High voltage relay for ESD simulator

[fcb]

I can't justify buying an NSG 435 https://www.teseq.com/products/NSG-435.php (Schaffner/Teseq), and I haven't seen any on eBay. So I'm building one.

The high voltage generator, and 'discharge network' are within my capability. But I'm stuck on the 'relay' that connects the output of the discharge network to the tip.

So with an ESD simulator there are two main types of event you simulate:
1. air discharge
2. contact discharge

With the air discharge you (rapidly!) approach the DUT target area with the tip of the simulator at it's charged potential  (upto +/- 15kV) and the idea is that discharge will find the nearest weak spot and discharge.

With the contact discharge you put the tip of the simulator on the DUT target area (e.g. the metal back-shell of a connector) and pull the trigger. This closes the relay and injects your ESD event into the DUT.

I can see how I could use a high-voltage relay in the first case, as current wouldn't (or at least minimal current) be flowing at this point.  But in the second case, I can only think that a spark would form across the relay contacts as the relay closes.

Anyone know how this is done?

 

[gasmeter]

Hi

I'm looking at it too.

Did you come up with a solution ?

I think there are two ways I can think of 

1/ is to use two relays like in this video https://youtu.be/RzUaV8E8mPA (my preferred method)

2/ The gun must be removed between each contact test.
Then it can simply be the capacitor that is charged and the relay is open before a test discharge ? Single relay solution

Not a good video but that's he appears to be doing here https://youtu.be/IrVvRWvVU5k


Peter

[mikeselectricstuff]

I can't justify buying an NSG 435 https://www.teseq.com/products/NSG-435.php (Schaffner/Teseq), and I haven't seen any on eBay. So I'm building one.

The high voltage generator, and 'discharge network' are within my capability. But I'm stuck on the 'relay' that connects the output of the discharge network to the tip.

So with an ESD simulator there are two main types of event you simulate:
1. air discharge
2. contact discharge

With the air discharge you (rapidly!) approach the DUT target area with the tip of the simulator at it's charged potential  (upto +/- 15kV) and the idea is that discharge will find the nearest weak spot and discharge.

With the contact discharge you put the tip of the simulator on the DUT target area (e.g. the metal back-shell of a connector) and pull the trigger. This closes the relay and injects your ESD event into the DUT.

I can see how I could use a high-voltage relay in the first case, as current wouldn't (or at least minimal current) be flowing at this point.  But in the second case, I can only think that a spark would form across the relay contacts as the relay closes.

Anyone know how this is done?

vacuum or SF6-filled relay

[NiHaoMike]

Maybe use a thyratron or other switching tube?

[Le_Bassiste]

you might try to stack MOSFETs in series up to the desired voltage (say, 15 x 1kV VDS rating for a 10kV switch), SMD mounted ones probably do the trick.
sounds silly? have a look at http://www.behlke.com/. that's exactly what they do. i had the pleasure to use those switches in an EMC Burst-tester design. they work like a charm and make for extremely good waveforms compared to the requirements layed out in the standards, albeit frikking expensive.

edit:
and there's of course also this kind of stuff:
http://www.stg-germany.de/products/high-voltage-reed-relays

[Gyro]

I remember opening up a contact discharge probe and finding a kilovac vacuum relay inside.

You could try one of these Pickering 12.5kV reed relays of ebay. I'm not sure if it uses vacuum, SF6, or other high pressure filling to meet the voltage rating, but as you're in the UK, it certainly looks cheap enough to give it a try....  https://www.ebay.co.uk/itm/Pickering-Electronics-Reed-Relay-High-Voltage-60-1-A-24-3-Free-P-P-50W-12-5kV/303099432854


EDIT: Datasheet:  https://www.pickeringrelay.com/product/60-1-a-24-1/

[Neilm]

The big problem with rolling your own tester is that it will probably not conform to the standard. The waveform rise and fall times are specified. You may find devices you test with your equipment may pass, but if they are tested with a complient system they fail. I am assuming that you have the ground planes required by the standard for your testing.

[T3sl4co1l]

Pretty sure they're relay.

Most ESD guns I've seen/used make an obvious relay click when cycling.

EFT isn't really practical to do mechanically, even with a fast bouncing contact (i.e., to implement the older 1kHz PRF).  Pretty sure they're using hydrogen thyratrons, or the newer ones perhaps FET stacks.

Don't expect it to be cheap.  You're asking for test equipment, after all!

If you want it cheap, just do air discharge.  Or do it completely uncal with a piezo generator.

Consistency isn't so soooo important for pre-cert.  The main advantage is being able to identify problems that occur infrequently at a certain threshold, and which are masked by problems at higher levels.  For example, say you find a comm interface that sometimes drops out at 8kV (contact), but at 10 or 12kV you knock out the power supply instead: separate problem, and you can't simply turn up the level to better analyze the one failure mode.

Tim

[MagicSmoker]

Here's a wacky idea: aim a UV LED (preferably the shortest wavelength you can find/afford) at a spark gap that is just barely wide enough to hold off the expected test voltage. When the LED shines on the gap it will ionize enough molecules to initiate discharge.

I know this works with low-pressure mercury lamps (basically fluorescent bulbs without a phosphor coating), but those emit almost entirely at 254nm so not sure if it will work here. Also, you have to mechanically adjust the air gap for different supply voltages and repeated operation will lead to premature firing.

[cat87]

Adding to the pool of already nifty things to use,   you can get an older  type of defibrilator (early 90s,  late 80s) because those have a relay,   usually encased in silicon gel, that is made to take a lot of energy and do that safely. I actually have one of those and it's a nice bit of kit....huge,  but nice

Later edit:
Added some pics of the one I have

[gasmeter]

I solved my problem for the time being .

I have a friend who can borrow a NSG435 test gun.
Prior to that I  plan to try with a electronic spark gun for lighting gas cookers etc as published on a previous post (sorry no name).

https://www.amazon.co.uk/gp/product/B002AT1F6C/ref=ppx_yo_dt_b_asin_title_o07_s00?ie=UTF8&psc=1

with the shield removed and a ground wire fitted.

It will be interesting to see a tear down of an esd gun one day

Peter

[Gyro]

One possible advantage of the squeeze type piezo lighters is they alternate spark polarity as you squeeze and release.

[doktor pyta]

https://www.tme.eu/de/en/details/h12-1a69/reed-electromagnetic-relays/meder/#

[fcb]

Actually made some progress on this project/idea.

I have now got some ultra high voltage reed switch samples in, made some drive coils. Will be testing these next week.

[EEngo]

Hi! which relay did you order and did you have any success so far? 

[floobydust]

A Trigatron gas switch is what I thought is used. A modern one could easily use a laser to ionize (fire) it. I haven't tried making one.

Anyone know what component joeqsmith uses in his transient generator? We never got a tour of the insides.

[fcb]

We got some fairly "standard" vacuum reed switches rated at 15kV. Results were a so-so, wave-shape reasonable to begin with but deteriorates (contacts being eaten) with use/abuse. Was able to conduct the testing I wanted to do and verify it with an NSG435.

Not really sure how to proceed - would love to do it with solid state as 15kV doesn't feel THAT high. Other thoughts are SF6 gas drenched mechanical DIY relay.  Or pony up the big bucks and buy something Gigavac.

Will have another look at it when I get some more free time.