martzloff f.d.the protection of industrial electronics and power conversion equipment against power...

7/27/2019 Martzloff F.D.the Protection of Industrial Electronics and Power Conversion Equipment Against Power Supply and Data Line Disturbances

1/14

The Protection of Industrial Electronics and Power Conversion

Equipment Against Power Supply and Data Line Disturbances

Franois D. Martzloff

General Elerctric Com pany

Schenectady NY

Reprinted, with permission, from Conference Record, IEEE ICPS(Industrial and Commercial Power Systems), May 1984

Significance

Part 6: Tutorials, textbooks and reviews

Tutorial presentation including a description of surge origins and standardization projects to characterize

them, including the now-abandoned IEC 664 concept of Installation Categories as contrasted with the

IEEE C62.41 concept of Location Categories. Brief discussions of issues such as the limitations of

transmission-line analysis, pitfalls in shield grounding practices, comparisons of voltage-switching vs.

voltage-limiting devices, and fusing options for failure modes.

Three examples are given in an Appendix on the erroneous expectation that an isolating transformer can

be a one-type-fits-all surge mitigation approach, on the implications of possible configurations of SPD

connections, and a case history of shifting reference voltages with retrofit of an appropriate remedy.


2/14

THE PROTECTION OF INDUSTRIAL ELECTRONICS AND POWER CONVERSIONEQUIPMENT AGAINST POWER LINE AND DATA LINE DISTURBANCESF r a n ~ o i sD. Mar tz lo f f , Fe l l ow IEEE

Corpora te Research a nd Deve lopmen tGenera l E lec t r i c CompanySchenec tady, N ew Yo rk 12345

AbstractThe i r resist ib le trend toward d istr ibuted comp uting systems for t he

contro l o f industr ia l power e lectronic equipme nt is adding new dimen -sions to the o ld problem of surge protect ion. Harmful surges may beproduced by l ightn ing, power switch ing, or b y d i f ferences i n groun dpotentia ls. The object ive o f the paper is to pro vide increased aware-ness and better understanding of the fundam enta ls o f protect ion aswell as the characteristics of protective devices.

1. I n t r o d u c t i o nThe con t i nu ing deve lopment o f power convers ion equ ipment ,

fur ther promo ted by the success o f microelectronics in contro l l ing th isequipment, raises the questio n o f vulnerabi l i ty to surge voltages orsurge currenls appear ing at the power or cont ro l inp ut terminals o f theequipm ent. Th e very origin of these surges (l ightning, sw itching, elec-trostatic discharges) makes i t somewhat difficult to predict wi th accu-racy thcir characteristics, so that surge protection has sometimes beenpresented as "art rather than science." The re is, however, enoughpractical experience and data avai lable to apply go od eng ineering judg-ment i n developing an understanding o f the practica l aspects o f surgeprotect ion based on fundamenta l concepts and on protect ive devicecharacteristics. Th e present paper is offered as a con trib utio n towardbetter surge protection, neither as an art nor as a science, but as a no-nonsense, sound, an d cost-effective engineering practice.

T h e paper presents First a discussion of the various and complexorigins o f l ightn ing and switch ing surges (electrostatic discharges areacknowledged b ut n ot addressed in a l imit ed space discussion). T h epaper also shows h ow standards can help simp li fy the appl ication ofprotect ion. A br ief discussion is given on aspects o f the propagationo f these surges, fo l lowed by a presentation o f some fundamenta lapproaches to protection. Basic characteristics o f prot ective devices areout l ined. Three examples are g iven on how th is knowledge andengineering judgment can be applied to answer specific questions andto avoid p i t fa l ls .

2 . The Or ig ins o f Trans ien t Overvo l tagesTransient overvoltages in power systems originate from one cause.

energy being in jected in to the power system, b ut fr om two sources:l ightn ing discharges. o r swi tch ing wi th in the power system In com-mu nica tion or data systems, there is anothe r source o f transients thecoupl ing of power system transients ln to the system Furth ermor e, al lsystems involving several connections to the environment have thepotent ial r isk of transient overvoltages associated with gro und poten tialr ise dur ing the f low o f surge currents As stated previously, staucdischarpe problems are not treated in this paper.L ig htn ing d ischarges may no t necessari ly mean d i rect term inat ion o fa l ightn ing stroke onto [he system A l ightn ing stroke terminat ingnear a power l ine. either by h ittin g a tall object or the bare earth, wi l lcreate a very fast-changing magnetic field that can induce voltages -and inject energy - in to the loop formed by the conductors of thesystem. Ligh tnin g can also inject overvoltages in a system by raisingthe ground potent!a l o n the surface o f the earth where the stroke ter-minates. whi le mo re d istant "ground" points rema in at a lower vo l r -age. closer to the poten tial o f "true earth." The l i terature providesinform ation o n the character ist ics o f l ightn in g d ischarges [ I-61.

Surges from power systeni switching create overvoltages as a resul to f t rapped energy in loads being swi tched off . or of restr ikes in theswitchgear. These uill be examined in greater deta i l in the fo l lowingparagraphs.2.1 Trans ien ts in Power Sy.tems

A transient is created whenever a sudden change occurs in a powercircuit. especial ly during power switching - either the closing or open-ing. I t is important to recognize the d i f ference between the in tendedswi tch ing ( the mechanical act ion o f the swi tch) and the actual happen-ing in the c i rcut i Du r in g the c losing sequence of a swi tch, the con-CH2040-4/84/0000-0018$01oO(C 1984 IEEE

tacts may bounce, prod ucing openings o f the circuit wl th reclosing byrestr ikes and reopening by clearing at the high-frequency current zeroPre stri kes can a150 occur just b efore t he contacts close, wi th a succes-sion o f c learings at the h~gh -freque ncy current zero, fol lowed by re-str ikes. Sim i lar ly~ dur ing an opening sequence o f o switch. restrikescan cause electrical closing(s ) oC the circ uit .Simple switching transients 171 include clrcuit closing transients,transients initi ate d by the clearing o f a short circuit, and transients pro-duced when the two c i rcu i ts on e i ther s ide or the swi tch being openedoscil la te at d i f feren t frequencies. In c i rcu i ts having induclance andcapacitance (a l l physical circuits have at least some, in the fo rm o fstray capacitance and inductance ) with l i tt le damping, these simpleswi tch ing transients are ~n her ent ly imi ted t o twice the peak amphtu deo f the steady-state sinusoidal voltage Wi th ou t a surge protect ivedevice, the current f lowing just before swi tch ing is available l o chargethe circu it capacitances at whatever voltage is require d l o store theinduc t i ve energy o f the cu r ren l by conver t ing i t into capacitive energy.

Several nlechdnisms are encountered In practical power circuits thatcan produce transient overvoltages Par in excess o f the theoretic altw i ce -no rma l l im i t r nen t~ oned bove. Tw o such mechan isms occur f re -quent ly - cur ren l chopping and restr ikes, the la t ter being especia l lytroublesome when capaci tor swi lch ing is involved

A simi lar scenar lo can un fo ld when an ungrounded power systemexperiences an arc lng ground faul t . The sw i tch ing act ion is then notthe resul t o f a del lherate part ing o f contacts but the in te rmi t ten t con-necl ion produced by the arc

These switching overvoltdges. high d5 they may be, are somewhatpredictable and can be estimated with reasonable accuracy from the cir-cu i t paramerers, once the mechanism lnvolved has been ident i f ied.There is sti l l some uncertainty as to where and when they occurbecause the worst offenders result f rom some abnormal behavior o f aci rcu i t e lement Lightning-induced transients dre much less predictablebecause there is a w d e range of coupl ing possib~l i t ies.

