El Supremo

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

Original poster: "K. C. Herrick by way of Terry Fritz <twftesla-at-qwest-dot-net>" <kchdlh-at-juno-dot-com>

I have proposed a new solid-state Primary apparatus, "El Supremo", based
on my previously-described Entertainment Machine and strongly influenced
by Terry Fritz' OLTC design.  See prior postings and also the schematic
at 
> http://hot-streamer-dot-com/temp/tch2.gif.  plus the accompanying driver
schematic at  
> http://hot-streamer-dot-com/temp/tch2drv.gif.  I'll repeat here features I
see as important & then refer you to two more posted drawings showing how
I envision the apparatus to be assembled.

1.  Very low impedance: just 1 equivalent turn.

2.  Highest quiescent voltage to ground = peak of mains voltage.  Safe as
a toaster (well...maybe almost).

3.  Non-resonant: operating frequency is controlled electronically and is
not a function of the associated energy-storage capacitance.

4.  Current-switching is by means of IGBTs.  Each IGBT sees only a
maximum of 2 x the mains voltage at any time.

5.  Uniquely flexibile in mounting multiple storage-capacitors.  They are
readily attached, added, removed, changed in value without affecting the
operating frequency.

6.  "Push-pull" switching design causes 2 x Vc to be applied to the
"coil" rather than the 1 x Vc of other designs.  That 2 x Vc is
multiplied by the quantity of capacitor-groups incorporated in a
particular configuration.

7.  The design is flexible in that more or fewer capacitor/IGBT segments
can be incorporated, in a differing configuration, without affecting the
operating-frequency capability.  That frequency's maximum is limited only
by the capabilities of the IGBTs.

8.  The IGBTs may be driven either from a simple pulse generator sync'd
with the mains frequency or, as I presently do, from pulses derived from
the secondary's instantaneous Fr, making the whole coil self-tuned. 
Non-synchronous operation is also possible, of course, if powered from
larger storage-capacitors that themselves are charged from the mains.

Refer now to:

http://hot-streamer-dot-com/temp/tchasy1.gif

http://hot-streamer-dot-com/temp/tchasy2.gif

Drawing -asy2 shows the plan view of what I have in mind: a set of
1"-diameter copper-pipe segments arrayed in a 16"-diameter circle
concentric with a ~12"-diameter secondary.  The pipe segments are
interconnected at collector/emitter plate-assemblies.  One of those
plate-assembly locations is shown, in elevation, in drawing -asy1.

A little study of the two drawings should show what I'm getting at.  In
-asy1, I mount 2 sets of up-to-4 paralleled IGBTs symmetrically with
respect to their "negative" pipe-circuits.  That they are asymmetrical
with respect to the "positive" pipe circuits is of no consequence.

I depend on using the full emitter lead-lengths as current-steering
resistors.  If that turns out not to be sufficient, then I can drill
holes in the emitter plates appropriately to increase the resistances
effectively in series with the emitters.  You need only enough resistance
to develop a few hundred mV at the peak current of each transistor, for
the other transistors to take up some of the load.

I show 3 discrete diodes associated with each IGBT group.  One per group
might do since they are nominally in parallel--but they're cheap so I put
in 3.  I need them because the fast IRG4PF50Ws don't include them
built-in.

Note how the multiple capacitors are to be mounted: just tack-soldered to
the outer peripheries of the 1"-diameter pipes.  There, they are out of
the way of sparks and there, there's plenty of room for lots of them. 
They can be added & removed willy-nilly without changing the Fr
adjustment.  Adding more, for instance, will merely increase the
exponential decay-time of the overall excitation during each spark event.
 That, of course, would mean more mains current drawn and thus a
reduction in the maximum spark-rate--in order to keep below the current
limit of the mains.  But that can be compensated for, to a degree, by
reducing the width of the energy-burst during each spark event.

All the pipe assemblies are to be held together using simple
non-conducting tie-plates that I don't show.  They are to be attached to
the pipes using sheet-metal screws: very easy to do.

Here's a nifty way to get power to the assembly (see the schematic,
-tch2):  Mount the whole thing on top of 4, 1"-diameter wood dowels
located at the 45 degree locations in -tch2.  Then merely wrap a dozen
turns of 18 ga. "speaker cable" around each dowel and use that as a
bifilar choke for supplying the rectified mains voltage to the adjacent
capacitors.

I think this--or a lot of it--has great promise.  Terry, what do you
think?

In the interim, I'm going to try my hand at modeling this notion using
Newbury Tech's SiMetrix program--before I start bending any pipe, for
sure.

Ken Herrick