Surge Suppressors

As an introduction to this section, let’s take a few moments to talk about terms and how we go about defining things. By now, we know that the term “spike” is a slang term for a positive-going impulse. We have used the term “noise” rather loosely and have defined it as any signal other than the desired signal.

So with that backdrop, we come to the term “surge.” We have described a surge as an rms voltage higher than the standard limits. The term rms (root-mean-square) is the key.

A surge describes a sine wave, or series of sine waves, that exceed standard values. This means the top of the waveform curve exceeds normal instantaneous values. A half cycle, for instance, lasts 8.34 milliseconds. Impulses are almost never longer than a dozen or so microseconds. An impulse and a surge are vastly different events. So, it's with this semantical incongruity that we tackle the subject of surge suppressors or surge protectors. These are devices that are designed to protect equipment from high energy impulses rather than surges. But that is not the only difficulty.

We are in the last half of the guide now, and the remainder of the guide will focus on specific product technologies and applications. Surge suppressors are the first of many devices that we will investigate. But we must be wary. We are in the world of PCs—consumer products. This means that packaging is as important as protection. It means that price is as important as performance. It means that persuasion is important to perceived value. To put it bluntly, there is more smoke and mirrors sold under the guise of protecting computers from lightning hits and such than in any other area of the industry. Why? Because the world is full of folks who think nothing of spending $2000 for the latest whiz-bang computer and , then, will start haggling over a $10.00 surge strip to protect the whole installation. You get what you pay for, as we shall soon see.

In this section, we are going to start with the simple and move toward the more complex. It will be impossible to cover every combination of components used in the various products on the market. We will not attempt to deal with the wild claims of some companies or the outright fraud of others. What we will do is arm the reader with enough understanding of the principles behind product designs that intelligent choices can be made.

Simple Surge Suppressors

A simple surge protector comes in one of several forms. But at the heart of the device is usually one and only one component—an MOV. Currently, this product appears in three incarnations. The first of these is the surge strip that we are all familiar with. This is a multi-outlet power bar. The second form of the same basic product is the single-outlet device that is plugged into the wall ahead of the CPU. The idea behind this form is that one of these units would be required for each of the pieces of equipment that needs protection. The third form of this product idea is the receptacle itself. Special receptacles can now be purchased so that surge suppression can be built right into the wall.

With a few exceptions, the only active component installed inside these products is a single MOV that is connected between the line and neutral. ill. 8-1 shows an oscilloscope display of an impulse from a test generator. It traces a 900-volt impulse (between the line and neutral). ill. 8-2 shows how the top of the same impulse has been clipped off (or clamped) by the MOV. it's this action of the MOV which diverts potentially damaging energy away from the load. ill. 8-3 shows how this appears in a circuit diagram.

ill. 8.1. In this photo of a 900-volt impulse, we see that the magnitude of the impulse is great enough that the peak of the pulse goes off the top of the screen.

ill. 8-2. With the insertion of an MOV between the line and neutral, the 900-volt impulse is clipped off or clamped.

A Test for Surge Suppressors

We mentioned that we used a test impulse voltage of about 900 volts. There is a good reason for this that is vitally important to the proper application of this product. At various points throughout this guide, we have made reference to the kinds of power disturbances we might expect at various locations in a building. That is why there are different stages of protection required, depending on the size of the event we might expect.

The MOV used in most surge suppressors is designed with the rationale that lightning arresters have been installed in accordance with the activity in the area. In addition, there seems to be a common misconception that surge protectors will stop most anything that comes down the line, and if they don’t, they are no good. Let’s put this into some orderly perspective.

ill. 8-3. How a simple surge suppressor works.

The IEEE 587 standard talks about 3000- or 6000-volt test impulses as does the new UL 1449 standard (discussed in the next section). Naturally, it would seem to be desirable to test products to these standards to see if they work. We couldn’t agree more! A recent magazine article that compared surge suppressors found that the typical 6000-volt, 200-ampere impulse showed no significant difference between products. Why? The only component reactive to this type of energy in most of the products tested was the same MOV manufactured by the same company.

