different substrate permittivities.
As illustrated in the figure, the higher the substrate relative permittivity,

the higher the characteristic impedance; the higher the W/h ratio, the lower the
characteristic impedance; the wider the strip width (when substrate height is
same), the lower the characteristic impedance; and the narrower the strip width
(when substrate height is same), the higher the characteristic impedance.
There are several empirical equations for the calculation of characteristic
impedance of a microstripline [3–5]. Table 4.5 lists a short MATLAB code that
plots characteristic impedance versus W/h ratio (as presented in Figure 4.30).
Microstripline circuits are too complex to solve analytically and obtain
closed-form solutions. It necessitates solution of Maxwell’s equations under
complex boundary and radiation conditions. Instead, empirical equations based
on measurements are widely used. Parallel to the increase in computer memory
and processor speeds, full wave computer simulations have also been developed
for the past few decades [4–6].
Table 4.5
Short MATLAB Code for the Calculating Z0 of a Microstripline
% Microstrip characteristic impedance plot
clc; clear all; close all;
er=9.6 % relative permittivity of the substrate
r=linspace(.01,1,100); % W/h <= 1
eff=(er+1)/2+((er-1)/2)*((1+12./r).^(-1/2)+0.04*(1-r).^2);
Zo=(60./sqrt(eff)).*log(8./r+r/4);
plot(r,Zo); xlabel(‘W/h ratio’); ylabel(‘Zo (Ohms)’); hold on;
grid on; title(‘Microstrip Characteristic Impedance versus W/h’)
r=linspace(1,5,400); % W/h > 1
eff=(er+1)/2+((er-1)/2)*(1+12./r).^(-1/2);
Zo=(120*pi)./(sqrt(eff).*(r+1.393+0.667*log(r+1.444)));
plot(r,Zo); % Program END
4.9.2 Basic Microstrip Circuits

Basic circuits such as filters, impedance matching networks, coupler, resonators,
antennas, and so on can easily be constructed with microstriplines. A two-port
approach is pictured in Figure 4.31. On the top, the 3D view is given; at the
bottom, a vertical slice is given. Input and output junctions, and a sample mi-
crostripline circuit in the middle are also shown.
Because of impedance mismatches at port 1 and 2, source-excited inci-
dent voltage wave is partially transmitted and reflected at port 1. S-parameters
of the circuit can be obtained using incident and reflected waves at both ports.
Source power can be given via a coaxial to microstripline connector. The top
view of a few microstripline circuits is given in Figure 4.32. A photo of a third-
order microstrip low pass filter is shown in Figure 4.33.
Microstrip circuits are more than a single straight TL. Lines are often bent
because of, for example, circuit topology and space limitations. Changes in rib-
bon width are often used when different characteristic impedances are required.
Junctions, intersections of more than two lines, are used in various ways. OC
stubs are used for impedance matching. A gap or a notch can be introduced
for different purposes. These are all abrupt discontinuities in microstrip lines,
which alter the electric and magnetic field distributions.
Another sample microstripline circuit (a third-order bandgap filter) is pre-
sented in Figure 4.34. IL versus frequency of this structure is also measured and
data is plotted. Analytical (lumped element) and numerical (FDTD) results are
also given [7–9].
Figure 4.31 A microstripline circuit as a two-port element.

Circuit Model 99
Figure 4.32 Typical microstripline circuits (top view).
Figure 4.33 A third-order microstripline LPF.
4.10 Crosstalk
Crosstalk is any phenomenon by which a signal transmitted on one cable cre-
ates an undesired effect in another cable. It is an internal EMC problem. It
is usually caused by undesired capacitive (electric field), inductive (magnetic
field), or conductive (common impedance) coupling from one cable to other
or one cable wire to another. Near end crosstalk (NEXT) is a measure of the
ability of cabling to reject crosstalk. Interference between two pairs in a cable
is measured at the same end of the cable as the interfering transmitter. Far end
crosstalk (FEXT) is the interference between two pairs of a cable measured at
the other end of the cable with respect to the interfering transmitter.
Generally speaking, a ground plane is a useful tool to combat crosstalk.
A good grounding may cause up to a 40–60 dB reduction in the common
Figure 4.34 Analysis results of a sample BSF.
impedance. The ground plane may also reduce mutual inductance coupling.
The capacitive coupling between cables or wires is not directly affected by the
ground plane, but the lowered common impedance indirectly reduces the ca-
pacitive crosstalk amplitude.
Crosstalk may be a major EMC problem in digital circuits.
References
[1] Electromagnetic Compatability Industry Association, journal articles of Armstrong, K.,
www.compliance-club.com. Last accessed, Oct. 21, 2016.
[2] Williams, T., EMC for Product Designers, 3rd Ed., Oxford: Newnes, 2001.
[3] Hong, J. G., and M. J. Lancaster, Microstrip Filters for RF/Microwave Applications, New
York: John Wiley and Sons, 2001.
[4] Sevgi, L., Electromagnetic Modeling and Simulation, New York: IEEE–John Wiley and
Sons, 2014.
[5] Sevgi, L., Complex Electromagnetic Problems and Numerical Simulation Approaches, New
York: IEEE–John Wiley Press, June 2003.
[6] Sevgi, L,. “Electromagnetic Modeling and Simulation: Challenges in Validation, Verifica-
tion and Calibration,” IEEE Trans on EMC, Vol. 56, No. 4, Aug. 2014, pp. 750–758.
[7] Çakır, G., S. Gündüz, and L. Sevgi, Geni Bantlı Filtre Tasarımı, Eksen Yayıncılık, Istanbul,
2006.
Circuit Model 101
[8] Gündüz, S., G. Çakır, and L. Sevgi, “A Generic Microstrip Structure for the Realization of
All-Type Broadband Filters,” Microwave and Optical Technology Letters, Vol. 48, No. 12,
Dec. 2006, pp. 2390–2393.
[9] Çakır, G., and L. Sevgi, “A Design of Microstrip Electromagnetic Double-Band Band-
gap (DB-EBG),” Microwave and Optical technology Letters, Vol. 48, No. 7, Jul. 2006, pp.
1405–1408.
Bibliography
Bowick, C., Circuit Design, Boston: Newness, 1982.
Besser, L., and R. Gilmore, Practical RF Circuit Design for Modern Wireless Systems Volume I,
Norwood, MA: Artech House, 2003.
Ludwig, R., and P. Bretchko, RF Circuit Design Theory and Applications, Englewood Cliffs, NJ:
Prentice Hall, 2000.
5
Antennas and Antenna Calibration
An antenna is a transducer [1]. A transducer is a device that transforms one type
of energy to another (e.g., mechanical to heat or kinetic to electrical, and so
on). Some may say that an antenna is not a transducer since it converts electri-
cal energy to electrical energy again. But it is appropriate to call an antenna a
transducer because it transforms electrical energy to EM energy and vice versa.
An antenna is a reciprocal device and behaves the same on transmit and
on receive [2]. The transmit antenna converts the voltage between its input ter-
minal, or the current flowing through to EM fields, while the receive antenna
does the opposite. It is a device that provides a transition from a guided wave on
a TL to a free-space EM wave (or vice versa). The IEEE definition of an antenna
is given as [3] “that part of a transmitting or receiving system that is designed to
radiate or receive EM waves.”
Antennas can focus and shape the radiated power in space by enhancing
the power in some desired directions and suppressing it in others. Antenna
theory is a well-established topic, so there are wonderful classical books (see
e.g., [4–15]).
Loosely speaking, antennas may be grouped into two classes depending
on their usage: communication antennas and EMC antennas. The fundamental
reason for this grouping is that communication and EMC antennas are used
in the far and near field regions, and mostly for narrowband and broadband
applications, respectively. Nobody deals with a TV and/or a radio broadcast an-
tenna once it is installed, except during periodic maintenance and failures. But
an EMC antenna is a device that is used many times a day in screened/anechoic
rooms on OFTS during emission measurements and immunity tests as either
receive or transmit elements.
A typical RE measurement setup is pictured in Figure 5.1. These mea-
surements are performed with broadband log-periodic, spiral, and biconic
103
Figure 5.1 A typical RE measurement setup.
antennas. EMC engineers perform tests and measurements almost every day
and write their reports. They record unwanted radiations from the EUT and
compare results with the threshold values in a specified standard. The objective
of immunity tests is to find out whether or not the EUT yields its designated
performance under the influence of EM interference (fast transients, ESD,
surges, voltage dips, and so on) at a specified level. No matter what the tests are,
EMC engineers usually do not pay much attention to fundamental antenna pa-
rameters such as standing wave ratio, gain, radiation pattern, input impedance,
and so on. The antenna factor (AF) and the transmit antenna factor (TAF) are
the two most important antenna parameters for the EMC engineers.
The primary design parameter for an EMC antenna is bandwidth;
VSWR, gain, efficiency, input impedance, and so on, are of secondary impor-
tance. This is mainly because EMC tests and measurements are performed over
broad frequency ranges specified by the standards. For example, RE measure-
ments of most commercial electrical and electronic devices are required in the
frequency range of 30 MHz–6 GHz. On the other hand, antenna engineers
who design communication antennas are accustomed to work with those sec-
ondary parameters.
The antenna is an interesting device. It combines well-established EM
and circuit theories. Therefore, an antenna engineer should be familiar with
both its EM and circuit equivalents.
Figure 5.2 depicts one of the simple models of a transmit/receive antenna
pair. The basic circuit representation of a transmit antenna consists of a com-
plex impedance ZA connected across a voltage source with an internal imped-
ance Zs . The radiated power Pradiated at the transmit side is determined by the
antenna input impedance and the voltage across its input terminal supplied
by the transmit circuit. The mismatch at the transmit antenna input deter-
mines the antenna input power PA extracted from the supplied power Ps (under
Antennas and Antenna Calibration 105
Figure 5.2 Simple antenna models.
matched conditions PA = Ps/2). The antenna input power is partially radiated;

the rest goes to antenna losses (PA = PLosses + Pradiated).
On the receive side, a voltage source and a complex impedance in series
can be used as a simple circuit model of the receive antenna. Again, antenna
input power is partially reflected back, and rest is fed into the receive block (PA
= PRefl + PRec) depending on the mismatch between ZA and the receiver input
impedance ZRec.
5.1 Fundamental Antenna Terms

A few important terms and definitions are included here for the sake of com-
pleteness. The reader is referred to the IEEE Standard Definitions of Terms for
Antennas, IEEE Std 145-1983, for further information. Also, more in-depth
explanations and formulas can be found in [4–15].
• Antenna: End (front) part of a transmitting (receiving) system that is

designed to radiate (receive) EM waves.
• Frequency bandwidth: Range of frequencies within which the perfor-
mance of the antenna conforms to a specified standard with respect to
specified characteristics.
• Input impedance: The impedance presented by an antenna across its
terminals.
• Reflection coefficient: The ratio of the voltages corresponding to the
reflected and incident waves at the antenna input terminal. It can be
calculated using the input impedance and the characteristic impedance

of the connecting feed line.
• Voltage standing wave ratio (VSWR): The ratio of the maximum/mini-
mum values of a standing wave pattern along a TL to which a load is
connected.
• Far-field region: That region of the field of an antenna where the an-
gular field distribution is essentially independent of the distance from a
specified point in the antenna region.
• Antenna polarization: In a specified direction from an antenna and at
a point in its far field, this is the polarization of the (locally) plane wave
used to represent the radiated wave at that point. It can be linear or el-
liptical (with circular being a special case).
• Antenna effective aperture: The ratio of the power delivered by the an-
tenna in the terminating resistance to the power density of the incident
EM plane wave; the wave being polarization matched to the antenna.
• Antenna pattern: The antenna pattern is a graphical representation of
the radiation of the antenna as a function of angular direction. Antenna
patterns are usually measured or computed in two orthogonal principal
planes (such as E-plane and H-plane or vertical and horizontal planes).
• Antenna directivity: The ratio of the maximum radiation intensity to
the average radiation intensity.
• Antenna efficiency: A parameter between (0–1) that accounts for the
mismatch as well as connector, cable, conductor, and dielectric losses.
• Antenna gain: Simply, the product of the directivity by efficiency, which
accounts for all cable, connector losses, and so on.
• Gain pattern: Normalizing the power/field to that of a reference an-
tenna yields a gain pattern. When the reference is an isotropic antenna,
the gain is expressed in dBi. When the reference is a half-wave dipole in
free space, the gain is expressed in dBd (1 dBd = 2.15 dBi).
• Half-power beamwidth: The angle between the two directions in which
the radiation intensity is half of the maximum value. It is also commonly
referred to as the 3-dB beamwidth.
• Isotropic radiator: A hypothetical, loss-free antenna (point source) hav-
ing equal radiation intensity in all directions.
• Omnidirectional antenna: An antenna having a nondirectional pattern
on a given plane.
5.2 Communication Antennas

Broadcast, radar and/or communication engineers desire to know the effective
isotropic radiated power (EIRP) at the transmit site and the received power at
the receive site. The EIRP is the product of the power radiated by the antenna
and its gain reduced by a loss factor:
EIRP = Ps + Gt − Ltot (5.1)
Here Ps is the source power in dB, Gt is the transmit antenna gain in dBi,
and Ltot is the total loss that accounts for impedance mismatch as well as cable
losses, insertion losses, and so on, in dB. Note that effective radiated power
(ERP) instead is used when the antenna gain is given in dBd.
The free-space path loss (or power dissipation, LFree) of an antenna range
determines the difference in power levels between the output of the transmit
antenna and the output of an isotropic (0 dBi) antenna located at the receive
site. This free-space loss is due to the dispersive nature of a transmitting antenna
(not atmospheric losses). A transmitting antenna radiates EM waves with a
spherical wavefront; only a portion of this spherical wavefront is captured by
the receiving antenna. For a free space, at a far-field range d and a given fre-
quency f, the path loss in dB is easily determined as follows:
LFree = 32.45 + 20Log 10 (d km ) + 20Log 10 ( f MHz ) (5.2)
The power PA in dB at the receive side is given in terms of EIRP in dB,

free-space path loss LFree in dB, and the receive antenna gain Gr in dBi as
(5.3)
PA = EIRP − LFree + Gr
All of these are derived from the well-known Friis formula:
2 2
 λ   4π d 
PA = Pt Gt Gr  ; L =  
 4 π d 
Free
λ  (5.4)
in which d is the distance and
G r λ2
Ae = (5.5)
4π
is used as the effective aperture of the receive antenna. Equation (5.4) is widely
used in signal power calculations in communication links assuming antennas
are matched in terms of both impedances and polarizations. In terms of source

and receiver powers, and the mismatches at both ends, the Friis equation given
in (5.4) can be modified as
2

 λ 
PA = Pt 
 4 π d 
(
Gt 1 − Γt
2
)G (1 − Γ )
r r
2
(5.6)
Note that the gains in (5.4) are as defined in the standard IEEE defini-
tions, while (modified) gains (used in the EMC society) derived from [14]
(
Gt' = Gt 1 − Γt
2
), (
Gr' = Gr 1 − Γr
2
) (5.7)
include mismatches. Here, Γ is the voltage reflection coefficient defined as the

ratio of reflected and incident voltages at a specified junction. Note that gain,
by the IEEE definition, includes directivity and antenna efficiency (ohmic loss-
es), but in EMC applications, it also includes additional mismatch factors.
Antenna radiation patterns, critical for communication engineers, show
the ability of energy radiating with respect to direction. Although not neces-
sary, antenna radiation patterns are usually plotted either in vertical plane or
in horizontal plane. The terms used for directive antennas are (as depicted in
Figure 5.3):
• Main lobe: The lobe along best radiation direction;

• Side lobe: Other lobes on both sides of the main lobe;
• Back lobe: The lobe just the opposite direction of the main lobe;
• MLSLR: Main lobe–side lobe power ratio;
• MLBLR: Main lobe–back lobe power ratio.
• Radiation beamwidth: The angular sector of the main beam between
half power (3 dB) angles.
5.3 EMC Antennas

EMC antennas are used for the recording of unwanted emissions as well as in
radiated immunity tests. Antenna calibration is a must for RE measurements
but this is not a requirement in radiated immunity tests. In immunity tests,
calibration of the EMC test systems as a whole is required; the specification of
the field strength in an environment where EUT is placed and its distribution
must be known. RE measurements and RI tests bring two important antenna
Figure 5.3 Antenna radiation pattern parameters.
parameters into consideration: AF for the emission measurements, and TAF

for the immunity tests. These two parameters and a simple circuit models are
depicted in Figures 5.4 and 5.5.
5.3.1 Receive Antenna and Antenna Factor

Antenna factor (AF) defined for a receive antenna is a critical parameter in
EMC tests and measurements. Theoretical AF can be obtained using the model
shown in Figure 5.4. In this model, the input impedance ZA of the antenna
consists of a real part RA , the combination of RL and Rr, representing power
losses and EM radiation, respectively, and, an imaginary part XA , representing
inductive or capacitive energy storage. The impedance mismatches at the two
discontinuities (i.e., antenna-cable and cable–EMI receiver connections) cause
Figure 5.4 A simple circuit model for the receive antenna.

Figure 5.5 A simple circuit model for the transmit antenna.
power reflection losses that are characterized by voltage reflection coefficient

(Γ) and VSWR.
The power delivered to the receiver input is
2
1 2 1 VA R Re c
PDelivered = I A R Rec =
2 2 (R A + R Re c )2 + ( X A + X Re c )2 (5.8)
and becomes maximum
1 ( VA )rms
2 2
1 VA
(PDelivered )max =
8 RA
or (PDelivered )max =
4 RA
(5.9)
when RA = RRec and XA = –XRec (here VA and (VA)rms are the peak and root-
mean-square values of the source and are related via (V A )rms = V A 2 for the
sinusoidal case). In many cases, the antenna impedance is not known; therefore,
VSWR is measured and used.
Voltage Standing Wave Ratio

The VSWR is a parameter related to the impedance mismatch and therefore to
the reflected power at the antenna terminals. It is defined as the ratio of voltage
minimum to maximum (it is a parameter that can easily be measured on slotted
TLs). The Γ and the VSWR at the antenna terminal having input impedance
ZA, connected to a TL with characteristic impedance Z0, are given as
V Re fl Z A − Z0 1+ | Γ |
Γ= = , VSWR = s = (5.10)
Vinc Z A + Z0 1− | Γ |
Note that Γ in (5.10) is for the direction toward the antenna; it becomes
Γ = (Z0 – ZA)/(Z0 + ZA) if the direction is toward the cable (connecting to the
source). But, the percentage of the reflected power does not change with direc-
tion, since it is proportional with |Γ|2 (reflection loss in dB is defined as RL =
–20log10|Γ|). If the power delivered to the antenna by the source is Ps, then
antenna power (the sum of radiated power and the power dissipated by the
antenna itself because of the losses) is PA = Ps(1 – |Γ|2)
Antenna Effective Aperture (Ae)

The antenna effective aperture is defined as the ratio of the power dissipated in
the receiver resistance RRec and the power density Pd of the incident EM wave
(see Figure 5.4):
Ae =
( )
R Rec I A2
rms
(5.11)
Pd
where (IA)rms is the rms value of the current flowing through the circuit shown
in Figure 5.4:
(V )
(I )
A rms
=
A rms
(R A + RRe c ) + j ( X A + X Re c ) (5.12)
Combining (5.9), (5.11), and (5.12), and using the fact that the maxi-
mum delivered power is the transferred power of the Friis transmission equa-
tion one obtains Ae as:
G A λ2 4 R A R Re c
Ae =
4 π (R A + R Re c )2 + ( X A + X Re c )2 (5.13)
This equation consists of a maximum aperture (first fraction) which is

equal to Ae given in (5.5) (note that GA = Gr ) and the second fraction represent-
ing the mismatch between the antenna and the receiver. It can also be written as
Ae =
G A λ2
4π
(
1 − Γ2 ) (5.14)
where the reflection coefficient is as defined in (5.10). Note that (5.10) is dif-
ferent than (5.4) by a factor of impedance mismatch; in other words, (5.5) does
not include the mismatch factor.
Antenna Factor (AF)

AF is defined as a parameter that relates electric field in the vicinity of the an-
tenna to the voltage across the receiver input connected to its terminals as (see
Figure 5.2) [15, 16]:
E
AF = m −1 
V Re c  
(5.15)
It presents the amount of voltage captured by the antenna (note that VRec
and VA are at the same junction if the antenna is connected directly to the
receiver without any intervening cable). As a logarithmic value, AF is given as
E [dBV / m ] = V Re c [dBV ] + AF [dB / m ] (5.16)
AF is a parameter related to the use of an antenna as a EM field measuring

probe. It may be defined for both OC and resistive termination conditions, and
in the near and far fields. However, most published AFs are defined for 50Ω
systems.
The AF of a receive antenna can be derived mathematically using simple
circuit models under idealistic conditions. The input power PA of the antenna
can be given in terms of effective antenna aperture Ae and the power density Pd
in the close vicinity:
G r λ2
PA = Pd × Ae [ W ], Ae = m 2  (5.17)
4π  
The power density can also be calculated if the electric field in the vicinity
of the antenna is known:
E2 E2
Pd = =  W m2  (5.18)
η0 120 π  
where η0 = 120π = 377Ω is the free-space wave impedance. This is valid un-
der far-field conditions (i.e., transmit antenna that causes this EM wave is far
from the receiver; this could be taken as the distance which satisfies d ≥ λ/2π
for simple and low-gain antennas. Use d ≥ 2D2/λ for the high-gain complex
antennas; D is the longest dimension of the antenna).
The antenna is usually connected to a receiver (e.g., an EMI device) over
a finite-length TL. Assuming (1) the frequency-independent receiver input im-
pedance is 50 Ω; (2) the characteristic impedance of loss-free TL is real and
equal to Z0 = 50Ω, allow the use of the simple circuit model depicted in Figure
5.2. In this case, the output power of the receive antenna (equivalently, the
input power of the receiver) PA can be given as
V A2 V A2 V A2 V A2
PA = = = = [W ] (5.19)
R Re c R A Z 0 50
Combining (5.17)–(5.19) yields
V A2 E2 G λ2
== × r [W ] (5.20)
R Re c 120 π 4π
from which AF is obtained as

E 480 π2
AF = = m −1  (5.21)
2  
VA R Re cGr λ
For a 50Ω system, this reduces to
9.73
AF = m −1  (5.22)
λ Gr  
or, in dB:
AF [dB / m ] = 19.8 − 20Log 10 ( λ) − 10Log 10 (Gr ) (5.23)
AF [dB / m ] = 20Log 10 ( f MHz ) − 10Log 10 (Gr ) − 29.79 (5.24)
It should be noted that this expression applies only to far-field (plane

wave) situations, valid only for 50Ω systems and under polarization- and im-
pedance-matched conditions. When measured, it includes the effects of mis-
matches at the antenna terminals and its associated equipment such as a balun
or a matching transformer (not the cable connecting antenna to the receiver).
Actually, AF is related to the effective aperture via
η0
AF = m −1  (5.25)
R Re c Ae  
meaning that effective aperture is a parameter representing power captured by

the antenna, while AF is a parameter representing voltage derived by the an-
tenna from the electric field.
Antennas with smaller AFs are more sensitive to the incident field. AF
generally increases with frequency so more sensitive receivers are needed at
higher frequencies in order to measure the same field level. AF values are nor-
mally provided by antenna manufactures or calibration labs as AF versus fre-
quency tables and/or plots.
5.3.2 Transmit Antenna Factor

TAF is defined as the ratio of electric field strength E at a distance d from the
transmit antenna and the voltage across its terminals VA · TAF = E/VA [m–1] the
same as given in (5.15) (see Figure 5.5). It can be calculated using VA , input
power, and Friis Formula.
If radiated power is Pt and antenna gain Gt, then power density at a range
d is equal to:
Pt Gt
Pd =  W m2  (5.26)
2  
4πd
The same power density can be derived from the electric field at that
point and the wave impedance as given in (5.18); therefore, the combination of
(5.18) and (5.26) yields
30Pt Gt
E=
d
[ V m] (5.27)
The input power may be derived from Ohm’s law using voltage across its
terminals and its input resistance as
V A2
Pt = [ W ], or V A = Pt R A (5.28)
RA
The combination of (5.27) and (5.28) yields the TAF as:
E 1 30Gt
TAF = = m −1  (5.29)
VA d RA  
For a RA = 50Ω system this reduces to
1
TAF = 0.6Gt m −1  (5.30)
d
or, in logarithmic form,
TAF [dB / m ] = 10Log 10 (Gt ) − 2.22 − 20Log 10 (d m ) (5.31)
As observed in (5.30), TAF is inversely proportional with the distance

from the transmit antenna and proportional with its gain.
Although AF and TAF have the same units, they are neither identical nor
reciprocal. They can be connected when the gains in both cases are identical
(i.e., use Gt = Gr and combine (5.22) and (5.29)) as:
TAF [dB ] = 20Log 10 ( f MHz ) + AF − 32.0 (5.32)
Note that (5.32) is valid only when one of AF-TAF is known and both
are obtained under the same conditions. For example, if AF is measured over a
ground plane, then TAF is requested over ground plane, too.
A typical AF graphic is given in Figure 5.6. Here, AFs of two different
models against frequency are plotted. As observed, AFs of different antennas
may be totally different. The important point is that AF (i.e., electric field–re-
ceive voltage transformation ratio) is known at every single frequency. If the
recorded voltage value is X dB and AF at this frequency is Y dB, then the mea-
sured electric field in the environment will be X + Y dB.
Figure 5.6 Typical producer-supplied AF versus frequency curves.

5.4 Antenna Calibration

Suppose you have just received your university degree and started to work with
the National Metrology Institute (NMI) as a new antenna design engineer. The
director of your department tells you that there are a few antennas left in the de-
pot for a long time and gives you your first task: You’re going to calibrate those
antennas, measure their antenna factors (AFs), and make them ready for EMC
measurements. It has only been a few months since you took your antennas and
propagation course. You remember that you enjoyed it a lot while exercising
useful virtual tools and antenna software. But antenna calibration and AF mea-
surements were new for you; you remember nothing about AF concepts. Where
do you need to start? What steps should you follow? What kind of procedures
should you apply? On what site and with what instruments should you perform
the measurements?
For example, you may start reading the tutorial in [17], revise antenna
terms and definitions, and learn how to calibrate an antenna. You may search
the literature and come up with some useful material [18–21]. Also, a short visit
to antenna companies’ websites such as [22, 23] would also be useful. But, of
course, the first step is to obtain related standards of international institutions,
such as CISPR [24–26], ANSI/IEEE [27–30], and ISO [31].
Antenna calibration necessitates a calibration test site (CALTS) as ex-
plained in [32], a calibration method, and certificated measurement devices.
The characteristic parameter for a CALTS is the normalized site attenuation
(NSA) as stated in [16]:
CALTS is an open field test site with metallic ground plane and tightly
specified site attenuation performance in horizontal and vertical polarization.
A CALTS is used for determining free-space antenna factor of an antenna.
Site attenuation measurements of a CALTS are used for comparison to
corresponding site attenuation measurement of a compliance test site, in
order to evaluate the performance of the compliance test site.
Antenna calibration is performed in antenna ranges: either outdoor

(OFTS) or indoor (anechoic chambers). OFTS suffers from environmental
conditions, while indoor ranges are space limited. There are several antenna
calibration methods.
The standard site method (SSM) is specified in ANSI/IEEE C63.5 [16]
and is best suited for dipole-like antennas. Here, the IL between transmit and
receive antennas is measured on a large, flat, and unobstructed conducting
ground plane. One antenna is set to be at a fixed height (e.g., 1m or 2m), while
the other one is scanned from 1m to 4m in height. The maximum response
between the two antennas is recorded. Typically, three antennas are needed to
perform such a calibration, and they are measured in three pairings. Calcula-
tions are then performed to derive the AFs. Although a ground plane is used,
the aim of the ANSI/IEEE C63.5 SSM is to obtain the free-space AF by theo-
retically removing the ground plane effect. The conducting ground plane is
there to establish a repeatable calibration environment.
SSM assumes that (1) the antennas under test have radiation patterns of
a short dipole (i.e., horizontally uniform and figure “∞” vertically), (2) anten-
nas and the metal ground plane are all uncoupled and, (3) antennas are in the
far field so that the physical size of an antenna has negligible effect. However,
these are not always acceptable. Depending on types of antennas, error intro-
duced from the SSM can be as large as 2.5 dB [16]. This means the total error
could be 5 dB (because an antenna pair is used for the site attenuation). A new
ANSI/IEEE C63.5-2006 addresses this limitation by providing correction fac-
tors (CFs) [28], which are based on numerical simulations.
Reference antenna method (RAM) is basically a substitution method speci-
fied in ANSI/IEEE C63.5 [16]. Here, the responses between two known an-
tennas are measured, and then one is replaced by the antenna under test. The
AF is derived from the difference. Again, the mutual coupling between the two
standard antennas and the antenna under test, as well as ground plane effects,
can be significantly different, leading to significant errors.
The equivalent capacitance substitution method (ECSM) is used to cali-
brate monopole antennas from 9 kHz to 30 MHz. In this method, a dummy
antenna consisting of a capacitor equal to the self capacitance of the monopole
is used in place of the actual element (see 2.4 of IEEE Std 291-1991).
The other methods are standard field method (SFM) (IEEE Std 291-
1991), standard antenna method (SAM) (IEEE Std 149-1979), and standard
transmitting loop method (STLM), a specialized version of the SAM (IEEE Std
291-1991) [29]. One important note is that even if a perfect free-space AF has
been obtained, one should still apply the CFs provided by ANSI/IEEE C63.5
for the NSA test [13], because NSA tests are not in free space. The CFs are
used to correct the influences from the test setup (i.e., the differences between
free space and the specific geometry of an NSA setup). Finally, antennas with
high gain/narrow beamwidth (such as horns) do not see the ground plane when
placed in a close distance; therefore, calibration is carried out under free-space
conditions.
Antenna calibration is simply measuring AF.
High gain antennas have narrow radiation beams and therefore may be
calibrated in free space when elevated slightly above the ground because an el-
evated narrowbeam antenna does not suffer from ground reflections.
5.5 Normalized Site Attenuation

Normalized site attenuation (NSA) is first obtained for the antenna calibration,
and then a correction factor is added.
An OFTS that complies with CISPR or ANSI/IEEE NSA specifications
is required in order to perform antenna calibrations. In addition, the following
instruments and components were required:
• A broadband signal generator and a spectrum analyzer;

• Two 10-dB padding attenuators, amplifiers, or preamplifiers;
• A mast capable of scanning the receive antenna from 1m to 4m in height
and a fixed mast for the transmit antenna;
• Floor absorbers for horn antennas.
NSA has become the standard method for determining the adequacy of
OFTS to perform EM emission and immunity measurements. NSA is defined
as the ratio of the power input of a matched, balanced, lossless, tuned dipole ra-
diator to that at the output of a similarly matched, balanced, lossless, tuned di-
pole receiving antenna for specified polarization, separation, and heights above
a flat electromagnetically reflecting surface [16]. It is a measure of the transmis-
sion path loss between two antennas.
The theoretical method proposed by ANSI/IEEE for the evaluation of
NSA is obtained with a geometrical-optics (GO) approximation based on
the Friis equation since Norton surface waves are negligible for these antenna
heights and frequency bands [21, 29]. Standards define the NSA as the available
input power to a short dipole for the field strength of 100 µV/m at a distance d
= 3m over a leveled conducting ground screen [29].
5.5.1 Theoretical NSA Calculations

There are two widely accepted NSA calculation approaches. The first one, con-
sidering the ground plane as infinitely conducting, proposed by ANSI/IEEE, is
based only on the far-field term. The second one takes also near-field contribu-
tions into account, assuming that far-field radiation in the 30–200 MHz range
is inappropriate when distance is equal to 3m or 10m as the test procedure in
OFTS imposes. The disadvantages of the ANSI/IEEE models are elimination
of near-field terms of the EM field as well as elimination of the Norton surface
wave component that can cause discrepancies up to 1.5 dB, especially for verti-
cal polarization [17, 20].
The theoretical NSA calculations is based on the 2-Ray GO scenario de-
picted in Figure 5.7 and is given as
Figure 5.7 The scenario for the 2-ray method.
279.1
NSA = (5.33)
f (MHz )E D max ( µV /m )
where
+ 2d1 d 2 ρh cos {φh − β (d 2 − d1 )}

2
49.2 d 22 + d12 ρh
E Dh = (5.34)
d1d 2
+ 2d13d 23 ρv cos {φv − β (d 2 − d1 )}

2
49.2d 2 d 26 + d16 ρv
E Dv = (5.35)
d13d 23
are horizontal and vertical electric field components [24]. The other parameters
in these equations are:
d1 = d 2 + (ht − hr ) (direct-ray path )

2
(5.36)
d 2 = d 2 + (ht + hr ) (reflected-ray path )

2
(5.37)
 h + hr 
g = tan −1  t
 d 
(reflection angle ) (5.38)
sin g − ( εr − j 60σλ) − cos2 g

ρh = (hor reflection coefficient ) (5.39)
sin g + ( εr − j 60σλ) − cos2 g
( εr − j 60σλ)
( εr − j 60σλ) sin g −
− cos 2 g
ρv = ( ver reflection coefficient ) (5.40)
( εr − j 60σλ)
( εr − j 60σλ) sin g +
− cos 2 g
2π  Im { ρh ,v } 
β= ; φh ,v = tan −1  
 Re { ρh ,v } 
λ (5.41)
In (5.33), EDmax is the maximum electric field strength (in µV/m) for the
receive antenna height scanned between 1m and 4m from a theoretical half-
wave dipole with 1 pW radiated power and 1.64 dB antenna gain calculated
from (5.34) and (5.35) for horizontal and vertical polarizations, respectively.
The calibration steps are as follows:
• Fix the transmit antenna height (ht).

• Lower the receive antenna height (hr) to the minimum height.
• For these heights, calculate the direct (d1) and ground-reflected (d2) ray
paths, and the angle of reflection (Γ).
• For the given electrical parameters of the ground (εr, σ) calculate hori-
zontal (ρh) and vertical (ρv) reflection coefficients. If the ground is PEC
then these will be: ρh = –1 and ρv = 1 .
• Use (5.34)–(5.35) and calculate EDh and EDv.
• Gradually change the receive antenna height from 1m to 4m and repeat
the above steps; record all EDh and EDv values. Then, choose maximum
EDh and EDv values and calculate NSA from (5.33).
A short MATLAB NSA calculator is prepared and is given in Table 5.1.

This may be used to prepare theoretical NSA tables where users can directly
produce frequency versus NSA, EDmax , and hmax data in four columns; here,
hmax is the maximum receive antenna height at the specified frequency in which
EDmax is recorded. This calculator can be used for any permittivity and con-
ductivity of the leveled ground. A typical example is given in Table 5.2. NSA
versus frequency for both vertical and horizontal polarizations at 10m distance
are shown in Figure 5.8.
NSA in free space, in the near field, can also be calculated from [25]
Table 5.1
MATLAB Code for Theoretical NSA Calculations
%---------------------- Program : LS_NSA.m--------------------------
clear all; clc;
fprintf(‘============================== \n’);
fprintf(‘(1) Ver Pol (TM) \n’); fprintf(‘(2) Hor Pol (TE) \n’);
fprintf(‘============================== \n’);
nx = input (‘Your Choice? : ‘); ht = 1; d = 10; er = 15; sigma = 0.0; k=0;
for freq=3e7:1e7:1e9 % Frequency loop
k=k+1; k0=2*pi*freq/3e8; lambda=3e8/freq; Edmax(k)=1e-6;
Hrmax(k)=1; ep0=8.854e-12; n2=er-j*60*lambda*sigma;
for hr=1:.02:4 % Receive antenna height
d1=sqrt(d^2+(ht-hr)^2); d2=sqrt(d^2+(ht+hr)^2); theta=(atan(ht+hr)/d);
if nx==1 % VER Polarizasyon
rho=(n2*sin(theta)-sqrt(n2-cos(theta)^2))/(n2*sin(theta)+sqrt(n2-cos(theta)^2));
rhom=abs(rho); rhoa=angle(rho); rh(k)=rho;
Ed=sqrt(49.2)*d^2*(d2^6+d1^6*rhom^2+2*d1^3*d2^3*...
rhom*cos(rhoa-k0*(d2-d1)))^(0.5)/(d1^3*d2^3);
elseif nx==2 % HOR Polarizasyon
rho=(sin(theta)-sqrt(n2-cos(theta)^2))/(sin(theta)+sqrt(n2-cos(theta)^2));
rhom=abs(rho); rhoa=angle(rho); rh(k)=rho;
Ed=sqrt(49.2)*(d2^2+d1^2*rhom^2+2*d1*d2*...
rhom*cos(rhoa-k0*(d2-d1)))^(0.5)/(d1*d2);
end
if Edmax(k)<Ed; Edmax(k)=Ed; Hrmax(k)=hr; end
end
f(k)=freq/1e6; nsa(k)=279.1/(f(k)*Edmax(k));
end
figure(1); plot(f,20*log10(nsa),’LineWidth’,2);
xlim([30,1000]); ylim([-20,30]); grid on;
xlabel(‘Frequency [MHz]’); ylabel(‘NSA [dB]’); % Program END
Table 5.2
NSA Versus Frequency over Metal
Ground (Ver Polarization, d = 10m,
ht = 1m)
Frequency Edmax NSA
[MHz] [dBµ/m] [dB]
30 2.671 16.702
50 2.641 12.295
80 2.567 8.286
100 2.499 6.416
200 1.917 0.978
400 1.009 -4.135
500 1.668 -6.732
600 2.021 -8.669
800 2.368 -11.515
900 2.457 -12.627
1000 2.523 -13.608
Figure 5.8 A typical NSA versus frequency variation over a metal ground.
 
 5Z 0   1   d

NSA = 
 2 π   f (MHz )   −4  (5.42)
 1 − ( βd ) + ( βd ) 
−2
for short antennas. In the far field, 1/d2 and 1/d4 terms in (5.42) can be ne-
glected and the free-space theoretical NSA equation becomes:
 5Z   d 
NSA =  0  
 2 π   f (MHz )  (5.43)
The logarithmic form of (5.43) for impedance matched (Z0 = 50Ω) sys-
tems will then be:
NSA = 32 + 20 log (d m ) − 20 log ( f MHz ) (5.44)
Table 5.3 lists NSA values computed from (5.9), (5.10), and (5.11) in the
free space. The free-space NSA from (5.9) may be obtained directly by setting
relative permittivity and conductivity values to one and zero, respectively.
As observed, using (5.11) instead of (5.10) introduces significant errors
that are not acceptable [20] for short range measurements. For d = 3m, this
error is 1 dB and maximum at 30 MHz. This error is less than 0.1 dB above
60 MHz at d = 5m, and above 110 MHz at d = 10m. Therefore, the free-space
NSA equation with only term can be used for EMC and AF measurements in
the frequency range of 30 MHz–1 GHz at d = 10m and d = 30m, but a CF
Table 5.3
NSA Versus Frequency in Free Space NSA1
(5.33), NSA2 (5.42), NSA3 (5.43), Ver-pol,
d=3m, hv=2m)
Frequency NSA1 NSA2 NSA3
[MHz] [dB] [dB] [dB]
30 12.00 12.98 12.00
50 7.56 7.97 7.56
70 4.64 4.94 4.64
90 2.45 2.59 2.46
110 0.71 0.80 0.71
130 –0.74 –0.68 –0.74
150 –1.98 –1.94 –1.98
170 –3.07 –3.03 –3.07
190 –4.04 –4.01 –4.03
200 –4.48 –4.46 –4.48
should be added at shorter ranges for lower end of this frequency band. The
CF includes not only the effects of 1/d 2 and 1/d 4 terms, but also nonuniform
radiation of receive antenna, mutual antenna couplings, and mutual coupling
between antennas and the ground [20]. Further investigation of theoretical
NSA models including Norton surface wave contributions is presented in [19].
5.5.2 NSA Measurements

NSA measurements can be performed in accordance with standards CIS-
PR16-1-4/5 [24, 25], ANSI/IEEE C63.4 [26], and ANSI/IEEE C63.5 [27].
Two antennas are set up on the test site in an appropriate geometry as shown
in Figure 5.7. The NSA procedure requires two different measurements of the
voltage VR at receiver. The first VR reading belongs to a case where the two
antennas are removed and coaxial cables are directly connected to each other
via an adapter. The second VR reading belongs to the case where antennas are
connected to their coaxial cable, and maximum signal is scanned in height. For
both of these measurements, the signal source VI, is kept constant. The first
reading of VR is called VDirect , and the second is VSite. Measured NSA is then
extracted from
NSA = V Direct − VSite − AFT − AFR − ∆AFTOT (5.45)
where AFT and AFR are antenna factors of the transmit and receive antennas (in
dBm–1), and AFTOT is the mutual impedance CF (in dB). Note that the dif-
ference VDirect – VSite is equal to the classic site attenuation.
The single-point NSA measurement stated earlier is valid only for OFTS.
For the NSA measurements in a semi-anechoic chamber, a single-point NSA
measurement is insufficient to pick up possible reflections from the construc-
tion or RF-absorbing materials comprising the walls and ceiling of the test site.
For anechoic chambers, a test volume is defined [27] as that volume traced out
by the largest equipment or system to be tested as it is rotated about its center
location through 360°, such as by a turntable. In evaluating the site, the trans-
mit antenna should be placed at various points within the test volume with
both horizontal and vertical polarizations, such as illustrated in Figure 5.9. This
may require a maximum of 20 separate site attenuation measurements—that is,
five positions in the horizontal plane (center, left, right, front, and rear, mea-
sured with respect to the center and a line drawn from the center to the position
of the measuring antenna), for two polarizations (horizontal and vertical), and
for two heights (1m and 2m, horizontal) (1m and 1.5m, vertical) [24, 27].
These measurements are carried out with broadband antennas. The dis-
tance d is measured from the center of transmit antenna to the center of receive
antenna. This distance shall be maintained for all measurements, which requires
that the receive antenna be moved along the line in the directions shown in the
figure. Also, transmit and receive antennas shall be aligned with the antenna
elements parallel to each other and orthogonal to the measurement axis. Fur-
thermore, the lower tip of the antenna shall be at a distance greater than 25 cm
from the floor, which may require the center of the antenna to be slightly higher
than 1m for the lowest height measurement.
Both CISPR and ANSI/IEEE necessitates NSA of an OFTS must be
within ±4 dB of the theoretical values in order to accredit a CALTS. It is rec-
ommended that this should be within ±1 dB to reduce the measurement un-
certainty significantly. There are currently no CALTS validation requirements
above 1 GHz. However, facilities suitable for measurements in the frequency
range 30 MHz to 1000 MHz are considered suitable for the frequency range 1
GHz to 40 GHz, including the presence of the reference (metal) ground plane.
Figure 5.9 NSA measurements in an anechoic chamber.