I n response lo these concerns. various committees and workinggroups have attem pted to describe ranges of transient occurrences orma xi mu m values occurr ing in power c i rcu i ts Three such auempts aredescribed in the nex t section Figures I . 2. and 3 show typical exam-ples of t ransients recorded In power systems

3. S ta n d a rd s o n T r a n s i e n t O v e r vo l ta g es i n P o w e r L i n e sSeveral standards or guides have been issued or proposed i n Europ e

and i n the U nit ed States, specifying a surge withstan d capabil i ty forspecific equipm ent or devices and specific conditions o f transients inpower or com munic at ion systems. Some of these speci f icat ionsrepresent early attempts to recognize and deal with the problem insplte o f insufficient data. As a growin g number o f organizat ionsaddress the p rob lem a nd as exchanges of in form atio n take place,impro veme nts are being made in the approach. Three o f these arebrlefly discussed here

0 5 10 20 30 4 0 752WICROSECONDS

Figu re 1. Lig ht n in g surge recorded on a120 V overhead d i s t r i bu t i on sys te p


3/14

Tab le 1I E C Re p o rt 6 6 4

0 5 10 20 30 40 I S 10 0MICROSECONDS

Figure 2. Sw itch ing surge recorded on a120 V res iden t ia l w i r ing sys tem

100050 0

VOL TS 050 0

1000

0 I 2 3 4MILLISECONDS

Figure 3. Capaci tor sw itching t ransie nt recorded on a 46 0 V system

3.1 The lE E E Surge Wi ths tand Capab i l i t y Tes t (SWC) 181One o f the earliest published documents w hich addressed new p rob-lems facing electronic equipment exposed to power system transientswas prepared by an IEEE commit tee deal ing with the exposure ofpower system relaying equipment to the harsh environ men t o f high-voltage substations.Because this useful document was released at a t ime when litt leother guidance was available, users attempted to apply the recommen-

dations of this documen t to situations where the environ men t o f ahigh-voltage substation did not exist. The revised version o f th is stan-dard, soon to be issued, recognizes the problem and attempts to bemore specif ic (and restrictive) i n its scope. Thus, an Impo rtant con-siderat ion i n the wr i t in g and publ ishing of documents deal ing wit htransients is a clear def in~ tio n of the scope and limitations of theapplication.3.2 T h e I E C 6 64 Re p o r t 191

The Insulat ion Coordinat ion Commit tee o f IEC included in i tsreport a table indicating the voltages that equipment must be capableof withstanding i n various system voltages and installatio n categories*(Table 1). The table spec~ fies its appl~ca bihty to a controlled voltagesituation, which implies that some surge-limiting device has been pro-vided - presumably a typical surge arrester wi th charac terist~csmatch-ing the system voltage i n each case. Th e waveshape specif ied fo rthese voltages is the 1.2/50 p s wave, a specif ication consistent with theinsulation withstand concerns of the group that prepared the document.N o source impedance is indicated, but fo ur "installation categories"are specified, each with decreasing voltage magnitude as the installationis fur ther removed from the outdoor environm ent. Thus, th is docu-men t prim arily addresses the concerns o f insula tion coo rdinat~ on; thespecification it imp lies for the environm ent is more the result o fefforts toward coordinating levels than efforts to describe the environ-men t and the occurrence o f transients. Th e latter approach has beenthat o f the IEE E Wo rking Gro up on Surge Voltages i n Low-VoltageA C Power Circuits, wh ich wi ll be reviewed i n detail.3. 3 The l E E E Gu ide on Surge Vo ltages(A N SV IE EE Std C62.41-1980) [ I013.3.1 Voltages and R ate of Occurrence. Data collected fro m anumber of sources Icd to plott ing a set of l ines representing a rate of

P R E F E R R E D S E R IE S OF V A L U ES O F I M P U L S EW I T H S T A N D V O L TA G E S F O R R A T E D V O L T A G E SB A S E D O N A C O N T R O L L E D VO L TA G E S I T U A T I O N

Der ived f rom Rated Wi ths tand Vo l tages i n

(V rms and dc)

occurrence as a functio n o f voltage for three types o f exposures (Fig-ure 4). These exposure levels are defined in general terms as follows:

Low Exposure - Systems in geographical areas known for lowl igh tn ing ac tiv it y , w i th l~ t t l e oad switching activity.Medium Exposure - Systems i n geographical areas kn ow n for hig hligh tnin g activity, w ith frequent and severe switching transients.High Exposure - Rare, but real, systems supplied by long overheadlines and subject to reflections at l ine ends, where the characteris-t ics of the ins tallation produce high sparkover levels o f theclearances.Th e two lower l ines o f Figure 4 have been drawn at the same slopesince the data base shows reasonable agreement among several sources

on that slope. A ll l ines may be truncated by sparkover of the clear-ances. at levels depend ing on the withsta nd voltage o f these clearances.T he high exposure l ine needs to be recognized, but it should not beindiscrim inately applied to all systems. Such application wo uld penalizethe vast majority of installations where the exposure is lower.Th e voltage and current ampl i tudes presented i n the Gu ide at temptto provide for the vast major i ty of l ightn ing str ikes but none should beconsid ered "worst case," as this concept cannot be det erm ine d rea listi-cally. I t is necessary to think I n terms of the statistical distr ib utio n ofstrikes, a nd to accept a reasonable upper li m it for mo st cases. Wh erethe consequences of a failure are not catastrophic but merely representan annoy ing economic loss. it is appropriate to m ake a tradeoff o f thecost of protect ion against the l ike l ih ood of a fa i lure caused by a h ighbut rare surge.

0 1 0 5 i 2 5 ( 0SURGE CREST L V

.In some locataons, s p a r k o v e r of clearances ma y lam01(h e overvol tages

Figure 4. Frequency of occurrence versus level fromAN SI / IEE E Std C62.41-1980


4/14

0pen.Clrcult Voltage,Current Defined by Teble IIIndoor Environment

Open Clrcult Vd trge Dlschrrga CurrentO u t d w r and Ne ar - a t do or E nvironment

Figure 5. Surges defined for power l ines by A NS W EE E Std C62.41-1980

3.3.2 Wav eshap e of the Surges. Many independent observations I l l -13) have established that the most frequent type of transient overvol-tage in ac power systems is a decaying oscillation, with frequenciesbetwee n 5 and 500 kHz. This finding is in contrast to earlier attem ptsto apply the unidirectional double-exponential voltage wave that is gen-erally described as 1.2/50. Indeed, the unidirectional voltage wave hasa long history of successful application in the field of dielectric with-stand tests and is representative of the surges propagating in powertransmission systems exposed to lightning. In order to combine themerits of both waveshape definitions and to specify them where theyare applicable, the Guide specifies an oscillatory waveshape insidebuildings, a unidirectional waveshape outside buildings, and both at theinterface (Figure 5 ) .The oscillatory waveshape simulates those transients affecting dev-ices that are sensitive to dvldt and to voltage reversals during conduc-tion [141, while the unidirectional voltage and current waveshapes,based on long-established ANSI standards for secondary valve arresters,simulate the transients where energy content is the significantparameter.3.3.3 Energy and Source Impedance.. Th e energy involved in theinteraction of a power system with a surge source and a surge prolctivedevice will divide between the source and the protective device inaccordance with the characteristics of the two impedances. In a gap-type protective device, the low impedance of the arc after sparkove rforces most of the energy to be dissipated elsewhere. e.g., in a resistoradded in series with the gap for limiting the power-follow current, oron the impedance of the circuit upstream of the protective device. Inan energy-absorber protective device, by its very nature, a substantialshare of the surge energy is dissipated in the suppressor, but its clamp-ing action does not involve the power-follow energy resulting from the--

ot e the difference belween ssr.?e Ct~e~Io~rcer the hne. also called characler~sltcimpedance. and the mipedanc~o rlre surge. Z , 41 F~gure8 d~h,rnl r anra "/R. L. Cor h e requenry correspondin g to rhe ssslye.

short-circuit action of a gap. It is therefore essential to the effectiveuse of surge protective devices that a realistic assumption be madeabout the source impedance of the surge whose effects are to beduplicated.Unfortunately, not enough data have been collected on what thisassumption should be for the source impedance of the transient. Stan-dards and recommendations, such as MIL STD-1399 or the IEC 664report. eith er ignore th e issue or indicate values applicable to limitedcases, such as the SWC test for high-voltage substation equipment.ANSIIIEEE Std C62.41-1980 attempts to relate impedance to categoriesof locations but unavoidably remains vague on their definitions(Table 2) .3.3.4 New Repo rts on Trans ient Occurrence. Th e importance ofdefining the environment has motivated a number of workers to installrecording devices on their power systems. and thus created a marketfor commercial transien t recorders in the last several years. Unfor-tunately. at least one such recorder contains a surge suppressor in irsown power supply (Figure 6 ) .

