The Dilemma of Surge Protection vs.

Overvoltage Scenarios:
Implications for Low-Voltage Surge-Protective Devices

Arshad Mansoor, Member, IEEE François Martzloff, Life Fellow, IEEE

Power Electronics Applications Center National Institute of Standards and Technology
Knoxville TN 37932 USA Gaithersburg MD 20899 USA
Amansoor@epri-peac.com f.martzloff@ieee.org

© 1998 IEEE
Reprinted, with permission, from
Proceedings, 8th Annual Conference on Harmonics and Quality of Power, Athens, October 1998

Significance:
Part 2 Development of standards
Part 7 Mitigation techniques

The application of surge-protective devices (SPDs) in low-voltage AC power circuits, in particular metal-oxide
varistors (MOVs) has been influenced by the perception that low-limiting voltage is a desirable characteristic.
Unfortunately, this low limiting voltage – intended for surge protection – makes the devices more susceptible to
fail under conditions of extended temporary overvoltage (TOV).

Like any electronic component, SPDs will fail if overstressed beyond reasonable limits, and this is not a cause for
rejecting their application, but a cause for concern on ensuring that the failure mode – rare as it might be – will be
acceptable.

This acceptability must also take into consideration the effect of the available fault current that the power system
can deliver at the point of connection of the SPD. This point needs to be more clearly and specifically stated in
emerging standards on SPD applications.
 The Dilemma of Surge Protection vs. Overvoltage Scenarios:
Implications for Low-Voltage Surge-ProtectiveDevices
Arshad Mansoor, Member, IEEE Fraqois Martzloff, Fellow, IEEE
Power Electronics Applications Center National Institute of Standards and Technology*
Knoxville TN 37932 USA Gaithersburg MD 20899 USA