5.5.3 Performing an Antenna Calibration

Antenna calibration is simply measuring the AF. The free space AF is the AF that
is not influenced by adjacent objects. The number and types of the antennas
at hand determines the method one should apply. Here are some possibilities:
• There are three or more antennas with the same frequency band; how-
ever, they may or may not be identical.
• There are two antennas with the same frequency band; however, they
may or may not be identical.
• There are two nonidentical antennas with the same frequency band. In
addition, one of them already has a recent calibration certificate from an
accredited lab.
The three-antenna calibration (TAC) method in accordance with the stan-

dard ANSI/IEEE C63.5 seems an appropriate approach. Also, one could use
the absolute gain method (AGM) provided in ANSI IEEE 149 in an alternative
way. The TAC method (based solely on horizontally polarized measurements)
provides AF measurements from 30 MHz to 1000 MHz. The distance is 10m,
the transmitting antenna height is 2m, and the receiving antenna scans heights
from 1m to 4m.
The TAC method requires a CALTS and provides near free-space AFs
for many antennas including biconical and tuned dipoles, log periodic dipole
arrays, and linearly polarized hybrid arrays. The free-space CFs for biconical
dipoles are provided in ANSI/IEEE C63.5. The mutual coupling CFs for tuned
dipoles are also provided in the same ANSI/IEEE standard. The near free-space
AFs for other types of antennas can be used without further corrections.
Using this method, AFs shall be determined only for horizontal polariza-
tion on an OFTS. Horizontal polarization measurements are relatively insensi-
tive to site variations and yield acceptable AFs even though the reflecting plane
does not create a free-space environment during calibration. Horizontal polar-
ization is preferred for antenna calibration because:
• Antenna-cable mutual coupling and scattering from the cable are neg-
ligible.
• Horizontally polarized ground reflection is less sensitive to differences
in the ground plane, as well as to differences in the ground conductivity
and permittivity.
• Ground screen edge diffractions are smaller for horizontal polarization.
The TAC method requires three site attenuation (A) measurements under
identical geometries (ht, hr, d), using three antennas at hand in pairs. The three
equations associated with these measurements are:
AF1 + AF2 = A1 − 48.92 − 20 log ( f MHz ) + E D max (5.46)
Here, AF1, AF2, and AF3 are the antenna factors of antennas 1, 2, and 3
(in dBm-1) to be calculated; A1, A2, and A3 are the measured site attenuations
(in dB). Rearranging these equations yields the AF (in dBm-1) of the three an-
tennas as:
E D max + A1 + A2 − A3
AF1 = 10 log ( f MHz ) − 24.46 + (5.49)
2
E D max + A1 + A3 − A2
AF2 = 10 log ( f MHz ) − 24.46 + (5.50)
2
E D max + A2 + A3 − A1
AF3 = 10 log ( f MHz ) − 24.46 + (5.51)
2
Antenna separation of 10m is recommended for the antenna calibration.

Values of EDmax over metal ground planes are theoretically calculated from
(5.33)–(5.35).
Antenna calibration is simply AF measurement based on theoretical NSA
calculations.
Broadband (narrowbeam) horn antennas may be calibrated using the same
TAC method, but receive antenna height scan is not necessary since the ground
reflection point is not within the beams. This could be verified by performing
the calibration at two significantly different heights and verifying that the dif-
ferences between the two resulting AFs are within the expected measurement
uncertainty. The calibration distance of horn antennas is d ≥ 2D2/l, where D
is the largest linear dimension of the aperture of the antenna and λ [m] is the
wavelength, both in meters. Note that antenna calibration at a smaller distance
shall only be used at the calibrated distance.
If two identical antennas are to be calibrated, their AF can be obtained

from a single site attenuation measurement using the following expression:
E D max + A
AF = 10 log ( f MHz ) − 24.46 + (5.52)
2
In practice, two antennas are never identical, and the AF calculated from
(5.20) is the geometric mean (in linear units) of the individual factor for each of
the two antennas. Certain antennas can be constructed to be so nearly identical
that their AFs are different by much less than the measurement uncertainty.
If the two antennas are not identical, an antenna that has the same frequency
range can be borrowed from another EMC lab to do the calibration by using
the TAC method. If there are two antennas with the same frequency band and
one of them already has a recent calibration certificate from an accredited lab,
then site attenuation A can be measured and then the AF of unknown antenna
can be calculated from
AF1 = A + 20 log ( f MHz ) − 48.92 + E D max − AF2 (5.53)
where AF2 is the AF of the certificated antenna. Table 5.4 lists a typical antenna
calibration result of a Schaffner BHA9118 horn antenna.
5.5.4 Antenna Calibration with Pattern Measurements

The ANSI/IEEE C63.5 standard aims only at the specification of AF. If other
parameters such as gain, directivity in any direction, dielectric, and conduc-
tive losses are of interest, then a new method—antenna pattern measurement
(APM)—presented here can be used [17]. The APM method seems to be a bet-
ter approach to calibrate antennas if extra information is required. The classical
ANSI/IEEE C63.5 method yields only the AF and absolute gain in the maxi-
mum radiation direction. It yields no information about directivity and gain in
any desired direction, dielectric/conductive losses, and radiation pattern.
Antenna pattern is a function of theta (θ) and phi (φ) angles for a con-
stant radial distance and frequency. The APM method is based on antenna-
received-power measurements around the antenna on the spherical surface en-
closing the antenna. A simple setup is shown in Figure 5.10. The number of
discrete points (discretization) in both azimuth and elevation is determined
from a specified model error limitation. The pattern of an antenna can be mea-
sured in both transmit or receive mode. If the antenna is reciprocal, as is the
case for most practical antennas, then either transmit or receive mode can be
Table 5.4
Typical Antenna Calibration Results
ANSI/IEEE
BHA 9118 ANSI/IEEE C63.5 149|G| [dB]
Horn Schaffner Specs TAC Method Method Difference
Freq. [GHz] AF [dB] |G| [dB] AF [dB] |G| [dB] |G| [dB] ∆G [dB]
1 24.37 5.83 24.50 5.71 5.73 0.022
2 27.28 8.94 27.09 9.14 9.16 0.022
3 29.94 9.80 29.87 9.87 9.91 0.024
4 32.62 9.62 32.45 9.80 9.83 0.023
5 33.01 11.16 33.76 10.43 10.45 0.022
6 34.27 11.49 34.48 11.29 11.31 0.022
Figure 5.10 APM method and power measurement setup.
utilized. However, the advantage of the transmit mode is to find the possibility
to measure the absolute radiated power in addition to the pattern, which in
turn also yields directivity and gain.
In addition to the instruments listed earlier in the TAC method, a cali-
brated receive antenna in the same frequency band, a positioning system for θ
and φ variations, a recording system, and data processing software are also re-
quired for the APM method. Radiation patterns of most practical antennas are
so complex that closed form mathematical expressions are not available.
The radiated power of an antenna can be calculated from the double inte-
gral over a sphere enclosing the antenna if the 3D directional radiation function
F(θ, φ) is known. In discrete form, this integral reduces to a double summation:
M N 
Prad = Bo ∆θ∆φ ∑  ∑ F ( θi , ϕ j )sin θi  (5.54)
j =1  i =1 
(B0 is constant) and can be used to calculate radiated power once the pat-
tern F(θi, φi) is measured at a number of points θi and φi:
π π π
N = , θi = i ∆θ = + (i − 1) , i = 1,2,, N (5.55)
∆θ N N
2π 2π 2π
M= , φ j = j ∆φ = + (i − 1) , j = 1,2,, M (5.56)
∆φ M M
It has been tested with a number of different antenna patterns that radia-
tion pattern discretization error (i.e., error introduced using (5.54) instead of
using the mathematical exact double-integral) is less than 0.5 dB if N and M are
between 10–15 and 20–30, respectively. Once the radiated power is found the
directivity at any direction can be calculated from [12]
4 πU ( θ, φ)
D ( θ, φ) = (5.57)
Prad
where radiation intensity U(θ, φ) is obtained from antenna pattern measure-

ments (maximum directivity Dmax corresponds to Umax).
The next step is to find the conductive and dielectric losses T (in dB),
which can be obtained from the transmit antenna input power. The setup
shown in Figure 5.11 can be used for the input power measurements. The con-
ductive and dielectric losses are solely the difference between radiated power
and the input power:
T = PInput − PRad
(5.58)
In power measurement setup shown in Figure 5.11, the antenna whose

pattern is measured is in the transmitting mode. The AF of the calibrated an-
tenna in the receiving mode and the loss of the coaxial cable that connects the
receiving antenna to the spectrum analyzer should be precisely known in order
to exactly determine the absolute radiated power.
The input power of the transmit antenna is then obtained from
PInput = PFwd − PRvs (5.59)
where PFwd and PRvs are the powers measured at ports 1 and 2 as shown in the
figure, respectively. Absolute gain of the antenna (in dB) is then equal to
Figure 5.11 The setup for antenna input power measurement.
G = D − T , G max = Dmax − T (5.60)
Finally, the AF of the antenna (in dBm–1) is calculated via [20]
AF = 20 log ( f MHz ) − G max − 29.79 (5.61)
It should be noted that as an AF measuring method, the APM needs

longer measurement time when compared with the other ANSI/IEEE C63.5
methods. But the ANSI/IEEE C63.5 TAC method gives only the AF whereas
the APM method yields many other antenna parameters (such as antenna loss-
es, directivity and gain in any direction, and 2D and/or 3D radiation patterns)
with a single measurement.
The final statement of the result of a measurement is complete only if
it contains both the value attributed to the measure and its uncertainty. An
uncertainty report should be given for every antenna measurement. Gener-
ally speaking, the uncertainty is expressed to no more than two significant
digits. Unless otherwise specified, the result of the measurement should be re-
ported, together with the expanded uncertainty appropriate to the 95% level of
confidence [31]. Components of uncertainty are evaluated by the appropriate
method, and each is expressed as a standard deviation and is referred to as a
standard uncertainty. The standard uncertainty components are combined to
produce an overall value of uncertainty, known as the combined standard un-
certainty. An expanded uncertainty provides a greater interval about the result
of a measurement than the standard uncertainty with a higher probability. It
is obtained by multiplying the combined standard uncertainty by a coverage
factor, k. Choice of the factor is based on the coverage probability or level of
confidence required. Table 5.5 lists components of the uncertainty with total
uncertainty budget. As shown there, combined and expanded uncertainties are
0.87 dB and 1.74 dB, respectively.

Table 5.5
Uncertainty Budget for Antenna Calibration
Ci Ui
Source of Probability (Sensitivity (Uncertainty
Sym Uncertainty Unit Value Distribution Divisor Coefficient) Component) U i2
dM Model Error dB 0.2 Rectangular 1,732 1 0.115 0.0132
dY Site reflections dB 0.2 Rectangular 1,732 1 0.115 0.0132
dIG Ap dB 0.1 Normal 2 2 1 0.05 0.0025
dIF dB 0.1 Rectangular 1,732 1 0.06 0.0036
dAG dB 0.2 Normal 2 2 1 0.100 0.01
dFH dB 0.2 Rectangular √3 1 0.115 0.0132
dP dB 0.1 Rectangular √3 1 0.060 0.004
dCP dB 0.1 Rectangular √3 1 0.060 0.004
dMS dB 0.3 Rectangular √3 1 0.173 0.030
dA dB 0.2 Rectangular √3 1 0.115 0.0132
dB dB 0.2 Rectangular 1 0.115 0.0132
Antennas and Antenna Calibration
√3
dAF dB 1 Normal 2 2 1 0.500 0.250
dKK dB 0.2 Normal 2 2 1 0.100 0.010
dU dB 0.5 U-Shaped √2 1 0.353 0.125
dYK dB 0.1 Rectangular √3 1 0.060 0.004
dUYK dB 0.1 U-Shaped √2 1 0.070 0.005
dT dB 0.5 Normal 1 1 1 0.500 0.250
Combined standard uncertainty: 0.87 dB. Expanded uncertainty: 1.74 dB
131
5.6 Loop Antennas

Loop antennas (magnetic field sensors) are widely used in military and civil
test applications in order to measure magnetic fields below 30 MHz. They are
important especially in EMC tests and measurements. This section is devoted
to loop antenna calibration.
Loop antennas are calibrated according to either of the standards IEEE
291 [33] and SAE-ARP 958 [34].
The standards CISPR 16-1-4 [25] and ANSI C63-4 [27] specify the
use of loop antennas for magnetic field emission measurements between 30
Hz–100 kHz. MIL-STD 461E/F [35, 36] requires the use of loop antennas
in several tests. One of the important tests, RE101, requires a shielded loop
antenna with 36 turns with the diameter of 13.3 cm. The aim of this test is
to measure magnetic RE from equipment and subsystem enclosures including
electrical cable interfaces in the frequency range 30 Hz–100 kHz. Limit levels
of the RE101 measurements are given in Figure 5.12 in terms of magnetic flux
density in dBpT with a measured sample device RE curve, which is below the
limits. As shown, the limit level for all army applications is 180 dBpT at 30 Hz
and reduces to 110 dBpT at 100 kHz. For the navy applications, these values
are 160 dBpT and 78 dBpT, respectively.
Calibration of RE101 loop antennas are defined in SAE-ARP 958 [34]
at 12 cm calibration distance by using a reduced calculation equation; there-
fore, calibration factors deviate from the normal calibration factors as will be
Figure 5.12 Typical RE101 measurement results and the thresholds.

experimentally shown in the following sections. The factors of the RE101 an-
tenna are often presented in unit of dBpT/µV since the limits in the standard
are in the unit of dBpT. The RE101 test setup is shown in Figure 5.13.
Another test in MIL-STD 461E/F that requires the use of loop antennas
is the RS101 test. This test defines two loop antennas. Unlike RE101 test, this
is an immunity test and aims to expose the EUT to magnetic radiated emis-
sions in the frequency range 30 Hz–100 kHz. One of required antennas by
the RS101 test is a transmitting loop antenna with 20 turns and a diameter
of 12 cm for exposing EUTs to magnetic fields. This loop is also chosen as a
transmitter by SAE-ARP 958 standard for the calibrations of RE101 receiving
loop antennas. The other antenna required by the RS101 test is a receiver loop
with 51 turns and a diameter of 4 cm for verifying the test setup before the
actual test. Unlike other immunity tests like RS103, during the RS101 test, no
feedback process is applied to measure the produced magnetic field, as shown
in Figure 5.14.
The current that flows on the transmitter loop is measured by means of
a current probe, and produced magnetic field on the EUT is then calculated
using this current. The precise measurement of the current therefore becomes
essential for the reliability of the test. Before the actual RS101 test, a verification
process must be applied to verify all the test setup by using the receiving loop
as shown in Figure 5.15.
In addition to military applications, loop antennas are commonly utilized
in SE measurements below 30 MHz [37] (see Figure 5.16). In SE measure-
ments, the calibration factors are not needed. The comparison of two measure-
ment cases with and without the target shield, as in Figure 5.16, yields the
desired SE value.
Figure 5.13 The setup for the RE101 measurements.

Figure 5.14 The setup for RS101 measurements.
Figure 5.15 The RS101 verification setup.

Figure 5.16 Screening effectiveness (SE) measurement setup [37].
5.7 Loop Antenna Calibration

As discussed in [1] (and earlier here), an antenna is a transducer. The transmit
antenna converts the voltage between its terminals to E-field (V → E). The re-
ceive antenna does the opposite. This transformation is modeled with a critical
parameter called AF. AF definitions of both E- and H-field sensors are shown
in Figure 5.17.
Loop antennas, in shapes such as circular, triangular, square, and ellipti-
cal, and so on, shapes are simple, compact, low cost, and versatile transduc-
ers, that are often used to capture EM energy, sensitive to the magnetic fields.
Electrically small loops (with circumference much less than the wavelength, l
<< λ) have very small radiation resistance; therefore, they are very often used
as receiving antennas (e.g., direction finding and/or EMC tests) with their very
poor efficiency below 30 MHz. Regardless of their shape, they have a far-field
pattern very similar to that of a small dipole (aligned normal to the plane of the
loop). Their input impedance is very low and near magnetic field is dominant.
As the circumference of the loop increases (i.e., when l ~ λ), the pattern maxi-
mum shifts toward the normal of the loop plane. The open circuit (OC) voltage
across a loop is [38]:
VOC = 2 πANfB (5.62)
where A, N, f, and B are the area of the loop in square meters, the number of
turns on the coil, the frequency in Hz, and the magnetic flux in Tesla, respec-
tively. The voltage VL across the load RL , when the loop antenna (with the coil
impedance ZC = RC + jXC) is loaded, will be
Figure 5.17 AFs of (a) electric and (b) magnetic sensor.
VL VOC
= (5.63)
RL RL + RC + jX C
from where loop antenna AF in pT/µV is obtained as
B (1 + RC / RL )2 + (2 πfL / RL )2
AFH = = (5.64)
VL 2 πfAN
(2 pT = 0.796µA/m or 1µA/m = 1.256pT; in other words, 0dBµA/m = 1.98

dBpT). The AFs of loop antennas are commonly expressed as AF = H[µA/m]/
V[µV] in dBS/m or as AF = B[pT]/V[µV] in dBpT/µV. The transformation
from dBpT/µV to dBS/m, or vice versa is possible as follows:
B [T ] = µo H [ A / m ] = 4 π 10 −7 H [ A / m ]
B [ pT ] = 4 π 10 −7 H [ A / m ]1012 = 4 π H [ A / m ]105
B [ pT ] H [A / m ] S
AFH [dBpT / µV ] = = 1.256 = 1.256( )
V [ µV ] V [V ] m
Hence, one subtracts 1.98 dB from dBpT/µV and obtains dBS/m. Some-
times, AFH is expressed in terms of equivalent electric field intensity (AFE =
E[µV/m]/V[µV]). In this case, the unit of the AF is dBm–1 (i.e.,AFE[dBm–1]
= AFE[dBS/m] + 51.5). A typical AF graph is given in Figure 5.18. In order
to obtain magnetic flux density at a given frequency, AFH value read from the
Figure 5.18 AFH graphics of the SE E-7334-1 antenna.
graphics and the dBµV value read from the EMI receiver are added. For ex-
ample, at 10 kHz, AFH is read from the graphics as 35 dBpT/µV. If the EMI
receiver shows 23 dBµV, then the measured magnetic flux density will be 58
dBpT.
Details of the loop antenna calibration according to IEEE 291 [33] and
SAE-ARP 958 [34] are given in [38].
5.8 Antenna Arrays

Radiation characteristics of an antenna can be changed by using multiple ele-
ments in the near field and by adjusting their feeding phases. These are called
antenna arrays or directive antennas. Radiation beam-forming and beam steer-
ing can be done both mechanically and electronically. Antenna systems with
electronic beam forming and beam steering capabilities are called phased arrays.
The pattern definitions given in Figure 5.3 are also valid for antenna arrays.
According to their radiation characteristics, antennas can be grouped into
two: broadside antennas and endfire antennas. Figure 5.19 presents these two
types. Here, the z-axis shows an antenna axis for a single antenna (e.g., wire
antenna) but the axis of the array for antenna arrays.
Antenna radiation patterns are given in 3D spherical coordinates with
respect to horizontal angle (φ) and vertical angle (θ). Two slices are of interest:
Figure 5.19 Basic antenna arrays.
horizontal and vertical. For example, in order to plot horizontal radiation pat-
tern, θ is kept constant (θ = 90°) and φ is changed from 0° to 360° with a given
angular resolution ∆φ. Similarly, φ angle is kept constant (φ = 0° for xz-plane
or φ = 90° for yz-plane) and θ angle is rotated 360° with an angular resolution
of ∆θ.
5.8.1 Arrays with Isotropic Radiators

The radiation pattern of a group of radiators (i.e., array) is determined by the
type of individual elements, their location, spacing, orientation, excitation am-
plitudes, and phases. In practice, the radiation pattern of an array is divided
into two: element pattern and array factor. Array factor is the radiation pattern
of an array, where all array elements are isotropic radiators. An isotropic radiator
is a hypothetical, lossless antenna occupying a point in space and when trans-
mitting radiates uniformly in all directions. Figure 5.20 shows geometrical pa-
rameters associated with this problem. In general, a radiation source is located
at (x1, y1, z1) with respect to a phase reference at the origin (0, 0, 0). When it
is an array of N-elements, the coordinates of each radiator with respect to the
phase reference are (xi, yi, zi) (i = 1, …, N) and the total electric field calculated
at a far field point P is [39]
N
E ( θ, ϕ) = ∑ I i e j ψi ( θ, ϕ) (5.65)
i =1
Here, Ii is the complex current for the ith element and ψi is phase con-
tribution at the far field point from the ith radiator with respect to the origin,
which can be expressed as:
Figure 5.20 An array of isotropic elements.

 r 
ψi = k  ri ⋅   (5.66)
 r 
Here, j = −1, k = 2π/λ is the wave number; λ is the wavelength,

   
ri = x i e x + yi e y + z i e z is the vector from the origin to the ith element (radia-
   
tor vector), r = xe x + ye y + ze z is the vector from the origin to the observation
  
point P (radiation direction vector), and (e x , e y , e z ) are the unit vectors along x,
y, and z, respectively. Using the spherical coordinates, with x = r sin θ cos φ, y
= r sinθ sin φ, and z = r cosθ, the phase of the ith radiator may be obtained as
ψi = k (x i sin θ cos φ + yi sin θ sin φ + z i cos θ ) (5.67)
For planar arrays located on the xy-plane, the radiators’ coordinates are xi = ri
cosφι, yi = ri sinθi, and zi = 0, where ri = x i2 + yi2 and φi = a tan–1(yi /xi). For
the planar arrays (i.e., θi = 90°) the total electric field, which is known as the
array factor, can be expressed as:
N
E ( θ, ϕ) = ∑ I i exp ( jkri sin θ cos( ϕi − ϕ)) (5.68)
i =1
By properly changing the phases of the feedings, this equation becomes
N
 
E ( θ, ϕ) = ∑ I i exp  jkri [ sin θ cos( ϕi − ϕ) − sin θ0 cos( ϕi − ϕ0 )]  (5.69)
i =1  
if maximum radiation is desired to point (θ0, φ0). Figure 5.21 shows broadside
and endfire radiation patterns obtained with the same array of five isotropic
radiators; only the phases of the radiators are changed accordingly.
5.8.2 A MATLAB-Based ARRAY Package

A MATLAB-based simple and useful array package, ARRAY, has been devel-
oped in [39]. It has been prepared for the visualization of radiation charac-
teristics of planar arrays of isotropic radiators. This package allows the user
to choose from a popup menu: (1) arbitrary array; (2) linear array; (3) planar
array; and (4) circular array, with a number of user-selected and located isotro-
pic radiators. The default array type is arbitrary. One of the exciting features of
ARRAY is to let the user locate the radiators arbitrarily. Once the number of
radiators (N), maximum radius (in meters) of the polar region where the radia-
tors will be located, and the operating frequency (in MHz) are specified, the
user may proceed to locate the radiators one by one after pressing the “Locate
Radiators” button. The coordinates of the located radiators are also displayed in
the list-box at the bottom.
With the help of the mouse, the user may design any kind of an array, but
the package also locates the radiators automatically when one of the other types
of arrays is selected and required parameters are supplied. If the array type is set
to linear, N-elements are located on the y-axis symmetrically, with an interele-
ment distance of d. If it is planar, N × M elements are located on xy-plane. In
this case, N and M are the number of the elements along y- and x-directions,
respectively. The distances between the radiators along x- and y-directions are
dx and dy, respectively. If the circular array is selected, N-elements are located
symmetrically on a circle whose center is at the origin with radius r.
After locating the radiators, the user may visualize a 2D radiation pattern
at the selected plane by pressing the “Draw Graph” button. The 2D radiation
pattern plane to be displayed is specified by the Theta/Phi Plane popup menu
Figure 5.21 Center-fed, five-element arrays: (left) broadside; (right) endfire.

and Theta/Phi value box. If the θ-plane is selected (vertical pattern), the angle
θ varies between [0°, 360°] at a fixed φ value (specified by the Theta/Phi box).
If the φ-plane (horizontal pattern) is selected, the angle φ varies between [0°,
360°] at a fixed θ-value specified by the user. All the 2D radiation patterns are
plotted with an angular resolution of ∆θ = ∆φ = 1°. Figure 5.22 presents a
circular array with maximum radiation along (θ0 = 45°, φ0 = 45°) direction.
Figure 5.23 shows the vertical radiation pattern of 10×10 element planar
array of isotropic radiators. Note that radiation patterns may also be plotted in
3D [39].
Another exciting feature of ARRAY is that the user can use sliding bars to
change N, f, θ0, φ0. For instance, by pressing the sliding bar of the frequency
continuously, the user can increase or decrease the frequency, see the corre-
sponding radiation pattern immediately, and may easily understand the effect
of frequency change on the radiation pattern. The package also allows the user
to save the radiation pattern data (field intensity versus angle) to a file named
Arraypattern.dat. The first column of this file corresponds to 361 observation
angle values (in radians), and the corresponding array factors are in the second
column.
Figure 5.22 The ARRAY package and circular array characteristics.

Figure 5.23 A 10×10 planar array and its radiation pattern on the yz-plane.
Extra line phasing capability has been included in ARRAY to change FBR
characteristics in the planar array option. The user may change the phase angle
∆β of the lined-up radiators and observe how FBR improves or gets worse. In
this case complex currents Ii in (5.65) is replaced with Iin as
M
I i → I in = ∑ I i exp { j (n − 1)∆β} (5.70)
n =1
In an N × M planar array, if for example, a phase of ∆β=30° is selected,

then the phases of the second row (M=2).lag 30° the phases of the first row
(M=1), the phases of the third row (M=3).lags 30° the phases of the second row
(M=2) and 60° the phases of the first row (M=1), and so on.
The next example in Figure 5.24 belongs to a 2×5 planar array. At left,
horizontal radiation pattern of this array with interelement distance, dy = 0.4m
and interarray distance of dx=0.25m is plotted without phasing the radiators.
This is a broadside radiation with FBR=0 dB. In many applications (e.g., a
shore-based high-frequency ocean surveillance radar), the pattern of the array
is desired to extend only in one direction (e.g., toward the ocean); therefore, a
Figure 5.24 A 2×5 array: (left) broadside; (right) endfire radiation.
large FBR is required. This may be achieved by phasing the second row radia-
tors (and the others if there are more than two rows) with respect to the first
row. This is illustrated at right in Figure 5.24, where a 90° phase-difference is
given for the second row radiators. As observed in the figure, a very good FBR
is achieved.
A short MATLAB code is also given in Table 5.6. This code plots radia-
tion pattern of a circular array of isotropic radiators. The user supplies num-
ber of radiators, radius of the circle, the frequency, and the beam direction. A
sample pattern plotted with this code is shown in Figure 5.25.
Note that side lobes of an array may also be reduced by changing the
radiators’ feeding amplitudes [8–12]. Two examples are given in Figures 5.26
and 5.27. Here, 1D and 2D arrays of isotropic radiators are fed with amplitude
values like in binomial series [12]. Binomial feeding is best in eliminating side
lobes; the payoff is the broader main lobe.
5.9 Antenna Types

An isotropic radiator (point source) is an antenna that can be modeled theoreti-
cally. It radiates equally in all directions. The second-order differential equations
for both vector and scalar potential functions, as discussed in Section 3.3.1, are
used in antenna modeling. With a isotropic radiator on the right-hand side, the
problem is called a Green’s function problem. Radiation patterns of such arrays
of isotropic radiators directly yield array patterns (array factors). Radiation pat-
terns of arrays with any other radiating elements are represented by the product
of element factor and array factor.
Antenna array radiation is the product of element pattern and array
pattern.
Table 5.6
MATLAB Code for a Circular Array Pattern Plot
%---------------------------------------------------------------------
% Program: Circular antenna array
%---------------------------------------------------------------------
clc; close all; clear all;
N = 4; % input(‘Number of radiators = ?’);
r = 0.25; %input(‘Radius [m]= ?’);
fr = 150; % input(‘Frequency [MHz] = ?’);
phi0 = 90; %input(‘Beam direction [degree]= ?’);
c=3e8; fr=fr*1e6; k=2*pi*fr/c; dtr = pi/180;
rtd = 180/pi; phi0r = phi0*dtr;
for i=1:N
phi=2*pi*(i-1)/N; phin(i)=phi; xn(i)=r*cos(phi); yn(i)=r*sin(phi);
end
i = 0;
for phi = 0:1:359
i = i+1; phir(i) = phi*dtr; Af(i) = 0;
for n = 1:N
Af(i)=Af(i)+exp(1i*k*r*(cos(phir(i)-phin(n))-cos(phi0r-phin(n))));
end
Af(i) = abs(Af(i));
end
Array = Af/max(Af);
polar(phir,Array); hold on;
plot(xn*0.25/max(xn),yn*0.25/max(xn),’ro’)
%--------------------PROGRAM END-------------------------
Figure 5.25 A four-element circular array.
The other two basic antennas that are widely used in antenna engineering
and important in EMC engineering are the electric and magnetic dipoles.
Figure 5.26 Side lobe reduction using binomial feeding: (left) 1-1-1-1-1-1-1-1; (right) 1-7-21-
35-35-21-7-1.
Figure 5.27 Side and back lobe reduction using binomial feeding.
5.9.1 Electric and Magnetic Dipoles

An electric dipole is a short wire on which an AC current flows. A magnetic
dipole is formed if the ends of this short wire are connected. In other words, a
magnetic dipole is an AC current carrying a wire loop. Figure 5.28 depicts both
dipoles. Their vertical and horizontal radiation patterns are as shown in Figure
5.29. Dominant radiations of electric and magnetic dipoles come from Eθ and
Hφ, respectively. In the near field, an electric dipole has a dominating electric
field component (high impedance wave); but a magnetic field component is
dominant near a magnetic dipole (low impedance wave).
Field components, radiation resistances, and simple/approximate electro-
magnetic emission formulas useful for EMC engineers are given in Figures 5.30
and 5.31. These formulas may be used to predict worst-case (maximum) emis-
sions of any disturbing EMI source once the type of the source is identified.
Figure 5.28 Electric and magnetic dipoles.
Figure 5.29 Electric and magnetic dipole radiation patterns.
Figure 5.30 Electric dipole characteristics.

Figure 5.31 Magnetic dipole characteristics.
Characteristics of electric and magnetic dipoles are just the opposite in the
near field. In the far field (once the contributions of r –2 and r –3 terms become
negligible), they are both assumed as a plane wave. The impedances of electric
and magnetic dipoles versus normalized distance are plotted in Figure 5.32. As
observed, the wave impedance of an electric dipole is very high (5 kΩ–10 kΩ),
but it is very low for a magnetic dipole (10Ω–30Ω) in the near field. Around
and beyond the normalized distance of λ/2π [m], wave impedances of both
types become free-space wave impedance (η0 = Z0 = |Eθ|/|Hφ| = 377 Ω).
In EMC engineering, the theoretical boundary of the near field is ac-
cepted as λ/2π.
Figure 5.32 Dipole impedances versus distance.

5.9.2 Wire Antennas

Early communication/broadcast antennas were single wire antennas, which
have the advantage of easy matching to the operating frequency and provide
efficient transmit and receive characteristics. They have been used both verti-
cally and horizontally above the Earth (either on the ground or elevated) from
a few hundred kHz up to several hundred MHz frequency range. The radia-
tion power of such antennas depends on their size and operating frequency
(i.e., l-λ relation). A wire antenna with a length much less than the wavelength
(short dipole) radiates almost nothing (merely because its radiation resistance is
very small), but is widely used as a radiating element because of its theoretical
importance.
The best radiation power of a wire antenna is achieved when its length is
half wavelength (l = λ/2). Significant radiation is also obtained for integer mul-
tiples of this length. These antennas are called resonating antennas. For a verti-
cally erected wire antenna, the horizontal radiation pattern is omnidirectional
but its vertical radiation pattern depends on l ∼ λ relation. Vertical patterns of a
vertically erected various-size wire antennas are shown in Figure 5.33.
These radiation characteristics are important in EMC engineering. EM
fields couple with cables and ground wires, induce currents, and create unde-
sired emissions. It is critical to predict the direction of these emissions in order
to choose a good protection approach. Very often, just relocating cables and
ground wires in a design may eliminate these undesired emissions.
Figure 5.33 Wire antenna radiation patterns.

5.9.3 Broadband EMC Antennas

Wire or loop antennas resonate at certain frequencies and radiate significantly.
When the frequency changes, the resonance disappears and radiation reduces
drastically. Several different-length wire antennas can be gathered together and
broadband antennas can be designed. Antennas shown in Figure 5.34 (log-pe-
riodic, conical, bi-log, biconical) are broadband antennas operate from several
MHz up to GHz frequencies. Design steps of such antennas can be found in
[4–15]. One of them, LPDA, is discussed in the next subsection.
5.9.4 Log-Periodic Dipole Antenna

A log-periodic dipole antenna (LPDA) is a broadband, multielement, unidirec-
tional antenna (see Figure 5.35). It has impedance and radiation characteristics
that are repetitive as a logarithmic function of the excitation frequency. The
individual components are dipoles. The length and spacing of the elements of
LPDA increase logarithmically from one end to the other. Activations of the
dipoles of the LPDA change from longest to shortest as the frequency changes
from lower end of the band to upper end. Design parameters are minimum and
maximum frequencies and the average gain [8]. A short MATLAB code listed
in Table 5.7 can be used to design a LPDA. A sample output of this code, for a
five-element, G = 6 dB LPDA for the band 5 MHz–15 MHz, is given in Figure
5.36.
Figure 5.34 Broadband EMC antennas.

Figure 5.35 The LPDA structure.
Table 5.7
Short MATLAB Code of an LPDA Design
%------------------------------------------------
% Program: LPDA.m
%------------------------------------------------
clc; clear all;
sig = [.139.149.158.160.163.168.172.174.176.177.180];
ta = [.782.822.843.868.882.900.920.952.948.956.949];
% Supply design parameters
fmin = input(‘Lower Freq [MHz]: ‘); fmax = input(‘Upper Freq [MHz]: ‘);
G = input(‘Enter Gain [6-11 dB]: ‘);
while G<6 || G>11; G = input(‘Enter GAIN [6-11 dB]: ‘); end
c = 300; sigma = sig(2*G-11); tau = ta(2*G-11);
alpha = 2 * atan((1-tau) / (4 * sigma)) * 180/pi;
lambda_lower = c / fmin; lambda_upper = c / fmax;
N = 1; L(N) = 0.5 * lambda_lower; Lu = 0.5 * lambda_upper;
fprintf(1, ‘\n Alpha degree is: %2.1f\n’, alpha);
fprintf(1, ‘ Tau for the optimum design: %1.3f\n’, tau);
fprintf(1, ‘ Sigma for the optimum design: %1.3f\n\n’, sigma);
while L(N) > Lu; L(N + 1) = tau * L(N); N = N + 1; end; fprintf(1, ‘ \n’)
for n = 1 : N-1
d(n) = 2 * sigma * L(n);
fprintf(1, ‘Dipole # %2.0f-- L = %3.3f-- d = %3.3f \n’, n, L(n), d(n)); end % Program END
Figure 5.37 shows vertical and horizontal radiation patterns of a 13-ele-

ment LPDA designed for the frequency band of 1.1 GHz (i.e., between 400
MHz and 1.5 GHz) with the desired gain of G=8 dB.
Figure 5.36 A sample LPDA designed with the code in Table 5.7.
Figure 5.37 (a) Vertical, and (b) horizontal radiation patterns of the LPDA at 800 MHz.
5.9.5 Horn Antenna

A horn antenna or microwave horn is an antenna that consists of a flaring
metal waveguide shaped like a horn to direct radio waves in a beam (see Figure
Figure 5.38 A typical horn antenna and its 3D radiation pattern.
5.38). Horns are widely used in EMC tests and measurements above 250–300
MHz frequencies. They can operate over a wide range of frequencies (e.g., from
1 GHz to 20 GHz) because they have no resonant elements. Their gains range
up to 25 dBi with 10–20 dBi being typical. Horns have very little loss; their
directivity are roughly equal to their gain and therefore are also used in antenna
calibration.
References
[1] Sevgi, L., “The Antenna as a Transducer: Simple Circuit and Electromagnetic Models,”
IEEE Antennas and Propagation Magazine, Vol. 49, 6, Dec. 2007, pp. 211–218.
[2] Sevgi, L., “Reciprocity: Some Remarks from Field Point of View,” IEEE Antennas and
Propagation Magazine, Vol. 52, 2, Apr. 2010, pp. 205–210.
[3] IEEE Standard Definitions of Terms for Antennas, IEEE Std 145–1983.
[4] King, R. P. W., The Theory of Linear Antennas, Cambridge, MA: Harvard University Press,
1956.
[5] Jordan, E., and K. Balmain, Electromagnetic Waves and Radiating Systems, 2nd Ed., Engle-
wood Cliffs, NJ: Prentice Hall, 1968.
[6] Elliot, R. S., Antenna Theory and Design, Englewood Cliffs, NJ: Prentice Hall, 1981.
[7] Colin, R. E., Antennas and Radiowave Propagation, New York: McGraw Hill, 1985.
[8] Milligan, Thomas A., Modern Antenna Design, New York: McGraw Hill, 1985.
[9] Kraus, J. D., Antennas, Second Ed., New York: McGraw Hill, 1988.
[10] Brookner, E., Practical Phased Array Systems, Norwood, MA: Artech House, 1991.
[11] Manser, G., Antenna Design: A Practical Guide, New York: McGraw Hill, 1996.
[12] Balanis, C. A., Antenna Theory, 2nd Ed., New York: John Wiley and Sons, 1997.
[13] Hansen, R. C., Phased Array Antennas, New York: John Wiley and Sons, 1997.
[14] Shutzman, W. L., and G. A. Thiele, Antenna Theory and Design, 2nd Edition, New York:
John Wiley and Sons, 1998.
[15] Macnamara, Thereza M., Handbook of EMC Antennas, Norwood, MA: Artech House,
1995.
[16] ANSI C63.5-1998, American National Standard for Electromagnetic Compatibility-
Radiated Emission Measurements in Electromagnetic Interference (EMI) Control
Calibration of Antennas (9kHz–40 GHz).
[18] Smith, A. A., R. F. German, and J. B. Pate, “Calculation of Site Attenuation from Antenna
Factors,” IEEE Trans on EMC, Vol. 24, No. 3, Aug. 1982, pp. 301–316.
[19] Trakadas, P. T., and C. N. Capsalis, “A Mixed Model for the Determination of Normalized
Site Attenuation in OATS,” IEEE Trans on EMC, Vol. 43, No. 1, Feb. 2001, pp. 29–36.
[20] Gavenda, J. D., “Near Field Corrections to Site Attenuation,” IEEE Trans on EMC, Vol.
36, No. 3, Aug. 1994, pp. 213–220.
[21] Askri, A., C. Vollaire, L. Nicolas, and D. Prebet, “Normalized Site Attenuation Standard
Correction from Numerical Computing,” IEEE Trans on EMC, Vol. 38, No. 2, Mar.
2002, pp. 693–696.
[22] Agilent Technologies, http://www.agilent.com.
[23] TDK RF Solutions, www.tdkrfsolutions.com.
[24] CISPR16-1-5:2003 Antenna calibration test sites (CALTS) for 30 MHz to 1000 MHz.
[25] CISPR 16-1-4, “Specification for Radio Interference Measuring Apparatus and
Measurement Methods,” IEC 2004.
[26] CISPR16-4-2:2003 Measuring Apparatus and Methods—Uncertainties, Statistics, and
Limit Modeling—Uncertainty in EMC Measurements.
[27] ANSI/IEEE Standard C63.4-1992, American National Standard for Methods of
Measurement of Radio-Noise Emissions from Low-Voltage Electrical and Electronic
Equipment in the Range of 9 kHz to 40 GHz.
[28] ANSI/IEEE C63.5-2006 (Revision of ANSI C63.5-2003), American National Standard
for Electromagnetic Compatibility–Radiated Emission Measurements in Electromagnetic
Interference (EMI) Control–Calibration of Antennas (9 kHz to 40 GHz).
[29] ANSI/IEEE Std 149-1979 (Revision of IEEE Std 149-1965) IEEE Standard Test
Procedures for Antennas.
[30] IEEE Standard 145-1983, IEEE Standard Definitions of Terms for Antennas.
[31] ISO/IEC/OIML/BIPM, Guide to the Expression of Uncertainty in Measurement, Prepared
by ISO/TAG 4/WG 3, January 1993.
[33] IEEE Std 291-1991, IEEE Standard Methods for Measuring Electromagnetic Field
Strength of Sinusoidal Continuous Waves, 30 Hz to 30 GHz.
[34] SAE-ARP 958 Rev. D, Electromagnetic Interference Measurement Antennas; Standard
Calibration Method.
[35] MIL-STD-461E-1999, Interference Standards Requirements for the Control of
Electromagnetic Interference Characteristics of Subsystems and Equipment, US
Department of Defense.
[36] MIL-STD-461F-2007, Interference Standards Requirements for the Control of
Electromagnetic Interference Characteristics of Subsystems and Equipment, US
Department of Defense.
[37] CENELEC EN50147-1: 1996, Anechoic Chambers Part 1: Shield Attenuation
Measurement.
[38] Çakır, S., R. Hamid, and L. Sevgi, “Loop Antenna Calibration,” IEEE Antennas and
Propagation Magazine, Vol. 53, No. 4, Oct. 2011, pp. 243–254.
[39] Sevgi, L., and Ç. Uluışık, “A MATLAB-Based Visualization Package for Planar Arrays of
Isotropic Radiators,” IEEE Antennas and Propagation Magazine, Vol. 47, No. 1, Feb. 2005,
pp. 156–163.
Bibliography
Paul, C. R., Introduction to Electromagnetic Compatibility, New York: Wiley Interscience, 1992.
McLeon, J., R. Sutton, and R. Hoffman, “Interpreting Antenna Performance Parameters for
EMC Applications,” Application Notes, TDK RF Solutions Inc., www.tdkrfsolutions.com.
IEEE Std 1309-2005, IEEE Standard for Calibration of Electromagnetic Field Sensors and
Probes, Excluding Antenna, from 9 kHz to 40 GHz.
6
Noise and Frequency Analysis
Suppose you are a 100m sprinter. You run on your own track; therefore, you
do not to worry about being obscured by the next runners. If you are a 3000m
runner, then the situation is different: after a short start on your track, like the
other runners, you prefer to continue on the innermost track. In that case, a
runner can obscure others. EMC problems are similar to this example. Two
or three decades ago, there was no EMI problem between a long-wave radio
transmitter and defense radar because their locations and frequencies were far
enough apart. Today, the electrical and electronics systems we use in our daily
lives (e.g., a laptop, a cell phone, hundreds of systems in automobiles, and fast-
trains) are all EMI-vulnerable because they are digital, broadband, and have to
operate close to each other.
Two parameters are important for all equipment operating in close prox-
imity: signal strength and frequency. In order not to interfere with each other
here are two simple EMC engineering tips:
• If frequencies of two pieces of equipment are very close, then the dis-
tance in between them should be far enough apart or some sort of isola-
tion should be applied.
• If two pieces of equipment need to operate very close to each other, they
must use different frequencies.
6.1 Fundamental Electromagnetic Signals

