Folks-
FWIW, last night I became the first person to run a SISG coil from a
pole pig.
Before anybody gets *too* excited, it was just my medium 6" coil, and
I
really hadn't started the evening with plans to use the Pig at all.
Yesterday I finished mounting the SISG4 PCB's I got from Mark Dunn.
All
six boards, since I was planning on running at high voltage. I was
originally
planning on running the SISG with my 12" pig coil, but I figured I could
give it a try on my 6" coil first. After all, I've never run anything but
spark-gap coils before, and if I was gonna fry all that silicon, I wanted
to at
least see it work!
I had been using a SRSG with this 6" coil, so I had a 15/60 NST and a
Maxwell 37667 cap (.030 uF). I hastily made up a bridge rectifier, with
40
diodes in series in each leg (UF4007, 1A 1000V, 75nS fast-recovery).
Added a 200
Ohm 100W power resistor in each DC leg to limit the peak current through
the
diodes if something went wrong.
Well, I was mighty disappointed when I couldn't get anything to
happen.
Only thing that happened as I cranked up the variac was noticeable arcing
on
the brushes near the top end. No noise or sparks from anywhere on the
coil.
But at least no smoke or tripped breakers.
A quick test showed that the rectifier was in fact working, as a nice
DC
arc set fire to an alligator clip gapped on the rectifier output. Hooked
everything back up and still nothing. This is where having test equipment
ready
to go would have been very helpful... and one of several times during the
evening that I almost threw in the towel.
I thought about it for a coupla minutes, and then figured I might as
well use the "bigger hammer" method of troubleshooting. After all, it was
a
clear, cold night outside on the driveway, and if something's gonna fail,
I want
some flames to show me where the problem is! So I disconnected the NST
and
rolled out the Pig.
As I was hooking up the pig, I realized I may have inadvertently
solved
the problem. I suddenly recalled the importance of having the cap tuned
to
the power supply with a SISG coil. But my .03uF LTR cap was only workable
with
the 15/60 because I had been running a SRSG. I paused to contemplate
decreasing the cap size to work with the NST and SISG, but then I would
have had to
take the time to increase the primary turns, and would've lost bang size.
I
figured hooking up the pig was easier, so that's what I did.
Well, as soon as I lit off the pig I at least got *some* kind of
response. The neighboring kids showed up, and were very helpful in
pointing out
where the sparks/flames were coming from. For some reason everything
wanted to
fail on the side of the coil facing *away* from me, and they were on the
"dark"
side of the coil not illuminated by the garage lights so they could see
better.
First off, the primary arced hot to the ground terminals a couple
inches
below it. For a few minutes I thought it was inter-turn arcing on the
primary (which *really* confused me!) until the kids managed to catch it
and point
it out. I re-spaced the primary away from this area and that took care of
that problem.
The *next* problem was interesting. The inner turn of the primary
arced
*hot* to the bottom (ground) wire of the secondary. I spaced the
secondary
up a coupla inches, and tried intervening insulators, to no avail. I
thought
the primary was arcing to the secondary, but on a little more reflection
I
realized that it was the secondary arcing to the primary! In my haste I
had
intentionally neglected to hook up a grounding wire from an RF ground to
the
ground terminal on the coil. I had never run this coil that way before. I
poked a
10" piece of rebar into the wet ground, ran a #10 wire to it four feet
away,
and problem solved! I never realized how important having *any* ground
reference was, but it was a night and day difference with this setup.
I finally got some streamers off the thing - all of 24", and somewhat
brush-like. That's when I had a sigh of relief that at least the SISG was
actually working. But it was pretty anemic. The control panel for the pig
has
ancient, low-tech meters - a moving vane AC ammeter and a voltmeter. It
was
interesting to see the voltmeter climb as I cranked up the ganged variac,
until
suddenly it hit the breakdown point of the SISG and the current went to
the
ballast limit (60 Amps).
The Pig, by the way, is a 10kVA 14/4kV unit I got from Resonance
Research. So putting 15,000 watts into a coil and only getting a 24"
brush discharge
was pretty pathetic. This wasn't the first time I had run the pig into
this
coil, though. A while back I had done almost the same thing, trying to
figure
out why I couldn't get the calculated output from this coil with the
SRSG.
