You are on page 1of 54



IEEE Guide on Surge Testing for

3quipment Connected to
Low-Voltage AC Power Circuits






Published by The Institute of Electrical and Electronics Engineers, Inc 345 East 47th Street, New York, NY 10017, USA
July J l . 1987


ANSI/IEEE C62.45-1987

An American National Standard

IEEE Guide on Surge Testing for
Equipment Connected to
Low-Voltage AC Power Circuits


Surge Protective Devices Committee of the

IEEE Power Engineering Society

Institute of Electrical and Electronics Engineers, Inc

National Electrical Manufacturers Association
Approved September 19, 1985

IEEE Standards Board

Approved October 29, 1986

American National Standards Institute

Copyright 1987 by

The Institute of Electrical and Electronics Engineers, Inc

345 East 47th Street, New York, NY 10017, USA
No part of this publication m a y be reproduced i n a n y form,
in a n electronic retrieval system or oth.erwise,
without the prior written permission of the publisher.

IEEE Standards documents are developed within the Technical Committees of the lEEE Societies and the Standards Coordinating Committees of the IEEE Standards Board. hlembers of the committees serve
voluntarily and without compensation. They are not necessarily members of the Institute. The standards developed within IEEE represent
a consensus of the broad expertise on the subject within the Institute
as well as those activities outside of IEEE which have expressed an interest in participating in the development of the standard.
Use of an IEEE Standard is wholly voluntary. The existence of an
IEEE Standard does not imply that there are no other ways to produce, test, measure, purchase, market, o r provide other goods and services related t o the scope of the IEEE Standard. Furthermore, the viewpoint expressed a t the time a standard is approved and issued is subject
to change brought about through developments iii the state of the art
and comments received from users of the standard. Every IEEE Standard is subjected t o review a t least once every five years for revision
o r reaffirmation. When a document is more than five years old, and has
n o t been reaffirmed, it is. reasonable t o conclude that its contents,
although still of some value, d o not wholly reflect the present state of
the art. Users are cautioned t o check t o determine that they have the
latest edition of any IEEE Standard.
Comments for revision of IEEE Standards are welcome from any
interested party, regardless of membership affiliation with IEEE. Suggestions for changes in documents should be in the form of a proposed
change of text, together with appropriate supporting comments.
Interpretations: Occasionally questions may arise regarding the meaning o f portions of standards as they relate t o specific applications. When
the need for interpretations is brought to the attention of IEEE, the
Institute will initiate action to prepare appropriate responses. Since
IEEE Standards represent a consensus of all concerned interests, it is
important to ensure that any interpretation has also received the concurrence of a balance of interests. For this reason IEEE and the members of its technical committees are not able to provide an instant response to interpretation requests except in those cases where the matter
has previously received formal consideration.
Comments on standards and requests for interpretations should be addressed to :
Secretary, IEEE Standards Board
345 East 47th Street
New York, N Y 10017

(This Foreword is not a p a r t of ANSI/IEEE C62.45-1987,IEEE Guide on Surge Testing for Equipment Connected t o Low-Voltage
AC Power Circuits.)

This document, intended to serve as a guide to appropriate surge test methods, covers a range of
diverse tests, each aimed at a particular aspect of design, quality control, acceptance, troubleshooting,
and so forth. The suggestions provided herein should not be construed as mandatory nor should they be
blindly applied to all types of equipment. The specifcation of performance or acceptance tests for any
particular type of equipment properly remains the prerogative of manufacturers, users, and standardwriting groups involved in the particular equipment.
To provide further assistance with applications, the present document offers guidance on surge testing
methods and includes specific surge test plan recommendations. The intent is to provide background
that can help determine whether specific equipment or a circuit has adequate withstand capability (that
is, maintains undisturbed operation, or merely is not damaged, depending on the criteria selected by the
user) in the surge environment defined in ANSI/IEEE C62.41-1980, IEC Pub 664-1980, or any other
applicable document specifylng the power-line surge environment. While lightning is a major source of
transient disturbances, it should be understood that there are many additional kinds of power disturbances, such as surges with extremely rapid rates of rise that are generated by other loads that may be
operating inside a building. Fast rising surges may require special considerations in addition to those
discussed in this guide. Manufacturers of sensitive equipment should consider additional tests to ensure
reliable performance of their instruments in actual operating conditions.
Because of the complexity of the subject, the variations in equipment tested, and the difficulty in
interpreting test results, the necessity for the reader to review and understand the entirety of this
document prior to commencing equipment tests cannot be over-emphasized.
The Summary given after this Foreword is intended only for a rapid overview and therefore is not
included as part of the guide.
Suggestions for improvement of this guide will be welcome. They should be sent to the Secretary, IEEE
Standards Board, Institute of Electrical and Electronics Engineers, 345 East 47th Street, New York, New
York 10017.
The Accredited Standards Committee C62 on Surge Arresters had the following membership when it
reviewed and approved this document:

E. J. Cohen, Vice Chuirman

J. L. Koepfinger, Chairman

John A. Gauthier,Secretary
Organization Represented
Association of American Railroads.. ..............................................................
Bonneville Power Administration .................................................................
Rural Electrification Administration ..............................................................
Electric Light and Power ........................................................................

Institute of Electrical a n d Electronics Engineers

National Electrical Manufacturers Association

Exchange Carriers Standards Association

Underwriters Laboratories





Canadian Standards Association .................................................................

Members-at-Large ........................

Name of Representative
R. W. McKnight
E. J. Yasuda
E. J. Cohen
R. A. Jones
W. R. Ossrnan
J. W. Wilson
W. J. Lannes
J. P. Markey (Alt)
J. L. Koepfinger
D.E. Hedrnan
S. S. Kershaw, J r
G. L. Gaibrois
E. J. Adolphson (Alt)
J. J. Keane (Alt)
R. D. Ball
S. W. Law
D. W. Lenk
M. G. Comber
A. Sweetna
M. Craddock (Alt)
L. H. Sessler
M. Parente
0. J. Gusella, J r (Alt)
P. Notarian
R. W. Seelbach (Alt)
D. M. Smith
J. Osterhout
F. D. Martzloff

At the time this guide was completed, the Working Group on Surge Characterization on Low-Voltage
Circuits had the following membership:

F. D. Martzloff, Chairman
D. W. Boehm
B. Braskich
J. Cawley
G. S. Clarke
E. J. Cohen
H. E. Foelker
D. A. Fuhrman
P. A. Goodwin
R. M. Henry
W. W. Hines

W. T. Rhoades
P. Richman
L. Shulman
P. D. Speranza
R. B. Standler
L. D. Sweeney
D. P. Symanski
M. Tetreault
S. J. Tharp
L. Williams

P. Jedlicka
C. J. Kawiecki
J. L. Koepfinger
E. Livermore
R. C. Mierendorf
W. Milwitt
J. J. Napiorkowski
V. B. Nolan
R. Odenberg
M. Parente

Other individuals who have contributed review and comments are:

V. C. Barnum
W. K. Boice
E. Cooper
S. Demircioglu
B. M. Epstein
A. J. Eriksson

C. L. Fisher
G. J. Hahn
B. W. Hayward
E. K. Howell
W. L. Jameson

H. J. Levine
J. K. Nelson
S. A. Potocny
E. E Vance
C. R. Wetter
E. A. Zivanov

The following persons were on the balloting committee that approved this document for submission to
the IEEE Standards Board:
E. J. Yasuda
R. D. Ball
J. A. Hetrick
L. S. Baker
C. L. Ballentine
G. A. Baril
F. G. Berg
E. W. Boehne
G. D. Breuer
J. J. Burke
E. J. Cohen
D. C. Dawson
R. W. Flugum
H. E. Foelker

G. L. Gaibrois
E. A. Goodman
C. D. Hansel1
G. S. Haralampu
D. E. Hedman
A. R. Hileman
W. W. Hines
D. W. Jackson
S. S. Kershaw
J. L. Koepfinger
J. A. Mambuca
E. H. Marrow
F. D. Martzloff

D. J. Melvold
J. J. Napiorkowski
R. Odenberg
W. R. Ossman
J. C. Osterhout
M. Parente
S. A. Potocny
P. Richman
P. D. Speranza
K. B. Stump
A. Sweetana
E. R. Taylor
A. C. Westrom
S. G. Whisenant

When the IEEE Standards Board approved this standard on September 19,1985,it had the following

John P. Riganati, Vice Chairman

John E. May, Chairman

Sava I. Sherr, Secretary
James H. Beall
Fletcher J. Buckley
Rene Castenschiold
Edward Chelotti
Edward J. Cohen
Paul G. Cummings
Donald C. Fleckenstein

*Member emeritus

Jay Forster
Daniel L. Goldberg
Kenneth D. Hendrix
Irvin N. Howell
Jack Kinn
Joseph L. Koepfinger *
Irving Kolodny
R. F. Lawrence

Lawrence V. McCall
Donald T. Michael'
Frank L. Rose
Clifford 0. Swanson
J. Richard Weger
W. B. Wilkens
Charles J. Wylie

(This Summary is not a p a r t of ANSI/IEEE C62 45-1987, IEEE Guide on Surge Testing for Equipment Connected t o Low-Voltage
AC Power Circuits.)

This document provides guidance on surge testing ac power interfaces of equipment connected to
low-voltage circuits that are subject to transient overvoltages. Signal and data lines are not addressed in
this document, nor are any specifications stated on the withstand levels that might be assigned to specific
equipments. An important objective of the document is to call attention to the safety aspects of surge
The power-line surge environment is described in standards and other documents as either a statistical
statement of measured occurrences or as a goal to be achieved through a coordinated control of the
overvoltages, involving interface devices. ANSI/IEEE C62.41-1980, Guide for Surge Voltages in LowVoltage AC Power Circuits, describes the power-line surge environment in which all equipment, with or
without protection, connected to an uncontrolled ac power line must operate; IEC Pub 664-1980, Insulation Co-ordination within Low-Voltage Systems, Including Clearances and Creepage Distances for
Equipment, sets forth a group of installation categories where the surge voltages are assumed to be
controlled in a descending staircase from the service entrance to the most remote equipment in a
building. The environment description of ANSIAEEE C62.41-1980 provides guidance on the waveshapes
of voltage as well as current surges that can be selected as representative of the environment. The
recommendations of IEC Pub 664-1980 are directed toward the coordination of insulation.
The phrase equipment connected to ac power lines covers a wide range of devices and circuits, for
which the need for surge testing can also vary over a wide range. Electromechanical devices tested only
for insulation can often be dealt with by a simple dielectric withstand test. Electronic circuits and
systems, however, may often require additional test complexity. It is important to recognize that specific
test levels must be defined by a consensus process involving both designer and user and not by the
present guide, which can only define procedures.
Planning a surge test program is a complex task involving many considerations, from the nature of the
equipment and purpose of the test to procedures and instrumentation.
Major issues to be recognized include:
(1) Outcome of the test and failure criteria: upset or damage?
(2) Type of tests: Design tests involve pushing stresses to the limits while production tests are more
likely to be pass/fail types. Diagnostic tests involve reproduction of field failures in the laboratory.
(a) Unpowered testing (simple dielectric withstand) versus powered testing, which is unavoidable
when testing for upset or for failure effects.
(b) Test schedule, including a range of stresses rather than just maximum ratings in order to avoid
blind spots, and the phase angle of the surge with respect to the power line sine wave. Limiting stresses is
required when repetitive tests are performed in order to obtain meaningful results.
(3) Appropriate selection of a voltage wave or of a current wave, or both, depending on the nature of
the test piece and on the expected pasdfail criteria.
(4) Implementation of the test:
(a) Selection of the appropriate test generator
(b) Coupling of the surge to the test circuit
(e) Decoupling of the surge from the ac power supply
(d) Points of application (coupling modes) of the surges
(e) Monitoring and measurements
( f ) Safety precautions in the facilities
Since many of these considerations involve detailed discussions, the document has been organized in
sections dealing with the major points of these considerations. References are made to detailed notes,
contained in the appendix, to provide a more comprehensive treatment of the issues.



1. Scope ......................................................................................................

2. Definitions .......

itions ..........................................
Special Word Usage ...............................................................................


3. References and Document in Preparation .............................................................

3.1 References .........................................................................................
Document in Preparation ........................................................................


4 . Planning of Surge Testing: Basic Objectives............................................................

Surge Environment ...............................................................................
Types of Tests .....................................................................................
Results and Consequences of the Test ..........................................................
Unpowered Testing versus Powered Testing ...................................................
4.5 Withstand Levels ..................................................................................
Waveshape, Voltage, Current .....................................................................
Safety ..............................................................................................


5. Implementation of Surge Testing: Test Equipment ...................................................

General .............................................................................................
Surge Generators ..................................................................................
Point of Test Surge Application ..................................................................
Coupling the Surge to the Equipment Under Test .............................................
Monitoring the Equipment Under Test ..........................................................



6 . Performance of Surge Testing: Test Procedures ....................................................... 17

General ............................................................................................. 17
Limiting Stresses ..................................................................................
Nature of the Equipment Under Test ...........................................................
................... 19
Safety .......
7. Applying the Test Surge: Coupling and Decoupling Circuits .........................................
ts for Surge Coupling ................................................................
Impedance Considerations .......................................................................
Requirements for Surge Decoupling .............................................................
Surge Coupling Modes ............................................................................


8. Grounding . . . .
......................................................................................... 22
8.1 Grounding for Safety .............................................................................. 22
Grounding Practices in Equipment Under Test ................................................

Fig 1
Fig 2
Fig 3
Fig 4
Fig 5
Fig 6
Fig 7

Guiding Considerations for Surge Testing ......................................................... 10

Surge Testing Equipment Considerations ......................................................... 16
Monitoring within Surged Equipment with Voltage Probes in Differential Connection ....... 18
Monitoring within Surged Equipment with Current Transformers ............................. 18
Elementary Diagram of Series Coupling ...........................................................
Elementary Diagram of Shunt Coupling .......................................................... 20
EUT Being Surge Tested, Showing Required Interfaces, Filters, or Reterminations . . . . . . . . . . . 22


Table 1
Table 2
Table 3


Selected Coupling Modes for Single-phase

(One Line and Neutral with Grounding Conductor). . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 3
Selected Coupling Modes for Single-phase
(Two Lines and Neutral with Grounding Conductor) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .24
Selected Coupling Modes for Three-phase
(Three Phase Wires and Neutral with Grounding Conductor) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .25


Notes on Definitions and Special Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27


Fig A1
Fig A2
Fig A3
Fig A4
Fig A5
Fig A6
Fig A7
Fig A8
Fig A9
Fig A10
Fig A1 1
Fig A12
Fig A13
Fig A14
Fig A1 5
Fig A16
Fig A17

Fig A18
Fig A19
Fig A20
Fig A21

Connection of the EUT with the Environment (a) Single Connection to the
AC Power System (b) Connections to Both AC Power System and Signal Lines . . . . . . . . . . .27
Example of Blind Spot Condition (a) With Single Protective Device (b) With
Multiple Protective Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . .29
Common Mode and Normal Mode in Communications Circui
in Power Circuits . . . .30
Difference of Potential Between Neutral and Grounding Conductor Resulting at
Line End from Surge Applied at Line Beginning.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30
Current Surge Test Example: Surging Ground or Common . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .31
Example of Protector with Multiple Devices . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .32
Simplified Test-Surge Generator Output Waveing Network . . . . . . . . . . . . . . . . . . . . . . . . . . . .34
Rate of Surge Occurrences versus Voltage Level at Unprotected Locations for
Different Exposures. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 5
Back Filter Added to Series Coupler . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . .36
. . . . . . . . . . . . . . . . . . . . . . . . .36
Back Filter Added to Shunt Coupler.. . . . . . . . . . . . . . . . . . . . . . . . . . . .
I Applications.. . . . . . . .36
Back Filter and Coupler with Input Isolation Transformer for
. . . . . . . . . . . . . . .37
Effect of Poor or Missing Connection to Grounding Conductor . . . . . . .
Floating Reference (Floating Ground) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .37
Single-Point Grounding (a) Parallel Single-Point Ground (b) Series Single-Point
Ground . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . .37
Multiple-Point Grounding. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . .38
Location Categories Defined in ANSI/IEEE C62.41-1980 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
Similarities and Differences Between the Location Categories Concept of
ANSIAEEE C62.41-1980 and the Installation Categories Concept of
IEC Pub 664-1980, Applied to a Typical Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
Phase Angle Effects.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
Series Coupling of Surge Generator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
Shunt Coupling of Surge Generator.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 6
Excitation of Nonlinear Protectors.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 8


Table A1 Description of Surge Environment in ANSI/IEEE C62.41-1980 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

Table A2 OCV/SCI (Peak Open-circuit Voltage)(Peak Short-circuit Current). . . . . . . . . . . . . . . . . . . . . . .33
Table A3 Suggested Wait Timesas a Function of Total Number of Surges for Surges
Applied as per Categories and Waves of ANSI/IEEE C62.41-1980... . . . . . . . . . . . . . . . . . . . . . . 4 5
INDEX ......................