1 -3 kV open-circuit \,kV open-circuit. 'with an aly ze r plugged in

4 1 , l V - -d(a ) 200 A short-circuit current (b) 500 A short-circuit current

Figure 6. Effect of conne cting a "Di sturb anc e Analyzer" on a powersystem subjected to the A N W IE EE C62.41 ring waveTable 2Description of Surge Environment in ANSI/IEEE Std C62.41-1980

LocationCategoryA Long branchCircuits andoutletsB Major feeders.short branchcircuits, and

load center

Comparable to ImpulseIEC No 664 Medium ExposureCategory Waveform AmplitudeI1 0.5 us-100 kHz 6 kV200 A

1. 2 X 50 ps 6 kV111 8 X 20 ps 3 kA0.5 ps-100 kHz 6 kV500 A

Typeof Specimenor LoadcircuitHigh impedancetLow impedance?. 3High impedance*Low impedance?High impedance*Low impedance+.8

Energy (joules)Deposited in a Suppressor*with C lamping Voltage o f500v lO0OV

(1 20 V System) (240 V System)

*Othe r suppressors ha nn g different clamp ing voltages would receive different energy levels.*For high-impedance test specimens o r load circuits. the voltage &own represents the surge voltage. In making simulation tests-that value for b e open-cimuit voltaee of the tes t generator.?For low-impedance test suecimens or load circuits. the current shown represents th e discharge current of the su qe (not the short-circuit current of the Power system). In making simulation tests. use that cu rrent fo r the short-circuit c urrent of the test g enerator.t ~ h e aximum amplitude (200 or 500 A ) is specified, but the exact waveform will be influenc ed by the load characteristics.


5/14

C o n s e q u e n t l y . if t h e l i n e b e i n g m o n i l o r e d i s t h a t in w h i c h t h e 4 . S u r g e P r o p a g a t i o na n a l y z e r h a s i t s p o w e r c o r d p l u gg e d . t h e t r a n s ie n t s t h a t o c c u r r e d o nt h e l i n e b e f o r e t h e a n a l y z e r w a s t n s ta l l e d d i sa p p ea r . T h i s i r o n y m a k e s 4 .1 T h e L i m i t a t i o n s o f A r b i t r a r y D i v i s i o n i n t o C a t e g or ie smy f a c e ti o u s r e m a r k o r 1 9 69 . ". . . [h e b e s t s u r g e s u p p r e ss o r i s a s u r g e T h e s t a n d a r ds c i t e d i n t h e p r e c e d i n g p a r a g ra p h s d e s c r ~ h e s u rg e sn ~ o n i t o r . " (121 t h e t r u t h . T h e r e f o r e . s o m e o f t h e r e ce n l l y p u b l i s h e d w h i c h m a y b e e x p e c t e d a t s p ec if ic p o i n t s o f a w i r i n g s ys te m ; t h e i m p l i -s u r v e y s o n t h e o c c u r r en c e o f s u rg e s m a y s h o w a l o w e r n u m b e r o f c a t i o n i s t h a t t h e s u r g e s w i l l p r oc e e d d o w n st re a m , a t t h e same a r n p l l -s u r g e s t h m i t c t u i ~ l l y c cu r . a n d t h er e fo r e d a t a s h o u l d b e e x a m i n e d w i t h t ud e , u n t i l s o m e i n t e r fa c e s o m e h o w p r o d u c e s a d e c re a s e i n a m p l i t u d ec a u t i o n u n t i l t h e a m b i g u i t y c a u s e d b y t h i s c o r d -c o n n e c te d s u p p r e ss o r i s d o w n t o t h e n e x t l o w e r l e v e l o f v o l ta g e o r lo t h e n e x t l o w e r l e v e l o fr e s o l v e d . c u r r e n t s p e c if i e d b y t h e n e x t d o w n s t r e a m c a t e g o r y ( F i g u r e 7 ) .

1. T H E ANSIIIEE,!? S T D C62.41-1980 C O N C E P T O F L O C A T I O N C A T E G O R IE S I N U N P R O T E C T E D C IR C U IT SLOCAT ION CATEGORY LOCA TION CATEGORY LOCA TION CATEGORYC B A

- 0 ~ t s t d e n d S e rw ce E n tr an c e" :Major Feeders and Short Branch Circuits" : "Out lets and Long Branch Circuits".V O L T A G E S 6 kv I mpu lse or Ring 6 kV RmgC U R R E N T S

2 , T Y P I C A L E X A M P L E S O F N D U S T RI A L O R R E S I D E NT I A L C I R C U I T SA l t s m a t s ~Undergroundcable serv ice

IOverhead Iservice

N o t e 1

Data l i n eS e e N o t e 3

C a b l e i n p u tS e e N o t e 3

IPI . Surge arrester ID lndustrml drwe systemLE GE N D : P3 Surge p r o tec t o r(second ary rat lngl M Owen motor ICS lndust ral cont rol system * P4 Surge protectorM Wall.hour meter FA Flxed applmnce WRZ W all receptacle wth PC Personal compu terP2 Surge ar rester WRI . Wail receplacle w~ tho ul t t enua t~on a t l e n t u a t m p r ow d e d HF Consumer elect ronc sfsecondarv ra t l n ~ l SE Serv~ce nt rance (may take m a w farms TV Transformer isolated

MB Mam br e ike rAW. Arc welding supplyP Transient p~oleclorCA Cord connec ted applianceCOMP Computer w lth buf lered nputLC Llne power cond!t?oner

3. T H E I E C R E P O R T 664-1980 C O N C E P T O F C O N T R O LL E D V O L T A G E Suncont r o l led

VOLTAGESlli: f4\25 kV

F i g u r e 7.S i m i l a r i t i e s a n d d if fe re nc e s b e t w e e n t h e l o c a t i o n c a te g o r i e s co n c e p t o f ANSI/IEEE S t d C 6 2 .4 1 -1 9 8 0

and t h e i n s t a l l a t i p n c a t e g o r i e s c o nc e p t o f IEC R e p o r t 6 6 4 -1 9 8 0, a p p l i e d t o a t y p i c a l e x a m p l e

INSTALLATION CATEGORYIVsupply level. Overhead linesand -ble systems including dlstr lbu-lion bu s and 11s associated overcurrentequipment.

. ,-INSTALLATION CATEGORY 1 11 (2.5)i 5 kV

111 I INSTALLATION CATEGORY 1 (1SP8 kVF w d ~ n st a ll a tm no l low +nginstal la l lon category IV. I1Appliances portable equipment etc INSTALLATION CATEGORYI f 0 l IW . n ~ s la l la l~o n a t egor y I l l I SPec...." N,.. ,.,..men1 following lns lal latlon category I .te lecomm unmt lon , etect ronlc, etc


6/14

T a b l e 3( N o t e s f o r F i g u r e 7)

I I THE ANSl i lEEE Sld C62.4 1-I980 CONCEPT OF LOCATION CATEGORIES IN UNPROTECTED CIRCUITSM T h e vo l tage leve ls shown i n the three IoCatiOn CategOrleS represenl h~ g h mpedance c i rcu l t cond~ t lons lgh l load~ng nd no

surge protective de v~ ce s I . P2. P3 or P4 The 10 kV vo l tage of Calegory C IS reduced lo a maxlmum of 6 kV In both Ca lego r~e sB a n d A by the t lke ly sparkover o f c learances, should a 10 kV surge Impinge on the servtce ent rance

(b)T he curren l leve ls shown In the three locat lon categ or~ es n a descendmg sla t rcase l rom C to A represe nl tow w npedance ct rcu l l cond thon for surges, su ch as the lnSlal la tlOn of one or more surge pro le cl lve devlces P I P2 P3 or P4 Another lowmpe dan ce c ondl l lon 1s the case of equtpment sparkover (installed equipment In an actual system or EUT durmg a test)

(c) I f mul l ip le surge protect ive devices are lns la l led on the system. the curren l waveform Imposed on the downstream pr otec t~vedevice is tn f luenced by the c lampl ing character is t ics c f the upst ream device

1 2 . TYPICAL EXAMPLES AND THE IEC REPORT 664 CONCEPT NOTES:(1) The Contro l led Vol tage St tua l lon of IEC Report 664 requires Ihe Presence of mlerfaces lhese can be surge protecl lve dev lces

such as P I . P2 P3 or P4 or the existence of wel l -def ined Impedance networks such as Z a n d C shown In the c l rcu i l d~a g ra mups t ream o f WR2

I Surge arresters or protectors P I . P2. P3 and P4 may be any pr0 leCl1ve de v~ ce u l lab le lor Ihe surge current leve ls expected atthat pomt o f the system P I and P2 are shown connecle d Ime-to ground P3 and P4 may be connected l tne to neulra l or be a