Abstract - The application of surge-protective devices in low- This dilemma of surge protection versus overvoltage scenarios
voltage systems faces the dilemma of providing effective limiting has been created by the industry's obsession with providing
against surges while not attempting to limit the temporary very low clamping voltages for surge mitigation (Martzloff &
overvoltages that do occur in a power system The paper illustrates M y , 1989[1]). And now, the need to ensure coordination of
this dilemma with specific scenarios and presents recommendations the "cascade" of the service-entrance SPD and the plug-in
for reconciling these two conditions through adequate design and SPDs has exacerbated this situation. The issue of cascade
more explicit standards. coordination has already been debated at length in the literature
(Martzloff & Lai, 1991 [2]); (Stonely & Stringfellow, 1991 [3]);
I. INTRODUCTION (Hostfet et al., 1992 [4]); (Rousseau & Perche, 1995 151).
Further debate or exhaustive references to the many papers on
The concept of b4Whole-House Surge Protection" has become that subject is not our purpose. It is mentioned here only as a
a popular subject of discussion and has in fact been contributing factor to the dilemma, but a factor that cannot be
implemented by several utilities in North America. In this ignored in a complete assessment.
approach, the utility will install a surge-protective device (SPD)
at the service entrance of the customer and provide additional 11. SURGE PROTECTIVE DEVICES
plug-in SPDs. These additional SPDs are presumed to be FOR LOW-VOLTAGESYSTEMS
wellcoordinated with the service-entrance SPD, and are
installed within the customer premises, presumably at the point I h e introduction of metal-oxide varistors (MOVs) in the
of connection of so-called "sensitive appliances" such as home seventies was a timely innovation, concurrent with the
entertainment,computers, and sophisticated kitchen appliances. increasing use of semiconductors in consumer products. While
these semiconductors opened new opportunities, their relatively
The main purpose of these SPDs is to protect sensitive low tolerance for surges created a strong demand for better
equipment against surges, a mission that they can accomplish surge-protective devices. Unfortunately, market competition
quite well. However, the failure mode of these devices under encouraged a downward "auction" that led to attributing high
temporary overvoltages (TOVs) that might be expected under value to low clamping voltages, a situation unwittingly
abnormal but possible conditions of the power system has encouraged by the listing of "transient suppression levels"
become cause of some concern for utilities and their customers.
stipulated in UL Standard 1449 [6]. The list begins at 330 V
for SPDs intended for 120 V circuits, although there is good
An SPD should not be expected to protect downstream evidence that most consumer loads do not need such a low level
equipment in the case of a TOV and then return to normal of protection (Anderson & Bowes, 1990 171); (Smith &
operation, as it does by definition for surge protection. The Standler, 1992 [8]).
dilemma for SPD designers is whether to select a maximum
continuous operating voltage (MCOV) high enough to survive The generic structure of typical low-voltage residential power
common TOVs - but at the price of diminished surge systems is shown in Figure 1 for the case of a detached home.
protection -or to select surge protection with a lower MCOV Underground service has similar characteristics. This system
- and then accept failure of the SPD for infrequent but extends from the outdoor line-side of the service drop all the
possible TOVs. In any case, one should expect that if a TOV way through the premises wiring, including plug-in type SPDs.
at any level would cause the SPD to fail, that failure mode Figure 1 also shows the various locations where an SPD can be
should be acceptable. installed.
* Electricity Division, Electronics mad Electrical Engineering Laboratory,
Technology Administrution, U.S.Del~urtmentof Commerce. Typically, there are six locations. The first three: O at the
Contributions~m the NufionulInstitute ojstundardr and Technolagy are not
outdoor weather-head, Q at the service entrance, and O on the
subject to U.S. copyright. line side of the main disconnect are within the scope of IEEE
Std C62.34 [9]. The next three are within the scope of ongoing
Paper accepted for presentation at the 8U International IEEE project P62.62 [lo]: @ at the
I load side of the panel
Conference on Harmonics and Quality of Power disconnect, 6 at a permanently wired receptacle, and 8 as a
ZCHQP '98,jointly organized by ZEEEIPES and NTUA, plug-in device. Locations O through Q are within the premises
Athens, Greece, October 14-16,1998 wiring and therefore under the control of the end-user, while
0-7803-5105-3/98/$10.00 0 1998 IEEE locations O and Q are under the control of the utility.