It seemed to me that it wanted more power. So I gave it "more cowbell"
with
the pig. The Pig had also hit the current limit on this occasion, but
with only
50" sparks. The power resistors in the Terry filter glowed bright orange,
and cracked their ceramic cores when they cooled down. So I knew I was
putting
a lot of power *into* the system, I just couldn't figure out where it was
going!
Last night I got a hunch and changed the Maxwell cap with another one
I
had gotten from the same source. But got the same results. Both caps
still
read the correct capacitance value, and the coil tuned with the primary
taps
where the calculations and the Terry Tuner said it should be. So I added
a .010
uF MMC string in parallel to the Maxwell.
After some retuning, *that* made a big difference! The tuning point is
surprisingly sharp with the SISG. At least, much sharper than it was with
the
SRSG. I think that's because the SISG just will not fire until the
voltage is
high enough, whereas the SRSG would allow lower-bang discharges to get
the
streamer going, which then would add enough topload to bring it into
tune.
So I managed 60" strikes to ground off the thing before the night was
over. I have no idea where the BPS was, but it seemed perfectly happy
running
almost eh same intensity from two breakout points simultaneously. My
8x36"
home-made toroid is so smooth, if I had no breakout point it wouldn't do
anything at all! Just faint corona off the strike rail! One piece of
aluminum foil
tape, taped *flat* to the edge of the toroid, suddenly allowed
full-length
sparks!
No primary strikes, but many hot ground strikes. I did run it for
about
half an hour total. Still pulled 60 Amps through the Pig primary. I did
have
another little flame when one of the diode strings in the rectifier arced
over, but after some physical re-location and no repairs the thing worked
exactly the same.
After all that, the IGBT heatsinks (all 24 of them!) were universally
"barely warm". At least they were all doing their part. All the caps were
cold.
But if I run it again tonight, I'm going to make the entire primary cap
an
MMC.
My first solid-state coil! I was very eager to see streamers without
the
distraction of a spark gap. I was amazed at how loud the arcs off the
topload were!
****Many thanks to Mark Dunn and Terry Fritz, for their advice and
patience!****
Here's my SISG tips:
1. "Maxclips" for mounting the IGBTs work fine. I mounted the heatsinks to
the board, then put compound on the IGBTs and heatsinks, then dropped the
IGBTs
into place on the board, then screwed on the clip, then went back and
soldered the IGBTs in. I figured this way, everything was in a happy
clamped
position before it got soldered solid.
2. The heatsinks are electrically hot on the SISG4 boards, so I took the
time to round off the corners to reduce any corona. Probably
unnecessary...
3. I used a 47 Ohm (DigiKey P47W-2BK 2W 2% metal film) "charging"
resistor.
I used a trim pot for the "turn-off" resistor (DigiKey 3296Y-502LF, 5k,
one-turn, top-adjust). I measured all of them before I installed them, and
from
the factory, they are all set to the "middle", at 2.222 k +/- 1 Ohm!
Turned
down to "20%" is about 400 Ohms. I ran them all at the factory setting of
2.222
k.
4. I put everything together with a 140W soldering gun. I never understood
the "use a low-wattage iron with heat-sensitive components" theory. I have
more
luck blasting it quickly, before the heat has time to spread up the
lead...
5. "Balancing" resistors are 1% Metal Film (DigiKey 1.00MXBK 1/4W). I
figured these should be as precise as possible, after all I'm running 24
sections
in series...
6. When I first went to order, DigiKey was out of the .001 uF caps.And the
.0012 caps! But the next night, they suddenly had a big stock of the *new*
.001's (DigiKey PF2104, the -JB RoHS style). So it looks like that little
supply-chain hiccup has been resolved.
I'll let everybody know what happens when I run the "standard" SISG
with
the "real" 12" pig coil. By my reckoning, the SISG shouldn't last long!
-Phil LaBudde
Center for the Advanced Study of Ballistic Improbabilities
**************Biggest Grammy Award surprises of all time on AOL Music.
(http://music.aol.com/grammys/pictures/never-won-a-grammy?NCID=aolcmp003000000025
48)
_______________________________________________
Tesla mailing list
Tesla@xxxxxxxxxxxxxx
http://www.pupman.com/mailman/listinfo/tesla