Voltage transient generator circuit and/or device

[phaset]

Greetings all,

I'm wanting to do durability testing on prospective power supply replacements from our current vendor. Because of the number of failures we've seen in the field, names and application details are witheld to protect the guilty

These supplies operate on a single leg of 480 leg-leg of 3 phase delta circuits, or at 277 on a 3phase wye configuration. In situ measurements of on-site power at various industrial client locations have shown V_peak of 1000V+ for durations of 1-5 milliseconds, plus some secondary ringing. The tend to coincide with contactor-based events.

I'm hoping to build/buy a widget to be able to whack potential PSU suitors with transients (either with or without reactors/snubbers) and see how tough they really are. Before I go try to mock something up, I thought I reach out to the community to see what you recommend, as I assume that creating these kinds of spikes in a predictable fashion is a solved problem in certain circles.

I don't need a fancy solution, just something that will help quickly thin the herd as we investigate other vendors.

Thanks as always!

-Phil

P.S. These supplies are implemented in the field with line reactor/snubber pairs, and we've still been through multiple revisions on the existing manufacturers product to knock down failures.

 

[Jraef]

Greetings all,

I'm wanting to do durability testing on prospective power supply replacements from our current vendor. Because of the number of failures we've seen in the field, names and application details are witheld to protect the guilty

These supplies operate on a single leg of 480 leg-leg of 3 phase delta circuits, or at 277 on a 3phase wye configuration. In situ measurements of on-site power at various industrial client locations have shown V_peak of 1000V+ for durations of 1-5 milliseconds, plus some secondary ringing. The tend to coincide with contactor-based events.

I'm hoping to build/buy a widget to be able to whack potential PSU suitors with transients (either with or without reactors/snubbers) and see how tough they really are. Before I go try to mock something up, I thought I reach out to the community to see what you recommend, as I assume that creating these kinds of spikes in a predictable fashion is a solved problem in certain circles.

I don't need a fancy solution, just something that will help quickly thin the herd as we investigate other vendors.

Thanks as always!

-Phil

P.S. These supplies are implemented in the field with line reactor/snubber pairs, and we've still been through multiple revisions on the existing manufacturers product to knock down failures.

Hi-Pot tester?

HIPotronics is kind of the standard for the industry.

http://www.hipotronics.com/products-hipot-testers.htm

 

[phaset]

Hi-Pot tester?

HIPotronics is kind of the standard for the industry.

http://www.hipotronics.com/products-hipot-testers.htm

These will certainly get the job done, but are a bit fancy. At the end of the product development cycle the NRTL will do their own battery on the final supply. I just need to (try to) blow up some test units from the various manufacturers trying to tell me that their version is "the best."

Before I cobble together my own boost circuit, I'd see if someone has done a simpler version than the HiPot.

 

[ATSman]

These will certainly get the job done, but are a bit fancy. At the end of the product development cycle the NRTL will do their own battery on the final supply. I just need to (try to) blow up some test units from the various manufacturers trying to tell me that their version is "the best."

Before I cobble together my own boost circuit, I'd see if someone has done a simpler version than the HiPot.

Check out the Biddle line of test equip.
I think they were bought out by AVO but not sure. They used to make a 5KV megger which may be more in your price range.
eBay also sells used test equip.

Tony

 

[ELA]

This may be too fancy for you but it is the correct type of equipment for what you want to accomplish:

http://www.emtest.com/products/product/135120100000006575.php

We use a Ring Wave tester to test equipment against UL864 and other EMC compliance standards.

 

[__dan]

Greetings all,

I'm wanting to do durability testing on prospective power supply replacements from our current vendor. Because of the number of failures we've seen in the field, names and application details are witheld to protect the guilty

These supplies operate on a single leg of 480 leg-leg of 3 phase delta circuits, or at 277 on a 3phase wye configuration. In situ measurements of on-site power at various industrial client locations have shown V_peak of 1000V+ for durations of 1-5 milliseconds, plus some secondary ringing. The tend to coincide with contactor-based events.

I'm hoping to build/buy a widget to be able to whack potential PSU suitors with transients (either with or without reactors/snubbers) and see how tough they really are. Before I go try to mock something up, I thought I reach out to the community to see what you recommend, as I assume that creating these kinds of spikes in a predictable fashion is a solved problem in certain circles.

I don't need a fancy solution, just something that will help quickly thin the herd as we investigate other vendors.

Thanks as always!

-Phil

P.S. These supplies are implemented in the field with line reactor/snubber pairs, and we've still been through multiple revisions on the existing manufacturers product to knock down failures.

The nature of the application would help understanding what is happening. Is the load high or low kVA, and does the load run clean or make a lot of hf noise of its own.

What you described are industrial plant sized switching transients, a half second or so of high voltage, high frequency ringing, and you must have some data of what is failing, the PS input section or the PS output.

Guessing, but if it's a small PS getting hit with industrial switching transients and it has its own internal TVSS transient suppressors, MOVs, the transients could be too big for it and the MOVs could fail destructively, and at 480 V LL, flash over. A lot depends on what part of the PS is failing, and if it's the internal TVSS at the input. Also, the site would be guilty of feeding the PS voltages outside its rating of 480 V, 1 ph, 60 Hz.

Some of your choices would be a low pass filter in front of the PS, maybe an isolating transformer, TVSS at the source gear big enough for that gear, and refeeding from a source without heavy switching transient on its bus.

As far as doing your own destruction testing, sounds like you're talking about charging a cap bank and discharging it through the PS as a short, not that I would recommend that. I would recommend determining the cause of the failures (switching transients at the source bus?), determining what part of the PS is susceptible to that (the internal TVSS MOVs?) and working forward from there.