, combmallon of Ime-to-neutral wtth addlhonal neutral-to-groundSurge a r reste r P2 may a lso be connec led on the load s lde o f t ne ma ln c t rc u~ l reake r ME I n lha l case MB wou ld then be con1 s ~ d e r e do be m In ntallatton Category IV(2)Vol tage leve ls to t low~ngh e d e s ~ g n a l ~ o nl Installallon Category IIV Ill 11 or I ) are shown m Parentheses lor a system wlth 300 vphase-to-ground voltage. and next lor 150 V Phase - lo -g round vo l tage The vo ltages shown a re ~ m ~ l t e ds I /5 0 m p u l s e sExample IV (6 )4 means 6 kV 1 2/50 p s lor a 240 V System 4 kV 1 2/50 p s tor a 120 V system See IEC Report 664-1 980 [g] lorthe complete (ab le o f leve ls correspondmg to system vo l tages f rom 50 lo 1000 V

( 3 ) T h ls d ~ a g r a mmakes no a l lowance fo r t he po ss ~b ~ l i l yf surges associated wl lh g round po ten l la l d~ f fe ren ces hat may occur, fo rinstance w ~ t h sensor connect ton to the ICS contro l system a cable TV connect ton to the line- so la led TV set etc . or the f lowof ground curre nt In the lmpedance of the groundmg conductors

(4)Ttansient pro lector P in the hne feedtng the welder A W la lypocal example of transient generator ml ern al to the system )Intended to protect the system lrom the welder rather than to prolect m e welder lrom Ih e system

(51 P o w e r l l n e c o n d ~ l ~ o n e rC whi le pertormlng the major task o f C ond~l io nmg he power supply to the computer may perform atunc t~on tml lar to that o f the protector Pa t the welder In blockm g conducte d mlerterence l ro m the load loward the system

(6)Many appl tances or electronic devtces might be equtpped wl lh In terna l surge orotecl lve devices and therefore be su l lab le lormsla l fa t ton In o ther cate gor ~eshan I I

Th is s~a i rcase ep resen la t ion i s use fu l l o s imp l i f y t he rea l wo r ld in toa manageab le set o f assumpt ions . bu t i t i s a simpl i f ica t ion that canmask ihe rea l i t y . Surges wi l l p ropagate in the system start ing a t thepo in t o f ent ry : vo l tage su rges w i l l be a t tenua ted to t he ex ten t t ha t t hese ries impedance be tween the po in t o f in te res t and the sou rce ( Z , .F i g u r e 8 ) and the shunt impedance ( Z : ) f o rn i a vo l t age d iv ide r . I ft he se r ies impedance i s low and the shun t impedance i s h igh ( l i gh lload ing o f t he sys tem) . t he vo l t age d iv ide r does no t p roduce h igha t tenua t ion o f t he vo l t age su rges. Converse ly, current surges, if t h e ya re the resuh o f a cu r ren t sou rce such as a l i gh tn ing s t r i ke . w i l l p ro -duce h igh vo l tages un less a low- impedance d ive r t ing pa th i s o f f e red lot h e f l o w o f c u rr e n t. I f the current surge in a sys tem is t he resu l t o f ac o m b i n e d c u r r e n t s o u r c e a n d m u l t i p a t h t o g r o u n d ( F i g u r e 9 ) , t he re i st h e n a d i v i s i o n o f t h e c u r r e n t a m o n g t h e p a l h s t h at i s g o ve r n e d b y t h einve rse ra t io o f t he impedances . I f a user has con t ro l ove r on ly oneo f t he pa ths , h e can decrease the amp l i t ude o f t he cu r ren t su rge in h i spa th on ly by fo rc ing a g rea te r sha re o f t he to ta l cu r ren t t o f lowthrough the other paths: thus, surge blocking i s l i ke ly t o be an exe rc isei n p as si ng t h e p r o b l e m f r o m o n e p o i n t o f I h e s y s te m t o a n ot h er . T h eso lu t ion l i es i n su rge drversion, o f fer ing to t he su rge a pa th whe re thecu r ren t f l ow can occu r ha rm less ly .

Vol tage~ 2 0 seen byU s e r

*F i g u r e 8. Vo l tage d iv ide r ef fect

4.2 T h e L i m i t a t i o n s o f T r a n s m i s s io n L i n e A n a l y s isIn qual i ta t ive d iscussions of surge propagat ion, the c lassica l behav ior

o f a t ransm iss ion l i ne i s o f t en ca l led upon to p rov id e exp lana t ions o fthe s i t ua t ion . In pa r ti cu la r , t he theo re t i ca l re f lec tions occu r r in g a t t hee n d o f a l i n e a r e c it e d: d o u b l i n g t h e i m p u l s e if t he l i ne i s open-ended,r e t u r n i n g a n in v e r s e i m p u l s e if t h e l i n e is sho rted. ~ o w e v e r T e s ed iscuss ions some t imes lose s igh t o f t he fac t t ha t t he concep t . cpn ,.k


7/14

Lightning Strokeor Lightning.Induced Current

User Has No ControlOver This Path

User'sY =I1I1. 11 Equipment 1% r Ground

User's World

orV = IIZ1 - 2Z' ~f2' is SignificantFigure 9. Mult ipsth current division

applied only i f the line length is sufficient to contain all of the surgefront. I f the surge has a rise time longer than the propagation timealong the line, the point is moot and, by the time the surge reachesits peak, the voltage at the receiving end of the line does not differfrom the voltage applied at the sending end. Figure 10 illustrates thissituation showing the voltage at the sending end (SD) and the voltageat the receiving end (RC) of a conduit-enclosed three-wire line(B: black or phase conductor, W: white or neutral conductor, G: greenor grounding conductor). There is a minor difference during the rise,where the ini tial front is doubled at the receiving end, but the crestsare the same. Thus, for many installations in buildings, the linelengths are short compared to the length required to contain, say, aI s ront traveling at a typical speed of 200 m l ~ s .A more detailed report describing some of the aspects of the propa-gation of surges and their implications is given i n Reference IS.4.3 Common Mode or Normal Mode?

Another aspect of the propagation o f surges concerns the dichotomynormal mode/common mode. I n other words, the issue is whethersignificant surges occur line-to-line (black-to-white, or phase-to-neutral), the situation described by "normal mode." or whether theyoccur between any - or both - of the lines and ground (black-to-green, white-to-green, or [black-and-whitel-to-green), the situationdescribed by "common mode."

The answer to that dichotomy is that, i n most cases, both modesmust be considered, since one often converts into the other, dependingupon the coupling, the wir ing practices, and the attempts made atsuppressing the mode perceived as the greatest threat. Here again, thepervasive and pervert reality is too often that a solution aimed atsuppressing one effect only displaces the problem. Examples 1 and 2,i n the Appendix, show how controversies on the method of suppress-ing a surge assumed to occur between two wires can lead to partial ormisleading solutions when the realities of the situation are notrecognized.

5. Fundamental Protection TechniquesThe protection o f a power conversion system or an electronic black

box against the threats of the surge environment can be accomplishedin different ways. There is no single truth or magic cure ensuringimmuni ty and success, but, rather, there are a number of validapproaches that can be combined as necessary to achieve the goal.The competent protection engineer can contribute his knowledge andperception to the choice of approaches against a threat which is impre-cise and unpredictable, keeping in mind the balance between thetechnical goal of maximum protection and the economic goal of realis-tic protection at an acceptable cost. However, just as in the case ofaccident insurance, the cost of the premium appears high before theaccident, not after.A discussion o f fundamental protection techniques that is limited i nspace'and scope has the risk o f becoming an inventory of a bag oftricks: yet. there are a few fundamental principles and fundamentaltechniques that can be useful in obtaining transient immunity, espe-cially at the design stages of a computer system or circuit. Al l toooften. the need for protection becomes apparent at a late stage, when

Figure 10. Voltages at sending and receiving ends of a line

it is much more difficu lt to apply the fundamental techniques whichare most effective and economical when implemented at the outset.5.1 Basic Techniques

Protection techniques can be classified into several categoriesaccording to the purpose and the system level at which the engineer isworking. For the system as a whole, protection is primarily a preven-tive effort. One must consider the physical exposure to transients -in particular, the indirect effects of lightning resulting from buildingdesign, location, physical spread, and coupling to other disturbancesources - as well as such inherent susceptibility characteristics as fre-quency response and nominal voltage. A system depending on low-voltage signals, high-impedance circuits, and instal led over a wide areawould present much more serious problems than the same systemconfined to a single building or a single cubicle

For the system components or electronic black boxes. the environ-ment is often beyond the control of the designer or user, and protec-tion becomes a curative effort - learning to live and survive in anenvironment which is imposed. Quite often this effort is motivated byfield failures. and retrofi t is needed. The techniques involved heretend to be the application o f protective devices to circuits or a searchfor inherent immunity rather than the elimination o f surges at theirorigin.

