Just thrown together out of some spare parts. +/-15V bench supply, an HV power transistor, a few coils and capacitors, a resistor and a bunch of jumpers. About 10kV worth, making a 3/8" long corona discharge.
The night of experimenting began last night when I thought I'd conduct an experiment I've seen online explaining how Nikola Tesla intended to provide wireless power transmission. It amounts to using a Tesla coil to feed energy to the electric field of the Earth/ionosphere system (since it's a continuous surface, the signal will circulate around the Earth and resonante around 80kHz as I recall). Well anyway I got out a TIP31C convieniently mounted on a heatsink, hooked up some coils, +15V and the signal generator and started screwing around. This particular coil is self-resonant at 1.5MHz (and with a Q of around 150 you can imagine how tricky it is to tune to it externally!) and I quickly got up a small bit of corona. Then it got late and I got really tired so I left it at that and went to sleep.
Today I returned with a 2SC3164 (some 800Vceo, 5A Ic and 20MHz Ft), upped the voltage (by using + to - instead of + to ground on the PS) and eventually set it up in self-exciting mode. No more constant adjustment! :D Aside from the plate, ermmm, collector tuning with a little variable air cap.
Oh, I suppose I should entertain you with a few pictures. These are with the camera feeding video webcam-style to the computer, capped at 640x480.
Low lighting showing the signal generator at 1.48MHz, the output sinewave on the scope (probe just floating in midair :) and the small discharge (the bright spot below the 'scope trace that looks like another panel light but is above the coils).
Closeup of the discharge. It mostly forks out in two directions, presumably because I cut the wire with diagonal cutters, leaving a sharp chisel shape on the end (with two sharp corners).
This is what happens when I reduce power slightly (usually by tuning it off resonance). It turns into a single, sharp line discharge that often buzzes or squeaks for why-I-don't-know. Oh, and all these discharges are actually purple. The camera just sees them as blue.
Hehe. I soldered a coil (about um 100 turns?) to this nightlight bulb (you can see the top of the coil). Alone it will give a hazy cyan-white discharge inside the glass but with the coil I can heat up the filament. Wireless power, anyone? :)
With the coil deeper inside I can extract more power. This is about 3W (the bulb being rated at 4W), although it looks white to the camera. About equal to 100VDC across it I think.
Needs a bit more voltage to be a plasma globe so to see any slow argon discharges I have to drag the corona over to a point inside this chunky 300W bulb which is near the metal bits inside to start an arc, after that I can drag it around on the surface. It looks a lot more graceful than the blue-white 'flame' seen here.
I've also tortured a crystal-oscillator-in-rectangular-can style chip with this, the most fun part being hanging it from some masking tape and letting it soak down inside the secondary for a while. It acts as a shorted turn (causing what is otherwise known as eddy currents), sucking down power and getting hot in the process. After five minutes it's more than hot enough to go TSSsss.. when you touch it with a licked finger. ;->