. .. . . . . . . .. .. . . .. .. .. . . . . . . . . . . . . . . . .. .. . . .. . . .. . .. .. .. .. .. . . .. . . . . . . . . . . . . . 5 0

An American National Standard

IEEE Guide on Surge Testing for

Equipment Connected to
Low-Voltage AC Power Circuits

1. Scope

figuration of the power system within a building.

The environment description of ANSI/IEEE
C62.41-1980 [ 101 provides guidance on the waveshapes of current as well as voltage surges that
can be selected as representative of the environment for different location categories. The
recommendations of IEC Pub 664-1980 [ 171, provide the basis for insulation coordination. Other
standards or specifications for surge withstand
capability might also require surge testing to be
performed, for which the guidance of the present
document may also be applicable. In any case,
equipment connected to the power system has to
be capable of satisfactory operation or survival, or
both, under these surge voltages, with or without
additional protection. Surge testing is therefore
required to demonstrate this capability.
The flow chart of Fig 1 shows the basic considerations involved in planning and performing
surge tests in accordance with the scope of this
guide. Note that the assignment of surge withstand levels to various equipment is not included
in the scope of this guide.

The purpose of this document is to provide guidance on applying surge testing to ac power
interfaces' of equipment connected to lowvoltage (up to 1000 V rms) ac power circuits that
are subject to transient overvoltages. These transients may be defined by any applicable document, such as ANSI/IEEE C62.41-1980
Pub 664-1980 [17]. Other documents (see [19],
[20], [26],and [33]) may also be used as appropriate. While frequent reference is made to documents [ 101 and [ 171,the application of this guide
is not limited to these two documents; the information they provide is primarily for line-to-neutral
surges, while the guidance provided here applies
to all configurations and all magnitudes and waveshapes of surges. Signal and data lines are not
included in the scope of this guide, but should not
be overlooked in the complete evaluation of specific equipment.
In ANWIEEE C62.41-1980 [ l o ] , a description
is given of the surge environment that can be
expected to prevail in low-voltage ac power circuits where no special precautions have been
taken to limit transient overvoltages; in IEC Pub
664-1980 [ 171,a recommendation made to establish a controlled overvoltage situation is justified
if there is compliance with appropriate interface
requirements, so that maximum overvoltage levels in a descending staircase can be assigned
according to a specified installation category
where such a category generally reflects the con-

2. Definitions
Systematic attempts have been made in this
guide to use the definitions provided by ANSI/
IEEE Std 100-1977 [3] and various IEC documents (see [12], [13], [14], and [15]) for the
terms used in the text. However, complementary
or more specific definitions are also necessary.
These definitions are given in this section and are
supported by appropriate discussion, as needed,
in the Appendix of this guide. When a term
covered by such a discussion appears for the first
time in each section, it is printed in bold italics
and Can be found in the alphabetical listing ofthe


See t h e Appendix for a discussion of terms indicated in

bold italic.
The numbers in brackets correspond to those of the references listed in Section 3.



To Be Tested


Qualification Test

Production Test

Diagnostic Test

- Find Limits

- Pass Fail

- Pass Fail
- Any Drift?

- What Happened?
- Retrofit OK?


and or

and or


and or


Upset (Susceptibility)
Momentary or Latch Up



Qualified Personnel



Voltage and Current

Phase Angle
Surge Coupling

Voltage Test


Progressive Tests
Number of Surges
Protector Life

Progressive Tests
Number of Surges

Test Circuits

Fig 1
Guiding Considerations for Surge Testing




2.1 Technical Definitions

3. References and Document in Preparation

back filter. A filter inserted in the power line

feeding an equipment to be surge tested; this fdter
has a dual purpose:
(1) To prevent the applied surge from being fed
back to the power source where it may cause
(2) To eliminate loading effects of the power
source on the surge generator.

The following publications shall be used in conjunction with this standard.

blind spot. A limited range within the total

domain of application of a device, generally at
values inferior to the maximum rating. Operation
of the equipment or the protective device itself
might fail in that limited range despite the device's
demonstration of satisfactory performance a t
maximum ratings.
coupler. A device, or combination of devices,
used to feed a surge from a generator to powered
equipment while limiting the flow of current
from the power source into the generator.
surge let-through. That part of the surge that
passes by a surge protective device with little or
no alteration.
surge remnant. That part of an applied surge
that remains downstream of one or several protective devices.
susceptibility. The inability of a device, equipment, or system to resist an electromagnetic disturbance.
NOTE: Susceptibility is the lack of immunity.

vulnerability. The characteristic of a device for

being damaged by an external influence, such as a
2.2 Special Word Usage. The words will, can,
may, and might are used in this guide with the
following meanings:
will. Denotes the certainty of an event to occur.

can. Denotes that it is possible for an event

to occur under the natural laws governing the
may. Denotes that an event is possible, or that it
is possible to exercise a choice a t the discretion of
the person conducting the test.
might. Denotes that an event is likely to occur, or
that a choice has been made previously by a third
party beyond the control of the person conducting the test, with actual occurrence uncertain.

3.1 References

[ 1] ANSI C2-1987, American National Standard

National Electrical Safety Code.

[a] ANSI C62.2-1981, American National Standard Guide for the Application of Valve Type
Surge Arresters for Alternating-Current Systems.
[3] ANSI/IEEE Std 100-1984, IEEE Standard
Dictionary of Electrical and Electronics Terms2
[4] ANSUIEEE Std 518-1982, IEEE Guide for the
Installation of Electrical Equipment to Minimize
Electrical Noise Inputs to Controllers from External Sources.
[5] ANSI/IEEE C37.90-1978, IEEE Standard
Relays and Relay Systems Associated with Electric Power Apparatus.
[6] ANSI/IEEE C62.1-1984, IEEE Standard for
Surge Arresters for AC Power Circuits.
(71 ANSI/IEEE C62.31-1984, IEEE Standard Test
Specifications for Gas-Tube Surge-Protective
[SI ANSI/IEEE C62.32-1981, IEEE Standard Test
Specifications for Low-Voltage Air Gap SurgeProtective Devices (Excluding Valve and Expulsion Type Devices).
[ 9 ] ANSI/IEEE C62.33-1982, IEEE Standard
Test Specifications for Varistor Surge-Protective

[ 101 ANSI/IEEE C62.41-1980, IEEE Guide for

Surge Voltages in Low-Voltage AC Power Circuits.
[ 111 ANSI/NFPA 70-1987, National Electrical
ANSI publications are available from the Sales Department, American National Standards Institute, 1430 Broadway,
New York, NY 10018.
ANSI/IEEE publications can be obtained from the Sales
Department, American National Standards Institute, 1430
Broadway, New York, NY 10018, or from t h e Institute of Electrical a n d Electronics Engineers, Service Center, Piscataway,
N J 08854.
ANSI/NFPA publications can be obtained from the Sales
Department, American ,National Standards Institute, 1430
Broadway, New York, NY 10018, or from Publication Sales,
National Fire Protection Association, Batterymarch Park,
Quincy, MA 02269.



[ 251 Characteristics of the Electricity Supply Voltage for Low Voltage Consumers. Report of the
UNIPEDE, from Revue Generale de l'Electricite,
Paris, Sept 1981, pp 613-622.

[ 121 IEC Multilingual Dictionary of Electricity.

Institute of Electrical and Electronic Engineers,
[13] IEC Pub 50-111-1982, Physics and Chemistry.

[26] FREY, 0.The Origins, the Effects, and the

Simulation of Transients as Well as Their International Standardization. Record of Professional
Program Session 12,Electronic Show and Convention, May 25-27,1982, Electronic Conventions,
El Segundo, California, Paper 12/5, pp 1-32.

[ 141 IEC Pub 50-131-1978, Electric and Magnetic

[ 151 IEC Pub 99, Lightning Arresters.
[ 161 IEC Pub 158-2, Semiconductor Contactors.

[27] GOLDSTEIN, M. and SPERANZA, P. D. The

Quality of U.S. Commercial Power. INTELEC Conference Proceedings, 1982, pp 28-33,

[ 171 IEC Pub 664-1980, Insulation Co-ordination

within Low-Voltage Systems, Including Clearances
and Creepage Distances for Equipment (including
IEC Pub 664A-1981, supplement).

[28] GRAF, W. and VANCE, E. F. A Topological

Approach to the Unification of Electromagnetic
Specifications and Standards. Proceedings of the
IEEE 1982 National Aerospace and Electronics
Conference, May 1982.

[ 181 MIL Std 461B, Electromagnetic Emission

and Susceptibility Requirements for the Control
of Electromagnetic I n t e r f e r e n ~ e . ~

[ 191 NEMA ICs 2-1983, Devices, Controllers, and

Assemblies for Industrial ControL6

[29] MARTZLOFF, F. D. Matching Surge Protective Devices to Their Environment. IEEE Transactions on Industry Applications, vol IA- 2 1,
no 1, Jan/Feb 1985.

[20] UL 943-1985, Safety Standard for Ground

Fault Circuit interrupter^.^

[ 21 ] ALLEN, G. W. and SEGALL, D. Monitoring of

Computer Installations for Power Line Disturbances. Conference Paper C74 199-6, IEEE PES
Winter Meeting, New York, 1974.

[30] MARTZLOFF, F. D. The Propagation and

Attenuation of Surge Voltage and Surge Currents
in Low-Voltage AC Power Circuits. IEEE Transactions on Power Apparatus and Systems, vol
PAS-102,no 5, May 1983, pp 1163-1170.

[22] BULL, J. H. Impedance of the Supply Mains

at Radio Frequencies. Proceedings of the First
Symposium on EMC, Montreux, Switzerland, May
1975, 75CH1012-4 Mont, pp 357-362.

[31] MARTZLOFF, F. D. and H A " , G. J. Surge

Voltage in Residential and Industrial Power Circuits. IEEE Transactions on Power Apparatus
and Systems, vol PAS-89, no 6, July/Aug 1970,pp
1049- 1056.

[23] CARROLL, R. L. Loop Transient Measurements in Cleveland, South Carolina. Bell System
Technical Journal, vol 59, no 9, Nov 1980, p p
1645- 1680.

[32] MORRISON, R. Crounding and Shielding

Techniques in Instrumentation. New York: John
Wiley & Sons, 1967.

[24] CARROLL, R. L. and MILLER, P. S. Loop

Transients at the Customer Station. Bell System
Technical Journal, vol 59, no 9, Nov 1980, pp
1609- 1643.

[33] ODENBERG, R. and BRASKICH, B. J. Measurements of Voltage and Current Surges on the
AC Power Line in Computer a n d Industrial
Environments. IEEE Transactions on Power
Apparatus and Systems, vol PAS-104, no 10,
Oct 1985, pp 2681-2691.

IEC publications are available in the US from the Sales

Department, American National Standards Institute, 1430
Broadway, New York, NY 10018.
MIL publications are available from US Navy Publications
and Forms, 5801 Tabor Avenue, Philadelphia, PA 19120.
NEMA publications are available from the National Electrical Manufacturers Association, 2101 L Street, NW, Washington, DC 20037.
UL publications are available from Publication Stock,
Underwriters Laboratories, Inc, 333 Pfingsten Rd, Northbrook,
IL 60062.

[34] OTT, H. Noise Reduction Techniques in

Electronic Systems. New York John Wiley & Sons,

[35] RHOADES, W. T. Development of Power Main

Transient Protection for Commercial Equipment.
Proceedings of the 22nd IEEE EMC Symposium,
OCt 7-9, 1980, pp 235-244.

C62.45- 1987


4.1 Surge Environment. Surge testing is generally performed to determine the surge withstand
capability of specific equipment ( E m , or equipment under test) that will be exposed to known
or unknown surge environments. Therefore, the
first decision or assumption to be made in planning a surge test concerns the nature of the surge
environment. ANSIAEEE C62.41-1980 [lo], IEC
Pub 664-1980 [ 171, or some other applicable document will be used to define the environment.
While the test procedures discussed in this
guide should be relevant to most surge tests, the
major concern here is with switching- and lightning-induced surges. Surges associated with
nuclear electromagnetic pulse (NEMP) and electrostatic discharges (ESD) involve rise times in
the order of a few nanoseconds, requiring instrumentation of different characteristics from those
discussed here. Furthermore, high-frequency
noise, generally a t amplitudes less than twice the
normal voltage, is the subject of other documents
(see [41, ~ 6 1 P, I , and ~ 9 1 ) .

[36] RHOADES, W. T. The Ratiocination of a

Commercial Power Main Conducted Susceptibility Standard. Proceedings of the 23rd IEEE EMC
Symposium, 1981, p p 269-276.
[37] RICHMAN, P. Changes to Classic Surge-test
Waves Required by Back Filters Used for Testing
Powered Equipment. Proceedings of the Sixth
S y m p o s i u m on EMC, Zurich, Mar 1985, p p
[38] RICHMAN, P. Conductive Surge Testing of
Circuits and Systems. Presented at the FAA-NASA
Symposium on Lightning Technology, Florida
Institute of Technology, Apr 1980.
[39] RICHMAN, P. Diagnostic Surge Testing, Parts
I & 11. Power Conversion, Sept/Oct 1979, pp
70-78, and Nov/Dec 1979, p p 64-67.
[40] RICHMAN, P. Precision Coil Impulse Testing
with 0.3 A, Micro-Breakdown Sensitivity. Power
Conversion International, vol8, no 5, May 1982,
pp 14-28.
[41] RICHMAN, P. Single-output, Voltage and Current Generation for Testing Electronic Systems.
Proceedings of the IEEE EMC Symposium, 1983,
pp 47-51.

4.2 Types of Tests. Figure 1 shows a branching

point involving the purpose of the surge test. Four
types of tests are identified, with their purpose
and interested parties, as follows.
Design tests are performed by an equipment
manufacturer for establishing or demonstrating
to others design margins and for optimizing the
design. These tests may involve pushing t h e
stresses to the limits until a failure is observed.
Qualification tests a r e performed by t h e
manufacturer, purchaser, or independent test
laboratory for demonstrating compliance with
specifications. These tests generally are limited to
a pass-fail criterion, but are more comprehensive
than tests carried out on a routine basis on production products.
Production tests are performed by an equipment manufacturer for verlfylng conformity and
consistency of the production process. These tests
generally involve some statistical evaluation.
Diagnostic tests are performed by the manufacturer or user for investigating difficulties encountered in service. These tests generally involve
attempts a t laboratory reproduction of the failure
modes observed in the field, followed by applying
the same test on equipment that has been redesigned or provided with retrofit protection.
Prior to conducting tests on an EUT, acceptance criteria should be defined in accordance
with the considerations of Fig 1, such as purpose
of the test and expected outcome.

I421 VANCE, E. F. Electromagnetic Interference

Control. IEEE Tra,nsactions on Electromagnetic
Compatibility,vol EMC-22, Nov 1980, pp 319-328.
[43] VANCE, E. F., NANEVICZ, J. E., and GRAF,
W. Unification of Electromagnetic Specifications
and Standards. Defense Nuclear Agency Report
DNA 5433F-1, Washington, DC, 1980, pp 49-50.
3.2 Document in Preparations

4. Planning of Surge Testing:

Basic Objectives
This section outlines the basic objectives to be
considered for planning surge testing, as indicated graphically in Fig 1. More detailed discussions of these considerations are provided in the

When the following document is completed, approved,

and published, it will become a part of the references of this
standard: PC62.35, IEEE Standard Test Specifications for Avalanche Junction Semiconductor Surge-Protective Devices.




