In recent years, the weaknesses of Metal-Oxide Varistors (MOVs) as applied to AC power surge protection have been widely publicized. Many articles have been written and published which have highlighted the weaknesses of MOV-based surge protection but MOVs still remain the primary component used in surge protectors today even though there are much more reliable alternatives.
MOVs were developed by GE in the early 1970’s primarily for relay contact suppression. The energies developed by even large inductive loads are minor compared with the energies that can be found on AC power lines during a lightning strike. MOVs were not originally intended to handle multiple occurrences of energies up to 90 Joules with currents of several thousand amps which, according to IEEE document C62.41-1991, can be found on AC power lines within a building during a lightning strike. Also, in the 1970’s, there was not the proliferation of electronic equipment and devices that there is now, sub-micron integrated circuits did not exist, and CMOS was just emerging. All branches of electronics have moved on, including surge protection.
Shunt Mode Surge Protectors
MOV based surge protectors come under the category of “Shunt Mode”. This is a term that describes their method of operation: they “shunt” the energy away. The diagram below shows a typical MOV based surge protector.
The idea here is that the MOVs turn on and shunt the surge energy on the live wire to ground and neutral when the voltage rises above the MOV’s threshold (usually around 200V peak). This very concept is flawed though: even if MOVs did not have weaknesses, this approach only works when the MOVs are situated very close to the building ground. Most of the energy in transients and surges exists in the hundreds-of-kilohertz region. When this technology is applied to branch circuits the impedance of the ground wire reduces the effectiveness of the surge protection dramatically. Take an example where the impedance of the ground wire is just one Ohm: if there was a moderate surge of only 1000 Amps, the equipment chassis potential would be raised by 1000 Volts. The surge would then take any and all paths to ground including those through interconnecting and signal cables! This phenomenon is a major drawback of all shunt mode surge protectors.
MOVs have two main limitations:
All MOVs have a property whereby they degrade internally each time they are subjected to a current pulse above a certain threshold. Typically this threshold is about 1% of the maximum pulse current, so for a 20 mm MOV with a maximum one-time current rating of 7000 Amps, the threshold would be only 70 Amps. What this means in practice is that the lifetime of an MOV depends not just on whether its maximum current rating is exceeded, but the number of times it is subjected to a certain current pulse. Basically, MOVs are sacrificial components. The chart below, taken from manufacturer’s data sheets, shows the endurance of a 20 mm MOV.
It can be seen from the above chart that a 20 mm MOV (the type most used in surge protectors) would be expected to fail after one 7000 Amp event, two 4000 Amp events, ten 2000 Amp events, a hundred 1000 Amp events or a thousand 400 Amp events. The author has tested several 20 mm MOVs and found the actual endurance to correlate well with these numbers. And once an MOV has failed, it is no longer providing any protection. If it fails short, the fuse will blow. If it fails open, there is no way of knowing this except by disconnecting the MOV from the AC, applying a DC voltage of around 300 Volts through a current limiting resistor, and then measuring the clamping voltage. I am not aware of any product that tests its own components in this way. The result for the user would be that your surge protector may not be providing any surge protection, and you are not aware of this fact!
2. Clamping Voltage
An MOV’s Voltage-Current characteristic is similar to an avalanche diode – when the voltage across it reaches a certain point the device begins to conduct. This is known as the clamping voltage, but is usually only specified at a few amps. Like all avalanche diodes, MOVs have a certain dynamic resistance which means that the clamping voltage rises as the current through the device rises. A 150 Volt AC MOV (the type most used in AC surge protectors) has a clamping voltage of around 200 Volts DC. UL 1449 second edition, the safety standard for Transient Voltage Surge Suppressors tests the clamping voltage (referred to as the let-through voltage) at 500 Amps. The best UL rating for let-through voltage is 330 Volts and the better MOV surge protectors can meet this test, but real-world surges within buildings (as defined by IEEE C62.41-1991) can be up to 3000 Amps – six times the UL test. The author has tested 20 mm MOVs at 1500, 2000 and 3000 Amps and found the let-through voltages to be 450, 500 and 600 Volts respectively.
