Re: IGBT long life at 3X rated Ip?? Re: 15kva 14.4 kV Transformer.. (fwd)

["Tesla list" <tesla@xxxxxxxxxx>]

Moderated and approved by: Gerry Reynolds <greynolds@xxxxxxxxxx>



---------- Forwarded message ----------
Date: Sun, 24 Dec 2006 16:56:44 -0700
From: Terrell Fritz <terrellfone@xxxxxxxxxxxxxxxxxxxxxxx>
To: Tesla list <tesla@xxxxxxxxxx>
Subject: Re: IGBT long life at 3X rated Ip??  Re: 15kva 14.4 kV   Transformer..
    (fwd)

Hi Finn,

At 05:56 AM 12/24/2006, you wrote:
>From: Finn Hammer <f-h@xxxx>
>Subject: Re: IGBT long life at 3X rated Ip??  Re: 15kva 14.4 kV  Transformer..
>     (fwd)
>
>
>Terry,
>
>"Us" that commercialized the solid state coil ;-} have been running
>multithousand amps for years.

:-)))


>A replica of Steve`s 2-brick OLTC motor is running in museum service,
>and it is counting 1 1/2 year now.

Perfect!!

>The coil is intermittent running, In short 116 ringdown long bursts, to
>simulate lightning strikes. These bursts are 0.3sec long at 300BPS. That
>ads up to around 1 hours total a day. So a very rough guess is that it
>is up to around 200 hours now. This coil is using PRX CM600HA-28H`s,
>doing up in the high 3000ths of amps.

:-))))))))  That gets us past the 100 hour mark which is very good!!


>If IR would supply me with replacements, I could swap them out on a
>service visit.

Leave them in!!  Wait till they fail (if ever).

Those bricks can be cut open very easily and they are only put under 
silicone gel so no special equipment is needed to gain excellent access to 
the die.  No boiling fuming nitric acid is needed ;-))  So "if" there is a 
failure, the die can be examined under a lower power microscope for the 
usual signs of high current damage.  But if they never fail, who cares ;-))

Your application is perfect for such a test and they have probably already 
done "good enough" so that most worries are satisfied.  1000 hours is the 
next big step...

Cheers,

         Terry



>But if they fail, I`l post the results, of course.
>
>Cheers, Finn Hammer
>
> > The only real concern is how IGBTs stand up to long term continuous 
> massive
> > over currents.  If anyone gets say 1000 hours on an IGBT with very well
> > known over stresses, we could probably send it to IR for analysis.  I used
> > to do that myself but I no longer have direct access to the "stuff".  I
> > have friends though who still do it.  Maybe IR or Motorola has already 
> done
> > this.  Probably not though...  If we had the parts, they would "love" to
> > know as much as we and it is simple to do.
> >
> > The issue really is what happens to the wire bonds and top metal layers on
> > the die.  Current pulses at that level can do really wild things and I am
> > not sure if they are "bad" in our time frame.  My DRSSTC IGBTs have copper
> > layers on the die which is an odd twist and the SISG die might use copper
> > too....
> >
> > DC's 24 hour test tells a lot, but not at the say 100 day range.  I am
> > slightly worried that some problem might be lurking out there "sooner than
> > we like" that might kill are poor IGBTs in very long term applications.  I
> > was thinking Finn was safe using CM600 bricks, till he started running 
> 4500
> > amps!!! :o))))
> >
> > So somebody has to do something someday to figure out something about
> > anything...
> >
> > Cheers,
> >
> >          Terry
> >
> >