One of the core immunity tests that electrical & electronic equipment is subjected to is input surge testing. The test method for this is IEC61000-4-5 and the limits are defined by end system requirements.
The basics of the test are to subject the system to voltage spikes on top of the nominal voltage input to the system. These spikes are representative of voltage fluctuations that occur from causes such as large motor drives, nearby lightning strikes, etc. These high voltage deviations can cause a variety of issues when applied to a system not designed to handle them. Testing ensures that the end equipment will give the level of performance needed for the intended application.
In some applications it may be acceptable if a system shuts down and a user needs to manually reset the device after a surge event. In other, more mission critical systems, that cannot be allowed. The system needs to continue operating without any errors during the entire event. The response of the system to the application of the surge is determined using performance criteria as outline below.
The test levels applied during the test are defined by the Installation Class that is called out for the end system. Most commercially available power supplies independently tested to Installation Class 3 which calls for 2kV surges between Line/Neutral and Ground (common mode) and 1kV surges between Line and Neutral (differential mode).
The performance criteria defines the level of performance an end system needs to have in response to the application of the surge. They are graded on a simple A, B, and C rating. Performance Criteria A means that there is no change in the system because of the test. Performance Criteria B means that the system does change during the test, but auto-recovers after the surge event. And finally, Performance Criteria C means that the user must intervene with the system after the event in some manner. This may mean anything from clearing an error code to restarting the system. If the system is damaged from the surge this would be considered a failure.
For an end equipment it's fairly easy to determine whether the performance criteria is A, B, or C. For a power supply tested in isolation it is a little less straightforward to determine essential performance. IEC61000-4-5 details how the voltage spike is generated, when it occurs, how often it occurs and the duration between spikes. But determination of the performance criteria (A, B or C) is left to the device or equipment manufacturer. The power supply industry in conjunction with the accredited test laboratories has long relied on the practice of monitoring the output with a standard moving coil or digital voltmeter, watching the output to see if it deviates during and after the test. This practice has been commonplace since the start of the standard and for the clear majority of situations it is adequate to determine essential performance showing that the power supply has continued to operate without interruption to the DC output. Occasionally, issues arise where the end equipment is sensitive to momentary voltage deviations, or ground disturbances. Due to the input to output capacitance of the power supply such momentary disturbances may exist which would not be detected with a standard voltmeter. To see detect these disturbances requires the use of an oscilloscope, though this is challenging in its own right as the surge voltages are high and with enough energy to create radiated disturbances and ground disturbances that are detected at the oscilloscope. A great deal of care with the connection of the oscilloscope probe and the system & measurement grounding is required to attain an accurate result with less than adequate measurement set ups leading to incorrect conclusions regarding the performance of the supply.
My system has surge failures. What do I do?
The first thing to check and verify is that the surge test is being run correctly and that the test interval between surges is performed at 60 seconds. The complete test is relatively long so the standard does allow you to speed up the test by reducing the interval between surge applications. However, in the event of a system failing the test at reduced intervals, the 60s interval should be applied allowing sufficient time for discharge between surges.
The surge test stresses certain components within the supply which are degraded by repeatedly testing the same individual unit. This applies to MOVs for example. Testing a system over and over again with the same power supply may eventually cause the performance to degrade.
If the end equipment is affected by momentary disturbances on the DC supply or ground plane, a capacitor connected across the DC supply close to the load connection point often resolves the issue, very effectively presenting a low impedance at the critical connection point. This can reduce the magnitude of any disturbances seen on the voltage inside the system.
If the system has an earth ground connection a ferrite on the earth ground wire with 2-3 turns through it as close to the AC input in the system can help. This will reduce the magnitude of the surge voltage thereby reducing the stress to the supply. We have seen good results using this approach in sensitive applications.
Finally, a common source of issues is the power cable routing within the system. It is best to keep AC input wiring and DC cabling well apart and away from sensitive low voltage circuitry.