The problems encountered in trying to provide surge protection by shunting large, high-frequency currents have been outlined above. Is there another way? What if instead of trying to divert the energy we block it. IEEE C62.41-1991 documents the worst-case expected voltage on building wiring for category C1/B3 of 6000 Volts. (This limitation is primarily due to arc-over at the service entrance and within the branch circuits.) So, if we are able to withstand 6000 Volts for a brief period of time we can actually block a surge. It is also a fact that most of the surge energy is up in the hundreds-of-kilohertz region, so we also have a clue of how to block 6000 Volt surges: If we can produce a low-pass filter with components able to withstand 6000 Volts and pass realistic load currents of 15 to 20 Amps then we have a totally new approach to surge protection. This is the concept behind SurgeX surge protection.
SurgeX Surge Protectors
The figure below shows a simplified block diagram of a SurgeX surge protector.
The simplified diagram shows the similarity to a low pass filter consisting of a series inductor and a parallel capacitor. The surge reactor is a physically large inductor which is capable of passing 20 Amps of current at 60 Hz and will withstand in excess of 6000 Volts without breaking down. The surge reactor is actually more than just a simple inductor in order to provide the exact parameters required but, for the purpose of understanding the technology, it can be thought of as an inductor. The dynamic clamping circuitry was designed by both computer modeling and bench testing to operate in conjunction with the surge reactor. The inductive reactance of the surge reactor provides the series impedance to incoming transients and surges. The time-constants and capacitors in the clamping circuitry provide the low impedance parallel path. The clamping circuitry is activated when the sensing circuit detects a rapid rise of 2 Volts above the power wave and the remaining surge energy is pulled into a bank of capacitors. By using this technique, a much lower let-through voltage can be achieved than by using a simple MOV or avalanche diode clamp. Note that there is no connection to ground. Ground is not part of the system. This means that the ground is not contaminated by SurgeX, and also that a low impedance ground is not required in order to provide fully effective protection.
To summarize: the surge reactor blocks the high-frequency energy and also limits the current. It slows down surge energy so that the damaging fast rising spike is prevented from reaching equipment downstream. The remaining low-frequency energy is then contained by the clamping circuitry.
Advantages of Series Mode
1. Protects against the worst-case surges found within a building.
2. Contains no sacrificial components.
3. Does not degrade or fail with repeated use.
4. Produces no ground contamination.
5. Does not require a low impedance ground for full efficiency.
6. Due to the fact that SurgeX is a low-pass filter, transients and noise are also attenuated.
Although this article is primarily about surge protection, a short discussion about power conditioning may also be of interest to the reader. Power conditioning is a much maligned term, being used to describe anything from a unit with a single MOV to a comprehensive system consisting of top-quality surge and transient protection including EMI and RFI filtering. In my opinion, one MOV (or even three MOVs) does not constitute power conditioning – this is surge protection only (and dubious surge protection at that). Power conditioning should provide clean AC free of damaging surges, transients, EMI and RFI; and also should not contaminate the ground wire. The importance of true power conditioning cannot be over-emphasized for applications where microprocessor based equipment is located in the same building as heavy factory equipment. Transients produced by large inductive loads can interfere with data-logging equipment and cause microprocessor equipment to suffer memory corruption, re-booting or hang-up problems. SurgeX has been used successfully to solve all these problems.
AC power EMI/RFI filters present an interesting topic of discussion by themselves. Most EMI/RFI filters are designed for the bench source and load impedance of 50 Ohms. How many AC lines are 50 Ohms impedance, and how many loads are 50 Ohms impedance? The answer to the first is, “None”, and the answer to the second is, “Only if the load is around 300 Watts”. Computer modeling has shown the severity of such mismatches on a standard EMI/RFI filter: at certain frequencies the “filter” can resonate and actually amplify the noise instead of attenuating it. By adding additional components, a filter can be designed that is tolerant of source and load mismatch providing superior performance in real-world applications. The SurgeX? line of surge protectors incorporate such an EMI/RFI filter as well as totally effective surge protection.
