This is still at the theory stage, but I thought I'd run it by the many knowledgable members here.
1st, I favour the Stanley Meyers tube in tube style.
We all are aware that lots of amps at the optimum volts, produces more gas.
Now the Meyers unit had multiple tube pairs, all connected in parallel, driven by supposedly High-Voltage 20 KHz pulses, using plain water, no electrolyte. (I quite like the idea of No electrolyte.)
My theory is to Not connect the tube-pairs in parallel, but to only common all the +Ve center electrodes, via a large inductor to the fixed +Ve supply. All the outer -Ve tubes having individual connections, wired through the bottom of the cell (similar to Meyers cell)
After testing a single tube pair, with a gap of about 2MM, to find the optimum pulse frequency, voltage & drive current, Hopefully, this would be a High Voltage, at lower current?
Now, hopefully, having found this 'ideal' combination the electronics would consist of distributing this power to all the tube-pairs, lets say Ten tube-pairs, a simple decade counter, with a variable clock, would distribute this power to each tube cell in turn, at a fairly high frequency, via N-Channel Power Mosfets and possibly via a ferrite Choke, to tune to the reasonance of the cell's natural capacitance?. This may end up as straight square waves? or using a similar circuit to send 'gated' Pulse bursts to each cell at a different frequency, hopefully producing a useful harmonic to rip the water apart? this is fairly simple, in terms of just the electronics, but needs to be evaluated.
Now I can guess what you may be thinking, I only have one cell energised at any one time? but research into the Meyer cell seemed to imply that cell 'recovery' time between the pulses was important, this method would achieve a period of recovery naturally, due to the sequencial timing period, also, each burst of a square pulse, (or gated pulses) would shock each cell with some force, hopefully shaking off more Hydrogen?
A possible advantage, the 'Individual' cell tubes are smaller in area than the sum of the usual 'plate' types, these should accept a much higher voltage and lower current than when using Sodium Hydroxide due to the higher impedance of each cell.
I'm not sure if it's feasible that the sequencial nature of the applied power, in this rotary fashion could enduce a resonant stirring effect?
The total power required should, in theory, be lower than conventional cells? Possibly 20V to 50V, but at lower current, possibly only about 1 Amp? Using a bog standard step-up Switch Mode PSU as the main power source, this would still run off the more normal 12V Battery.
I would be very interested in your thoughts on this approach, yup, a bit bizarre, driven by my aim to avoid electrolyte, mega amps, and running cool

Steve