I started with a secondary coil that I already had (28.2 mH), an
adjustable top load that resonates with it at ~300 kHz (10.4 pF), and a
22 nF MMC primary capacitor made with 9 22 nF 1600 V capacitors and 10
Mohm bleeding resistors. With mode 37:39:41 this results in primary
inductance of 13.5 uH and coupling coefficient k = 0.102. With
excitation at 294 kHz using a half-bridge powered by +/- 180 V, the
expected maximum output voltage is of 205 kV, and the maximum input
current of 90.5 A.
I made the bridge using two IRFP460 mosfets, assuming that they can
handle this current for brief periods (they can), and other parts taken
from PC power supplies, including the case, blower, bus capacitors, and
a switching power supply for the blower and other uses. The timing
control is a quite overcomplicated circuit made by my students, that
gives a specified number of cycles at specified periods. There is no
feedback, but just a button to adjust the driver frequency manually.
The designed operation required 10-cycles bursts. It worked well in this
setting, producing the expected waveforms, but unimpressive sparks and
streamers. I tried then to increase the burst length, without changing
anything else. The operation is then not so efficient, but the driver
was supporting the increased power. With 20-cycles bursts I obtained this:
The input current didn't increase excessively, starting as in the
10-cycles case, but reaching 170 A in something as a second beat. The
streamer loading turns difficult to calculate if this is the expected.
The driver becomes just noticeably warm.
When I tried to see the driver output voltage with an oscilloscope,
something strange happened. I see the expected bursts, but also a
continuous 1 MHZ oscillation appears. The problem appears to be local to
the driver, and the coil operates normally, as the high-frequency
doesn't transfer energy to it. Some parasitic coupling turning the
driver into an oscillator, but just when I connect the scope (?). Has
someone observed something similar? I have another primary circuit, with
higher impedance, that does not cause this.