Time-frequency characteristics of electrical, electronics, and EM signals deter-
mine the way EMC problems should be handled from design to market. The
155
frequency characteristics of a signal are modeled using the Fourier transform

(FT) [1–5]. FT has been widely used in circuit analysis and synthesis, from
filter design to signal processing, image reconstruction, stochastic modeling,
and nondestructive measurements. FT has also been widely used in EM from
antenna theory to radiowave propagation modeling, radar cross-section predic-
tion., and multisensor system design. FT is used in all types of signals such as
useful signal, noise, clutter, and interference.
Signal is the useful component under investigation—narrow- or broad-
band, continuous or transient, low or high power, deterministic or stochastic,
and so on. Usually, it is a correlated signal (i.e., occupies a finite frequency band).
Noise is a floor signal that limits the sensitivity (i.e., the smallest signal that can
be detected or measured) of the EMI receiver. Noise is present in all electronic
circuits, although it is often so small compared to useful signals. It may be inter-
nal or external. Noise is uncorrelated (i.e., present in all frequencies). Clutter is a
correlated signal that is not under investigation. It may be echoes from ground,
sea, rain, and so on, from other transmitters, communication means, and so on.
Finally, interference is a disturbing signal that should be mitigated [3–5].
Note that all of these signals, except noise, are correlated, which means
they occupy certain frequency bands. This is important in EMC engineering
because their mitigation approaches are different: Noise can be eliminated by
oversampling and time-averaging; others can only be filtered out.
6.2 Noise
Noise [5] is an uncorrelated signal that limits signal detection. Figure 6.1 pres-
ents typical noise spectra with manmade or natural components (sun burst,
magnetic storms, cosmic noise, atmospheric effects, etc.).
In terms of EMC engineering, noise can be grouped into two types: envi-
ronmental noise and thermal noise (also called internal noise, electronic noise,
Gaussian noise, and white noise). The thick line in the figure belongs to ther-
mal noise. As observed, above a few hundred MHz frequencies thermal noise
is dominant. Roughly speaking, EMC engineers do not need to worry about
environmental noise if frequencies are above 300 MHz–500 MHz; the only
noise they have to pay attention to is thermal noise. On the contrary, for the
frequencies below 200 MHz–300 MHz, thermal noise is negligible; EMC en-
gineers have to take care of the environmental noise. At high frequencies (above
500 MHz) thermal noise is critical in EMC engineering.
Thermal noise exists in all circuits and devices caused by random varia-
tions in current or voltage by random movement of charges. Thermal noise can
be reduced by reducing the temperature. This phenomenon limits the mini-
mum signal level that any receiver can usefully respond to, because there will
Noise and Frequency Analysis 157
Figure 6.1 Typical noise sources and noise spectra.
always be a small but significant amount of thermal noise arising in its input
circuits. There are several other sources of noise in electronic circuits such as
shot noise, seen in very low-level signals where the finite number of energy-
carrying particles becomes significant, or flicker noise (1/f noise) in semicon-
ductor devices.
The sensitivity (minimum detectable signal) of an EMI receiver is deter-
mined by the noise level. Thermal noise is uncorrelated; this means it appears at
all frequencies. It is a random signal and therefore necessitates stochastic mod-
eling. Thermal noise is Gaussian with zero average value. It fluctuates around
useful signal. Noise power can roughly be predicted from
Pn = kTB [ W ] (6.1)
where T is the temperature in degrees Kelvin (°K), k = 1.6 × 10–23 [J/°K] is

the Boltzman constant, and B [Hz] is the bandwidth. As observed, thermal
noise depends on the signal bandwidth, not a carrier frequency. For example,
the thermal noise level of a signal with 1-MHz bandwidth modulated with a
10-GHz carrier is Pn =10 pW (Pn = –110 dBm). For a 50Ω system receiver,
this noise power corresponds to (Pn = Vn2 50) Vn = –22 µV (13 dBµV) noise
voltage.
Signal-to-noise ratio (SNR) is an important parameter in signal theory. It
is given in terms of signal and noise powers (SNR = Ps /Pn). The lowest signal
level Ps = Pn corresponds to the sensitivity of the receiver. For a simple sinusoidal

signal s(t) = Vm sin(2πft), the SNR is:
Vm2  V2 
SNR = or SNRdB = 10Log 10  m  (6.2)
2kTB  2kTB 
Any of these three (SNR, signal voltage, and noise voltage) can be cal-
culated if the other two are known or given. The noise voltage, for example, is
obtained from
Vm
Vn = (6.3)
2SNR
Thermal noise, which is Gaussian and uncorrelated, can be generated syn-

thetically as shown in Figure 6.2. Here noise alone, signal alone, and signal +
noise samples generated with simple MATLAB commands is presented.
The average and standard variation of the Gaussian distribution corre-
spond to DC (µn = 0) and AC (σn = Vn) noise voltages, respectively. There are
1000 samples in the figure. First, 1000 random numbers (x1000) are generated
using x=randn(1,1000) command. This command generates Gaussian random
numbers with zero average and unit variance. Then, random noise samples hav-
ing noise voltage Vn are obtained via n = 2Vnx –Vn. Signal samples are directly
generated from s(t) = Vm sin(2πft). For example, sample signals in a 25 period
Figure 6.2 SNR versus time of a sinusoidal signal with Gaussian noise (SNR=10 dB, Vm = 3V,
f = 100 Hz, T = 0.25s).
can be generated with ∆t = 0.1 ms sampling for a duration of T = 0.25s if the

signal frequency is f=100 Hz. The addition of these two yields the total signal
(y(t) = s(t) + n(t)). In the figure, the SNR is 10 dB. Even for this value, it is hard
to distinguish the signal from the noise. In practical receivers, for reliable signal
detection, required SNR values may be in the range of 13–50 dB, depending
on the application.
Noise reduction or mitigation is a major problem in EMC engineering.
EMC engineers either reduce the bandwidth (i.e., use matched filter) or apply
simple averaging. The effects of averaging is shown in Figure 6.3 for SNR=0
dB.
The dashed curve belongs to noisy signal with SNR = 0 dB. The thick
solid line belongs to the signal without noise. The thick line belongs to noisy
signal after 10 spectra averaging (i.e., 10 samples are recorded and averaged
for every single data in the figure). Comparing these two figures, one can see
how noise is reduced with this method (nearly 10 dB signal processing gain is
obtained in this case).
6.3 Frequency Analysis and Fourier Transforms

Analysis of a real-world signal is a fundamental problem for scientists, especially
for electrical engineers since almost every real-world signal is transformed into
an electrical signal by means of transducers (e.g., accelerometers in mechanical
engineering, EEG electrodes and blood pressure probes in biomedical engineer-
Figure 6.3 Noise reduction using spectra averaging.

ing, seismic transducers in earth sciences, antennas in EM, and microphones in

communication engineering).
The traditional way of observing and analyzing signals is to view them in
time domain. Baron Jean Baptiste Fourier [1] more than a century ago showed
that any waveform that exists in the real world can be represented (i.e., gener-
ated) by adding up sinusoidal waves. Since then, we have been able to build
(break down) our real-world time signal in terms of (into) these sinusoidal
waves. It is shown that the combination of sinusoidal waves is unique; any real-
world signal can be represented by only one sinusoidal combination [2].
The reader should keep in mind that the time domain–frequency domain
relations in EM are similar to the relations between spatial and wave number
domains. A simplest propagating (e.g., along z) plane wave is in the form of
exp{–j(kz – ωt)} (where k and z are the wavenumber and position, respectively),
and whatever characteristics exp(jωt) have are also applicable to exp(–jkz). A
few important observations are:
• A rectangular time (frequency) window corresponds to a beam type

(Sinc(.) function) variation in frequency (time) domain.
• Similarly, a rectangular aperture (array) in spatial domain corresponds to
a beam type (Sinc(.) function) variation in wavenumber domain.
• The wider the antenna aperture, the narrower the antenna beam, or the
narrower the pulse in time domain, the wider the frequency band.
• An impulse in the time (frequency) domain corresponds to a flat re-
sponse in the frequency (time) domains.
• TIME × BANDWIDTH product is constant.
Therefore, FT has also been used in EM from antenna analysis to imaging

and nondestructive measurements, even in propagation problems. For example,
the split-step parabolic equation method (which is simply the beam propaga-
tion method in optics) has been in use for nearly a century and is based on
sequential FT operations between the spatial and wavenumber domains. Two-
and three-dimensional propagation problems with nonflat realistic terrain pro-
files and inhomogeneous atmospheric variations above have been solved with
this method successfully [6–8].
Mathematically, FT is defined as
1 ∞ − j ωt
S ( ω) = ∫ s (t )e dt
2 π −∞
(6.4)
1 ∞
s (t ) = ∫
2 π −∞
S ( ω)e j ωt dt (6.5)
Here, s(t) is the time signal, S(ω) is its FT, and ω=2πf is the angular fre-
quency. FT is defined for continuous time (analog) signals. To obtain FT of a
time function at every frequency with infinitesimal frequency steps, the time
signal must be observed forever.
A few functions and their FT are given in Figure 6.4. FT of a triangle
function is square of a Sinc() function. FT of a Gauss pulse is also a Gauss pulse.
As the pulse gets narrower in the time domain, its FT gets broader. FT of an
impulse train is also an impulse train. As the time interval between impulses
gets narrower, the frequency intervals gets broader. For a single impulse, the FT
is constant. In communication society, it is called impulse response.
Knowing impulse response of a circuit means knowing all frequency
responses.
Fourier transform is used for energy signals (i.e., for signals which are non-
∞
2
zero in a certain period of time and goes to zero when t → ±∞ ( ∫ s (t ) dt is
−∞
finite)). Fourier series (FS) representation is used for periodic (power) signals. A
v(t) function in a finite T1 ≤ t ≤ T2can be represented with FS in terms of sine
and cosine functions:
A0 ∞  2 πnt   2 πnt 
v (t ) ≈ + ∑ An cos   + Bn sin 
 T2 − T1 
(6.6)
2 n =1  T2 − T1 
Figure 6.4 Time and frequency variations of some signals.

where Fourier coefficients An and Bn are derived from

T
2 2
T2 − T1 T∫
A0 = v (t )dt (6.7a)
1
T2
2  2 πnt 
An =
T2 − T1 T∫ v (t )cos  dt
 T2 − T1  (6.7b)
1
T2
2  2 πnt 
Bn =
T2 − T1 T∫ v (t )sin  dt
 T2 − T1  (6.7c)
1
The FS of a function represents the function only in T1 ≤ t ≤ T2 interval.

The FS representation is periodic outside this interval, but the function may
totally be different.
Table 6.1 lists a short MATLAB code that plots a trapezoidal function and
its FS representation for a given sets of parameters.
When this code is run, first the function is plotted, then the first set of
FS coefficients (Ao, A1 and B1) is calculated, and FS representation is plotted.
Each time the user presses a key, the next set of coefficients (Ai and Bi) are cal-
culated and added to the FS representation. Figure 6.5 presents an output of
this code for a symmetric trapezoidal pulse with 0.5s rise and fall time and 2s
pulse duration.
The plots are given on the left; the calculated FS coefficients are shown
on the right. On the top, the pulse and its FS representation with the first set
of coefficients are shown. The mid-plot belongs to the results with the first
10 FS coefficients. At the bottom, results with the first 18 terms are given. As
observed, less than 20 terms are enough to represent this pulse if rise/fall time
is 25% of the pulse duration. Note that the trapezoidal pulse in this example is
zero at the origin and 4s shifted in the time. It is not time-symmetric; therefore,
most of the FS coefficients are nonzero.
The code in Table 6.1 is rerun for a trapezoidal pulse symmetric with
respect to the time origin, and results are given in Figure 6.6. As observed, the
number of FS terms is almost the same, but in the symmetric case all Bn become
zero. For an even-symmetric function all the coefficients of (6.7c) would be
zero (because sine function is odd-symmetric). Similarly, for an odd-symmetric
function all the coefficients of (6.7b) would be zero (because cosine function
is even-symmetric). The last example given in Figure 6.7 belongs to the same
Table 6.1
MATLAB Code for Fourier Series Summation of a Trapezoidal Pulse
�%-----------------------------------------------------------------------------------
% PROGRAM: Fourier Series representation of a trapezoidal function
%-----------------------------------------------------------------------------------
clear all; clc;
N=400; xf=0; xl=10; P=abs(xl-xf); delx=P/N; x0=2.5; xT=.4;
tr=0.5; tau=2; % Pulse rise time and duration
for j=1:N
xx(j)=xf+(j-1)*delx;
% Trapezoidal pulse
if xx(j)<=2*tau ; fnr(j)=0; elseif xx(j)<=2*tau+tr; fnr(j)=(xx(j)-2*tau)/tr;
elseif xx(j)<=3*tau+tr ; fnr(j)=1; elseif xx(j)<=3*tau+2*tr
fnr(j)=3-(xx(j)-3*tau+tr)/tr; else; fnr(j)=0; end; end
A0=trapz(xx,fnr)*2./P; % Calculating A0
for j=1:N; fn(j)=A0/2; end
OK=1; k=0; Nt=0;
fprintf(1, ‘--CoefficientNo---------An----------Bn---\n’);
fprintf(1, ‘------------------------------------------------------\n ‘);
while OK>0 % Fourier series of the function
Nt=Nt+1;
for k=1:N % Calculate A(k) and B(k)
g1(k)=fnr(k)*cos(2*pi*Nt*xx(k)/P); g2(k)=fnr(k)*sin(2*pi*Nt*xx(k)/P);
end
An=trapz(xx,g1)*2./P; Bn=trapz(xx,g2)*2./P; % Use trapz integral function
fprintf(1, ‘------------------------------------------------------\n ‘);
fprintf(1, ‘ %3d %12.8f %12.8f \n’, Nt, An, Bn);
fprintf(1, ‘------------------------------------------------------\n ‘);
for j=1:N % Calculate the function
fn(j)=fn(j)+An*cos(2*Nt*pi*xx(j)/P)+Bn*sin(2*Nt*pi*xx(j)/P);
end
plot(xx,fnr,’k’,xx,fn, ‘r--’,’LineWidth’,2); grid; xlabel(‘Time [s])’); ylabel(‘f(t)’);
title(ttle); legend(‘Function’,’FSeries’); pause; end % Program END
even-symmetric trapezoidal pulse with sharper rise and fall times. As observed,
the FS representation is still insufficient even after the first 80 terms.
6.4 Discrete Fourier Transform

The FT is a mathematical operation that decomposes a signal into its constitu-
ent frequencies. The original signal depends on time and therefore is called the
time domain representation of the signal, whereas the FT depends on frequency
and is called the frequency domain representation of the signal. The term FT
refers both to the frequency domain representation of the signal and the process
that transforms the signal to its frequency domain representation. In effect, the
FT decomposes a function into oscillatory functions. The FT and its general-
izations are subject of Fourier analysis; both the time and frequency domains
are unbounded linear continua.
Figure 6.5 Trapezoidal pulse and its FS representation: (a) 1 term; (b) 10 terms; (c) 18 terms.
Figure 6.6 Symmetric trapezoidal pulse and its Fourier series representation.
Figure 6.7 Rectangular pulse and its Fourier series representation.
To compute the FT numerically on a computer, discretization plus nu-

merical integration are required. This is an approximation of the true (i.e.,
mathematical), analytically defined FT in a synthetic (digital) environment,
and is called the discrete Fourier transform (DFT). There are three difficulties
with the numerical computation of the FT:
• Discretization (introduces periodicity in both the time and the frequen-

cy domains);
• Numerical integration (introduces numerical error, approximation);

• Finite time duration (introduces maximum frequency and resolution
limitations).
Kept in mind when performing DFT that
• Multiplication in the time domain corresponds to a convolution in the

frequency domain.
• The FT of an impulse train in the time domain is also an impulse train
in the frequency domain with the frequency samples separated by T0 =
1/f 0.
• The narrower the distance between impulses (T0) in the time domain,
the wider the distance between impulses (f 0) in the frequency domain
(and vice versa).
• The sampling rate must be greater than twice the highest frequency of
the time record (i.e.,∆t ≥ 1/(2f max) Nyquist sampling criterion).
• Since time-bandwidth product is constant, narrow transients in the time
domain possess wide bandwidths in the frequency domain.
• In the limit, the frequency spectrum of an impulse is constant and cov-
ers the whole frequency domain (that’s why an impulse response of a
system is enough to find out the response of any arbitrary input).
Performing FT in a discrete environment introduces artificial effects.

These are called aliasing effects, spectral leakage, and scalloping loss.
Aliasing occurs when the time sampling rate is lower than the Nyquist
rate. Two signals are said to alias if the difference of their frequencies falls in the
frequency range of interest, which is always generated in the process of sam-
pling (aliasing is not always bad; it is called mixing or heterodyning in analog
electronics and is commonly used in tuning radios and TV channels). It should
be noted that although obeying Nyquist sampling criterion is sufficient to avoid
aliasing, it does not give high-quality display in time domain record. Figure 6.8
shows effects of aliasing. The signal spectrum is shown on top. Overlapping of
duplicates of signal bands is shown in the middle. Sufficient sampling at the
bottom separates duplicates and all (but the fundamental) bands can be filtered
out.
If a sinusoid existing in the time signal is not bin-centered (i.e., if its fre-
quency is not equal to any of the frequency samples) in the frequency domain,
spectral leakage occurs. In addition, there is a reduction in coherent gain if the
Figure 6.8 Effects of sufficient and insufficient sampling.
frequency of the sinusoid differs in value from the frequency samples, which is
termed scalloping loss.
The DFT of a continuous time signal sampled over the period of T, with
a sampling rate of t can be given as
N −1
T
S (m ∆f ) =
N
∑ s (n ∆t )e − j 2 πm ∆f n ∆t (6.8)
n =0
where ∆f =1/T is valid at frequencies up to f max = 1/(2∆t). The maximum fre-

quency depends on the sampling interval, and the frequency resolution is deter-
mined by the signal record length. That is, N samples of a time signal recorded
during a finite duration of T with a sampling period of ∆t (N = T/∆t) can be
transformed into N samples in the frequency domain between –f max and +f max
(also between +f max and +3f max, –3f max and –f max, and so on) according to
1 1
f max = , ∆f = (6.9)
2 ∆t T
Since sampling interval and signal record lengths are finite in numeri-
cal computations maximum frequency and the resolution are also finite. This
means:
• Any frequency component fc beyond +fmax cannot be observed in its

actual frequency; instead, it enters from left because of rotational sym-
metry and periodicity and appears at –fmax + fD where fD = fc – f max.
• Similarly, any frequency component –fc beyond –fmax cannot be ob-
served in its actual frequency; instead, it enters from left because of ro-
tational symmetry and periodicity and appears at fmax –fD where fD =
|fc – fmax|.
Tables 6.2 and 6.3 list short MATLAB codes for the computations of
DFT and FFT of given functions. Note that the code in Table 6.2 uses (6.4) in
discrete form, so any number of frequency samples () can be used to plot a fre-
quency spectrum. Unfortunately, (6.8) still holds for the DFT. In other words,
for example, if one wants to discriminate two sinusoids with 50 Hz and 52 Hz
in the frequency domain, the frequency resolution must be less than 2 Hz. This
necessitates minimum T = 0.5s. On the other hand, the DFT of a signal can be
taken using any ∆f, but the capability of discriminating two nearby sinusoids in
the frequency domain is limited by the length of the time record. The FFT code
in Table 6.3 uses built-in MATLAB commands fft, ifft, and fftshift.
A typical spectrum obtained with these codes (either DFT or FFT) is
presented in Figure 6.9. Here, two sine functions with 35 Hz and 50 Hz (s(t) =
sin(100πt) + sin(70πt)) are used as the time function. The sampling parameters
are ∆t = 10 ms and T = 1 s. As observed, only 35-Hz components appear in
the spectrum. The other (±50 Hz) components cannot be observed since they
Table 6.2
Short MATLAB DFT Code
%----------------------------------------------------------
% Program: DFT.m
%----------------------------------------------------------
% Supply f1, f2, T, dt, df, fmax
a1=1; a2=0.5; N=T/dt; w1=2*pi*fr1; w2=2*pi*fr2; M=fmax/df;
% build the input time series
for k=1:N; t(k)=k*dt; st(k)=a1*sin(w1*dt*k)+a2*sin(w2*dt*k); end
for k=1:M % Apply the DFT with M points
Sf(k)=complex(0,0);
for n=1:N; Sf(k)=Sf(k)+st(n)*exp(-i*2*pi*n*dt*k*df); end
Sf(k)=Sf(k)*dt;
End
% Prepare the frequency samples
for k=1:M; f(k)=k*df; end
plot(t,st)); title(‘The Sum of two Sinusoids’); xlabel(‘Time [s]’); ylabel(‘Amplitude’)
figure(2); plot(f,abs(Sf)); title(‘The DFT of the Sum of two Sinusoids’)
xlabel(‘Frequency [Hz]’); ylabel(‘Amplitude’); % Program END
Table 6.3
Short MATLAB FFT Code
%-----------------------------------------------
% Program: FFT.m
%-----------------------------------------------
% Get the input parameters f1, f2, dt, T
N=floor(T/dt); df=1/T; fmax=1/(2*dt); a1=1.0; a2=0.5;
for n=1:N
t(n)=(n-1)*dt; f(n)=-fmax+(n-1)*df;
st(n)=a1*sin(2*pi*f1*t(n))+a2*sin(2*pi*f2*t(n));
end
Sf=fftshift(fft(st)*dt); % Apply FFT Swap the left and right halves
plot(t,st); % plot the output in the time domain
title(‘Sum of two Sinusoids’); xlabel(‘Time [s]’); ylabel(‘Amplitude’)
figure(2); plot(f,abs(Sf)); title(‘The DFT of the function’)
xlabel(‘Frequency [Hz]’); ylabel(‘Amplitude’); % Program END
exactly appear at ±fmax. One can run either of these codes and see that the cor-
rect spectrum will be obtained by just reducing the sampling rate, for example,
to ∆t = 5 ms (which means fmax = ±100 Hz).
Any function may be approximated by series summation of sine and co-
sine functions. The number of terms required in the FS representation depends
on the smoothness of the function and the specified accuracy. The degree of
smoothness of the function determines the number of terms in its FS represen-
tation. In addition, only sine or cosine terms contribute the function if it is odd
or even symmetric.
The frequency content of the rectangular pulse should be well analyzed in
order to understand digital communication concepts. A symmetric rectangular
pulse is defined as
 T T
1 if − ≤t ≤ +
 t   2 2
Re ct   =  (6.10)
T   T
0 if t>
 2
An infinite number of terms is required to fully represent this function

with the FS summation. The terms are called harmonics. It is interesting to vi-
sualize term-by-term contributions in the FS representation. This is illustrated
in Figure 6.10 for a 2s rectangular pulse between 4 ≤ t ≤ 6s. On the top, rectan-
gular pulse and 3-term FS representations are shown. At the bottom, the same
plot is given with 51-term FS representation. As observed, artificial oscillations
Figure 6.9 The spectrum of s(t) = sin(100πt) + sin(70πt) with ∆t = 10 ms and T =1 s.
occur at both edges of the pulse. These are called overshoot or ringing (i.e., Gibbs
effect in communication, after mathematical physicist Josiah Gibbs, who ex-
plained the phenomenon in 1899). They occur even with a very high number
of terms in the FS representation.
Full FS representation of a rectangular pulse requires an infinite number
of terms. In other words, a rectangular pulse has infinite harmonics.
Digital signals have an infinite number of harmonics and create severe
EMC problems. Because of this, do not upgrade your processor unless it is
essential.
As noted earlier, the FS representation represents the function only in
the given finite period; beyond that, the signal and its FS representation may
totally be different. Figure 6.11 illustrates this. Here, fmax = (x2sin(3x))/ex/2 is
chosen as an example. Its FS representation is given between 0 ≤ x ≤ 10. Both
the function and its 17-term FS representation are plotted for 0 ≤ x ≤ 2 0. As
observed, they match very well in the first period, but are totally different in
the second period.
A few signals widely used in electronic systems and number of significant
FS terms are shown in Table 6.4. The table shows why a triangular pulse is
preferred against square or rectangular pulse in order to mitigate EMC prob-
lems. The last column presents the number of FS terms required for 10% error
between the function and its FS representation.
Question: What is the maximum frequency of a triangular pulse with 50-
µs pulse repetition period assuming 10% and more harmonic contributions?
Figure 6.10 Two rectangular pulses and their FS representations.
Answer: Fundamental signal frequency is f = 1/T = 1/50 µs = 20 kHz.

The significant harmonics of this pulse for 10% error is the third harmonic;
therefore, maximum frequency will be 3 × 20 kHz = 60 kHz.
6.5 Harmonic Generation

A load in a circuit may be linear or nonlinear according to its current and wave-
form. The voltage of a linear load is directly proportional with its current. A
few examples for nonlinear loads are transformers, convertors, TVs, computers,
semi-conductors, inductive ballasts, photovoltaic systems, gas discharc lamps,
transportation systems, and so on.
Nonlinear elements draw nonsinusoidal currents even under pure applied
sinusoidal voltages. The current has undesired components with integer mul-
tiples of the frequency of the supplied voltage called harmonics. The first four
Figure 6.11 A function and its FS in two periods of time.
Table 6.4
Waveforms and Significant Fourier Components
Harmonics
Waveform First Second Third Fourth Fifth Sixth Total (10%)
Sine 1 0 0 0 0 0 1
Square 1.273 0.0 0.424 0.0 0.255 0.0 9
Triangle 0.811 0.0 0.090 0.0 0.032 0.0 3
Pulse (50%) 0.637 0.0 0.212 0.0 0.127 0.0 9
Pulse (25%) 0.450 0.318 0.150 0.0 0.090 0.105 14
Pulse (10%) 0.197 0.187 0.172 0.151 0.127 0.101 26
harmonics of a sinusoidal signal and total waveforms when they add up are
shown in Figure 6.12.
6.6 AC Mains and Power Quality

Power quality (PQ) deals with the interaction of electrical power and electrical
equipment. If the equipment operates as planned without being damaged or
stressed, it is said that the electrical power is of good quality. Otherwise, it is
presumed that the PQ is poor. In general, any deviation from normal of a DC
or AC voltage source may be thought as a PQ problem. Fast transients, voltage
impulses, high-frequency noise, waveshape faults, voltage swells and sags, and
total power loss are some of important PQ problems.
For users, power quality is defined at the source/feed point.
Figure 6.12 Pure sine wave and effects of harmonics.
PQ necessitates a good understanding of load-source interactions. The

IEEE Std 1100-1999 [9] provides a basis for this understanding. Electrical and
electronics devices are susceptible to PQ problems; in the meantime, they may
also be the source of PQ problems. Some of the important concerns are wave-
form distortions like harmonics, transients, voltage fluctuations (e.g., voltage
sags and swells), interruptions (e.g., outages and blinks). Important areas of
load-source interactions are electrical wiring and grounding of a facility, surge
(transient) protection systems, harmonics proliferation demands, and electrical
power reliability.
Power quality problems are considered at low frequencies (below 9 kHz).
6.6.1 Distortions in AC Mains

Power quality issues of AC mains are defined as follows [9]:
• Power quality: The concept of powering and grounding electronic

equipment in a manner that is suitable to the operation of that equip-
ment and compatible with the premise wiring system and other con-
nected equipment.
• Power disturbance: Any deviation from the nominal value (or from
some selected thresholds based on load tolerance) of the input AC power
characteristics.
• Power factor, total: The ratio of the total power input in [W], to the
total voltampere input.
• Transient: A subcycle disturbance in the AC waveform that is evidenced
by a sharp, brief discontinuity of the waveform. May be of either polar-
ity and may be additive to, or subtractive from, the nominal waveform.
• Interrupt: The complete loss of voltage for a time period.
• Outage: The complete loss of voltage for at least half a cycle. For exam-
ple, in 50-Hz AC mains, outage means loss of voltage for at least 10 ms.
• Voltage distortion: Any deviation from the nominal sine waveform of
the AC line voltage.
• Voltage impulse: From 50V to 5 kV, voltage pulses with durations be-
tween 0.5 ms and 20 ms.
• Notches: A switching (or other) disturbance of the normal power volt-
age waveform, lasting less than a half cycle, which is initially of oppo-
site polarity than the waveform and is thus subtractive from the normal
waveform in terms of the peak value of the disturbance voltage. This
includes complete loss of voltage for up to a half cycle.
• Harmonic distortion: The mathematical representation of the distortion
of the pure sine waveform.
• Distortion factor: The ratio of the root square value of the harmonic
content to the root square value of the fundamental quantity, expressed
as a percent of the fundamental.
• Sag: An rms reduction in the AC voltage, at the power frequency, for du-
rations from a half cycle to a few seconds. In other words, voltage drops
up to 80% of nominal value for a period longer than a full cycle (20 ms).
• Undervoltage: When used to describe a specific type of long duration
variation, refers to an rms decrease in the AC voltage, at the power fre-
quency, for a period of time greater than 1 min.
• Voltage regulation: The degree of control or stability of the rms voltage
at the load. Often specified in relation to other parameters, such as input
voltage changes, load changes, or temperature changes.
• Swell: An increase in rms voltage or current at the power frequency for
durations from a half cycle to 1.0 min. In other words, more than 110%
increase of the voltage or current for a period longer than a full cycle
(i.e., longer than 20 ms).
• Frequency change: The change in nominal frequency (i.e., frequency
fluctuations around 50 Hz or 60 Hz AC mains).
• Flicker: A variation of input voltage, either magnitude or frequency,
sufficient in duration to allow visual observation of a change in electric
light source intensity. In other words, it is defined as a periodic voltage
decrease in nearly a 6–7 full cycle (i.e., 8–9 Hz). The human eye is sensi-
tive at 8.2 Hz and suffers from signals around this frequency.
• DC voltage component: The inequality of positive and negative halves
of the AC voltage.
• Common-mode noise (longitudinal): The noise voltage that appears
equally, and in phase, from each current carrying conductor to ground.
• Differential-mode noise (transverse-mode noise): Noise signals mea-
surable between or among active circuit conductors feeding the subject
load, but not between the equipment grounding conductor or associated
signal reference structure and the active circuit conductors.
• Crest factor (of a periodic function): The ratio of the peak value of a
periodic function to the rms value.
• Ground: (a) A conducting connection, whether intentional or acciden-
tal, by which an electric circuit or equipment is connected to Earth or
to some conducting body of relatively large extent that serves in place of
the Earth; (b) high-frequency reference or approximately that potential
on conductors connected to it and for conducting ground currents to
and from Earth (or the conducting body).
IEC 61000-4-30:201 5 [10] defines the methods for measurement and

interpretation of results for PQ parameters in AC mains with a declared fun-
damental frequency of 50 Hz or 60 Hz. Measurement methods are described
for each relevant parameter in terms that give reliable and repeatable results,
regardless of the method’s implementation. The standard addresses measure-
ment methods for in situ measurements. Measurement of parameters is lim-
ited to conducted phenomena in power systems. The PQ parameters are power
frequency, magnitude of the supply voltage, flicker, supply voltage dips and
swells, voltage interruptions, transient voltages, supply voltage unbalance, volt-
age harmonics and interharmonics, mains signaling on the supply voltage, rapid
voltage changes, and current measurements.
Note that some of the words, such as blackout, brownout, clean ground,
clean power, conducting barriers, dedicated ground, dirty ground, dirty power,
frequency shift, glitch, power surge, shared ground, spike, and subcycle outages,
have different histories of usage, and some may have specific definitions for
other applications. Therefore, they are avoided in [9].
6.6.2 Harmonic Distortions and Standards

Mathematically, a nonsinosoidal voltage signal can be given as:
V (t ) = V 0 + V1 sin( ωt + Φ1 ) + V 2 sin( ωt + Φ 2 )
(6.11)
+ V 3 sin( ωt + Φ 3 ) + ...... + Vn sin( ωt + Φn )
where
V0(t): Instant value of the voltage

V0: DC voltage component
V1: Amplitude of the fundamental voltage component
V2: Amplitude of the second harmonics
Vn: Amplitude of the nth harmonics
Φ: Phase difference between voltage and current
The EU standard for the protection of the AC mains from the connected
equipment is EN61000-3-2 (1998) [11]. This standard gives the limits for har-
monic current emissions for equipment with input current less than 16A per
phase. According to this standard, voltage harmonic distortion [HD]V and total
harmonic distortion [THD]V are defined as:
Vk
∑ Vk2
(6.12)
[HD ]V = , [THD ]V = k =2
V1 V1
Similarly, harmonic current distortion [HD]I and total harmonic current

distortion [THD]I, defined according to SC case, are given as:
I
∑ I k2
(6.13)
[HD ]I = k , [THD ]I = k =2
I1 I1
6.6.3 Harmonics Measurements and Evaluation

Nonlinear equipment, such as electronic power supplies, refrigerators, and au-
to-program washing machines and dishwashers, do not draw pure sinusoidal
currents from AC mains. They create undesired harmonics in the power supply
systems. Therefore, the standard EN61000-3-2 [11] became mandatory after
July 1, 1998.
The limiting values in this standard are given for four different classes:
class B for portable equipment, class C for lightning protection devices, class
D for 600W and lower power PC and TV receivers, and class A for everything
else powered from AC mains. The harmonic current limits in EN61000-3-2 are
listed in Table 6.5.
The standard EN61000-3-2 covers everything powered by 50–60 Hz
220/380V, 230/400V, and 240/415V power systems. This standard requests
measurements up to the 40th harmonics (i.e., up to 2 kHz and 2.4 kHz for 50-
Hz and 60-Hz systems, respectively). According to this standard, if a gradual
decrease is observed above the 19th harmonics, then upper harmonics may not
be measured. Moreover, harmonics less than 6% input current and/or less then
5 mA are assumed negligible.
The following remarks are important for harmonic measurements of an
equipment during its full operation:
• Class A, B, C, and D limits are given in the tables for steady state opera-
tions (see Table 6.5).
• Harmonics observed/recorded for a period less than 10s, no matter what
their amplitudes are, are not taken into account.
Table 6.5
Harmonic Current Limits in EN61000-3-2
Class A Class B Class C Class D
Harmonic No [A rms] [A rms] (%) [mA/W]
2 1.08 1.62 2 —
3 2.30 3.45 30 3.4
4 0.43 0.645 — —
5 1.14 1.71 10 1.9
6 0.30 0.45 — --
7 0.77 1.115 7 1.0
8 ≤ n ≤ 40 1.84 / n 2.76 / n — —
9 0.40 0.60 5 0.5
11 0.33 0.495 3 0.35
13 0.21 3.15 3 0.296
15 ≤ n ≤ 39 2.25 / n 3.375 / n 3 3.85 / n
• In a 150s (2.5 min) measurement time, the amplitudes of even-ordered

harmonics between 2–10 and odd-ordered harmonics between 3–19
may be 1.5 times more than the limit values for 15s (10% of the obser-
vation period).
Any device with frequency selectivity can be used in harmonic measure-

ments. A measurement error may not exceed 5 % of the given limit value or
more than 2% of EUT’s current. The impedance of the measuring device shall
not cause a voltage drop of peak-to-peak 0.15V. Harmonic selectivities of fn for
both f1 = 50 and f1 = 60 Hz systems are 30 dB for 2 f1 < fn < 12 f1, 20 dB for 12
f1 < fn < 20 f1, and 15 dB for 20 f1 < fn < 40 f1.
The standard states different conditions for all TV receivers, video re-
corders, lamps, vacuum cleaners, washing machines, microwave ovens, and
ITE. The test voltage for both single- and tri-phase systems should be between
230–400V. Maximum allowable values for the test voltage and its frequency are
±2.0% V and ±0.5% Hz, respectively. Required test voltages for the terminals
of the EUT are as follows:
• 3rd harmonics: 0.9%;

• 5th harmonics: 0.4%;
• Even harmonics between 2–10: 0.2%;
• Harmonics between 11–40: 0.1%.
The best way for a reliable harmonic analysis of an installation or EUT is

to perform harmonic measurements. Harmonic analyzers may be used for this
purpose. Although comments and evaluations differ from device to device, the
following observations may be made for an installation:
• The installation does not disturb AC mains if [THD]V < 2.5% and
[THD]I < 10%.
• The installation may disturb AC mains if [THD]V ≅ 2.5%–3 % and
[THD]I > 10%. It might be necessary to use harmonic filters.
• The installation does disturb AC mains if [THD]V > 3%. The use of
harmonic filters would be essential technically and financially.
6.6.4 Effects of Harmonic Currents and Voltages

EMC problems related to harmonics can be grouped into two categories:
• Current harmonics:
• Increase in power losses, decrease in usable power, low power factor;
• Excess currents in triphase systems;
• Excess heating in transformers and generators;
• Increase in acoustics noise;
• Interference with telephone lines.
• Voltage harmonics:
• Excess heating in transformers and generators;
• Excess heating in capacitors and motors;
• Stress decrease in isolation systems;
• Resonance occurrence or high-voltage breakdown;
• Mechanical oscillations or induction problems in motors.
The best way in harmonic protection is to use active and/or passive filters.
Filtration is a very effective EMC protection approach, not only for harmonics
but also for several other EMC problems.
Signal behaviors determine EMC/EMI problems and protection ap-
proaches. The time-domain EM transients are modeled in terms of Laplace
transforms. Steady-state EM signals are well understood by means of Fourier
transforms. A Fourier transform is a mathematical definition. Electronic sys-
tems are digital; therefore, DFT is used for spectrum analyzes. FT and DFT are
totally different operations; one needs to be very careful in dealing with discrete
signals. DFT introduces periodicity in both the time and frequency domains.
Aliasing and spectrum leakage are two artificial effects of the DFT.
References
[1] Baron Jean Baptiste Fourier, http://bartleby.com/65/fo/Fouriers.html.
[2] Hewlett Packard, The Fundamentals of Signal Analysis, Application Note 243, 1994.
[3] Sevgi, L., “Numerical Fourier Transforms DFT and FFT,” IEEE Antennas and Propagation
Magazine, Vol. 49, No. 3, Jun. 2007, pp. 238–243.
[4] Sevgi, L. “Synthetic Radar Signal Environment: Computer Generation of Signal, Noise,
and Clutter,” IEEE Antennas and Propagation Magazine, Vol. 49, No. 5, Oct. 2007, pp.
192–198.
[5] Sevgi, L., Electromagnetic Modeling and Simulation, New York: IEEE Press–John Wiley
and Sons, April 2014.
[6] Sevgi, L., F. Akleman, L. B. Felsen, “Ground Wave Propagation Modeling: Problem-
Matched Analytical Formulations and Direct Numerical Techniques,” IEEE Antennas and
Propagation Magazine, Vol. 44, No. 1, Feb. 2002, pp. 55–75.
York: IEEE Press–John Wiley and Sons, June 2003.
[8] M. Levy, Parabolic equation methods for electromagnetic wave propagation, IEE, Institution
of Electrical Engineers, 2000
[9] IEEE Std 1100-1999 -Recommended Practice for Powering and Grounding Electronic
Equipment.
[10] IEC 61000-4-30 Ed. 3.0 b(2015) Electromagnetic Compatibility (EMC)-Part 4-30:
Testing and Measurement Techniques -Power Quality Measurement Methods.
[11] EN 61000-3-2 (1998): Electromagnetic Compatibility (EMC)-Part 3: Limits-Section 2:
Limits for Harmonic Current Emissions (Equipment Input Current <=16 A Per Phase).
Bibliography
F. J. Harris, “On the Use of Windows for Harmonic Analysis with the Discrete Fourier Trans-
form,” Proc. IEEE, Vol. 66, No. 1, Jan 1978, pp. 51–83.
Sevgi, L., and Ç. Uluışık, “A Labview-Based Virtual Instrument for Engineering Education: A
Numerical Fourier Transform Tool,” ELEKTRIK, Turkish J. of Electrical Engineering and Comput-
er Sciences (Special issue on Electrical and Computer Engineering Education in the 21st Century:
Issues, Perspectives and Challenges), Vol. 14, No. 1, 2006, pp. 129–152.
7
EMC Test and Measurement
Environments
EMC tests and measurements are performed in two different environments:
inside and outside. The place where outside tests and measurements are per-
formed is called open area test site (OATS). Inside tests and measurements are
performed in screened or shielded rooms and anechoic chambers. Inside or out-
side, the main problem in tests and measurements is the control/elimination of
environmental interference: up to a certain extent, free-space conditions must
be satisfied. Screened rooms isolate inside and outside but do not eliminate
inside reflections. Moreover, tests and measurements may yield values 30–50
dB higher than EUT’s emission because of resonance effects. Anechoic cham-
bers both isolate inside/outside and eliminate internal reflections (emulate free-
space conditions). OATS must be located in an electromagnetically quiet zone.
Whether or not such tests and measurements areas satisfy these requirements is
controlled and traced via calibration.
7.1 Open Area Test Site

Both RE measurements and antenna calibrations are performed in OATS. An-
tenna calibration is based on AF measurements and is performed in CALTS.
They both suffer from environmental conditions; therefore, site calibration
is essential before conducting any measurement and/or test. Note that both
OATS and CALTS can also be used in immunity tests if calibrated together
with ground absorbers. The calibration procedures are presented in CISPR [1,
2], FCC [3], and ANSI/IEEE [4, 5] standards. The calibration of an OATS can
be done according to the procedure given in [1, 3, 4]. CALTS, on the other
181
hand, is also an OATS used for determining the free-space AF and can be cali-
brated/validated according to [2].
OATS calibration is performed through site measurements and compari-
sons against a reference site. The ideal or reference site is simply a perfectly con-
ducting, large plane over leveled ground in an electrically quiet area. In [1], the
reference or standard site is defined as “A site comprised of a flat, open-area,
devoid of nearby scatterers such as trees, power lines, and fences that has a large
metallic ground plane.” There are many papers on OATS calibration in the
literature; a few of them may be found in [1–3].
A typical picture of an OATS is sketched in Figure 7.1. The site is an
elliptical open, leveled area with the major and minor axes of 2R and R√3, re-
spectively. The distance between the antenna and EUT is R. The measurement
distance R may be 3m, 10m, or 30m.
The standard site calibration is based on the measurement of the IL be-
tween transmit and receive antennas, on a large, flat, and unobstructed con-
ducting ground plane. As specified in Clause 5.8.2 of CISPR 16-1-4 [1], site
performance may be validated by the NSA method. One antenna is set to be
at a fixed height (e.g., 1m, 2m), while the other one is scanned from 1m to 4m
in height. The maximum response between the two antennas is recorded. The
conducting ground plane is there to ensure the repeatability of the calibration
measurements.
The critical parameter in OATS calibration is NSA.
Figure 7.1 A typical open field test site (OFTS) sketch.

