EFT and Surge

by William D. Kimmel, P.E.

and Daryl D. Gerke, P.E.

Kimmel Gerke Associates, Ltd.

Electrical Fast Transient (EFT) and Surge are two companion pieces, sitting side-by-side in the IEC and EU requirements list. So why the difference? And how do you handle the differences? Both requirements are aimed primarily on AC power input to your equipment, even though EFT also covers signal cables.

Let's do a comparison of the two: what the sources are, how the transients differ, and how the treatment differs.

Nature of the Two Transients

Simply put, surge is a high energy, low frequency transient and EFT is a low energy high frequency transient. Even though the test amplitudes are approximately the same, the causes and effects are significantly different.

Lightning is the main reason for the surge requirement, but cycling heavy equipment and facility power faults will also cause surge transients. A bolt of lightning striking a nearby power pole will cause a power surge (or dropout). A direct hit to your facility will bounce ground, effectively resulting in a voltage between neutral/phase and ground.

The result is several thousand volts on your input power, with a risetime of a little over one microsecond, and a pulse duration of 20 to 50 microseconds, depending on the load impedance.

Using the conventional Fourier conversion from risetime to frequency content

F = 1/ pi*tr

we find a surge with risetime of one microsecond produces significant equivalent frequencies of up to about 300kHz. This is not a particularly high frequency, but your protection needs to be able to absorb a substantial amount of energy.

An electrical fast transient (EFT) arises from a much different situation, basically that of interrupting current in an inductive load. As your switch opens, the back voltage across the inductor rises until an arc is established, at which time the current resumes until the voltage collapses and the arc goes out. At that time the voltage rises again, re-igniting the arc. This cycle continues until the contacts finally open far enough that the diminishing energy can no longer sustain the arc. This "burst of transients is sometimes called a "showering arc." If you have ever been listening to an AM radio when turning off an inductive load, a room fan or an electric drill, you heard a burst of static, not just a single click -- that's the EFT.

The EFT is a much faster phenomena than the surge, having a risetime of 5nS and a pulse width of 50nS. This will produce equivalent frequencies of up to 60MHz, more than two orders of magnitude faster than the surge. Consequently, your protection needs to be fast, but not particularly high energy capacity. Note that because of the faster transient, EFT has limited range -- series inductance prevents the transient traveling too far. You won't expect to see an EFT unless you are looking at the same power receptacle as the interference source.

A side note before moving on, power quality monitors usually have a bandwidth of no more than 2MHz, which is too slow to catch an EFT. If you suspect an EFT, you need to use a high frequency oscilloscope to catch the transient.

What Are the Fixes?

Basically, there are two methods of suppressing the overvoltage, with transient suppressors and filtering. Transient suppression works with threshold limiting, whereas filtering uses frequency discrimination.

Transient Suppressors

There are three primary transient voltage suppression devices (TVSS): the crowbar (gas discharge tube) and two clamps, the Zener diode and the MOV.

The gas discharge device is the most robust -- it is basically nonconductive until the voltage reaches a threshold, at which time the device breaks down and drops into a low impedance state. The device can then sink copious amounts of current without overheating. This is basically the same concept as primary lightning control, where the arcing path is open air.

The major limitation with gas discharge devices is the speed -- it is too slow to suppress an EFT. A secondary issue is that once the device has fired, it will remain on until the source voltage is removed. This poses no problem for AC, but the device cannot be used to protect DC voltages unless supplementary electronics is added to switch the device off after the transient has passed.

These devices are mostly used for power input to a building, often installed as facility protection, rather than device protection.

The next device is the Zener diode clamp. As is well known, Zener diodes break down when the reverse voltage exceeds a design voltage level, clamping the voltage to that level. The breakdown is very abrupt, then the breakdown voltage is reached. The voltage will not rise much above that level. Zener diodes used for transient voltage suppression (well known by the trade name Tranzorb) are built with a large junction cross section, to accommodate the high surge current.

These devices are very fast, able to shunt very fast transients, including EFT (and ESD, as well) provided the lead inductance to the device is kept to an absolute minimum. Since the Zener diode clamps to the breakdown voltage, it burns up lots of energy -- energy capacity is the principal limitation and, if exceeded, will burn up. Hence, the device is most useful for EFT and ESD but are at risk for lightning surges.

Zeners are naturally polar in nature, well suited for DC. For AC, two diodes are placed head to foot, available in a single package.

Capacity can be increased by combining a Zener with an SCR. When the Zener threshold is reached, the SCR fires and drops the voltage down to saturation low level, greatly reducing the energy burned in the device. However, this device will not shut off by itself, much like the arc device discussed above.

The second clamping device is the MOV, or metal-oxide-varistor -- basically a highly nonlinear resistor. It is symmetrical, so it works well for AC as well as DC. In operation, it functions much like two Zener diodes placed head to foot (see Figure 1). The major difference is that the breakdown has a much softer knee than the Zener combination, so its voltage breakdown limit is not as precise.

The traditional MOV is very effective for surge in the power line -- it is the most common device employed in surge protectors for computers. It is, however, too slow for EFT. Early units exhibited a problem with breakdown voltage drifting with age -- it turned out that the problem was essentially eliminated if you selected a breakdown voltage high enough to ensure it didn't conduct (and overheat) with nominal AC supply voltage.

More recently, the multilayer varistor has proven to be much faster and, if mounted correctly, is fast enough for EFT and even ESD.

For a given component size, the MOV can take more heat than the Zener diode. The soft breakdown curve, however, limits use to cases where precise breakdown is not needed.

Filters

Filters can also be used to suppress transients. Two conditions are of import -- the devices must be able to withstand the transient voltage and current, and the transient bandwidth must be significantly higher than the signal bandwidth.

For filtering power, bandwidth is not an issue so, in principal, you can use enough filtering to control both surge and EFT. Audio frequency signal lines can usually be filtered adequately as well. Higher speed signal lines are another story, however. Surge, having a bandwidth of 300kHz, requires a filter cut-off of 30kHz or less in order to provide any kind of protection. EFT, having a bandwidth of 60MHz, would need a cutoff of 6MHz or less -- acceptable for many RS 232 lines, but too low for many digital data streams.

If you are using shunt capacitors as the primary filter element, breakdown voltage may not be an issue. You need to make sure you have adequate capacitance to shunt the current while still keeping the voltage below rated level. Basically, this means you need a fairly big capacitor to handle surge.

If you are using a large series inductor directly in the line, your design must assume that the entire transient voltage will be dropped across the inductor.

With these limitations noted, we prefer to use filters in equipment wherever possible.

Summary

When dealing with transient voltages, you can use transient suppression devices or filters. Make sure the device is fast enough to shunt the current, has adequate energy and voltage rating -- and don't kill the signal.