The testers evidently didn’t suspect this so they increased the amperage of the impulse to 500 amperes. Here’s a typical example of what happened to one product.

…produced excellent results through the first half of the tests, clamping 6-kV spikes down to as low as 60 volts. Then, poof—the unit croaked. We were not impressed.

The unit croaked because the MOV inside the device finally gave up its life after handling repeated hits (about 192 in this case) of energy that possibly were in excess of its specified ratings. Actually, the performance described was very good and would have done exactly what the owner bought it for. The surge suppressor that took all those 6000-volt, 500- ampere hits before it “croaked” was priced at less than $40.00. That’s an incredible performance for the money. Tests need to be designed that properly reflect the design differences between products, not just the survivability of the same type of MOV from one product to another. We believe that the new UL 1449 standard does this.

It is for these reasons that we choose the more realistic level of about 900 volts for our tests. We wanted to show the performance characteristics of differing technologies, not blow up black boxes. As we will see in this and later sections, this test is just as dramatic at 900 volts as it's at 6000 volts.

MOV Failure Modes

For an investment of under $40.00, we should be happy to see our simple surge suppressor give up its life. Considering the alternative, $40.00 is a small price to pay. When we say give up its life, we mean the MOV has failed, and MOVs always fail “short.” This means that as they fail, they become a short circuit as the peak current capability of the device is exceeded. Current beyond this failure level will cause the leads of the device to melt or will cause the casing to explode.

This short-circuit characteristic is an important advantage as well as a major disadvantage. it's an advantage in that surge current will continue to be diverted away from the computer as the MOV fails. it's a disadvantage in that once this failure has taken place, most simple products have no way of warning the user that the protection is no longer available.

There are many MOV-based products plugged into sockets all over the country that have long ago given up what measure of protection they were designed to give, but have no way of telling their owners. Meanwhile, the user has a false sense of security and continues to use them.

How can we take care of this problem? A properly selected fuse, placed in series with the MOV, can trigger a status indicator to annunciate the failure of the MOV. Without this indicator, it's impossible to tell if the MOV is still operational without opening the case of the product. Beware, how ever—there are many products on the market that have some sort of status light which has nothing whatsoever to do with the health of the MOVs or any other internal suppression device. The additional circuitry needed to provide the user with this information increases the price of the product. Suffice it to say, most simple surge suppressors have no status indicator. Without this feature, the device is useless and a waste of money.

Noise-Mode Conversion

We have shown that the typical positioning of an MOV in a simple surge suppressor would be from the line to neutral. Let’s take a few moments to consider the implications of this.

As the MOV conducts current through the line-to-neutral noise path, a noise voltage drop occurs that is in parallel with the neutral and ground wires. ill. 8-4 shows this in simplified form. It almost appears as if the MOV has clipped off a large part of the impulse and has placed it on the neutral wire. Strictly speaking, this is not exactly what happens, but the net effect amounts to the same thing, due to the relationships between impedance, current, and voltage. The MOV is driving current through an impedance that is seen to drop from neutral to ground.

ill. 8.4. Normal- to common-mode conversion as created by a surge suppressor.

This conversion action has taken a normal-mode impulse and , by diverting it through the neutral-line impedance, created a common-mode impulse. The MOV has literally converted normal-mode noise into common-mode noise. ill. 8-5 shows a display of our test impulse appearing in the normal mode, while we see nothing in the common mode. ill. 8-6 shows what happens if we insert a surge suppressor into the line. The common-mode impulse that is created may be either smaller in peak voltage or larger, depending on the impedance of the lines and the current of the source.

