Many types of surge protective devices (SPDs) and technologies are available on the market. To get the most effective protection at the best value, you need to make a selection based on the most important technical performance specifications. While indication, physical and environmental issues, and standards compliance are important, the following performance characteristics are the most critical:

  • Maximum continuous operating voltage (MCOV)

  • Voltage protection level

  • Surge rating

  • Protection modes

Let's take a detailed look at each characteristic.

MCOV. This specification relates to the maximum steady state voltage the SPD can withstand without becoming a fire or safety hazard.

Traditionally, SPDs couldn't differentiate between slower overvoltages and the faster transient voltages. NEMA, in its standard LS1-1992, “Low Voltage Surge Protection Devices,” selected an MCOV of 20% above the normal voltage to cover expected normal utility regulation limits.

In many cases, however, the nominal voltage has exceeded this 20% threshold, causing the SPD to clamp on each half cycle and build up sufficient heat to become a hazard. Overvoltages can be caused by poor power regulation and wiring faults like disconnected neutrals in unbalanced 3-phase wye systems. SPDs can also be subjected to overvoltages by incorrect installation.

In 1998, Underwriters Laboratory (UL) issued Edition 2 of UL 1449 with new requirements to address this growing problem. It specifies that when an overvoltage of 110% of nominal voltage is applied, the device must remain functional and safe. Second, when an abnormal overvoltage of 125% is applied, the device is allowed to permanently stop functioning, but must not become unsafe. Finally, when a full-phase test voltage as listed in Table 1 is applied, the device is allowed to permanently stop functioning, but must not become a fire or safety hazard.

Although this full-phase voltage test is an extreme overvoltage test, for safety reasons, you should select only UL 1449 Edition 2-recognized products because wiring faults and accidents can occur.

For sites where poor regulation is a possibility, you should select a technology that's not only UL 1499 Edition 2 recognized, but also doesn't permanently disconnect during the full-phase voltage test. This performance capability avoids the cost and trouble of having to replace the SPD every time the site voltage exceeds 25% of nominal. Modern technologies are available that offer high MCOV withstand without the traditional disadvantage of increased voltage protection levels.

Voltage protection level. An SPD is responsible for clamping the transient voltage to a safe level so it won't affect the protected equipment. But no SPD device can completely remove all impulses. Some small amount of transient will still reach the equipment the device is intended to protect. This is acceptable provided the let-through voltage is low enough for the protected equipment to withstand.

The voltage protection level specifies the SPD's ability to limit a given impulse magnitude and waveshape. The larger the current, the worse the protection level. Commonly, 500A 8/20µsec (UL 1449 SVR rating), ANSI/IEEE B3 3kA 8/20µsec, and ANSI/IEEE C3 10kA 8/20µsec results are given. For a well-constructed 150V MOV-based device with a clamping voltage of 400V at 3kA 8/20µsec, this may increase to 600V at 20kA 8/20µsec.

When comparing data, you should ensure that the same waveshape, impulse current magnitude, and connection method are being compared. Normally, results are measured with 6 inches of wiring connection to the SPD. Lower results will be obtained if measurements are taken at the device terminals.

Surge rating. The third critical performance characteristic of an SPD is surge rating, which is typically defined by a short-duration, high-current impulse defined by an 8µsec rise time and a 20µsec decay time. The selection of a suitable surge rating for the intended application is key to ensuring longer service life of the product. Guidance on selecting the correct surge rating for each application is well defined in ANSI/IEEE C62.41.2. Only 1% of direct lightning strikes exceed 130kA 8/20µsec, making surge ratings above 100kA 8/20µsec excessive in most cases.

Manufacturers around the world use different methods for detailing the surge rating of their products. For example, a manufacturer could claim the same product as an 80kA-per-line SPD (connected to each line is two 40kA devices), a 3-mode 40kA (per mode) SPD, or a 120kA SPD (simply the total of all the supplied surge material.) So for exactly the same surge material, three claimed ratings are possible (Fig. 1).

It's critical that when you compare surge ratings for devices with multiple modes, you fully understand how each manufacturer has arrived at its claimed ratings to allow true comparison. Surge ratings are easily tested within numerous certified laboratories. It's a good idea to request the manufacturer to supply test reports that prove claimed surge ratings are met.

Modes of protection. Most suppliers offer line-to-neutral (L-N), line-to-ground (L-G), and neutral-to-ground (N-G) protection within their SPDs. And some now offer line-to-line (L-L) protection. The argument is that because you don't know where the transient will occur, having all modes protected will ensure no damage occurs. However, equipment is more sensitive to transients in some modes than others. L-N and N-G mode protection is an acceptable minimum, while L-G modes can actually make the SPD more susceptible to overvoltage failure. In multiple line power systems, L-N connected SPD modes also provide protection against L-L transients. Hence, a more reliable, less complex “reduced mode” SPD protects all modes.

Don't be misled. Some parameters, while somewhat important, can be very misleading. Be careful with performance characteristics like speed of response, energy (joule) rating, and technology.