4.3 Results and Consequences of the Test.

While surge protective devices are generally provided for damage avoidance, they can also serve
to prevent upset in the operation of an EUT.
Therefore, a treatise on surge testing must include
subject matter dealing with the evaluation of the
test results. Any given surge test will produce one
of four results:
(1) upset (susceptibility)
(2) damage (uulnerability)
( 3 ) no observed change, or
(4) an unforeseen consequence elsewhere in
the equipment environment.
The last result actually involves consideration
of circumstances external to the EUT proper that
might be overlooked or considered irrelevant to
the scope of surge testing in the laboratory. Setting aside any consideration of unforeseen consequences on that basis would be a severe error. The
note unforeseen consequences in the Appendix
gives some scenarios that illustrate this concern.
Depending upon the nature and function of the
EUT, six different outcomes of a surge test should
be evaluated wherever direct, local results are
(1) No apparent response in the EUT - neither
upset nor damage
(2) Temporary upset of the EUT operation
( 3 ) Upset with trip-out or latch-up of the EUT
(4) Flashover of clearances without apparent
permanent damage
(a) with no power-follow or with a selfclearing power-follow, which might seem a benign
(b) with power-follow resulting in operation
of an overcurrent protective device (fuse or
breaker), an occurrence similar to outcome (3)
(5) Insulation degradation or breakdown due
to partial discharges across the surfaces or in
solid insulation, or both
(6) Insulation breakdown or permanent component damage requiring replacement or repair.
The first outcome (no upset or damage) may
represent a success from the point of view of
acceptance, but yields incomplete information
since the actual design margin is not determined
until further tests, at higher stress, are performed.
The next two outcomes (upsets) are mainly
concerned with control or data circuits, and are
related to the susceptibility of the equipment. The
electrical noise required to produce them can be
quite low; in fact, low-level noise can be sufficient to upset sensitive circuits (see [4], [16],
[18], and [19]). The emphasis in this guide,

however, is on surges, generally implying voltage

levels of twice or more the normal voltage of the
The fourth outcome (flashover) might involve
both control and power circuits, and is expected
to occur at surge levels significantly above the
normal circuit voltages. As long as no permanent
damage or insulation tracking occurs as a result
of the sparkover and eventual power-follow, this
outcome is still in the category of susceptibility.
Some EUTs might be insensitive to or unaffected
by the flashover, while others would definitely be
considered as having been upset by the flashover.
The fifth outcome (insulation degradation or
breakdown) might occur across the surfaces of
insulation or within solid insulation as a result of
partial discharges, especially if multiple tests are
The sixth outcome (permanent damage) might
occur in either control or power circuits, as a
result of sparkover with or without power-follow
producing a permanent degradation, or as a result
of semiconductor failure or excessive energy
deposition leading to component or etch burnout.
The fifth and sixth outcomes describe the vulnerability of the equipment.
Some of the outcomes can occur in combination so that the distinction made here might not
be so clearcut in reality, but which is nevertheless
useful as a starting point. It should be remembered that any surge test is potentially destructive
to the EUT, so that appropriate precautions
should be taken.
4.4 Unpowered Testing versus Powered Testing. Test surges may be applied to the EUT in two
(1) With normal operating power disconnected
from the EUT (unpowered testing), and
( 2 ) With normal operating power applied to
the EUT @owered testing).
The intended purpose of the test will determine
whether one approach is sufficient or whether
both are advisable.
Unpowered testing is sufficient in situations for
which a test outcome does not depend on the
evaluation of EUT performance during the surge,
and for which power-follow is not considered to
be a significant factor in regard to vulnerability.
For instance, clearance flashover of an electromechanical device may be the selected failure criterion; in that case, there should usually be no
need to power the EUT. Unpowered testing is
usually necessary as a preliminary to powered
testing, for design and diagnostic testing.



Powered testing is necessary under two circumstances:

(1) When a test outcome depends on the evaluation of EUT functional performance during the
surge. Thus, a test for susceptibility implies normal equipment functioning prior to the surge;
therefore, the EUT can only be checked in the
powered mode.
(2) When determination of EUT vulnerability
involves the likelihood or consequence of a powerfollow (which might also depend in part on the
phase angle at which the surge is applied with
respect to the line voltage wave).

While this aspect of surge testing is not directly

within the scope of low-voltage ac power circuits,
it should be recognized. Some discussion is provided in the Appendix under current surging.
Therefore the waveshape for both voltage and
current tests should be included in specifying a
test procedure. ANSI/IEEE C62.41-1980 [ 101
attempts to make such a distinction between current and voltage tests.
The specific selection of withstand levels, for
voltage as well as current, depends on the exposure to transients as well as on the consequences
of a failure to withstand the transient. This guide
provides some perspective in selecting appropriate
levels but the final choice must be made by the
user of this document. This guide therefore includes a wide range of test alternatives to provide
guidance in making such decisions with maximum flexibility. The guidance given here is applicable to all waveforms, including longer or shorter
waves if the particulars of the application indicate that these waves are appropriate (see (261,
[33], and data base in Section A1 of Appendixes
to (101).

4.5 Withstand Levels. Surge testing is ordinarily

carried out at different stages in a product life
cycle, such as design, quality control, and protection retrofit. The extent and severity of the test
will depend on the particular stage involved. A
design test is likely to involve testing to failure
while a production test must carefully avoid
creating incipient failures. The voltage surge
environment (see ANSI/IEEE C62.41-1980 [ l o ] )
is described only in statistical terms without
imposing a fixed withstand level. On the other
hand, IEC Pub 664-1980 [ 171 sets forth maximum
levels for the various installation categories and
system voltages in a controlled overvoltage situation. Thus, if equipment is to be classified as suitable for various installation or location categories,
it will be desirable to establish durability or withstand levels for comparison purposes.
The withstand levels should be expressed in
terms of voltage for equipment exhibiting high
impedance to a surge; for those EUT that contain
a surge protective device, the withstand level
should be expressed in terms of current in order
to give consideration to energy deposition, as discussed in 6.2.

4.7 Safety. Surge testing of electrical or electronic equipment presents potentially hazardous
situations for both personnel and equipment. The
surge tester generates potentially lethal voltage
surges. Furthermore, failure of the EUT might
result in a fire or explosion. Only qualified personnel should perform the tests, with safety precautions enforced according to national codes as
well as the normal safety directives of the organization conducting the test. More specific aspects
of safety are discussed in 6.4.

5. Implementation of Surge Testing:

Test Equipment

4.6 Waveshape, Voltage, Current. The nature of

the EUT will affect its response to an applied test
surge. A high-impedance EUT, such as a winding,
a clearance, or a semiconductor in the blocking
mode, will be stressed by a voltage surge. The
energy associated with the surge is not significant
here. A low-impedance EUT, such as a circuit containing fiter capacitors or surge-divertingprotective devices, will be stressed by a current surge.
The energy deposited in the components becomes
a significant factor in this type of surge occurrence.
Current test surges applied between different
points of the grounding connections can also provide essential information on the EUT capability.

5.1 General. Implementation of surge testing

can proceed after the considerations discussed in
Section 4 have been addressed. This section provides guidance on the major aspects of the equipment requirements, for generating and applying
the surge as well as for monitoring the performance of the EUT (Fig 2). Occasional users might
encounter difficulty and should obtain guidance
from qualified sources. Test equipment that does
not meet the requirements of applicable standards might give misleading results (see [30]).
5.2 Surge Generators. The test surge to be applied to the EUT will be produced by a surge
generator capable of delivering the specified

(6) Possible Outcomes

Surge AC Interface
Surge Other Points*
Surge Component Terminals

Unpowered vs Powered

Unknown EUT Impedance

High EUT Impedance
Low EUT Impedance
EUT With Impedance that
Changes During the Test

waveshape at any specifledphase angle of the ac

voltage sine wave at the EUT terminals. Capability
for bidirectional surge polarity simplifies the
general test procedure. A number of commercia!
surge generators are available with specific waveshapes that meet the various standards in existence. High-voltage laboratories are also generally
equipped and staffed so that the generation of a
test wave is not a problem. Surge testing of electronic equipment is different from a simple dielectric test on an insulation system or a simple surge
current test on an individual surge protective
component (see [2], [6], [7], [ 8 ] , [9], and 3.2).
An EUT that contains a surge protective device or
that might experience an insulation breakdown
during the test will exhibit an impedance change
during the surge. A surge generator inherently
capable of delivering a specified voltage or a specified current during a single surge test as required
by the EUT impedance will yield information on
the EUT performance that cannot be obtained by

two separate tests, one for voltage, the other for

current (see [41] and [43]). Further discussion of
these considerations is given in the Appendix
under waveshape: voltage versus current.

5.3 Point of Test Surge Application. According

to the scope of this guide, the ac interface of an
EUT is the point of application of the test surge.
In the process of evaluating the performance of
the EUT, other terminals may also be subjected to
surges. Interconnected or distributed systems may
have to be broken into separate subsystems, or
the whole system may have to be treated as an
EUT. Therefore, the nature of the EUT will affect
the points a t which the surge is to be applied, and
thus the method of coupling the surge.

5.4 Coupling the Surge to the Equipment Under

Test. In the case of an unpowered testing, the
coupling is quite simple. The input port of the

C62.45- 1987


EUT is merely connected to the output terminals

of the surge generator, but further precautions
are required. All other terminals or outputs of the
EUT, including its grounding condudor(s), should
be isolated to prevent damage to other equipment.
In the case of apowered testing, the coupling
becomes a complex matter, which is discussed in
detail in Section 7. This complexity is the result of
the need to apply the surge to the power supply
line of the EUT, maintaining the specified waveform, but without feeding the surge back into the
laboratory ac power supply where it might damage other loads in the laboratory. Thus, a back
filter is needed to prevent this feedback. In addition, there is a need to isolate the power supply
line, lest it load the surge generator, thereby
reducing the generator output below the required
levels. The availability of a separate power supply
generator, often used in specialized laboratories,
can alleviate some of these problems.
Isolated components or simple two-terminal
devices can be subjected to the surge in a simple
configuration; multiterminal devices, including a
simple balanced two-input EUT with ground, require careful attention to speclfylng which terminals are surged with respect to which others. This
area of coupling techniques and configurations is
treated in greater detail in Section 7.

Monitoring within the EUT may also be necessary in order to understand the failure mechanism under the surge, to control one or more
critical voltages within the EUT, or to check the
amount of surge remnant or surge let-through
reaching specific critical components.
5.5.1 Monitoring with Voltage Probes. A reliable and safe method for monitoringvoltages within the EUT is to use a differential connection of
two matched voltage probes (Fig 3). This type of
connection, shown in the figure for the case of a
surge applied between line and neutral of the
EUT, enables the use of a safely grounded oscilloscope or peak detector. The high voltage probes
have no ground leads attached to the EUT, while
the chassis of the instruments are safely grounded
by the grounding conductors of their power cords.
5.5.2 Monitoring with Current Transformers.
Properly applied current transformers can be
useful for monitoring surge currents. It is often
desirable, if not necessary, to monitor current
during application of a voltage surge, in order to
detect breakdowns or to v e r a EUT performance.
A current transformer enables complete isolation
of the current-monitoring channel of the oscilloscope, in contrast with a current-viewing coaxial
shunt, which can only be inserted very near the
grounding reference point.
Voltages can also be monitored with a current
transformer: a high resistance (noninductive resistor, with appropriate surge voltage rating) is connected between the two points where voltage is to
be monitored and a current transformer is used
to monitor the current in the resistor, hence the
voltage difference at its points of connection.
Figure 4 shows these two applications of current transformers. Note, however, that special
precautions are required to provide safe and valid
measurements, which are discussed in detail in
the Appendix under current transformers.

5.5 Monitoringthe Equipment Under Test. Both

the applied surge and the output, as appropriate,
of the EUT may need to be monitored; monitoring
may also be required within the EUT. Current, as
well as voltage, should be monitored to provide
complete information on the EUT performance
(see [40]).
The need to monitor the input surge is axiomatic since this will v e r a the characteristics of
the applied surge, both open-circuit and modified
by the load. For simple failure modes of isolated
components, such as insulation breakdown or
permanent semiconductor damage, monitoring
the applied surge also reveals a failure because
the observed applied voltage wave will appear
chopped. On the other hand, a surge applied to an
EUT being powered by the line might show extensive distortion and ringing, making diagnosis by
waveform inspection nearly impossible (see [30]).
Checking a complex EUT for susceptibility
under surge requires more extensive instrumentation in order to detect misoperation. (That
instrumentation itself must be immune to the disturbances created in the area by production of
the test surge.)

6. Performance of Surge Testing:

Test Procedures
6.1 General. This section is intended primarily
for the guidance of those individuals involved in
performing surge tests. Those who do the testing
are presumed to be familiar with safety procedures and with the general techniques of highvoltage and high-frequency (impulse) instrumentation. Specific guidance is therefore aimed at the
specialized aspects of surge techniques.
(1) A surge test is a single event. Thus, once the
surge has been applied to the EUT, any damage

C62.45 1987


Minimize area cross hatched in order

to avoid contaminating measurement by
voltage induced in probe loop

Insulation or spacing
for twice pe=_


CRO frame should not

touch any other frames
AC Power


("A ..... B " ) *


o f oscilloscope

Isolate or
disconnect All
other connections
to EUT

o f generator

'Or suitable differential probe/amplifier

NOTE: The bonding of probe shields to each other at the measuring end of the probes (they are both connected to the
oscilloscope chassis at the other end) has been advocated by some workers, while discouraged by others. The benefit of either
method might be dependent upon the particulars of the situation. A check for noise background (both probes connected to the
same point) will provide the basis for a selection tailored to the particular situation.

Fig 3
Monitoring within Surged Equipment
with Voltage Probes in Differential Connection

Insulation or spacing
for twice peak surge

CRO frame should not

touch anv other frames


AC Power


Frame of EUT

of oscilloscope

Isoiate or

of generator

A .^^^^^^^ I

A I ,

other connections
to EUT

See par. 5.5.2 for requirements

for current transformers CTI, CTV
and resistor R

Fig 4
Monitoring within Surged Equipment with Current Transformers

C62.45 1987


sensitive to the phase angle of the applied surge

with respect to the power supply.

that occurs has to be repaired and the most

probable cause determined before the next surge
test is run, possibly at a lower level.
(2) Voltage and energy levels required to duplicate the equipment surge environment are necessarily high enough to be a personnel hazard.
(3) Performance of some surge protectors is
statistical in nature, and performance of virtually
all surge protective devices is a strong function of
the surge waves applied.

6.4 Safety. Many of the tests indicated in this

guide are inherently hazardous; the safeguards
for personnel and property described in this section are essential to achieve safety. Observance of
the prescriptions of the ANSVNFPA 70-1987 [ 11]
(National Electrical Code [NEC]) and ANSI C21987 [ 11 (National Electrical Safety Code [NESC]),
of course, is a requirement.
Surge testing is best conducted only in an area
dedicated solely to that purpose. The boundaries
of the area should be clearly defined and appropriately marked. Where possible, the area should
be fenced in and provided with electrical or
mechanical interlocks, or both, on all entrances
into the test area and removable barrier panels.
All metal fences or barriers, or both, should be
grounded. Care should be taken to ensure that all
of the EUT is within the assigned area. Consideration should be given to the possibility of the surge
flashing over to circuits or metallic parts that
were not intended to be surged. The surge test
area should be kept free of all materials, meters,
test setups, and so forth that, are not associated
with the surge test being conducted.
When the EUT can be enclosed in an effective
barrier, the preceding requirements are easier to
satisfy. This barrier itself may be sufficient
separation -including separation from the floor,
which should be presumed to contain conduit or
other metal. Alternatively, the entire barrier may
be made up of physical insulation. In either case,
it should be complete except where penetrated
for insertion of input or output lines and measurement probes, and it has to be safe for a peak
volt,age equal to at least twice the peak of the
incident test surge. (Circuits in breakdown at or
near the surge peak can oscillate at high frequencies. Such oscillatory flashovers can thereby
increase effective applied peaks by a factor
approaching two.) Interlocks should be provided
to allow safe access between tests.
Capacitors used in the fdter or coupler might
retain a trapped charge; suitable bleeders or
short-circuiting devices should be provided to
ensure operator safety against any such trapped
charge after passage of the test surge.
Consideration should also be given to the possibility of ignition or explosion within the EUT.
Where an examination of the EUT indicates a likelihood of ignition, factors to be considered are
(1) The amount of combustible materials likely
to be involved initially.

6.2 Limiting Stresses. Pass/fail qualification

tests andproduction tests may consist of a single
surge application. On the other hand, design
tests are generally applied by increasing the surge
levels in several steps starting from the operating
voltage level and increasing to the goal, in order to
obtain meaningful data and reveal blind spots.
However, these many steps result in an accumulation of energy deposition that needs to be recognized and possibly limited (see [2], [6], (71, [ 8 ] ,
[9], and 3.2).
All surge protectors have not only surge performance specifications but also maximum
averagepower limitations. Furthermore, a series
of repeated tests can consume (expend) part of
the protector life. Therefore, it is very important
that consideration be given to limiting integrated
stress in multiple pulse tests as well as to life
consumption, average power, and repetition rate,
especially for making repeated tests for blind spot
In the absence of specific information on the
failure modes of the EUT, several surges may be
required of each polarity and at each selected
phase angle to ensure that the defect will be
marked or that ac follow current will finally cause
an arcing fault. Thus, efforts should be made to
reduce a large number of surges by considering
the failure modes and applying good engineering
6.3 Nature of the Equipment Under Test. The
nature of the EUT has an influence on the test
procedure. Single components, or simple systems
without multiple built-in protective devices, can
be tested with a few increasing steps, often in an
unpowered configuration. On the other hand,
complex systems, especially those containing several successive protective devices, require more
comprehensive test procedures. There can be blind
spots in the protection; that is, satisfactory performance a t high stress does not guarantee satisfactory performance at lower stresses, or for different wave shapes. Likewise, some EUTs might be



(2) The probable rate of propagation.