Another distinction can be made in classiiying protective tech-niques. Whi le surges are unavoidable, one can attempt to block them.divert them, or strive to withstand them; the latter, however, is gen-erally difficult to achieve alone.5.2 Shielding, Bonding, and Grounding .

Shielding, bonding, and grounding are three interrelated methodsfor protecting a circui t from external transients. Shielding consists ofenclosing the circuit wiring in a conductive enclosure, which theoreti-cally cancels out any electromagnetic field inside the enclosure: actu-ally. it is more an attenuation than a cancellation. Bonding is thepractice of providing low-impedance connections between adjacentmetal parts, such as the panels of a shield. cabinets in an electronicrack, or rebars in a concrete structure. Grounding is the practice ofproviding a low impedance to earth. through various methods of driv-ing conductors into the soil. Each of these techniques has its limi ta-tions, and each can sometimes be overemphasized. Many texts andpapers discuss the subject at length, so that a detailed discussion is notnecessary here. with one exception, which follows.5.3 Ground One End or Both Ends of the Shield?

Shielding conductors by wrapping them in a grounded sheath orshielding an electronic circuit by enclosing it i n a grounded conductivebox is a defensive measure that occurs very naturally to the systemdesigner or the laboratory experimenter anticipating a hostile elec-tromagnetic environment . Difficulties arise. however, when [he con-cept of "grounded" is examined in detail. Difficu lties also arise whenthe goals of shielding for noise immunity conflict with the goals ofshielding for lightning surge immunity.

A shield can be the size of a matchbox or an airplane fuselage: i tcan cover a few inches of wire. or miles of buried or overhead cables.Grounding these diverse shields is not an easy thing to do because theimpedance to earth of the grounding connection must be acknowl-edged. The situation is made even more controversial because of t h rconflict between the often-proclaimed design rule "ground cable shieldsat one end only" - a rule justified by noise immunity performance, inparticular common mode noise reduction - and the harsh reality ofcurrent flow and Ohm's law when lightning strikes.

23


8/14

The difficulty may be caused by a perceptlon on the part of thenoi\e prevention designers that the shield serves as an electrostaticshield in which longitudinal currents associated with common modenoire coupling should not flow. This concept is exemplified in the ter-minology of shielded cable users. when they describe the shield con-struction of some cable design as having a foil plus drairr wire. as ifthere were elecrrosrarrc charges that needed to be removed (drairred.Indeed, electrostatic charges can be drained by connecting only oneend of the shield. Furthermore, if the two ends of the shield of acable spanning some distance are connected to the local ground at eachend, there is a definite possibility that some power frequency currentmay flow in the shield. For low-level signals. this current would pro-duce nolse ( hu m) in the signals. For that reason. many systemdesigners will insist on the one-end-only grounding rule, and they arecorrect from that point of view. Sometimes the shield is used as areturn path for the c~rcuir , n which case shield currents can cause vol-tage drops added to the signal. But the fact is that, when surgecurrents flow near the circuits, they will unavoidably inject magneticflux variations into the circuits, hence mduced voltages. Worse yet, inthe case of a lightning stroke injecting current into the earth in thearea spanned by the one-end-only grounded shield, the potential ofone end cf the cable defined as "ground" is not the same as the"ground" at the other end of the cable. Very high voltages can bedeveloped (Figure 11) between the floating end of the shield and thelocal ground. No practical insulation can withstand these levels, andbreakdown yill occur, allowing surge currents to flow in spite of thedesigners' intent to prevent them; the path of these currents will bedetermined by the components most likely to fail when the voltagerises - the low-level logic circuits, of course. In contrast, by deli-berately allowing part of these surge currents to flow in the shields,-~one obtains a cancellation of the voltages that otherwise would beinduced in the circuits, and the currents will follow a welldefined paththat can be designed to provide harmless effects.This conflict is actually very simple to resolve if recognized in time:~rovidean outer shield, grounded at both ends (and at any possibleintermediate points), mside this shield the electronic designer is thenfree to enforce his single-point grounding rules. The only drawback tothis approach is the hardware cost of double shields. In many installa-tions. however. there is a metallic conduit through which the cablesare pulled; with simple but close attention to maintaining the con-tinu~tyof this conduit path, through all the joints and junction boxes,a very effective outer shield is obtained at negligible additional cost.In the case of underground conduit runs, the most frequent practice isto use plastic conduit, which unfortunately breaks the continuity. Sys-tem designers would be well advised to require metal conduits wherethe circuits are sensi tive or, at a minimum, to pull a shielded cable inrhe plasuc conduit where the shield is used to maintain continuitybetween the above-ground metal conduits. That additional cost, then,is the insurance premium, which is well worth accepting. Example 3in the Appendix shows the consequence of the mtsapplied "groundone end only" rule

Figure 11. Voltage developed by a lightning strokeat the ungrounded end of a shield

6. Transient SuppressorsVarious devices have been developed for protecting electrical andelectronic equipment against transients. They are often called "tran-sient suppressors" although, for accuracy, they should be called "tran-

sient limiters," "clamps," or "diverters" because they cannot reallysuppress transients; rather, they limit transients to acceptable levels ormake them harmless by diverting them to ground.

There are two categories of transient suppressors: lhose that blocktransients, preventing their propagation toward sensitive circuits, andthose that divert transients. limiting residual voltages. Since many ofthe transients originate from a current source, the blocking of a tran-sient may not always be possible: thus. diverting the transient is morelikely to find general application. A combination of diverting andblocking can be a very effective approach. This approach generallytakes the form of a multistage circuit, where a first device diverts thetransient toward ground, a second device - an impedance or resis-tance - offers a restricted path to the transient propagation but anacceptable path to the signal or power. and a third device clamps theresidual transient (Figure 12). Thus, we are primarily interested in thediverting devices. These diverting devices can be of two kinds:voltage-clamping devices or short-circuiting devices (crowbar). Bothinvolve some nonlinearity, either frequency nonlinearity (as in filters)or, more usually, voltage nonlinearity. This voltage nonlinearity is theresult of two different mechanisms - a continuous change in the dev-ice conductivity as the current increases, or an abrupt switching as thevoltage increases.Because the technical and trade literature contains many articles onthese devices, a discussion of the details will be limited. Some com-parisons will be made, however, to point out the significant differencesin performance.

RESTRICTDIVERT CLAMP

PROTECTEOCIRCUITuigure 12 . Three-step approach for surge suppression

6.1 Crowbar DevicesThe principle of crowbar devices is quite simple: upon occurrenceof an overvoltage, the device changes from a high-impedance state to alow-impedance state, offering a low-impedance path to divert the surgeto ground. This switching can be inherent to the device, as in thecase of spark gaps involving the breakdown of a gas. Some applica-tions have also been made of triggered devices, such as triggeredvacuum gaps in high-voltage technology or thyristors in low-voltage cir-

cuits when a control circuit senses the rising voltage and turns on thepower-rated device to divert the surge.The major advantage of the crowbar device is that its lowimpedance allows the flow of substantial surge currents without thedevelopmenc of high energy within the device itselr; the energy has tobe spent elsewhere in the circuit. This "reflection" of the impingingsurge can also be a disadvantage in some circuits when the transiectdisturbance associated with the gap firing is being considered. Wherethere is no problem of power-follow (discussed below), sueh as insome communication circuits, the spark gap has the advantage of verysimple construction with potentially low cost.Th e crowbar device, however, has three major limitations. The firstlimitation concerns the volt-time sensitivity of the breakdown process.As the voltage increases across a spark gap, a significant conduction ofcurrent - and hence the voltage limitation of a surge - cannot takeplace until the transition occurs to the arc mode of conduction, byavalanche breakdown of the gas between the electrodes. The load islefl unprotected during the initial rise because of this delay time (typi-cally in microseconds). Considerable variation exists in the sparkovervoltage achieved in successive operations, since the process is statisticalin nature. In addition, this sparkover voltage can be substantiallyhigher after a long period of rest than after successive discharges.From t he physical nature of the process, it is difficult to produce con-sistent sparkover voltage for low voltage ratings.The second limitation is associated with the sharpness of the spark-over, which produces fast current rises in the circuits and. thus, objec-tionable noise. A classic example is found in oscillograms recordingthe sparkover of a gap where the trace exhibits an anomaly before thesparkover (Figure 13). This anomaly is due to the delay introduced inthe oscilloscope circuits to provide an advanced trigger of -weep.What the trace shows is the event delayed by a few nanoseconds, sothat in real time, the gap sparkover occurs and noise enters the oscillo-scope by stray coupling while the electron beam is still writing the pre-sparkover rise.

