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

Re: DRSSTC thoughts...



Original poster: Terry Fritz <teslalist-at-twfpowerelectronics-dot-com>

Hi Steve,

At 09:03 PM 9/28/2004, you wrote:
>Hey Terry,
>
>Some comments:
>
>
> >
> > Dan ­ I see you used the model for streamer loading of 220K + 1pF/foot for
> > a streamer load in your SPICE models.  I was just wondering how well that
> > worked for you?  There is some concern that model is just "too simple" for
> > some things, so I was curious if it worked well in this case?
>
>Steve Conner has been battling this one for a long time.  A real
>streamer model is the #1 hinderance in simulations i feel.  Its just
>hard to get it exact.  And your results can change wildly depending on
>what you used for a streamer model!  Am i gonna see 1000A or 2500A?..
>was my question at one time :-P.

I guess it is the "best guess " out there....  Any info on the "troubles" 
would be useful to go to the "next step"...


> > I see that super high current IGBTs are really in demand now ;-))  But
> > monsters like the fabulous Powerex CM600HA-24H my be too heavily designed
> > for "power" when what we need is "current".  It can "officially" do 1200
> > amps peak which is nice but I know we don't need the 4200 watts of power
> > dissipation ;-))
>
>Yeah exactly.  They also didnt strive to minimize stray L inside the
>bricks either.  Greg Leyh wrote an article on this... cant remember
>where to find it though.  As we push these IGBTs harder and harder the
>little things like stray L inside the IGBT will add up quick!  And
>heck, we'd be doing something bad if we dissipated over 50W per IGBT
>(unless it was a REALLY big system).  Even with making 11 footers, my
>IGBTs stay cold to the touch.

Stray emitter inductances and higher frequencies and giant currents can far 
exceed our gate drive voltages...  Thus, the "Kelvin" connections....  The 
big boy makers will "feel our pain" here and make them for us ;-)))


>   So I wish to remind of the simple IR IRG4PH50UD.  A lowly
> > 24 amp rated TO247AC IGBT with diode.  But it uses a great big die in a
> > small package thus it's 180 Amp peak rating ;-))
>
>Yeah, thats a tough little guy!
>
>   Note that some people
> > have put a lot more current through them like this one happily eating 700+
> > amps :o))
> >
>
>If only the bigger IGBTs could withstand such abuse.  I only run my
>CM300s at 2X their peak rating.
>
> >
> > Above about 740 amps, the signal starts "flat topping" do to the fact that
> > the gate structure simply cannot support more current no matter how high
> > the gate voltage is.  BTW ­ I note a lot of folks no longer worry much
> > about the maximum gate voltage specs :o))
>
>Seems that 30V or so is common now, at least thats what i always use.

50V is the "killer" there....  But above 30V does no addled good...


>   But suppose we just
> > can't find them "cheap" or want new IGBTs without spending $1300...  Then
> > using a few of the little IRG4PH50UD IGBTs may start to be a good
> > deal.  They cost $16 a "pop".  If you get $1300 worth of them, you have 20
> > IGBTs per leg at 3600 amps peak "rated" and 14000 amps "unrated"!!  So it
> > may be useful to consider a "multi-mini IGBT" (MMIGBT ;-)) array in some
> > cases.
>
>Yeah, that certianly is impressive.  And it seems that you could take
>more control of problems with stray L.  Then again, how to minimize
>inductance on a layout consisting of so many individual devices is
>quite a feat!

Most BIG IGBTs are a "bungled mess" when it comes to higher frequencies...


>  However, Dan's data shows the primary current can almost
> > double during a ground strike!!  That is a critical bit of info!!  Do the
> > models show that behavior too?
>
>Ive simulated this in some of my models.  My big coil has active
>current limiting to keep things "safe" under these conditions.
>Shorter, heavier arcs make things much worse as they present lower
>impedances.

Cool!!!  We will prepare for that!!!!