(3) The consequences of such propagation, that
is, the probability of extension beyond the EUT.
Appropriate precautions should be taken to
keep these factors within manageable limits. These
precautions may consist of suitable extinguishing
agents in sufficient quantity, physical separation
from other combustibles, or other appropriate
measures. In evaluating the possibility of explosion, consideration should be given to component
failure whenever hazardous materials are available in enough quantity to create an explosive
All surge testing has to be conducted by technically qualified personnel who are aware of the
hazards of such testing. The voltage and current
levels generally associated with surge testing are
well above those considered lethal. Some considerations are the possibility of an accidental discharge of the surge generator, the consequences
of a flashover to an unfavorable circuit, the possibility of a charge being trapped in the EUT, or the
consequences of a violent component failure.
Testing personnel should never stand in the line
of sight of components on printed circuit boards
or panels with the enclosure open during EUT
surge testing. On occasion, a component will fail
in an explosive manner during surge testing.
Fragments of the ruptured case and the component might cause injury to personnel in the vicinity. If visual observation is desired, a suitable
transparent barrier of sufficient thickness should
be provided.
The importance of conducting surge tests in a
prudent manner cannot be overstressed; safeguarding personnel has to be the prime consideration.

the scope of this guide, but should not be overlooked (see [23], [24], and the discussion of ac
interface and communications interface in the

7.2 Requirements for Surge Coupling. Two

basic methods may be used for coupling: series
coupling connects the test surge generator in series with the conductor being surged (Fig 5), and
shunt coupling connects the test surge generator
in parallel with one or several lines (Fig 6). The
most frequently used method is shunt coupling,
but both have advantages, as discussed in the
To apply the output of a test surge generator to
a powered EUT, it is almost always necessary to
use a surge coupling device, or coupler. The
coupler should conduct the surge energy, with
reasonable waveform fidelity, from the test surge
generator into the EUT. The requirements include
(1) Minimizing cross-loading and power dissipation in the surge generator output network.
Appropriate impedances in the coupler network
have to be provided.
(2) Permitting the EUT to function normally
before and after the test surge. Coupling gaps
can also be used to provide coupling of the generator only during the surge.
(3) Permitting different modes of coupling, as
required by the test schedule.
(4) Providing bleeder action to discharge any
residual voltage trapped inside after the test.

Fig 5
Elementary Diagram of Series Coupling

7. Applying the Test Surge:

Coupling and Decoupling Circuits

2 E line

0 neutral


7.1 General. For independent equipment,the test

surges will be applied to the power lines supplying
the EUT. For interconnected or distributed systems, the testing of the individual units should be
evaluated with regard to the rest of the system.
Testing a complete system might not be possible
or economical. Each unit comprising the system
may be tested as a unit, provided its functional
integrity can be monitored during the test. The
test surges are to be applied to the cable ports
that are connected to cables that are routed to
other areas. All ac power inputs and outputs
within an interconnected system should be surge
tested. Signal and data lines are not included in

T = Surge Coupling Transformel

Fig 6
Elementary Diagram of Shunt Coupling


l ,


C = Surge Coupling Capacitor



7.3 Impedance Considerations. In general, the

output impedance of a test surge generator will
be that of its output wave-shaping network as
seen through the coupler. Since such networks
usually involve inductors or capacitors, or both,
the output impedance will be complex, involving
both a real and an imaginary component.
For convenience, an effective output impedance is defined for a surge generator and its
coupler, if appropriate, by calculating the ratio of
peak open-circuit output voltage (OCV) to peak
short-circuit output current (SCI), or as an OCV/
SCI ratio at the injection points. Ideally, this
impedance, when combined with the back filter
impedance, should represent the ac power system
impedance for the incoming surge (see [22] and
[35]). Typical values range from near 0 SZ at the
power line frequency to 200 SZ above 100 kHz.

7.4 Requirements for Surge Decoupling. The

simplified surge coupling of Fig 5 and Fig 6 shows
the requirements of back filters and decoupling.
Without these, the low line impedance [22] would
load the generator and prevent it from delivering
the full voltage. Furthermore, all other equipment
connected to the same power line would be subjected to the surge, with attendant equipment
damage and personnel hazard resulting. Further
yet, some other equipment connected to the line
might include a surge protective device, defeating
the test. Thus, surge decoupling, generally in the
form of surge filters in the power line (referred to
as back filters), is required to eliminate these
limitations. The insertion of the filter raises the
question, however, of reduced available fault
current because of the added impedance of the
It is necessary to use back filters in all lines into
and out of the EUT, excluding the grounding
conductor, that will not be disconnected for the
surge tests. To avoid possible damage to interconnected equipment, leads to such equipment
should be disconnected, and the associated EUT
port should be terminated with a representative
equivalent circuit.
The neutral conductor is treated just like the
other lines; that is, the back filter will decouple it
from ground during testing. This conductor might,
in fact, be the most susceptible of the connections
to the EUT.
While the signal lines into or out of the EUT are
technically outside of the scope of this guide, their
presence should be recognized. As shown in Fig 7,
they should be disconnected, back filtered, or

C62.45- 1987

reterminated with impedances or grounds that

simulate operating conditions.
The presence of these filters might produce
some interference with ground fault protection
systems, which needs to be recognized.

7.5 Surge Coupling. For a given number of power

supply lines, there are a specific number of ways
that a test surge may be applied. For shunt coupling, the surges can be applied between the
following groups of conductors:
Line(s) to neutral (Fig 6)
Line(s) to line(s)
Line(s) to ground
Neutral to ground
Line(s) and neutral, simultaneously, to ground
For series coupling, the surges can be applied in
series with one conductor, as shown in Fig 5.
Tests should be performed with surges applied
between each group of conductors to evaluate the
capability of an EUT to withstand the surges that
can be encountered in its application.
In tests involving neutral to ground or line to
ground, the available data on characteristics of
these surges are insufficient to recommend specific waveshapes and magnitudes (see current
surging and common mode in the Appendix). It
is expected that surges between either neutral
and ground or between line and ground might be
different from the surges between line and neutral. In situations where the bond between the
neutral and grounding conductor is near the EUT,
common mode surges would be expected to be
smaller than normal mode surges. In situations,
such as European practice, where t h e bond
between neutral and earth is located remotely
from the building service entrance, common mode
surges would be expected to be substantially
larger than normal mode surges (although the
common mode surge will have a relatively large
source impedance). Neither ANSI/IEEE C62.411980 [ l o ] nor IEC Pub 664-1980 [17] addresses
that issue. Other reported measurements provide
some information (see [21], [25], [27], [31],and
[35]). The EUT susceptibility levels can be well
below the EUT withstand level if a decaying ringing surge occurs (see [5] and [35]). Therefore a
comprehensive test program should provide for
some evaluation of neutral-to-ground surge
For test purposes, usually the low terminal of
an ungrounded test surge generator is connected
directly to one of the power lines or the grounding
conductor. The surge generator high terminal is
then connected to the other power line(s) via



Insulation or soacina7
for twice p e a k s u r g e
voltage applied

Line 1
Line 2
Line 3

Neu t ra I





T l l I






Ground to

(1) Signal or power conductors, or both, to other equipment

(2) The crosses (

) indicate one or more of the following:

A. Complete disconnect of the conductors

B. Insertion of a surge filter similar to the back filter

C. Disconnect of the conductors, with addition of a representative termination.
indicates disconnection of grounding conductors to downstream equipment in order to avoid passing
(3) The cross
o n a surge. However, a grounding connection to that downstream equipment must be re-established, bypassing the
EUT test setuD.

Fig 7
EUT Being Surge Tested, Showing Required Interfaces, Filters, or Reterminations
7.5.2 Three Phase. In three-phase systems
from three to five wires may be involved. The tests
marked Basic in Table 3 should routinely be
performed. The tests marked Supplemental in
Table 3 may be performed to obtain additional
information on vulnerability of the EUT to surges.
The tests marked Diagnostic in Table 3 may be
performed in the course of a major investigation.
Table 3 does not show a full complement of
diagnostic tests. Additional diagnostic tests may
be useful in testing specific E m s .

capacitors. A separate capacitor is used for each

line. Note that the use of capacitors presents the
hazard of trapped charges after the test; bleeders
or discharge interlocks should be provided.
In all cases, surge testing should be performed
with both polarities.

7.5.1 Single Phase. In single-phase power systems, the EUT is powered either by two wires, line
and neutral (see Table l), or by two lines plus a
center-tapped neutral (see Table 2). In both cases
a grounding conductor may or may not be present.
The tests marked Basic in Tables 1 and 2
should routinely be performed. The tests marked
Supplemental in Tables 1 and 2 may be
performed to obtain additional information on
vulnerability of the EUT to surges. The tests
marked Diagnostic in Tables 1 and 2 may be
performed in the course of a major investigation
(see [38] and [39]).
Table 2 does not show a full complement of
diagnostic tests. Additional diagnostic tests may
be useful in testing specific Ems.

8. Grounding
8.1 Grounding for Safety. Since the voltages
involved in surge testing are hazardous, appropriate precautions are required for equipment
other than the EUT and for personnel. One basic
precaution is the correct application of grounding conductors in the test setup. See 6.4 for
general safety considerations. Barriers or separation between the EUT and other parts accessible




Connection of Generator
Test Type




Basic 1
Basic 2



Supplemental 1
Supplemental 2

Diagnostic 1
Diagnostic 2






@L %






Example of Connection Diagram

for Shunt Coupling, Basic 2





the conductors shown

to personnel may be used; however, the most

effective protection is obtained by grounding surrounding objects and having one point of the test
circuit maintained at this safety ground potential.
Figures 3 and 4 show the recommended configurations for applying and monitoring the surge,
with safe connections.
In test facilities where there are permanently
connected grounding conductors, always check
for possible defects in the ground system before
each surge test. For instance, in US installations
regulated by ANSI/NFPA 70-1987 [ 111 (National
Electrical Code [NEC]), all grounding is to be in
accordance with this Code. The grounding conductor and the neutral conductor must be bonded
together at the output of every separately derived
source. This bond must not be broken. See also Fig
A12(b) under the note grounding conductor in
the Appendix.
All other connections to equipment that is not
part of the EUT are to be removed (see Fig 7). If it
is not possible to do so, then the connections
should be filtered like the lines actually being
surged, since flashover occurring within the EUT
might be conducted to any port. To ensure proper
testing of internal insulation, these terminals may
have to be locally reterminated with impedances
or grounds that simulate operating conditions.

The test equipment has to include backfilters

to prevent surges from disturbing the power line.
If a ground fault circuit interrupter (GFCI) is
used, the filters are likely to cause operation of
the GFCI. An isolation transformer might be required to avoid this problem. However, the safety
ground is to be bonded to the EUT grounding stud
or equivalent. All common mode tests are performed using the grounding conductor serving as
the reference.
8.2 Grounding Practices in Equipment Under
Test. Electronic circuits are grounded for three
basic reasons:
(1) Safety. Safety grounds with low impedance
provide return paths to the current source for
fault currents, thereby ensuring the rapid and
reliable operation of overcurrent protective devices. The application of common grounds also
provides equalization of equipment potentials to
ensure personnel protection.
When this function is associated with ac power
faults, the term earth ground is frequently used
to distinguish the safety ground from other usages
of the generic term ground, and the establishment of such a connection between equipment
chassis and earth ground is called earthing.




Table 2
Selected Coupling for Single-phase Systems
(Two Lines and Neutral with Grounding Conductor)
Connection of Generator
Test Type



Basic 1
Basic 2
Basic 3


Supplemental 1
Supplemental 2
Supplemental 3



Diagnostic 1
Diagnostic 2


Example of Connection Diagram

for Shunt Coupling, Diagnostic 1

Line 1 Line 2


= connection to surge generator low (Lo)


connection to surge generator high (Hi) by coupling capacitor C,

= connection to surge generator high (Hi) by coupling capacitor C,

= connection to surge generator high (Hi) by coupling capacitor C2

For each test type shown horizontally in the table, the surge generator is to be connected as indicated in the four "Connection of
Generator" columns. The connection diagram in the table shows as an example the jumpers required to obtain the shunt
coupling for test Diagnostic 1.
When several H's appear on one horizontal row of the table, shunt coupling requires several coupling capacitors, shown as C,
C , , C,, between each of the conductors indicated and the surge generator high, in order to apply the surge simultaneously to
the conductors shown.

(2) Signal Voltage Reference. The concept of

a signal, circuit, or logic ground relates to a
common equipotential reference against which
the various circuit components operate, thereby ensuring that the intended signal voltage
levels are consistently and properly recognized
throughout t h e equipment. The relationship
between this common reference and the equipment chassis or frame is a function of the
equipment design and its intended operating
( 3 ) Static Charges. Grounding provides a
means for bleeding off electrostatic charges.
In surge testing, safety precautions are of prime
importance because surge testing involves the use
of potentially dangerous voltages along with the
necessity of making accurate measurements. For

these reasons, it is imperative that the grounding

configuration, not only of the EUT, but of the
entire test setup, be understood.
Unfortunately, t h e two requirements for
grounding are not always compatible (see [as],
[32], [34], and [42]). Power safety grounds are
often very noisy, thereby limiting their use as
signal references. Also, signal reference grounds
are sometimes required to be a t some potential
other than earth, whereas power safety ground
is generally referenced t o earth. Equipment
grounding configurations found in EUTs tend to
fall into one or some combination of four general
schemes: floating reference, single-point ground,
multiple-point ground, and isolated ground.
These are discussed in detail under the heading
grounding practices in the Appendix.




Table 3
Selected Coupling for Three-phase Systems
(Three Phase Wires and Neutral with Grounding Conductor)

Test Type

Connection of Generator

Basic 1
Basic 2
Basic 3
Basic 4

Example of Connection Diagram

for Shunt Coupling, Diagnostic 1


Line 1 Line 2 Line 3



Supplemental 1
Supplemental 2
Supplemental 3
Supplemental 4


Diagnostic 1
Diagnostic 2





= connection to surge generator low (Lo)

= connection to surge generator high (Hi) by coupling capacitor CN

= connection to surge generator high (Hi) by coupling capacitor C ,

= connection to surge generator high (Hi) by coupling capacitor C2
= connection to surge generator high (Hi) by coupling capacitor C,

For each test type shown horizontally in the table, the surge generator is to be connected as indicated in the five Connection of
Genera:.>r columns. The connection diagram in the table shows as an example the jumpers required to obtain the shunt
coupling for test Diagnostic 1.
When several Hs appear on one horizontal row of the table, shunt coupling requiresseveral coupling capacitors, shown as CN,
C , , C2, or C,, between each of the conductors indicated and the surge generator high, in order to apply the surge simultaneously to the conductors shown.


C62.45- 1987

Notes on Definitions and Epecial Considerations
(This Appendix is not a p a r t of khISIIIEEE C62.45-1987,IEEE Guide on Surge Testing for Equipment Connected to Low-Voltage
AC Power Circuits.)

ac power interface. The scope of this guide limits

the equipment under test (EUT) to that which is
connected to low-voltage ac power circuits. Available data on the occurrence of surges, whether
measurements or specifications, concern surge
voltages imposed on the EUT at its connection to
the ac power supply.
In the case of EUTs that are simple loads, the ac
power interface can be understood as the connection to the power system, and the surges to be
taken into consideration are those appearing
among the line(s), neutral, and grounding conductors of the EUT ac power supply [Fig A1 (a) ].
These surges can be applied in the common mode
or in the normal mode, but are limited to the
combinations of conductors on that ac power
supply connection.
There is, however, another category of equipment that involves signal or data lines in addition
to the ac power supply [(Fig Al(b)]. Typical
examples of these EUT are an industrial process
control system with remote sensors and actuators, a TV set with community antenna input, a
computer system with remote terminals, and telephone equipment such as answering machines or
cordless set base stations.
For these types of EUTs, the ac power interface
concept not only includes all of the possible combinations involved in the Fig A1 (a) connection, but
also the surge voltages that might appear between
the ac connection on one hand, and the connection to the signal line on the other hand. This type
of surge exposure is now recognized as one of the
major sources of problems, but little data have
been collected, except on public telephone lines
(see [23] and [24]).9One of the reasons for this
lack of general data may be that each application
depends on the local conditions of installation.
Users of this guide should not ignore the potential problems of surges appearing between the ac
power connection and the signal lines, although
the guide does not cover such surges.





