A very upsetting, and embarrassing, development.

Once upon a time you said you were using IR22141SS gate drivers. Did that change?

Did you get around to doing any actual welding, or were you still running your inverter into a short?

I do not really want to share this, but feel that I must. I fried my gate drivers today. I think that I miswired the snubber backwards when I tried installing the snubber boards that were on these IGBTs originally that I bought off ebay.

Fortunately, the seller of gate driver boards has something similar, but it is highly disappointing. Everything actually did work under power, just yesterday, very nicely. And now I have this disgusting smell in my garage.

What a sad day.

I went through *eight* sticks of MOSFET's getting a 10 kV, 10 A solid state switch to work. This was for a magnetron driver. I had a glass jar for the dead MOSFETS on my desk labelled "progress" (as measured by the liter).

Bosses were not happy as they already sold the thing before we knew it would work. Luckily, the buyer went bust.

Then I met the father of a friend - an old guy used to work in BAE-Marconi:

He did the exact same thing with transistors and we had the exact same experience, that the simulation works, the theory works, the lower voltage prototypes works, one can fire it straight into a spark gap and it works. Then one start on the real thing and the bin starts filling with dead transistors ...

Because - as one will eventually discover - a magnetron can turn on so fast that the Source bonding wire has enough inductance to blow out the gate.

I'm just glad the selenium rectifiers have gone!

Alan

This reminds me of a story. I think it was about Ted Taylor, one of the atomic bomb guys. When he finally had a bomb fizzle (most of his bombs were very successful) someone like Edward Teller said "Congratulations Ted, now you are finally learning something!" or words to that effect.

I think I read this in _The Curve of Binding Energy_ which is a really good book, BTW.

--Mac

I remember those. Little dark spots would form on the plates, followed by a curl of very nauseous smoke ;-)

...Jim Thompson

I guess... The somewhat consoling part of this is that I had something working prior to messing it up. I would be much more upset if I made something that never was working prior to frying it. I am just as sad, but a little calmer now.

ROTFLMAO

Very interesting...

Yes, I used Semikron 23's eventually.

Still was running it into a short, but it ran fine at 200 amps, for a couple of minutes. Then, the next day, I messed things up by incorrectly altering the snubber.

My favourite was 'The Greatest Power on Earth' - awesome!

Heh, heh! I think it went off, but the yield was very low. Apparently they call that a "fizzle yield" or something like that.

--Mac

Was it Ted who had to go out and see why one didn't go off? That'd be a little unnerving.... "Oh it's okay, the fuse was a little wet....." :-)

Ken

ten bucks says I have done stupider things, that cost more :)

Like forgetting to press the single-shot button on a scope before doing a destructive test on $1800 worth of IGBTs.

A boss of mine once killed ten 87C552 micros (in the ceramic, windowed package) in a single week, by carrying them in his hand from his desk to the programmer and back again. they were NZ$200 each.

Cheers Terry

PS welcome to the exciting world of power electronics. None of this piddly "gosh, my pA reference is drifitng with time" nonsense, just sizeable detonations.