Tank cap charging

["sundog" <sundog-at-timeship-dot-net> (by way of Terry Fritz <twftesla-at-uswest-dot-net>)]

  Hi all,

     In case you're wondering, Yes, I have quite a bit of free time while at
work (it's my job to sit and wait for the network to break so I can fix it,
but thanks to Unix, I have all this time ;)  Unfortunately, they'd frown
highly on me soldering on my desk, or test-firing rotary gaps, so I think.
And question.  But I digress...

   The high bps rotary is an interesting thing to tinker with.  I plan on
converting the charging circuit to DC (full wave rectification) to hopefully
provide a smoother supply for the caps.

   But even with full wave there will be some ripple when the voltage
crosses 0.  How large of a stiffening cap do I need (Voltage &
capacitance-wise) and would an MMC from the 1.2kv .220uf panasonic caps
stand up to that abuse?    Being on the transformer side of the rectifier,
they should only really see the 60hz signal, right?

   I was planning on running 2x the peak voltage per "diode" (using a
multi-mini-diode), but how well will they handle the kickback from the gap?
Also, using DC I'll eliminate resonant charging, so I'll be able to run a
manual gap and measure the voltage across the cap.

      Do I have to remove the bleeder resistors for DC useage?  If I'm
constantly pumping the caps up, won't the bleeders try to dissapate the
energy?  While I am not bothered too much by the burning of the resistor
itself, the smell is horrible.  I intend to measue the voltage across the
cap and see for myself how long it takes to bleed down.  But, should I
measure the voltage from the positive side of the cap to ground, or to the
negative side of the cap? (using a voltage divider of 10mohm resistors)  I
usually put them all in series, and measure the voltage across the last
resistor closest to ground, keeping the total V measured at the resistor
~200v.  Normally I hook the divider to the source and then to ground.  It's
worked for the GMHEICSLR setup (4/30 nst, 25 resistors in series from the +
side of the rectifier feeding the cap.  V goes from 0 to ~225v and then the
NST output peeters out (fully charged cap to ~5656v, no bleeders), then
*BANG* when I short the gap.  The ratio works out to 25:1, so I can pretty
accurately charge my cap to whatever I like.  A string of resistors between
the rectifier and the cap keeps the charging current down so I can just bump
up the variac till I hit the voltage I want, then spin the variac back down.
It's served me well, and requires nothing more than a bunch of resistors,
and an analog (safer from kickbacks) multimeter.

      I'd love to buy a surplus HV probe, and use a scope, but mine is dead,
and money for one is hard to come by.  That, and I can work a multimeter a
whole lot better than a scope ;)
									Caio!
									Sundog