For the oscillator, I looked at the 555 timer, and designed a circuit that generates a 72 Hz square wave signal with a duty cycle of 50%

Note that is a really low frequency to operate a CW multiplier and would require about 1uF of capacitance per stage, even with let's say a modest 5kV input it would be huge and expensive. For a better visual look at the size of a 2kV microwave doubler capacitor at 60Hz. The small ferrite core transformers I assume you are looking at will better operate in the 20kHz range.

Your CW multiplier will still work with pulsed DC as opposed to AC sine wave but you can also buy square wave alternating -+ modules like so https://www.aliexpress.com/item/1005004977783673.html?spm=a2g0o.order_detail.order_detail_item.5.1b7ff19crJk8B4 if you want sine wave AC its easiest to use the ZVS drivers like Milstar suggested.

I've seen a few times that adding "resonant capacitor" in parallel with the primary coil will result in improved performance. Now I don't know what is meaned by "improved", be it more efficient or higher voltage etc. I don't understand the physics of it but I ran a test anyway.

Nothing of practical interest was obtained, I plugged the inductance and frequency into an online tank circuit calculator 16kHz, 100uH = 1uF capacitor. Well, it definitely had an effect just nothing useful.

The transformer was just a 1:1 ratio on a small ferrite core. Driver was a 12V variable duty 16kHz square wave.

12V 16kHz 5% duty turned into low voltage low current sine wave. Interesting but with a 0.08A short circuit not good for much.

12V 16kHz 80% duty gave a drop in voltage and very high current draw, it killed 2 caps before I could photograph it. FETs and caps got very hot quick.

Is there any actual purpose to adding a capacitor to the primary input?

/>
This is a mosfet driven from a signal generator which seems to work quite well, at least for ignition coils.

I have an older version schematic as attached, but am wondering what the improvements in the video one are? I see a capacitor has been added but where and for what purpose?.

This is the only other one I've found with both variable freq. and duty. https://www.elecbee.com/en-31968-YF-22-PWM-Dimming-Speed-Controller-Module-Frequency-Duty-Ratio-Pulse-Adjustable-Square-Wave-Rectangu?search=frequency I've ordered but isn't here yet.

The above didn't work, set to 16kHz 50% it was a weak driver that quickly went up in a puff of smoke.

I ran another test,

Ordered

1 x    AT2078/20941=BSC27-5906=JF0501-21835   $10.80
1 x    BSC26-01N4035=JF0501-21953                           $9.60

I received a JF0501-2135 (same thing?) and correctly a JF0501-21953.

It's actually pretty legit. Unless they are just adding printed labels to generic transformers haha.

For arguments sake, if you could have a "universal driver" would you prefer a variable square wave +10/0, an alternating square +10/-10, or a sinusoidal wave? or other?

I've noticed some square waves perform very well like dimmers, and some not at all or extremely weakly like motor speed controllers. I've tried numerous types and I'm not sure why.

Thankyou Mads, it seems to perform well. So its a 12v driver that switches rectified mains like the RMcybernetics one above but looks to actually give a square wave.

What limited its lowest frequency to 38kHz? Is it a huge leap to get that down to 10kHz or so?

Oh, sorry swap the mA for uA, giving 200uA meter in series with 1G resistor for a 0-200kV meter.

200uA digital meter https://www.ebay.com.au/itm/382508507644 , yes I hadn't thought about the interference problems, but I guess I'll test it anyway. Only the digital meter will fit where I need it. Wondering how ripple will affect a digital meter compared to analogue too, I'm assuming it will require full wave rectification and capacitor.

I've been trying to upload pictures but SnapDrop is down so I cannot get it off my phone. I've been testing a 50kV analogue meter with 500MOhm resistor as bought off Aliexpress. It came with the resistor.

It doesn't work as planned, Attached to a 20kV source (as determined by probe and 20mm spark gap) it reads 10kV. Now if I remove the analogue meter and attach the probe to the end of the resistor it also reads 10kV. So the resistor is halving the voltage :/

Yeh I think you are right, I hadn't bargained on this being so hard. I thought I'd be able to switch the 30kV before the multiplier but attempting to draw it out wasn't so successful at least not with a single DPST relay. The only sensible solution left is swapping leads around, if you switch it at the 150kV side but leave it energised it'll still be dumping corona.

Yeh I had noticed the right hand side was going to be even more a problem then the left after posting that. I was hoping there was a solution that I wasn't seeing, but I guess there isn't.

A >30kV vacuum relay is looking to be about $2k, not really feasible at this time. There's a few open 40kV used Ross relays getting about for not much cheaper and generally 115v, not practical here. Making one may be an option but not one I'm enthusiastic about.

I've no practical purpose in mind for this I just want it.