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

Thanks, I have a very old Cyber Tig. It is pictured here:

I just went out to weld a little bit (with stick) and looked at current. You are right, it fluctuates a great deal around the setting.

I do, in fact, have manuals for it, with a few missing pages here and there.

Whew, the "reactor" is larger and heavier than the transformer! That should also make a hell of a lot of high-energy RFI and serious high dV/dt spiking as the arc is under way. Creaming Ignoramus1797's delicate IGBT power-control electronics. :>(

I don't know about that reactor, but its size shows that it'll store a considerable amount of energy. And I have experienced massive blown-out electronics (vaporized traces, bad ICs, etc.), whose only sin was to be grounded, via the ac power cord, to a large telescope frame that had something arc welded to it on the other side of the frame during the night shift. Arc welder RFI currents are powerful, and will flow wherever they can!

You're going to have to select the Rce, Cce or the IGBTs, and thoroughly study the operating range of the power elements. It is far more involved than simply plugging values into their app circuit and reading "typical" graphs. If the damned drive is that expensive then go to a GMR and actually quantify the current with complete isolation and more than sufficient accuracy and bandwidth.

If your manual is incomplete then go to the link I provided and get a complete manual (free) which will include the schematic and volt-ampere curves. I would have taken a look but I can't read the third digit in the serial number on your nameplate photo.

Just because your unit is rated at 200 A does not mean it will not put out well over that into a short, and just because it is rated at 70 V open circuit does not mean that the "inductive kick" of that big "reactor" will not push the voltage well over 1200 V when your IGBT attempts to turn off (probably with a big bang, one time only) - you had better provide someplace for that "reactor" energy to go when the IGBTs turn off, if you want them to turn off more than once!

Also note that the volt-ampere curves refer to the RMS value of DC current, because there is a fair bit of ripple, with peak currents higher than RMS value.

My model is old, so I called Hobart. They were confused, since my spec number applied to a Cyber Tig 150, and yet my Cyber Tig is clearly a

Someone will hopefully call me back and email the manual, if available.

You are 100% right. I realized that last night, when I looked at the amp gauge while welding and intentionally creating shorts.

That means, pretty much, that I have to parallel my 200A IGBTs. I am glad that I have 4 of them.

As far as protection from overvoltage, I plan to implement a snubber circuit.

Yes, that's true.

I appreciate your insightful comment.

Hi Winfield, I am very interested to know what you know about that reactor, its inductance etc. Thanks a lot.

Oh, I see. That's separate from the mere inductance issue, is that right? I am going to find some info on whatever it is.

By the way, I uploaded the manuals for my welder (although, surprisingly, the manual is for a Cyber Tig 150, but they insist that it is my "spec number".

That's why he must make sure his bridge shorts rather than opens during switching.

robert

I don't have a TIG welder, but if I (hopefully) ever have a good enough excuse to buy one, it'll be a Hobart.

I simply love companies that publish manuals on the Net. It's also a great service not only to the unfortunate owner who lost his manual, but also to the would-be customer who can see exactly what he's gonna get.

That said -- your Ignoramus's welder looks like a massive EMP device. No electronics is going to survive this thing.

robert

I think the design resistance for serial lines is about 300 ohms. at 40Khz that's equivalent to about 260nF what sort of capacitance do these IGBT gates have ?

Bye. Jasen

Listen up my ignorant sonny, your enthusiasm is noted, but I was checking out IR's ir2114 and ir21141 (600V), and their ir2214 and ir22141 (1200V) when you were still dribbling in kneepants. The datasheet is now rev C. The oldest version in my computer (out of six files) dates to before there were revs, when it was marked "Advance Data." I even have IR's original press release, which said, "Pricing begins at US $3.50 each in 10,000-unit quantities."

So there!

BTW, with your welder's over-active reactor (flyback city), and your 1200V IGBTs, you should chuck those wimply free ir2114 parts (who wants 'em), and upgrade to the manly ir22141 chips, complete with the all-important sought-after active-bias feature for those awkward, embarrassing desaturation-detector coupled-noise moments.

Dreadful indeed, but the iR2214 family at 2/3A isn't much better, when one is considering the huge IGBTs under discussion here.

Winfield, sorry, that was a typo. I did mean the IR22141SS chips, in fact. That's what I will, hopefully, receive probably tomorrow. I apologize for the confusion.

yes, but isn't that some enthusiasm? Maybe I'm a poor judge because I always lean to the pessimistic side, but I've never seen anything as doomed as this TIG welder project. In the end, Ignoramus will be out 50 bucks for his IGBTs and a few dollars for his gate drive circuitry, but that'll be a small price for the knowledge gained.

So... power to him!

robert

wrong. Typically detect Vce > 10V whilst device is on - called "desat" detection, as IGBTs "de-saturate" and limit current (to about 10x rated, depends on Vg and device) when switched into a short.

there are plenty of ways of creating such a short. Firstly, transient thermal events eventually cause IGBT modules to fail (ditto for T0-247, T0-220 etc), and they fail short-circuit (initially - add in enough energy and they fail "vaporised"). Desat detection can significantly limit damage, greatly reducing MTTR and $TTR.

Its not that hard to arrange a short-circuited load, either - forklifts are particularly useful. Likewise choke insulation can and does fail.....

I reverse-engineered one a few years back. They also play sneaky buggers during fault turn-off, to limit dI/dt. Nice little drivers.

Cheers Terry

its an adiabatic thing. typically, you have about 10us to shut the thing off, and current is limited (by the IGBT) to about 10x its rated value - IOW 2,000A for a 200A IGBT.

yep

Cheers Terry

Note though, that my switching frequency is extremely low. That gives me a little leeway here. The International Rectifier engineer who I asked for help, did not think that driving these particular IGBTs would be difficult, and suggested a particular bootstrap capacitor.

if you mess up the PCB layout, the drivers will fail catastrophically, and take out your IGBTs :)

I first looked at IR 1200V 1/2-bridge driver chips around 1997 or so. IR were trying hard to convince us to buy IR parts (they never succeeded, mostly because they were too damned expensive, and of course only make tiny IGBTs). Our techs annihilated 3 test boards, and about 9 chips, doing nothing nasty.

Apparently they got an idiot to lay out the test PCB, and it was done so badly that they pretty much all failed. oops.

I have also re-worked several designs using these types of chips, that failed for similar reasons.

Cheers Terry