Hi Peterae,
I like thinks like this! The faster the diode the more unidirectional that pulse will be.. ie. less times the cap will ring through the arc in the sg. When the diode does its job you kind of have to sit around and wait for the 75ns to elapse before it is ready to cycle again. lol..
I had always wanted to put say five diodes in parallel and in line with each diode put a resistor with a slightly different value than the next. R , R+1 , R+2 , etc...
The proposed result being the lowest value resistor R would focus cycle 1 through its diode holding off all the other diodes then say 5ns later the next pulse would see the first diode still trying to recover so it would hit the next path of least resistance R+1 and hit its diode...
Sort of a diode cascade.. I dont know if it would work but it may be worth a try..
|---WWW---->|------| R |---WWW---->|------| R+1 |---WWW---->|------| R+2 |---WWW---->|------| R+3
|