where are the 200c transistors?

They're all mil spec. Did you have a cost target in mind? :^)

Tim

Well, "we" (Oil 4 Less LLC) sell a solid-state shunt regulator that is guaranteed to 400F/204C, programmable from 400V up to 2500V; sand power, not rock power.

Cost is not an issue Tim.

Well, I've seen those brazed gold packages on Digikey... GeneSiC and such. Looks like they drop off around 225C, not 250C. I don't think anyone makes anything SiC under 600V, because... why bother, I guess? I expect it can be done (what's Vebo on a typical SiC BJT?), it's just not.

As for what you *would* use at 250C and lower voltages... does... Raytheon still make subminiature vacuum tubes? :o

I expect you're asking for something so rarefied that manufacturers don't bother talking about it. Useless to ask here (unless someone here has already done it -- AND isn't under an NDA!), or even at suppliers, likely...

I'm not familiar with the vast world of downhole (and related) electronics, but it's true in other market sectors that, you're probably stuck with asking mfg reps and quoting and sampling and NDAing and all that bothersome old timey stuff.

Tim

I suspect you're right. Transistor dice _should_ work at those temperatures, but with reduced lifetimes. You need to find someone who bothers to put them into packages that work, and to characterize them (including lifetimes, which will be reduced) at those temperatures.

I know a couple of people who do high-temperature design. Contact me off- list and I'll see if I can hook you up.

Bob, a "600V" Mosfet will still work at 100V -- it'll just have a lot of overhead, and have a high Rds_on for its size and gate capacitance.

I suspect there's mainly package limitations. The old 2N3055 in a welded steel package is rated for 200C junction temperature, though the performance is less than spectacular. Metal-case DO5 diodes, and some glass-metal DO-35 diodes, are also rated that high.

Back in the days when I was running the Dickson Electronics (now Siemens, Scottsdale) hybrid facility, Schlumberger was one of my favorite customers... send one of my hybrids down-hole, pull it back up, take it out of its socket, toss it and replace with a new one.

(Gold-plated can dual-inline, double-stack alumina boards inside.)

So I had them send me back all the "used" parts, which I de-lidded and tested thoroughly... diagnosis... NO circuit damage except the gold-plated can was discolored ;-) ...Jim Thompson

They don't make them anymore, but when they did they made about ten million of them and there are still plenty available:

I have no clue where you'd find them, but bond-out and ceramic packages is something that could be done in a two-car garage (if you didn't want to park a care there, ever).

And you can buy bare dice.

Ergo, the solution is out there if you can just find it.

And an intrinsic diode with an AWFUL reverse recovery spec, unfortunately.

boB

If the voltage is low enough you can parallel the body diode with an external Schottky and it won't get much of a chance to misbehave.

Best regards, Spehro Pefhany

At 200C?

Begging the question... :^)

I think even SiC schottkys aren't the greatest at 200C..

Tim

Right ! From what I have found in past years is that lower voltage (less than about 100V) aren't such a problem as hundreds of volt FETs and their diodes. The problem with paralleling fast diodes like SiC or real Schottkys with power MOSFETs is that they don't have a low enough voltage drop to do much about the reverse recovery of the MOSFET diode. But if you heat them up real good, then they sure do. Real high voltage Schottkys anyway. (200+ volt) I've tried just letting the paralleled diodes be not heat sinked and that doesn't work very well either because their starting voltage drop was just too high compared with the MOSFETs' diode. I wonder just now if I was to heat them up first maybe if they would stay hot ? Hmmm...

I do like SiC diodes for high voltage but they still have a high-ish drop. Synchronous rectification can work with Sic FETs but they are so expensive compared to power FETs. Need just a few more years on them I guess so they can be a bit more affordable for me anyway.

boB K7IQ

Why does this

have a fuse in it? ;-)

Simple. That makes it a PROTECTED electronic fuse device.

The feed wires should be reduced in size to one just above what is needed for the maximum controlled current level, however.

Unlike an AC feed line, we do not need to think about line drop here.

Get some SPC wire at the local mil surplus shop, or a small roll from the wire mfg boys.

Feed a 3 amp circuit with 5 amp wire, not 20 amp wire.

The outfeed can be big and short to reduce R loss, but unless really long, it won't be a factor either. PVC sucks when SPC is so cheap these days.

I know, hence the winky.

A real fuse, in case the fake fuse doesn't ;-)

Also in case of reverse operation.

The waveforms I've measured in testing it are pretty neat... I discovered my bench supply delivers over 40A short circuit, for the few microseconds before the fuse turns off. A wad of electrolytics will spike over 100A when switched externally, before the current limit kicks in (a half a microsecond later), and the fuse turns off (three microseconds later).

Tim

Den onsdag den 21. januar 2015 kl. 00.07.58 UTC+1 skrev Tim Williams:

useyou two fets in reverse so it'll work for AC too, I've done as a main switch to protect a bunch of Triacs that couldn't handle a dead short for a full half period

-Lasse