RE: Gap turn off at zero current and TVI

	If you think about it, TVI _HAS_ to stem not from the main
tank resonance, but from higher frequency parasitic resonance's.  I measured
a self-resonance in my primary coil in the 24 MHz range, but there are
probably many other, higher resonance's as well.  Quoting from my NST
protection web page (inspired by Terry's research findings):

Gap conduction is not really continuous, but ceases conduction briefly each
time the tank current crosses zero as it rings at its resonant frequency. At
the moment that the gap ceases conduction, the voltage across both the tank
capacitor and the primary coil are at maximum, or Vtank. The primary coil
has some parasitic self-capacitance, C-self, associated and in parallel with
it, and it is also charged up to V-tank. As soon as the gap opens, the
primary coil and it's C-self ring at their own resonant frequency, about 24
MHz in the case of my coil. This ringing is in series with the tank
capacitor, sitting at Vtank, so the combination of the tank capacitor and
the ringing primary inductor in series now present a voltage of 2Vtank
across the non-conducting spark gap, and to the NST if no protection network

	I suspect that this 2Vtank effect may be responsible for
reigniting the gap following its zero current shutoff.  It's unclear what
all of the ramifications of primary C-self are.  It may play a significant
roll in how the gap quenches, and it clearly affects VHF resonance's and
TVI.  If larger primary C-self's store more energy and are in fact
responsible for gap reignition, that would be bad for quenching.  I was
considering a new primary made from closely wound copper ribbon, perhaps
that's not such a good idea, as it would have a much higher C-self.  Hmmm.
It would be an interesting experiment to add a few pF in parallel with the
primary to see if it affects quenching.  

	Back to the original point, I think one can at best minimize
TVI by making the tank as compact and free of parasitic L's and C's as
possible, but some will always be present.  It may also help if all wires
to/from the coil (AC mains) have an RF filter on them.  Maybe it would also
help to add some ferrite filters (allows 200KHz resonance to pass but hi-z
at TVI frequencies) on the RF ground to keep that wire from being a

		Regards, Gary Lau
		Waltham, MA USA