24


9/14

F igure 13. Anomaly i n recording showing gap inter ferenceA th ird l imitat ion occurs when a power current f rom the steady-state voltage source follows the surge discharge (follow-current, orpower-follow). I n ac circuits, this power-follow current may or maydot be cleared at a natural current zero. In dc circuits, clearing iseven mor e uncertain. Add ittonal means, therefore, m ust be provided toopen the power circuit i f the crowbar device is not designed to provideself-clearing action within specif ied l imits of surge energy, system volt-age, and power-follgw curre nt. This combina tion o f a gap with a non-

linear varistor that l imits the power-follow current has been very suc-cessful in the uti l i ty industry as a surge arrester or surge diverter.6 .2 Voltage-Clamping Devices

Voltage-clamping devices will exhibit a variable impedance, depend-ing on the current f lowing through the device or the voltage across itsterminals. These components show a nonlinear characteristic - thatis, Ohm's law can be applied. b ut the equa tion has a variable R.Impedance variation is monotonic and does not contain discontinuit ies.in contrast t o the crowbar device, which shows a turn -on action. Asfar as volt-am per e characteristics are concerned, these com pone nts aretime-dependent to a certain degree. However, unl ike the sparkover ofa gap or the tr iggering o f a thyristor, t ime delay is no t involve d.When a voltage-clamping device is installed, the circuit remainsessentially unaffected by the device before and after the transient forany steady-state voltage below clam ping level. Increased curren t drawnthro ugh the device as the surge voltage attempts to rise results i nvoltage-clamping action. Nonlin ear impedance is the result if thiscurrent rise is greater than the voltage increase. The increased voltagedrop (IR) i n the source impedance due to higher current results in theapparent clamping of the voltage. It shou ld be emphasized that thedevice depends on the source impedance to produce clamping. A volt-age divider action is at work where the ratio of the divider as not con-stant, but changing. I f the source impedance is very low, the ratio islow, and eventually the suppressor could n ot work at all with a zerosource impedance (Figure 14). I n contrast, a crowbar-type deviceeffectively short circuits the transient to ground; once established, how-ever, this short circuit wil l contin ue un til the current (the surgecurrent as well as any power-follow current supplied by the power sys-tem) is brought to a low level .Th e prin ciple o f voltage clam ping can be achieved wit h any deviceexh ibit i ng this nonlinear impedance. Tw o categories of devices, havin gthe same effect but operating on very different physical processes, havefound acceptance In the indus try: polycrystalline varistors and single-junction avalanche diodes. Anoth er technology, selenium rectif iers. hasbeen pract ical ly e l iminated f rom the f ie ld because of the improvedo f m odern varistors.

F igure 14 . Voltage divid er effect of shunt-connected suppressor

6.3 Ava lanche DiodesAvalanche diodes, or Zener diodes, were init ially applied as voltageclamps, a natural outg rowth o f their application as voltage regulators.Imp rov ed construction, specif ically aimed at surge absorption, has made

these diodes very effective suppressors. Large diameter junction s andlow thermal impedance connections are used to deal with the inherentproblem o f dissipating the heat deposited by the surge in the sma llvolume of a very thin single-layer junction.The advantage of th e avalanche diode, generally a P -N sil icon junc-tion, is the possibil ity of achieving low clamping voltage and a nearlyflat volt-ampere characteristic over its useful power range. Therefo re,these diodes are widely used in low -voltage electronic circuits for theprotect ion of 5 o r 15 V logic circuits, for instance. For higher volt-ages, the heat generation pr oble m associated wi th single junctions canbe overcome by stacking a number of lower voltage junctions.6.4 Varistors

The term varislor is derived fr om its funct ion as a variable resistor.I t is also called a voltage-dependent resisior, bu t that descrip tion tends toimp ly that the voltage is the independent parameter i n surge protec-tion . Tw o very different devices have been successfully developed asvaristors: sil icon carbide disks have been used for years in the surgearrester industry, and more recently, metal oxide varistor technologyhas come of age.Metal oxide varistors depend on the conduction process occurring atthe bound aries between the large grains o f oxide (typically zinc oxide)g rown in a carefully controlled sintering process. The physics of thenonlinear cond uction m echanism have been described i n the literature(16-201.Because the prime function of a varistor is to provide the nonlineareffect, o the r parameters are generally the resu lt of tradeoffs i n designand inherent characteristics. The electr ical beh av~ orof a varistor canbe understood by exam ination of the equivalent circuit o f Figure 15.The major element is the varistor proper, R,,, hose V - I characteris-tic is assumed to be the perfect power law I kV". In para lle l w ~ t hthis varistor, there is a capacitor, C, and a leakage resistance, R I nseries with this three-componen t group, there is the bulk resistance o fthe zinc oxide grains, R, , and the inductance of the leads, L.

Figure 15. Equivalent c i rcui t of a var istorUnder dc conditions, at low-current densit ies because obviously novaristor could stand the high energy deposited by dc currents of high

density, only the varistor element and the parallel leakage resistanceare signif icant. Und er pulse conditions at high-current densit ies, allbut the leakage resistance are significant: the varistor provides lowimpedance to the flow of current, but eventually the series reststancewi l l p roduce an up tu rn in the V - I characteristic; the lead inductancecan give rise to spurious overshoot problems i f i t is not deal t wi thproperly; the capacitance can offer either a welcome additional pathwith fast transients or an objectionable loading at high frequency,depending on the application.6.4.1 V - I Charac te r is ti c . When the Y - characteristic is plotted on alog-log graph, the curve of Figure 16 IS obtained. Three regions resultf rom the dominance of R,, R, . , an d R , as the current i n the devicegoes from nanoamperes to kiloamperes.T h e V - l characteristic is then the basic application design tool forselecting a device in order to perform a protective function . For a---successful application, however, other tactora, discussed in detail in th einfo rma tion available from manufacturers, must also be taken into con-sideration. Som e o f these factors are:


10/14

Figure 16. V -I chara cterist ic of a varistorSelection of the appropriate nominal voltage for the line volt-age of the applicationSelection of the energy-handling capability (including con-sideration of the source impedance of the transient, the wave-shape, and the number of occurrences 1211)Heat dissipationProper installation in the circuit (lead length) [221

6,s Packaged SuppressorsThe need for protection and the opportunity to provide packagedprotective devices to concerned computer users has prompted themarketing of many packaged suppressors, ranging from the very simpleand inexpen sive to the complicated (not necessari ly m uch better) andexpen sive. Th e field has also seen a numb er of devices claimingenergy savings in conjunction with transient suppression; there is nofoundation for such a claim, and the issue hopefully has been settledin a study published by EPRl 1231.Line conditioners using a ferro-resonant transformer are offered prl-marily as line-voltage regulators, but also provide high attenuation oftransient overvoltages, a performance that isolating transformers do notprovide. Example No. 2, in the Appendix, gives examples of the poorperformance of isolat ing transformers and of the effective performanceof line conditioners in dealing with normal mode transients

