High Voltage isolation testing question

[Circuitous]

I’m looking for an approach to test HV isolation on a board.  The board will control and monitor the navigation lights on a 400 foot radio tower, one that seems to get a lot of lightning activity... not necessarily direct strikes coming down the power lines, but surges that are big enough to trip a 20Amp circuit breaker that feeds the lights.

I have seen posts from Mike and others about needing to test with 8Kv surges.  I can get a 5Kv insulation tester, but I’m not sure that’s the right tool.  Any recommendations on generating the appropriate spikes? 

[jeroen74]

You need to look for a surge tester that generates pulses according to the IEC 64000-4-5 standard. These can generate 2 to 8kV 8/20u pulses. I've done these tests in the past, mainly to test varistor endurance. Results are spectacular if a MOV fails

This is specialized equipment and you need a shielded box for safety as things can literally explode. Dangerous HV stuff so you need to be careful.

edit:
An insulation tester applies a voltage between the housing and all of the connections for one minute. The current is measured during this time and when it becomes higher than a certain level the product fails the test.

[G7PSK]

Megger make insulation testers that go to 10KV I expect that such things can be hired as well as purchased. I am sure that there are other makes as well.

http://www.megger.com/uk/products/ProductDetailsBySubGroup.php?PSGC=P107&BS=P

[jeroen74]

I believe the OP does not need an insulation tester, but a surge tester. You can have perfect insulation but still fail an easy 1kV surge.

For fun I once tested one sample without the varistor mounted. Most of components exploded at the first surge Completely hopeless without the varistor. It's simply amazing really how much abuse these MOVs can take, I've seen them handle peaks of 1 megawatt.

[Circuitous]

Yes,  based on reading and other posts, I need to use 8Kv 8/20 pulses.  My impression is that commercial products to generate the 8KV pulses can be quite expensive, even to rent.  I saw an example circuit to generate the 8/20 pulse in a write-up by Littelfuse http://www.littelfuse.com/data/en/Application_Notes/an9773.pdf

But, it leaves out charging the capacitor to 8+ Kv.  So, I'm looking for suggestions for circuits to charge the capacitor to 10 KV or thereabouts.  Then I can run that through the example circuit from Littelfuse.
I saw some HV examples using microwave oven transformers (MOT), but haven't pursued that yet.

Any suggestions for charging the capacitor and sources for the capacitor?

[jeroen74]

I wouldn't want to build this high energy/voltage stuff myself I used two different tester back then at work and both were serious pieces of gear. Remember that the circuits in that appnote are just simplifications. An actual tester is much more complicated.

Isn't there a test house that can do the tests for you? It's expensive I guess, but a necessary part of the development. And at these levels, a simple varistor probably is not going to cut it. At the place I worked they made electronic drivers for HID lamps and such, and the surge and lighting protection for one product range was quite a substantial circuit, multiple VDRs and GDTs.

[G7PSK]

Why charge the capacitor to such a high voltage if you used an ignition coil or an exciter coil from a flash tube you would attain the EHT without actually having to handle it other than the spark gap or whatever is involved.

[Circuitous]

The reason for the capacitor and related discharge circuit is to get the 8/20 pulse and proper current.  But, I am thinking of skipping that initially and just using a simple flash to get the surge.  If it passes that I may try to do more accurate testing later.

One approach may be to just deploy the circuit in the field and see how it holds up.  I'm sure the tower will get hit (or a near miss) when the Spring thunderstorms arrive.