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

Primary Qs



Hi everyone,
             After some discussion on this subject and (as usual) 
some misgivings about past measurements, I decided to take some 
more to check. A resistance of about 150kOhms was used in series
with the DC supply. Measurements were taken using 2 coils and 2 
capacitors. A 50Hz bandstop filter was used to block stray mains
waveforms from the scope front end.

Circuits :  11.8uH/100nF    350uH/100nF     350uH/25nF
Fr       :    144.6kHz        27.15kHz        38.5kHz
Q (3dB)  :      35             54               64   
Poor isolation shows itself. Capacitor quality is not great either.

Measurements were then taken at 4 different voltages for each 
combination:

Cap Voltage         3kV    6kV    9kV    12kV 
11.8uH/100nF (Q)    8.2    10.8   10.9   11.5
350uH/100nF   "     11.5   13.7   13.6   14.4
350uH/25nF    "     10.1   13.0   14.4   14.4

Max number of rings before gap extinguished : 32 for 350uH/25nF -at- 12kV
Min    "                                 : 15.5 for 11.8uH/100nF -at- 3kV
   
Q does trend up with increased gap current showing a drop in gap 
resistance. It would be useful to get readings for much higher 
voltages but the supply and caps on hand wouldn't allow it. 

More interesting stuff: Gap at 12kV = 0.16"  (160 mils), (4.07mm)
This translates to 75kV/inch on a linear basis but there would
be few if any "stepped leaders" from these electrodes :)
Gap electrodes are 1/2" diameter rounded tungsten carbide tips.
It would probably fire a couple of kV higher if a fast rising 
waveform was used based on past experience.
    
    The gap _always_ extinguished on a waveform peak (when Ip = 0)
As Greg said, the gap does look better than it has a right to. The Q's
are generally disappointing though. From the lowest L/C combination
at 3kV to the best at 12kV, there is less than a doubling of Q.

I guess the recipe is : high L/C ratio at a high voltage for lowest 
losses. There is hope yet.

Malcolm