 III. FAILURE MODES UNDER SURGE
CONDITIONS
For a correctly applied SPD, failure under surge conditions
should be a very rare occurrence. Nevertheless, one can
enumerate the following failure scenarios in a field application
-including misapplications:
A single, large, and not anticipated impinging surge
exceeds the capability of the SPD. An example of this
situation can be the presence of switched capacitor banks;
A succession of surges, such as multiple lightning strokes,
Branch arcuit
exceeds the capability of the SPD. This situation has been
identified for distribution arresters (Darveniza, 1997 [I31)
and might also occur for low-voltage SPDs;
A thermal runaway is launched in an SPD exposed to high
ambient temperatures at the time when a surge (within
specifications for normal ambient) occurs;
0 Loadside d Service dmp @ Load side d mPh Overevnntd8W In the questioriable scenario of an alleged "degraded" SPD
@ Meter base adapter @ Penanently cumded maplade (Stringfellow, 1992 [14]), a thermal runaway is launched
@Linesirkjdmainovercurrentdevke @Ccmicme&dwplug.ln
Nat shown: SPDs incorporatedin equipment power port
by the heat generated during a within-specifications surge.
Figure 1 - Possible SPD locations for a residential building For all these scenarios, the ultimate failure mode depends on
the fault current that the power system can deliver at the point
of connection of the SPD. For maderate fault currents, such as
An emerging requirement in standards for application of the on branch circuits, the disconnector can generally provide
low-voltage SPDs under development at the International protection. The fuse design can still be a challenge: carry the
ElectrotechnicalCommission (IEC 61643-1 [I 11) as well as at load current, carry the specified surge current, but melt in case
the IEEE (P62.62 [lo]) is the provision of a "disconnector" of a power-frequency fault current resulting from failure of the
intended to discomect a failing SPD. Failure under conditions SPD component. For very high available fault currents, such
exceeding the SPD capability is recognized as unavoidable, but as that prevailing at some service entrances close to a large
the consequences of such failure are made acceptable thanks to distribution transformer, successful clearing may be a greater
the action of the disconnector. challenge. Furthermore, coordination of overcurrent protection
is more difficult, compared to branch circuits inside the
Unfortunately, some ambiguity has crept in the interpretation building where the wiring impedance and the rating of circuit
of this requirement. In some cases is has been interpreted as breakers in the panel can ensure proper coordination.
only disconnectingthe failed SPD component from the power
system, but leaving the load energized -and without further IV. FAILURE MODES UNDER TOV CONDITIONS
surge protection (Martzloff, 1998 [12]). If the disconnector is
of this latter type, the SPDs components of an SPD package Three major types of TOV-induced failures can be identified
will fail under TOV conditions, presumably in a safe manner, for low-voltage SPDs:
but then allow the TOV to be applied to the downstream load.
This is undesirable for the typical user who values equipment Moderate TOVs associated with power system faults, such as
protection above continuity of operation. Finally, there is the a line-toearth fault in a three-phase system, creating a 1.73
worst case, as reported in many anecdotal instances, where the times n o d line voltage in the other phases. Ferroresonance
can also produce moderate but significant overvoltages.
failure mode of SPDs under TOV conditions has not been
graceful, to say the least. These instances, while not very Extreme TOVs associated with the commingling scenario
frequent, have led to new testing requirements for failure (accidental fall of conductors of a higher voltage upon
modes from the Underwriters Laboratories in the updated conductors of a lesser voltage). No conventional, varistor-only
Second Edition of their UL 1449 [6] Standard for low-voltage SPD can be expected to survive such a scenario.
surge-protectivedevices ("TVSS" in the industry jargon).
Double voltage TOVs associated with the loss of neutral in a
Given this unsettled situation, it will be useful to review the single-phase, threeconductor, earthed center-tap system such
scenarios that can lead to failure of an SPD component in an as the 1201240 V service typical of North American systems.
SPD package, in particular under TOV conditions. For the Depending on the philosophy of the system designer, in
sake of completeness, we will describe first some failure particular the utility for the case of a service-entrance SPD.
scenarios under surge conditions, as they have some bearing on survival or expected but acceptable failure can be stipulated for
the disconnector design. the loss-of-neutral scenario.
a) Moderate TOVs: System Faults

