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

Re: DRSSTC - I think I heard Cross Conduction...



Original poster: Terry Fritz <teslalist@xxxxxxxxxxxxxxxxxxxxxxx>

Hi All,

I got my noise filter caps in and first tried them on the remote filter buss. The noise was still tearing off the H-bridge, but the filter board is an easy place to experiment:

No cap:
http://drsstc.com/~terrell/pictures/FilterBuss-nocap.gif

Ceramic cap:
http://drsstc.com/~terrell/pictures/FilterBuss-CeramicCap.gif

Mica cap:
http://drsstc.com/~terrell/pictures/FilterBuss-MicaCap.gif

The mica cap does just slightly better but probably not enough to worry about or justify their expense. The plain old ceramics should do just fine.

So before pulling the H-bridge to add the caps, I tried the caps across the output first to see how well they stop the noise at that point.

That caused a different kind of noise...

Sort of a magnum handgun noise. Sharp super loud BANG!!!! =:O Hmmmm. Hey! sounded like cross conduction :-(

Of course, if I had thought... Putting caps directly across the output of the DRSSTC was a real BAD idea :-P The first IGBT started up and it's current went "nice and high!". When the other side came up it all hung on and dumped everything into the two cross conducting IGBTS!! A dumb mistake....

But it did give me a chance to test that case :o)

As expected, the little IGBTs blew VERY loudly and made a nice flash. However, they did very little collateral damage:

http://drsstc.com/~terrell/pictures/BlownIGBT-01.JPG

http://drsstc.com/~terrell/pictures/BlownIGBT-02.JPG

I have not checked the drives yet but they should be fine. A small soot smudge on the case wiped right off. So it appears that a full cross conduction event did little damage to anything else. The lights did not flicker and the scope and such nearby did not even blip. So everything reacted just like it should have :-))

I was hoping a simple IGBT set replacement would have it right back on line. However, a trace on the bottom exploded too:

http://drsstc.com/~terrell/pictures/TraceBlown.JPG

The trace is vaporized all over the caps (cover them with tape next time ;-)) The traces that are not blown have intense pulse current cutting in the corners and cluster cracking... Probably went to may 10's of thousands of amps... The board just cannot be fixed reliably...

I kept wondering if I should get another set of boards incase but kept putting it off incase I wanted to change something (which I do). So It will be about a week to get new boards. I am going to tear all the parts here apart to see if there is any hidden damage. I will see if I can beef up the traces to keep the explosion in the IGBTs where it does little damage, but the current is enormous!!! Not sure if anything I do would help that. One lead of the IGBT is also "gone". I think it is splattered on one of the transorbs.

So my little buss noise experiment did not get finished. Now it is the cross conduction test... But no big deal other than having to wait about a week to get the new parts. But I do think the ceramics across the H-bridge buss is a wonderful thing that should really help the noise in the box.

So all you other DRSSTC guys out there can now sleep soundly knowing Terry finally blew his IGBTs up too :o)))

Science marches boldly on ;-))

Cheers,

	Terry