@tango
This is the general idea. I do not think this circuit is operating at perfect resonance, but a large voltage gain is seen on the highlighted wire just above the far right 5pf capacitor. It takes several seconds (real time) to see it on the scope, but it would be nearly instantaneous if circuit was operated. That is just to do with the limitations of simulation time on the program with the pulse square wave I'm using.
Also, his specifications call for a pulse train (group of pulses separated by rest period repeated continuously). This is not incorporated into my circuit here, but can be simulated by releasing the switch to battery, whereupon doing so an even larger voltage spike comes off when the switch has been on for several seconds then disconnected (up to 4Kv depending on connection time).
As you can see, the starting voltage source is only 9v and in just a little while the voltage is up between 240V and 320V (far right capacitor) and the power consumption for a 9v battery is only in the milli-amp range. So power consumption is low to create a gain from 9v to between 240 and 320 V, with large spikes up to 4kV coming off when switch disconnected.
I'm sure this is far from perfect or ideal, but it provides the general idea. The capacitor on the right is the one with water as the dialectric, receiving the largest voltage gain and voltage spikes when switch is disconnected.
EveryCircuit - VIC replica 1st try