> >
> > Of course, if off the self IGBTs just are not right, check this DigiKey
> > part out ;-))
> >
> > IRG4CC50UB-ND
> >
> > 600V 55A for a buck ;-))  I bet you can find a wire bonder on E-bay for a
> > song (Yipps!! guess not
> > 
> (http://cgi.ebay-dot-com/ws/eBayISAPI.dll?ViewItem&category=45045&item=3842974611)...
> > Just use the silver epoxy for the pad bonding too...)...  Silver filled
> > epoxy for mounting, and your there baby!! :o)))
>
>Oooh, now THAT can get interesting.  Talking to Jimmy Hynes, the idea
>came up to make our own bricks with some dies, but instead of
>designing for lots of power, design it to withstand very high
>temperature transients.  That is, put heatsinking on BOTH sides of the
>die to remove heat even faster.  Or even just some sort of small
>aluminum block on top of the dies to wick away heat faster... then IT
>can be dealt with via forced air cooling.  I wonder if they come in
>1200V versions :-).

"Top of die" heat sinking is typically a "giant" disaster in the 
industry...  Cray tried it but.... Disaster...   Best to use old "thin die" 
bottom thermal dissipation teck there...  Happily, within reach of 
"basement shops" these days ;-)))


> >
> > Also note that the primary cap strings could be distributed between the
> > IGBTs to force current sharing like the OLTC does if that would help:
>
>Not sure if there would be much reason for that one, though if we were
>paralleling a bunch of small IGBTs it may somehow work.

It was a "giant deal" for the OLTC...  Maybe not a big deal in the "end 
game"...


> > Apparently, primary to secondary arcing is a  problem.  This is often due
> > to incorrect primary to secondary tuning.
>
>Or your secondary throwing sparks 3X its length!  Thats when my coil
>starts to show some signs of being stressed out ;-)

Jimmy had computerized circuits there ;-)))  Can we say "predictive 
impedance compensations" and such ;-)))


>   Since DRSSTCs are self tuning,
> > there may need to be a little "tweaking" the that circuit to better
> > match/lock the primary and secondary frequencies.
>
>Well, the tank circuit can't tune itself.  I think that is what was
>meant.  Tweaking the primary L is what is needed... or just turn down
>the variac a bit ;)

Yep!!  Need to thinkies there...

> >
> > The SPICE models can be used to force all kinds of fun fault conditions
> > too.  Primary to secondary arcs and such are just t timed switch
> > away.  These studies tend to reassure one that such events usually are not
> > too bad and can also find a few real doozies!
>
>Ooh, i will have to try that out in my next sim.

";-))))  Yaeh!!!  Tell the model to predict what happens when things go all 
wrong...


> >
> > One may "possibly" be able to use 60Hz resonant primary charging for the
> > DRSSTC like the OLTC used.  That would eliminate a lot of the DC charging
> > and giant cap stuff.  But I do note that "I" seem to be the only one
> > foolish to go that way ;-))  But I have not had any problems with it...
>
>I dont think *my* line can supply 1200A for even very short periods of
>time.  I think we need these big DC caps there to be a low Z power
>supply.  Maybe im wrong...

An interesting area there...  Big inverters draw giant currents for tiny 
bits of time off the AC line and seem to "get away with it" very well...



> >
> > It would be neat if there were a fairly off-the-shelf system that could be
> > hooked up to any already made coil to get rid of the HV stuff and simply
> > convert it to DRSSTC operation right then and there.
>
>The main problems are 1) tank impedance for DRSSTCs are usually less
>than that of a SGTC.  2) operating frequencies can tend to be too high
>for efficient use of a solid state driver using slugish IGBTs (though
>they are indeed getting much faster!).
>
>I too must think that the days of the
> > spark gap coil are numbered.  Not because there "are" good SSTCs out there,
> > but because the will soon simply be the "best" TCs...  Even "big" systems
> > are coming to speed fast!!  The record for a point to point arc is 59 
> feet...
>
>59' for what???  a solid state system?  I was pretty sure that SGTCs
>were beyond those spark lengths for some of the larger systems, but
>maybe not?

That was 59 "FEET" ;-)))   Ain't NO solid state system gotten that far :o))))

Cheers,

         Terry


>Steve
>
> >
> > Maybe this all will be a bit of help, forget any parts that are obviously
> > stupid :o))
> >
> > Cheers,
> >
> >        Terry
> >
> >