AC Interface

(a) Single Connection to the AC Power System


AC Interface

(b) Connections to both AC Power System

and Signal Lines

Fig AI
Connection of the EUT with the Environment

instrumentation capability, it is unlikely that all

types of protectors that may be used within an
EUT will be able to withstand such treatment for
very many surges (see [a], [6], (71, (81, [9], 3.2,
(All, and (A2]).9
Tests should take into consideration realistic
intended use of the EUT. A life test should not be
compressed with an excessive test-surge repetition rate, for the sake of saving test time, if stress
caused by excessive average power would result
in apparent failure of an adequate EUT protection design.
The stress level-voltage, current, or energyof each individual test surge is no guide to the
rate at which the test may be repeated. What is significant is the relationship of that stress level to the

average power (overstressing). It is possible to

inadvertently exceed the rated maximum average
power of protectors that may be in use within an
EUT if the interval between test surges is too
short. Thus, while surge generation at a rate of
once every few seconds may be an available

The numbers in brackets correspond to those of the references listed in Srction 3 of this standard; when preceded by
A, they correspond to those references listed at the end of this




is of long duration [37],while not excessively limiting the power-frequencycurrent. Both the normal
operating power-line current drawn by the EUT
and the fault current that is available in the event
of internal EUT flashover are significant factors to
be taken into account when test conditions are
set up. See Fig A10 and Fig A l l under the note
ground fault protection for examples of back
filters, and further discussion under the note
fault current.

average power ratings of the internal EUT protectors. For example, consider an EUT protected for
the current surge and typical energy levels of
Category B (Table Al). If the EUT is tested to the
Category A levels, the repetition rate can be
higher than if the test were made to a Category B
Note that service requirements for the EUT can
include multiple pukes. For instance, daring a
single lightning flash there might be several
strokes; in the case of switching surges, several
can occur within a fraction of a second when
abnormal switch behavior is involved.

blind spots. Blind spots can exist in EUTs that

contain surge protective devices. The protective
device performs well at maximum stress (voltage,
current, rise time), but at some intermediate level
the protective device might not perform as intended and the circuits expected to be protected
might in fact be subjected to greater stress than
at the maximum surge levels.
As a first example, consider the EUT of Fig
A2(a) containing a crowbar-type of protective
device, CT. Assume that an impulse voltage of up
to 6000 V is to be applied to the EUT and that the
crowbar has a breakdown voltage of 1000 V. The
application of a 6000 V test impulse operates the
crowbar on the leading edge of the surge; only a
small surge remnant reaches the protected circuit and the equipment survives the test. On the
other hand, a test voltage of 950 V does not operate the crowbar; a large surge let-through reaches

back filter. A back filter is defined in this document as a filter inserted into the ac power line
supplying the equipment to be surge-tested. Such
a filter is required to prevent the surge (assumed
coupled in shunt to the ac line input terminals
of the EUT) from travelling upstream toward
the power source, where it might damage other
devices connected to the same power source. In
addition, in the absence of the filter, the low
impedance of the power source [22] wouid load
the surge generator, which might therefore be
unable to deliver the high peak voltage required
by the test plan to the EUT.
A compromise is necessary, however, between
having a filter that presents an adequately high
impedance to the surge, in particular if the surge

Table A1
Description of Surge Environment in ANSI/IEEE C62.41-1980 [lo]

A Long branch
Circuits and

Comparable to
IEC Pub 664-1980 (171



0.5 ~ s - ~ OkHZ

B Major feeders,

1 . 2 X 50.m
8. x 20 ps

short branch
circuits, and
load center



0.5 p - 1 0 0 kHZ

Medium Exposure

Energy (joules)
Deposited in a Suppressor.
with Clamping Voltage of
500 V
1000 v
( I 2 0 V System) (240 V System)

of Specimen
or Load

6 kV
200 A

High impedance?
Low impedance?,

6 kV
3 kA
6 kV
500 A

High impedance?
Low impedance?
High impedance?
Low impedance?,





*Other suppressors having different clamping voltages would receive different energy levels.
?For high-impedance test specimens or load circuits, the voltage shown represents the surge voltage. In making simulation tests, use
that value for the open-circuit voltage of the test generator.
?For low-impedance test specimens or load circuits, the current shown represents tbe discharge current of the surge (not the shortcircuit current of the power system). In making simulation tests. use that current for the short-circuit current of the test generator.
The maximum amplitude (200 or 500 A) is specified, but the exact waveform will be influenced by the load characteristics.

Note that ANSIAEEE C62.41-1980[lo]states in 6.2that the values shown in the table above represent the maximum range and
correspond to the medium exposure situation of the exposures defined in that standard. For less exposed systems, or when the
prospect of failure is not highly objectionable, one could specify lower values of open-circuit voltages with corresponding
reduction in the discharge current.





Therefore, it will be desirable to test the EUT

over the complete range of impulses for the system, whenever possible, in order to reveal such
blind spots.

3 i i - p ~ ~ ~

common mode (and normal mode). ANWIEEE

Std 100-1984 [3] gives the following definitions of
common mode and normal mode (and information on the mechanisms leading to the coupling of
these unwanted signals):
Common mode noise (longitudinal) (cable systems in power generating stations). The noise voltage which appears equally and in phase from
each signal conductor to ground.. . .
Normal mode noise (transverse or differential)
(cable systems in power generating stations). The
noise voltage which appears differentially between
two signal wires and which acts on the signal
sensing circuit in the same manner as the desired
signal.. . .
In the context of describing surges in ac power
distribution systems, it is useful to examine how
the concepts first defined in communications circuits can be applied to power circuits. Consider
the circuit of Fig A3(a), showing a simplified typical balanced circuit in which the sender and
receiver transformer windings have a grounded
center tap. Note that no metallic conductor is
explicitly provided between the two ground connections. A surge current external to this circuit
along a path including the two points of ground
connection can couple simultaneous impulses on
the two transmitting conductors. Voltmeter VI,
connected as shown, will not read a voltage difference, while the voltmeters V, will. This is classified as common mode coupling.
On the other hand, if a surge is coupled onto
only one of the wires, as shown in simplified form
in Fig A3(b), then all the VI and V, voltmeters will
read a voltage difference. This is classified as
normal mode coupling.
There is a difference in the common mode of ac
power circuits, because in these circuits a grounding conductor is provided, in contrast to the balanced communications circuits in which it is not.
When simultaneous impulses are coupled in the
circuit of Fig A3(c) by an external surge, all conductors are involved. In this situation, simplified
here by assuming equal coupling with all conductors, none of the voltmeters shown would read a
voltage difference. This is in contrast with the
case of Fig A3(a), where both Y2s would show a
voltage difference. On the other hand, if a remote
ground were to be provided, as shown in the dotted line in Fig A3(c), then one could have the

(a) With Single Protective Device



- 1

To Protected

I -

(b) With Multiple Protective Devices

Fig A2
Example of Blind Spot Condition

the protected circuit and the equipment might be

damaged. A similar situation, with large surge letthrough, can occur at higher test voltages than the
crowbar breakdown if the protected circuit has
too low an impedance to allow the tube to operate.
As a second example, consider the multiple protective device arrangement of Fig A2(b) with a
crowbar, CT, and a clamp-type protective device,
P, separated by a series combination of resistance
R and inductance L. This combination can be
implemented in a single package incorporated to
the EUT, or can be implemented with separate
components working, as shown in Fig A2(b), but
actually some distance apart in the case of a complex EUT. When the proper components are
selected for a n application with a specified surge
current amplitude and waveform, the voltage drop
across R and L added to the clamping voltage of P
becomes sufficient to cause sparkover of CT,relieving protector P from diverting the full surge current. However, at some intermediate surge current
amplitude or at some slow rise of the current, the
voltage developed across R and L might be insufficient to produce sparkover of CT so that clamp P
would then have to carry the full current surge, an
event that might not have been foreseen if the
design or test parameters were limited to maximum values.

C62.45 - 1987



- -


L dildt

Voltage added as a
result of surge
current flow
through protective device

1j k----



roun ing



* Grounding
to earth


r @
1 0
1 0 8





Fig A3
Common Mode and Normal Mode
in Communications Circuits
and in Power Circuits

same definition for common mode as given for Fig

For normal mode coupling, the same definition
may be applied to communications and to power
Asurge protective device installed at the end of
a long line for the purpose of protecting the
equipment at the end of the line can produce a
voltage between neutral and grounding conductor, in addition to the clamping voltage between
line and neutral, when a surge current is flowing
in the line (Fig A4). The inductive voltage drop,
L di/dt,caused by the surge current in the neutral conductor, can elevate the neutral terminal of
the equipment above the grounding conductor
potential and cause stress of neutral-to-ground
insulation. In such a situation, the addition of a
protector between neutral and grounding conductor will limit this voltage. Adding the protective device in this location, as distinct from the
alternative of adding it from line to the grounding
conductor, minimizes the amount of steady-state
power frequency current injected into the ground
plane of the equipment to which the grounding
conductor is connected.

communications interface. While the power line

environment of equipment might be in a low or
moderate exposure or duty level, the addition of

Fig A4
Difference of Potential Between Neutral
and Grounding Conductor Resulting at Line
End from Surge Applied at Line Beginning

signal lines with undefined ground potential

changes constitutes an additional threat to survival of the system in its normal environment.
Examples of this situation are a TV set with a
cable TV signal input, a process computer with
far-reaching sensor lines, and communication
equipment powered by the ac line and connected
to the telephone wire-line plant.
Therefore, tests limited to power-line impinging
transients might not ensure adequacy of the
system (see acpower interface).A central project or system responsibility should be provided to
guard against uncoordinated practices and incomplete test specifications.

controlled overvoltage situation. IEC Pub 6641980 [ 171 proposes a method of insulation
coordination based on clearances in the wiring
and connected equipment. Overvoltages can be
controlled in many ways. In addition to interface
devices provided specifically for this purpose, the
existing clearances can also act as voltage-limiting
means by flashing over. A preferred series of
values is specified in Table I of IEC Pub 664-1980
[ 171.Accordingly, an orderly and controlled staircase of voltages would be established, progressing from the service entrance toward the load
(see Fig A17).
In present systems, a staircase also exists since
clearances in present equipment generally decrease from service entrance equipment to wiring
receptacles and electrical load apparatus. The
problem is that the present system is uncoordinated. In other words, interfaces do exist in
present systems but have not been recognized as
such and are neither controlled nor coordinated.



coupler. The coupler is a device, or combination

of devices, used to feed a surge from a generator
into powered equipment, at the same time limiting the flow of current from the power source into
the surge generator. Typically, capacitors are used
to perform the coupling function. These capacitors should be noninductive, capable of carrying
the pulse currents involved, and have appropriate
voltage ratings. Precautions are necessary to ensure operator safety against any trapped charge
that might exist in these capacitors after they
have passed on the test surge.
coupling gap. A convenient method for coupling
the surge generator to the EUT only during the
surge is to use a gap that sparks over during the
initial rise of the pulse. The gap subsequently provides direct coupling and disconnects the generator from the EUT after the surge.
However, it should be recognized that this
method applies to the EUT a voltage wave with
very steep front as the gap sparks over. Some
EUTs might respond to this unintended steep
front in a manner that could be misleading-for
example, a logic circuit upset caused by the steep
front. Selecting a gap with very low sparkover voltage and fast response will minimize the problem.

current surging. Surge current flow, especially in

ground and neutral conductors, can have particularly severe effects on connected equipment.
Damage is usually due to the fact that several
interconnected circuits or devices are not grounded at the same point. While this aspect of surge
testing is not directly related to the explicit scope
of surges impinging from the ac power line interface, experience has shown that one of the causes
of common mode voltages appearing on ac lines
is lightning ground current flow that couples into
the ac system. Explicit ground and neutral current surge testing can be a separate requirement
in many cases, therefore, during the testing of distributed subsystems such as circuit boards and
other subsystem elements.
Fig A5 shows one of the ways in which a surge
generator can be used to simulate ground current
flow. Two devices, EUT #1 and EUT #2, are interconnected by signal or power leads, or both. In
addition, they are also connected to a common
system ground at points G , and G , .
Significant ground current can flow due to an
external surge. This current may be simulated by
applying the output of surge generator between
points G , and G , on the ground system, after first
disconnecting all other equipment not included in

Fig A5
Current Surge Test Example: Surging Ground or Common



Signal andicr Power






prior to

Surge Is




Ground to



the test. The voltage, E,, between points G , and

G , on "ground" will be a function of the resistance,
R,, and inductance, L,, in the line between these
points and the surge current, I,:


Example. To illustrate the need for performing

tests with well-characterized samples of protective devices, refer to Fig A6. A protective scheme is
shown, with a gap device as primary protector,
followed by an impedance and a secondary protector, which may be an avalanche junction device
or a varistor. This scheme is predicated on the
expectation that, for large currents, the clamping
voltage of the secondary protector plus the voltage drop in the impedance will reach the sparkover voltage of the primary protector. This coordination requires evaluation with the primary
protector at the upper limit of the tolerance band
and the secondary protector at the lower limit in
order to reveal any blind spot in the coordination.
A test made with the reverse situation (low primary with high secondary) might provide misleading optimistic results. This situation parallels
the concern discussed under blind spots. A volttime curve would also be useful to define the circuit performance.
The selection of sample protectors at the limit
of tolerance band may not always be possible. An
alternative approach is to determine from manufacturers' data the level of the surge remnant at
the upper limits, remove the randomly selected
protector from the circuit, and apply from a test
generator a voltage waveform and level duplicating the remnant voltage determined above. This
approach is a significant simpllfication of the test
program, provided that reliable information on
the surge remnant is available.

+ L,(dI,/dt)

The test surge current should be a function of

the installation or location category in which the
EUT is assumed to be located, as well as the
values selected for L , and R , in Fig A5 (see [24] ).
In the present state of knowledge, these parameters are not well defined.

current transformers. As described in 5.5.2, current transformers offer the advantage of isolation
between the test circuit and the monitoring oscilloscope. However, the current transformer used
in this case must have adequate response and be
inserted in the EUT with adequate insulation for
the voltages involved in the surge test:
(a) The transformer has adequate response
time to assure that the surge will be faithfully
(b) The transformer will not overload (saturate) because of the high current encountered.
(c) Adequate insulation is provided between
the wire carrying the current to be measured
(primary lead) and the viewing coil (secondary
(d) Adequate insulation is provided between
the viewing coil case to any other nearby component or lead at high voltage.
(e) No undesired magnetic or capacitive coupling should exist between the viewing coil and
the EUT component, other than the magnetic
coupling to the primary lead.

diagnostic test. I t might be necessary to attempt

reproducing, in the laboratory, surge failure modes
that have occurred in the field. The next steps are
to modlfy the equipment to withstand the surges
that caused the failures and then to retest to
ensure success of the redesign. This sequence of
activities is referred to as diagnostic testing.

design test. An important point about design

testing when it is done during the engineering
design phase of an equipment development is
that it may include testing for failure as well as
testing for success. By extending the applied
surge stress beyond the specification the equipment is required to meet, the protection design
margin can be evaluated. This extended stress may
also include repetitive tests to establish the EUT
limits on stress recurrence (see average power).
When protectors exhibiting performance variations are included within the EUT, protectors
with limit values should be substituted during the
test program. Limit values are usually selected to
be at the 3 U point in their expected clamping or
crowbar ranges, at the 99th percentile, or at some
other worst,-caselevel recommended by the manufacturer.

Fig A6
Example of Protector with Multiple Devices




Commercial oscilloscope preamplfiers offer a

wide choice of differential mode operation. One
method is to use a two-channel preamplifier in its
add mode, with one channel inverted. This method
is limited by the capability of the preamplifier for
high-frequency common mode rejection.
A preferred method is to use a true differential
preamplfier built specifically for high common
mode rejection. Many such preamplfiers achieve
this objective at the expense of bandwidth; therefore, the preamplifier should be selected from
those having both a high common mode rejection
ratio (CMRR) and appropriate bandwidth for the
impulses being monitored at high frequencies as
well as at dc. Furthermore, CMRR specifications
are sometimes given for dc, so that additional
care is required in selecting the preamplifiers. The
purchaser should consult the oscilloscope manufacturer for assistance in making the appropriate
For the measurement to be valid, the two
attenuator probes for the differential input have
to be matched; that is, their attenuation must be
equal over the frequency range of interest. Typical
high-attenuation probes are provided with a compensation box that makes possible such a match,
when properly adjusted. A complete single-unit
differential probe may also be used if available in
the complement of the laboratory equipment.