EMC Test and Measurement Environments 183
7.2 Open Area Test Site Calibration

Calibration of an OATS is given via an example here. A new OATS has been
constructed according to [9–11]. Photos of the OATS to be calibrated are given
in Figure 7.2. As shown there, there are nearby buildings, trees, and a hill with
a long 1m-high side wall. The sketch showing the location, nearby obstacles,
and their distances are shown in Figure 7.3. As shown in the sketch, the metal
ground is a 12m × 20m area. The nearest building is 10m away. The distance
between the metal ground and the control room is 9.5m. The nearby ground
supporting wall is 3m away.
The NSA of the OATS is measured and compared with the theoretical
NSA value for the calibration. If measured and calculated NSA values are with-
in ±4 dB uncertainty strip, then the OATS is said to be calibrated.
Two broadband antennas can be used in NSA measurements. The trans-
mit antenna has its reference point at the measurement positions of the test
Figure 7.2 Photos of the OFTA to be calibrated.

Figure 7.3 Diagram of the OFTS to be calibrated.
volume and the receive antenna is outside this test volume at a prescribed
orientation and position. The transmit antenna shall have an approximately
omnidirectional H-plane pattern. Typical receive antennas are hybrid anten-
nas (biconicaI/LPD combination) for 30 to 1000 MHz, or separate biconical
antennas (for 30 MHz to 200 MHz) and LPDA (for 200 MHz to 1000 MHz).
The equipment used during the NSA measurements is listed in Table 7.1. Due
to their large dimensions and phase center problems hybrid antennas are not
recommended for NSA measurements.
As described in [1] the NSA method is used for the calibration of OATS
having interantenna distance greater than 5m. Free-space AFs of the antennas
Table 7.1
Equipment Used During the Calibration
Equipment Firm Model Characteristics
Biconic antenna Schwarzbeck VHA 9103 30–300 MHz
Biconic antenna EMCO 3109 30–300 MHz
Log-periodic antenna Schwarzbeck VUSPL 9111 200–2000 MHz
Bi-log antenna Schaffner CBL 6141A 30–2000 MHz
Spectrum analyzer HP 8560E 30 Hz–2.9 GHz
Coaxial cable – RG213U 18m, 4m, 10m
Signal generator HP 8657B 0.1–2000 MHz
Signal generator Anritsu MG3633A 0.01–2700 MHz
used in the calibration are needed in this method. For each frequency site, cali-
bration is performed with the following steps:
• Adjust the output level of the signal generator to give a received voltage
display well above ambient and measured receiver or spectrum analyzer
noise.
• Raise the receiving antenna on the mast and scan 1–4m.
• Record the maximum signal level. This value is VSite in (5.45) in Chapter
5.
• Disconnect transmit and receive cables from their antennas. Directly
connect these cables with a straight-through adapter.
• Record the signal level when transmit and receive cables connected. This
value is VDirect in (5.45) of Chapter 5.
• At each frequency and for each polarization record VSite and VDirect .
• Insert transmit and receive AFs at the measurement frequency as shown
in (5.45) of Chapter 5.
• Insert the mutual impedance correction factor ∆AFtot , which applies
only for the specific geometry of horizontal polarization using tunable
dipoles separated by 3m. ∆AFtot = 0 for all other geometries.
• Solve (5.45) of Chapter 5 for AN, which is the NSA for the measure-
ment frequency and polarization used. Then, subtract this AN from the
NSA value.
• If the result is less then ±4 dB, the site is considered validated at that
frequency and polarization.
• Repeat these steps for the next frequency and polarization combination.
Measurements are performed as follows:
• The spectrum analyzer is located in the control room.

• The signal generator is placed at the furthest point behind the transmit
antenna. The signal generator and spectrum analyzer are directly con-
nected via coaxial cables and suitable connectors.
• The signal generator is set to 120 dBµV, and frequency is set to 30
MHz.
• A coaxial cable is connected to the RF input of the spectrum analyzer
with center frequency set to 30 MHz, with 200 kHz SPAN, RB = 10
kHz, VB = 100 Hz, and sweep time of 200 ms.
• The reference level is set to a suitable level according to the signal

strength to be measured.
• The value at the spectrum analyzer is recorded as VDirect at 30 MHz.
This procedure is repeated for every measurement frequency without
changing the level at the signal generator.
• Then, coaxial cables are disconnected. The signal generator and its cable
are connected to the transmit antenna over a 10-dB attenuator; the spec-
trum analyzer’s cable is connected to the receive antenna over another 10
dB attenuator via an adaptor.
• The transmit antenna is located to the mast and fixed at 1m height
above the ground plane.
• The coaxial cable is fixed horizontally and extended by at least 2m be-
hind the antenna before dropping to the ground and connected to the
signal generator.
• The receiver antenna is located 3m or 10m away from the transmit
antenna.
• Signal generator is set to dBµV and frequency is set to 30 MHz. MAX
HOLD is selected at the spectrum analyzer.
• The receiver antenna is moved vertically between 1m and 4m by scan-
ning from the control room.
• Use MARK PEAK to record maximum field value.
• Repeat the procedure for other frequencies.
Measurements are repeated for 3m and 10m for both polarizations. A

typical measurement chart is given in Table 7.2. The results of the OATS cal-
ibration are given in Figures 7.4–7.7. These figures belong to 3m and 10m
calibration measurements for both horizontal and vertical polarizations. The
dashed lines belong to theoretical NSA calculations; solid lines are the real
OATS measurement results; the dots show the ±4 dB margins. As observed,
the newly established OATS successfully completes the calibration procedure.
Note that the antennas used in OATS calibration are calibrated in the Na-
tional Metrology Institute (UME, www.ume.tubitak.gov.tr) according to [5];
therefore, AF versus frequency values listed in Table 7.2 belong to this calibra-
tion data. In fact, OATS calibration indirectly verifies the antenna calibration
performed in UME.
Note also that OATS calibration is an important EMC engineering task.
The keywords for OATS calibration are traceability, accreditation, repeatability
(precision), and accuracy. Both CISPR and ANSI/IEEE standards [12, 13] pres-

Table 7.2
NSA Measurement Table (d=10m, Hor-Pol)
EMC Test and Measurement Environments
187
Figure 7.4 NSA versus frequency and uncertainty limits (Hor-Pol, d=10m).
Figure 7.5 NSA versus frequency and uncertainty limits (Ver-Pol, d=10m).
Figure 7.6 NSA versus frequency and uncertainty limits (Hor-Pol, d=3m).
Figure 7.7 NSA versus frequency and uncertainty limits (Ver-Pol, d=3m).
ent every detail of the calibration procedure. NSA is an important parameter

for the OATS calibration.
7.3 Screened Room and Resonance Effects

Some EMC tests and measurements are conducted in a screened (shielded)
room. The screened room is simply a Faraday cage; the walls, floor, and ceiling
are covered with conducting plates. The screened room isolates the inside from
the outside. Outside EM fields from various sources induce surface currents on
these walls. If the thickness of these conducting plates is much bigger than the
skin depth, then outside EM fields cannot penetrate inside. For a good isolation,
the thickness of the metal plates should be at least 4–5 times the skin depth. A
screened room gives more than 100–150 dB isolation if it is mono-block. Un-
fortunately, every screened room has a door, grounding/cabling holes, and the
like, which drastically reduce the isolation. This may be prevented somewhat by
using shielding paints, conductive coated plastics, and so on.
Usually, 40–60 dB screening is assumed average; a good screening means
at least 80–100 dB isolation. Two different commercial screened rooms and
their screening effectiveness are pictured in Figures 7.8 and 7.9. In Figure 7.8,
a screened room made of one- and two-layer aluminum folio is shown. This is
simple, portable, and affordable for precompliance tests that can be performed
in a company. As shown, 35–80 dB of isolation can be obtained with this sim-
ple screened room.
Figure 7.9 shows a certified screened room. Good isolation is obtained in
these kinds of screened rooms with 1–3 mm thick ferromagnetic metals. Addi-
tional chemical processes are usually applied to prevent corrosion, rusting, and
so on. Here, a 2-mm thick MuFerro steel is used. As seen, more than 120-dB
shielding is obtained for a wide range of frequencies (from 120 kHz up to 6–7
GHz). Shielding performances of screened rooms degrade for both very low
and very high frequencies.
In addition to EMC tests and measurements, screened rooms are also
used in data security, in biomedical applications like EEG and MR, in military
applications (TEMPEST), in EMP nuclear shelters, in noise-free rooms for
sensitive electronic devices, in microwave radars and communication systems,
and so on.
A screened room provides inside/outside isolation but brings in undesired
resonance effects. Any finite medium (a wire in 1D, a plate in 2D, or a room/
enclosure in 3D) is a resonator. Suppose you have a piece of conducting wire
(with length l) inside your design (a device or equipment) connected, for ex-
ample, with the enclosure at one end and a printed board at the other. You use
this to ground the board, but you also introduce a resonating antenna. This
Figure 7.8 An aluminum tent screened room.
wire will resonate at frequencies frm = mc/2l · (m = 1, 2, 3, …) and strongly

radiate EM waves inside the enclosure. Fields along the wire at resonances are
called modes. If there are radiating fields inside the enclosure they also strongly
couple with this wire and cause CM currents. If, for example, l = 10 cm, the
first (dominant) resonance frequency is 1.5 GHz, and the rest are integer mul-
tiples of this frequency (i.e., 3 GHz, 4.5 GHz, 6 GHz, and so on).
Note that to observe a resonance, it must be excited and received at the
same time. This means both the source and the receiver points must not be lo-
cated at null-points (i.e., zero-crossing) of that resonance. For example, the null
points of the first resonance (i.e., the dominant mode, m = 1) is at both left and
right terminations, and its amplitude is maximum in the middle of the wire).
On the other hand, the second resonance (m = 2) has null points at both ends
as well as in the middle, along the wire. This is same for all other higher even-
order modes (i.e., m = 4, m = 6, and so on.) This means if either of the source
or observation points is located at midpoint, then these modes are either not
excited or not observed. These are also valid for 2D and 3D enclosures. By just
Figure 7.9 Metal construction screened room.
choosing proper locations for your source and receiver, you may excite a reso-
nance or you may get rid of that resonance. This is critical in EMC engineering.
The resonance frequencies of a screened room with dimensions a, b, and
c can be calculated from
2 2 2
c m n  p
f mnp =   +   +   (7.1)
2 a b c
Table 7.3 lists a short MATLAB code that calculates the first several reso-
nances of a screened room. A sample output of this code, for a screened room
with a = 4.5m, b = 2.5m, and c = 2.5m, is given in Figure 7.10.
It is possible to locate absorbers (e.g., ferrite tiles) in a screened room and
damp some of the resonances that are critical [14]. The damped screened room
may then be used as a precompliance test facility for both radiated emission and
immunity testing.
Table 7.3
MATLAB Code Calculating Resonances in a Screened Room
%--------------------------------------
% Program: RezDamping.m
%--------------------------------------
clc; clear all; close all;
fmin = 30e6; fmax = 230e6; c=3e8;k=1; a=4.5; b=2.5; d=2.5;
fprintf(‘ \n’);
fprintf(‘---Lowest Order Resonances of the screened room--- \n’);
fprintf(‘ (a = %5.2f [m] b = %5.2f [m] d = %5.2f [m]) \n’,a,b,d);
fprintf(‘---------------------------------------------------- \n’);
fprintf(‘ \n’);
for m=0:4; for n=0:2; for p=1:1
f(k)=(c/2)*sqrt((m/a)^2+(n/b)^2+(p/d)^2); f(k)=f(k)/1e6;
fprintf(‘ m = %2d n = %2d p = %2d f [MHz] = %8.2f \n’,m,n,p,f(k));
k=k+1;
end; end; end % Program END
Figure 7.10 Sample output of the code given in Table 7.1.

Carbon-loaded absorbers may be placed, for example, on the floor at po-

sitions of electric field maxima within the chamber. Similarly, the magnetic
fields for all modes in a resonant enclosure have maxima that are tangential
to the walls. This means that a lossy magnetic material can be used to absorb
energy when it is placed on the walls. This leaves the volume of the room free
of obstruction. Ferrite tiles sold as wave absorbers are suitable for this purpose.
Figure 7.11 shows recorded field strengths of a screened room before and after
resonance damping is applied [14].
The lowest resonance is called dominant mode. A few dominant modes
of a screened room with length, width, and height a, b, and c are pictured in
Figure 7.12.
Theoretically a screened room has infinite resonances, but only those that
fall in the signal frequency band are critical. At a resonance frequency, EM
signals inside a screened room may be amplified up to 20–40 dB. Figure 7.13
pictures a typical case that compares RE measurement results performed inside
a screened room and anechoic chamber. Resonance effects are clearly observed
in this figure.
7.4 Anechoic Chamber

A fundamental issue in EMC tests and measurements is to choose a site that
guarantees that whatever received with the receive antenna and recorded with
the EMI device belongs only the EUT. Calibrated OATS in electronically quiet
areas satisfy this condition. It can also be satisfied in an indoor site if (1) ex-
ternal noise and interference are isolated; (2) indoor reflections are eliminated.
Anechoic chambers satisfy these two and are used as indoor tests and measure-
ments sites. Figures 7.14 and 7.15 present photos of two anechoic chambers
used for military and commercial EMC tests and measurements, respectively.
Figure 7.11 Recorded field strengths before and after resonance damping.
Figure 7.12 A resonator and its resonance frequencies.
Figure 7.13 RE measurements in a screened room and anechoic chamber.
All four side walls, the ceiling, and the floor are covered with absorbing
materials in a fully anechoic chamber (FAC). As shown in Figures 7.12 and
7.13, tests and measurements are performed for heavy vehicles where the floor
is covered with metal plates. These types of chambers are semi-anechoic (SAC).
CISPR16-1-4 [1] explains all the requirements required to be satisfied by an-
echoic chambers.
An anechoic chamber that operates down to 30 MHz is either very large
with up to 2.5m depth of carbon-loaded absorber over all surfaces or is smaller
but requires greater strength to be lined with ferrite tiles. The ferrite tiles are
expensive, and the room strength required to support the weight of the tiles
adds to the cost. Ferrite tiles on their own are unlikely to give the performance
Figure 7.14 Anechoic chamber and an RE measurement setup (ASELSAN/Ankara).
Figure 7.15 Anechoic room and an RI measurement setup (OTOKAR/Turkey).
of an open area test site, especially in small rooms, although for precompliance
work this is not necessarily a problem. Composite absorbers (carbon-loaded
foam with a ferrite backing) fall midway between the extremes of weight and
size and can also be used.
7.5 TEM/GTEM Cell

Transverse electromagnetic (TEM) transmission-line cells are devices used for
establishing standard electromagnetic fields in a shielded environment (see Fig-
ure 7.16). The cell consists of a section of rectangular coaxial transmission line
tapered at each end to adapt to standard coaxial connectors. TEM cells are used
for emission testing of small equipment, for calibration of RF probes, and for
biomedical experiments. The wave traveling through the cell has essentially a
free-space impedance (377W), thus providing a close approximation to a far-
field plane propagating in free space. The cell also has limitations, among which
is that the upper useful frequency is bound by its physical dimensions, which,
in turn, constrain the size of an item you can test with the cell.
Gigahertz transverse electromagnetic (GTEM) cells are anechoic cham-
bers used for EMC tests and measurements (mostly immunity tests) of small,
portable high-frequency equipment. A typical GTEM cell is pictured in Figure
7.17. GTEM cells are designed to produce high intensity but homogeneous
EM fields (i.e., plane waves). Requirements for emission measurements and im-
munity tests inside a GTEM cell are given in EN61000-4-20 standard. GTEM
cell is a transmission structure using a TEM cell approach. A slightly spherical
wave propagates from the source into a 50W rectangular coaxial transmission
line and its distributed hybrid termination without geometrical distortion of
the TEM wave. Since the opening angle of the waveguide is small, the undis-
torted spherical wave can be considered a plane wave.
Photos of commercial TEM and GTEM cells are shown in Figure 7.18.
Figure 7.16 Side view of a TEM cell. The height of the cell (h) is approximately three times
of the height of the EUT. The width (cross section) of the cell is d. The maximum frequency of
the cell depends on d and h.
Figure 7.17 Side view of a GTEM cell.
Figure 7.18 Photos of commercial (a) TEM, and (b) GTEM cells (www.teseq.com).
7.6 Reverberation Chamber

Another chamber used in EMC tests and measurements (mostly in immunity
tests) is the reverberation chamber or mode-stirred chamber. A reverberation
chamber is simply an empty resonator with high-quality factor. The aim is to
create high EM fields with high voltage standing wave ratios. It looks like a
screened room; the difference is that there are passive, metal reflectors and stir-
rers in reverberation chambers. With the help of these reflectors and stirrers, it
is possible to create high-intensity EM fields with low/medium input powers.
The minimum usable frequency of a reverberation chamber depends on the
room sizes as well as reflector/stirrer design. A photo of a typical reverberation
chamber is shown in Figure 7.19.
Figure 7.19 A reverberation chamber used in immunity tests.
References
[1] CISPR 16-1-4: 2003 Radio Disturbance and Immunity Measuring Apparatus—Ancillary
Equipment—Radiated Disturbances.
[2] CISPR16-1-5: 2003 Antenna Calibration Test Sites (CALTS) for 30 MHz to 1000 MHz.
[3] Federal Communications Commission, “Calibration of a Radiation Measurement Site-
Site Attenuation,” in Bull. OCE 44. Washington, DC: US Government. Printing Office,
Sept. 1977, Docket 21371.
[4] ANSI/IEEE C63.4-1992, American National Standard Guide for Methods of Measure-
ment of Radio-Noise Emissions from Low-Voltage Electrical and Electronic Equipment
in the Range of 9 kHz to 40 GHz.
[5] ANSI/IEEE C63.5-2006 (Revision of C63.5-2003) American National Standard Guide
for Electromagnetic Compatibility—Radiated Emission Measurements in Electromag-
netic Interference (EMI) Control—Calibration of Antennas (9 kHz to 40 GHz).
[6] Smith, A. A., R. F. German, and J. B. Pate, “Calculation of Site Attenuation from Antenna
Factors,” IEEE Trans on EMC, Vol. 24, No. 3, Aug. 1982, pp. 301–316.
[7] Trakadas, P. T., and C. N. Capsalis, “A Mixed Model for the Determination of Normalized
Site Attenuation in OATS,” IEEE Trans. on EMC, Vol. 43, No. 1, Feb. 2001, pp. 29–36.
[8] Askri, A., C. Vollaire, L. Nicolas, and D. Prebet, “Normalized Site Attenuation Standard
Correction from Numerical Computing,” IEEE Trans. on EMC, Vol. 38, No. 2, Mar.
2002, pp. 693–696.
[9] Sevgi, L., “The Antenna as a Transducer: Simple Circuit and Electromagnetic Models,”
IEEE Antennas and Propagation Magazine, Vol. 49, No. 6, Dec. 2007, pp. 211–218.
[12] ANSI C63.6-1996 (Revision of C63.6-1988), American National Standard Guide for the
Computation of Errors in Open Area Test Site Measurements.
[13] ANSI C63.7-1992 (Revision of C63.7-1988), American National Standard Guide for
Construction of Open Area Test Sites for Performing Radiated Emission Measurements.
[14] Dawson, L., J. F. Dawson, A. C. Marwin, and D. Welsh, “Damping Resonances Within
a Screened Enclosure,” IEEE Trans. on Electromagnetic Compatibility, Vol. 43, No. 1, Mar.
2001, pp. 45–55.
Bibliography
EN 61000-4-20:2010: Electromagnetic Compatibility (EMC)—Testing and Measurement
Techniques. Emission and Immunity Testing in Transverse Electromagnetic (TEM) Waveguides.
8
EMC Test and Measurement Devices
Fundamental measuring equipment used in electrical and electronics engineer-
ing are a multimeter (MM), oscilloscope, spectrum analyzer (SA), and network
analyzer (NA). A MM measures resistance, voltage, and current. It is also used
to check a fault in a circuit, and to find out whether the source supplies energy.
An oscilloscope displays time variations of DC and AC signals. SA and EMI
receivers display signal versus frequency (an EMI receiver is a slightly modified
SA). An NA is used to investigate input/output characteristics (e.g., insertion
loss, reflection loss, or voltage transfer functions of n-port circuits), impedance
versus frequency variation on Smith chart, and so on.
Generally speaking, MMs and EMI receivers are mostly used in EMC
tests and measurements.
8.1 Digital Multimeter

The first thing freshmen learn in basic electrical engineering circuit theory labs
is how to measure the voltage across a resistor, the current through a resis-
tor, and the resistance itself. An MM is a voltmeter, ammeter, and ohmmeter
combined into a single device [1, 2]. It is used to measure potential in [V],
current in [A], and resistance in [Ω]. MMs are connected in parallel (series)
when measuring voltage between (current through) two nodes. MMs are also
used to identify electrical and electronics problems. Advanced instruments have
additional features such as diode and IC testing modes, and so on. These tests
check to see if a diode or an IC is good or bad. Digital multimeters (DMMs)
have replaced analog ones.
201
DMMs are classified according to their accuracy, sensitivity, precision,

and resolution. Very often, accuracy, sensitivity, precision, and resolution are
misused and confused with each other [3] (see Section 2.6 for the definitions of
accuracy, precision, and sensitivity).
Resolution is the ability to resolve, to discriminate. It is also defined as the
smallest physically indicated division that an instrument displays or is marked.
The resolution of DMMs is often expressed in bits, such as 8, 10, 12,
and so on [4]. Bits specifically refer to the performance of the analog-to-digital
converter (ADC). The resolution ∆ of an n-bit ADC is ∆ = Vpp/2n–1, where Vpp
is peak-to-peak (or in general, full scale) voltage. Here, 2n is the least significant
bit (LSB). For example, the resolution of a 16-bit ADC is ∆ = 1/216 = 1/65,536
(i.e., LSB=65,536). If a 20V full-scale (20V pp ⇒ 10V p ) is being used, then the
resolution is 20/65,536 = 305 µV. It means the smallest change that can be
detected by a device with this scale is 305 µV.
Digits of resolution (DOR) can be obtained from LSB using DOR =
log10 (LSB). A DMM with a 12-bit ADC has a digit of resolution of DOR
= log10(4,096) (note that noise may reduce the value of LSB, therefore reduc-
ing DOR). For example, with a four-digit DMM, the resolution is 1/10,000;
therefore, it must be using more than 14-bit ADC (213 = 8182, 214 = 16384).
Since quantization noise power is obtained directly from the resolution ∆ as
PqN = ∆2/12, signal to quantization noise ratio (SQNR) of an n-bit ADC, for
a 1V pure sinusoidal signal is SQNR = (1/2) ÷ (∆2/12) = 3 × 22n/2 or SQNRdB
= 1.76 + 6n (there is a 6-dB improvement for every extra bit). Note that the rms
quantization noise voltage is Vrms = ∆ / 12 . For example, an 8-bit A/D that
can measure voltages from –1V to +1V has a resolution of 2/256 = 7.8 mV, and
expected rms noise voltage is then equal to 2.3 mV.
8.2 Noise-Limited Detection

The theoretical limit of sensitivity in a measurement is limited by the noise.
Sensitivity of an electronic receiver shows the lowest signal level that is measur-
able with the device (called the floor signal). Usually, receiver sensitivity (i.e., the
floor signal) is determined by either internal or external EM disturbances. EM
disturbances caused by internal and/or external, manmade, or natural sources
are generally called noise. For example, manmade sources, such as power lines,
electric razors, blow dryers, and so on, and, natural sources, such as lightning,
electrical storms, galactic effects, and so on, are typical external noise sources.
On the other hand, power amplifiers, mixers, diodes, and transistors are some
of the internal noise sources because of random collisions of the electrons. A
variety of noise sources and their effective frequencies are shown in Figure 6.1
of Section 6.2 [5].
EMC Test and Measurement Devices 203
Electronic circuits and receivers are affected by a variety of noise sources.

In terms of electronic receivers, noise can be separated into two groups: thermal
noise (also called internal noise, caused by electronic devices themselves) and
the environmental noise (atmospheric, cosmic, manmade, and so on). Roughly
speaking, thermal noise dominates the others above a few hundred MHz fre-
quencies. Therefore, noise-limited electronic detection can be grouped into
thermal noise-limited detection (e.g., microwave receivers) and environmental
noise-limited detection (e.g., HF receivers in 3–30 MHz band).
As discussed in Section 6.2, thermal noise power may be calculated from
PN = kTB [W], where k = 1.3810–23 J/°K is Boltzman’s constant and T is the
temperature in degrees Kelvin [°K]. At room temperature (i.e., for T = 290°K)
and in dB this equation becomes PN = −204 + BdBHz [dBW]. For example, noise
thresholds for an FM signal with 10 kHz bandwidth, a GPS signal with 2-MHz
bandwidth, and a TV channel with 6-MHz bandwidth, respectively, are PN =
–134 dBm, PN = –111 dBm, and PN = –106 dBm (X dBW=X+30 dBm). The
noise threshold voltage for a 50Ω receiver is V N = PN × 50 . For example, 10-
pW noise power (i.e., PN = –110 dBm) corresponds to VN = –22µW (13 dBµV)
noise voltage.
The environmental noise (day- and nighttime atmospheric noise and
manmade noise in rural and urban areas) determine the noise floor at HF fre-
quencies (i.e., between 3 MHz and 30 MHz). As shown in Figure 6.1, night-
time atmospheric noise level is much higher than daytime, especially at lower
HF frequencies, which degrades the performances of the HF radars at night-
times. Environmental noise can be expressed by the empirical formula [5]
 f 
N e = N 0 − 12.6 ln  MHz   +10Log 10 (B ) [dBW ] (8.1)
  3 
where, No is equal to the manmade noise level of the measurement site (typi-
cally, –136, –148, or –164 in residential, rural, or remote sites, respectively) and
the term inside the first parenthesis is the noise density given in dBW/Hz. The
noise power of an HF receiver operating at 5 MHz with 1-MHz bandwidth
at residential and remote sites will be −82.4 dBW and −110.4 dBW, respec-
tively (which are nearly 55–70 dB higher than thermal noise levels). They are
dominant at different frequencies. At microwaves, environmental noise can be
neglected, but below, thermal noise is negligible.
Note that both thermal and environmental noise powers depend on sig-
nal bandwidth (i.e., the bandwidth of the measurement). Therefore, one may
expect DMM’s DC voltage measurement noise to be zero (because the fre-
quency of the signal being measured is zero). This is not true. The measurement
bandwidth of the DMM is determined by the number of (measurement) sam-

ples per second, and this varies from a few samples to a few mega-samples (e.g.,
if the DMM is digitizing at a 1 kilo-sample per second, then the measurement
bandwidth is 1 kHz).
8.3 The Loading Effect

There are many sources of errors when doing measurements with DMMs. Al-
though incomplete, the list includes RFI, thermoelectric voltages (generated
when circuit connections are made between dissimilar metals at different tem-
peratures), nearby magnetic fields, imperfect grounding errors such as noise and
offset voltages, and the loading error.
In general, when two circuits are connected, the voltages and currents in
both circuits change. These circuits are usually referred to as source and load.
When an electronic measurement is being made, the circuit under test is the
source, and the measuring device becomes the load. The measuring device
always has an effect on the circuit. A source under test may generally be as-
sumed as a Thevenin source or may be represented with a Thevenin equivalent
circuit. A Thevenin equivalent circuit is made up of a voltage source and a
series source resistance. A voltage measuring device may be represented with
an input impedance in parallel to the voltage display (see Figure 8.1). There-
fore, the voltage measurement device displays the measured voltage correctly
if and only if the voltage input resistance is negligibly low and devices input
impedance is significantly high. Otherwise, an error is introduced called load-
ing error [1, 2].
As shown in Figure 8.1, instead of measuring VS the MM measures Vm =
Vs × Rm/(Rs + Rm), and ±∆ Vm is called the loading error(DVm = |Vm – Vs| = Vs
× RS / (Rs + Rm)). To reduce the loading error, RS should be low or Rm should
be high.
Figure 8.1 DMM connection and voltage measurements.

8.4 Digital Multimeters and Specs

DMMs are basic measurement/test instruments in EMC engineering. Different
applications require different DMMs with different specifications. The main
purpose of the specs is to establish the measurement uncertainty (i.e., error
bounds). A good understanding of these specifications is critical for the confi-
dence that the readings accurately reflect reality.
DMMs display measurement data on the readout with a given number of
DOR (e.g., 4-digit, 5.5-digit, 7-digit, 8.5-digit). A 5.5-digit DMM has 5 full
digits with the display values from 0 to 9, and one half digit that could display
only 0 or 1. Today, there are excellent DMMs with both high speed and high
accuracy (see, e.g., [6–10] for data sheets of such DMMs). Measurements from
10,000 to 100,000 samples per second are achievable with high level preci-
sions of 7–9 digits of measurement resolutions, 0.1 ppm accuracies (1 ppm =
0.0001%), and complete self-calibration capabilities.
Baseline uncertainty specifications are usually given as ±(percent of read-
ing + number of digits) or ±(percent of reading + number of counts). The count
for a DMM is analogous to LSB for an ADC. In some specification, data sheets
use the form ±(percent of reading + percent of range). High-accuracy, high-
speed DMM also uses the form ±(ppm of reading + ppm of range). “Digit” and
“count” are used in the same meaning and they indicate the value of LSB for a
specified range. Here “percent of reading” is named as the scaling error or gain
error, and “count” is called the quantization error. Scaling error is proportional
to, but quantization error is independent of, the measured value. Total error of
the DMM is the addition of scaling and quantization errors.
All electronic instruments (including DMM) change with time; therefore,
specs are usually given for a period of time beyond which accuracy cannot be
guaranteed. Uncertainty is modified accordingly using the specifications which
account for environmental and time factors. Specifications may apply for 24
hours, 90 days, or 1 year after the last calibration.
Temperature also affects accuracy of a measurement, and DMM specifica-
tions are generally given for a range of temperatures (i.e., Tcal ± 1°C or Tcal
± 5°C, where Tcal is the calibration temperature). The error introduced for a
measurement performed outside this temperature range is usually given as a
temperature coefficient (tempco) such as ±(0.0005%+0.1count)/°C or ±(5ppm
of reading+1ppm of range)/°C. A typical spec for a DC voltage measurement
is given in Table 8.1 [7].
Suppose that 8V DC voltage is measured with 10V range for 90-day ac-
curacy specifications where Tcal = 23°C. The accuracy for this case is given as
±(ppm of reading + ppm of range). Accuracy specifications in Table 8.1 are said
to be relative to the calibration standards. Absolute accuracy is determined by
adding these relative accuracies to the traceability of the calibration standard.
Traceability error is the absolute error relative to the standards used. For DC
voltage measurements, the factory traceability error of this product is given to
be 2 ppm [7].
Question 1: What is the relative accuracy if 24-hr operating temperature
is Tcal ± 1°C?
Answer: The accuracy specification in Table 8.1 for this case is ±(0.5 ppm
of reading + 0.05 ppm of range). Using these specifications, relative measure-
ment error is obtained as ±(0.5/1,000,000 × 8) + (0.05/1,000,000 × 10) = ±4.5
µV or 0.45 ppm of 10V.
Question 2: What is the relative accuracy if the measurement is performed
after 24 hours and the operating temperature is 27°C?
Answer: As shown in the table, the optimum technical specifications of
this DMM are based on auto-calibration (ACAL) of the instrument within the
previous 24 hours and for the operating temperatures of Tcal ± 1°C. After 24
hours, the accuracy specification is ±(4.1 ppm of reading + 0.05 ppm of range).
This yields a total relative error of ±(4.1 ppm × 8V) + (0.05 ppm × 10V) =
±33.3 µV. There is no temperature correction for this example.
Table 8.1
Technical Specs of Agilent 3458 DMM
DC Voltage
Temperature Coefficient (ppm
Maximum Input or Reading + ppm of Range)/°C
Range Full Scale Resolution Impedance Without ACAL1 With ACAL2
100 mV 120.00000 10 nV >10 GΩ 1.2 + 1 0.15 + 1
1V 1.20000000 10 nV >10 GΩ 1.2 +0.1 0.15 + 0.1
10 V 12.0000000 100 nV >10 GΩ 0.5 + 0.01 0.15 +0.01
100 V 120.0000000 1 µV 10 MΩ ± 1% 2 + 0.4 0.15 + 0.1
1000 V 1050.00000 10 µV 10 MΩ ± 1% 2 + 0.04 0.15 + 0.01
Accuracy3 [ppm of Reading (ppm of Reading for Option 002) + ppm of Range]
Range 24 Hour4 90 Day5 1 Year5 2 Year5
100 mv 2.5 + 3 5.0 (3.5) + 3 9 (5) + 3 14 (10) + 3
1V 1.5 + 0.3 4.6 (3.1) + 0.3 8 (4) + 0.3 14 (10) + 0.3
10 V 0.5 + 0.05 4.1 (2.6) + 0.05 8 (4) + 0.05 14 (10) + 0.05
100 V 2.5 + 0.3 6.0 (4.5) + 0.3 10 (6) + 0.3 14 (10) + 0.3
1000 V6 2.5 + 0.1 6.0 (4.5) + 0.1 10 (6) + 0.1 14 (10) + 0.1
1Additional error for Tcal or last ACAL±5°C.
2Additional error for Tcal±5°C.
(Tcal: calibration temperature; Acal: Auto calibration in the last 24 hours when the temperature change is less
than ±1°C).
Question 3: What is the relative accuracy with and without ACAL if the
measurement is performed after 24 hours and the operating temperature is
35°C?
Answer: The total relative accuracy for this case is the same with the pre-
vious example: ±(4.1 ppm × 8V) + (0.05 ppm × 10V) = ±33.3 µV. The tem-
perature coefficient, without ACAL, is equal to ±61.2 µV (±((0.5 ppm × 8V)
+ (0.01 ppm × 10V)) ×12°C). The total relative error is the addition of these
two: (33.3+61.2) = ±94.5 µV. Assuming the same conditions, but using ACAL,
significantly reduce the error due to temperature difference from calibration
temperature. Since the measurement temperature is 7°C beyond the standard
range of Tcal ± 5°C the temperature coefficient is then equal to ±((0.15 ppm
× 8V) + (0.01 ppm × 10V)) ×7°C = ±9.1 µV. Finally, the total relative error is
(33.3+9.1) = ±42.4 µV. If the absolute error is of interest for the last two ex-
amples, the traceability error of (2 ppm × 8) = ±16 µV should be added.
Absolute errors of the past two examples can be calculated by adding the
traceability error (2 ppm × 8) = ±16 µV to the relative errors.
DMM resolution is inversely proportional with the measurement speed.
Speeding up the measurements makes the resolution worse. For example, the
DMM given in [7] supplies 100,000 readings per second with only 4.5-digit
resolution; this is reduced to 6 readings per second if the resolution is raised to
8.5-digit.
Aperture (or integration) time is a period during which the ADC reads the
input signal. Increasing the aperture time potentially improves the accuracy of
the measurement. This is because more samples are internally averaged during a
longer aperture time (noise is sample-to-sample uncorrelated signal; therefore,
simple averaging reduces the noise level and increases SNR). Configuring the
DMM’s aperture time to be an integer multiple of power line cycles also reduces
power line noise. This is called normal mode rejection (NMR). Aperture time
is often expressed in number of power line cycles (NPLCs), where 1 PLC for
60 Hz (50 Hz) is 16.67 ms (20 ms). The noise can be greatly reduced by set-
ting the aperture time to 1 PLC or greater. The larger the value of NPLC, the
greater the line noise reduction, but the longer the measurement period. If frej
and are NS the signal frequency to be rejected and the number of samples to be
averaged, respectively, the aperture time can be set to Ta = NPLC/(NS × frej) in
order to eliminate frej and DC voltage noise. For example, averaging 40 samples
for a NPLC of 3 acquired at a sampling interval of 1.25 ms (or 800 samples/s)
rejects 60 Hz power line noise.
Question 4: What do you know about the accuracy, resolution, and sensi-
tivity of the DMM in Table 8.2?
Answer: The DMM’s resolution is the smallest number called count that
can be displayed. Suppose a DC voltage is measured with the 3V scale of the
DMM described in Table 8.2. Since a DC voltage with 4-digit resolution can
Table 8.2
Technical Specs of a 4-Digit DMM
DC Voltage
Scale: 300 mV, 3V, 30V, 300V, 1000V
Accuracy: ±(0.75% of reading + 3 counts)
Input resistance: Min 10 MΩ
AC Voltage
Scale: 3V, 30V, 300V, 750V
Accuracy: ±(1.2% of reading + 4 count)
Max frequency 500 Hz
DC current
Scale: 300 µA, 3 mA, 30 mA, 300 mA, 10A
Resistance
Scale: 300 Ω, 3 kΩ, 30 kΩ, 300 kΩ, 3 MΩ, 30 MΩ
be displayed as, for example, 2.543V, or 1.902V, and so on, one count that can
be displayed (i.e., the resolution) is 0.001V. If, on the other hand, a DC voltage
is being measured by the 300V scale, then one count will be 0.1V (the DMM
displays a DC voltage of, for example, 267.3V, or 112.8V). This shows that the
resolution is a relative quantity.
Specifically, for the DMM described here, sensitivity of the device can be
assumed as the resolution under the lowest scale. For example, in this 4-digit
DMM, minimum DC voltage measurement scale is 300mV and one count is
equal to 0.1 mV or 100 µV, therefore sensitivity is said to be 100 µV.
Question 5: What do you expect to see on the display if you want to mea-
sure a 20V DC voltage source having an internal resistance of 5 kΩ, and 150
mV having an internal resistance of 1 kΩ? What are the resolutions and the
absolute and relative errors of these measurements?
Answer: The accuracy of this DMM for different quantity measurements
are specified as in the table. For example, for a DC voltage measurement it is
given to be “±(0.75% of reading + 3 count).”
Error introduced because of loading effect and total measurement error
because of scaling and quantization errors should be calculated for each case. In
general, error introduced because of loading effect is much less than the limited
accuracy of the device and is negligible.
To measure a DC voltage of 20V, the DMM should be best used in 30V
scale. Since the internal and input resistances are given to be 5 kΩ and 10 MΩ,
respectively, the loading error of this measurement will be 20×5/(10005)=0.01V.
The scaling error is calculated to be 20 × 0.0075 = 0.150V. The 4-digit DMM
displays this measurement result as, for example, 19.55. Since, one count is
0.01V, the quantization errors will be 0.03V. As a result, the total error will be
the addition of the loading, scaling, and quantization errors, ±∆Vm = 0.01 +
0.15 + 0.03 = 0.19V. The value of the measured DC voltage is then in the range
of 19.81V ≤ Vmeas ≤ 20.19V.
To measure a DC voltage of 150 mV, the DMM should be best used in
300 mV scale. In this case the three error values are:
• Loading error: 150×1/(10005)=0.015 mV.