Among abnormal conditions that can produce temporary

overvoltages, we give two examples of incidents resulting in
overvoltages not exceeding twice the normal voltage. Some
SPDs based on the misconception that a very low clamping
voltage is desirable might not survive such moderate TOVs.

Single-phase faults to earth on a three-phase system produce

a shift in the unfaulted phases. The severity of the voltage rise
depends on the fault location, the system impedance, and the
earthing practices. For a 'IT power system in steady-state, the Source: EPRl Report [l5]
neutral will by symmetry be at the same potential as earth, as in -
Figure 3 Temporary overvoltage caused by ferroresonance
Figure 2 (a). However, if an earth fault occu on one phase, it
will cause a shift of potential as the system attempts to maintain
balance. The neutral will be elevated and a corresponding shift b) Extreme TOVs: Commingling
will be experienced by the other two phases. In the worst case
of a completely isolated system with a bolted fault to earth on In this real-life scenario, the low-voltage SPD connected on the
one of the phases, we could have the situation depicted in secondary side of the distribution transformer, and normally
Figure 2 (b). This shows a neutral which has been elevated to energized at its rated voltage, is first brought to failure by the
one per unit, causing the unfaulted phases to drift up to 1.73per large overvoltage resulting from commingling conductors. In
unit with respect to earth. an overbuilt system, a collision of a vehicle with a pole, or
breaks caused by icing, the conductors of the higher voltage
For three-phase systems with an artificial neutral through an distribution system or sub-transmission system can fall on the
earthing transformer or a finite resistance, the earth impedance lower voltage distribution system. Such accidental contact
is high. This arrangement limits the fault current during a injects an intruding voltage for a few cycles, until the higher
singleline-toearth fault, but allows enough for fault detection voltage breaker clears the fault.
using overcurrent relays. It also inserts some impedance
between the system neutral and the actual earth so that the Figure 4 shows a simplified one-line diagram of the two
voltage on the unfaulted phases will shift toward somewhat less medium-voltage systems (the intruding MV1 and the victim
than line-line value as shown in Figure 2 (c). MV2) being accidentally commingled. In this figure, the bond
between the two systems earth connections is shown in dotted
line to present the generic case of commingling. In an overbuilt
system with common neutrals, a solid bond exists where the
dotted tine is shown in the diagram. In the scenario of a simple
crossing of two systems (not overbuilt along the right of way),
or delta systems, a solid bond might not be present, and the
fault current from MV1 will involve the earthing impedances
shown in the diagram In that case, the intruding voltage Vi
might be less than the system voltage MV1 but still enough to
precipitate failure of an SPD on the secondary.