Since it seeks to cure one or more specific surge

failure modes, it may involve tests that would not
be included in a normal qualification test program. This guide is therefore intended to be broad
enough to serve as a basis for extension from
qualification testing to diagnostic testing as may
be required.

differential connection. Since the ground or

common points within an EVT are likely to
assume different potentials during the flow of the
surge current, erroneous measurements would be
obtained if single-ended input circuitry were used
for instruments measuring the surge within the
Probes with safe peak-voltage margins for at
least twice the applied surge peak are required.
Ordinary low-voltage oscilloscope probes are
inappropriate, even if the protected circuit peak
voltages are thought to be just a few hundred
volts. This is due to fault conditions, in which an
internal EUT flashover, protector failure, or other
malfunction might apply enough voltage to destroy the probe, the monitor device input circuits,
and possibly even other equipment if the voltage
can enter the laboratory ground system via this
route. Consult the probe instruction sheet or consult the manufacturer before using these probes
in the test circuit. Commercial high-voltage probes
are available with sufficient voltage ratings for
surge tests on low-voltage equipment.
Oscilloscope (or other monitor) common mode
rejection should be carefully checked. Checking is
best accomplished with both input probes connected to the same point, monitonng first with
the point of application of the surge, and then the
EUT ground. In both checks, the oscilloscope or
monitor readings should be small compared to
the surge amplitude that is finally measured. Of
the two methods given in Figs 3 and 4, the use of
current probes (Fig 4) has almost no noise or
common-mode pickup. However, the current reading must be converted to voltage, which is subject
to errors if the current and voltage are not in
perfect phase or the resistance is not a constant
value with frequency.
Careful attention to the details of the instrumentation is required in order that the measurements obtained reflect only the electrical signals
present at the points of interest, and no more.
Oscilloscope probes have to be identical in model,
length, termination, and compensation; the probe
leads should be twisted together to minimize
induced error voltage.

effective output impedance. For convenience,

an effective output impedance is defined for a
surge generator by calculating the ratio of peak
open-circuit output voltage (OCV) to peak shortcircuit output current (SCI), or OCV/SCI. Note,
however, that this is only a convenient simplification because the peaks of the voltage and current
waves are generally not simultaneous. For instance, for the three waves of ANSIAEEE C62.411980 [ l o ] ,these effective output impedances are
given in Table A2.

Table A2
(Peak Open-circuit Voltage)
(Peak Short-circuit Current)

Category B Impulse

6kV/3 kA




direct connection of the output E , across a 220 V

ac line, would be one sixth of 24 kW calculated
above for the simple 2 R output impedance, that
is, 4 kW. Thus, isolation of the surge network from
the ac line is clearly a requirement.

A judicious addition of circuit reactance,

either alone or in combination with nonlinear circuit elements, can be used to eliminate two potential problems:
(1) Excessive pawer dissipation in the network
of the surge generator.
(2) The unreasonable drain it might impose on
the power source supplying the EUT. Often, a
simple surge-coupling capacitor can perform this
For instance, it follows from Table A2 that,
while there is a range of 15 to 1 in the OCV/SCI
ratio, all of the waves derive from sources that
might not tolerate sustained imposition of ac line
voltage. With 220 V, for example, the dissipations
t h a t would result if the effective output impedances were purely resistive range from over
1.6 kW for the 30 R of Category A ring wave
source to over 24 kW for the 2 R of Category B
impulse source. These situations would seldom be
practical for either the surge generator of the
power line.
To illustrate this situation, Fig A7 shows a
simplified test surge generator output waveshaping network. C is the energy storage capacitor,
charged to peak voltage E,. When switch S is
closed, the test surge is generated across shunt
resistor R , with its leading edge shaped by inductor L. R, limits the output short-circuit current
to whatever value is required.
In Fig A7, if R , is 2 R (as in the case of the
Category B impulse shown in Table Al), and if R ,
is 5 * R,, for example, then the total R , + R ,
dissipation prior to closure of switch S , due to

equipment under test (EUT). In this testing

guide, EUT is a shorthand way of describing any
electrical equipment or group of equipment that
is ready to be exposed to a prescribed series of
overvoltage or overcurrent tests, or both. Some
examples of what an EUT could be follow:
(1) A single component
(2) A single group of protective devices
(3) A known protected or unprotected circuit
(4) An unknown protected or unprotected circuit (sometimes called a black box)
(5) Appliances
(6) Frames of electrical equipment making up
a system or subsystem
exposure. The selection of an exposure level may
be made either from the point of view of a specific
application, and as such would be the prerogative
of a user seeking proposals for equipment to be
installed in his facilities, or it may be made from
the point of view of an equipment manufacturer
seeking to offer a product for installation in a variety of locations. Selection of either will be guided
by economic limitations and will require a compromise between high withstand levels for maximum reliability and moderate withstand levels for
reasonable cost. Overprotection based on a few
difficult cases would penalize the majority of
As indicated by the sloping lines of Fig A8, the
high surges are less frequent for any given exposure; therefore, a limit has to be selected beyond
which failure or misoperation of speclfic equipment is deemed acceptable (under specified conditions of safety). This level will depend on the
type of application, on the consequences of failure
(direct or indirect), and on the cost of added protection for a few high exposures compared to
lower cost with lower withstand for lower exposures.

Fig A7
Simplified Test-Surge Generator
Output Wave-Shaping Network

fault current. It is conventional in performing

stress tests on powered equipment to spec@ a
minimum value of the fault current that will flow
from the power line in the event of flashover or
an arcing fault. Specified minimum fault currents
for residential circuits below 600 V will usually
range from 1- 10 kA, with higher requirements for
industrial circuits, up to 100 and 200 kA. However, the requirement for filtering during a surge


Energy storage capacitor

Initial charge on capacitor
Surge-generating switch
Wave-front shaping inductor
R I Shunt discharge resistor
R, Output current-limiting resistor
E, Output voltage across output terminals


C62.45- 1987





For t,esting protective devices that involve a

power-follow current, it would be desirable to
perform the test with full power-follow current,
that is, the available fault current. More powerful
surge generators and filters with lower impedance
can be used if required.
For EUTs other than the protector, the available fault current will flow only if the EUT has
failed, in which case the exact level of the fault
current may be of secondary concern, unless the
purpose of the test is explicitly to investigate failure modes and effects. As a matter of fact, in this
situation, it may be advantageous to perform the
test with reduced fault current in order not to
completely destroy evidence for the post-mortem.
Therefore, it is important to recognize the effect
of the unavoidable back filters and to make appropriate allowances for the difference between a
test with filters inserted into the line with the
surge injected downstream (away from the source
of power) from the filters, as opposed to performance in actual applications where the surge is
coming from the power system.


In some locations, sparkover of clearances may limit the


ground fault protection. If ground fault circuit

interrupters (GFCIs) are employed in the test
laboratory, difficulties may be experienced with
the back filters in the ac power lines. GFCIs
operate in a single-phase system to shut power
down when line and neutral currents differ by
more than a few milliamperes.
Since normally only a small voltage exists
between neutral and ground, a capacitor between
them (such as C , in FigA9 and Fig A10) will draw
far less current than one from line to ground
( Ci3). If the line-to-ground capacitor is omitted,
even tens of microfarads between neutral and
ground -C , -yield an effective imbalance current of only a few milliamperes due to the minimal potential difference normally existing between neutral and ground. By contrast, if the
line-to-ground capacitor C, is included, values as
low as 10-100 nF can produce a current sufficient to actuate the GFCI due to the line-toground potential difference that normally exists.
However, for some surge test configurations, omission of the line-to-ground filter capacitor will
result in higher residual transients on the nominally unsurged line, unless other capacitor values
are increased to compensate.
An alternative configuration to employ in more
stringent GFCI applications is shown in Fig A1 1.
Use of the power line isolation transformer, T,
permits virtually any degree of capacitor loading

Fig A8
Rate of Surge Occurrences versus
Voltage Level at Unprotected Locations
for Different Exposures

test generally implies the insertion of the series

impedance of the buck filter in the line. Typical
inductor values for L , and L , in Fig A10, for
example, range from microhenries to a few millihenries. Even with inductors as low as 0.5 mH, the
sum ofL, and L , constitutes a total impedance of
0.377 fl at 60 Hz. The maximum fault current is
therefore only 115/0.377 or about 300 A, for a 115
V ac line. For filters with higher series inductance,
fault currents will be proportionately lower. Thus,
the test schedule should recognize the effect of
the filter; some reduction of the available fault
current may be unavoidable.
If the back filters include shunt capacitors on
the EUT side of the series inductance of the filters, a flashover in the EUT will cause these
capacitors to discharge into the fault. This capacitor discharge might mask the effect that would
occur in the normal environment where these
shunt capacit,ors would not be present, and in
which only the power frequency fault current
would normally flow.




on both line and neutral conductors - in this case

t h e transformer secondary- with no adverse
effect on the power line GFCI. Insertion of the
impedance of the isolating transformer, however,
decreases the available fault current.
GFCIs are also designed to trip out if the neutral conductor (load side) becomes grounded
through an impedance of less than 2 - 5 a.Most
GFCIs sense grounding of the neutral Conductor
by use of a high-frequency ( 5 - 10 kHz) signal
induced in the neutral conductor. Therefore, a
capacitor connected between the neutral conductor and grounding conductor might cause the
GFCI to trip out (10 p F is about 2 a at 75 kHz).

T Surge coupling transformer

L Inductor (if secondary of T has insufficient inductance
for proper filter design)
C, Line-to-neutral filter capacitor
C, Neutral-to-ground filter capacitor
C, (Optional, Line-to.ground filter capacitor

Fig A9
Back Filter Added to Series Coupler

grounding conductor. The equipment grounding

conductor in this text describes the ground return
path of the E m . This conductor may be a copper
or aluminum wire or a metallic raceway (conduit,
wire tray, etc). In actual use, normal power current does not flow in this path. Momentary current, from line-to-ground faults or from lightninginduced current, can flow in this path.
Some instances might be encountered where a
surge protective device is connected between the
line and the grounding conductor [Fig A12(a)].
When surge tests are performed on such EUTs, it
is imperative that the grounding conductor be
properly connected to the system ground. Failure
to provide this proper connection would cause
the surge current, diverted as intended by the
protective device, to seek a return path in the
direction of the EUT chassis rather than in the
intended direction of the system ground [Fig
A12(b)]. The result of such unintended routing of
the surge current might cause upset or damage to
the EUT, which would not occur if the grounding
connection were made correctly.



Fiiter inductors Only L, is needed for normal mode surging

shown both L, and L, are needed for common mode
Surge generator output coupling capacitor
Line-to-neutral filter capacitor
Neutral to-ground filter capacitor
Line-to-ground filter capacitor (optional)

L, L,



Fig A10
Back Filter Added to Shunt Coupler

Back Filter








grounding practices. Equipment grounding configurations found in EUTs tend to fall into one of
four general schemes, or into some combination
of them:
(1) floating reference. Where the single reference for certain circuits is isolated from earth
and power reference (sometimes called floating
(2) single-point ground. Where only one connection is made to power ground reference.
( 3 ) multiplepoint ground. Where there are
many connections at various points to earth and
power ground references.
(4) isolatedgrund. Where a controlled impedance is installed between the power safety ground
and the signal reference ground.






T Isolation transformer, carrying full EUT input current

All other components identical to those of Fig A9.

Fig A l l
Back Filter and Coupler with Input Isolation
Transformer for GFCI Applications

C62.45- 1987



EUT Chassis

. ) +


l a r t u eN



I _ irouble!

Grounding Conductor

Fig A13
Floating Reference (Floating Ground)

Poor or /

Fig A12
Effect of Poor or Missing Connection
to Grounding Conductor



Facility Ground-2+

(a) Parallel Single-PointGround

Each of these grounding configurations presents

certain problems that have to be dealt with when
a surge test procedure is developed.
As described, a floating ground is isolated from
all other ground references. However, this does
not alleviate the need to provide protection. The
result is usually a configuration similar to that
shown in Fig A13. Notice that the equipment
chassis is bonded to the grounding conductor,
providing the necessary safety protection. It is
also important to realize that Circuit A is totally
isolated from the chassis and the grounding conductor. The important considerations here are
the possibilities of flashover between Circuit A
and the chassis, along with the unintentional
introduction of a ground connection via test equipment probes if these are not used in a differential connection mode.
Several schemes are shown in Fig A14 that
illustrate single-point connections to the power
ground reference, or earth. In Fig A14(a), the
equipment designer has utilized the chassis as the
internal common equipotential reference, and this
chassis g r o u n d is earthed at one point. In Fig
A14(b), the designer has elected to control the
structure of the signal-ground distribution within
the equipment, bonding it to the chassis at the
point selected for earth connection. With this
single-point grounding scheme, surge current flow
in the facility ground path does not pass through
the chassis.


AC Power



I Note:


Facility Groundf

Removal of center circuits for test purposes can

cause an inadvertant break in the ground chain

(b) Series Single-PointGround

Fig A14
Single-Point Grounding

Fig A15 is a representation of multiple-point

ground. As can be seen, ground connections are
made to several points in the configuration. The
distinction between this and single-point grounding is that the various signal references and
grounding conductors are all tied to different
ground points, none of which shares commonality
with the chassis ground. Thus, surge currents in
the facility ground can flow in and out of such
chassis, which permits voltage buildup exposing
the circuits in the chassis to flashover, upset, or



installation category. The two basic documents,

IEC Pub 664-1980 [17] and ANSI/IEEE C62.411980 [lo],call for relating the surge environment
to the physical/electrical position of the equipment in the power system. ANSI/IEEE C62.411980 [ l o ] defines location categories, and IEC
Pub 664-1980 [ 171 defines installation categories. (Subsequent IEC publications will refer to
this concept as overvoltage categories.)

Again, it is emphasized that these descriptions

do not aim a t recommending any particular
scheme for better noise immunity; rather, they
are described to forearm the individual performing surge tests on an existing EUT. Many EUT
designers strive to avoid noise problems by controlling the ways by which their circuits are
grounded (see 1281, 1321, [34], and 1421).

hybrid protector. This term is sometimes used to

describe a packaged surge protective device consisting of a parallel-series combination of different
devices. The use of the term is deprecated in favor
of the term multiple device protector.

The ANSI/IEEE C62.41-1980Concept. The three

categories,A, B, and C, defined as location categories in ANSI/IEEE C62.41-1980 [ 101 (see Fig A16
and Table A1 in this document), are aimed at
simpllfylng the choice between a high withstand
required for outdoor or near-outdoor systems
and the lower withstand sufficient for equipment
connected inside a building at some distance from
the service entrance. Depending upon the nature
of the equipment, one may predict that some will
practically never be exposed to the Category C
environment, and thus not penalize the user with
the cost associated with protection against that
Small appliances, for instance, can be expected
to be connected in Category A locations, but
exceptions can be found where these small appliances will be connected close t o t h e service
entrance and thus be assigned a Category B duty.

Fig A15
Multiple-Point Grounding


Fig A16
Location Categories Defined in ANSI/IEEE C62.41-1980 [lo]

A nI T i B l


Service Connection

Location Categories
A. Outlets and Long Branch Circuits
All outlets at more than 1 0 m (30 ft) from
Category B with wires R4 - 10
All outlets at more than 20 m (60ft) from
Category C with wires X14 - 10

B. Major Feeders and Short Branch Clrcults

Distribution panel devices

C. Outslde and Service Entrance

Service drop from pole to building entrance

Bus and feeder systems in industrial plants

Run between meter and distribution panel

Heavy applianceoutlets with "shod" con.

nections to the service entrance

Overhead line to detached buildings

Lighting systems in commercial buildings


Underground lines to well pumps



concept. The similarity, which is not an analogy, is

in the recognition of some decrease in the surge
environment severity from the outdoor environment toward t h e inner recesses of a building. The differences originate from the fact that
ANWIEEE C62.41-1980 [ 101 describes the condition where existing, uncoordinated clearances
and solid insulation determine the maximum
overvoltages that can occur, but in an uncontrolled manner (Fig A8). IEC Pub 664-1980 [17]
addresses primarily a coordinated insulation system where the voltage breakdown levels of clearances and solid insulation are known and controlled.
Voltage surges are the significant factor for
determination of clearances and solid insulation,
and hence insulation coordination. Therefore, in
the present uncontrolled situation, ANWIEEE
Std C62.41-1980 [ 101 conservatively recommends
consideration that voltage surges can propagate
unattenuated in an unloaded system [30], maintaining the same voltage level in both Categories B
and A. In IEC Pub 664-1980 [17], the need to
provide guidance for a simple and conservative
installation code might lead to the interpretation
of treating all branch circuits as Category 111,
unless a special surge protective device is selected
to lower the surge voltage to the Category 11 level.
Thus, concerning voltage surge levels in branch
circuits, there appears to be no disagreement
between the two concepts.
Current surge levels are the significant factor
for determination of the duty imposed on surge
protective devices. ANSI/IEEE C62.41-1980 [ 101
recognizes the increasing series impedance from
Location Category B to Location Category A, and
consequently proposes lower current levels in
Category A. Thus, when surge currents can flow
through a surge protective device installed in the
uncontrolled environment of ANSI/IEEE C62.411980 [ l o ] , a current staircase might be considered while, according to IEC Pub 664-1980 [ 171, a
voltage staircase can be considered in a controlled environment. In the development of an
application guide for IEC Pub 664-1980 [ 171, now
under way, the current surge levels are being considered for the design of the interface devices.
Thus, while the two concepts address different
aspects of surge protection, there is no conflict
between them.
These considerations have the following implications for speclfylng EUT performance levels:
If the EUT is the fixed wiring of an installation
(the concern of code-writing bodies), it is difficult

Most heavy industrial equipment is installed within

buildings served by a power system fed from a
substation with stepdown transformers, underground low-voltage feeders, and arresters on the
high-voltage side. This type of installation is therefore less exposed to lightning-induced surges in
the secondary; only primary-induced surges, as
coupled by the transformers, would affect the
low-voltage circuits. However, the occurrence of
switching surges associated with larger load being
switched might increase the severity of the exposure. Some larger installations might use extensive ground mats and multiple-ground practices,
which are more apt to inject surge ground currents into the system than into a smaller system
with single-point ground. A Category B environment could be appropriate.
Many commercial, office, and industrial-parktype facilities use a low-voltage power system
intermediate between residential and heavy industrial, with pole or pad-mounted transformers and
some overhead exposure. There, also, the Category B environment is likely to be found.

