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

Re: Scopes Away!

HI all,
        I must thank Greg for making me go back and scrutinize my 
notes more carefully. In what little data I did give there was an 
error. I stated the unloaded Q of the secondary as being 184. That 
was in fact without the terminal. With a terminal it dropped to 147
(bandwidth measurement). Which leads me to the question about the 
substrate zener in the MOSFETs.
     I recorded several discrepancies in my notes that could be 
attributed to this. The notes are meant to be complemented by the 
photos so I don't have full data available yet.
    (1) The decrement Q of the secondary was measured at 123. I 
remember puzzling over this at the time (vs 147 bandwidth).
    (2) The rings due to uncoupled energy were produced in a major 
fashion only when the quench was effected on one of the zero-
crossings of the primary waveform (from memory the positive going one
but I'll have to check the photos). Also, they were all positive with 
respect to ground (no clipping I think). Despite that, quench was 
effected cleanly on all portions of the waveform. Perhaps this 
shouldn't have worked but it did.
     As expected, the spikes got higher and higher as dwell time was 
cutoff more and more into the primary storage part of the envelope.
     I could detect the secondary ringing for well over a millisecond
when optimally quenched after the first transfer.
     I obtained half inch arcs with the power supply set to 30V, a 
current drain of about 1/4 amp and the charging choke and diode in 
place. The final Cp voltage I didn't record (on film) but I never
swung the supply above the Vds of the MOSFETs. The substrate zener
wouldn't have allowed it anyhow. This was with a repetition time of 
85uS. The dwell time is also on film. As I increased rep rates much 
more than this, I had to reduce dwell time. At the same time, other 
more complex reactions between the charging choke and primary cap and 
beat envelope were creeping in. I never got close to getting 45W into 
the secondary. I found that the "gap" was taking a lot of hammer with
dwell set to about 20uS and rep rate about twice that (50/50 duty 
cycle roughly). 
     With the terminal loaded by the scope probe (10MOhm//21pF - just 
checked the manual) I recorded a 200V peak ring in the secondary with
Vsupply = Vcap = 12V (and the 10k resistor feed). With secondary 
capacitances added to this I was getting loading around 37pF//10MOhm.
This was around a microJoule of energy in one half cycle of ring.
     Maximum Cp energy recorded was at 60V which with Cp=284nF came
to about 511 microJoules.
     In CW operation, the output amplitude soared when I was able to 
set the rep and dwell times optimally. At one stage I had to set the
pickup probe sensitivity to 50V/div to get the waveform on the screen.
For other tests it was set to 2V/div and 5V/div.
     I guess in summary, far from perfect. But it did work and I think
makes some of the waveforms accessible at least. I hope it's of some 
use to others.

Malcolm