• Scaling error: 150×0.0075=1.25 mV.
• Quantization error: 0.3 mV
• Total error: ±∆Vm=0.015+1.25+0.3 mV=1.565 mV=1.6 mV.
The measured DC voltage is then in the range of 148.4 mV ≤ Vmeas ≤

151.6 mV.
Question 6: What do you expect to see on the display if you want to mea-
sure a resistor of 75Ω and 75 kΩ? What are absolute and relative errors of these
measurements?
Answer: The total error in resistance measurements is given as ±(0.75% of
reading + 3 count). For a 75Ω resistor, the scaling and quantization errors are
75 × 0.0075 = 0.5625Ω, and 0.03Ω, respectively. The total error is ±0.5625
+ 0.03=0.59Ω. The value of the measured resistor is then between 74.41W ≤
Rmeas ≤ 75.59Ω.
For a 75-kΩ resistor, the scaling and quantization errors are 75 × 0.0075
= 0.5625 kΩ, and 0.03 kΩ, respectively. The total error is ±0.5625 + 0.03 =
0.59 kΩ. The value of the measured resistor is then between 74.41 kΩ ≤ Rmeas
≤ 75.59 kΩ.
Question 7: The resistance can also be measured using voltage-current
method and Ohm’s law. The DC voltage across the terminals of and the DC
current through the 75-kΩ resistor are measured to be 75V and 10 mA, respec-
tively. What do you expect the total (propagated) error to be in this resistance
measurement if the voltage-current method and Ohm’s law are used?
Answer: The total error in this case can be calculated using the error prop-
agation approach. Ohm’s law states that R=V/I; therefore, total relative error
will be the addition of relative errors of voltage and current measurements.
Total relative error of the DC voltage measurement:
• Scaling error: 75 × 0.0075 = 0.5625V.

• Quantization error: 0.03V (one count is 0.01V).
• Total relative error: 0.59/75 = 0.79%.
Total relative error of the DC current measurement:

• Scaling error: 10 × 0.015 = 0.15 mA.

• Quantization error: 0.05 mA (one count is 0.01 mA).
• Total relative error: 0.2/10 = 2%.
As a result, total relative error of the voltage-current method will be

2.79%. This yields an absolute error of 2.09 kΩ. The value of the measured
resistor is between 72.91 kΩ ≤ Rmeas ≤ 77.094 kΩ.
The ability to work with today’s high-speed DMM, and understanding
and using uncertainty specifications is a fundamental skill. This is so basic that
we usually neglect to look at the requirements such as accuracy, resolution,
precision, and so on.
Choosing the right DMM for your measurements is important; it directly
affects the integrity and repeatability of the measurements. One needs to well
understand instrument categories and know physical and mathematical mean-
ings of the specs of the instrument at hand.
DMMs are usually adequate if the source resistance is around 1 MΩ or
if the desired resolution is no better than 0.1 µV (with low source resistance).
For low-level signals or high-source resistances, some other specialized instru-
ments should be used. Nanovoltmeters can be used for low voltages up to a few
nanovolts. Electrometers are DC MMs with input resistance up to 1016 Ω and
input current as low as 10–17 A. They can also measure charge and resistance.
Picoammeters are used when only current must be measured. Microohmmeters
are designed for low-level resistance measurements; they can resolve resistances
as low as 10 µΩ.
8.5 Measurement Receivers

The two critical measurement instruments in EMC engineering are the spec-
trum analyzers and the EMI receivers.
8.5.1 Spectrum Analyzer

Design engineers, manufacturers, and field service/repair technicians need a
tool that will help to analyze electrical signals that are passing through or being
transmitted by their system, device, or equipment. These electrical signals can
be measured without interference with a measurement device called a spectrum
analyzer (SA) [14]. A SA measures the magnitude of an input signal versus
frequency within the full frequency range of the instrument. It usually displays
raw, unprocessed signal information such as voltage, power, period, wave-shape,
sidebands, and frequency. The primary use is to measure the power of the spec-
trum of known and unknown signals. By analyzing the spectra of electrical sig-
nals, dominant frequency, power, distortion, harmonics, bandwidth, and other

spectral components of a signal can be observed that are not easily detectable in
time domain waveforms. These parameters are useful in the characterization of
electronic devices. Table 8.3 lists fundamental technical specs of a typical SA.
The display of a SA has frequency on the horizontal axis and the amplitude
displayed on the vertical axis. In EMC tests and measurements, a SA is used for
basic precompliance testing; however, it cannot be used for full EMC tests and
measurements for certification. Instead, an EMI receiver is used.
According to the IEEE Std. 748-1979 [11], SA manufacturers should list,
as a minimum, performance specifications such as frequency range, frequency
span, residual FM, resolution bandwidth, input impedance, dynamic range,
spurious response, sensitivity, shape factor, noise sidebands, display modes (e.g.,
linear, log, square law), and so on.
Definitions of such SA parameters are given in [11] as follows:
• Center frequency: That frequency which corresponds to the center of a

frequency span [Hz].
• Frequency band: A continuous range of frequencies extending between
two limiting frequencies.
• Frequency linearity: The linearity of the relationship between the fre-
quency of an input signal and the displayed frequency.
• Frequency drift: Gradual shift or change in displayed frequency over a
period of time due to internal changes in the spectrum analyzer (Hz/s,
Hz/°C, and so on).
• Dynamic range: The maximum ratio of two signals simultaneously pres-
ent at the input that can be measured to a specified accuracy.
• Full span (max span): A mode of operation in which the spectrum ana-
lyzer scans an entire selected frequency band.
Table 8.3
Basic Technical Specs of the Spectrum Analyzer
Frequency accuracy ±0.2% or 5 MHz + 20% scan resolution
Long-term drift 3 kHz/10 min after the first 30 min
Noise side-bands –75 dBc for 30 × resolution drift from center frequency
Residual FM 1 kHz peak to peak for 2ms
Input sensitivity –115 dBm for 1 kHz bandwidth
3rd order intermodulation (TOIP) 70 dB below 2 full screen signal
Resolution bandwidth –6 dB points from 1 MHz to 1 kHz
• Deflection factor: The ratio of the input signal amplitude to the resul-
tant output indication. The ratio may be in terms of rms volts per divi-
sion, dB per division, watts per division, or any other specified factor.
• Resolution bandwidth: The width in Hz of the SA’s response to a CW
signal. This width is usually defined as the frequency difference at speci-
fied points on the response curve, such as the 3 dB or 6 dB down points.
The manufacturer will specify the dB down points to be used.
• Residual response: A spurious response in the absence of an input, not
including noise and zero pip.
• Residual FM: Short-term displayed frequency instability (jitter) of the
SA caused by instability of the local oscillators (given in terms of PP
frequency deviation).
• Noise sidebands: Undesired response caused by noise internal to the SA
appearing on the display around a desired response.
• Shape factor (skirt selectivity): A measure of the asymptotic shape of the
resolution bandwidth response curve of a SA. Shape factor is defined
as the ratio between bandwidths at two widely spaced points on the
response curve, such as the 3-dB and 60-dB down points.
• Spurious response (spurii, spur): A characteristic of a SA wherein the
displayed frequency does not conform to the input frequency.
A commercial SA is shown in Figure 8.2. In addition to input/output

connectors and warning lamps, there are more than two dozen function but-
tons in today’s SAs. This results in hundred-page user manuals, several pages
of which belong only to technical specs. A typical page showing frequency and
resolution bandwidth specs is presented in Figure 8.3.
Figure 8.2 A typical Rohde and Schwarz SA.

Figure 8.3 Specs of the Rohde and Schwarz SA.
Specifications describe the performance of parameters covered by the

product warranty (usually, temperature values from 0°C to 55°C, unless oth-
erwise noted). “Typical” values describe additional product performance but
do not include measurement uncertainty. “Nominal” values indicate expected
performance or describe product performance that is useful in the application
of the product, but is not covered by the product warranty.
According to this spec, for example, the frequency readout error may be
obtained as follows:
Frequency readout accuracy = ± Frequency indication X frequency

reference accuracy
= + 1% span
= + 0.5 × span/ (sweep point-1)
= + Hz
Span in this formula is the frequency range between the first and last
frequencies. Span/(sweep points-1) is defined as the horizontal resolution. Al-
though internal reference accuracy is given as “≤ 5% ppm/year (within two
years of adjustment),” it may need to be calculated, for example, from:
Frequency reference accuracy = 1.8 10 –7 /year

= 1.5 10 –7 (aging)
+ 0.1 × 10 –7 (temperature stability)
+ 0.1 × 10 –7 (setability)
+ 0.1 × 10 –7 (15 warm up)
Question: What is the error if a measurement is performed at 1 GHz with

400 kHz screen sweep and RBW = 3 kHz (assuming the screen is swept at 100
points)?
Answer: As shown from the following calculations, the frequency readout
error (i.e., the total uncertainty) is ±1532 Hz. The measurement result given
as 1 GHz may be any value in between (1 GHz ± 1532 Hz). As stated by the
producer of the EMI receiver, nothing more can be said with this EMI receiver.
(1 × 10 ) × (±1.8 × 10
9 −7
year ) = 180
400 kHz × %0.25 = 1000
3kHz × %5 = 150
2Hz + 0.5 × 400 EkHz (1000 − 1) = 202
Total uncertainty = ±1532 Hz
A typical display of a spectrum analyzer is given in Figure 8.4. Here, a

1-GHz span between 150 MHz (the first frequency) and 1.15 GHz (the last
frequency) is shown. Definitions of spectral purity, harmonic suppression ratio,
and phase noise are pictured in Figure 8.5. Figure 8.6 explains the definition of
selectivity. Selectivity is an important parameter to distinguish different ampli-
tude nearby signals. It is defined as 60 dB BW/3 dB BW.
Example: Suppose there are two equal amplitude signals 10 kHz apart
in the spectrum. They can be distinguished and displayed on the screen with
a RBW equal to or less than 10 kHz. If one of the signals is 50 dB below the
other, then it may be masked and cannot be displayed even at further frequen-
cies. If, for example, selectivity is 15:1 and RBW is 3 kHz, then 60 dB below
of the filter will be (15 × 3 kHz =) 45 kHz. If a filter with RBW = 1 kHz is
selected, then 60 dB below point will be (15 × 1 kHz =) 15 kHz, and the weak
signal can easily be detected (half of 60 dB bandwidth is 7.5 kHz). So, the RBW
≤ 1 kHz should be selected for this measurement (see Figure 8.7).
Figure 8.4 A main window of a SA.
Figure 8.5 Spectral purity of a SA.
8.5.2 EMI Receiver

EMI tests and measurements require a different approach than other types of
general tests and measurements (e.g., RF signal measurements). EMC engi-
neers never quite know what signals are in the environment. Moreover, the
EUTs they test every day will be different; therefore, having the correct tools
for characterization of the EMI signals is of key importance. Also, knowing the
strengths and weaknesses of the tools being used is vital. For any EMI testing,
both spectrum analyzers and test receivers may be used, but each requires a dif-
ferent approach to the test, and each has advantages and disadvantages.
Figure 8.6 Selectivity of a SA.
Figure 8.7 Required selectivity for the previous example.
Although both instruments measure the amplitude of signals in the fre-

quency domain, the units are not the same and require knowledge to extract
correct measurements.
EMI receivers are designed to comply with the detailed equipment re-
quirements of measurement standards for RFI, such as the civilian specification
CISPR 16-1-1 or the military specification MIL-STD 461. The EMI receiver
has defined IF bandwidths, typically 200 Hz, 9 kHz, 120 kHz, and 1 MHz
(usually, SAs do not have all of these), and standardized detector modes—peak
(PK), quasi-peak (QP), and average (AV). Usually, SAs do not have QP detec-
tors used to quantify the annoyance factor of a signal by weighing its repetition
rate in addition to its frequency and amplitude. They use a preselection for an
improved dynamic range. Figure 8.8 shows a photo of Rohde and Schwarz EMI
receiver.
Typical parameters of SAs are first and last frequencies of a sweep, resolu-
tion bandwidth of 3 dB or 6 dB, detectors, sweep time, and video bandwidth.
Although manual control is allowed, resolution and video bandwidths, and
sweep time of SAs are kept constant during a measurement. Typical parameters
of EMI receivers are first and last frequencies of a sweep, RWB of 3 dB or 6 dB,
detectors, dwell time, and frequency step.
General SAs measure either power or voltage; EMI receivers measure and
display only field strengths. EMC standards specify limit values in terms of field
strength at some distance (i.e., 1m, 3m, 10m, 30m) from the EUT. To arrive
at a measurement that is comparable to a specified limit value, one needs to be
sure that the measured value belongs to only the EUT. This is why the AF of the
EMI receive antenna also includes cables and connectors [12].
A typical EMI receiver display is shown in Figure 8.9. Here, emissions
measured in the frequency range of 30 MHz – 1 GHz with RBW = 120 kHz
are displayed as field amplitude, in dBµV/m, versus frequency. A PK detector
is used with 23.67 µs measurement at each frequency. The table at the bottom
of the display lists the PK values marked on the graphics. At the right of the
display, recorded PK, QP, and AV field values of a selected frequency are dis-
played. On the top of the display, PASS/FAIL message according to the related
CISPR is also given. As observed, this EUT fails to comply with the related
CISPR standard.
Note that it is important to be an experienced user in EMC test and
measurements. Measurement devices from simple DMM to EMI receivers have
many parameters that are critical in any test and measurement [3, 12, 13].
Figure 8.8 A typical Rohde and Schwarz EMI receiver.

Figure 8.9 A typical display of an EMI receiver.
The CISPR 16-1-1 standard [14] gives the requirements such as detector
type, resolution bandwidth, selectivity, harmonic suppression, input imped-
ance, dynamic range, noise threshold, and so on, of voltage, current, and EM
field measuring devices in the frequency range of 9 kHz–18 GHz. Commercial
EMI receivers fully comply with CISPR 16-1-X standard family.
8.6 Other Test/Measurement Instruments

In addition to all these measurement instruments, current probes/clamps, cur-
rent-injection clamps, and line impedance stabilization networks (LISN) are
widely used in EMC tests and measurements.
8.6.1 Current Probes and Current-Injection Clamps

A current probe or a current clamp probe is an electrical device measuring the
electric current in a conductor without having to make physical contact with
it or to disconnect it for insertion through the probe. They are usually used to
read the magnitude of an AC current up to 1 kA and more.
A current-injection probe does the opposite of the current probe; it in-
jects RF current into conductors and cables. It provides a way of applying a
controlled RF stress level to an EUT/DUT through interconnecting cables or

power cables without requiring a direct connection to them.
Figure 8.10 presents Ampere’s law and a commercial current probe and a
commercial current-injection clamp. As illustrated in Figure 8.10(a), magnetic
field is induced around a current carrying cable, and the direction of this field
may be found by the right-hand rule (if the thumb shows the direction of the
current, then the other four fingers enclosing the cable show the direction of
the magnetic field). This forms the basis of both current probe and current-
injection clamp. If you capture the magnetic field via a ferromagnetic element
around the cable, then you control/measure the flowing current. In addition, if
you generate a magnetic field around this cable, you inject a current along the
wire. These both can be done without cutting the cable.
8.6.2 Line Impedance Stabilization Network

A line impedance stabilization network (LISN), also called an artificial mains
network (AMN), is used to measure disturbance voltages on the mains cord
of an electrical instrument (ANSI C63.4 uses the term LISN, and CISPR 22
uses AMN). It is mainly used for CE testing and is inserted between the power
mains and the EUT. As a passive circuit, LISN serves to isolate the test system
with a reference impedance and provide a measurement point to monitor the
CE. The measurement bandwidth is in the 9 kHz to 30 MHz region. A spec-
trum analyzer or RF noise meter is used to monitor the CE of the EUT via an
RF (BNC) connector to a (test receiver) port on the LISN. Figure 8.11 show a
typical LISN circuit used for the frequency range of 10 kHz to 150 kHz (which
may also be used up to 30 MHz if constructed carefully). A variety of LISNs are
available for different applications and standard requirements.
Figure 8.10 (a) Ampere’s law, (b) fluke current clamp, (c) Teseq current-injection clamp.
Figure 8.11 A typical LISN circuit.
8.7 Network Analyzer

Although they are not directly used in EMC test and measurements, network
analyzers are also critical in many applications in EMC engineering. Network
analyzers (NAs) are used to measure components, devices, circuits, and subas-
semblies under known source and receiver, and generally display amplitude and
phase information (frequency or power sweeps). Network analyzers are often
used to characterize two-port networks such as amplifiers and filters, but they
can be used on networks with an arbitrary number of ports. A NA always looks
at known signals, in terms of frequency, since it is a stimulus response system.
It measures input/output port responses such as IL and/or RF versus frequency
(i.e., S-parameters), input/output impedances versus frequency on the Smith
chart, and so on. A typical Rohde & Schwarz NA is shown in Figure 8.12.
The two basic types of network analyzers are scalar network analyz-
er (SNA), which measures amplitude properties only, and vector network ana-
lyzer (VNA), which measures both amplitude and phase properties.
Figure 8.12 A Rohde and Schwarz network analyzer.

References
[1] Witte, R. A., Electronic Test Instruments: Theory and Applications, Englewood Cliffs, NJ:
Prentice Hall, 1993.
[2] Dunn, P. F., Measurement and Data Analysis for Engineering and Science, New York: Mc-
Graw-Hill, 2005.
[3] Sevgi, L., “Innumeracy: The Meaning of the Numbers We Use,” IEEE Antennas and Prop-
agation Magazine, Vol. 49, No. 2, Apr. 2007, pp.195–190.
[4] National Instruments Tutorial, “Digital Multimeter Measurement Fundamentals,” Feb.
2006, http://zone.ni.com/devzone.
[5] Rec.P–372 “Radio Noise,” ITU (formerly CCIR)-Recs., 2003, or Rec. 322-3, “Character-
istics and Applications of� Atmospheric Radio Noise Data,” CCIR Recs., 1988.
[6] Agilent Technologies Application Note, “System Cabling Errors and DC Voltage Mea-
surement Errors in Digital Multimeters,” AN-1389-1, Jan. 2005, www.agilent.com.
[7] Agilent Technologies 3458 Multimeter Data Sheet, 5965-4971E, 2005.
[8] Fluke Application Note “Understanding Specifications for Precision Multimeters,”
2547797 A-EN-N, Rev A, 2006, www.fluke.com/library.
[9] National Instruments Tutorial, “Digital Multimeter Measurement Cycle,” Feb. 2006,
http://zone.ni.com/devzone.
[10] Cawley, Kevin, “Choosing DMMs and More for High-Performance Applications,”
Keithley Instruments, Inc., No 25260604, Jun. 2004, www.keithley.com.
[11] IEEE Std 748-1979: IEEE Standard for Spectrum Analyzers.
[12] L. Sevgi, S. Çakır, and G. Çakır, “Antenna Calibration for EMC Tests and Measurements,”
[13] Sevgi, L., “Digital Multimeters and Basic Measurements,” IEEE Antennas and Propagation
Magazine, Vol. 49, No. 4, Aug. 2007, pp. 232–237.
[14] CISPR 16-1-4: 2003, Radio Disturbance and Immunity Measuring Apparatus—
Measuring Apparatus.
9
EMC Tests and Measurements
EMC requires that systems/equipment must be able to tolerate a specified de-
gree of interference and not generate more than a specified amount of interfer-
ence. These are monitored via EMC tests and measurements. EMC tests and
measurements are performed to certify that a product complies with all the
requirements specified in related standards. A product has to be designed and
fabricated accordingly; it must not cause unintentional, undesired emissions
that may affect or be susceptible to other products in the environment.
The three major components of EMC tests and measurements of EUT
are (1) finding out related standards; (2) knowing operational status during tests
and measurements; and (3) understanding the methods, measurement devices,
and procedures of tests and measurements. EMC tests and measurements must
be reliable, realizable, and repeatable. These are the issues that are controlled via
traceability. Traceability is the possibility of a test or measurement result avail-
able for a comparison within a chain of national/international references within
a specified uncertainty.
Figure 9.1 shows main EMC tests and measurements. As shown in the
figure, emission is measured but susceptibility (immunity) is testes.
Emission measurements are divided into two parts: conducted emissions
(CE) and radiated emissions (RE). CE measurements are performed over EUT
cables up to 30 MHz frequencies, and the voltage values are recorded in dBµV.
A line impedance stabilization network is used during all measurements to sta-
bilize impedances. CE measurements are performed inside a screened room. RE
measurements, performed either in anechoic chamber or on OATS, are usually
requested between the 30-MHz and (up to) 6-GHz frequency range. RE val-
ues are given in terms of electric field in dBµV/m. In addition to CE and RE
measurement, harmonics, flicker, and clicks are also measured. Immunity tests
are also performed conducted and radiated: conducted immunity (CI) below
223
Figure 9.1 Fundamental EMC tests and measurements.
30 MHz and radiated immunity (RE) above 30 MHz. In CI tests, high cur-
rents are injected through cables and EUT’s performance is observed. RI tests
are performed under high homogeneous electric and magnetic fields described
in related standards. Immunity tests also include against ESD, EFT, and surge.
EMC emission and immunity standards are prepared by internat‑ional
expert institutions. These standards specify conditions to be satisfied by equip-
ment that operates in an environment including other equipment. EMC tests
and measurements are complex and time consuming, but not complicated. Ev-
ery detail of the test and measurement procedures are given in related standards.
Table 9.1 lists mandatory EU EMC standards. Some of the fundamental
CISPR standards referred in EU standards are given in Table 9.2.
The CISPR standards given in Table 9.2 are about test and measurement
methods, environments and devices [1–15]. The EU standards in Table 9.1,
on the other hand, present EMC tests and measurements for various products
in our lives [16–29]. For example, the EU standards for conducted emissions
EN55022, EN55011, EN550141, EN61000−6−3, and EN61000−6−4 refer
to CISPR 14 and CISPR 16 as reference standards (see, for example, EMC
standards for IT products in Figure 9.2.)
The starting point of EMC tests and measurements is to identify related
standards.
9.1 Risk Analysis in EMC Tests and Measurements

The progress in technology, new designs, new equipment, and novel applica-
tions bring two types of events in our lives: negative events are called risks;
EMC Tests and Measurements 225
Table 9.1
EU Standards for Fundamental EMC Tests and Measurements
Conducted EN55022, EN55011, EN55014−1, EN61000−6−3, EN61000−6−4
Emission
EN55022, EN55011, EN61000−6−3, EN61000−6−4, 95/54/EC,
Radiated Emission 72/245/EEC, 75/322/EEC
Power Parasitics EN55014−1
Click EN 55014−1
EN61000−4−3, EN55024, EN61000−6−1, EN61000−6−2,
Radiated Immunity EN55014−2, 95/54/EC, 72/245/EEC, 75/322/EEC
Conducted EN61000−4−6, EN55024, EN55014−2, EN61000−6−1,
Immunity EN61000−6−2
EN61000−4−4, EN55024, EN55014−2, EN61000−6−1,
EFT / Burst EN61000−6−2
EN61000−4−5 , EN55024, EN55014−2, EN61000−6−1,
Surge EN61000−6−2
EN61000−4−2 , EN55024, EN55014−2, EN61000−6−1,
ESD EN61000−6−2
Magnetic Field EN61000−4−8 , EN55024, EN55014−2, EN61000−6−1,
Immunity EN61000−6−2
Voltage outage EN61000−4−11 , EN55024, EN55014−2, EN61000−6−1,
and dip EN61000−6−2
Harmonic Currents EN61000−3−2
Harmonic Immunity EN61000−4−13
Flicker EN61000−3−3
Table 9.2
Basic CISPR EMC Test and Measurement Standards
CISPR 11 EMC thresholds and measurement methods—ISM RF devices
CISPR 14-1 EMC requirements—emission (Household appl, electrical equipment)
CISPR 14-2 EMC requirements—immunity (Household appl, electrical equipment)
CISPR 16-1 EMC requirements 1—emission and immunity measurement setups
CISPR 16-2 EMC requirements 2—emission and immunity measurement methods
CISPR 16-3 EMC requirements 3—reporting and CISPR recommendations
CISPR 16-4 EMC requirements 4—uncertainty and statistical evaluations
CISPR 24 Information technology devices—immunity, thresholds, and test methods
positive events are called opportunities. Risks may come from various sources
including uncertainty, threats at any phase in design, development, production,
legal liabilities, accidents, natural causes and disasters, deliberate attacks, and
Figure 9.2 IEC and CENELEC standards for IT products.
so on. Living with those risks is called risk perception and risk management
(see Section 1.2.5). Risk management is the identification, assessment, and pri-
oritization of risks (defined in [30] as the effect of uncertainty on objectives)
followed by coordinated and economical application of resources to minimize,
monitor, and control the probability and/or impact of unfortunate events or to
maximize the realization of opportunities.
Risk is never zero. It is expressed with a real number between zero and one
if calculated using a kind of risk model. If not, risk is expressed as “high,” “low,”
“negligibly low,” “acceptable,” and so on. According to the definition of risk, a
risk is the possibility that an event will occur and adversely affect the achieve-
ment of an objective. Therefore, risk itself has uncertainty. Risk management
aims to control the risk.
An important issue in EMC tests and measurements is the risk analysis.
The degrees of risks in EMC tests and measurements are listed in Table 9.3.
Unless EMC tests and measurements are repeated for every product, there
will always be a risk. The important issue is to minimize this. Depending on
the size of the company and type of the product, it may not be economic to es-
tablish an in-house EMC lab. But it may also not be economic to continuously
send a sample product to a commercial lab. As in all engineering problems,
here, an optimization is necessary.
9.2 Emission Measurements

CE and RE measurements shall be performed below and above 30 MHz, re-
spectively. CE measurements are performed in a screened room; RE measure-
ments are done in anechoic chambers and/or OATSs.
Table 9.3
Risks in EMC Tests and Measurements
Least risk Establish a complete EMC test center and test every product
Low risk Establish a complete EMC test center and test by sampling
Medium risk (1) Establish a complete EMC test center and test one product
Medium risk (2) Establish a pre-compliance EMC test center and test a product
High risk Do not test—pray!
9.2.1 Detector Types

EMI receivers have peak (PK), quasi-peak (QP) and average (AV) detectors. PK
detector records maximum value of a signal in a certain measurement period.
An AV detector calculates the average value of number of measurements made
in one measurement period. QP detectors weigh signals according to their rep-
etition rate, which is a way of measuring their annoyance factor. The following
remarks are worth mentioning:
• QP detector output signals depend on the amplitude and frequency of

the signal.
• QP detector readings will always be less than or equal to the PK detec-
tion.
• QP readings are much slower (a few orders of magnitude) than PK read-
ings.
• It is suggested to scan initially with the PK detection first, and then if
this is marginal or fails, to switch and run the QP measurement against
the limits.
• QP and PK detector outputs are the same for a sinusoidal input signal.
• The difference between QP and PK detector outputs are significant for
low-repetition pulsed signals.
9.2.2 Basic Factors in EMC Tests and Measurements

Basic EMC tests and measurements factors are as follows:
• Control components: Effects of physical environment on the EMC tests

and measurements;
• Coupling components: Undesired coupling of receive antenna, its ca-
bles, and connectors with the EUT;
• EMI receiver: Calibration and use of EMC tests and measurements

setup.
Contributions of all these components must be well understood and in-

vestigated. For example, a screened room has resonances, and, depending on
the conditions, the recorded emissions may be a few orders lower and/or higher
than the actual value. Similarly, the EUT should be well isolated from both AC
mains and the receive antenna. AC power supply may be a terrible interference
source. It is mandatory to show that whatever recorded during the measure-
ments belongs solely to the EUT. This is possible only after calibration of the
test and measurement environment, equipment, and personnel.
Important steps for EMC tests and measurements are finding the right
standard, calibration of test/measurement environments and equipment (e.g.,
EMI receiver + antennas.), understanding boundaries of the test/measurement
method, accreditation of the lab and personnel, error/uncertainty analysis, and
finally reporting.
9.2.3 Performing Emission Measurements

CE and RE measurements are performed in the frequency ranges 150 kHz–30
MHz and 30 MHz–6 GHz, respectively. The boundary between these two is
30 MHz. Standards define two different RBWs for emissions measurements in
the frequency range of 150 kHz to 6 GHz. These bandwidths and measurement
times are given in Table 9.4. As seen there, 9 kHz RBW is used during the CE
measurements, but RBW is 120 kHz for RE measurements (1 MHz RBW is
used above 1 GHz). Measurement times with PK and QP detectors are also
given. CE measurements with PK detector takes only 2 minutes, but it takes
100 minutes with QP detector. Because of this, standards allow fast-scan with
PK detector and then additional scans with QP detector around critical fre-
quencies. If PK detector values are 6–10 dB below the limit values, then there
is no need to repeat the measurements with QP detectors.
Values in Table 9.4 also clarify narrowband/wideband signal definitions.
A signal with bandwidth less than 9 kHz is assumed narrow band for CE
Table 9.4
Bandwidths and Measurement Times for Emission Measurements
Frequency Band Step # of Meas Meas
[MHz] [kHz] [kHz] steps time (PK) time (QP)
0.15–30 9 5 5970 2 dak 100 dak
30–1000 120 50 19400 6.5 dak 323 dak
measurements (i.e., in the frequency range of 150 kHz–30 MHz). On the other
hand, signals having bandwidths more than 120 kHz are assumed broadband
in RE measurements up to 1 MHz. Above this, signals with more than 1 MHz
bandwidth are broadband.
CE measurements are performed between 150 kHz–30 MHz. Mandatory
standards are EN55016-2-1, EN55022, EN55011, and EN55014-1. Results
are given as voltage in dBµV versus frequency. Both the EUT and EMI receiv-
ers are connected via a LISN. AC power supply is also connected to LISN. The
three mission of the LISN are (1) to filter out AC mains and supply the EUT
clean enery; (2) to direct EUT-generated emissions to reach only to the EMI
receiver; (3) to prevent other reflections by impedance matching between AC
mains and the EUT in the frequency range of 150 kHz–30 MHz. Table 9.5 lists
equipment necessary for CE measurements.
To satisfy traceability, CE measurements are performed under the follow-
ing conditions specified in the standards:
• A vertical/horizontal ground plane shall extend at least 0.5m beyond the
projection of the test arrangement, but shall nevertheless have a mini-
mum size of 2m × 2m.
• Bottom and rear of the EUT shall be at a controlled distance of 40 cm
from the ground plane. This ground plane is normally the wall or floor
of a shielded room.
• The EUT is placed on an 80-cm-high nonconducting table. Floor-
standing EUTs are subject to the same provisions as tabletop equipment
with the exception that they should be placed on a floor, with the points
of contact being consistent with normal use.
• Interconnecting cables that hang closer than 40 cm to the ground plane
shall be folded back and forth forming a bundle 40 cm long or less,
Table 9.5
Equipment Necessary for CE Measurements
Equipment Specification
EMI receiver CISPR 16-1-1 compliant
LISN 9 KHz-30 MHz, CISPR 16-1-2
Coaxial cable —
Reference ground plane At least 2m by 2 m.
Shielded enclosure (optional) Depends on the EUT dimension
RF mains filters (optional) Broadband RF filter
Voltage probe (CISPR 16-1-2) 9 KHz–30 MHz, high impedance
hanging approximately in the middle between the ground plane and the
table. The minimum bend radius of the cable should not be exceeded.
• I/O cables that are connected to a peripheral should be bundled in
the center. The end of the cable may be terminated if required using
correct terminating impedance. The total length should not exceed 1m,
if possible.
A typical CE measurement result is shown in Figure 9.3 [31]. Here, a scan

results with 9 kHz RBW in the frequency range of 150 kHz–30 MHz, com-
pleted in 382 µs, is shown. As stated on the display, this EUT fails to comply
with the CISPR standard.
RE measurements shall be made in an anechoic chamber or OATS with a
QP detector in the frequency range 30 MHz and (up to) 6 GHz. Related stan-
dards are EN55016-2-3, EN55022, EN55011, and EN55014-1. Results are
given as field strength in dBµV/m versus frequency. A typical RE measurement
setup is pictured in Figure 9.4. To reduce the testing time, a PK detector may
be used instead of a QP detector. In case of dispute, measurement with a QP
detector will take precedence.
The followings are specifically mentioned in the standards:
• The receive antenna shall be a balanced dipole.
Figure 9.3 A typical CE measurement result.

Figure 9.4 A typical RE measurement setup.
• For frequencies 80 MHz or above, the antenna shall be resonant in

length, but below 80 MHz it shall have a length equal to the 80-MHz
resonant length.
• The antenna-EUT distance shall be 10m. If this cannot be satisfied be-
cause of high ambient noise or some other reasons, a closer distance
(e.g., 3m), can be selected.
• The antenna shall be adjusted between 1m and 4m, for both vertical
and horizontal polarizations, in height above the ground plane for maxi-
mum meter reading at each test frequency.
• A measurement site shall be considered acceptable if the horizontal and
vertical site attenuation measurements are within ±4 dB of the theoreti-
cal NSA of an ideal site.
• The test site shall be flat; free of overhead wires and nearby reflecting
structures; sufficiently large to permit antenna placing at the specified
distance; and provide adequate separation between antenna, EUT, and
reflecting structures.
• A conducting ground plane shall extend at least 1m beyond the periph-
ery of the EUT and the largest measuring antenna, and cover the entire
area between the EUT and the antenna. It should be of metal with no
holes or gaps having dimensions larger than one-tenth of a wavelength
at the highest frequency of measurement.
• AC mains cables shall drape to the ground reference plane. They shall
then be routed to the mains power outlet. AC mains power outlet shall
be bonded to, and should not protrude above, the ground reference
plane.
A typical RE measurement result is given in Figure 9.5 [31]. Here, a scan

results with 120 kHz RBW in the frequency range of 30 MHz–300 MHz,
completed in less than 24 µs, is shown. As stated on the display this EUT fails
to comply with the CISPR standard.
9.3 Immunity/Susceptibility Tests

Immunity tests are also divided into two: conducted immunity (CI) below 80
MHz (according to EN61000-4-6) and radiated immunity (RI) above 80 MHz
(EN61000-4-3). In addition, immunity against EDS (EN61000-4-2), EFT/
burst (EN61000-4-4), surge (EN61000-4-5), magnetic fields (EN61000-4-8),
and voltage dips (EN61000-4-11) are also mandatory.
Immunity tests are difficult to perform, may be dangerous, and neces-
sitate expensive test setups. These tests are performed under high (but homoge-
neous) electric and magnetic fields, high-intensity ESD pulses, EFT/bursts, and
so on, which may be harmful for EMC test personnel. Immunity test evalua-
tion is also difficult; therefore, standards define different performance criteria
(see Section 2.5.4).
Figure 9.5 Typical RE measurement results.

Conducted immunity tests are performed according to the EN61000-4-6

standard between 150 kHz–80 MHz frequencies. During this test, the immu-
nity of the equipment for induced or conducted EM fields is checked under
the typical operating conditions of the EUT. The EUT should be tested within
its intended operating and climatic conditions. The temperature and relative
humidity should be recorded in the test report. The test should be performed
with the test generator connected to each of the coupling devices (coupling/
decoupling network-CDN, EM clamp, and so on) in turn. The main CDN
parameter is the CM impedance seen at the EUT-port.
The following are important according to EN61000-4-6 standard:
• The equipment to be tested is placed on an insulating support of 10 cm

height above a ground reference plane.
• All cables exiting the EUT shall be supported at a height of at least 30
mm above the ground reference plane.
• All other cables not under test, shall either be disconnected (when func-
tionally allowed) or provided with CDNs.
• lf the equipment is designed to be mounted in a panel, rack, or cabinet,
then it shall be tested in this configuration.
• CDN shall be located between 10 cm and 30 cm from the EUT. This
distance is to be measured horizontally from the projection of the EUT
on to the ground reference plane to the CDN device.
• CDN shall be used for appropriate coupling of the disturbing signal to
the various cables connected to the EUT and for preventing applied test
signals from affecting other devices and systems that are not under test.
• The frequency range is swept from 150 kHz to 80 (230) MHz, using
the signal levels established during calibration process and with the dis-
turbance signal 80% AM with a 1 kHz sine wave.
• The step size shall not exceed 1% of the preceding frequency value.
• The dwell time of the amplitude modulated carrier at each frequency
shall not be less than the time necessary for the EYT to be exercised and
to respond, but shall in no case be less than 0.5s.
• In any one test, only two 150Ω networks are required. The network
used for injection of the test signal can be moved between different ports
as they are tested.
• If the EUT has multiple identical ports (same I/O electronic circuit,
loads connected to the EUT), at least one of these ports shall be tested.
• If the EUT has only one port, that port is connected to the CDN used
for injection.
Radiated immunity tests are performed with the typical RI test setup given
in EN 61000−4−3 standard (see Figure 9.6). For most of the commercial prod-
ucts, RI tests are mandatory from 80 MHz up to 1 GHz frequency. Table 9.6
lists necessary test equipment. Test levels for commercial products are listed in
Table 9.7 (levels of special and/or military products may be higher sometimes
mandatory by the standards, very often is requested by the producer).
The vertical area where homogeneous high-level electric fields will be gen-
erated is shown in Figure 9.6. This is called calibration plane or uniform field
area (UFA). The calibration plane should be larger than the EUT. Immunity
tests can also be performed by shifting and relocating the EUT if it is larger than
UFA. The critical issue in RI tests is the generation of this field. This should
be verified before every RE test. The generated fields is measured at 16 points
(on a 4 × 4 grid). The four anomalous values are left out, and the field values
on the other 12 points are averaged. According to EN61000-4-3, this average
must be within ± 6 dB of the requested level. If this is satisfied, then RI tests
will be performed.
The following are described in the EN61000-4-3 standard:
• During the test, it should be verified that the EUT has sufficient im-
munity against radiated EM fields. All testing of equipment shall be
Figure 9.6 A typical RI test setup.

