Designing the plate capacitor to be 20Kv rated, I come up with a dielectric of .050" plastic and air gaps of 2/3 cm.

So I tested .0078" aluminum separated with .0625" polyethylene and 1.125" air gaps and came up with about 38 layers to get the desired 20nf capacitance.

However, when I mocked up 3 plates (1.5nf) and hooked it up to the 10Kv rms OBIT, I got arcing through the air gaps. At 30Kv per cm of air, that comes out to 84Kv. WTF?

Not every time. Simply plugging in the OBIT and leaving it seemed ok with just a buzzing noise and no apparent arcing. It was when intermittently powering the OBIT (say 3 times per second) that the problem occurred.

Anybody know the cause and/or how to correct it?

How worried about it should I be if the goal is a good-running SGTC?

I'm looking to shoot 16 to 24 inch arcs in random directions from the top load continuously, and am fairly convinced that a spark gap build is the way to go. My first step is to build the power supply, but I don't know how big it needs to be.

The OBIT says to deliver 6KV and 35mA (RMS) continuous. Can I expect to need 1,2, or 10 of these in parallel to meet the objective?

Is the popular dual MOT rig enough? Can I short circuit protect it with a dimmer switch?

Thanks in advance.