[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Please help ! IGBT exploded ...



Original poster: Jan Wagner <jwagner-at-cc.hut.fi> 


On Mon, 8 Mar 2004, Tesla list wrote:
 > Yesterday I was experimenting with an old IGBT - BSM100GB60DLC, which I got
 > from my friend, for my SSTC. I used the same gate driver I designed for
 > IRFP460 MOSFET H bridge.
 > I connected the terminals to 220V line and neutral and switched ON my gate
 > drive circuit. The IGBT exploded immediately ! I did not give any load
 > to IGBT. The H Bridge's output terminals were not connected to anything!
 > The gate drive was at 70kHz approx. I observed that the gate drive
 > waveform was not clean like a square wave.
 > Any advise for me ?

A sad loss, that one, a good IGBT it was... :| Better luck next time! ;-)
Suggestions:
  - use at least a 60W light bulb in series with the mains circuit before
    you turn it on, that way it won't explode and you can check whether
    the bridge shorts out or not (no primary connected but the bulb
    glows brightly => there's a problem in the circuit)
  - a more professional approach: use a variac/autotransformer, a mains
    isolation transformer, and a current meter (ok, the variac might
    already have an amps display...)
    With these you can bring the circuit up slowly, and check how
    steeply the current rises when you increase the voltage

About the IGBT - are you sure it was in ok condition before you
plugged it in?

The datasheet states 4.3nF as a typical input capacitance and 0.4nF for
the reverse transfer capacitance, which IIRC are pretty similar to
the IRFP460. Meaning, the gate signal should have looked almost the same
as with an IRFP460 - if the IGBT was ok.

If it was ok, maybe it exploded because of the long turn-off delay.
The specs state a 150ns-typical for an inductive load. That's no problem
at 70kHz if the load is inductive. But if the load is resistive or
capacitive (e.g., no primary connected) it's very much possible the
turn-off delay is much longer than those 150ns. For instance, I've some
older Siemens 3-phase IGBT bricks, the inductive turn-off rating looks
pretty, but the resistive and capacitive turn-off delays are in the
order of ~3...5 times longer. So there's a high risk that one IGBT is
still on (while it should be off), and the other is turning on, and you
get a short circuit. With no current limiting, the IGBTs explode.

So if you still got some of those IGBTs, one thing you can do to
troubleshoot is to add some dead time into the gate drive signals (==a
short pause during which both pulses are OFF). Hopefully that'll help.

Good luck!
cheers,
  - Jan

--
****************************************************
  Helsinki University of Technology
  Dept. of Electrical and Communications Engineering
  http://www.hut.fi/~jwagner/ - jwagner-at-cc.hut.fi