Table 9.6
RI Test Equipment
Equipment Specification
Signal generator 80–1000/2700 MHz or higher
Power amplifier Broadband, 80–1000 MHz
Power amplifier Broadband, 1000–2700 MHz
Power meter 80–2700 MHz or higher
Power sensor(s) –60 dBm +20 dBm
Directional coupler(s) 40/60 dB
Function generator (for PM) —
Broadband antenna(s), bilog 80–1000 MHz or higher
Double-ridged/horn antenna 1000–2700 MHz or higher
Anechoic chamber Dimension depends on the EUT.
Isotropic EM probe 80–2700 MHz or higher
RF immunity software —
Coaxial cable(s) —
Computer —
Table 9.7
RI Test Levels Given in EN
6100-4-3 Standard
Electric Field
Test Level Level
1 1 V/m
2 3 V/m
3 10 V/m
4 30 V/m
performed in a configuration as close as possible to actual installation

conditions.
• Wiring shall be consistent with the manufacturer’s recommended pro-
cedures, and the equipment shall be in its housing with all covers and
access panels in place unless otherwise stated.
• If the equipment is designed to be mounted in a panel, rack, or cabinet,
it shall be tested in this configuration. Grounding of housing or case
of the equipment shall be consistent with the manufacturer’s installa-
tion recommendations. When an EUT consists of floor-standing and
tabletop components, the correct relative positions shall be maintained.
• The equipment to be tested is placed in the test facility on a nonconduc-
tive table 0.8m high. The equipment is then connected to power and
signal wires according to relevant installation instructions.
• Floor-standing equipment should be mounted on a nonconductive sup-

port 5 cm to 15 cm above the supporting plane. The use of nonconduc-
tive supports prevents accidental Earthing of the EUT and distortion of
the field. To ensure the latter, the support shall be bulk nonconducting,
rather than an insulating coating on a metallic structure.
• Floor-standing equipment may be put on the table if it is not too large
or heavy, or where its elevation would not create a safety hazard, but this
must be stated in the test report.
• The equipment is then connected to power and signal wires accord-
ing to relevant installation instructions. If possible, a minimum of 1m
of cable is exposed to the electromagnetic field. Excess length of cables
interconnecting units of the EUT shall be bundled low-inductively in
the approximate center of the cable to form a bundle 30 cm to 40 cm
in length.
After setup is completed, the immunity tests are performed as follows:
• The EUT is initially placed with one face coincident with the calibra-
tion plane.
• The EUT face being illuminated shall be contained within the UFA.
• The dwell time of the amplitude modulated carrier at each frequency
shall not be less than the time necessary for the EUT to be exercised and
to respond, but shall in no case be less than 0.5s.
• The test shall normally be performed with the generating antenna facing
each side of the EUT.
• When equipment can be used in different orientations, all sides shall be
exposed to the field during the test.
• The test shall be performed for both horizontal and vertical antenna
polarization for each side of the EUT.
9.4 Harmonic Measurements

The legislative requirement to measure harmonics on equipment rated up to
16A rms is EN61000-3-2 standard. It defines limits for levels of harmonic cur-
rent in equipment rated up to 75A. The test methods are basically identical;
only different limits have been specified. Both standards aim to protect AC
mains from the equipment powered by the mains. EN61000-3-2 is general
and valid for all AC mains (50−60 Hz 220/380V, 230/400V, and 240/415V).
EN61000-3-2 standard makes harmonic current measurements mandatory up
to 40th harmonics. If the amplitudes of the harmonics monotonically decrease,
then measurements may be stopped at the 19th harmonics.
9.5 Surge and Flicker Tests

Surge is a transient over voltage on AC power system and is produced by events
such as load switching, capacitor bank switching, equipment faults, and light-
ning discharges. They are usually of short duration, from several microseconds
to a few milliseconds. Surge waveforms can be oscillatory or impulsive with the
rising wavefront usually on the order of 0.5 µs to 10 µs. The highest surges are
caused by direct lightning strikes to overhead power lines. Surges entering a fa-
cility typically range from 10 kV to 50 kV. Surge voltage and current attenuate
as they propagate.
EN61000-4-5 is the EMC standard for surge immunity tests. Voltage
surge testing is typically done with commercially available surge test equip-
ment. The test equipment typically has a surge generator and a CDN. Com-
bination wave generators produce a specified open circuit voltage waveform
and a specified SC current waveform. CDNs couple a surge generator to the
equipment being tested and prevent dangerous voltages from being sent back
into the ac power system.
Even though power surges are so brief that they are measured in nano-
seconds, they can cause considerable damage to electronic equipment; there-
fore, surge protectors are widely used. A surge protector is a device that shields
electrical and electronic devices from surges. It attempts to limit the voltage
supplied to an electric device by either blocking or shorting to ground any un-
wanted voltages above a safe threshold. IEC/EN 61643 is the standard of low
voltage surge protector devices.
Flicker is determined by the amplitude and frequency of AC power line
voltage fluctuations. Human (eye) sensitivity is maximum around 8.8-Hz fre-
quency (1058 fluctuations per minute). EN61000-3-3 standard is concerned
with the limitation of voltage fluctuations and flicker impressed on the public
low-voltage system. It specifies limits of voltage changes that may be produced
by equipment tested under specified conditions and gives guidance on methods
of assessment. It is applicable to electrical and electronic equipment having an
input current equal to or less than 16A per phase, intended to be connected
to public low-voltage distribution systems of between 220V and 250V line to
neutral at 50 Hz and not subject to conditional connection. The standard con-
tains normative references, definitions, assessment of voltage changes, voltage
fluctuations and flicker, limits, and test conditions. Flicker meters are used in
flicker tests.
9.6 Electrostatic Discharge Tests

Electrostatic discharge (ESD) is the sudden and momentary electric current
that flows between two objects at different electrical potentials. The term is
usually used in the electronics and other industries to describe momentary un-
wanted currents that may cause damage to electronic equipment. ESD is a seri-
ous issue in solid state electronics, such as integrated circuits. Integrated circuits
are made from semiconductor materials such as silicon and insulating materials
such as silicon dioxide. Either of these materials can suffer permanent damage
when subjected to high voltages; as a result, there are now a number of antistatic
devices that help prevent static build up. A static event starts with a very slow
buildup of energy (seconds or minutes) that is stored in the capacitance of a
structure (e.g., a human body, furniture, or unconnected cable). This charge is
followed by a very rapid breakdown (typically nanoseconds). With this pulse in
the nanosecond range, the discharged energy can produce EMI in the frequency
range of hundreds of megahertz to beyond 1 GHz. Figure 9.7 shows typical
ESD discharges through a human finger, hand, and handheld device.
A static event from a human can exhibit rise times ranging from approxi-
mately 200 ps to greater than 10 ns with peak impulse currents from a few
amperes to greater than 30A. Because of its high-speed, high-frequency spectral
distribution, ESD energy can damage circuits, bounce grounds, and even cause
upsets through EM coupling up to 200-MHz frequencies.
Equivalent of a typical ESD pulse can be generated via 200-pF capacitor
and a 20-kV voltage source suddenly discharged over a 500-ohm load resistor
Figure 9.7 Typical human ESD levels.

[8]. ESD immunity tests are performed with ESD simulators/guns manufac-
tured with this model.
ESD pulses (between 4 kV and 8 kV) during ESD immunity tests, as
described in details in EN61000-4-2 standards, are applied directly (through
contacts) and indirectly (over air). The standards describe everything including
how an ESD gun is held, at what distance, how many consecutive pulses will
be applied, and so on. For the sake of repeatability (traceability), ESD immu-
nity tests are performed in a screened room on a conducting/metal plate. The
conducting plate must be, at all sides, 50 cm wider than the size of the EUT.
ESD pulses are applied only to those points and surfaces of the EUT that are
expected to be touched during usual operation, including user access, as speci-
fied in the user manual (e.g., cleaning or adding consumables when the EUT is
powered). The EUT is exposed to at least 200 pulses, 100 each at negative and
positive polarity, at a minimum of four test points. For tabletop equipment,
one of the test points should be the center-front edge of the horizontal coupling
plane, which should be subjected to at least 50 indirect discharges (25 of each
polarity). All other test points should each receive at least 50 direct contact dis-
charges (25 of each polarity). All areas normally touched by the user should be
tested. A photo of a typical ESD immunity test is given in Figure 9.8.
ESD occurs because of potential differences; therefore, the main ESD
protection approach is to ground or bring all elements to the same potential.
ANSI ESD S20.20 and EN100-015/1-1991 standards cover the requirements
for designing, establishing, implementing, and maintaining ESD protection.
Figure 9.8 A typical ESD test (Esim Test Center).

9.7 Electrical Fast Transients

Circuit elements like switches, natural events such as lightning, and sudden
cable OC/SC faults may cause instant and high-level voltages, disturbing sig-
nals called electrical fast transients (EFTs). Burst or EFT represents transients
generated from circuit breakers, nonprotected relays, and so on. EN61000-4-4
standard defines an idealized burst with relatively long duration (15 ms) re-
peated at a rate of a few kHz. The pulse amplitude of the transients during the
burst is usually 0.5–4 kV peak, but since the duration of each transient is fairly
short, the bursts have fairly low energy. The pulse rise time is typically 5 ns, and
the pulse width 50 ns. They are generated with a burst generator and injected
into cables through a coupling network or a capacitive clamp. The bursts are
injected into both power lines (both CM and DM and I/O lines, if longer than
3m).
Note that if the EUT contains several ports with the same particular inter-
face, only one shall be tested. Multiconductor cables, such as a 50-pair telecom-
munication cable, shall be tested as a single cable. Cables should not be split or
divided into groups of conductors for this test. The standard is applicable only
to cables that according to the manufacturer’s specification support communi-
cation on cable lengths greater than 3m. The cable length between the EUT
and the coupling device should be as short as possible in the range 0.5m to 3m.
9.8 Measurements of Spurious

As shown in Figure 9.9, all signals like harmonics, oscillations, internal prod-
ucts, and so on, except EM carriers are called parasitic/spurious. An important
issue in EMC tests and measurements is the measurement of spurious.
Spurious measurements are different than EMI measurements. First, the
measurement bandwidth is adjusted at the receiver (on the other hand, they
are prefixed to 200 Hz, 9 kHz and 120 kHz in EMI receivers). The resolution
bandwidth of EMI receivers is defined at 6 dB of the center frequency, but
this is 3 dB in spurious measurements. Moreover, QP detector is used in EMI
receivers; PK detector is used in parasitic measurements. Because of all these,
spurious measurements are done with SA.
An important standard for spurious measurements is ITU-R SM.328
[31].
9.9 Error Analysis and Uncertainty

If you want to be sure, never measure twice! This is a phrase well known by
EMC engineers. The value obtained after multiple measurements will show a
Figure 9.9 Parasitic signals (1: useful signal, 2: out-of-band signals, 3: parasitics).
distribution. In general, the result of a measurement is only an approximation

or estimate of the value of the specific quantity subject to measurement (i.e.,
the measurand), and thus the result is complete only when accompanied by
a quantitative statement of its uncertainty. The uncertainty of the result of a
measurement generally consists of several components that may be grouped
into two categories according to the method used to estimate their numerical
values: those evaluated by statistical methods (random), and those evaluated by
other means (systematic).
Several factors affecting the measurements but may be grouped into three
as follows:
• Environmental uncertainties (reflections, interfering signals, wind, tem-

perature, humidity, etc.);
• Test and measurement device/setup uncertainties (EMI errors, antenna
positioning and polarization errors, cable and connector losses, etc.);
• Operator uncertainties (reading, positioning errors, etc.).
As a result, there is always a difference between the real/true value and

measured value. It is called error. Error may be systematic or random. System-
atic error is an error that plagues experiments or calculations caused by negative
factors. For example, a DC voltage component, which unintentionally is pres-
ent (e.g., because of a failure on the blockage capacitor during an AC voltage
measurement) is a systematic error. Another example would be an ammeter
that displays only 85% of the true voltage because of calibration problems.
Systematic errors can be complex, but can be removed once understood or
discovered via careful controls and calibration. Random error is an error that is
always present, but varies unpredictably in size and direction. They are related
to scatter in the data obtained under fixed conditions, which determines the
repeatability (precision) of the measurement. Random errors (fortunately) fol-
low well-behaved statistical rules. Random error can be reduced by repeating
the measurement as often as possible. Error can be given in one of two ways.
Absolute error is an error that is expressed in physical units. It is the absolute
value of the difference between the measured value and the true value (or the
average value if the true value is not known) of a quantity. Relative error is an
error expressed as a fraction of the absolute error to the true (or average) value
of a quantity. It is always given as a percentage. Uncertainty is a range that is
likely to contain the true value of a quantity being measured or calculated. Un-
certainty can be expressed in absolute or relative terms.
Note that the difference between error and uncertainty [30] should always
be kept in mind. For example, the result of a measurement after correction can
unknowably be very close to the unknown value of the measurand, and thus
have negligible error, even though it may have a large uncertainty.
Error and uncertainty analysis of EMC tests and measurements are done
statistically. Two important parameters in statistics are the average value and
standard deviation. The average value of a measurement performed N times is
given as:
1 x1 + x 2 + x 3 + x 4 + ... + x N
∑ k =1
N
x av = xk = (9.1)
N N
Very often, this is assumed to be the true value of the measurand. The
standard deviation (σ) (i.e., absolute error –∆x) of this measurement repeated
N times will be:
∑ k =1 (x k − x av )
N 2
(9.2)
σ = ∆x =
N −1
At the end, the result of these measurements is given as xav ± ∆x.

Propagation error is the error in the succeeding steps of a measurement due
to an occurrence of an earlier error. Consider a result of a measurement given as
A ± ∆a. Here, A is the value, ∆a > 0 is the absolute error, and ∆a/A is the rela-
tive error. If a quantity is expressed in terms of the addition or subtraction of
two other quantities, either measured or computed, then the total (propagated)
error is equal to the addition of absolute errors:
C = ( A ± B ) ± ∆c ⇒ ∆c = ∆a + ∆b (9.3)
If the quantity is given as the multiplication/division of two measured

quantities, then the total (propagated) relative error is equal to the addition of
relative errors:
∆c ∆a ∆b
C = A ×B or C = A / B ⇒ = + (9.4)
C A B
Uncertainty of an EMC measurement is critical!

Each component of uncertainty that contributes to the uncertainty of a
measurement is represented by an estimated standard deviation, termed stan-
dard uncertainty with suggested symbol ui and equal to the positive square root
of the estimated variance ui2.
The combined standard uncertainty of a measurement result, suggested
symbol, uc, is taken to represent the estimated standard deviation of the result.
It is obtained by combining the individual standard uncertainties ui. Although
the combined standard uncertainty uc is used to express the uncertainty of
many measurements, for some commercial, industrial, and regulatory applica-
tions, what is often required is a measure of uncertainty that defines an interval
about the measurement result y within which the value of the measurand Y is
confidently believed to lie. The measure of uncertainty intended to meet this re-
quirement is termed expanded uncertainty, suggested symbol, U, and is obtained
by multiplying uc by a coverage factor, suggested symbol k. In general, the value
of the coverage factor k is chosen on the basis of the desired level of confidence
to be associated with the interval defined by U = kuc. Typically, k is in the range
2 to 3. When the normal distribution applies, k=2 correspond to nearly 95%
confidence level; k=3 belongs to confidence level greater than 99%.
The standard deviation measures the width of the distribution and is used
as a measure of uncertainty. For normal distribution, any single measurement
has a 68.3% chance of falling within one standard deviation of the true value.
Often, data is given with error bars of ± one sigma, meaning that the true value
has a 68.3% chance of falling in the stated range. Values with (± 2 sigma) and
(± 3 sigma) correspond, respectively, to 95.0% and 99.7% chances that the true
value is in the stated ranges.
A measurement result should be expressed by giving the measured value,
expanded uncertainty, confidence level, and its coverage factor.
Example: Suppose you measured electric field strength as 99 dBµV/m
(with a nominal value 100 dBµV/m) and calculated the combined uncertainty
uc = 0.27 dBµV/m. You present the result as 99 ± 0.54 dBµV/m with cover-
age factor k=2. Since it can be assumed that the measurand is approximately
normally distributed, the unknown value of the measurand is believed to lie in

this interval with a confidence level of approximately 95%.
References
[1] IEC/EN61000-1-1, Electromagnetic Compatibility (EMC)—Part 1: General-Section 1:
Application and Interpretation of Fundamental Definitions and Terms.
[2] IEC/EN61000-2-1, Electromagnetic Compatibility (EMC)—Part 2: Environment-Sec-
tion 1: Description of the Environment-Electromagnetic Environment for Low-Frequen-
cy Conducted Disturbances and Signaling in Public Power Supply Systems.
[3] IEC/EN61000-2-3, Electromagnetic compatibility (EMC)—Part 2: Environment-Sec-
tion 3: Description of the Environment-Radiated and Non-Network-Frequency-Related
Conducted Ohenomena.
[4] IEC EN61000-3-2, Electromagnetic Compatibility (EMC)—Part 3-2-Limits—Limita-
tion of Harmonic Current Emissions (equipment input current ≤ 16A per phase).
[5] IEC EN61000-3-3, Electromagnetic Compatibility (EMC)—Part 3-3-Limits—Limita-
tion of Voltage Changes, Voltage Fluctuations and Flicker in Public Low-Voltage Supply
Systems (equipment with rated current ≤ 16A per phase).
[6] IEC/EN61000-3-4, Electromagnetic Compatibility (EMC)—Part 3-4: Limits—Limita-
tion of Emission of Harmonic Currents in Low-Voltage Power Supply Systems (equip-
ment with rated current greater than 16A).
[7] IEC/EN61000-3-5, Electromagnetic Compatibility (EMC)—Part 3: Limits—Section 5:
Limitation of Voltage Fluctuations and Flicker in Low-Voltage Power Supply Systems
(equipment with rated current greater than 16A).
[8] IEC/EN61000-4-2, Electromagnetic Compatibility (EMC)—Part 4-2: Testing and Mea-
surement Techniques-ESD Immunity Test.
[9] IEC/EN61000-4-3, Electromagnetic Compatibility (EMC)—Part 4-3: Testing and Mea-
surement Techniques-Radiated, Radio-Frequency, EM Field Immunity Test.
[10] IEC/EN61000-4-4, Electromagnetic Compatibility (EMC)—Part 4-4: Testing and
Measurement Techniques-Electrical Fast Transient / Burst Immunity Test.
Measurement Techniques-Surge Immunity Test.
Measurement Techniques-Immunity to Conducted Disturbances, Induced by Radio-
Frequency Fields.
[13] IEC/EN61000-4-7, Electromagnetic Compatibility (EMC)—Part 4-7: Testing
and Measurement Techniques-General Guide on Harmonics and Inter-Harmonics
Measurements and Instrumentation, for Power Supply Systems and Equipment Connected
Thereto.

Measurement Techniques-Power Frequency Magnetic Field Immunity Test.
Measurement Techniques-Pulse Magnetic Field Immunity Test.
Measurement Techniques-Voltage Dips, Short Interruptions and Voltage Variations
Immunity Tests.
[17] CISPR 16-1-4: 2003 Radio Disturbance and Immunity Measuring Apparatus—Ancillary
Equipment—Radiated Disturbances.
[18] CISPR 11, Industrial, Scientific and Medical (ISM) Radio-Frequency Equipment-
Electromagnetic Disturbance Characteristics-Limits and Methods of Measurement.
[19] CISPR 12, Vehicles, Boats and Internal Combustion Engine Driven Devices-Radio
Disturbance Characteristics-Limits and Methods of Measurement for the Protection of
Receivers Except those Installed in the Vehicle/Boat/Device itself or in Adjacent Vehicles/
Boats/Devices.
[20] CISPR 14-1, Electromagnetic Compatibility-Requirements for Household Appliances,
Electric Tools and Similar Apparatus-Part 1: Emission.
[21] CISPR 14-2, Electromagnetic Compatibility-Requirements for Household Appliances,
Electric Tools and Similar Apparatus-Part 2: Immunity-Product Family Standard.
[22] CISPR 15, Limits and Methods of Measurement of Radio Disturbance Characteristics of
Electrical Lighting and Similar Equipment
[23] CISPR 16-1, Specification for Radio Disturbance and Immunity Measurement Apparatus
and Methods-Part 1: Radio Disturbance and Immunity Measuring Apparatus.
and Methods-Part 2: Methods of Measurement of Disturbances and Immunity.
and Methods-Part 3: Reports and Recommendations of CISPR.
[26] CISPR 16-4, Part 4-1: Uncertainties, Statistics and Limit Modeling — Uncertainties in
Standardized EMC Tests.
[27] CISPR 22, Information Technology Equipment-Radio Disturbance Characteristics-
Limits and Methods of Measurement.
[28] CISPR 24, Information Technology Equipment-Immunity Characteristics-Limits and
Methods of Measurement.
[29] CISPR 25, Vehicles, Boats and Internal Combustion Engines-Radio Disturbance
Characteristics-Limits and Methods of Measurement for the Protection of on-board
Receivers.
[30] ISO 31000:2009—Risk Management—Principles and Guidelines.
[31] Agilent Application Note, Making Conducted and Radiated Emissions Measurements,
www.agilent.com.
Bibliography
NIST Technical Note 1297-1997: Guidelines for Evaluating and Expressing the Uncertainty of
NIST Measurement Results.
ITU-Rec SM.328-10 Spectra and Bandwidth of Emissions.
10
EMC and Protection
In order to avoid EMC problems, (1) EMI sources must be identified and sup-
pressed, (2) EMI victims must be strengthened, (3) finally, the coupling means
and mechanisms must be removed. EMC protection should take into account
all of these approaches. Well-known EMC protection approaches are (1) filter-
ing, (2) screening, (3) cabling and grounding, and (4) impedance matching.
10.1 Filtering: Butterworth Filters

Filters have important roles in communication/radar systems and their usage is
unavoidable when rejection of unwanted frequency range is required. Filtering
is also a major approach in EMC engineering for the cancellation of noise and
interference.
Functionally, filters can be grouped into four categories: low pass filters
(LPFs), high-pass filters (HPFs), band pass filters (BPFs), and band-stop filters
(BSFs). Various sets of analytical functions satisfy given filter specifications,
but Butterworth, Chebyshev, Cauer, and Bessel functions, with their pros and
cons, are the ones widely used in RF/microwave filter design. For example,
Butterworth filters are maximally flat in the passband but their out-of-band at-
tenuation slopes are not good. Chebyshev filters have sharper attenuation slopes
(compared to Buttorworth filters), but the payoff is the ripple inside the pass-
band. Elliptic filters have the sharpest out-of-band attenuation, but they have
undesired ripples both in and out of the passband. Here, the Butterworth filter
design is taken into consideration.
Figure 10.1 shows four filter types and the key parameter χ that specifies
out-of-band attenuation curve. The relations of the key parameter χ for four
different filter types are also given in the figure. The characteristics of the filters
247
Figure 10.1 Basic filter characteristics and design parameters.
designed according to the given specifications can then be given via frequency
responses. This may be presented as either IL versus frequency or RL versus
frequency. They are defined in Section 4.2. Serial and parallel LC resonance
circuits are also discussed in Section 4.3. As shown in Figure 4.13, four types
of filters can be realized with only inductors (L) and capacitors (C). While
parallel combinations of L and C elements are enough for LPF and HPF, LC
resonance circuits are required for the realization of BPF and BSF. Any L (C)
element, connected serially between input and output of a two-port circuit, acts
as LPF (HPF). They act oppositely when connected in parallel between input
and output. Similarly, any serial (parallel) LC resonance circuit connected seri-
ally between input and output of a two-port circuit acts as BPF (BSF). They act
oppositely when connected in parallel between input and output.
10.1.1 Butterworth LC Filters

Butterworth filter design method is a systematic form of the application ex-
plained in Figure 4.10 (i.e., using ladder-type structures) and is used when pass-
band needs to be maximally flat [1]. Details of lumped element (LC) filter the-
ory can be found in many classical sources [1–5]; therefore, Butterworth filter
design approach is briefly summarized here only for the sake of completeness.
Filter design specifications are the order of the filter (n), the 3 dB cut-off
(corner) frequency (fc) and the attenuation (AdB) that determines out-of-band
slope of the filter. For the Butterworth method, the squared magnitude of the
EMC and Protection 249
normalized transfer function is chosen to be 1/[1+χ2n], which has its poles in

quadrant-symmetry on the unit circle; AdB thus becomes
f
AdB = 10 log 1 + χ 2n  ; χ = (10.1)
fc
The user needs only to specify the cut-off frequency, and either the at-
tenuation or the order of the filter. The order of filter n corresponds to the
number of poles of the transfer function and changes with the bandwidth as
well as the attenuation slope. A (four-element) Butterworth prototype LPF is
given in Figure 10.2, which is valid when the source and load impedances are
normalized to 1 Ohm and 3 dB corner frequency to 1/2π Hz. The prototype
lumped element values Ak can be calculated from (A1 → L1 , A2 → C2, A3 →
L3, and so on):
Ak = 2 sin
(2k − 1) π ; k = 1,2,3,...,n (10.2)
2n
Table 10.1 lists prototype element values for the first several filter orders.
Equations (10.1) and (10.2) are enough to design a LPF. Generally, the user
doesn’t have to deal with the order of the filter, unless there is strict space and
cost requirements. All he or she wants is a passband characteristics and the de-
gree of out-of-band attenuation. For example, a statement such as a cut-off fre-
quency of 1000 MHz and 25 dB attenuation at 1500 MHz is enough to design
a LPF. Once the prototype values are known, the real inductor and capacitor
values of the desired filter are then obtained from
RL An An
Ln = H, C n = F (10.3)
2 πf c 2 πf c RL
Figure 10.2 Four-element (n=4) LPF prototype.

Table 10.1
Butterworth Prototype LPF Coefficients
n A1 A2 A3 A4 A5 A6 A7 A8 A9 A10
1 2.0000 1.0 — — — — — — — —
2 1.4142 1.4142 1.0 — — — — — — —
3 1.0000 2.0000 1.0000 1.0 — — — — — —
4 0.7654 1.8478 1.8478 0.7654 1.0 — — — — —
5 0.6180 1.6180 2.0000 1.6180 0.6180 1.0 — — — —
6 0.5176 1.4142 1.9318 1.9318 1.4142 0.5176 1.0 — — —
7 0.4450 1.2470 1.8019 2.0000 1.8019 1.2470 0.4450 1.0 — —
8 0.3902 1.1111 1.6629 1.9616 1.9616 1.6629 1.1111 0.3902 1.0 —
9 0.3473 1.0000 1.5324 1.8794 2.0000 1.8794 1.5321 1.0000 0.3473 1.0
A short MATLAB code, which automates the calculation of the Butter-

worth LPF prototype, is given in Table 10.2. The user needs only to supply
3-dB cut-off frequencdy, and the two out-of-band decay parameters; the rest is
automatically done by the code.
The LPF in Figure 10.2 is prototype and values belong to normalized
source and load resistances (Rs = RL = 1) as well as normalized cut-off frequency
(ωc = 1). To obtain the real filter elements in (10.3), one needs source and load
resistances, as well as the cut-off frequency. Note that the prototype LPF shown
in Figure 10.2 has inductors on the serial arm and capacitors on the parallel
arms; it starts with an inductor at the source end. An LPF prototype that starts
with parallel capacitor at the source end may also be used in the design. This
will only change filter element values; the design steps are exactly the same.
The other three filter types, HPF, BPF, and BSF, may directly be obtained
from the LPF design procedure. All that is needed are LPF, BPF, or BSF design
Table 10.2
Short MATLAB Code for Butterworth LPF Design
%-------------------------------------------------------
% Butterworth LPF Design
%-------------------------------------------------------
fc = input (‘Cut-off Frequency? ‘); AdB = input (‘Filter Attenuation, AdB? ‘);
f1 = input (‘Attenuation frequency? ‘);
n=ceil(0.5*log10(10^(AdB/10)-1)/log10(f1/fc)) ;
for k=1:n; g(k)=2*sin(((2*k)-1)*pi/(2*n)); end;
if mod(n,2) ~= 0; mm = n-1; else; mm = n; end; nn=200; df=(fmax-fmin)/nn;
for k=1:nn % Frequency loop
f(k) = (fmin+(k-1)*df); omega = 2*pi*f(k);
for kk=1:n
if mod(kk,2) ~= 0
ll=g(kk).*rload./(2*pi*fc); z(kk)=i*omega*ll;
else
cc=g(kk)./(2*pi*fc*rload); z(kk)=-i/(omega*cc);
end; end; end % Program END
parameters as shown in Figure 10.1. Once χ is extracted from given design

parameters, the rest of the procedure is the same as the procedure for LPF [1].
For example, to design HPF one needs to use χ = fc /f (instead of χ = f /fc) as
the key parameter and then calculates filter order and prototype values as if he
or she is designing a LPF. Then, the prototype HPF is obtained via the trans-
formation shown in Figure 10.3; serial inductances and parallel capacitors are
interchanged and the reciprocal values (AHPF = 1/ALPF) are used.
The inductor and capacitor of the desired HPF are obtained from
RL 1
Ln = , H, C n = F (10.4)
An 2 πf c An 2 πf c RL
The BPF and BSF design procedures are also straightforward. Figure 10.4
shows the transformation from LPF to BPF and/or BSF. Here, inductors and
capacitors of the LPF are replaced with resonance circuits according to the re-
quirements. For the BPF, serial elements must show low-impedance and parallel
elements high-impedance inside the passband B. This may be achieved by using
serial resonance circuits at serial arms and parallel resonance circuits at parallel
arms of the circuit. For the BSF, serial elements must show high-impedance and
parallel elements low-impedance inside the stopband B. Serial resonance circuit
elements of the real BSF are obtained via
An RL B
Cn = F, Ln = H (10.5)
2 π RL B 2 π f c 2 An
and parallel resonance circuit elements of the real BSF are obtained
B R A
Cn = 2
F, Ln = L n H (10.6)
2 π f c RL An 2 πB
Figure 10.3 LPF-HPF transformation.

Figure 10.4 (a) LPF → BPF; (b) LPF → BSF transformations.
Finally, real values of the BPF on the serial arms are:
B R A
Cn = 2
F, Ln = L n F (10.7)
2 π f c RL An 2 πB
and on the parallel arms are:
An RL B
Cn = F, Ln = H (10.8)
2 π RL B 2 π f c 2 An
Note that input/output impedance mismatch is also important in filter-

ing. A simple inductor (capacitor) as a filter element gives good results up to
35–40 dB attenuation at low (high) impedance circuits but is quite useless in
high (low) impedance circuits.
Suppose a BSF with center frequency 800 MHz, 3 dB (stop), and 20 dB
(pass) bandwidths of 400 MHz and 200 MHz, respectively, is requested. The
order of this filter will then be n = 4. Figure 10.5 shows the designed filter. Its
frequency characteristic is given in Figure 10.6.
Figure 10.5 Butterworth filter circuit for the given example.
Figure 10.6 Frequency response of the LPF given in Figure 10.5.
EMC protection filters are widely used in harmonic filtering in AC mains,

DC power supply filtering, and communication channel filtering. AC mains
filters are usually linear and put input of a power supply inside a screen in order
to protect AC mains from EMI effects. DC power supply filters are similar but
their cut-off frequencies are usually selected very low. Communication filters
may be any type.
Filters provide insertion loss by creating an impedance mismatch between
the AC mains and the equipment; therefore, the choice of filter circuit should
take into account the impedances of both source and load. IL versus frequency
curves indicate the performance and hence quality of a filter and is usually
supplied by the manufacturer. However, precise performance can only be estab-
lished by testing the filter under real conditions.
The components in a filter are classified for a range of climatic and me-
chanical requirements, including lower temperature limit, over temperature
limit, and relative humidity. Characteristics of CM and DM currents are also
important in classification.
A filter should be positioned as close as possible to the connection to the
supply—ideally before the AC mains switch in the front panel and fuses. Oth-
erwise, the connecting cables could provide a coupling path via stray induction
to the unfiltered cables. A good ground connection (with large contact surfaces,
without paint) is also essential.
10.1.2 Transmission Line Filters

As shown in Section 4.4 (see Figure 4.17), any inductor and/or capacitor can be
obtained via a piece of SC and/or OC TL (0 ≤ l ≤ λ/4 or 4 ≤ l ≤ λ/2). The im-
pedance across terminals of a SC transmission line (TL) having a characteristic
impedance of Z0 and a length of l is
Z SC = jZ 0 tan θ (10.9)
where θ = βl is the electrical length, and β = 2π/λ is the phase constant (λ is the
wavelength). This impedance becomes
ZOC = − jZ 0 cot θ (10.10)
if the line is OC. These two form the basis of LC to TL filter transformation.
Transforming lumped elements to distributed elements is desired espe-
cially when the physical length of lumped element becomes comparable with
the signal wavelength [2, 6]. Richards’ transformation may be used to convert
lumped prototype elements into ideal TLs that have equal commensurate
length. Commensurate lengths can be λ/4 or λ/8, but generally λ/8 is being
used to obtain a more compact filter in size. Using Richards’ transformation,
capacitors can be replaced with OC TLs and inductors with SC TLs. The char-
acteristic impedances of the TLs with SC and OC terminations are given as
RL An R tan θ
Z SC = , ZOC = L (10.11)
tan θ An
Serial OC or SC TLs are not desired in distributed circuits, especially

in PCBs, because their implementation is quite difficult and they are not so
efficient. Therefore, Kuroda identities are used, and lumped filter prototypes
are converted into distributed forms [7]. As shown in Table 10.3, as the order
of the LC filter increases, the number required Kuroda transforms drastically
Table 10.3
Filter Order Versus Number of Required Kuroda Transformations
Filter Order 2 3 4 5 6 7 8 9 10 11 12
Kuroda 1 2 5 8 13 18 25 32 41 50 61
Transformations
increases. When converting lumped element prototype values into TLs using
Kuroda identities, one needs to separate TL elements via unit elements. Add-
ing a unit element having the same impedance of the scaled filter and the same
commensurate length at both sides of the filter does not change the frequency
response. Kuroda identities are heuristic, and the number of transforms re-
quired in converting lumped element prototype into a distributed form is a
disadvantage [7].
Example: Design a LPF with a 500-MHz cut-off frequency and a 30-dB
attenuation at 700 MHz. The order of the corresponding LC filter is found
from (10.1) at n = 5. The TL equivalent has five stubs. Both LC and TL filters
are given in Figure 10.7.
Tables 10.4 and 10.5 list short MATLAB scripts which plot IL versus
frequency characteristics of LC and TL LPFs, respectively, for another LPF with
100-MHz cut-off frequency and 30-dB attenuation at 200 MHz (see Figure
10.8).
As observed, LC LPF has the exact desired performance. On the oth-
er hand, TL LPF is equivalent to LC LPF in the passband and at the cut-off
Figure 10.7 Fifth-order LC LPF and its TL equivalent.

Table 10.4
Short MATLAB Butterworth LC LPF Code
%-------------------------------------------------------------------------------------
% Program: LC_LPF.m (fc = 100 MHz; 30dB attenuation at 200 MHz)
%-------------------------------------------------------------------------------------
clc; clear all; tpi=2*pi; Vs = 100; Zs = 50; Zl = 50; % source/ load impedance
f = 0:1e6:250e6; IL = zeros(1,length(f)); % define vectors for freq/IL
% Define low-pass filter component values
L1 = 49.2e-9; C2 = 51.5e-12; L3 = 159.2e-9; C4 = C2; L5 = L1;
for n=1:length(f)
Z1=1i*tpi*f(n)*L1; Z2=-1i/( tpi*f(n)*C2); Z3=1i*tpi*f(n)*L3; Z4=Z2; Z5=Z1;
Ze=Z5+Zl; Zd=Ze*Z4/(Ze+Z4); Zc=Z3+Zd; Zb=Z2*Zc/(Z2+Zc);
Za=Z1+Zb; Zeq=Zs+Za; % Calculate equivalent impedances
I = Vs/Zeq; Ib = I/(1+Zc/Z2); Ie = Ib/(1+Ze/Z4); % Calculate currents
% Calculate voltage at the load with or without the filter present
V2 = Zl*Ie; V2bar = Vs*Zl/(Zl+Zs); IL(n)=20*log10(abs(V2)/abs(V2bar));
end; plot(f/1e6,IL); xlabel(‘Freq [MHz]’); ylabel(‘IL [dB]’); ylim([-50 10]); % END
Table 10.5
Short MATLAB Butterworth TL LPF Code
%------------------------------------------------------------------------------------
% Program: TL_LPF.m (fc = 100 MHz; 30dB attenuation at 200 MHz)
%------------------------------------------------------------------------------------
clc; clear all; Vs = 100; Zs = 50; Zl = 50; % source/load impedance
f = 0:1e6:250e6; IL = zeros(1,length(f)); % define vectors for freq/IL
% Define low-pass filter component values
fc=1e8; lambdac=3e8/fc; A1 = 0.618; A2 = 1.618; A3 = 2.000;
A4 = 1.618; A5 = 0.618; teta=2*pi*fc/3e8*lambdac/8;
Z01 = A1*Zl/tan(teta); Z03 = A3*Zl/tan(teta); Z05 = A5*Zl/tan(teta);
Z02 = Zl*tan(teta)/A2; Z04 = Zl*tan(teta)/A4;
for n=1:length(f)
teta=2*pi*f(n)/3e8*lambdac/8;
Z1 = 1i*Z01*tan(teta); Z3 = 1i*Z03*tan(teta); Z5 = 1i*Z05*tan(teta);
Z2 =-1i*Z02/tan(teta); Z4 =-1i*Z04/tan(teta);
Ze = Z5+Zl; Zd = Ze*Z4/(Ze+Z4); Zc = Z3+Zd;
Zb = Z2*Zc/(Z2+Zc); Za = Z1+Zb; Zeq = Zs+Za;
I = Vs/Zeq; Ib = I/(1+Zc/Z2); Ie = Ib/(1+Ze/Z4); % currents
V2 = Zl*Ie; V2bar = Vs*Zl/(Zl+Zs); % Load voltages
IL(n)=20*log10(abs(V2)/abs(V2bar));
end; plot(f/1e6,IL); xlabel(‘Freq [MHz]’); ylabel(‘IL [dB]’); ylim([-50 10]); % END
frequency. Its out-of-band attenuation is much sharper. Moreover, TL filters’

performances are periodic with the period of 2fc. Generalized codes of But-
torworth LC LPF and TL LPF are given in Tables 10.6 and 10.7, respectively.
10.1.3 Microstripline (ML) Transformation

A microstripline is a TL. It is generally characterized with its dielectric constant,
conductor thickness, the height of the dielectric material, and the dissipation
factor (also called the loss tangent). Analytical solutions of MLs require a treat-
Figure 10.8 IL versus frequency of the LPF obtained with the codes in Tables 10.4 and 10.5.
ment of Maxwell equations under complex boundary conditions. This is why

no closed-form solution has appeared yet. Instead, many sets of empirical for-
mulas can calculate the characteristic impedance of a ML for a given substrate
thickness, relative permittivity, and strip width and thickness. Once the charac-
teristic impedance is calculated, TL-ML transformation is straightforward [7].
10.1.4 MATLAB-Based Filter Package

