A low power high voltage supply is a useful general purpose tool. It can be used to light up Geissler tubes and other gas-filled tubes, to diagnose failed diodes and triodes, and for anything else that needs a fairly non-destructive spark. I made one using a small high voltage module, and made a box for it out of leftover poplar from making the variable battery. The finished box can be seen above, next to the variable battery for size comparison.
Operation is fairly straightforward. Pressing the button on the left connects four D batteries in series to the high voltage module, which appears to contain some form of charge pump. The output of this module is connected to the copper pillars on the right, through which pointed steel rods are mounted. These rods form an adjustable spark gap, which sets the maximum voltage of the module. A tube or other device can then be connected in parallel with this spark gap, using alligator clips.
The device works well and can make a loud, crackling spark up to 1 inch in length. I connected a small neon bulb across the terminals and it appears that the electrode on the left (viewed from the button side) is negative. I then connected various failed triodes, and judging by the glow I discovered that the pressure inside the majority of these tubes was unexpectedly high (at least 10 torr, with one tube at almost atmospheric pressure). This leads me to believe that the failure mode of the majority of my tubes has been leakage, not outgassing as I had expected. I will need to improve my pinch seal methodology.
The one problem with this design is that the high voltage module I chose to use is cheaply made and tempermental. I have had two fail for no apparent reason, one of which was already mounted inside the box and had to be replaced. My hope is that if I limit usage to only short bursts, the current module will last longer than previous ones. The inside of the box can be seen below, with the high voltage module on the left next to the batteries. The wires are fairly close together, but the insulation seems to be holding up well.
I have since replaced the high voltage module with a more reliable one, seen below. The layout of this module also allowed for greater physical separation between the high voltage and low voltage leads.