Re: Simulation of a conventional Tesla coil

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

Original poster: "Malcolm Watts by way of Terry Fritz <teslalist-at-qwest-dot-net>" <m.j.watts-at-massey.ac.nz>

HI Ken,

On 18 May 2003, at 16:15, Tesla list wrote:

 > Original poster: "K. C. Herrick by way of Terry Fritz 
<teslalist-at-qwest-dot-net>" <kchdlh-at-juno-dot-com>
 >
 > Terry (& all)-
 >
 > Thanks for the heads-up on gap resistance.  Oh yes...2-3 ohms kills my
 > "Marx" scheme for sure.  With a revised simulation and only 40 m-ohms
 > total around the loop, I get 200 KV out at 125 KHz and with 115 V 60 Hz
 > in.  But 2-3 ohms per gap?...forget it!  I'm glad it's mostly only in
 > simulation so far.  I'd started to build part of it for testing but
 > thought better of it before going very far.

I've been following your thoughts with interest. Before you drop the
idea, it might be as well to remember that the 2 - 3 Ohms figure
amounts to some sort of rule-of-thumb average and is not truly
representative of a real runtime situation. The gap resistance is a
dynamic animal which varies with gap current, and that of course is
changing all the time the gap is conducting. I have run coils with a
primary inductance of less than 5uH and although they were abysmal
performers by any standard, they did run. From an experimental point
of view, it might be worth at least doing some tests. It is always
going to be a bear running with such low inductance coils as the poor
L/C ratio puts you on the back foot to start with.

Some encouragement,
Malcolm

 > With a 2:500 turns ratio, 45 mH-secondary simulation, I get 3 KA peak gap
 > current so perhaps that would be feasible for IGBTs (6 of them in the
 > Marx scheme).  I think it would be, also, for Triacs since the way I have
 > it configured now, going back to the switched mains-sine-wave I'd started
 > with, there's a period of ~4 ms between 60-Hz half-cycles where gap
 > voltage and gap current are zero at the same time.  Thus, Triacs would
 > shut off each time as required.  With 2 primary turns, however, the mains
 > current is wimpy: just 7 A or so.
 >
 > But with a 1:500 ratio I get 8 KA peak at the 1st 1/2 cycle dropping to 4
 > at the next 1/2 cycle.  There's a 2 1/2 ms shut-off time-period and the
 > mains current is 17 A rms at 115 V--about right for me.  Perhaps still OK
 > for Triacs.  But I doubt I'll ever make it to hardware.  Anyone else want
 > to try?
 >
 > That extra mains current goes into the duration of the Fr oscillation and
 > not the peak voltage which is not much higher.  Sparks louder but not
 > longer??
 >
 > Tell me more about the ohms per gap.  Is that ohms per 1/4"-or-so arc, or
 > ohms per arc + hardware, or... what?  And what, I wonder, would be the
 > resistance of a 0.03" gap--with negligible hardware resistance?  That
 > might well be a lot better.
 >
 > Thanks for the correction on the SIMetrix URL; I'd forgotten that they'd
 > changed.
 >
 > There's about 50 KV difference in the voltage, either side of the 200K
 > resistor; thanks for that reminder, too.
 >
 > Ken Herrick
 >
 > On Sat, 17 May 2003 18:56:16 -0600 "Tesla list" <tesla-at-pupman-dot-com>
 > writes:
 >  > Original poster: "Terry Fritz" <teslalist-at-qwest-dot-net>
 >  >
 >  > Hi Ken,
 >  >
 >  > I have been catching up on your recent posts.  I was not able to
 >  > follow in
 >  > real time but I just went through them.
 >  >
 >  > I think the URL for the SiMetrix program is:
 >  >
 >  > http://www.catena.uk-dot-com/Pages/download.html
 >  >
 >  > One little correction, the coil's top terminal voltage (Probe1 NODE)
 >  > should
 >  > be taken on the other side of the 220k resistor instead of between
 >  > the
 >  > resistor and 5pF cap.  Probably does not make much difference.
 >  >
 >  > Another thing that is maybe real important is the spark gap loss.  I
 >  >
 >  > normally put the spark gap resistance at 2 to 3 ohms.  That looses a
 >  > LOT of
 >  > power but I think it is realistic.  The spark gap loss may have been
 >  > a
 >  > serious problem in you single/double turn Marx coil configuration.
 >  > The
 >  > loss in all those gaps may damp the coil's resonance into oblivion.
 >  > But I
 >  > am not sure about say 4000 amp gaps and what the losses are then,
 >  > but I bet
 >  > they are high.
 >  >
 >  > One reason that I thought the OLTC could work in the first place is
 >  > that
 >  > these big IGBTs can switch the currents with very low loss.  Far
 >  > less loss
 >  > than any spark gap.  I and Steve report that our IGBTs only get a
 >  > little
 >  > warm so the losses are far less than if there were a spark gap
 >  > there.  If
 >  > spark a spark gap were used on my OLTC, the loss (like 2 ohms) would
 >  >
 >  > completely kill the resonance since the loop impedance now is only
 >  > 0.10 ohms!
 >  >
 >  > Neat that you are having so much fun with the simulators!  It really
 >  > is
 >  > cool and a great help to be able to "test" things with a computer.
 >  > It is
 >  > also fun to try "faults" like a cap suddenly shorting or the
 >  > streamer
 >  > hitting the primary (that does not do much ;-)).
 >  >
 >  > Cheers,
 >  >
 >  >          Terry
 >  >
 >  > At 02:47 PM 5/17/2003 -0700, you wrote:
 >  > >Still playing with Newbury Tech's SIMetrix, and pleased with my
 >  > notion
 >  > >for using their "transformer" backward, I've come up with
 >  > >http://hot-streamer-dot-com/temp/conv-tc.sxsch,
 >  > >http://hot-streamer-dot-com/temp/conv-tc.sxsch.pdf and
 >  > >http://hot-streamer-dot-com/temp/conv-tc-grph.pdf.  Newbies & others
 >  > alike
 >  > >might want to go to www.newburytech-dot-com and download a free version
 >  > of
 >  > >SIMetrix.  You can then put the .sxsch file into its Work folder,
 >  > run it,
 >  > >and play with it yourselves.  Generate 500,000 volts with the click
 >  > of a
 >  > >mouse-button!
 >  > >
 >  > >If you have the cash, buy SIMetrix; I believe it's about $1000.  If
 >  > I
 >  > >were gainfully using it, I would.
 >  > >
 >  > >You can view the schematic and the first few cycles of primary
 >  > current
 >  > >and output voltage from the pdf files using Acrobat.
 >  > >
 >  > >This may be old-hat to spark-gap regulars but since I've been
 >  > >(masochistically) into solid-state, I've perhaps missed it.
 >  > >
 >  > >SIMetrix is so much easier to fathom than MicroSim's kludge that I
 >  > can't
 >  > >believe it.  It also now incorporates another simulation program
 >  > >(licensed from someone else) which I haven't tried out.
 >  > >
 >  > >Ken Herrick
 >  >
 >  >
 >  >
 >  >
 >
 >
 >