Switching Power Supplies
A special mention must be made concerning protecting switching power supplies and MOV based surge protectors. The front-end of a switching power supply is a bridge rectifier and a large capacitor bank. This presents a very low impedance to the AC line, and computer modeling has shown that, in the event of a surge, the power supply actually ends up protecting the MOV! There are two reasons for this: firstly, the voltage has to rise to the point where the MOV begins clamping which can be from 300 to 400 Volts depending on the quality of the surge protector. During this time the power supply is absorbing energy. Then the voltage continues to rise due to the dynamic resistance of the MOV up to 400 to 600 Volts. Again, the power supply is absorbing energy during this time. The bottom line is that, even when they are working, MOV based surge protectors do a particularly poor job of protecting switching power supplies.
This document is a compilation of several studies on surges in and around buildings and contains valuable information on surge voltage, current, energy and waveforms. It is the foundation on which other standards were built. The SurgeX? line of surge protectors were designed to handle category C1 and B3 surges (from the service entrance along to branch circuits).
UL 1449 Second Edition
The safety standard for Transient Voltage Surge Suppressors. However, this standard tests only for safety: it says, “This surge protector shouldn’t start a fire and burn your building down”. The only performance test is to measure the let-through voltage at 500 Amps, which is not a real-world situation. There is now also Adjunct Endurance Testing, which is optional – manufacturers have to request this testing, but this does guarantee performance. In order to pass the Adjunct Endurance Testing a product must survive 1000 applied surges of a specified voltage (typically 6000 volts) with a let-through voltage no more than a specified voltage (typically 330 volts) with no degradation or failure.
Federal Commercial Item Description (CID) A-A-55818
This Federal CID was created in order that federal purchasing agents may purchase surge protectors with known tested, guaranteed performance. It uses the a system of Grades, Classes and Modes which are derived from IEEE 62.41-1991 Adjunct Endurance Testing. There are three Grades which designate the type of applied surge. Grade A, the highest, is 6000 Volts and 3000 Amps, Grade B is 4000 Volts and 2000 Amps, Grade C is 2000 Volts and 1000 Amps. There are three Classes which designate the Let-through Voltage. Class 1, the best, is 330 Volts, Class 2 is 400 Volts and Class 3 is 500 Volts. Finally, there are two Mode designations, 1 and 2. Mode 1does not contaminate ground, Mode 2 does contaminate ground. The creation of this CID was an important step in standardizing surge protection specifications. Hopefully, all manufacturers of surge protectors will eventually submit their products for Endurance Testing. The SurgeX? line of surge protectors are rated A-1-1.
MOVs were not originally developed to be used for AC power surge protection. Not only are they unreliable, providing mediocre protection at best, but the very concept of shunting surges and high-frequency transients to ground on branch circuits is flawed. SurgeX protection meets the most stringent requirements of the Federal CID, was designed to withstand an unlimited number of category C1 surges as defined by IEEE 62.41-1991, filters out transients, and does not require close proximity to the building-ground to operate with total effectiveness.
Andrew Benton received his engineering degree at Cambridge University, England in 1976. Within three years he was head of Electronics at Kent Industrial Measurements in Cambridge, and moved to the United States in 1980 to work at General Electric in Lynchburg, VA. He moved to New Jersey three years later to work for Burroughs (now Unisys) and became head of the Storage Products Group. During his nine years at Unisys Andrew received three awards including the corporate-level Achievement Award for Excellence. He left Unisys in 1992 accepting the position of Director of Engineering at Pacific Rim, Inc. and two years later, with two partners, he formed New Frontier Electronics. Andrew’s hobbies include: hiking, cosmology, reading (non-fiction), natural history, the environment, global warming and restoring classic British cars.