Figure 2 - Effect of a single phase-to-earth fault

Series resonances occur in a power system when a series

circuit consisting of an inductance and a capacitance is excited
at its natural frequency. As an example of this condition,
Figure 3 shows a portion of a temporary overvoltage resulting
from a power system switching incident (restoring power phase
by phase after interruption [15]), which produced a sustained Figure 4 - Commingling overvoltage applied to the distribution
rms voltage exceeding 150%and lasting four seconds. transformer primary and reflected on the LV side
c) Double-voltageTOV (Lossof neutral) 1996 National Electrical Code, 230-65, Availabk Short-circuit
Cumnt
There are many situations where loss of neutral can occur. "Service equipment shall be suitable for the short-circuit
Some can be of a transient nature, such as a loose connection, current available at the supply terminals."
while some might be permanent until repaired, such as a
mechanical break or corrosion of the neutral conductor. While IEEE C62.341996, Standard for Performance of Low-Voltage
the latter might not linguistically fall under the label of Surge-ProtectiveDevices (Secondary Arresters), Article 7.6
'temporary', the consequences are the same. In a three-wire, "If the manufacturerclaims a fault current withstand rating,
then that rating shall be verified."
single-phase system typical of North American practice, this
condition has been observed many times.
IEC 61643-1 (FDIS November 1997) Surge protective devices
Figure 5 shows a system where one side of the supply (Ll) is
-
connected to low-voltage power distribution systems Part I:
Performance requirements and testing methodr, Artick 6.2.11
lightly loaded, while the other side (L2) is heavily loaded, "The SPD shall be able to carry the paver short-circuit
Under normal conditions, the two sides remain at normal current until it is interrupted either by the SPD itself; by an
voltage. Should the neutral connection be lost, then the voltage internal or external overcurrent disconnector, or by the backup
at mid-point is determined by the ratio of impedances on the overcurrent protection."
two sides: the L1 side experiences an overvoltage that can
approach twice nonnal. Any SPD connected to this side will It is noteworthy that none of these documents specify a value
then be exposed to the twice-normal voltage, with an available for the available short circuit, but rather leave it to the
current determined by the impedance 22. This current, by the discretion of the manufacturer, while imposing criteria of
very design of the circuit, will not be interrupted by the acceptability after the failure. The NEC tersely requires the
overcurrent protection and is available to generate substantial device to be "suitable," without elaboration. The two standards
heat in the SPD that has failed as the result of the overvoltage. applicable to a service-entrance SPD - where the available
A disconnector designed to clear larger fault currents, such as fault current can be quite high -acknowledge the possibility
those occumng if the SPD fails while the neutral is connected, of failure and significance of the available fault current, but do
might not be capable of clearing the limited current. Other not stipulate specific values.
schemes are then necessary to ensure an acceptable failure
mode, such as a thermally-activateddisconnect. In an attempt to obtain information on what levels of fault
currents should be considered, an informal survey was
conducted among a few utility engineers. It turns out that very
few utilities limit the available fault current at the service point
of residential customers but many people are under the
misconception, as cited below, that all residential service load
centers and breakers are limited to an available fault current of
10 kA and therefore they assume that actual available fault
current must be less than 10 kA.

Interestingly, none of the individuals consulted on this issue

Figure 5 - Three-wire, singlephasesystem could cite a standard or even a document with the status of
where a broken or intermittent neutral connection a consensus guide that does stipulate a current value. Two
creates a voltage imbalance perceptions seem to prevail among the individuals consulted, as
quoted below:
Based on the observation that breakers for typical
V. AVAILABLE FAULT CURRENT residential service panels (up to 200 A) generally have an
interrupting capacity of 10 kA, it would be logical that 'the
As described in the preceding paragraphs, several scenarios can
availablefault current at the service entrance would be in
produce failure of the SPD, ultimately offering a very low (but
the same order of magnitude.
not zero) impedance to the flow of the fault current that the
power system can deliver. For each scenario, the available fault Based on the reality that a service entrance connected
current at the point of connection of the SPD will have a very clo,e to a large distribution transformer(such as a garden
significant effect on the failure mode. apartment or high-rise)will have availablefault currents
in excess of I0 kA, it would be prudent to review the
The significance of available fault current is recognized in specifics of the situation.
standards, but the value that should be considered is generally
left undefined. The following statements can be found in Furthermore, these standards do not require that the SPD
application data state a limit of acceptable fault current.
published standards.
 VI. EMERGING STANDARDS From this brief overview of emerging standards, it appears that
on the low side of available fault current, both the UL and the
a) SPDs installed downstream from the service entrance IEEE standards have recogruzed the issues of acceptable failure
modes associated with temporary (or quasi-permanent)
In recognition of the issues raised by failure modes of the SPDs overvoltages. On the other hand, at the high side of available
within the scope of UL Standard 1449 - downstream of the fault currents, it seems that insufficient recognition of the issue
main disconnect -the second edition of this standard [6] now still prevails, as discussed in the preceding section on the
requires demonstration of an acceptable behavior in the failure significance of the level of available fault currents.
mode of SPDs exposed to various overvoltage scenarios.
Specifically, among the many tests required by UL,three tests Nevertheless, Figure 6, offered as background information in
address the issue: a tutorial addressing considerations on the revenue meter
environment (Ward, 1980 [16]), shows typical values of fault
Temporary overvoltage with high available current - This test current as a function of distribution transformer size and length
is stipulated un&r article 37.2 of UL 1449, calling for extended of service drop. That figure clearly shows values in excess of
exposure to 125% of normal line voltage with either an 10 kA, but somehow this information has not been fully
acceptable temperature equilibrium being attained or until an recognized by the SPD community.
internal disconnect device operates. This test will demonstrate
capability of sustaining moderate overvoltages, primarily - -

