[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: solid state simulation
Original poster: "Terry Fritz" <twftesla-at-uswest-dot-net>
At 04:49 PM 4/3/2001 +1200, you wrote:
>While thinking about computer simulation of the primary circuit, I wondered,
>why can't you do that for real? (putting two back to back zener diodes in
>for the static spark gap was not my idea, I read it on this list ;-)
Voltage devices like those will open up again once the voltage gets below
their rating. So If it is set to 10000 volts, it will conduct at 11000
volts but will stop conducting again once the voltage lowers to 10000
volts. Spark gaps will super heat the gap and stay conducting right down
to zero volts. They open for a short time when the current reaches zero
and begin conducting again very soon after do to the hot plasma in the gap.
It may be this zero crossing opening that gives the ring down the
logarithmic shape. That log shape is far less pronounced in solid state
gaps (see below).
>Can you replace your spark gap with a solid state device (triggered switch
>for an SRSG or a pair of back to back zener diodes for a static gap) and
>replace the transformer with a, say, 10 to 100V AC source, and then use a
>frequency probe on a DMM or ocilliscope to measure the frequency that the
>capacitor and primary coil resonate at?
Yes, but it may be much easier to use other methods. I use an IGBT and
reverse diode to do the switching function.
>This would be done to get the primary voltage down to a level an Oscope or
>(especially.) a DMM can handle. Then either of these could be used to
>measure the frequency of the primary circuit.
In my lower power testing of the SSgap project this is what I do.
>10-100 VAC source charges tank capacitor. At point where SRSG fires (solid
>state switch is triggered) or static gap fires (zener diode becomes forward
>biased) the primary cap resonates with the primary coil. Frequnency counter
The problem with using a standard frequency counter is that the coil fires
in short bursts lasting say 100uS and then the gap is quiet for the next
9900us. So the frequency counter will only read say 20 cycles of each
burst multiplied by the break rate per second which would be say 100. So
it may only real 2000 cycles. A scope can catch the fast burst waveform
and read off the distance (time) for a single oscillation and give you the
frequency (see below).
>Question - does resonant frequency change, even slightly, with voltage?
>(15kV rms down to say, 50VAC rms)
Surprisingly, very little. Streamers add top capacitance which lowers the
frequency about 5% but if there is no breakout the frequency seems to stay
very constant regardless of the drive voltage.
>Can this be done? Will it work?
I posted a scope capture of my work last night showing pretty much what you
Since a frequency counter counts the "number of cycles per second", if this
waveform were only repeated once per second the counter wound read 7Hz. I
didn't have any pics of a frequency readout handy but you can see the time
per cycle is ~8uS for a frequency of 125kHz. If one lays a straight edge
across the peaks you will see it is a line indicting very low loss (15 amps
peak with the primary cap charged to about 700 volts. The IGBT stay stone
cold despite having only a tiny heat sink.
>Thank you in advance,