Is there a reason why have to be alternator with brushes?

Resonance with 2.5V input.

120kV P-P

If you were following my messages earlier, I said the sizing of original spoked electrodes anode with area dimensions , came from taking the Patent cell drawing  to an office copying machine with ruler, and doing successive enlargements , so as the TR2 metal core was exact 3/4", as stated in Patent". This way the other dimensions can be revealed. Are you building a cell+ power supply also?

What's important is the "time line"-  Speed of charging as faster than the discharging frequency of capacitor discharges. 51Khz full wave DC ripple frequency has period 1/F. It takes 2 pulses to fully recharge cap in between discharges. Each discharge of Cap is same, - .054Watt. Times 16,700 cycles for separate discharging oscillator to get the 900 Peak watts. Use scientific calculator with formulas. There has to be a "window" of time microseconds after the pulse stops, for the available off time , so as complete recharges happen for 330vDC. A feedback loop wiring and zeners + optocoupler as hooked back to the Master IC, SG3525 keeps voltage regulated, so as parts don't burn out with over voltage. 23.7usec cap stops through diode of primary, , with difference to 60 useconds of the discharge oscillator available off time. The 22 amps pulse is under 30useconds and then off to the 60usecs. This is reflected to the TR2 secondary side as POSITIVE DC pulse with off time in the water. 1/4 wavelength is used in formula for cap discharge time as primary has diode. The charging section at other end has about 11.3 uhenry inductor and series diode also , so as just 330 volts is applied and not 660v. .Charge time was 10.6 microseconds in 1/2 wavelength. Separate digital DISCHARGE oscillator  module set duty cycle about 30 microseconds. You can also send short questions to Grok AI, but need to repeat with different wording so as a better reply is made. . A computer cannot think, just relay tech data that's all ready out there.

the powersupply of horvarth was made to dump pulses.
The only way that such a cell can consume high amp pulses is when a higher voltage is applied and or if the water is higly conductive.Horvarth talks about hundreds of amps.
Thats the challange here...

Just finishing up the built power supply for the Horvath Patent Hydrogen Generator.  It includes the efficiency methods added , as in the Patent. 900W peak pulses into Primary TR2 transformer for positive DC, #12gauge magnet wire,  #6 gauge cable for secondary side.  50% duty cycle with off time , for the cell to be built later. Pulsing on + off dc capacitor pulses at 16.7khz, N + S magnetic fields to speed ions transport, and 12 spoked / flutes/ ribs, as roughened with machined knurling. US Patent  # 3980053. More modern components were used to replace technology of mid 1970's. Battery power draw as about 450 watts as due to the off time , designed in, for  the later used aquarium water cell  .Is there anyone interested in this method self powered engine on Hydrogen. I have Briggs one cyl and Geo Metro 3 cyl engine on engine stand. Questions?