We are now back to the controversy of grounding. If we are convinced that the PC is immune to common-mode noise, then we have no real concern with this scenario. If, on the other hand, we worry about the vulnerability of the chips inside the PC to common-mode offsets, then we might just rather risk the power supply rather than the logic circuits. More on this later. Suffice it to say that surge suppressors, by their designed action on normal-mode noise, will generate common-mode noise.

If we refer back to ill. 7-10, we see that an MOV in position 2 will also divert energy away from the normal mode. This can create even larger voltage drops because of the added impedance of the ground line, which is not a part of the noise path. it's almost as if the designers forgot that the neutral and ground wires come together at some point. However, if MOV 2 were combined with MOV 3 (in ill. 7-10), we have some very useful protection. We will talk about that later.

ill. 8-5. The test impulse appears in the normal mode without a surge suppressor.

ill. 8-6. The addition of a surge suppressor has created common-mode noise.

The Anatomy of a Spike

The simple surge suppressor we have been describing typically retails for under $20.00. By adding more components, we can do more things to eliminate the potentially harmful effects of high-frequency noise and high- energy impulses. Before we see what these additional components are, let’s take a closer look at the impulse itself.

We have mentioned frequency here and in other places. But where noise impulses are concerned, there are really two aspects of frequency. One, we are already familiar with. That is the rate of oscillation of a signal. The power waveform has a frequency of 60 Hz because it oscillates back and forth 60 times a second. A noise frequency of 10 MHz oscillates 10 million times a second. If we were concerned about a certain band of noise frequencies, we would design a filter to block those frequencies. We discussed this earlier.

There is another aspect of frequency, however, that we need to explore. it's the rate of change. In other words, how fast is a given thing changing at any given instant. A drag-race car has a varying rate of change in-acceleration. During the early stage of a race—the first few seconds—the rate of change is slow. Then, as traction with the road is established and the rpm’s of the engine build, the rate of change in acceleration dramatically increases. Seconds later, the dragster reaches its peak acceleration and the rate of change slows. ill. 8-7 shows what a graph of this action might look like.

ill. 8.7. What a graph of acceleration might look like during a drag race.

Bear in mind that the actual speed of the dragster is increasing through out the whole race. The rate of change is what we are getting at here. At some point along the curve, we see that the frequency of change is far different than the frequency of change at other points. ill. 8-8 shows what a high-energy impulse might look like. Notice that it has differing rates of change, or frequencies. This can be defined by plotting the instantaneous slope of the curve. The steeper the slope of the curve, the higher the frequency of change in voltage.

ill. 8-8. The waveform of an impulse has many instantaneous rates of change, or frequency components.

By now, you may be able to see where this discussion is heading. A transient voltage impulse contains many frequency components. Usually, it's the fast rise time or high-frequency components that can be the most troublesome. They are the parts that find strange paths of previously nonexistent inductance and capacitance inside computers. it's the high- frequency components of the impulses that do the damage. Obviously, it would be nice to change the shape of the impulse so as to reduce as much as possible its high-frequency nature.

This can be done two ways. One way is to divert high-frequency noise through capacitors. Another way is to slow down the impulse by forcing it through an inductance which will set up a bucking effect that is counter to the impulse current. Through the use of chokes or coils of wire, the shape of the wave can be rounded out, stripping away many high-frequency elements of the impulse—those elements that encounter high impedance at frequencies that match those of the choke.

Hybrid Circuits

We have discussed the relative strengths and weaknesses of both the MOV and zener diode. In general, MOVs are better able to handle high-magnitude, high- energy impulses while zener diodes have lower breakdown voltages. These characteristics can be used to advantage in a surge-suppression product to create a barrier to the propagation of damaging events. ill. 8-9 shows us a hybrid circuit that combines diverting elements with blocking elements. It also combines high-energy handling with low breakdown voltages.

ill. 8.9. A hybrid circuit combines the advantages of MOVs and zener diodes with a choke for more protection.