A MATLAB-based filter design package MWFilterDesigner.m has been devel-
oped to simplify and automate all steps of the filter design procedure. The
package is substantially different than the one in [8, 9] in that it (a) designs and
pictures a lumped element (LC) filter for given specifications (filter type, cut-
off frequency/bandwidth, and out-of-band attenuation slope); (b) transforms
and pictures the LC filter to its TL equivalent; (c) transforms and pictures the
TL filter into its ML equivalent using Richard’s transformation and Kuroda
identities systematically (for a given ML parameters of relative permittivity and
substrate thickness); (d) calculates and plots insertion IL of all LC, TL, and ML
filters and (e) automatically exports the ML layout to a GDSII stream format
that makes EM simulations of the ML filter possible with various commercial
packages (GDS II stream format is a database file format, which is the integrat-
ed circuit industry default standard for integrated circuit layout data exchange).
The main window of the filter package is shown in Figure 10.9.
Table 10.6
Short MATLAB Code for IL Calculation of a LC BPF
%---------------------------------------------------------------
% Program : BW_BPF.m
%---------------------------------------------------------------
clc; clear all; rs = 50; rload=rs; vl2=rload/(rload+rs);
fc = input (‘BPF--Cut-off Frequency [MHz] : ‘); fc = fc*1e6;
Band = input (‘Pass Band [MHz] : ‘); Band = Band*1e6;
SBand= input (‘Stop Band (should be more larger Pass Band) [MHz] : ‘);
SBand = SBand*1e6; ff=SBand/Band; AdB = input (‘Out of band atten [dB] : ‘);
display(‘Frequency range for the graphics:’);
fmin = input(‘Min Frequency [kHz] : ‘); fmin=fmin*1e3;
fmax = input(‘Max Frequency [MHz] : ‘); fmax=fmax*1e6;
n=ceil(0.5*log10(10^(AdB/10)-1)/log10(ff)) ; % n=Filter order
display(‘The order of the Filter: ‘); n
%.........Prototype Filter elements..............
for k=1:n; g(k)=2*sin(((2*k)-1)*pi/(2*n)); s(k)=1/g(k); end
display(‘Element values of the LPF prototype filter: ‘); g
if mod(n,2) ~= 0; mm = n-1; else; mm = n; end; nn=200; df=(fmax-fmin)/nn;
for k=1:nn % Frequency loop
f(k) = (fmin+(k-1)*df); omega = 2*pi*f(k);
for kk=1:n
if mod(kk,2) ~= 0
cc=Band/(2*pi*fc*fc*rload*g(kk)); ll=rload*g(kk)/(2*pi*Band);
z(kk)=i*omega*ll-i/(omega*cc);
else
cc=g(kk)/(2*pi*rload*Band); ll=rload*Band/(2*pi*fc*fc*g(kk));
z(kk)=(i*omega*ll)*(-i/(omega*cc))/(i*omega*ll-i/(omega*cc));
end; end
% Checkt the last element and load
if mod(n,2) ~= 0; ZA = rload + z(n); else; ZA = rload; end;
ZLast = ZA; % Calculate total input impedance
for kk=mm:-1:1
if mod(kk,2) == 0; ZT(kk)=z(kk)*ZLast/(z(kk)+ZLast); else;
ZT(kk)=z(kk)+ZLast; end;
ZLast = ZT(kk); end
iload = 1/(rs+ZLast); % Calculate source current
% Distribute and find out load current
for kk=2:2:mm-1; iload=iload*z(kk)/(z(kk)+ZT(kk+1)); end;
iload=iload*z(mm)/(z(mm)+ZA); vl1=iload*rload;
iloss(k)=-20*log10(abs(vl2/vl1));
end % End of frequency loop
plot(f/1e6,iloss); ylim([-40 10]); ylabel(‘IL[dB]’); xlabel(‘Freq [MHz]’) % END
The user only selects the filter type and inputs the two design parameters.
The ML material properties (substrate height of the dielectric material and its
relative permittivity) are also supplied by the user. The rest is handled in the
program automatically, and the LC filter, the TL filter, the ML filter, and all
element values and dimensions are given by the program. The layout of the ML
filter and the frequency responses are also plotted.
The ML equivalent of the filter given in Figure 10.7, obtained with this
package, is shown in Figure 10.10 Both the top view and dimensions are given
Table 10.7
Short MATLAB Code for IL Calculation of a TL with Stubs
%------------------------------------------------------------------------------------------
% Program : LS_IL_TRLine.m
% Parameters:
% ls(ns) : lengths of ns stubs
% d(ns) : lengths of ns nodes that ls(ns) are connected
% type(ns) : ‘oc’, ‘sc’ for open and short circuit stubs
% Z0 : Characteristic impedance; ZL : Load impedance
% Zin(ns) : Equivalent impedance of the ns node looking towards the load
% Zs(ns) : parallel effects of the ns stubs at the nodes
% | d3 | d2 | d1 |
%---------------\----------\-----------\---------------|
% line Zo \ \ \ ZL
%------------\----\-----\----\------\----\-------------|
% l3 \ \ l2 \ \ l1 \ \
% \____\ \____\ \____\
%-----------------------------------------------------------------------------------------------
clear all; clc; close all; % ZL=75+j*100;
Z0=50; ZL=50; Zsource=50; Vsource=1; VL1=Vsource*ZL/(Zsource+ZL);
fmin = input(‘\n Min Freq [MHz] : ‘); fmax = input(‘\n Max Freq [MHz] : ‘);
fmax=fmax*1e6; fmin=fmin*1e6; nf=200; df=(fmax-fmin)/nf;
ns = input(‘\n Number of stubs [] : ‘);
for k=1:ns
fprintf(‘Give d(%1d) in [cm] \n’,k); d(k) = input(‘\n ‘)*1e-2;
fprintf(‘Give stub length ls(%1d) in [cm] \n’,k);
ls(k) = input(‘\n ‘)*1e-2; tip = input(‘Stub termination [oc] or [sc] \n’, ‘s’);
if tip ==’oc’; type(k)=1; else; type(k)=0; end
end
for n=1:200 % Frequency Loop
freq(n)=fmin+(n-1)*df; beta=2*pi*freq(n)/3e8;
for k=1:ns % Calculate equivalent stub impedances
if type(k) == 1; Zs(k) =-j*Z0/tan(beta*ls(k)); else; Zs(k) = j*Z0*tan(beta*ls(k));
end; end
Ztot=ZL; % Calculate equivalent input impedances at each node
for k=1:ns
Zin(k) = Z0*(Ztot+j*Z0*tan(beta*d(k)))/(Z0+j*Ztot*tan(beta*d(k)));
Ztot = Zin(k)*Zs(k)/(Zin(k)+Zs(k));
end
Isource = Vsource/(Zsource+Ztot); Vleft = Isource*Ztot; Iright = Isource;
for k=ns:-1:1
Ileft = Iright*Zs(k)/(Zs(k)+Zin(k));
Vright = cos(beta*d(k))*Vleft-j*Z0*sin(beta*d(k))*Ileft;
Iright =-j*sin(beta*d(k))*Vleft/Z0+cos(beta*d(k))*Ileft; Vleft = Vright; end
VL2 = Vright; IL(n) = 20*log10(abs(VL2/VL1));
end; plot(freq/1e6,IL); xlabel(‘Freq [MHz]’); ylabel(‘IL [dB]’); % Program END
in this figure. IL versus frequency of all these (LC, TL, and ML) filters is plotted
in Figure 10.11.
Note that frequency responses of LC and TL filters are automatically cal-
culated with this package. On the other hand, the ML structure needs a full
wave simulation package. The GDSII-format file may be used together with the
fullwave FDTD-based package MSTRIP [10] and filter characteristics can be
Figure 10.9 MWFilterDesigner package.
Figure 10.10 Microstrip equivalent of the designed LPF.
obtained. The MS curve in Figure 10.10 is obtained in this way. Both MWFil-
terDesigner and MSTRIP tools can be downloaded from www.leventsevgi.net.
10.2 Ferrites
A ferrite bead is a passive electric component used to suppress high-frequency
noise in electronic circuits. It is a specific type of electronic choke. Ferrite beads
employ the mechanism of high dissipation of high-frequency currents in a fer-
rite to build high-frequency noise suppression devices. Ferrite beads may also
Figure 10.11 Filter characteristics of the designed LPF.
be called ferrite rings, ferrite EMI filters, or ferrite chokes. Ferrite beads are
used as passive bandpass filters. The geometry and EM properties of coiled
wire over the ferrite bead result in a high resistive impedance (resistance) for
high-frequency signals, attenuating high-frequency EMI/RFI electronic noise.
Typical ferrite filters are shown in Figure 10.12.
Ferrite, with high permeability (µ) and very low conductivity (σ), is a
ceramic material, made like pottery or bricks. It contains iron and other fer-
romagnetic elements. If a cable is surrounded by ferrite, then the magnetic
field encircling the cable due to CM current in the cable magnetizes the fer-
rite. Because of its high permeability, the amount of energy stored magnetically
Figure 10.12 Typical ferrite filters commercially available.

in the ferrite is very high. Thus, the inductance per unit length of cable sur-
rounded by ferrite is very high. A CM choke is an inductor having a high value
of inductance.
The energy is either reflected back up the cable or absorbed resistively
within the ferrite core and dissipated as heat. Only in extreme cases will the
heat be noticeable. A pure inductor does not dissipate energy; it only absorbs
energy from the circuit and returns it at a later time. A ferrite bead, by design,
filters out the high-frequency noise in the circuit by dissipating it as heat. The
ferrite bead is effectively an inductor with a very small Q factor. When electrical
noise within the target frequency range travels in the signal cable, an opposite
electromotor force is induced in the ferrite bead because of its high inductance.
The material used to construct the ferrite bead becomes highly resistive
at the design frequency range, and the induced current inside the bead is dissi-
pated as heat instead of inducing an opposing current back in the signal cable. It
is for this reason that the specific circuit characteristics as well as the frequency
band of the noise need to be taken into account when the ferrite bead is in-
stalled as a noise filter.
The right ferrite filter is selected according to the frequency range where
maximum IL is needed, the amount of attenuation, the cable type and struc-
ture, temperature requirements, and DC currents. It may be mono-block or in
pieces.
In two-port circuits, ferrites are modeled as serial impedances.
10.3 Screening/Shielding
An effective EM protection method is screening or shielding. Shielding is the
degree of isolating two pieces of equipment, devices, or environments [11–14].
This could be satisfied in two ways: (1) EM fields exponentially decay with the
distance, so increasing the distance between an EMI source and victim may sup-
ply required isolation; (2) when this is impossible and/or undesired, then the
isolation is supplied by putting obstacles between the EMI source and victim.
Enclosing devices and/or circuitry in a shielded enclosure is a good way of
controlling EM emissions.
Figure 10.13 shows a scenario for the shielding effectiveness (SE) defini-
tion. SE is a degree of isolation of an enclosure between EM interfering source
and a victim:
E  H 
SE = 20Log 10  0  = 20Log 10  0  (10.12)
 E1   H1 
Figure 10.13 Definition of the shielding effectiveness (SE).
where E0(H0) and E1(H1) are electric (magnetic) field strength values measured
or calculated at the victim position without and with the enclosure, respectively.
If the enclosure is absent or fully transparent in terms of EM waves, then E1 =
E0 and the SE is 0 dB; otherwise, positive and negative SE corresponds to the
attenuation and amplification, respectively, of the EM interfering signal.
The value of SE depends upon a number of factors: EM interfering source
characteristics, such as wave impedance; frequency; amplitude; polarization;
enclosure characteristics such as thickness, material permittivity, permeability,
conductivity, and so on. Among all, the factor related to the type of EM inter-
fering source is usually dominant.
Choosing the right shielding approach depends on identification of the
EMI source (i.e., the distance and type of disturbing signal). In the far field,
EM waves act as plane waves; therefore, electrical shielding may always be ap-
plied. In the near field, shielding approaches are totally different for electrical
and magnetic disturbing sources. As discussed in Section 3.5, all disturbing
signals act either as an electric dipole or a magnetic dipole. Electric dipole has
strong electric fields in the near field; magnetic fields are dominant in the near
field of magnetic dipole. This is explained in Figure 10.14, where a nonmag-
netic (µ = µ0), mono-block metal sphere with thickness d is used as a shield in
the vicinity of both strong electric and magnetic disturbing signals.
Simple SE expressions are also given in the figure. As observed, SE of a
nonmagnetic metal conductor enclosure is quite different for electric and mag-
netic fields in the near field. For high electric fields, electrical shielding is high at
low frequencies; as the frequency increases, SE decreases, reaching a minimum
at medium frequencies, and then increases again toward high frequencies. For
high magnetic fields, SE is almost zero (the enclosure is almost transparent) at
low and medium frequencies and starts increasing at high frequencies.
Figure 10.14 E and H field screening with nonmagnetic conductors.
10.3.1 Shielding Effectiveness—Ideal Case

Assume an infinite conducting wall, having thickness “t,” separates two regions.
In this ideal case, theoretical SE expressions may be obtained. As shown in Fig-
ure 10.15, isolation is supplied with three different mechanisms: (1) via reflec-
tions on the air-metal boundary; (2) via absorptions through the metal; (3) via
multiple internal reflections. Usually, the first two contributions are taken into
account in theoretical SE calculations; shielding from multiple internal reflec-
tions may be neglected.
Reflection Losses
The SE from reflections can be calculated in terms of wave impedance Zw and
conductor impedance Zs:
E   ( Z + Z )2 
SE R = 20Log 10  i  = 20Log 10  s w
 (10.13)
 Et   4 Z Z
s w 
where Ei and Et are field strengths incident upon and transmitted through the
screen, respectively. The screen impedance is Z s = ωµ σ Ω. For the far field,
the (plane-) wave impedance in free space is Zw = Z0 = 120π Ω. Therefore,
(10.13) reduces to
Figure 10.15 Reflection and transmission losses in SE calculations.
 1 σr   σr 
SE R = 20Log 10   = 168 + 10Log 10   (10.14)
 4 µr f   µr f 
But in the near field, the type of interfering source is important. The wave
impedance of electrical (i.e., dipole-like emitting) source is Zwe = Z0(l/2πd) Ω,
while the wave impedance of magnetic (i.e., loop-like emitting) source is Zwm =
Z0(l/2πd) Ω ( (see Section 5.9.1 and Figure 5.31). Therefore, the SER in [dB]
in the near field is
 σ r 
SE Re = 332 + 10Log 10  3 2 (10.15)
 µr f d 
for electrical sources, and for the magnetic sources it is given as:
 fd σ r 
2
SE Rm = 14.6 + 10Log 10   (10.16)
 µr 
Reflection loss in the near field is different for electric and magnetic
dipoles.
Absorption Losses
EM waves barely penetrate conductors. The degree of this penetration is speci-
fied in terms of skin (penetration) depth δ = 2 ωσµ[m] (see Section 3.4), and
absorption losses are calculated in terms of skin depth as:
  t 
SE A = 20Log 10 exp    = 131.4 t µr σ r f (10.17)
  δ 
The total SE is then obtained from the addition of SER and SEA. A sim-
ple, MATLAB-based SE calculator code based on (10.12)–(10.17) is given in
Table 10.8. A simple virtual SE calculator tool is also designed for this pur-
pose [15]. The user only needs to choose shielding material type and supply
its thickness and the interfering source distance. The virtual SE tool yields SE
versus frequency plots of both electrical and magnetic interfering sources. After
specifying input parameters and the type of annoying source, one calculates SE
caused by the absorption using (10.17), then checks whether or not the inter-
fering source is in the near or far field region and calculates SE caused by the
reflections accordingly either from (10.14) or (10.15)–(10.16), and finally finds
out the total SE by adding these two contributions.
Two typical examples are given in Figures 10.16 and 10.17 for 0.01-mm-
thick screen from a 30-cm distant interfering source. Figures 10.16 and 10.17
belong to copper (µr = 1, σr = 1.0) and mumetal (µr = 30,000, σr = 0.03) shields,
Table 10.8
Short MATLAB Code for SE Calculations
%--------------------------------------------------------------------
% Program : Shielding Effectiveness (SE) Calculator
%--------------------------------------------------------------------
clear all; clc; format long; sigmar=0.1; mur=500; N=250;
dt=input(‘Material thickness [mm] = ? ‘); d=input(‘\n Screen Distance [cm] = ? ‘);
dt=dt/1e3; d=d/1e2; eps0=10^(-9)/(36*pi); mu0=4*pi*10^(-7); z0=sqrt(mu0/eps0);
fmin=1e3; fmax=1e9; df=(fmax-fmin)/N; df=(fmax-fmin)/N;
for k=1:2*N % Frekans çevrimi
fr(k)=fmin+k*df; wavel=3e8/fr(k);
if d < wavel/(2*pi) % Near Field
SRe=322+10*log10(sigmar/(mur*fr(k)^3*d^2));
SRm=14.6+10*log10(fr(k)*d^2*sigmar/mur);
SA=131.4*dt*sqrt(fr(k)*sigmar*mur);
else % Far Field
SRf=168-10*log10(mur*fr(k)/sigmar); SA=131.4*dt*sqrt(fr(k)*sigmar*mur);
SEf(k)=SRf+SA;
end
SEe(k)=SRe+SA; SEm(k)=SRm+SA;
end
semilogx(fr,SEe,’k’,fr,SEm, ‘r--’); grid; xlabel(‘Freq [MHz])’); ylabel(‘SE [dB]’);
legend(‘Electrical Shielding’,’Magnetic Shielding’); % Program END
Figure 10.16 SE versus frequency for a copper shield at 30 cm.
Figure 10.17 SE versus frequency for a mumetal shield at 31 cm.
respectively (here, σr is the relative conductivity with respect to copper; σr = σ/

σcu; σcu = 5.8107 S/m). SE versus frequency plots for both electrical and magnet-
ic interfering sources are shown in the figures. As observed, electrical shielding
is effective at low frequencies, whereas magnetic shielding is not. After a critical
frequency (fc = 3108/2πd) determined from the near field–far field boundary (d
= λc /2π), both electrical and magnetic screenings are the same.
Theoretically, a few hundred dB SE can be reached at low frequencies

even with a few micrometer thick shield against electrical interfering sources,
but this may be as low as 0 dB for the magnetic sources as seen in the earlier ex-
amples. Starting from low frequencies, SE decreases (increases) with frequency
for electrical (magnetic) shielding.
In practice, depending on the criticality of the problem under consid-
eration, SE values of 30–60 dB are considered acceptable, whereas SE values
in the range of 70–90 dB represent quite high-quality shields. The shielding
performance of a metal box with no holes or seams exceeds 100 dB, which is
usually too high to measure. If the shield has holes, slots, joints, vents, windows,
or other discontinuities, the SE can only be as good as allowed by such shield-
ing imperfections. One rule of thumb is to find the longest dimension of these
discontinuities and determine the frequency at which that length represents
a half-wavelength. As a worst-case approximation, the shield can be assumed
transparent beyond that frequency. The following remarks are important:
• Shielding mostly occurs because of reflections and absorptions at low

and high frequencies, respectively.
• One should decide whether or not both electrical and magnetic shield-
ing is necessary. In many cases, only electrical screening would be neces-
sary.
• Shielding occurs because of absorption of magnetic fields at low fre-
quencies.
• High conductivity is good for both reflections and absorptions.
• High permeability causes high absorptions but low reflections.
• Shielding of low frequency magnetic fields necessitates high permeabil-
ity materials.
• Absorption is directly proportional with the screen thickness.
• While thick screens are required to shield magnetic fields, very thin ma-
terials (e.g., aluminum foil) may be enough for electric fields.
• The distance between EMI source and victim determines reflection
losses.
• Interfering electric and magnetic sources are located near and far from
the screen, respectively.
SE performance of a screen depends on the material parameters, frequen-

cy, and type of interfering sources. In practice, connections, galvanic effects,
bonds, gaskets, holes, and apertures all affect SE.
Magnetic Screening/Filtering
Magnetic screening is done via creating magnetic fields opposite to magnetic
field lines of the disturbing signals. Enclosures like Faraday cage do not work
for magnetic fields, since field lines form closed loops. Roughly speaking, mag-
netic screening/filtering is important below 30 MHz. Holes and apertures are
barely important.
Electrical Screening
Electrical screening is applied by blocking field lines that start at positive charg-
es and end at negative charges. Electrical screening is important for frequencies
above 30 MHz. Note that isolated pieces of conductors act as an antenna. Cable
connectors, holes, and apertures reduce electrical screening performance. Con-
tact resistance between different conductors is also important. Several empirical
SE expressions can be used to get a feeling of a degree of screening. One of them
given for an aperture is SEdB = 20log10(λ/2d ), where λ and d are the wavelength
and the longest aperture size. As a rule of thumb, d ≤ λ/20 is necessary to reach
at least 20 dB SE. Experience tells us that in electrical shielding, material type
and connections/bonding are important below 10 MHz; above this frequency,
holes and apertures are critical. Shielding materials can be grouped into three
categories:
• High performance materials: Materials like copper, steel, stainless mono-

block enclosures (supply 80–120 dB SE);
• Standard performance materials: Conducting plates and/or metal-in-
jected plastics (supply 20–40 dB SE);
• Low performance materials: Aluminum foil, conducting polymers, and
fabrics (supply 15–30 dB SE).
10.3.2 Shielding in Real Life

In practice, it is neither possible to reach theoretical SE values nor is it easy to
calculate the SE of realistic objects. The previously shown expressions belong
either to ideal cases or very simplistic ones. SE predictions for real enclosures
are extremely complex and mostly impossible to model analytically; therefore,
they necessitate the use of numerical models [12, 13]. International IEEE and
ASTM SE standards are in [14, 16]. A simulation example obtained with the
FDTD SE package presented in [12] is shown in Figure 10.18. Here, SE versus
frequency variations of a rectangular PEC box with one and two apertures are
presented. In each figure, the dashed lines belong to a single aperture (on the
Figure 10.18 SE versus frequency of a metal box with various apertures.
right side) enclosure. The solid lines represent two-aperture cases. Some obser-
vations from this figure may be listed as follows:
• Positive SE means EM isolation; SE = 0 dB means full transparency. But

negative SE means amplification! This happens in enclosures because of
ringing effects in the time domain corresponding to resonances in the
frequency domain.
• The peaks and nulls in the SE versus frequency plots correspond to
certain dimensions of the enclosures, aperture sizes, and the interfer-
ing sources, mutually resonating; therefore, attention should be paid to
these frequencies.
• EM interactions in these structures are very complex, so it is almost im-
possible to make statements such as “as the frequency increases SE …,”
or “as the number of apertures and holes increases SE …,” and so on.
Suppose a λ/2 dipole is located inside a PEC enclosure with an aperture

on one of its vertical facets. Another λ/2 dipole is located outside right across
the aperture. If the dipole inside radiates and the one outside receives, this is a
RE problem. On the other hand, an AC current on the outside dipole induces
a voltage on the inside dipole. This is a typical RI problem. Are they the same
for the SE characteristics of the enclosure? Does it matter which dipole radi-
ates? The answer of this question is given in Figure 10.19, where SE versus
frequency of both cases is simulated [15]. Here, a 10-cm cubical enclosure with
2-cm by 6-cm rectangular aperture is used. Two, infinitely thin, 6-cm-long,
center-fed dipoles are taken into account. The inside antenna is located right
across the aperture vertically at the center of the enclosure. The outside antenna
is located vertically 15 cm away from the aperture. Simulations are performed
Figure 10.19 SE versus frequency for inside-outside and outside-inside emissions.

twice; inside-outside (inside antenna transmits, outside antenna receives) and

outside-inside (outside antenna transmits, inside antenna receives), with and
without the enclosure. As observed, the results perfectly matched.
When available, mathematical models are best because they are exact—or
the contributions of the approximations made there are exactly known and help
one to understand the physics of the problem. Unfortunately, they are limited
to only a few canonical structures. The virtual tools such as those supplied in
[13] are simple and may be used in education and training. But when it comes
to the simulation of realistic problems, it is best to refer to commercial packages.
10.3.3 Practical Protection Tips

EM emissions depend on frequency, signal strength, and the geometry. Equal
and opposite (differential mode) currents belong to useful signals and decrease
undesired emissions. Equal currents flowing on both wires of a cable (CM cur-
rents) complete their loop over the ground or ground-like enclosures, surface
conductors, and so on. They are the disturbing signals and create EMI problems.
The frequency or the size of a system/equipment does not mean much; the
relation between frequency (in other words, wavelength) of the disturbing sig-
nal and geometry of the system/equipment (λ-l ) is critical for EMC engineers.
Disturbing EM signal may be transient or permanent. These two create
quite different EMC problems. Usually, it is easy to deal with EMC problems in
the frequency domain; unfortunately, today, almost everything is digital. This
makes EMC problems broadband. Moreover, the trend in EMI receivers is to-
ward time domain measurements. Frequency domain behaviors are obtained
through discrete Fourier transforms (DFTs) or fast Fourier transforms (FFTs).
DFTs and FFTs are synthetic processes that are totally different than the math-
ematical definition of Fourier transforms (see Section 6.4).
Theoretically, discrete signals have infinite harmonics. During a design,
test, or measurement, EMC engineers should take into account frequencies (at
least) up to 10th harmonics. For AC mains, the standards request up to 40th
harmonics.
Cables and connectors act as wire antennas; holes and apertures act as
aperture antennas. They both create EMI problems. A practical way to mitigate
the emissions they create is to keep their size much less than the wavelength of
the maximum frequency of the disturbing signal (l << λmin, e.g., l ≤ λmin/20).
10.4 Cables and Connectors

Cables are critical in EMC engineering because they are usually the longest
parts in a circuit and act as antennas. Undesired effects are twofold: they pick
up energy from a disturbing signal, and they radiate undesired EM energy and
disturb others.
The capacitive or electric coupling results from the interaction of electric
fields between circuits. The inductive, or magnetic, coupling results from the
interaction between the magnetic fields of two circuits. The combination of
electric and magnetic fields is appropriately called EM coupling or radiation.
To prevent radiation of a magnetic field from a conductor grounded at both
ends, the conductor should be shielded, and the shield should be grounded at
both ends.
Common-Impedance Coupling
When a disturbing signal shares a ground connection with a victim, then its
current flowing through the common impedance may develop an unwanted
victim voltage. The common impedance increases with high-frequency signals
or signals with high di/dt. A solution is to separate the connections (no com-
mon current path), and eliminate common impedance. Grounds are the most
usual source of common impedance. The inductance of a straight length of wire
is L = 0.0051/2.3ln [mH], where l and d are length and diameter in inches (a
useful rule of thumb is 20 nH/in).
Magnetic Induction
Even with no direct connection, the fields resulting from circuit operation will
allow coupling between two adjacent circuits. This is known as reactive or near-
field coupling. An AC current flowing in one conductor creates a magnetic
field that may couple with a nearby conductor and induce a voltage in it. The
voltage induced in the victim conductor is V = –M diL/dt, where M is the mu-
tual inductance in [H]. M depends on the areas of the source and victim cur-
rent loops, their orientation and separation, and the presence of any magnetic
screening (note that this is not the same as electric field screening). Typical
values for short lengths of cable loomed together are 0.1 mH to 3mH.
Electrical Induction
Changing voltage on one conductor creates an electric field that may couple
with a nearby conductor and induce a voltage on it. The voltage induced on
the victim conductor is V = –CdVL /dtZin , where C is the coupling capacitance
and Zin is the impedance (to ground) of the victim circuit. The value of C is a
function of the distance between the conductors, their effective areas, and the
presence and point of connection of any electric screening material. Typically,
two parallel wires, 1–3 mm apart, show a coupling capacitance of about 50
pF/m. Note that stray capacitance will complete the coupling path even if the
two circuits are not directly referenced to each other.
Transfer Impedance
The quality of a cable at low frequencies is the transfer impedance and at high
frequencies is screening attenuation. Transfer impedance is defined as the ratio
of the voltage induced on the inner conductor to the current flowing along the
outer conductor because of the external effects (see Figure 10.20).
10.4.1 Cables
Cables in a circuit guide signals from one end to the other. Performances of
cables depend on their screen material and type of the grounding used for their
screen. Coaxial cables are naturally screened. The critical issue for the coaxial
cables is their connection and grounding. Pair or twisted-pair cables must be
screened. Two typical cables are pictured in Figure 10.21. On top, a multicon-
ductor cable with copper screen and ferrite isolator layers is shown. Electrical
screening of this cable is good. Even though ferrite cover is used, magnetic
screening of this cable is poor because of multiconductor. On the other hand,
the cable shown at the bottom is good in terms of both electrical and magnetic
screenings. As much as 100–110 dB electric and magnetic isolation could be
obtained with this cable if used properly in a circuit.
Undesired currents are induced on the cable screen because of environ-
mental EMI. These currents must not couple internal conductors. The perfor-
mance of the cable screen depends on preventing this coupling. A voltage differ-
ence also occurs along the screen of the cable if there is a magnetic field around.
The amount of this voltage difference depends on the unit-length resistance
and inductance of the cable. If the cable is screenless, then this voltage increases
as the frequency increases and may cause significant noise voltages. This occurs
only up to the cable cut-off frequency if the cable is screened. Electrical screen-
ing is effective above cable cut-off frequency. Magnetic screening becomes effec-
tive for frequencies several times higher than the cut-off frequency. Table 10.9
lists a few cables and their cut-off frequencies.
Figure 10.20 Definition of the cable transfer impedance.

Table 10.9
Cut-Off Frequencies of Some Cables
Cable Type Imp [Ω] fc [kHz] 5fc [kHz] Explanation
RG-6A coaxial 75 0.6 3.0 Double shielded
RG-213 coaxial 50 0.7 3.5
RG-214 coaxial 50 0.7 3.5 Double shielded
Figure 10.21 Single/double shielded cables.
Cable grounding is as critical as cable screening. Usually, single-end

grounding is enough at low frequencies, but when the frequency is high,
grounding at both ends of the cable may be essential. Different grounding ex-
amples are pictured in Figure 10.22. As explained there, wide ground strips
with tight connections supply the best grounding.
Electrical screening is usually achieved via grounding of a screened cable
at the load end. On the other hand, magnetic screening necessitates grounding
at both source and load ends. This also results in 10–20 dB increases in electri-
cal shielding. Twisted pair cables naturally supply good magnetic screening.
Cables act as antennas and suffer from intrinsic resistance, capacitance,
and inductance.
Cables in a system/device should be grouped according to their function,
such as power, analog, digital, and data. Data (sensitive) cables and power (in-
sensitive) cables should be located as far apart as possible. Separate connector
Figure 10.22 Cable grounding examples.
assemblies should be used for analog/digital signals, and they should be located
as far apart as possible.
Lightning and pulsed sources may cause high-energy transients into
cables.
10.4.2 Connectors
Connectors are also important in EMC engineering. A tight connection be-
tween two cables and surfaces mitigate EMI. EM signals penetrate materials
like air, water, light, and sound. Leakage points where air, sound, water, or light
leak should be closed. This may not be enough, because EM waves also pen-
etrate solids where air, sound, water, or light cannot leak.
Discontinuities in system/device, such as caps, input panels, ventilation
holes and apertures, and so on, are inevitable. In such cases, connectors, and
bonding materials (i.e., gaskets, seals) are used to prevent EMI. Tight contact is
critical when connecting cables with each other as well as to a panel or surface.
In addition, wide strips are best for grounding conductors. Better/tight contact
for the doors or windows may be achieved by metal doped plastics, gaskets, and
fabrics.
Corrosion is also important in surface mounts. Galvanic corrosion is an
electrochemical process in which one metal corrodes preferentially to another
when both metals are in electrical contact and immersed in an electrolyte. Dis-
similar metals and alloys have different electrode potentials, and when two or
more come into contact in an electrolyte a galvanic couple is set up. Some met-
als such as gold, silver, copper, and nickel are cathodic; others (e.g., aluminum,
zinc, magnesium, and iron) are anodic. One way in reducing or preventing the
corrosion is to electrically insulate the two metals from each other. This can be
done using plastic or another insulator. Another way is to keep the metals dry
and/or shielded from salts, acids, and so on. Painting, coating, or choosing met-
als with similar potentials or the same metal for all construction also reduces
the corrosion.
Note that BNC connectors developed in Bell Labs back in the 1990s have
50Ω–75Ω characteristic impedances and can be used up to 500–600 MHz
frequencies. SMA connectors extend this range up to 20–25 GHz frequencies.
10.5 Switched-Mode Power Supplies

A switched-mode power supply (SMPS) is a DC-DC and/or DC-AC power
supplier that applies switching principles. It eliminates physically massive power
transformers and other power line magnetics. They are small, light, and cheap,
and they eliminate the 50-Hz component AC mains. The power received from
AC mains is rectified and filtered as high-voltage DC. It is then switched at
a rate of minimum speed of 15 kHz to 50 kHz and a maximum speed of 70
kHz–90 kHz, and fed to the primary side of the step-down transformer. The
step-down transformer is only a fraction of the size of a comparable 50-Hz unit,
thus relieving the size and weight problems. The output at the secondary side
of the transformer is rectified and filtered. Then it is sent to the output of the
power supply.
A disadvantage of a SMPS is that it may be a terrible disturbing EMI
source. Switching means narrow time pulses, which in turn means an infinite
number of harmonics. A typical CE measurement result of 70 kHz SMPS, to-
gether with CISPR 22 limits, is shown in Figure 10.23. As observed, emissions
cover frequencies from 150 kHz up to 30 MHz.
Improper choice of an SMPS may cause severe EMC problems.
10.6 Practical Design Tips

EMC engineers well know that EMC is a problem from design to market and
necessitates system or box level approaches. Digital circuits are more likely to
be the source of emissions due to periodic waveforms and fast clock/switching
rates. Analog circuits are more likely to be the susceptible victims due to higher
gain functions.
In many cases, both EMI source and victim are electronic components.
Active components are usually the sources of EMI; passive components often
Figure 10.23 Emissions from a typical SMPS.
contribute to it, depending on the signal frequencies and component’s charac-

teristics. For example, an inductor can become predominantly capacitive due
to the high-frequency parasitic coupling between windings. A capacitor can
develop parasitic series inductance due to its internal inductance and external
lead inductance at high fundamental and harmonic frequencies.
Active components’ EMI problems can occur because of a device’s input/
output paths at low frequencies, but they may become a direct radiator or re-
ceptor at high frequencies.
CM currents are usually small compared to DM currents, but they can
be the main cause of radiated emissions. Emissions and susceptibility that are
typical in single-layer PCB design can be greatly improved by using multilayer
PC boards with power planes. High capacitance between a forward signal and
its return path (ground plane) provides containment of the electric field. Low
inductance of the paths provides for magnetic flux cancellation. A suggested
PCB design to mitigate EMI is pictured in Figure 10.24. The following are
some important PCB level EMC tips:
• PCB stack-up design is important in creating EM fields, but it provides

additional power bypassing/decoupling and therefore minimizes bus
voltage transients.
• Since the return currents in the power planes are equal and opposite
polarity to the associated signal currents, their EM fields tend to cancel.
Figure 10.24 A typical device/component placement on a PCB.
Power planes can also reduce the loop areas of signal and power traces,
causing a decrease of EMI emissions/susceptibility.
• A ground plane can lower the overall ground impedance and can re-
duce high-frequency ground bounce. Also, the impedance between the
ground and voltage planes is lowered at the high frequencies, and this
reduces power bus ringing.
• The capacitors must be properly selected in their frequency response to
deliver the energy needed. However, as the number of decoupling paths
increase, so do the number of voltage drops across them, and this can
result in power bus transients along with associated CM emissions. This
problem can be minimized with proper power plane design.
• The power plane acts as an effective high-frequency capacitor and, con-
sequently, as an additional energy source needed for cleaner IC outputs.
• The delay for each subsection should be no larger than 1/10th the signal
rise time.
• Multilayer PCBs are better than single-layer PCBS in terms of EMI
mitigation. If a single-layer board must be used, a ground plane should
be utilized to help reduce radiation.
• Top and bottom ground planes can help reduce radiation from multi-
layer boards by at least 10 dB.
• Segmented PC board ground planes are useful for reducing cable radia-
tion due to common mode currents.
• Power and return planes should be located on opposite sides of a multi-

layer PCB. Effective power planes are low in inductance. Therefore, any
transients that may develop on the power planes will be at lower levels,
resulting in lower common mode EMI.
• Analog and digital circuits are susceptible to interaction when located in
close proximity to each other. These should be located on different layers
of the PCB, whenever possible.
• High frequency traces, such as those used for clock and oscillator cir-
cuits, should be contained by two ground planes. This provides for
maximum isolation. The reactance of a trace or conductor can easily
exceed its dc resistance as frequency increases. Additional EMI preven-
tive measures for clock/oscillator traces include the utilization of guard
traces grounded to the ground plane at several locations. The shielding
of clock and oscillator components with foil or small metallic enclosures
may also be needed.
• Overall, circuit crosstalk increases by a factor of two whenever the clock
rate is doubled. EMI radiation and crosstalk may be reduced by mini-
mizing the PC board trace height above the ground plane.
• PCB edge radiation may be the result of traces being located too close to
the board edge. This can be minimized by keeping traces at a distance of
at least three times the board thickness away from the board edge.
• PC board trace stacking should be avoided if possible. Otherwise, it
should be limited to one trace height to reduce radiation, crosstalk, and
impedance mismatches.
• Parallel traces are often susceptible to crosstalk. These should be sepa-
rated by at least two trace widths for crosstalk reduction.
• Two or more ground planes would be better on PCB design above 25–
30 MHz. Slots in the ground and power planes should be avoided. All
heat sinks should be grounded.
10.7 Impedance Matching

Every discontinuity means an impedance difference, which in turn means pow-
er loss and EM emissions (through reflections and diffractions). Impedance
matching is an important problem in EMC engineering. Impedance matching
in a circuit, between the source and load, allows maximum power transfer. It
means an increase in SNR in receiver circuits and receive antenna. It reduces

phase and amplitude errors in antenna arrays.
10.7.1 Lumped Element Circuits

Different source, load, and TL impedances can be matched as shown in Figure
10.25. Here, the load impedance ZL will be matched to the line impedance Z0
via a simple impedance matching circuit that contains two impedances: (a) Z1
= jB parallel to the load and then Z2 = jX on the serial ar,m or (b) (a) Z1 = jX
serial to the load and Z2 = jB on the parallel arm.
The choice of which of these circuits depends on the impedances of the
load and TL. If the load impedance is greater, then Figure 10.25(a) is chosen;
otherwise, Figure 10.25(b) is used. Suppose the load impedance is ZL = RL +
jXL and RL > Z0. As shown in Figure 10.25(a), the equivalent impedance on the
left of the matching circuit must be equal to the characteristic impedance of the
TL; that is:
1
Z 0 = jX + (10.18)
jB + 1 (RL + jX L )
This equation is satisfied only with the following two expressions:
X y ± RL / Z 0 RL2 + X L2 − Z 0RL
B= (10.19)
RL2 + X L2
Figure 10.25 Impedance matching of lumped element circuits.

1 X LZ0 Z0
X = + − (10.20)
B RL BRL
Equation (10.19) yields two solutions for B and X, which means both B
and X may be inductive (positive) or capacitive (negative).
The circuit in Figure 10.25(b) (i.e., for RL < Z0) may be solved in a similar
approach and the following equations may be obtained:
1 1
= jB +
Z0 RL + j ( X + X L ) (10.21)
X = ± RL ( Z 0 − RL ) − X L , (10.22)
( Z 0 − RL ) / RL
B=± (10.23)
Z0
For example, in order to match the load impedance ZL = 200 –j100Ω to

100Ω TL at 500 MHz, normalized (b = B/Z0) and (x = X/Z0) values are found,
respectively, as 0.29 and 1.22. The corresponding real capacitor and inductor
values will be computed as C = b/2πfZ0 = 0.92 pF and L = XZ0/2πf = 38.8 nH.
10.7.2 Scattered Parameter Circuits

High frequency impedance matching (e.g., in PCBs) is done with ML. ML is a
TL; therefore, impedance matching is achieved by using SC and/or OC termi-
nated stubs connected between the load and source in serial and/or in parallel.
Figure 10.26 shows impedance matching with an OC stub connected in paral-
lel. This is called impedance matching with a single stub.
As discussed in Section 4.4, reflections occur on a circuit if source and/
or load impedances are different than the TL characteristic impedance. The
voltage, current, and equivalent impedance at any point along the circuit are
different and periodically repeat themselves on the TL at every half-wavelength.
If the impedances are matched, no reflections occur. As shown in Figure 10.26,
single-stub impedance matching is similar to the lumped element matching
approach shown in Figure 10.25. The load-stub distance (d ) and length of the
stub (l ) are requested. Although not necessary, usually characteristic imped-
ances of the main line and the stub are taken equal. Table 10.10 lists a short
MATLAB script that computes these two parameters for a given load and line
Figure 10.26 TL impedance matching with a single stub.
Table 10.10
Short MATLAB Code for a Single-Stub Impedance Matching
%------------------------------------------------------------------------------------
% Program: ImpedanceMatch.m; Usage: ImpedanceMatch(50,15+10j)
% Result = [distance, length] (in wavelengths)
%-------------------------------------------------------------------------------------
function stub = ImpedanceMatch(Z0,ZL)
if ZL == Z0
disp(‘Load is already matched.’);
else
% Voltage reflection coefficient at the LOAD
GammaL = (ZL-Z0)/(ZL+Z0); thL = angle(GammaL);
bl = thL/2 + [1;-1]*acos(-abs(GammaL))/2;
bd = acot(tan(2*bl-thL)/2);
end
position = bl/2/pi; stublength = bd/2/pi; disp(‘ ‘);
disp(‘Distance and Stub length in wavelengths:’);
stub = mod([position,stublength], 0.5);
end %------------------PROGRAM END------------------------
impedances. If, for example, ZL = 15 + j10Ω load is going to be matched to a

50Ω TL, then all needed is to type ImpedanceMatch (50, 15 + 10j) in the com-
mand line window. The program automatically yields these two parameters. A
typical example is given in Figure 10.27.
10.7.3 Impedance Matching on the Smith Chart

The Smith chart (see Figure 10.28), invented by Phillip H. Smith (1905–1987),
is a graphical aid or nomogram designed for electrical engineers specializing in
RF engineering to assist in solving problems with TLs and matching circuits.
It can be used to simultaneously display multiple parameters including imped-
Figure 10.27 TL impedance matching with a single stub.
Figure 10.28 The Smith chart.
ances, admittances, reflection coefficients, scattering parameters, noise figure

circles, and constant gain contours.
The Smith chart is a graphic showing the relation between voltage reflec-
tion coefficient (Γ) and normalized TL impedance (z = Zy /Z0 = r ± x). It is
plotted on the complex reflection coefficient plane in 2D and scaled with the
normalized impedance.
Polar coordinates on the chart are the amplitude and phase angle of the re-
flection coefficient. The Cartesian coordinates are the real and imaginary parts
of the reflection coefficient. The entire chart lies within the unit circle |Γ| = 1.
The chart also shows curves of normalized constant resistances and constant
reactance. Normalized impedance is defined as the ratio of impedance over
characteristic impedance of the line used. The horizontal axis, which divides
the chart into top and bottom semi-circles, is the normalized resistance axis.
It starts from zero at left (SC point) and goes to infinity at right (OC point).
The circles that are tangent to the right boundary are normalized resis-
tance = constant circles. The center of the chart is 1, where ZL = Z0. On the left,
ZL < Z0; on right, ZL > Z0. The circle that is tangent l to both center point and
right boundary is the loci, where real part of the normalized impedance is 1.
The arcs above the resistance axis (initiated from right) are normalized induc-
tance values (i.e., XL /Z0 = ωL/Z0). The arcs below the resistance axis correspond
to normalized capacitances (i.e., YL /Z0 = 1/ωCZ0). Therefore, normalized com-
plex impedance with (positive) negative imaginary parts are located at the top
(bottom) semi-circle.
The center of |Γ| = constant circle is the origin. The phase angle of the
reflection coefficient may be directly read on the chart measured from the right
end of the resistance axis up to the line connecting the origin to the normalized
impedance point. For a lossless TL, moving from load toward the source along
the line correspond to turning clockwise on the |Γ| = constant circle, or vice
versa. A complete tour on |Γ| = constant circle corresponds to half wavelength.
That is, impedances at λ/2, λ, 3λ/2, 2λ, and so on, away from the load are also
ZL . The point on |Γ| = constant circle, which is symmetric with respect to the
origin, is the normalized admittance at the same point or normalized imped-
ance at λ/4 away. At the origin, the reflection coefficient is zero (VSWR is equal
to 1). The philosophy of impedance matching is to bring any point on the chart
to the center.
Parallel Stub Impedance Matching

Suppose ZL = A – jB will be matched with a TL having characteristic impedance
Z0. Single-stub impedance matching is performed with the following steps:
• Step 1: Mark the normalized load impedance ZL = a – jb on the Smith

chart.
• Step 2: Draw the |Γ| = constant circle. This is the circle with its center
located at the center of the chart, and the radius of the circle is equal to
the distance between the normalized impedance point and the center of
the chart. Moving from the load toward the source along the TL cor-
responds to turning on this circle in the clockwise direction.
• Step 3: Mark the normalized admittance y L = a ′ + jb ′, which is just the
symmetric of normalized impedance with respect to the origin.
• Step 4: Turn on the |Γ| = constant circle and go to the intersection point
between |Γ| = constant circle and the unit circle. The real part of nor-
malized impedance is equal to 1 (z′ = 1 + jc). Read the distance in terms
of wavelength between these two points on the outer scale of the chart.
This is the distance d, where series stub will be connected.
• Step 5: Add an open-ended parallel stub at this point, which has a ca-
pacitive effect of jc. Since parallel admittances are directly added, the
total normalized admittance at this point will then be yt = 1.
• Step 6: The stub is open ended (OC) at one end and jc at the other. The
problem is reduced to find out the length of the stub with infinite nor-
malized admittance at one and jc at the other. Read this length on the
chart as the distance from SC (OC impedance means SC admittance) to
jc in the clockwise direction.
10.7.4 Practical/Useful Information

The following are clues when choosing the right impedance matching circuit:
• Simplicity: Choose the simplest circuit that satisfies the requirements.