6.6 Failur e ModesFailure of an electrical component can occur because its capabilitywas exceeded by the applied stress or because some latent defect inthe comp onent went unnoticed in the quality control processes. Whilethis situation 1s well recognized for ordinary components, a surge pro-tective device, which is no exception to these limitations, tends to beexpected to perform miracles, or at least to fail graciously in a fail-safemode. Th e term "fail-safe," however, may mean different failuremodes to different users and, therefore, should not be used. To someusers. fail-safe means that the protected hardware must never beexposed to an overvoltage, so that failure of the protective devicemust be in the fail-short mode, even if it puts the system out ofoperatio n. To others, fail-safe mean s that the funcrion must be main-tained, even if the hardware is left temporarily unprotected, so thatfailure of the protective device must be in the open- circuit mode . Itis more accurate and less misleading to describe failure modes as fail-short or fail-open, as the case may be.When the diverting path is a crowbar-type device. little energy isdissipated in the crowbar. as noted earlier. In a voltage-clamping dev-ice. because m ore energy is dep os ~t ed in the device. the e nergy-handling capability of a candidate protective device is an importantparam eter to consider in the design of a protection scheme. Withnonlinear devices. an error made in the assumed value of the currentsurge produces little error on the voltage developed across the protec-t ive device and thus applied to the protected c~rcuit ,but the error isdirectly reflected in the amount of energy which the protective devicehas to absorb. At worst, when surge curre nts in excess of theprotective device capability are imposed by the env iron me nt, e.g ., anerror made in the assumption, a human error in the use of the device,or because nature tends to support Murphy's law, the circuit In needof protection can generally be protected at the price of failure in theshort-circuit mode of the protective device. However, if substantialpwe r-frequ ency curren ts can be supplied by the power system, the

fail-short protective device generally terminates as fail-open when them w er system f ault in the failed device is not quickly cleared by aseries overcurrent protective device (fuse or breaker)

With the failure mode of a suppressor being of the fail-short type.the system protection with fuses can take two forms (Figure 17) . Fo rthe user conce rned with maintaining the protection of expen sive equip -ment, even if failure of the protector means the loss of the function,Al tern ativ eA must be selected. Conversely. I T the function isparamount, Alternative B must be selected

a ) P r o t e c t ~ o n a in ta in edf u n c t io n i n t e r r u p t e dmb) F u n c l ~ o n a i n ta ~ n e dp r o t e c h o n l o st

Figure 17. Fus ing a l te rna t ives for suppress ors7. Conclusions

Power system disturbances can inject damaging overvoltages inpower lines as well as data lines. Lig htn ~n g surges can be equallydamaging. by direct termination of a stroke, or by induction or evendifferences in groun d potential caused by the flow of the curre nt.Beware of differential ground potential rise!

Fund amen tal prec autions , best applied in the design an d constru c-tion stages. can provide effective protection a1 a small cost comparedto the alternativ e of failures and later relrofils. T h e cost of in su ra nc epremiums always seems high before the accident.Shielding. bonding, and g rou nd~ ng are the classical preventivemethods at the system and component level . Conflicts between tradi-tional grounding practices for noise reduction can be reconciled w ~ t hthe requirements of surge protection. Grou nding the shields at oneend only invites trouble.A combined approach of fundamental precautions and protectivedevices can provide effective protection over the range of natural andman-made d isturbances. However, these devices must be applied aspart of a concerted effort. Coo rdination of protective devices is th ekey to functional and cost-effective protection.

AcknowledgmentsMotivation for presenting this paper was provided by the reported casehistor~ es and the penetrat ing questions raised by students at theUniversity of Wisconsin annua l confere nces on surge protection, aswell as by d~scu ssions with members of the IEEE Su rge ProtectiveD e v ~ c e sCom mittee . Catharine Fisher and Elizabeth Zivanov of C R Dcontributed valuable revhews and comments toward development of thefinal text.

References1. J .H. Hagenguth and J . G . A nd er so n, "L ig htn in g t o k E m p i r eSta te Building," Part 111, A I E E Transactions 71 PAS, August 1952.2 . N . Cianos and E.T. Pierce, A Ground Llghlning Envrronment forEngineerrng Usage, Stanford Research Institute Technical Report,

August 1972.26


11/14


12/14

To illustrate the preceding discussion, some practical examples aregiven in this appendix as the basis for sound design approaches.Example No. 1 - Does a n i so la ti ng t r ~ s f o r m e r e lp ? T h e au th o rhas witnessed and engaged In many discussions on the merits of isolat-ing transformers, sparked by the misconception that "spikes are attenu-ated by transformers" or "spikes do not pass through transformers."Figures 18 and 19 are offered to support th e position that these quota-tion s are misconceptions. Whe n properly applied, isolating transform -ers are useful to break ground loops and reduce common mode effect.but they do not by themselves attenuate spikes that occur line-to-linein the normal mode.Figure 18 shows the propagation - or worse, the enhancement -of a voltage impulse in a 1.1 isolating transformer. Th e 6 kV, 0.5-100 kHz im pinging wave of ANSIIIEEE Std C62.41-1980 is applied tothe primary of the transformer, H1 H3 to H 2H4 . The output voltage.measu red at XIX, to X2X4, appears as a 7 kV crest on the secondaryside of this "isolating*' transformer.

0.5-lrn

BWG

1WA GENERALPURPOSE120112~-TRANSFORMER I I 1 J I I I I -Figure 18. Surge propagation through isolating transforme r

Figure 18 was recorded with no load on the transformer secondary,which represents the extreme case of the low-power electronic controlin the standby mode. Figure 19 shows the primary an d secondaryvoltages of the transform er with a 1 0 W (1500 fl) and a 100 W(150 R ) load on the secondary side, at the same generator setting asFigure 18. With the 10 W load that might be typical of an electroniccontrol in standby mode, the combined series reactance of thetransformer and shunt resistance of the load produce the output shownin Figure 19A, still slightly higher than the input.With the 100 W load shown in Figure 19B, the attenuation is nowapparent, but is only 2'1. Capacitive loads would, of course, produce agreater attenuation than resistive loads for the inductive seriesimpedance of the transformer, a1 the frequency spectrum of this fast2 ps-wide spike. For surges of longer duration, the attenuation wouldbe even smaller.These examples show that, unless a well-defined load is connectedto the transformer, expecting attenuation from the transformer mayprove to be hazardous to the health of low-power electronics connectkdon the secondary side of a transformer.

Figure 19. Effect of loading o n th e secondary sideIn contrast, decoupling is possible with a ferro-resonant line condi-tioner which is primarily intended for line voltage regulation, butwhich also provides a high degree of surge suppres sion. Figure 20shows the 6 kV incoming wave being attenuated to 60 V (100:l) onthe secondary side of the unloaded line conditioner, and to 40 V

(150:l) with a load of only 10%; at full load, an attenuation to lessthan LO V was observed. The nature of the ferro-resonant line condi-tioner is such that the decoupling improved with loading, while the

AP P E NDIX - E XAM P L E S

Figure 20. Surge decoupling by ferro-resonant line conditioner

simple transformer of Figure 18 can only act as linear dividers withload chang es. Conv ersely, the decoupling between primary andsecondary sides of the line conditioner is further seen on the oscillo-gram recorded on the input side of the line conditioner. This oscillo-gram is, in fact, a photograph of two successwe measurements, onewith no load on the line conditioner and one with a 100 W load. Theinput waves are exactly superimposed.This decoupling reflects the nonlinear behavior of the ferro-resonantline conditioner, which is significant in this case, compared to thelinear behavior of transform ers: For surge sources of lower impedancethan the generator used in these tests, or for frequencies lower thanthat contained in the 0.5 p s - 100 kHz sp ike, the transform erattenuation would become lower, in direct proportion to thecorrespo nding impedan ce change, while the ferro-resonant line condi-tioner would keep the dec?upling unchanged.FOJ worst-case demonstration, the two oscillograms of the outputwere recorded with the spike timed to occur at the peak of the 60 Hzline voltage. demonstration. The peak-to-peak amplitude of the l~n evoltage js indicated by the gray band recorded on the oscillograms byphotographically superimposing repetitive traces of the line voltage.For timings other than the peak, the small voltage oscillaiion on theoutput voltage would be completely contained within the normal peak-40-peak band of the 60 H z line voltage.Th ere is at presen t a trend to an upward spiral in specifying dB's ofattenuation in line condition ers. In the author's opinion. the point ofciting a 120 dB attenuation is moot because typical installation practicewill degrade that level of decoupling.Example No. 2 - Connections options for suppressors an effects onresidua l voltages. The author has w~tnes sed l~vely cpntroverslesamong varlous appl~cat~onnforrnat~on sources on the most effect~vetranslent suppression c onfigura t~on to be apphed Taklng, as asmphfied example, the task of spec~fy~nghe protection of a s~ ngl e -phase equlpment connected at the end of a h e w ~ t h o opportunity tod ~ v e rt he tran slent closer to the source (for instance, at the servlceentr anc e), the optlons wo&d be to connect one. two or three varistorsbetween the thre e wlres (black, wh ~te . and green ) at the end of theline However, addm onal mfo rma t~on needs to be known WtII theImpinging surge be In the normal mode (black to w h~ te ) or In th ecommon mode ([black-and-wh~tel-to-green)? Wh ere In th; equ lpm entis the most sensi twe component h e - t o - h e (most I~kely) or h e(black OR whtte)-to-greej17 Clearly, the sltuatlon ts con fus~ ltg;- andthere wtll not be a smgle, s~ mp le nswer apphcable md~ sc r~m ~n ate lyoall cases The Natlonal Electr~calCode 1241 spec~fically llows the connectlon of surge arresters (Artlcle 280-22) ~ f the tnterconnectlon occursonly by operation of the surge arrester dur~ ng he surge S m e the2 8


