Re: Recent s.s.t.c work

["Tesla list" <tesla@xxxxxxxxxx>]

Original poster: "K. C. Herrick" <kchdlh@xxxxxxx>

You may recall my postings of 2 months ago.  After 2 abortive attempts 
following the failure-by-"improvement" of my first--& only--successful s.s. 
Tesla coil, I've summoned up sufficient residual motivation to try just 
once more.

Slowly & I hope more or less surely I've now reached the stage where I have 
to plug it all in & find out if I can make it play.  Always the most 
exasperating part of the whole process and I don't look forward to 
it.  Murphy is forever hanging around my shop!  I will ask Terry to post 2 
group portraits--virgin ones, in the sense that not one single thing has 
yet been fired up--as

http://hot-streamer.com/temp/KCH_TCH1.jpg

and

http://hot-streamer.com/temp/KCH_TCH2.jpg.

In TCH1, the 24"-square platform at the right holds the series-resonant 
primary assembly and the driving electronics.  As I wrote previously, the 
6-turn, paralleled-1/4" tubing primary coil is wound inside an 18 quart 
butyrate food-service bucket, along with the resonating capacitors.  The 
latter are physically arrayed so as to constitute (more-or-less) an 
additional inductive turn.  The bucket has no conductive holes in it and 
its material is almost 1/8" thick.  Inside the bottom end of the 100 KHz 
secondary (see TCH2) I've mounted the major part of an identical 
bucket--but made of (cheaper) polyethylene--so that it and its attached 
secondary fit snugly down over the primary's bucket.  That gives me two 
bucket-thicknesses plus the Sonotube thickness between the primary and 
secondary coils, which will likely provide sufficient voltage-withstanding 
capability.  I used the rim-portion of the secondary's bucket as mounting 
for the capacitors.

In front of the coil/capacitor bucket is the driving electronics, all 
mounted on a large rectangular heat sink with the whole assembly merely 
sitting loose inside four corner brackets.  The four blue capacitors are 
each 4000 uF/450 WV with each paralleled pair charged via a rectifier 
driven from the variac, one to + ac peak and the other to -. Two H-bridge 
IGBT "bricks" are mounted between the capacitor pairs and under the 
mostly-wire-wrapped circuit board that holds the low-voltage 
electronics.  On that board I have 4 identical NPN/PNP "totem pole" drivers 
for the IGBT gates.  They, in turn, are driven from a 1:2:2:2:2 toroidal 
transformer just visible under the right corner of the board.  The 
transformer is driven from another pair of NPN/PNPs which in turn are 
driven from a UCC37321/UCC27322 pair of IC drivers.  Those NPN/PNPs will 
deliver ~2x10.5 V to the transformer primary, which is stepped up to ~42 V 
p-p (absent losses) at the secondarys.  Asymmetrical charging of the 
drivers' source-capacitors will yield ~+25/-15 V to be applied to the IGBT 
gates.

The bricks (courtesy of Steve Ward) are rather modest in capability & I 
hope, if all goes well, to be in the market for more robust ones e.g. the 
Powerex CM300DY-24H, for which I've provided space.  Anyone offering those, 
used but good?...

At the far center of the l.v. assembly I have a Rogowski coil, through 
which pass both H-bridge Cu-braid output leads (in opposite 
directions).  That coil's output connects to an over-current sensing 
circuit and thence to a 1:1 toroidal ferrite-core transformer.  That, in 
turn, is ac coupled to a diode array that clamps its output to +5/-0.7 V 
for driving signal-conditioning circuits that output to the previously 
mentioned IC drivers.  I've incorporated a "pilot oscillator" there to 
maintain excitation through the 5-winding transformer at all times, for the 
purpose of charging the source-capacitors for the totem-pole IGBT 
drivers.  Those capacitors are charged continuously from the signal 
windings of the transformer.  As soon as the primary develops current, its 
Rogowski-coupled signal will swamp out the pilot oscillator's signal, 
providing the positive feedback to sustain oscillation.  At least, that's 
what the simulation shows...

In each of the 4 driver circuits, I have an opto-isolator whose purpose is 
to block plus-going IGBT drive signals until a spark is to be 
generated.  For each spark event, all 4 opto-isolators are turned on 
simultaneously for the duration of the event.  But at all times, the 4 
transformer-outputs are present, both to keep the drive-source capacitors 
charged and the IGBT gates at about -15 V.

The conditioning circuits also include a synchronizing circuit to ensure 
that the spark events end at primary-current zero-crossing.

To the left in TCH1 is the variac assembly.  It, plus its cables for 
storage, will drop down inside a fiberboard box (behind, that was 
originally used to ship magnetrons).  In that assembly I have a 20 A variac 
(courtesy of Dave Leddon, and I thank him again for that!), a mains 
contactor, switches & circuit breakers, the 5 V/12 V power supply and a 
small board that holds the spark-burst controller.  The latter includes a 
simple oscillator and pot for setting the spark rate.  For setting the 
spark duration, I count primary cycles and cut off the drive after the 
selected quantity.  Selection goes in 5 steps from 64 to 1024 cycles.  The 
variac assembly connects to the main assembly thru a power cable + a 
control cable and is itself to be powered from the 115 V mains.  The power 
supply is a beautiful little jewel I picked up surplus; I'm going to mount 
a spare one on a little pedestal and display it as an artwork.

If & when that happy day arrives when I actually make sparks, I'll add a 
triangular strike-screen above the electronics assembly: plexiglas to keep 
stray crud off the electronics and wire screen to keep stray electricity off.

I've simulated a good part of my l.v. circuits, so perhaps that will save 
me just a bit of the frustration normally to be expected.  I'm determined 
that this will be my last--my very last!--Tesla project; one has to stop 
some time and well-past-77 is more than a good age for that.  Seventy years 
as an electricity-nut...it's hard to believe!

Ken Herrick