The IECPub 664-1980 Concept. The four installation categories of IEC Pub 664-1980 [17],I, 11, 111,
and IV (see Fig A17 in this document), are
intended to allow exact specifications of the category for which particular equipment is suitable.
This means that equipment could be designed to
have known surge voltage withstand capability. It
should be kept in mind that in the absence of such
specfications, as in present systems, the surge
voltage withstand capability is unknown since it
depends upon the selection of clearances and
solid insulation by the designer. Thus, in present
systems, insulation coordination is unlikely.
When specfic information is available on the
exact nature of the interface devices that provide
the staircase of voltages, more precise evaluation
can be obtained by considering the total system;
starting with an initial assumption for the impinging surge at the first interface, the remnant voltage at subsequent levels can then be determined
and compared to the assumed 1.2/50 ps voltages
defined in IEC Pub 664-1980 [17] for each category. This comparison is necessary because the
remnant voltages will not have the 1.2/50 ps
waveform assumed for the impinging surge (see
design test).
Similarities and oiffwences. As illustrated in
Fig A17, there are both similarities and differences between the ANSI/IEEE Std C62.41-1980
[ l o ] concept and the IEC Pub 664-1980 [17]



THE ANSI/IEEE C62 41-1980


Outside and Servlce Entrance

Major Feeders a n j Short Branch Circuits

Outlets and Long Branch Circuits

10 kV or more


6 kV Ring

6 kV Impulse or Ring
10 kA or more


3 kA Impulse 500 A Ring

200 A Ring


cable service

See Note 3



M Watt hour meter

P , Surge arrester

(secondary rating)
Surge arrester
(secondary rating)


Arc welding supply

Fixed appliance
Industrial drive system
Driven motor
Main breaker
Transient protector
SE Service entrance (may take many forms
depending on specific case of system)
WR, Wall receptacle without attenuation

CA Cord connected appliance

COMP Computer with buffered input
ICs Industrial control system
LC Line power conditioner
P Surge protector
WR: Wall receptacle with
attentuation provided
by 2-series impedance
C-shunt impedance




Consumer electronics
Surge protector
Personal computer
Optional built-in
surge protector

I l l (4)25hV

Primary supply level Overhead lines
and cable systems including dislribu
tlon bus and 11sassociated Overcurrent
protection equipment


I L-


Fixed installation following

installation category IV

11 (2 5)t 5 kV
I (1 5 ) O B kV

Appliances portable equ pment etc

lollow ng nitallation category 1 1 1


Special equipment l o r parts 01 equ p
men1 lollow ng nslallation category I1
telecommunicat on electron c etc

Fig A17
Similarities and Differences Between the Location Categories Concept of
ANSI/IEEE C62.41-1980 [lo] and the Installation Categories Concept of
IEC Pub 664-1980 [17], Applied to a Typical Example




Notes for Fig A17

The ANSIAEEE C62.41-1980 [ l o ] Concept of Location Categories in Unprotected Circuits:
.The voltage levels shown in the three location categories represent high impedance circuit conditions: light loading and no
surge protective devices, PI, P,, P, or P4. The 10 kV voltage of Category C is reduced to a maximum of 6 kV in both
Categories B and A by the likely sparkover of clearances, should a 10 kV surge impinge on the service entrance.
.The current levels shown in the three location categories, in a descending staircase from C to A, represent low impedance
circuit condition for surges, such as the installation of one or more surge protective devices, P,, P,. P,, or P,. Another low
impedance condition is the case of equipment sparkover (installed equipment in an actual system or EUT during a test).

If multiple surge protective devices are installed on the system, the current waveform imposed on the downstream protective
device is influenced by the clamping characteristics of the upstream device (see blind spots).

Typical Examples and the IEC Pub 664-1980 [17] Concept Notes:
(1)The controlled voltage situation of IEC Pub 664-1980 [17] requires the presence of interfaces; these can be surge
protective devices such as PI, P,, P, or P,, or the existence of well-defined impedance networks such as Z and C shown
in the circuit diagram upstream of WR,.
Surge arresters or protectors PI. P,, P, and P, may be any protective device suitable for the surge current levels
expected at that point of the system. PI and P, are shown connected line-to-ground. P, and P, may be connected
line-to-neutral. or be a combination of line-to-neutral with additional neutral-to-ground (see common mode).
Surge arrester P, may also be connected on the load side of the main circuit breaker ME. In that case, MB would then be
considered to be in Installation Category IV.
(2) Voltage levels following the designation of Installation Category (IV, I l l , 11, or I ) are shown in parentheses for a system with
300 V phase-to-ground voltage, and outside of parentheses for 150 V phase-to-ground voltage. The voltages shown are
implied as 1.2/50 ps impulses.
Example: IV: (6)4 means 6 kV 1.2/50 ps for a 240 V system; 4 kV 1.260 ps for a 120 V system. See IEC Pub 664-1980 (171
for the complete table of levels corresponding to system voltages from 50-1000 V.

(3) This diagram makes no allowance for the possibility of surges associated with ground potential differences that may
occur, for instance, with a sensor connection to the I C s control system, a cable TV connection to the line-isolated TV set,
etc, or the flow of ground current in the impedance of the grounding conductors.
(4) Transient protector P in the line feeding the welder AW (a typical example of transient generator internal to the system) is
intended to protect the system from the welder, rather than to protect the welder from the system.
(5) Power line conditioner LC, while performing the major task of conditioning the power supply to the computer, might
perform a function similar to that of the protector P at the welder in blocking conducted interference from the load toward
the system.
(6) Many appliances or electronic devices might be equipped with internal surge protective devices and therefore be suitable
for installation in other categories than II.
(7)The use of the term installation category in Fig A17 and the text is based on IEC Pub 664-1980 [17]. This term will be
supplanted by overvoltage categories in subsequent IEC publications.
Note: Independent of its location in the above figure, a device or equipment should remain safe (no fires, no personnel hazard)
over the full range of available surges at any point within the installation. It may also be desirable, under particular circumstances
and for specific devices, t o proscribe damage as a result of testing at higher levels than might be suggested by its typical

to differentiate between short and long branch

circuits. Thus, application of IEC Pub 664-1980
[17] might result in classlfylng all branch circuits
as Installation Category 111.That approach is consistent with ANWIEEE C62.41-1980 (101,where
the same voltage is indicated in Categories B
and A.
If the EUT is the load equipment to be connected at an undefined point of the installation,
conservative design would point toward spec@ing for the EUT a voltage withstand or current
handling capability, or both, based on the most

severe indoor conditions; that is, IEC Category I11

and IEEE Category B.
However, economic factors and the statistical
nature of the surge environment are likely to motivate equipment manufacturers and users to consider lower levels. These lower levels nevertheless
need to be consistent with the most likely use, the
risk, the consequences of failure, and safety considerations. This selection of the appropriate level
is the prerogative of the equipment manufacturer
and the equipment user, both correctly informed,
as discussed under withstand level.



Therefore, the test criteria should include, as

appropriate, an evaluation of this mechanism.

insulation coordination. Insulation coordination

is a concept in which solid insulation in electrical
equipment has an impulse withstand voltage level
in excess of either the protective voltage level of
associated surge protective devices or the breakdown of the clearances existing in the equipment.
This coordination ensures that the surge protective devices or the clearances in air ( a renewable
insulation) protect the nonrenewable solid insulation.
A coordinated insulation system implies the
selection of the electrical insulation characteristics of equipment with regard to its application
and its relation to its surroundings. It is necessary
to consider the voltages that can appear within
the system, the location and characteristics of the
surge protective devices, the continuity of service
desired, and the safety of persons and property,
so that the probability of undesired incidents due
to voltage stresses is reduced to ensure an economically and operationally acceptable performance.
Most existing low-voltage product standards,
based primarily on experience, specify dimensional requirements for equipment design but do
not require tests measuring the actual performance of the equipment insulation. As a result, the
insulation systems of various equipment a r e
uncoordinated and the sparkover of clearances
or the breakdown of solid insulation can occur a t
random locations at unknown transient overvoltage levels. New standards are being proposed that
are based on the concept of insulation coordination and rely on performance tests rather than
the traditional dimension standards.

insulation tracking. The flashover of a clearance,

at the first occurrence or after multiple pulses,
might affect adjacent solid organic insulation or
weaken the long-term withstand capability of the
EUT, and therefore might not be such a benign
occurrence. In addition, a high voltage, either of a
transient nature or continuously recurring, might
result in partial discharges within solid insulation or on surfaces of insulation. Repetitive partial
discharges might result in degradation of the
insulation material.
A post-test examination should be performed
to attempt to ascertain the location and effect of
a flashover that occurred in the EUT. Observation
during the surge will also help in locating a flashover, and a repeat test where flashover would
occur at a lower voltage would indicate some
change in the EUT. Other components subjected
to the surge stress might also be weakened but
appear to perform normallyfor some time, while
others might not even withstand normal line voltage. Therefore, some life testing after surge testing
is desirable.
life consumption. It is possible that a significant
portion of rated protector life might be consumed
during a surge test program (see multiplepulse).
This consumption is far more likely to occur during the engineering phase of EUT development
than duringproduction tests. If it does occur, the
protector(s) may be replaced by new ones during
tests or following. test completion, with a small
portion of the tests then being repeated.
Performing tests with built-in deliberate defects
in the EUT may also provide information on life
consumption and integrated stress for the protectors.

insulation degradation. The majority of the EUT

failures under a voltage transient may be very
broadly considered as insulation failures. This
insulation failure can occur in wiring insulation,
within the bulk of solid insulation, at the edges
of printed circuit boards, within semiconductor
material, or along the etch runs of an integrated
circuit. Two modes of failure might be encountered:
(1) A destructive single flashover that leads to
permanent damage, either by the energy of the
surge alone or by the energy of the power-follow
initiated by the surge sparkover.
(2) A progressive loss of quality of the insulation where several successive sparkovers along
the surface o r p a r t i a l discharges within the bulk
of the insulation eventually lead to grossly visible
and permanent damage.

location category. See the note on installation

low voltage. The term low-voltage in this document has the same meaning as low-voltage system
(electric power) in ANSI/IEEE Std 100-1984 [3],
that is, up to 1000 V rms for ac circuits. For electronic designers, the term may have a different
meaning, but in keeping with ANSI/IEEE Std 1001984 [3], this document uses low voltage only as
stated above.

monitoring. ANSI/IEEE C62.41-1980 [ 101 defines an open-circuit voltage and a short-circuit

current; the impedance of the test circuit determines the choice between voltage and current.

C62.45 -1987


rarily functional, perhaps even wit,hin its specifications. Therefore, some life testing should follow
surge testing.

Therefore, the monitoring of the applied surge

should include both voltage and current capabilities. For instance, if the EUT is presumed to be a
high-impedance circuit, the test would aim at
applying a voltage surge a t a value set by adjusting the surge generator with the EUT disconnected, following which the EUT is reconnected
without a change in the generator setting. If the
EUT is presumed to be a low-impedance circuit
(because it includes a surge-protective device
expected to operate under the applied surge), the
test would aim at exercising that protective device
by the injection of a current surge, with an available generator short-circuit current selected on
the basis of the ANSI/IEEE C62.41-1980 [ l o ]
environment. If the EUT contains surge protective
devices with negative impedance characteristics,
such as a gas tube, the EUT impedance will switch
from high to low during the test, at the time the
gas tube sparks over (see waveshape: voltage
versus current).
Monitoring the test outcome may include such
criteria as voltage withstand of the EUT (absence
of a breakdown that would otherwise chop the
applied wave), absence of corona or partial discharge in the circuits during the impulse, surge
let-through voltage or surge remnant passed on
to devices connected at the output of the EUT, or
any other parameter significant to the evaluation
of a specific EUT.

noise. This guide on surge testing primarily concerns those surges with amplitudes that typically
exceed twice the normal circuit voltages. These
surges are mainly caused by lightning and power
system switching. It is recognized that the range
of amplitudes extends downward, and that switching transients can be injected in ac power lines
with low amplitude but very short rise time (that
is, below 100 ns). These fast transients can produce severe interference problems and should not
be ignored. Another source of interference is
mobile communication equipment, which is increasingly used in industrial plants. The electromagnetic compatibility (EMC) literature and
standards should be consulted for guidance in
this area (see [4], [16], [18], [19], and [ 3 5 ] ) .
normal mode. See common mode.
partial discharge. When an electrically stressed
dielectric structure exhibits one or more areas of
very local field intensification, it is possible that
electrical (corona) discharges might be present in
the high field regions without the failure of the
complete insulation. The most common circumstance leading to the formation of such partial
discharges occurs in solid dielectrics in which
small cavities exist. Under alternating voltage conditions the electric stress in the gaseous voids will
be enhanced because of the ratio of dielectric
constants between solid and occluded gas. Discharges in the gas-filled cavity will commence
when the voltage on the complete structure
reaches a value known as the inception voltage,
which will generally be considerably less than that
required for complete short-term failure. The
magnitude of the discharges, usually measured in
picocoulombs, is dependent on cavity dimension
and their repetition rate on the extent to which
the inception voltage is exceeded. Cavities are
often present in solid dielectrics as manufacturing defects or by virtue of the operation of thermal
or mechanical stresses. In liquid impregnated systems, partial discharges can occur in bubbles
formed in the stressed impregnant fluid. Other
forms of stress enhancement, such as conducting
inclusions, can also generate partial discharges.
Under direct voltage conditions, the distribution
of stress in a solid is determined by the volume
resistivity and not by the relative permittivities.
Under these circumstances the time constant for

multiple pulse. In service, the protectors within

the EUT will be subjected to one or at most a few
surge events at a time. Typically, a long recovery
period of minutes, hours, or even days will follow
before they are required to withstand the next
surge event. However, during thorough surge testing it might be necessary to apply a sequence of
tens or even hundreds of pulses or surge events.
Surging at all network phase angles versus the
ac line and in both polarities on a progressive
stress basis may result in a large number of successive test waves.
In production tests, it is important to limit the
number of surges applied to devices intended for
shipment in order to avoid degradation of the
It is possible that a component would be stressed
during a single transient and continue to perform
normally for some time. The life of this component might be shortened so that it might fail
much earlier than it would have without the transient. The device parameters might be altered by
the transient, but the device would remain tempo43



The likelihood of such an occurrence depends on

a number of factors, one of which is the instantaneous value of power system voltage at the time
of sparkover. If this voltage is low (early or late in
the cycle), power-follow might not result. If this
voltage is sufficient to establish a power-follow
(later, but not too late in the cycle), the powerfrequency current integral is a function of the
time remaining in the cycle until current zero.
Thus, both the occurrence and consequence of a
power-follow depend on the phase angle of the
For semiconductors, the phenomenon appears
related to the conduction state of EUT semiconductor devices at the time the surge occurs. Semiconductor parameters that might be involved
include forward and reverse recovery characteristics and second-breakdown performance. The
devices most likely to fail in a phase-related way
are semiconductors involved in the power-input
circuitry. Others, in different areas of the EUT,
might also exhibit such failure modes if the EUT
power-input circuits let some or all of the surge
pass through to them.

charging the capacitance of a small cavity to a

voltage at which it will discharge is large, and thus
discharges under dc conditions are much less frequent. For surge voltages, the partial discharge
behavior is critically dependent on whether the
impulses are unipolar or involve voltage reversals.
Since charges might be trapped within a dielectric,
reversing the polarity of the applied voltage is
much more severe than a single polarity test.
Partial discharges represent a power loss and
also constitute an aging mechanism, especially for
organic insulation. Over a long period of time the
erosion caused by the discharges and their byproducts at multiple sites can lead to complete
failure of the insulating structure. Large discharges from a plurality of sites may be readily
detected by their integrated effect on dielectric
loss determined by a Schering bridge or loss analyzer measurement. Commercial equipment is
available to detect and measure individual discharges reliably to a sensitivity of better than
1 pC. Location of partial discharges in equipment
is more difficult but may be accomplished by
pulse reflectometry, or by acoustic, optical, or
chemical means. Subcommittee SC28A of the
International Electrotechnical Commission (see
[17]) is investigating test circuits and test
methods for evaluating the withstand capability
of solid insulation.

power-follow. Power-follow is defined in ANSI/

IEEE Std 100-1984 [3] under follow (power)
surge arresters as The current from the connected power source that flows through an arrester during and following the passage of discharge
current. In this guide, in addition to the quoted
definition, power-follow is also used to refer to a
similar current from the power source resulting
from any sparkover in the EUT- intentional or

phase angle. Many surge-related equipment failures depend on the phase angle of the ac voltage
sine wave at which the surge is applied (Fig A18).
When an EUT clearance sparks over during an
impulse (surge) test, power-follow might occur.
Fig A18
Phase Angle Effects

powered testing. There are several reasons for

performing powered testing:
( 1 ) From the standpoint of good practice, it is
best to perform laboratory tests in a manner that
most closely simulates the actual service environment.
(2) It is the applied ac that furnishes the
energy following the surge, that can establish
sustained arcing faults, tracking on insulation, destruction of printed wiring, and so on (seephase
(3) The application of normal ac power generally raises the EUT to an initial level of stress.
Without power current following a surge-induced
flashover, the resulting defect might not be detected. Even repeated, unpowered surging might
not mark the defect well enough to make it


Line voltage is
insufficient to
establish power.follow.

