Re: Terry's DRSSTC - Odd IGBT damage??

["Tesla list" <tesla@xxxxxxxxxx>]

Original poster: Terry Fritz <teslalist@xxxxxxxxxxxxxxxxxxxxxxx>

Hi,

These days, the voltage waveforms (blue) from my DRSSTC look like this:

http://hot-streamer.com/temp/DRSSTC-ModeSwitch-01.gif

The used to look like this:

http://drsstc.com/~terrell/pictures/FullSystem-02.gif

Ever since I tried running with the deadtimes removed on the IGBTs, the 
corners of the voltage waveforms have gotten very noisy and the ~400nS flat 
portions of the waveform have gone away.  I put the dead time back in but 
"things have changed".  I have checked and rechecked all the drive 
electronics and the dead times still seem to be there even if one really 
looks close even at the output waveform (the current blips due to lack of 
current drive).  So the corners are noisy and the turnoff is slow and messy 
now.  I am also getting spikes that go well over the rail voltages (like 
100V!) which should not be possible.  The IGBTs seem to have lost their 
"warp"...  I double checked all the test equipment and probes too.

I am thinking the anti-parallel diodes have gone away.  Strange since one 
would think the IGBTs themselves would get beat up from tiny dead times and 
fast switching rather than the separate diodes....

So I am going to switch out the set of IGBTs and see if it all returns to 
normal.  If so, I will have the die checked to find what blew up.

Interesting that the thing still runs fine from an overall perspective.  If 
the diodes are gone, it is just pulling down the voltage with the opposite 
IGBT now instead of doing it the nice way.

So for the smaller high speed co-pack IGBTs, running with very low dead 
time my not be a good idea.  I also have just 5.1 ohm gate resistors in 
there but that does not seem to be a problem according to the 
literature....  I think I just have to get the IGBTs out and start testing 
them specifically to pinpoint what is going on.

So I "think" I blew something up!!  But it still "runs" fine :o))  Hard to 
say but I think the noise my be getting a little worse with time which is 
probably to be expected...

I just played with the coil at low power last night and there are all kinds 
of mysteries there to be played with ;-))  The very first burst seems to 
really "make the sparks",  Tuning and all that mode stuff have a pretty big 
effect too...  I wonder about our tuning of the primary to a lower 
frequency in regular coils.  Perhaps is not so much a streamer loading 
compensation but rather forcing the coil into the lower mode where things 
seem to run better...  It does seem to help the DRSSTC, but those effects 
should be the same in any coil!

Much to ponder...

Cheers,

	Terry