Hi there! I have spent years accumulating components and parts and building
small coils and having fun with the learning of this concept. I'm not an
electronics engineer but i should have been. My dad was but my interest in
electronics came late and it all started with an interest in high frequency
high voltage stuff. I wanted to see some sparks but in a relatively safe
way.
This leads me to my current build. It's a drsstc and I have these
components: 4 mica/silver capacitors made by custom electronics that are
pulse rated for 20Kv each with .0153UF capacitance each. That will be the
tank capacity for the primary. I have more from that same company but they
are smaller and have a higher rating. Best to suck with 4 matching ones.
These are top of the line and designed especially for this application. But
they are all o have so i don't want to ruin them.
I have a dozen, (but picked out 4 that have identical electrical
characteristics), CM600HA-24 IGBT modules with the plan to make an H Bridge
rectifier that runs on a generator to keep transients away from mains power
and pissing off the electric company and neighbors. It will likely be
somewhere around 340 volts DC being pulsed at 117KHz.
I put together a small drummer for it that has optical isolation- fiber
optic cable so the little 555 timer inside won't have a hiccup and melt
down those expensive igbt modules. The timer runs a dedicated driver ic, a
TC4422 that should be capable of slamming on and off those big modules
without getting any of them caught up in the Miller plateau and causing
catastrophic shoot through... if it's done correctly. Each one of those
modules has a huge gate capacitance that has to be instantly charged up and
pulled off so I also thought about using some large power transistors in a
totem pole configuration to do the job- that's all up in the air. I have a
4.81 inch coil that I meticulously wrapped with 32 gage high quality magnet
wire and it's been coated twice with that enamel they use on bar counter
tops.. you know the clear stuff that people like to put coins in and so on.
I also made a little frequency counter with a 555 timer and since it sends
square wave pulses through the coil, that harmonics ring clearly and not
just the native frequency as i thought. That's partly from me accidentally
choosing a timing cap that was 11 times bigger than I intended and found
this out by chance and it's useful so i left it like that. It's almost
better than putting the correct size in it because the timer stays well
within its comfort zone and just purrs along at a pace that's octaves below
the coil's resonant frequency but shows easily on the scope. This coil
demonstrates a high Q judging by the response and how quickly it goes flat
line with the slightest change of the input frequency-- it's tight.
I have 50 feet of 300mm copper water line for a helical primary that I have
yet to make. i plan to wrap it on a pvc pipe coupler that has fins attached
to it with notches cut into them at the correct spacing, so all turns are
equally spaced and exactly the same distance from the secondary coil. I
think minimum distance for the 117KHz is just less than 2 inches so I'll
use cutting board strips and put tic marks to reflect a quarter inch gap
between turns and make the whole 360 degree span end up one step higher for
the last strip, then all the others will fall into place as they transition
around the coupler.i made a low profile tap using ground strap and the
strap itself is molded to the tube for the primary using solder and a heat
sink to keep it from wicking at a section to make the hinge and i figure
that a silicone band threaded onto the primary coil itself can be the
holder that maintains a good amount of pressure on that clip. It is a
really solid connection that's ready to move along the primary tube.
I am having some trouble with the H bridge design. I plan to use three
layers of .1mm copper sheet laminated in alternating fashion both
polarities, one atop the other with kapton tape insulation between each
layer to cancel out the noise.. So the sheets that go to the collector of
module1 and the other collector of the moduk 3 for example, will get
pinched together at the terminal location and sweated together with solder
at that predrilled hole. The sheets that do not terminate there will travel
in that direction to within an inch or so and be rounded off prior to going
any further, and the kapton will extend for another half inch beyond where
the copper stops. Do the same for the other connections and I should have
made a good bus bar that accounts for the normal trouble that high
frequency electronics has. The skin depth will be around .18mm or close and
therefore will use the entire thickness of each sheet without causing stray
inductance to gather in the unused portion of copper because there won't be
any unused portion at all. I feel that using 3 redundant sheets for each
part of the three zones will be enough to account for all of the current
and provide a great highway for the electrons to travel. But only if that
kapton film will do as it states it will do and not punch through. Each
sheet will be intimately close to opposing fields with only two layers of
tape between them so I'm a little concerned. Also figuring out the best
way to orientate the modules on the large heat sink they will share is
another variable that I've been toying with. It seems that there's always a
connection that is beyond one that's already involved and in the same axis
plane...i wanted to avoid that. i realize that equal distance and symmetry
is very important to keep out stray inductance. Oh, i almost forgot to
mention that i have two giant MOV's made by one of Cornell's subsidiaries,
the CKE class and i plan to put those across the big cap bank i have for
providing the necessary voltage and current for large streamers. That way
there's no lag from the generator. No hiccups. The snubber caps are 450v
with 525 peak at 4700UF each so I'll parallel 6 of those I think will be
sufficient and not too much as to cause an explosion if something misfires
but plenty to cover for the bigger spikes. The logic stuff will run on a
separate transformer or battery. it'll need at least 16 volts or more,
depending on which driver gets involved, totem pole, or the TC4422. The
data sheet for the igbts states that nominal switching is done with 15
volts at the gates. Gate capacitance is huge! It's around 700+ nano farads
total to switch all 4 of them, pulling that out of two and dumping it into
the other pair without delay.. it's gotta be a beefy driver. Well what do
you guys think?? Will this methodology be able to pull it off? Am I missing
something critical? Does anyone want to get in on it? I'm knowledgeable but
not formally educated in electronics.. I'm pretty sure this is the correct
way to build it, but not feeling super confident. It's slightly over my
pay scale but I'm learning a lot about the nuances of high voltage
electronics. These are not just straight forward like most things... I
know about the skin depth and so on but have never made anything remotely
close to this kind of power bar. I have twisted wires to help cancel out
EMF before and that's standard practice, but laminating multiple thin
sheets together is entirely new to me. It sounds like an awesome way to do
it though. I need to test the idea with something comparable in voltage ..
that would be best.
Thanks for reading!
Patrick
_______________________________________________
Tesla mailing list -- tcml@xxxxxxxxxxxxxxx
To unsubscribe send an email to tcml-leave@xxxxxxxxxxxxxxx