[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [TCML] Spark gaps, Solid state switches and diodes

Thank you Mark. Now I understand the situation that would cause you to attempt some off time adjustment. And yes, I do appreciate that this was some time ago in the infancy of SISG technology (all your hard work and Terry's). A cause and effect aspect is the large C values increasing charging time significantly for any given supply, but I agree, the small time frame should not help "unless" you just happen to be right at that edge within that window (maybe you were on your coil).

I think with Roff fixed at 4700 ohms, "all" coils will perform just fine without problems as long as the power supply is sufficient to charge Cp within the 120bps range. In my case, plenty of power for the 6 boards and also using standard Cp, so no charging issues. I'm using Terry's 51 ohm value for Ron (3W). I could play with pot's there, but given the quantity of boards and the need of 24 pots, that would be a nightmare to mess with.

Again, thanks for the explanation. That greatly helped me realize what it was you were trying to do, and all makes sense now.

Best regards,

Mark R Dunn wrote:

An SISG coil fires at twice the line frequency, in my case 120 Hz.  As one
increases the SISG threshold voltage (by adding SISG circuits) there comes a
point where one approaches the peak voltage of the power supply and the cap
is not sufficiently charged to fire the SISG in one cycle.  As a result the
firing rate is halved because the SISG fires on every other cycle.  On my
coil I reached a point where it would sometimes run at 120 Hz and other
times run at 60 Hz.  It seemed somewhat random as to which frequency it
chose.  I began to decrease the turn-off time to see if it would add enough
charging time to push the charging over the edge and return to consistent
120 Hz operation.  It seemed to work at the time, but in hindsight the time
gained was so theoretically small that it should not have made any
difference.  I encountered no difficulties, so I left the resistors at the
final value of 680 ohms.  Of course, the "minimum" value depends on the ring
down time for a particular coil.

You might also appreciate that this testing was done in June/July 2006.  I
believe at that time only Terry and I had working SISG coils and mine was
running with (4) and (5) boards at 14 to 18 kV.  Thus what at the time
seemed like an important test protocol for at brand new technology seems a
lot less important today.


Hi Mark,

The 680 ohm seems "on the edge" and specific to your coil. Were you trying to perform solid state forced quenching? This would be ok for some coils to a degree, but could be very risky to say my 70kHz coil that I'm running my SISG on where 330us is 3rd notch primary quenching. The 4700 ohm provides plenty of ring time without an IGBT energy crash and doesn't compromise charging. I'm just curious what you actually gain from setting Roff so low?

Best regards,

Mark R Dunn wrote:

I still have boards available.  Contact me off list if you are interested.

In the original SISG design Terry has a 4700 ohm gate shutdown resistor.
This gave a shutdown time of 901 uSec which was plenty of time for full
decay.  I ran a series of tests with progressively smaller resistors (1/2
value at each step) and stopped when I reached 680 ohms which gave a
shutdown time of 130 uSec.  My test coil has a resonant frequency of 168K
and rings down in about 80 to 100 uSeconds.  I did not go lower because I
was afraid of destroying IGBT's if I shutdown before full decay.

Waveforms can be found here:

I have never destroyed an IGBT on any of my SISG based Tesla Coils.


Recently, my boys and I built a Gauss Canon.  We used the SISG to trigger
it.  I misjudged the time it would take to discharge the cap because of
very low resonant frequency of this device (only 212 Hz!! for our unit).
we blew IGBT's until we increased the shutdown resistor value to 100K ohms
(shutdown time 19 milliseconds).  With the coil magnetic field wound up
energy when the IGBT shutdown the voltage would skyrocket and blow the
As soon as we changed the resistor, the canon began firing 1/2" steel
30 to 40 feet.  We are using one SISG circuit and firing at 900 volts.
Current is approximately 300 amps and each bang is about 180 Joules.

Mark Dunn

-----Original Message-----
From: tesla-bounces@xxxxxxxxxx [mailto:tesla-bounces@xxxxxxxxxx] On
Of Gary Peterson
Sent: Monday, February 09, 2009 10:26 AM
To: 'Tesla Coil Mailing List'
Subject: RE: [TCML] Spark gaps, Solid state switches and diodes
Thanks Bart.
I understand the SISG switch is usually left closed for two or more few
cycles in order to wring more energy out of the primary capacitor, and
had thought it could not be switched off any faster than that due to
switching-speed constraints imposed by the IGBT itself.  It appears if
operating frequency of the Tesla coil is sufficiently low, then the IGBT
be switched off at the "first notch."  Attempting to switch it off any
sooner than that would place a strain on the device itself, correct?
Here is another question.  I understand that Mark Dunn designed a printed
circuit board a while back that embodies Terry's SISG circuit design (see
http://www.teslaboys.com/SISG/SISG4PCB.pdf ).  Was this PCB ever mass
produced, and if so is it still available?
Tesla mailing list

Tesla mailing list

Tesla mailing list

Tesla mailing list