RE: 48kW DRSSTC

["Tesla list" <tesla@xxxxxxxxxx>]

Original poster: "Hooper, Christopher AZ" <christopher.az.hooper@xxxxxxxxx>

Hi Steve,

I agree that the design is the number issue!  I have really only lost
one IGBT (in my life of DRSSTC's) in little red @
http://users.cableaz.com/~chooper/images/1igbt.jpg due to a primary
strike as this is an older design with no OCR and feedback was on the RF
ground as now using primary feedback (why I bought so many 40N60's I do
not know, need some he, he). That is now why I only pump sparks out the
top and put the primary cover protector on little red, however; working
in the semi conductor field for a long time, we have a thing called bath
tube curve as all silicon will fail someday and we try to accelerate
this with a burnin process of putting the chips in a oven under bias and
run functional patterns to exercise the device. This pushes the bath
tube curve out 2x (trade secret) years and screens out 9x% (trade
secret) and failures x% (trade secret) is allow for commercial product.
When I worked in the Military division, this was not allowed as the
silicon was in war heads (nukes), space station, etc. and the silicon
was never allowed to fail. To ensure this, we had extensive testing and
long burnin times in the oven (125c) for a xxx hours (trade secret) to
push and the failure rate out 2xx (trade secret) years. Would be bad if
the Military pushed a button and the nuke fell over and popped
(whoooops) sorry Kansas, he he. Anyway, commercial product (not 833C
compliant) is allowed to fail and will someday. Electron tunneling will
put small holes in silicon no mater what you do (run at spec), and
someday a path to metal will happen. If you go with 883C compliant
parts, this will be pushed out 2xxx (trade secret) and no concerns but
if I remember a /S device (Space) for a first generation Pentium was
($20,000) a part. So to be able to repair a coil in front of 100 people
in less than a minute to me is worth the design cost, but you know me
anal thing.

Happy DRSSTCing indeed
christopher

-----Original Message-----
From: Tesla list [mailto:tesla@xxxxxxxxxx]
Sent: Friday, December 02, 2005 12:10 PM
To: tesla@xxxxxxxxxx
Subject: RE: 48kW DRSSTC

Original poster: Steve Conner <steve@xxxxxxxxxxxx>


  I just want
to bring the coiling of DRSSTC to a more run time and less bench time
repairing blow bricks and such.

Well so do I. But my plan of attack is to try and design a system
that won't blow IGBTs in the first place.

My driver has overcurrent protection and reacts in a sane way to
missing or bad feedback signals. I also believe in pushing the IGBTs
somewhat less hard than other DRSSTC builders do. I try to avoid
going much over the manufacturer's absolute maximum current rating.
It may mean using almost twice as much silicon for a given spark
length, but I bet I will end up spending less on silicon in the long
run, if I never blow any.

The bigger the coil gets, the more important it is to do it right
first time. By the time you get to the 48kW level, I reckon each
inverter blow-out will cost you a couple of thousand dollars, if you
include the cost of the silicon, shrapnel damage to nearby
components, the time that your hired football field and generator
truck spend sitting idle while you do repairs, and the damage to your
flak jacket. :-o

On my current research coil, I used 40N60s and set the primary
current limit to 400A. (the frequency is around 220kHz.) I have got
sparks up to 4.3 times the length of my secondary, and too many
flashovers and primary strikes to count, but never lost any IGBTs
yet. You can't get less bench time than none.

Steve Conner
http://www.scopeboy.com/