A MOV is the first stage of protection. It has a somewhat slower reaction time than the zener so some of the fast-rising initial slope of an impulse might pass through during this extremely short interval. However, the MOV is not going to allow much more than a peak of about 400 volts to pass through after its clamping action. Its large current-handling capability diverts most of the energy away from the sensitive load.

Next, the choke will present a nearly open-circuit reactance to those high-frequency components of the front slope of the impulse that make it past the MOV The impulse current generates an opposing current due to the current being induced by the choke’s own magnetic field. Much energy is blocked by the choke and is dissipated as heat.

The energy impulse that remains is rounder, slower, and lower in magnitude. Now that the impulse has been knocked down to size, the diodes can clean up what is left, as well as that which made it through before the MOV was activated. We should note that both MOV manufacturers and avalanche diode makers have various models that might suffice in our circuit. But our point, here, was to use the advantages of different technologies to accomplish some unique form of protection. and , in fact, this type of design is often used in commercial surge protectors.

However, a note of caution is now appropriate. Hybrid designs were discussed in my earlier guide, Computer Electrical Power Requirements, as they related to large systems. A hybrid design can be constructed by using a gas-discharge tube as the first stage and , then, using the building wiring as the inductive cushion or choke. But, as we have stated earlier, gas-discharge tubes function in such a way that their effects can be nearly as disruptive as the spike that might trigger them. This is especially true in the PC environment since events are of a smaller nature anyway.

There are tiny, little, gas tubes on the market that work just as well as the large ones. But, again, solid-state transient-voltage devices have far more desirable characteristics. Yes, there are gas tubes used in many consumer surge protectors, and the literature furnished with them gives the illusion that some protection against enormous events is available. The question is, “Do they represent the highest level in design sophistication?” The belief is that because of the device’s high current-handling capability, they are a good first stage of protection. Usually, they are placed in circuits along with MOVs and chokes. But because of their response time, other devices usually deal with the impulse long before the gas tube can operate. The response time of the three devices we have discussed are: zeners first, MOVs second, and gas tubes last. Gas tubes may very well take a relatively unimportant impulse and , by arcing to ground, create a damaging, fast, rise-time transient. This action should never take place near the computer’s grounding point.

Total Protection

By now, we have some idea about how different elements might be used to construct a product. We have discussed the transient suppression devices themselves. We have talked about the use of capacitance and inductance, both as they apply to filtering and as they are used in the hybrid-circuit concept. We have also talked about normal- to common-mode conversion. Before we see how the various elements come together to form some of the more sophisticated product types, we need to talk about one more thing— common-mode protection.

One of the points we have emphasized several times in this guide is the vulnerability of microcomputer-based products to noise associated with logic ground. As we have stated, this is one of those grey areas of controversy and folklore. In fact, if we all got on an airplane and flew to Europe, we would find a philosophy of grounding that might make no sense whatsoever. The laws of physics don’t change from one continent to another, but the practical application of electrical practices and terminology are so different that we might think so. Of course, after we had learned the European system, it would make perfect sense from a scientific point of view.

How does this discussion apply to surge suppressors? We have shown that the simple surge suppressor is a common-mode noise generator. We may feel warm and fuzzy knowing that a high-energy impulse has been diverted; it has been prevented from damaging our power supply. But, most people have no idea that electricity travels in circuits and that diverted energy can enter the logic circuits through stray capacitance or ground loops.

So what can be done? ill. 7-10 shows how all three MOVs might be applied to divert both normal- and common-mode noise. Remember, common-mode noise is a voltage that appears common to the line and neutral with respect to ground. In ill. 7-10, we see that MOV 1 will take care of diverting the normal- mode noise, a voltage measured from line to neutral. MOVs 2 and 3 are needed for common-mode protection. Either one will operate if the common-mode voltage appearing across them exceeds the threshold of the MOV

MOV 2 alone doesn’t help much. It’s a poor choice for normal-mode protection and it leaves the neutral line untouched for common-mode protection. MOV 3 is a good choice to use in combination with MOV 1 since it will divert energy from the neutral line that is routed there by MOV 1. In essence, MOV 3 bonds the neutral and ground lines together, providing common-mode protection, and thereby eliminating the impedance in the neutral-to-ground loop. MOV 2 can be added to further divert common-mode voltages. But, before this happens, the MOVs will let enough voltage appear across the PC’S circuits that processing errors or actual chip damage may occur.