13/14

s(andby current of varistors is very low. this requirement can be met;furthermore. there will not be any interference with the operation of~ ~ o u n dault Circuit Interrupters if there IS only a small number ofp ro t ec to nThe set of measurements recorded in Figure 21 shows an exampleof these many options with increasing protection, albeit at increasingw t . from a single varistor to three varistors. Th e selection woulddepend on the vulnerability level and location of the equipment to beprotected. Th e impinging surge is assumed to be black-to-jwhite-and-#reen], since white and Breen are tied together at the service entrance.The line is a 75 m line and the surge is that available from the gen-orator set for a ZOO0 A 8/20 p s short-circuit impulse. Rather thanattempt to modify the setting of the generator for each case in orderto maintain constant current crest for the various configurations (animpossible task if wave form is also to be maintained). the generatorwas left unchanged. to discharge a constant total energy in the sys-tem - not a bad hypothesis for the real world. Th e current crests area]] in the range of 300 to 380 A. which is not a significant change forcomparing varistor clamping 'rollages.If only one varistor is allocated to protect the equipment. the black-to-white varistor connection (first row) affords maximum protection forthe cltetronics which are also likely to be connected black-to-white.However. the voltages between either black or white and green arelarge: this is the stress that will be applied to the clearances of theequipment. This is a good example of conversion of a normal mod etransient into a common mode.Th e configuration w ith varistor black-to-green (second row) doe snot afford very good protection for com pone nts connected black-to-white: therefore. it should be used only if there is a special need tod a m p black-to-green at a low voltage with only one varistor availableAn improved protection is obtalned with a varistor connected black-to-white. complem~ntedby a second varistor connected white-to-green(third row). Th e ultimate protection is, of course. one varistor inevery position (fourth row). but this should be required only forexceptionally sensitive loads

Example NO. 3 - Ground potential rise on data lines. A distrib-uted computer system had been installed with a central processing unitand remote terminals located in separate buildings. In a span of5 weeks during the first summer after commissioning the system, threelightning storms occurred in the area; no direct strikes were reportedon the buildings, but extensive damage was done to the circuit boardson terminals and CP U inputs.

After the first occurrence, power line surges were suspected andsome precautions were applied, when access to the hardware was possi-ble. by pulling out the ac power plugs from the CP U or terminal atthe onset of a lightning storm. This did not help. Next, isolatingtransformers were considered but, again, did not help. At this point.the author was called in for consultation, and the following proposeddiagnostic was established: the surges were not coming from the aclines, but rather were due to differences in the ground potential exist-ing between the separate buildings during flow of lightning currents.The data cables had been run in PVC conduit buried between thebuildings and, true to one tenet of steady-state noise prevention, onlyone end of the shield of the wire pairs had been grounded, with theothe r left floating. Figure 22 shows how this arrangement can producehigh voltage between a floating end of the shield and the local ground,a practice which is bound to produce a flashover and flow of surgecurr ents along unwanted paths in the circuit componen ts. Thu s, theproblem was no1 power line surges, but differential ground potential.Worse. by pulling out the ac line plugs but leaving the incoming datacables connected, the operators had unwittingly removed the localgrounding connection from the hardware frame, leaving only the datacables coming in, with a possibility of raising the complete hardwareseveral thousand volts above local grounds in the room - a dangerouscondition.

A solution to the problem could take several approache s. Radicalsolutions, such as a microwave link or a fiber optics cable, wouldindeed have eliminated the differential ground potential problems, butwere considered too expensive or too long to install.

Figure 21. Connee tlons options and effect on eommon mode or no rmal mode overvoltages29


14/14

Building I3

?=- m~ F & ~ ~ g t / ~ata Pair

F i g u r e 22. 2004 correction, caption should have read:P r o b l e m o f s i n g l e - e n d e d s h i e l d

Incidentally. part of the original puzzlement at the failures was thenotio n that opto-isolators provided in the data link ro ute should haveavoided problems. Close scrutiny o f the circuits. however, disclosedthat the opto-isolators had been provided for some other purpose: infact, the gro und p otential loop was closed by the power supply to theopto-isolator feeding the amplifiers from a local source rather than theremote source, negating the isolation function.An oth er solution, really the most simple and effective. would havebeen the replacement of the plast ic underground conduit by a cont inu-ous steel conduit linking the steel conduits used inside the buildings.This would have provided equalization of the ground potentials alongthe data cable, while allow ing the desired use of shields with one endonly at ground. Howe ver, that solution was not acceptable to the plantfacility organization.I n this particular location. a spare condu it had been buried next tothe data line cond uit. This offered the possibility o f pulling a heavyground cable in this conduit, close to the data cables. w hich could thentie the two corners of the buildings at the point of entry of the datacables, as a first step toward reducing differential ground potentials.A t first. this concept was somewhat difficult to sell to plant fac il~ties :because there is a ground grid tying lo two buildings for 60 Hz faults.further ties between the two buildings did not seem necessary. H o w -ever. our thesis was that this grid would have too high an impedanceto serve the purpose. and furthermore that the data cable run. locatedaway from the ground grid. would form a f lux-collect ing loop with thegro un d grid. Aft er leng thy discussions. the thesis was accepted andthe cable was installed.Simultaneously, the concept of tying the two ends of the cableshields to ground was proposed. with the provision of a barrier thatwould avoid the circulat ion o f power-frequency currents 1251: insertingan array o f diodes (Figure 23) at one end o f the shields reconciled theneeds of noise prevention under normal operation and the requirementof grounding at both ends during l ightning events. The forward dropof the two diodes in series (1.5 V) was enough to block any 60 Hz

Shield Grounded Twistedat Other End Data Paw

Forward drop of d iodesisolate shie ld end a t M Hz. -!.ocal GroundConduct ion of d iodes dur inga sur e effect ively groundsshiel!.F i g u r e 23.2004 correction, caption should have read:

S i n g le - e n d ed s h i e l d c o r r e c t e d by s e c o n d bondwith d i o d e scirculating current that would inject noise in the data cables. Du rin g alightning strike. however. the diodes would allow flow of current tocompensate and cancel the ground potenlial direre nces.These two cures were implemented during the first summer, and nofurth er problems occurred for the rcst o f the lightn ing season. Wh ilethese two solutions mig ht have been sufficient. the concern overanother possible fai lure of the system was suf ic ient to motivate thedesign of further protect ion: the insert ion o f a voltage clamp i n eachdata pair. This solution required some design and acceptance testingfrom the system manufacturer. so that it was not implem ented u nt ilthe next lightning season. Thus. the system survived the remainder ofthe first lightn ing season with only the first tw o remedies.Experience has shown that conclusions on the effectiveness oflightn ing protection schemes should wait perhaps as m uch as 10 yearsbefore being proclaimed. because of the large variations in lightningactivity. However, after three summers of troub le-lre e operation com-pared to three major failures in 5 weeks, the cure would seemeffective. Hopefu lly. these words will not have to be eaten by theauthor in a few years.

In retrospect, then. the following recommendations can be drawnfrom this horror tale. for retrofits or new installations:I Data cables linking separate buildings or spanning beyond a singleroom within one building should have a shield t ied to local groundat both ends of the cable I f the f irst shield provided with thecables must be left with one end floating by dikht of the systemvendor. then these cables should be ins talled with in a conrintrousmetal shield. This continu ous shield can be either a double shieldof the flexible cable. or simply a metal conduit. wi th both endsi rounded2. Substantial relief can be obtained i n retrof i ts by grounding bothends of existing shields through a low-voltage clamp. such as adiode array, that will block noise-inducing power frequencycurrents, but wil l allow the flow o f ground -potential equalizingcurrents during surges.3. The ultimate protection may be the insertion of surge-protectivedevices in each line. Howe ver, this solutio n requires careruldesign so that degradation of the signals does not occur, and resid-ual spikes are not allowed to uass thrmvh

martzloff f.d.the protection of industrial electronics and power conversion equipment against power...

Documents