RE: DRSSTC driver tests- Dual resonance disaster

["Tesla list" <tesla@xxxxxxxxxx>]

Original poster: "Steve Conner" <steve.conner@xxxxxxxxxxx>

>This might be a good opportunity to bypass the PLL and just run it
>straight off its feedback signal and compare.
>I assume you tried altering the coupling between your coils?

You're right, it would be a good opportunity to test a self-resonant driver
while I have the mock-up dual resonator. But I'd basically need to build
another driver board, the circuit I have just now needs the oscillator
running all the time to function.


I did some more messing about with it last night. I tried altering the
tuning and coupling and I was eventually able to get it to lock to either
pole. I bypassed the interrupter and ran it CW as the pulsed operation was
just confusing things.

I think the problem was that the secondary was tuned pretty low. This
increases the primary current at the upper pole and decreases it at the
lower one. So the lower pole wasn't giving a strong enough feedback signal
to get the PLL to lock. When I tuned them both exactly together so the two
poles were the same size, I could get it locked onto either one.

The upper pole still seems more stable. I think this is because the IGBTs
switch cleanest with slightly lagging current, and it's easiest to get that
at the upper pole. I found if I tried to dial in lots of lagging current at
the lower pole, the PLL would just hop over to the upper pole.

Because of this it's tempting to run on the upper pole. But streamer
detuning will increase the primary current and decrease the output voltage
here. On the lower pole it works the opposite, adding capacitance to the top
makes the primary current go down and the output voltage go up. These may be
good things or bad things, I can't really imagine how it would interact with
the streamer load.


I also managed to test some of Antonio's tunings. I discovered if I reversed
the feedback connections I could get the PLL to sit midway between the two
poles in pulsed mode. I got waveforms that looked just like Antonio and
Jimmy H's simulations, I'll post some screenshots tomorrow.

I tried detuning the secondary in various ways and I was surprised how
robust these tunings were. I tried a heavy simulated streamer load of a ~200
ohm light bulb in series with 52nF (the secondary capacitance is also 52nF)
and it hardly changed the frequency needed for soft switching. Shorting the
secondary altogether didn't change the frequency much either.

I saw things that suggested these tunings could be used with longer burst
lengths. When I added the load, the beats disappeared leaving a decent
steady state response. The one problem would be- if using longer burst
lengths the coil must break out during the first beat or it goes completely
nuts. I think the sign of the feedback reverses or something.


Steve C.