• Bandwidth: Beware of bandwidth requirements. Usually, it is easy to
match impedances at a single frequency or in a narrowband. Unfortu-
nately, in practice, impedance matching is usually requested in a broad-
band.
• Practical realization: Watch out for the circuit where impedance match-
ing will be used. It may be a TL, PCB, or a waveguide. Choose the one
that best fits practical realization.
• Connectivity: Watch out for input/output locations and the place of the
impedance matching circuit. Connectivity may also be crucial.
References
[1] Bowick, C., Circuit Design, Newness, Boston, 1982.
[2] Pozar, M. D., Microwave Engineering, Addison-Wesley, Menlo Park, 1990.

[3] Besser, L., and R. Gilmore, Practical RF Circuit Design for Modern Wireless Systems Volume
I, Norwood, MA: Artech House, 2003.
[4] Mathaei, G. L., L.Young and E. M. T. Jones, Microwave Filter, Impedance Matching Net-
works and Coupling Structures, Dedham, MA: Artech House, 1980.
[5] Ludwig, R., and P. Bretchko, RF Circuit Design Theory and Applications, Prentice Hall,
New Jersey, 2000.
[6] Hong, J. G., and M. J. Lancaster, Microstrip Filters for RF/Microwave Applications, John
Wiley & Sons, Inc., New York, 2001.
[7] Uslu, A., and L. Sevgi, “Matlab-Based Filter Design Program: From Lumped Elements to
Microstriplines,” IEEE Antennas and Propagation Magazine, Vol. 53, No. 1, Feb. 2011, pp.
212–224.
[8] Gündüz, S., “Broadband Microstrip Filter Design,” M.S.E.E Thesis, Kocaeli University,
2005.
[9] Gündüz, S., G. Çakır, L. Sevgi, “A Generic Microstrip Structure for the Realization of All-
type Broadband Filters,” Microwave and Optical technology Letters, Vol. 48, No. 12, Dec.
2006, pp. 2390–2393.
[10] Çakır, G., M. Çakır, and L. Sevgi, “A Novel Virtual FDTD-Based Microstrip Circuit
Design and Analysis Tool,” IEEE Antennas and Propagation Magazine, Vol. 48, No. 6,
Dec. 2006, pp. 161–173.
[11] Paul, C. R., Introduction to Electromagnetic Compatibility, John Wiley InterScience (Series
in Microwave and Optical Engineering), New York: John Wiley and Sons, 2006.
York: IEEE Press-John Wiley and Sons, 2003.
[13] Sevgi, L., “Electromagnetic Screening and Shielding Effectiveness Modeling,” IEEE
Antennas and Propagation Magazine, Vol. 51, No. 1, Feb. 2009, pp. 211–216.
[14] IEEE STD -299-1997, IEEE Standard for Measuring the Effectiveness of Electromagnetic
Shielding Enclosures.
[15] Sevgi, L., “Reciprocity: Some Remarks from Field Point of View,” IEEE Antennas and
Propagation Magazine, Vol. 52, No. 2, Apr. 2010, pp. 205–210.
[16] ASTM D-4935, Measuring the Electromagnetic Shielding Effectiveness of Planar
Materials.
Bibliography
Nguyen, C., Analysis Methods for RF, Microwave, and Millimeter-Wave Planar Transmission Line
Structures, New York: John Wiley and Sons, 2000.
Sobol, H., “Applications of Integrated Circuit Technology to Microwave Frequencies,” Proceed-
ings of the IEEE, 1971.
Williams, T. “Emissions Measurements on Switch-Mode Power Supplies,” Technical Note, Elmac

Services, www.elmac.co.uk.
Uslu, A., and L. Sevgi, “Matlab-Based Filter Design Program: From Lumped Elements to Mi-
crostriplines,” IEEE Antennas and Propagation Magazine, Vol. 53, No. 1, Feb. 2011, pp. 213–224.
Bibliography
Suggested Readings
The author has been the associate editor and a writer of the “Testing Ourselves”
column of the IEEE Antennas and Propagation Magazine since 2007. The fol-
lowing tutorials have been published, and several free EM virtual tools have
been shared with international EM community. Some of these are very basic,
but some are highly comprehensive. All of them have been used all around the
world in undergraduate- and graduate-level courses as well as in research.
Sevgi, L., “Electromagnetic Modeling and Simulation: Challenges in Validation, Verification and
Calibration,” IEEE Trans on EMC, Vol. 56, Aug. 2014, pp. 750–758.
Uslu, A., and L. Sevgi, “MATLAB-Based Filter Design Program: From Lumped Elements to Mi-
crostriplines,” IEEE Antennas and Propagation Magazine, Vol. 53, No. 1, Feb. 2011, pp. 213–224.
Eser, S., and L. Sevgi, “Open Area Test Site (OATS) Calibration,” IEEE Antennas and Propagation
Magazine, Vol. 52, No. 3, Jun. 2010, pp. 204–212.
Sevgi, L., “Biostatistics and Epidemiology: Hypothetical Tests on Cell Phone Users,” IEEE An-
tennas and Propagation Magazine, Vol. 52, No. 1, Feb. 2010, pp. 267–273.
Sevgi, L., “Electromagnetic Screening and Shielding Effectiveness (SE) Modeling,” IEEE Anten-
nas and Propagation Magazine, Vol. 51, No. 1, Feb. 2009, pp. 211–216.
Sevgi, L., “Electromagnetic Compatibility Engineering Education: Problems, Challenges and
Perspectives,” (special issue) TJEECS, Turkish Journal of Electrical Engineering and Computer Sci-
ences, Vol. 17, No. 3, Dec. 2009, pp. 273–278.
Sevgi, L., “The Antenna as a Transducer: Simple Circuit and Electromagnetic Models,” IEEE
Antennas and Propagation Magazine, Vol. 49, No. 6, Dec. 2007, pp. 211–218.
289
Sevgi, L., “Digital Multimeters and Basic Measurements,” IEEE Antennas and Propagation Maga-
zine, Vol. 49, No. 4, Aug. 2007, pp.232–237.
Sevgi, L., “Numerical Fourier Transforms DFT and FFT,” IEEE Antennas and Propagation Maga-
zine, Vol. 49, No. 3, Jun. 2007, pp.238–243.
Sevgi, L., “Innumeracy: The Meaning of the Numbers We Use,” IEEE Antennas and Propagation
Magazine, Vol. 49, No. 2, Apr. 2007, pp.195–190.
Sevgi, L., and Ç. Uluıık, “A MATLAB-based Transmission Line Virtual Tool: Finite-Difference
Time-Domain Reflectometer,” IEEE Antennas and Propagation Magazine, Vol. 48, No. 1, Feb.
2006, pp. 141–145.
Sevgi, L., “Transmission Line Fault Analysis Using a MATLAB-Based Virtual Time Domain
Reflectometer,” IEEE EMC Newsletter, Fall 2006, pp. 67–72.
Sevgi, L., “EMC and BEM Engineering Education: Physics based Modeling, Hands-on Train-
ing and Challenges,” IEEE Antennas and Propagation Magazine, Vol. 46, No. 2, Apr. 2004, pp.
140–145.
Suggested Books
The followings are several useful EMC books for the EMC engineers who
would like to go beyond the scope of this book.
Paul, C. R., Introduction to Electromagnetic Compatibility, Hoboken, NJ: Wiley, 2006.

Williams, T., EMC for Product Designers, 4th Ed., Oxford: Elsevier, 2007.
Monrose, M. I., and E. M. Nakauchi, Testing for EMC Compliance: Approaches and Techniques,
New York: Wiley-IEEE Press, 2004.
Kodali, W. P., Engineering Electromagnetic Compatibility: Principles, Measurements, Technologies,
and Computer Models, New York: Wiley-IEEE Press, 2001.
Peterson, A., and G. Durgin, Transient Signals on Transmission Lines: An Introduction to Non-Ideal
Effects and Signal Integrity Issues in Electrical Systems, London: Morgan and Claypool, 2009.
Paul, C. R., Transmission Lines in Digital Systems for EMC Practitioners, New York: Wiley-IEEE
Press, 2012.
Huray, P. G., The Foundations of Signal Integrity, New York: Wiley-IEEE Press, 2010.
Celozzi, S., R. Araneo, and G. Lovat, Electromagnetic Shielding, New York: Wiley-IEEE Press,
2008.
Ulrich, R. K., W. D. Brown, Advanced Electronic Packaging, New York: Wiley-IEEE Press, 2006.
Boxleitner, W., Electrostatic Discharge and Electronic Equipment: A Practical Guide for Designing to
Prevent ESD Problems, New York: Wiley-IEEE Press, 1989.
Monrose, M., EMC and the Printed Circuit Board: Design, Theory, and Layout Made Simple, New
York: Wiley-IEEE Press, 1999.
Bibliography 291
Hemming, L. H., Architectural Electromagnetic Shielding Handbook: A Design and Specification

Guide, New York: Wiley-IEEE Press, 1992.
Hemming, L. H., Electromagnetic Anechoic Chambers: A Fundamental Design and Specification
Guide, New York: Wiley-IEEE Press, 2002.
Baba, Y., and V. A. Rakov, Electromagnetic Computation Methods for Lightning Surge Protection
Studies, New York: Wiley-IEEE Press, 2016
Morrison, R., Grounding and Shielding: Circuits and Interference, New York: Wiley-IEEE Press,
2007.
Kantartzis, N. V., and T. D. Tsiboukis, Modern EMC Analysis Techniques Volume I: Time-Domain
Computational Schemes, Volume II: Models and Applications, London: Morgan and Claypool,
2008.
Morgan, D., A Handbook for EMC Testing and Measurement, IET Electrical Measurement Series,
Stevenage, UK: 1994.
Ott, H., Electromagnetic Compatibility Engineering, New York: John Wiley and Sons, 2009.
Ott, H., Noise Reduction Techniques in Electronic Systems, 2nd Ed., New York: John Wiley and
Sons, 1998.
Suggested Sites
ANSI, American National Standardization Institute, www.ansi.org
CENELEC, European Committee for Electrotechnical Standardization, www.cenelec.eu
CISPR, International Special Committee on Radio Interference, www.iec.ch/cispr
FCC, Federal Communications Commission, www.fcc.gov
EMC Academy, www.emcacademy.org
EMC Information Center, www.compliance-club.com
EMC Banana Skins, www.compliance-club.com/archive/old_archive/Bananaskins.htm
EN Standards, http://ec.europa.eu/growth/single-market/european-standards/
harmonised-standards/
EU EMC Directive, http://ec.europa.eu/growth/single-market/european-standards/
harmonised-standards/electromagnetic-compatibility/
IAF, International Accreditation Forum, www.iaf.nu
ICNIRP, International Commission on Non-Ionizing Radiation Protection, www.icnirp.org
IEC, International Electrotechnical Commission, www.iec.ch
IEC ACEC, Advisory Committee on EMC, www.iec.ch/acec
IEEE, Institute for Electrical and Electronics Engineers, www.ieee.org
ILAC, International Laboratory Accreditation Cooperation, www.ilac.org
ISO, International Standardization Organization, www.iso.org

NIST, National Institute of Standards and Technology, www.nist.gov
RAPEX, Rapid Alarm System, http://ec.europa.eu/consumers/consumers_safety/ safety_prod-
ucts/rapex/alerts/main/?event=main.listNotifications
WADA, World Anti-Doping Agency, www.wada-ama.org
WHO, World Health Organization, www.who.int
About the Author
Levent Sevgi was born in Akhisar, Turkey, in 1958. He graduated from the
Electronics and Communications Engineering Department of Istanbul Tech-
nical University (ITU) and completed his PhD studies with Prof. Leopold B.
Felsen at Weber Research Institute/New York Polytechnic University.
He became an assistant, associate, and full professor in 1991, 1996, and
2002, respectively. In 1990, he returned to the electrical and electronics en-
gineering faculty of the ITU. He was with the Center for Defense Studies,
ITUV-SAM, from 1993 to 1997, for the Long Horizon (UZUN UFUK) Proj-
ect studies for the Turkish Navy. He was with the Scientific Research Group of
Raytheon Systems Canada from September 1998 until Jun 1999, during the
field trials of the Canadian East Coast Integrated Maritime Surveillance Sys-
tem based on surface wave HF radars. He joined TUBITAK-MRC Information
Technologies Research Institute of the Turkish Scientific Research and Technol-
ogy Council as the chair of the Electronic Systems Department in June 1999
and spent nearly two years there. He was also with the Center for Defense Stud-
ies, ITUV-SAM for Vessel Traffic System installation for Turkish Straits from
2000 to 2002. Prof. Sevgi had been with the electronics and communication
engineering department of the engineering faculty in Dogus University from
September 2001 to September 2014. He was a full-time visiting professor at the
University of Massachusetts at Lowell from September 2012 to August 2013.
Since September 2014, he has been with the electrical and electronics engineer-
ing department in the faculty of engineering and architecture, Okan University.
His research study has focused on propagation in complex environments,
analytical and numerical methods in electromagnetic, radar systems, EMC/
EMI modeling and measurement, surface wave HF radars, FDTD, TLM, SSPE
and MoM techniques and their applications, RCS modeling, and bioelectro-
magnetics. He is also interested in novel approaches in engineering education
293
and teaching electromagnetics via virtual tools. He also teaches popular science
lectures in science, technology, and society.
He is a Fellow of the IEEE, a member of Turkish Chamber of Electrical
Engineers (EMO), an associate editor of the IEEE Antennas and Propagation
magazine’s “Testing Ourselves” column, a member of the IEEE Antennas and
Propagation Society Education Committee, a column writer in the IEEE Re-
gion 8 Newsletter’s “Scientific Literacy” column, on the editorial boards of IEEE
Access, ELEKTRIK, the Turkish Journal of Electrical Engineering and Computer
Sciences, Wiley’s RF and Microwave Computer Aided Engineering, the Journal of
Applied Computational Electromagnetics (ACES), the American Journal of Food
Technology, IJAP, the International Journal of Antennas and Propagation, and
Asian Journal of Management.
He is the author or coauthor of more than 175 journal and 100 inter-
national conference papers. He has published several books in English and
Turkish.
His two books, Complex Electromagnetic Problems and Numerical Simula-
tion Approaches and Electromagnetic Modeling and Simulation, were published
by the IEEE Press (Wiley) in 2003 and 2014, respectively.
Index
Absolute errors, 242–43 RE measurement setup, 196
Absorption losses, 266–68 RI measurement setup, 196
AC circuits, 65 use of, 194
Accreditation, 33–49 Antenna arrays
benefits of, 36 ARRAY package, 140–43
certification versus, 36 broadside, 137, 140, 143
as dynamic process, 36 circular, 144
EMC tests and measurements, 40–44 defined, 137
institutions, 36–38 endfire, 137, 140, 143
interlab comparisons, 44–49 illustrated, 137
introduction to, 33–35 with isotropic radiators, 138–40
lab, 41 phased, 137
MLA coverage, 37 planar, 139
philosophy of, 35 radiation patterns, 137–38
proficiency testing, 44–49 Antenna calibration, 116–17
regional bodies, structure of, 38 AF measurement, 117, 126
TÜRKAK, 38–39 antenna pattern measurement (APM),
Accuracy, proficiency testing and, 47, 48 127–31
AC mains, power quality issues of, 173–76 CALTS, 116, 124, 125
Admittance equivalent capacitance substitution
high frequency equivalent, 71 method (ECSM), 117
parameters, 90 horizontal polarization and, 125
Advisory Committee of Electromagnetic performing, 125–27
Compatibility (ACEC), 20 referencce antenna method (RAM), 116
Aliasing, 166 standard site method (SSM), 116
American wire gauge (AWG) standards, 66 three-antenna calibration (TAC)
Ampere’s law, 52 method, 125, 126
Analog-to-digital converter (ADC), 201–2 uncertainty budget for, 131
Anechoic chamber Antenna effective aperture, 111
composite absorbers, 196 Antenna factor (AF), 112–14
illustrated, 196 defined, 112
operation, 195 effective aperture, 113–14
295
Antenna factor (continued) ARRAY package, 140–43

power density, 112 Artificial mains network (AMN), 219
producer-supplied, 115 Asia Pacific Laboratory Accreditation
receive antenna, 112–14 Cooperation (APLAC), 36, 37
transmit (TAF), 114–15 Assessment factor, 12
Antenna pattern Average value, 242
defined, 106
parameters, 109 Bandwidth
Antenna pattern measurement (APM) antenna, 104, 105
combined standard uncertainty, 130 DMM measurement, 204
coverage factor, 130 emission measurements, 228
defined, 127 impedance matching and, 286
expanded uncertainty, 130 resonance circuits, 76
power measurement setup, 128–30 Bioelectromagnetics (BEM) engineering
standard uncertainty, 130 defined, 2
Antennas, 103–52 device-human interaction, 4
bandwidth, 104, 105 EM field-tissue interaction, 4
broadband, 149 introduction to, 1
communication, 103, 107–8 target of, 3
defined, 103 Broadband EMC antennas, 149
dipole-like, 116 Broadside radiation, 137, 140, 143
directivity, 106 Butterworth LC filters
effective aperture, 106 BPF and BSF design, 251–52
electric dipole, 145–47 circuit, 253
EMC, 103, 108–15 components, 254
far-field region, 106 defined, 248
fundamental terms, 105–6 frequency response, 253
gain, 106 HPF transformation, 251
half-power beamwidth, 106 inductor and capacitor of HPF, 251
horn, 151–52 insertion loss, 253
input impedance, 105 MATLAB code for, 250
isotropic radiator, 106, 143 positioning, 254
linear, 140 prototype, 249–50
log-periodic dipole (LPDA), 149–51 See also EMC protection
loop, 132–37
magnetic dipole, 145–47 Cables and connectors
normalized site attenuation (NSA), cable grounding, 275–76
118–31 cables, 274–76
omnidirectional, 106 cable transfer impedance, 274
planar, 140 common-impedance coupling, 273
polarization, 106 connectors, 276–77
ranges, 116 electrical induction, 273
reflection coefficient, 105–6 magnetic induction, 273
RE measurement setup, 103–4 overview, 272–73
simple models, 104–5 single/double shielded cables, 275
transmit, 110 transfer impedance, 273
types of, 143–52 Cables and transmission line model, 77
VSWR, 106, 110–11 Calibration
wire, 148 antenna, 116–17, 125–31
Index 297
equipment, 27–28, 42–43, 115 Commercial EMC standards, 21–22

OATS, 182, 183–90 Common-impedance coupling, 273
standard site, 182 Common mode (CM) current, 86–87, 88,
CALTS, 116, 124, 125, 181 278
Capacitive effect of conductor wire, 68–69 Communication antennas, 103, 107–8
Capacitors, 70, 75 Conducted emissions (CE), 223
CE marking Conducted immunity (CI), 223–24
affixing, 17 Conducted immunity (CI) tests, 232,
defined, 17 233–34
misuse of, 19 Conductor wires, 66–67
process, 17–20 capacitive effect of, 68–69
responsibility for, 18 inductive effect of, 68
rules, 18–19 resistances, 67
as self-certification scheme, 18 Confidence level, proficiency testing and, 48
steps, 19 Conformity Assessment Boards (CABs),
tests and measurements and, 42–43 38–39
CE measurements, 44, 228, 229–30 Conformity Europe (CE) mark, 2
Certification, 36 Connectors, 276–77
CE thresholds, 24 Correlated signals, 156
Characteristic impedance, 78–79 Corrosion, 276
Circuit model, 65–100 Coupling, 3
cables and transmission line model, Coverage factor, 243
77–83 Crosstalk, 99–100
common mode and differential mode Current-injection clamps, 218–19
currents, 86–87 Current probes, 218–19
crosstalk, 99–100 Current waves, 77, 80
grounding, 83–86 Cylindrical waves, 55
lumped parameter circuit elements,
65–70 DC circuits, 65
microstripline circuits, 95–99 Differential mode (DM) current, 86–87,
nonlinearity effects, 87–89 88, 278
resonance circuits, 74–76 Digital multimeters (DMMs)
two-port circuit definitions, 71–73 4-digit, technical specs, 208
two-port circuits and S-parameters, accuracy, 208
89–95 aperture, 207
Circuits classification, 202
capacitive effect, 68–69 DC current measurement, 209
conductor wires, 66–67 DC voltage measurement noise, 203
elements, 65–70 defined, 201, 205
IL versus frequency, 73 error sources, 204
inductive effect, 68d loading effect, 204
lumped element, 281–82 measurement bandwidth, 204
microstripline, 95–99 noise-limited detection, 202–4
realistic R/L/C models, 69–70 resolution, 207
resonance, 74–76 selection, 210
scattered parameter, 282–83 source resistance and, 210
S-parameters and, 89–95 specs and, 205–10
two-port definitions, 71–73 temperature and, 205
Circular polarization, 57 See also EMC tests and measurements
Combined standard uncertainty, 130, 243
Discrete Fourier transform (DFT), 163–71 power density, 57

computation of, 168–69 propagation modes, 59
of continuous time signal, 167 reflection, 54, 55
defined, 165 refraction, 54, 55
performing, 166 scattering, 53–54
Distortion factor, 174 skin depth and, 61
Div operator, 53 Electromagnetic compatibility. See EMC
Electromagnetic model, 51–64
EC EMC directive, 40 electrical engineering theories and,
Eigenfunctions, 55 51–52
Eigenvalues, 55 electric and magnetic dipoles, 61–62
Electrical engineering theories, 51–52 EM coupling mechanisms, 63–64
Electrical fast transients (EFTs), 240 EM materials and skin depth, 59–61
Electrical induction, 273 EM scattering, diffraction, and
Electrical screening, 269 propagation, 53–59
Electric dipole, 61–62, 86 Maxwell equations and, 52–53
Electric dipole antenna, 145–47 typical emissions, 62–63
Electric fields, 6–7 Electromagnetic signals, 155–56
defined, 53 Electrostatic discharge (ESD)
spatial variations of, 52 defined, 238
See also Electromagnetic (EM) fields pulses, 239
Electromagnetic (EM) fields tests, 238–39
electric field, 6 typical human levels, 238
in environment, 5–16 EMC
human interaction, 11, 13 behaviors of devices and systems, 29
human tissue penetration, 6 commercial standards, 21–22
magnetic field, 6 defined, 2
penetration of materials, 63 device-device interaction, 3
power absorption from, 7–8 EU directives, 16–17, 21
public concern, 10 limiting values, 23–26
radiation part, 5–6 military standards, 22–23
risk assessment and precautionary overview, 2–4
principle, 12–13 potential solutions, 4
role of, 1 problem elements, 2
simple calculations, 14–16 requirements, 2
time of exposure versus body currents, standards, 20–23
16 typical scenario, 3
time-varying, 7 EMC antennas
Electromagnetic (EM) levels, in antenna effective aperture, 111
environment, 8–10 antenna factor (AF), 112–14
Electromagnetic (EM) signals, 2 broadband, 149
Electromagnetic (EM) spectrum, 5 overview, 108–9
Electromagnetic (EM) waves receive, 109–14
circular polarization, 57 uses of, 103
cylindrical, 55 VSWR, 110–11
diffraction, 54 See also Antennas
linear polarization, 56–57 EMC engineering
penetration of materials, 60 approach selection, 31
plane, 57 cables and connectors, 272–77
polarization, 56–57 complexity of, 4
Index 299
field as black magic, 29–30 network analyzer and, 220

impedance matching, 280–86 open area test site (OATS), 181–90
interference source identification in, 62 radiated, 27
introduction to, 1–2 reporting and product certification,
noise and, 156 43–44
philosophy, 28–29 reverberation chamber, 198–99
practical design tips, 277–80 risk analysis in, 224–26
suggested approach, 29–31 screened room and resonance effects,
switched-mode power supplies (SMPS), 190–94
277, 278 spectrum analyzer, 210–15
target of, 2 spurious measurements, 240
tests and measurements, 26 surge and flicker tests, 237–38
EM coupling mechanisms, 63–64 TEM/GTEM cell, 197–98
EMC protection, 247–86 traceability, 40–41, 223
Butterworth filters, 247–54 EMI mitigation, 4
ferrites, 260–62 EMI receivers, 215–18
filtering, 247–60 defined, 216
MATLAB-based filter package, 257–60 detector types, 227
ML transformation, 256–57 display, 217, 218
screening/shielding, 262–72 parameters, 217
tips, 272 Rohde and Schwarz, 217
transmission line filters, 254–56 Emission measurements
EMC tests and measurements, 223–44 bandwidths and measurement times,
accreditation and, 40–44 228
anechoic chamber, 194–96 basic factors, 227–28
basic factors in, 227–28 CE, 226, 229–30
calibration, 27–28, 42–43 detector types, 227
components of, 26, 223 parts of, 223
conducted, 27 performing, 228–32
current probes and current-injection RE, 228, 230–32
clamps, 218–19 setup, 41
devices, 201–20 Emissions, 3, 31
digital multimeters (DMMs), 201–10 conducted (CE), 223
electrical fast transients (EFTs), 240 from electric dipole, 62
electrostatic discharge (ESD) tests, radiated (RE), 223
238–39 from SMPS, 278
EMI receiver, 215–18 standards, 224
emission, 27, 41 tests and measurements, 27, 41
emission measurements, 226–32 typical, 62–63
environments, 181–99 EM materials and skin depth, 59–61
error analysis and uncertainty, 240–44 EM point and line sources, 55–56
factors of, 26 EM problem group, 57
features of, 40–42 Endfire radiation, 137, 140, 143
harmonic measurements, 236–37 Energy signals, 161
immunity, 27 Environment
immunity/susceptibility tests, 232–36 electromagnetic levels in, 8–10
important steps for, 228 EMC tests and measurements, 181–99
line impedance stabilization network Equipment under test (EUT), 43
(LISN), 219–20 Equivalent capacitance substitution method
measurement receivers, 210–18 (ECSM), 117
Error effects, 85
absolute, 242–43 function of, 83–86
analysis, 240–44 good, tips for, 85
defined, 241 ground categories, 83–84
propagation, 242–43 lightening effects and, 84
random, 241–42 myths, 86
uncertainty and, 242–43 problem, 84, 85
EU EMC directives, 16–17, 21 wire, 84–85, 86
European Accreditation (EA), 36, 37
Expanded uncertainty, 130, 243 Half-power beamwidth, 106
Extremely low frequency (ELF) radiation, 5 Harmonic distortions, 174, 176
Harmonic generation, 88, 89, 171–72
Faraday’s law, 52 Harmonics
Far end crosstalk (FEXT), 99 analysis, 178
Ferrites, 260–62 current and voltages effects, 179
Filters defined, 169, 171
Butterworth LC, 248–54 effects of, 173
design parameters, 248 measurement of, 236–37
ferrite, 260–62 measurements, 177–78
transmission line, 254–56 Horn antennas, 151–52
types of, 247
See also EMC protection IEEE/ANSI ASC-C63 EMC standards, 23
Flicker Immunity, 3, 31
defined, 175, 237 conducted (CI), 223–24
tests, 237–38 measurements, 27, 43
Fourier coefficients, 161 radiated (RI), 224
Fourier series (FS), 163–71 standards, 224
of a function, 162 Immunity tests, 27, 43
rectangular pulse representation, 165, conducted, 232, 233–34
171 difficulty in performing, 232
summation, MATLAB code for, 163 performing, 236
trapezoidal pulse representation, 164 radiated, 232, 234–36
Fourier transform (FT) types of, 232
defined, 160–61 Impedance
discrete (DFT), 163–71 dipole, 147
for energy signals, 161 high frequency equivalent, 72
numerical computation difficulties, 165 in lumped element model, 71
performing, 166 mismatch, 81
Frequency analysis, 160 parameters, 90
Impedance matching
Gain, 106 bandwidth and, 286
Gain pattern, 106 connectivity and, 286
Galvanic corrosion, 276 importance, 280
Gauss’s law, 52 lumped element circuits, 281–82
Gigahertz transverse electromagnetic parallel stub, 285–86
(GTEM) cells, 197–98 practical realization, 286
Green’s function problem, 55–56, 143 scattered parameter circuits, 282–83
Grounding, 83–86 simplicity and, 286
cable, 275–76 single-stub, MATLAB code, 283
Index 301
on Smith chart, 283–85 measurement setup, 135

transmission line (TL), 283, 284 shapes, 135
Inductive effect of conductor wire, 68 standards, 132
Inductors use of, 132
high frequency equivalent, 71 See also Antennas
transfer characteristics, 75 Lorentz force laws, 52
Infrared radiation (IR), 5 Low-frequency magnetic field coupling, 7
Inter-American Accreditation Cooperation Low Voltage Directive (LVD), 17
(IAAC), 36, 37 Lumped element circuits, 281–82
Interference, 156 Lumped parameter circuit elements, 65–70
Interlab comparisons, 44–49
Internal products (IPs), 88, 89 Magnetic dipole, 61–62, 86
International Accreditation Forum (IAF), Magnetic dipole antenna, 145–47
36, 37 Magnetic fields, 6–7, 15–16
International Bureau of Weights and defined, 53
Measures (IBWM), 43 time variation of, 52
International Committee on Non-Ionising See also Electromagnetic (EM) fields
Radiation Protection (ICNIRP), Magnetic induction, 273
8–9 Magnetic screening, 269
International Electrotechnical Commission Matched termination, 91
(IEC), 20–23 MATLAB-based filter package, 257–60
International Laboratory Accreditation Maxwell equations, 52–53
Cooperation (ILAC), 36 Measurements. See EMC tests and
ISO/IEC 17025, 44–49 measurements
Isotropic radiator, 106 Microstripline (ML) transformation, 256–57
Microstripline circuits
Kuroda identities, 254, 255 basic, 98–99
as bent, 98
Labs characteristic impedance versus
accreditation, 41 frequency, 96–97
interlab comparisons, 44–49 characteristics of, 95–97
performance of, 45 defined, 95
proficiency testing, 44–49 illustrated, 96, 99
QP performance, 45 MATLAB code for, 97
Laser hazards, 6 third-order, 98, 99
Least significant bit (LSB), 201–2 as two-port element, 98
Linear polarization, 56–57 Microwave (MW) radiation, 5
Line impedance stabilization network Military EMC standards, 22–23
(LISN), 219–20 Multilateral agreements (MLAs), 36–37
Load, 71 MWFilterDesigner, 257, 260
Loading error, 204
Log-periodic dipole antennas (LPDA), National Metrology Institutes (NMIs), 43,
149–51 116
Loop antennas National Toxicology Program (NTP), 11–12
calibration, 135–37 Near end crosstalk (NEXT), 99
RE101 test setup, 133 Network analyzer, 220
RS101 test setup, 133, 134 Noise
RS101 verification setup, 134 common-mode, 175
screening effectiveness (SE) defined, 156
Noise (continued) Phase noise, 214, 215

differential-mode, 175 Plane waves, 55
EMC engineering and, 156 Power density, 14, 15, 57, 112
phase, 214, 215 Power quality (PQ)
reduction or mitigation, 159 of AC mains, 173–76
SNR, 157–58 defined, 171
sources, 157 load source interactions and, 173
spectra, 157 Precautionary principle, 12–13
thermal, 156 Precision, proficiency testing and, 47, 48
as uncorrelated, 156 Printed circuit board (PCB)
Noise-limited detection, 202–4 design tips, 278–80
Nonlinearity effects, 87–89 device/component placement, 279
Normalized site attenuation (NSA) edge radiation, 280
2-ray method, 118–19 multilayer, 279, 280
calibration steps, 120 stack-up design, 278
defined, 118 Proficiency testing, 44–49
frequency versus, 121, 122, 123 accuracy, 47, 48
MATLAB calculations, 120, 121 confidence level, 48
measurements, 123 defined, 44
in OFTS adequacy determination, 118 format, 44–45
single-point measurement, 124 precision, 47, 48
theoretical calculations, 118–23 results, 45, 46
values, 122 suspicious values, 46
See also Antennas See also Labs
Normal mode rejection (NMR), 207 Propagation constant, 79
Propagation error, 242–43
OATS, 41, 42
Ohm’s law, 65 Radiated emission (RE), 223
Open area test site (OATS), 181–90 Radiated immunity (RI), 224
calibration, 182, 183–90 Radiated immunity (RI) tests, 232, 234–36
calibration steps, 185 Radio frequency (RF) radiation, 5
defined, 181 Random error, 241–42
diagram of, 184 RAPEX, 34
equipment used during calibration, 184 Receive antenna factor, 109–14
in free-space AF determination, 182 Reference antenna method (RAM), 116
in immunity tests, 181 Reflection, 54, 55
location of, 181 Reflection coefficient, 80–81, 105–6
measurements, 185–86 Reflection losses, 264–65
NSA measurement table, 187 Refraction, 54, 55
NSA versus frequency, 188–89 RE measurements, 41, 228, 230–32
photo, 183 Resistors, 70
sketch, 182 Resonance circuits, 74–76
standard site calibration, 182 bandwidth, 76
See also EMC tests and measurements losses, 76
Overshoot, 170 parallel, 74, 75, 76
quality factor, 76
Parallel stub impedance matching, 285–86 serial, 74, 75, 76
Parasitic signals, 241 shape factor, 76
Phased arrays, 137 RE thresholds, 25
Reverberation chamber, 198–99
Index 303
Richard’s transformation, 254 Signal integrity, 31

Ringing, 170 Signals, 155–56
Risk, 12 EM, 2
Risk assessment energy, 161
assessment factor, 12 parasitic, 241
defined, 12 Signal-to-noise ratio (SNR), 157–58
EMC tests and measurements and, Smith chart, 283–85
224–26 S-parameters
precautionary principle and, 12–13 calculations, 92
R/L/C models, 69–70 defined, 91
Rot operator, 53 definitions of two-port circuit, 91
MATLAB code for, 92, 93, 94
Safety ground, 83–84 of multiple devices, 95
Scalloping loss, 167 Specific absorption rate (SAR), 10, 12
Scattered parameter circuits, 282–83 Spectrum analyzers (SAs)
Screened room, 190–94 defined, 210
aluminum tent, 191 frequency readout error, 213
carbon-loaded absorbers, 194 frequency reference accuracy, 214
in data security, 190 harmonic suppression ratio, 214
defined, 190 main window, 215
dominant mode, 194 parameters, 211–12
inside/outside isolation, 190 phase noise, 214, 215
MATLAB code calculating resonances Rohde and Schwarz, 212, 213
in, 193 selectivity, 214, 216
metal construction, 192 spectral purity, 214, 215
RE measurements in, 195 technical specs, 211
resonance effects, 191–94 See also EMC tests and measurements
Shielding Spurious measurements, 240
absorption losses, 266–68 Standard deviation, 242
approaches, choosing, 263 Standard field method (SFM), 117
defined, 262 Standard site calibration, 182
electrical screening, 269 Standard site method (SSM), 116
magnetic screening, 269 Standard transmitting loop method (STLM),
occurrence, 268 117
in real life, 269–72 Standard uncertainty, 130, 243
Shielding effectiveness (SE) Surge tests, 237
ASTM standards, 269 Susceptibility, 3
defined, 262–63 Swell, 174–75
FDTD package, 269 Switched-mode power supplies (SMPS),
frequency versus, 267, 270, 271 277, 278
ideal case, 264–69
MATLAB code for calculations, 266 TEM/GTEM cell, 197–98
modeling, 63 Tests. See EMC tests and measurements
performance, 268 Third order internal products (TOIP), 88,
positive, 270 899
reflection and transmission losses and, Three-antenna calibration (TAC) method,
265 125, 126
reflection losses, 264–65 Time-varying electromagnetic (EM) fields, 7
values, 263, 268 Total harmonic distortion, 176
Signal grounds, 83–84 Toys Safety Directive (TSD), 17
Traceability, 40–41 Two-port circuits

Transfer impedance, 273 definitions, 71–73
Transmission line (TL) impedance parameters for, 90
characteristic impedance, 78–79 S-parameters and, 89–95
field distributions, 79 See also Circuits
filters, 254–56
impedance matching, 283, 284 Ultraviolet (UV) radiation, 6
mathematical relations, 77 Uncertainty, 130–31, 242–43
model, 77–83 Uncorrelated signals, 156
primary parameters, 78 Unit under test (UUT), 43
propagation constant, 79
secondary parameters, 78 Visible light radiation, 5
short-/open-circuit terminations, 82 Voltage distortion, 174
theory, 77 Voltage reflection coefficient, 80–81
voltage reflection coefficient, 80–81 Voltage regulation, 174
voltage standing wave ratio (VSWR), Voltage standing wave ratio (VSWR), 81–83,
81–83 106, 110–11
Transmit antenna factor (TAF), 104, 114–15 Voltage waves, 77, 80
Transverse electric (TE) mode, 59
Transverse electromagnetic (TEM) cells,
Wave impedance, 57
197–98
Wire antennas, 148
Transverse electromagnetic (TEM) mode, 59
World Anti-Doping Agency (WADA), 33,
Transverse magnetic (TM) mode, 59
34
TÜRKAK, 38–39

A Practical Guide To EMC Engineering (PDFDrive)

Uploaded by

Copyright:

Available Formats

A Practical Guide To EMC Engineering (PDFDrive)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

A Practical Guide To EMC Engineering (PDFDrive)

Uploaded by

Copyright:

Available Formats

A Practical Guide to EMC Engineering

For a complete listing of titles in the

British Library Cataloguing in Publication Data

Cover design by John Gomes

ISBN 13: 978-1-63081-383-3

© 2017 ARTECH HOUSE

1.1 Electromagnetic Compatibility 2

1.2 EM Fields in Our Environment 5

1.3 EU EMC Directives 16

1.4 CE Marking Process 17

1.5 EMC Institutions and EMC Standards 20

1.6 EMC Limiting Values 23

1.7 EMC Tests and Measurements 26

1.8 EMC Engineering Philosophy 28

1.9 Suggested EMC Approach 29

2.3 Accreditation Institutions 36

2.5 EMC Tests and Measurements 40

2.6 Proficiency Testing and Interlab Comparisons 44

3.1 Basic Electrical Engineering Theories 51

3.2 Maxwell Equations 52

3.3 EM Scattering, Diffraction, and Propagation 53

3.4 EM Materials and Skin Depth 59

3.5 Electric and Magnetic Dipoles 61

3.6 Typical Emissions 62

3.7 EM Coupling Mechanisms 63

4.1 Lumped Parameter Circuit Elements 65

4.2 Two-Port Circuit Definitions 71

4.3 Resonance Circuits 74

4.4 Cables and Transmission Line Model 77

4.6 Common Mode and Differential Mode Currents 86

4.7 Nonlinearity Effects 87

4.8 Two-Port Circuits and S-Parameters 89

4.9 Microstipline Circuits 95

5 Antennas and Antenna Calibration 103

5.1 Fundamental Antenna Terms 105

5.2 Communication Antennas 107

5.3 EMC Antennas 108

5.3.1 Receive Antenna and Antenna Factor 109

5.4 Antenna Calibration 116

5.5 Normalized Site Attenuation 118

5.6 Loop Antennas 132

5.7 Loop Antenna Calibration 135

5.8 Antenna Arrays 137

5.9 Antenna Types 143

6 Noise and Frequency Analysis 155

6.1 Fundamental Electromagnetic Signals 155

6.2 Noise 156

6.3 Frequency Analysis and Fourier Transforms 159

6.4 Discrete Fourier Transform 163

6.5 Harmonic Generation 171

6.6 AC Mains and Power Quality 172

6.6.4 Effects of Harmonic Currents and Voltages 179

7 EMC Test and Measurement Environments 181

7.1 Open Area Test Site 181

7.2 Open Area Test Site Calibration 183

7.3 Screened Room and Resonance Effects 190

7.4 Anechoic Chamber 194

7.5 TEM/GTEM Cell 197

7.6 Reverberation Chamber 198