SSgap update

["Tesla list" <tesla-at-pupman-dot-com>]

Original poster: "Terry Fritz" <twftesla-at-uswest-dot-net>

Hi All,

I have been real busy working on my solid state gap.  Still mostly on paper
but all is looking good.  I have now gone to three IGBTs per 1000 volt
section.  That gives a peak switching current of 500 amps (at least, still
well under ratings ;-)).  The full blown machine should have 24 sections
for 24kV at 500 amps peak or a whooping 12,000,000 VA of high voltage
switching pleasure :-))  That will easily run a 15/60 coil at full LTR
power.  That's 24 rows of 3 IGBTs at 1200 volts and 180 amps each :-))
Makes one respect simple spark gaps more...

I simplified and took a lot of cost out of the modules but they will still
run about $50 each (buying IGBTs in bulk ;-)).  I never thought this would
be cheap $:-O  However, the modules are real tough and should be blowup
proof.  The only situation that worries me is if 23 modules fire and one
decides not to (a bad choice ;-))  The odd module will take one heck of a
hit (especially if it is repeated at 120 BPS).  That should not happen, but
Tesla coils have a habit of things not going quite right ;-)  I thought it
was better to really beef up the modules and make them bullet proof and
worry with making them cheaper (and more fragile) later.  This costs more
up front but helps prevent a lot of blown IGBTs during the "figuring it
out" stage...  The 72 IGBTs and 72 high power MOVs could easily handle 1000
watts of pure power dissipation all by themselves ;-))  Of course, the key
is to dissipate as little power as possible so all the power can go to the
streamers.  I have pretty nice heat sinks but they should run stone cold.
Mostly worrying about what can go wrong rather than right...  Amazingly,
the gap electronics can probably take direct streamer hits.  A miracle of
modern TVS technology.  I also have to be wary of stray capacitance and
inductance in the sections...

The fiber optics is being finicky but I just ordered up the latest and
greatest (but still very cheap) today.  I think I will still go to yet
another type but something will work eventually.  I had to go to red LEDs
to get more distance and signal strength in the cheap plastic fiber.  They
also have to be faster types since propagation and switching delays can
cause some modules to turn on sooner than others which just stresses (but
hopefully not breaks) things.  There is a lot of fiber optics in the thing
but that cost is not bad at all.

I worked a lot on the controller design today.  It has a quench dial so I
can set the gap to turn off at any time I wish (hehehe!!).  First notch
quenching has many advantages including taking heat, loss, and stress off
the IGBTs and sending more power to the streamers.  I will have it do
single shot, static, sync, or async (any BPS I want) modes at the turn of a
dial.  Since it is all electronic, I can pretty much do what I want.  It is
all linked by digital fiber optics and run off batteries at low drain
currents.  I will have a zero crossing and high voltage detector from the
NST and HV rectifier to feed fiber optic signals back to the controller for
static and sync timing.  It will also have over voltage protection clamp
trips to save the MOVs if something goes bad (like I turn up the variac too
much ;-))  The controller is rather complex since I am making it fancy with
signal lights, many firing modes, and scope outputs but it could really be
very very simple.  Once the best control method is found, most of it could
be eliminated.  I just made it do what I wanted for the time being, and I
wanted a lot ;-))  It is all high speed, heavily protected, shielded, and
bypassed CMOS logic so it should work in a TC environment.  If it all goes
haywire, all the stuff should protect itself so no real damage will be
done.  It is important to make all this as fail safe as possible!

I am suddenly faced with a few theoretical oddities.  Like running an LTR
coil from DC.  I figure it will still work.  Being suddenly faced with many
new options is kind of surprising, but I guess I will just see what works
and what does not.  No time to computer model out every last detail...  The
logic is all on proto board so it pulls out real fast ;-))  Even though I
have not spent a lot of time calculating stuff out, I find knowing the
electrical characteristics of these coils so well is a giant help.  So many
of the "numbers questions" are simply second nature...

Still very "paper only" but much real hardware is on its way.  Since stuff
is changing so fast I won't post schematics but I will get it all out
eventually.  Definitely a pretty complex toy but mostly simple stuff
repeated many times.  Right now, I will try and get a single module tested
for the voltage and current rating.  Then I will string a few together with
a MOT supply.  My small coil uses only a 9kV NST so it should be the first
to really get converted.  Then my big 15/60.  Starting to think about how
to stuff all this onto sheets of Lexan so I can lift it and fit it into the
car...

Still unknown what strange properties a fully controlled big solid state
disruptive coil will have.  Even though the gap will now be flashless and
quiet.  I wonder if the primary currents will make noise surging in the
primary coil.  I also wonder what "just" the streamers sound like...

Cheers,

	Terry