providing a margin against high system voltage, but not the AVAILABLE FAULT CURRENT
WITH rOorr SERVICE COR 3
higher levels of temporary overvoltages covered in the next TRANSFOWER SIZES
test. The available fault current specified for this test is defined
as a function of the ampere rating of the service over a range of
200 A to 25 000 A.
Fullphase voltage with high availablefault current - This test
is stipulated under article 37.3 of UL 1449, calling for exposure
to the "full phase voltage" as shown on Figure 2(b) of this
paper. The same criteria as above apply, namely acceptable
temperature equilibriumor operation of an internal disconnect.
The available fault current specified for this test is &fined as
a function of the ampere rating of the service over a range of tm AL SERVICE -CABLE LENOW -FEET
200 A to 25 000 A. Source: (Ward, 1980 [I 61)
Overvoltage with limited current - This test is stipulated under figure 6 - Fault current amplitude as a function of.
article 37.4 of UL 1449, with overvoltage values presumably transformer rating and length and size of service drop
corresponding to a loss of neutral scenario, and the associated
low values of current supplied by the connected load, as in
Figure 5. vn. POSSIBLESOLUTIONS
It is noteworthy that in the stipulations of these three tests, the
emotionally charged word "failure" is not used. Instead, a list Among possible solutions, two approaches may be
of unacceptable conditions is given, including emission of considered: making the SPD less sensitive to TOVs, and,
flame, molten metal, flaming particles, charring of adjacent in any case, ensure that if failure is unavoidable under
material, ignition of enclosure, or creation of openings leaving extreme stress, it will be in an acceptable mode.
live parts accessible.
The obvious way to desensitize SPDs to TOVs is to
b) SPDs installed upstream from the service entrance design them with a higher MCOV. However, as the
higher MCOV in a varistorsnly SPD means a higher
For SPDs connected upstream from the service panel, the surge-limiting voltage, there is a limit beyond which such
recently published IEEE Standard C62.34 does describes a an SPD becomes useless (op cit., [2-51).
loss-of-neutral scenario with limited current, similar to the UL
37.4 test. However, a demonstration test is not mandated, as
the consensus development process settled for a weaker A pos:.ble solution may be in reviving the concept of a
statement: "if the manufacturer claims a loss of neutral gapped arrester for the upstream SPD of a "whole house"
withstand capability ... then that capability shall be venjied ..." scheme (Mansoor et al., 1998 [17]). There, the initial
Thus, the inference might be made that if no claim for loss-of- let-through associated with the gap volt-time response
neutral withstand capability is made, no demonstration test is can easily be mitigated by the downstream SPD, while
required. However, an additional paragraph in the standard the gap prevents the SPD from becoming involved with
does mention "...must fail in an acceptable manner." moderate TOVs.
Of course, for the (rare) commingling scenario, little can Rousseau, A. and Perche, T., "Coordination of Surge Arresters
be done but to ensure a graceful failure. This condition in the Low-Voltage Field," Proceedings. 7th International
TelecommunicationsConference (IhTELEC), 9SCH35824.
should be an implicit requirement, but, as discussed in the
section on available fault current, the implications of such UL 1449 Standard for Safety for Transient Voltage Surge
Suppressors, Second Edition, August 1996.
a requirement apparently have not been recognized by all
interested parties. Even among the community of SPD Anderson. L.M. and Bowes, K.B.,'The Effects of Power-Line
Disturbances on Consumer Electronic Equipment." IEEE
engineers, there has been some reluctance to accept the Transactions PWRD-5, No.2, April 1990.
concept that temporary overvoltages should be addressed
Smith, S.B. and Standler, R.B., 'The Effects of Surges on
in documents discussing the surge environment. Electronic Appliances," IEEE Transactions PWRD-7, No.3,
July 1992.
VIII. CONCLUSIONS IEEE Std C62.34, IEEE Standard for Pe$ormance of Low-
Voltage Surge Protective Devices (Secondary Arresters), 1996.
The dilemma of providing a suitable surge protection of IEEE Project P62.62 - Application Guide - Low-Voltage Surge-
load equipment by means of surge-protective devices, Protective Devices (ongoing).
while ensuring acceptable response of these surge-
protective devices to unavoidable temporary overvoltages IEC Document 37A163EDIS -Draft IEC 6 1643-1: Surge-
protective devices comected to low-voltage distribution systems
raises several application issues that demand attention. Part 1: Performance requirements and testing methods, November
"Acceptable response" can be interpreted either as survival 1997.
of the SPD (a challenge to the coordination of cascades) or Martzloff, F.D., "Interpretation and Misinterpretation of TVSS
as accepting failure, but within well-defined conditions of Discomector Indications," Scheduled for publication in Power
the failure mode (a challenge for applications where high Quality Magazine, July/ August 1998.
available fault currents prevail). Darveniza, M., Turnma, L.R., Richter, B., and Roby, D.A.,
Coordinating a cascade of surge-protective devices can be "Multiple Lightning Currents and Metal-Oxide Arresters," IEEE
Transactions PWRD-12, No.3. July 1997.
solved by providing a gapped arrester at the service
entrance, which will coordinate with the de facto situation Stringfellow, M.F., "Fire Hazard of Surge Suppressors,"
of low limiting voltage SPDs inside the building. Proceedings, Fiflh Annual Power Quality Conference, Irvine
CA, 1992.
The need for a service-entrance arrester to withstand the EPRI Report An Assessment of Distribution System Power
scenario of lost neutral can be satisfied by a gapped Qualify,Electric Power Research Institute. Palo Alto CA, 1996.
arrester having sufficient maximum continuous operating
voltage capability. Ward, D.J., "Secondary Fault Currents at the Service Entrance."
Proceedings, 55* Annual Electric Meter School and Conference,
Emerging standards for low-voltage SPDS have given new University of Florida, Gainesville FL,1980.
recognition to the importance of taking into consideration Mansoor, A., Martzloff, F.D.. and Phipps, K.O.. "Gapped
temporary overvoltages in the design of SPDs. Arresters Revisited: A Solution to Cascade Coordination" Paper
PE-114-PWRD-0-12-1997, IEEE Winter Power Meeting, 1998.
Notwithstandingconclusion (3,SPD application standards
as well as performance and test standards should be more Arshad Mansoor (M' 1995) is an
explicit in defining how to deal with the issues raised by Electrical Systems Engineer at the EPRI
available fault current in case of unavoidable SPD failure. Power Electronics Applications Center
(PEAC). He received his MS and Ph.D.
in electrical engineering from the
IX. REFERENCES University of Texas, Austin in 1992 and
1994 respectively. His areas of interest
Martzloff, ED.and Leedy. T.F.,"Selecting Varistor Clamping include power Quality, power systems
Voltage: Lower Is Not Better!" Proceedings, 1989 Zurich transients analysis, harmonics, surge
International EMC Symposium. propagation and protection, and EMTP
- -
Martzloff, F.D. and Lai, J.S., "Cascading Surge-Protective model development.
Devices: Coordination versus the IEC 664 Staircase,"
Proceedings. PQA 91 Conference. Francois Martzloff (M'1956, F'1983)
Born and educated in France, with
Stringfellow, M.F. and Stonely, B.T., "Coordination of Surge additional MS degrees from Georgia
Suppressors in Low-Voltage AC Power Circuits," Proceedings, Tech and Union College, worked at
Fonun on Surge Protection Application, NISTIR-4657. August General Electric for 29 years and now 13
1991. years at the National Institute of
Hostfet, O.T., Hervland, T., Nansen. B. and Huse, J., Standards and Technology. He is
"Coordination of surge-protective devices in power supply contributing to several technical com-
systems: Needs for secondary protection," Proceedings, mittees for the development of standards
International Conference on Lightning Protection, Berlin, on EMC, surge protection and Power
September 1992. Quality in the IEEE and the IEC.