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

THOR PROJECT: SUCCESS!!!



Hello all.

After the last power supply breakdown I talked to a guy from Intersil: he told
me that if the IGBTs fail with gate, collector and emitter shorted it is likely
a gate-emitter overvoltage condition causing of the failure.

Therefore, I decreased the IGBT drive voltage from 17V to 15V, added a fast
over-current protection limiting the IGBT current to 33A and added also a
couple
of decoupling capacitors on the controller board (the one generating the drive
pulses to the IGBT bridges).

As the design is modular (two 5kV modules at the moment) it also took a second
to wire it to give +/- 5 kV referred to ground, instead of +10 kV referred to
ground. I had thus a simmetrical power supply. That halves the HV stress on the
power supply transformers.

Result: this morning I had several 30 second runs with the power supply
connected to my old TC.

IT WORKED FINE! The power supply was NOT damaged! Long, steady ramificating
streamers!

Pulse voltage was 11 kV DC, repetitions rate was about 660 Hz: that makes a
power of about 2kW (with the 0.05 uF cap I had).

I also measured all with a scope and printed a few snapshots: In a few days
I'll
scan them and put them on the web.

Next step: modify the PROBE board to measure two voltages (positive to ground
and negative to ground), sum them and also subtract them. The sum will be the
net output voltage measurement, the difference will be an alarm signal warning
if the positive and negative outputs difference crosses a safety threshold (if
the two module outputs get too much unbalanced).

That will include also embedding the HV divider networks into epoxy and a new
PCB layout design for the PROBE board: maybe another month all together.

Regards