Speed of response. Although it's sometimes identified as a performance measure, speed of response is actually a misleading specification that you can ignore. Instead, focus on the voltage protection level because this includes “speed of response” and other more important criteria. Total SPD performance under real world conditions, not solely the internal component's speed, is what's important.

Response times for shunt diverters less than 1 nanosecond (1 billionth of a second) or 5 nanoseconds aren't uncommon. Generally, silicon devices are quoted at 1 nanosecond to 5 nanoseconds. MOVs are generally quoted at 5 nanoseconds to 25 nanoseconds, while spark gaps are quoted at 100 nanoseconds.

Theoretically, speed of response will make a slight difference (10 nanoseconds equates to about 43V to 48V) if all other specifications, such as technology, layout, and construction details, are exactly identical. However, this is never the case, and as shown by the example in Fig. 2, other effects are much more dominant.

Fig. 2 is a simplified example because it assumes the devices start to turn on at 400V and will be fully on by the end of the speed of response time. It also only includes the effects of inductance, not resistance. However, it clearly shows that speed of response is included in the let-through voltage, but is a secondary effect with regard to internal construction. It's also less important than technology when choosing an SPD.

The calculations in Fig. 2 show that as the speed of the pulse increases (i.e. “dt” becomes smaller), the clamping voltage across the internal leads will increase proportionally. For faster pulses, the clamping voltage due just to the inductance of the internal wiring will be dramatically larger, swamping any difference between the fast and slow internal components.

The British Standard BS6651:1992 supports this by stating “…let-through voltage takes into account the response time of the device, i.e. a slow response time will result in a higher voltage… the response time of a parallel connected protector will often be overshadowed by the inductive voltage drops on the connecting leads.”

Energy (joule) rating. Another misleading specification is device energy ratings. Energy capability is sometimes specified in joules and perhaps as energy diverted and/or absorbed, but unless detailed information is provided, it's difficult to compare just the joule rating. A larger joule rating isn't always better. Instead, you should focus on the surge rating (in kA) and the resultant voltage protection level, using the following two equations:

Power = voltage×current

(Pwatts = V× I)

Energy=voltage×current×time

(Ejoules 5V×I×t)

So if two devices are offered that have the same joule rating, are they equal in performance? Yes, but only if the manufacturers use the same pulse waveshape (time) and currents. Basically, you need more information to truly compare the performance characteristic.

Look at the simplified example in Fig. 3. Based on the supplied data, you would expect that SPD B would be twice as good as SPD A. But with more test data (as per Test 1), SPD A proves to be the better device, even though it has half the joule rating. This is because both devices were tested with the same 1,000A 2msec pulse and SPD A managed to clamp this to half the let-through of SPD B. Since SPD A has clamped at a lower voltage, its energy dissipation rating is actually less.

Alternatively, suppose you had test data per Test 2, in which the current is altered so both devices have the same let-through voltage result. This would make SPD B the better device because it takes twice the surge current to produce the same let-through voltage as SPD A. Disregard energy ratings and just consider surge rating and voltage protection levels.

Technology. The type of technology an SPD uses can also be a misleading parameter for simple evaluation. Most international standards treat an SPD device as a “black box.” That is, the standard doesn't care what type of technology the manufacturer uses within its SPD. Instead, the performance and suitability of the tested SPD is defined only by its ability to remain safe and to protect the downstream equipment in a given number of defined input conditions.

Fig. 4 provides some general comparisons between three different technologies. Table 2 lists comparative performance of the three SPD technologies for the same amount (size) of protection. Admittedly, this is a crude approximation, giving the values as a set of multiplying factors, but it does provide a comparison basis. For example, silicon devices will typically have a 20% better voltage protection level than MOVs, but can be as much as five times the price and have 12.5% the surge rating.

Silicon avalanche diode (SAD) SPDs include devices like transzorbs and zeners, and are typically characterized by low clamping voltage, low surge rating, high speed, long life, and high cost. With this type of SPD, you should specifically ask for NEMA LS1 (tested) maximum 8/20µsec surge rating of each individual protection stage and each individual mode. Don't accept just a 10/1,000µsec or joule rating. Also, ask for the cost of replacement of each protection mode.

Although MOVs are generally well accepted in the industry, they may not last long if not sufficiently rated. With this type of device, you should ask for the NEMA LS1 (tested) maximum 8/20µsec surge rating of each individual mode and specify life cycle testing compliance to ANSI/IEEE C62.45 (1,000 impulses).

Spark gaps are ventilated air gaps (not gas arresters) that contain a low-pressure inert gas to lower the firing voltage. With this type device, you should pay close attention to voltage protection levels, as these can be in the order of 3,000V to 4,000V. Specifically ask for details on follow-On currents and life at the expected short-circuit rating of the supply.

Be careful when investing in SPDs. The more feature-rich they get, the more you'll need to do your research before purchase to make sure that you need everything you're paying for and you're getting everything you need.

Beech is senior applications engineer at the Critech Division of ERICO in Solon, Ohio.