Power-follow can be
established, energy
until next current
zero is high

Power-follow is
established but
energy is low.


Transient in same direction

as reverse bias during o f f
state can be less harmful than..



or oscillating) occurring
during conducting 112 cycle.




production test. It may be desirable to set up an

acceptable quality level (AQL) surge-test program
based, as far as possible, on knowledge of EUT
surge sensitivities.
In production tests, it is important to limit the
number of surges applied to devices intended for
shipment in order to avoid degradation of the
Tests made on some samples in AQL programs
might damage components or the EUT. All samples used in potentially destructive tests should
be scrapped.

particular piece of equipment should be surged

more slowly than the table suggests in no way
implies that it provides less surge protection in
the field. The slower rate applies only during
surge test. There might be repetition rate requirements for surges to be withstood in actual service,
such as multiple discharges, and these should be
determined at the outset if they exist. The repetition capabilities of the surge test generator should
also be considered.
In the case of the EUT being a computer system,
a very high repetition rate at low energy level
might be necessary to check for software susceptibility, whereas a one-shot surge would be
needed to check for hardware vulnerability. (See
further comments under noise.)

qualification test. It will generally be u p to the

equipment specifier and manufacturer to agree
on which tests will be performed. Tests performed
to qualm a new equipment design or a major
modification to an existing one will ordinarily be
more complete than tests carried out on a routine
basis on production products.

series coupling. Fig A19 shows an example of

the series approach. Coupling transformer T is
driven by the test-surge generator, and its secondary is inserted in series with the line side of the
single-phase mains. Note that the line side is, in
effect, being surged in this example with respect
to both neutral and ground. Series coupling has
this property: the line in which the coupling is
inserted is always effectively surged with respect
to all other lines (power or signal, or both) into
and out of the EUT.

repetition rate. A maximum allowable surge

repetition rate cannot be determined without the
evaluation of the EUT protection design. For this
reason it is strongly recommended that the maximum allowable repetition rate for pulse trains of
varying length be incorporated into the original
equipment protection specification. In the absence
of other requirements, it is suggested that the figures of Table A3 be incorporated.
Note that all equipment does not need to meet
the suggestions of Table A3, which are given
mainly for new equipment designs. The fact that a

shunt coupling. Fig A20 gives an example of

shunt coupling. In this case, called normal mode,
the surge is applied via coupling capacitor C
between line and neutral. Shunt coupling is there-

Table A3
Suggested Wait Times as a Function
of Total Number of Surges for Surges Applied
per Categories and Waves of ANSI/IEEE C62.41-1980 [lo]


Wave Type


No. of


0.5 /L.s-~OO kHz

Ring Wave




100 to 1000

6 to 20

0.5 / L S - ~ O O kHz
Ring Wave




100 to 1000

10 to 30

1.2/50 Impulse
8/20 Impulse




100 to 1000



open-circuit voltage
short-circuit current


30 to 120



the source of the surge) is a surge remnant. The

front part of a surge before a protective device
conducts surge current is also a surge remnant.
When an EUT contains built-in surge protective
devices, the EUT should be surge tested with the
stress relevant to the expected performance of
the protective device, with the components downstream of the protective device stressed only to
the voltage level allowed to pass by the protective
device. Whole-system surge tests need be concerned with internal protector remnants only to
the extent that, during engineering or qualification tests, or both, it is usually desirable to employ
limit protectors. Carrying the process a step
further, it is possible then to select the subsequent
protector to ensure that it, in turn, is maximally
stressed. This extension implies selection of a performance point that will generally be the opposite
3 U point from that which transmits maximum
stress to the protector or circuit that follows in
turn. Various combinations of these situations
should be employed to ensure a design that will
remain valid over the entire range of the anticipated protector performance distribution (see
blind spots, design test, and surge let-through).


1 1 I sz


T = Surge Coupling Transformel

Fig A19
Series Coupling of Surge Generator




0 a0l

m a


C = Surge Coupling Capacitor

Fig A20
Shunt Coupling of Surge Generator

susceptibility. A test for susceptibility implies

normal equipment functioning prior to the surge;
therefore, equipment can only be checked in the
powered configuration. In order to segment the
surge test of a larger system, it may be necessary
to use simulations, for inputs as well as outputs.
As a result, an evaluation of misoperation may
require measurements to be taken on either simulated outputs (for performance) or on simulated
inputs (for unusual loadings). In addition, it is
always necessary to consider the possibility that
the EUT might carry some portion of the applied
surge to either inputs or outputs, or both, whether
or not the EUT itself demonstrates susceptibility.
For instance, a power supply might be insensitive
to a surge applied to both of its inputs versus
ground, but the common mode surge might well
pass unattenuated to the circuitry being driven by
the supply.
The importance of susceptibility depends on
the EUT and its function: an audible click in a
telephone system might not be important, but bit
dropouts with consequent corruption of data in a
computer or shutdown of a digital processor
might be highly significant.

fore different from series coupling; other lines to

the EUT need not be intentionally involved with
the surge path. (A flashover during testing might,
however, involve these other lines.)

surge event. A surge event, as understood here,

may be a single pulse or a train of pulses occurring with little time between them (see multiple
pulse). Therefore, for power or energy considerations, the train of pulses should be considered; for
voltage clamping, a single pulse can provide the
necessary information.
surge let-through. A surge let-through is that
part of a surge that passes through a surge protective device or a power supply, with little or no
alteration. For instance, a power supply might be
unaffected by a line-to-neutral surge or a surge
along the grounding conductor; however, if the
power supply allows the surge to pass through,
the circuit it powers will be subjected to this letthrough surge (see surge remnant).
surge remnant. A surge remnant is that portion
of the surge that remains downstream of one or
more devices. The clamping voltage of a protective device, propagating downstream (away from

unforeseen consequences. An unforeseen consequence may be the outcome of a surge test performed in the laboratory, or of a surge occurring



waveshape: voltage versus current. Two different voltage waveshapes and two different current
waveshapes are proposed by ANSI/IEEE C62.411980 [ l o ] ,depending on the installation category
of interest and the nature of the EUT (high or low
impedance). If the particulars of the application
warrant it, a different waveform, longer or shorter,
may also be appropriate (see [26], (331, and data
base in Section A1 of Appendixes to [lo]). If the
characteristics of the specific EUTs are known,
one may spec@ the appropriate test wave, voltage, or current, depending on the input impedance of the EUT under surge conditions. Pass-fail
acceptance tests on insulation, for instance, are
not concerned with events following the insulation breakdown, and therefore the simple voltage
test should be sufficient. Other tests are so concerned, since the EUT impedance might be affected, temporarily or permanently, by the very
test surge being applied. The following aspects of
surge testing electronic systems are relevant to
the selection of a test method:
(1) Testing for failure modes that involve flashover are influenced by the surge current that
would flow after flashover.
(2) The surge let-through of a protective device
depends on the applied voltage front, and the
surge remnant depends on the ensuing discharge
( 3 ) The surge current front and magnitude
after operation of a crowbar device can affect the
coupling of additional voltages in downstream
(away from the source of the surge) circuits.
(4) The response of a crowbar-type protective
device, subjected to an intended current test, will
be influenced by the voltage front applied by the
generator, sensing a high-impedance test piece,
until operation of the crowbar.
(5) The voltage compliance of an intended current test generator will influence the surge remnant of a clamp-type surge protective device.
Therefore, the generator should be capable of a
dual role:
First, for what is nominally a voltage test,
such as the 1.2/50 ps wave ofANSI/IEEE C62.411980 [ l o ] , sudden impedance drop has to be
accompanied by a minimum, specified, surge current to ensure both marking the defect and initiatingpower-follow current. It is also desirable that
the short-circuit current waveshape (such as the
8/20 p s wave) be the prescribed value for the
same reason.
Second, for what is nominally a current test
(such as the 8/20 ps wave), the voltages involved
should be held between a low and a high limit in

in the field and impinging a device for which this

type of outcome might not have been previously
recognized, despite the fact that surge tests had
been performed during the design phases of that
device. Some examples follow:
(1) A power supply is not damaged by the
surge, but passes it downstream causing failure of
connected equipment in the field, a consequence
that was not recognized during a test focusing
only on the EUT.
(2) A device deemed most likely to be installed
in a Category A location might in fact be installed
in a Category B environment where it might fail in
an unacceptable mode.
( 3 ) A solid-state power-control EUT fails in a
mode that leaves power on full; hence, a motor
overspeed, a load burnout, etc.
(4) A surge protective device fails under longterm exposure (failure to reseal, thermal runaway, etc).
The phenomenon is systemic in nature and results from testing an EUT as an isolated device
rather than testing it in its intended application.
Downstream damage results from s u r g e letthrough, which is not observable unless systemlevel testing is performed.

unpowered testing. As a precaution against

severe damage of the test piece that might be difficult to analyze, unpowered testing should always
precede powered testing. For some types of EUT,
where it is judged that powered testing would not
provide additional insights on the outcome, unpowered testing only is quite appropriate. For
instance, in an insulation coordination test, a
clearance sparkover is sufficient to signal a failure
without the more complicated powered test.
vulnerability. Vulnerability refers to EUT damage that has to be repaired before the normal
equipment function can be reestablished. It might
involve components such as semiconductors, or it
might result in open runs (conducting paths) on a
printed circuit board, in arcing between runs that
leaves a heavy conductive carbon track, or in
other destructive phenomena that prevent normal EUT performance.
Testing for susceptibility is elective, and the
need for it should be decided by the user or
designer of the EUT or both, taking into consideration its application; many other criteria are
defined in electromagnetic compatibility (EMC)
standards (see [4], [16], [18],and [19]). In the
context of ANSI/IEEE C62.41-1980 [ l o ] and IEC
Pub 664-1980 (171 testing for vulnerability is
always required.

C62.45 1987


order to respectively initiate current flow by a

sparkover-type device, while not exceeding an
appropriate upper limit. The prescribed opencircuit voltage waveshape (or at least rise time,
such as 1.2 p s ) is preferable to ensure crowbar
operation or internal flashover at a repeatable
Such tests can be accomplished with a surge
generator having the inherent capability of applying a selected voltage wave to a high-impedance
specimen and of delivering a selected short-circuit
current wave into a very low-impedance test specimen. While the need for providing conditions 1
and 2 above might not always be apparent, ignoring these conditions can lead to meaningless testing, for instance, where a surge generator of high
internal impedance or low stored energy would be
applied to conduct a voltage test on an EUT having relatively low input impedance.
Further reference to this need for providing the
dual capability, if and when the test piece experiences flashover, crowbar, or clamping action,
can be found in [41] and [43].FigA21, excerpted
from [43],illustrates the case where the test piece
is a protector containing at least a gas tube, with
the following descriptive comments:
However, if we examine the test requirements more carefully, we observe that
these lines are usually provided with
spark-gap surge arresters that fire at a
few kilovolts. After the surge arrester
fires, it behaves somewhat as a voltage
regulator; the most important parameter,
then, is the current delivered to the surge
arrester. Thus, as illustrated in the figure, it is necessary to simulate the proper
impedance and voltage for these protected lines only until thesurge arrester
fires; thereafter, only the current need be
If a generator with dual capability is not available, separate current and voltage surge tests
need to be performed in succession with two
separate generators. However, this approach
might not detect performance limitations such as
blind spots and energy-sharing difficulties in
multiple-device protective systems (see [41] and
On the other hand, it may be desirable to limit
the surge current after a breakdown under a voltage surge test in order to avoid complete destruction around the breakdown path to the point
where diagnosis might be difficult (this is a different purpose from that of the test for failure




SOURCE Vance Nanevlcz. and Graf [431

Fig A21
Excitation of Nonlinear Protectors
modes and effects). The person conducting the
test should select the waveshape or combination
of waveshapes that represents the type of surges
to which the EUT should be subjected. Both the
nature of the EUT and the nature of the surges
should be considered.

withstand level. As pointed out in the Foreword

and the Scope, assignment of specific withstand
levels is not the purpose of this guide. The selection process will be performed with agreement
among the concerned parties, typically by technical committees concerned with specific products,
or on the basis of the speclfic situation. Note also
that the selection of withstand levels, when clearance flashover is involved, requires consideration
of altitude effects both in speclfylng the level and
in performing the tests. ANSI/IEEE C62.41-1980
[ l o ] mentions in 6.2 that the values shown in
Table A1 represent the maximum range and correspond to the medium-exposure situation of
Fig A8, and states:
For less exposed systems, or when the
prospect of failure is not highly objectionable, one could speclfy lower values
of open-circuit voltages with corresponding reductions in the discharge currents.
In evaluating the results of a test for withstand
level, it is necessary to recognize the statistical
aspects of the EUTbehavior. It is assumed that if
a single test specimen is provided (a frequent




Appendix References

situation when the EUT is a complex and expensive piece of equipment), a specimen has been
selected at random from the population of interest. Typically,this single piece of equipment is subjected to increasing voltage tests until the failure
criterion is observed; the test result is then a single number, the voltage surge or current surge
that produced the failure.
The result for a single random sample provides
a statistical best estima.te of the characteristic in
question for the sampled population. This result,
of course, provides no information about the
variability of this characteristic from equipment
to equipmeni. Therefore, unless other relevant
prior information about variability is available, it
is not possible to make any statement concerning
how close the values of the individual equipment
in the population are to that of the single unit
that was sampled or, more generally, about the
statistical distribution of such samples. To obtain
such information about variability, additional
;amples are required (see [A41 and [A5]).

[All CHEN, C. L. Lightning Protection DewiCes.

FAA Report FAA-RD-74-104,April 1974.
[A21 COHEN, E. J., EPPES, J. B., and FISHER, E. L.
Gas Tube Arresters. International Confwence on
Communications, June 19-21, 1972, p p 43-1 to

[A31 MARTZLOFF, F. D. Coordination of Surge

Protectors in Low-Voltage AC Power Circuits.
IEEE Transactions on Power Apparatus and Systems,vol PAS-99,no 1,JanIFeb 1980, p p 129- 133.
[A41 HA", G. J. Random Samplings: How Large
a Sample Do I Need for 95% Confidence? Chemical Technology, vol5, January 1975, pp 61-62.
(A51 HA", G. J. and SHAPIRO, S. S. Statistical
Models in Engineering. New York: John Wiley &
Sons, 1967, p p 75-76.


C62.45- 1987



Mentioned on Page

Discussed in the
Appendix on Page

AC power interface
Average power (overstressing)
Back filter
Blind spots
Common mode

9, 20
17, 21, 23, 35
19, 32, 41, 46
21, 23, 27, 31, 33, 43, 46


Communications interface
Controlled overvoltage situation
Coupling gap
Current surging

9, 15
15, 21


Current transformers
Design test
Diagnostic test
Differential connection
Effective output impedance

13, 19, 39, 46


Equipment under test (EUT)

Fault current
Ground fault protection
Grounding conductor

13-17, 19-24, 27-29, 31-36, 38-49

15, 28
21, 23, 28, 36
17,21-23,29, 37,46


Grounding practices
Hybrid protector
Installation category
Insulation coordination
Insulation degradation

9, 15
9, 30, 39, 47


Insulation tracking
Life consumption
Location category
Low voltage

15, 23, 33


Multiple pulse
Normal mode
Partial discharge
Phase angle

19, 28, 42
13, 38, 45
21, 27, 45
14, 42, 43
15, 19, 43, 44


Powered testing
Production test
Qualification test
Repetition rate

14, 35, 42, 44, 47

14, 17
13, 19, 42
13, 19
19, 27


Series coupling
Shunt coupling
Surge event
Surge let-through
Surge remnant

17, 43, 46
17, 32, 43, 47





Unforeseen consequences
Unpowered testing
Waveshape: voltage versus current
Withstand level

Mentioned on Page
14, 16
9, 15, 34, 41


Discussed in the
Appendix on Page