In less expensive surge suppressors, we might find only MOV 1 or MOV 2— sometimes both. The solution for total protection is the use of all three MOVs. Now, no matter where the voltage appears, there is a device that can be a short circuit for excess energy.

The use of three MOVs for total protection is the exception not the rule. Most of the simple surge protectors not only generate common-mode noise when they operate, they offer no common-mode protection.

Complex Surge Circuits

Now we come to the real effective products, which have had some true engineering thought put into their design. These products often cost in excess of $100. ill. 8-10 shows an example of what the circuit of one might look like. This is just one example. There are any number of combinations that might be just as effective. What we are attempting to show here is the combination of all the elements we have talked about so far.

ill. 8.10. What the schematic of a complex surge-protector circuit might look like.

Let’s analyze this circuit as it might react to a very fast impulse as it appears across the surge suppressor. In terms of surge suppression, the zener diode would be the first of the devices activated. Its fast response time would begin conduction of the impulse as soon as the impulse’s voltage reached the zener’s breakdown voltage. Next, the MOV would be activated, and the bulk of the current would flow through these components.

Meanwhile, inductance L1 and capacitance C1 are stripping away the high-frequency components of the impulse’s leading edge. Inductance L2 is a choke that blocks the flow of noise toward the load. In practice, we would probably not see both Li and L2 in the circuit. The filter section might be constructed of just one choke on the line. Also, MOVs 2 and 3, along with capacitors C2 and C3 and coil L3, have been added for complete common- mode protection.

Two Additional Points

For the sake of simplicity, we have left a few things out of the circuit shown in ill. 8-10. These missing items don't affect the operation of the circuit, but they are essential. Fuses or circuit breakers and status indicators must be used for complete product utility. If the MOVs have shorted out, or somehow given up their life, we must have some indication that the unit needs attention. It does no good to spend $100 on a product that can’t tell you when it has done its job and must be repaired or replaced. After all, it’s the computer we are worried about. The surge-suppressor manufacturer must design his product with this worry in mind also.

Another fact that must be highlighted before we move on is that not all surge protectors come packaged as multi-outlet devices. Some come packaged with wire pigtails or lugs so they may be inserted at the building service or mounted near a breaker panel, somewhere upstream from the computer.

This is a very good idea from a couple of standpoints. First, it takes advantage of the inductive cushioning of the wiring between the panel and the computer. At high frequencies, this presents an impedance that is roughly equivalent to the L2 and L3 inductance of our circuit given in ill. 8-10. Also, if the surge device is lacking in common-mode protection, placing it nearer the neutral-to-ground bonding point presents a relatively lower impedance path toward the nearest power source than what might be presented by a path through stray capacitance or a ground loop in the computer system.

Of course, most of us don't have the skill or bravery to go mucking about in the breaker panel. So the product that plugs into a wall outlet is the typical purchase. it's good to know that these other devices are available, both since their application in front of a panel will protect every branch circuit and for the reasons stated above.

But surge protectors have limitations no matter how fancy they are. Product life, ruggedness, noise bandwidths with the inability to attenuate transients at all points on the sine wave, and common-mode noise conversion are among those limitations. As we will see in the next section, there are products that overcome these limitations entirely. But the single most important difficulty that a buyer must face when selecting a product is to see through all the hype and fancy packaging so as to compare designs, power ratings in joules, and see the protection per dollar that is being delivered.

NEXT: Power-Line Conditioners

PREV: Transient Suppression Devices

All related articles   Top of Page   Home