Driver for IGBT? (-ve off and +ve on from CMOS)

Yes. Intereference will travel in on all possible paths.

Great, show us your calculations when you are done.

Alright. So, there is some degree of what I hope will protect my circuit against both paths (EM waves traveling in space and by wire)

Yes, I definitely will post a schematic and all calculations (exported to excel format for compatibility), links to all datasheets, etc etc.

Once I do the layout, I will post that as well.

I will probably use a breadboard first, being mindful to revisit the layout and use of breadboard before testing my inverter inside the welder.

Things like this, as far as I am concerned, are not done in one stage and are not plugged to do the work right after assembly.

Good idea. I expect he can get away without them while testing with his low power supply if he is careful, but at full power the same transients that will destroy his IGBTs could easily fry his scope at the same time. While he is looking for diff probes on eBay he might want to keep an eye out for suitable current transformers as an alternative to measuring currents with a shunt and a diff probe.

Well at least he should be well under a kJ. Although I previously guessed that his inductor was on the order of 10 uH, that was a typo, I meant 10 mH. I see that Hobart uses the classic 6 phase double-Y connection with interphase transformer, so the SCRs fire 6 times per cycle or every 2.8 ms, and if the inductor is sized to allow current decay of 5 amps at 20 volts in 2.5 msec then the inductor would be 20 volts / 2000 volt/sec = 10 mH. They might easily have designed for less ripple but probably not less than 1 amp, for a 50 mH reactor. So with a short circuit current of say 400 amps he only has 20 joules to dump somewhere unless I screwed up again :-).

Well, tell me the format that you want... Like I said, I will actually be using Gnumeric under linux.

Vaguely reminiscent of sky-diving ground school - the instructor is taking you through the training, while you're hanging on the harness, and he says, "Malfunction!" You look at the instructor, like "WTF?" and he says, "Don't look at me! When you're up there, there's nobody to talk to but God, and he don't talk to skydivers, 'cause he thinks we're crazy."

Cheers! Rich

I will have to check tonight if the welder's "work" terminal is in fact connected to the frame of the welder (and equipment ground).

A few days ago, I did look inside the welding machine at the commutator and stuff around it, and did not see any such connection. Since there is a commutator, the connection of work terminal to ground would need to be done after the commutator, and there was not much after the commutator, between it and the work terminal. But perhaps I missed something obvious.

I will definitely check with an ohmmeter, for a definite answer.

That would be a bummer.

That's a valid question, but, on the other hand, this particular difficulty can be overcome. I mean, I know how to tell one from another by using test equipment. :)

Thank you.

That's for sure. On the other hand, I am not doing it completely alone, there is plenty of people who already showed their ability to set me on the straight and narrow.

I will do calculations in excel... the schematic I will draw by hand...

I will do calculations in excel... the schematic I will draw by hand...

I hope i is paying closer attention than I am! And I hope he wears good protective gear, sooner or later those IGBTs are bound to try to shut off at high current.

I don't understand how you can do a schematic in Excel anyway.

I read in sci.electronics.design that Ignoramus13229 wrote (in ) about 'Driver for IGBT? (-ve off and +ve on from CMOS)', on Thu, 13 Oct 2005:

They are good at explaining how you walk a tightrope. But it will be you that is 50ft up without a safety net.

Enough with the TIG-- I'd rather spend the $10K to fly a MIG.

Best regards, Spehro Pefhany

That's actually incorrect. I verified it with an ohmmeter. The work terminal is not connected to equipment body/equipment ground.

There is infinite resistance between both work and stick terminals and the welder.

I can flip flop the polarity all I want, without affecting the welder's grounding in any way.

I kind of expected that to be the case, for obvious safety reasons.

I skydived once, it ws quite a harrowing experience. Well worth $75.

Let's come up with a way to weaponize IGBTs...

On Fri, 14 Oct 2005 02:43:19 GMT, Ignoramus25589 wrote in Msg.

Your TIG project will be a good start ;-)

robert

$10K for a MIG? Heck, go FCAW, you don't need the argon. ;-)

Remember when they invented the ICBM? Everybody thought it was the latest and greatest, until some congressman from Alaska said, "What's the big deal? The Eskimos have been having ICBMs for years!"

I read in sci.electronics.design that Rich the Newsgroup Wacko wrote (in ) about 'Driver for IGBT? (-ve off and +ve on from CMOS)', on Fri, 14 Oct 2005:

Must have been a Republican; he said 'Eskimo'. Very non-PC.(;-)