Plasma Cup Inverter

the plasma cup in action

I ordered one of these things from Think Geek a while back, they had them quite cheap.

A pretty cool toy, but like any good geek I just had to take it apart to inspect the guts!

Externally, the cup/envelope looks like soda/lime or flint glass, I don't believe it is Pyrex or any other kind of exotic glass, its density is pretty low. The envelope is quite thick and has withstood a few accidents, but I would imagine it would break pretty easily if dropped. The evacuation nipple is very large, but there is really little other place it could be. The spectra shows Neon alone, there does not appear to be any Helium or Argon present.

The base is some kind of black polymer, probably PE. There is a power switch on the back, and a DC power jack as an alternative to the many AA cells needed to run it. The depression in the middle centres and hold the glass well.

Four screws hold the case together, they are hidden under the rubber feet on the bottom of the unit. The feet give it good traction on the table top and won't allow it to move around once placed.

Once the screws are out the top lifts off revealing a small circuit board on one side. A short wire connects it to a plate held to the top shell by four plastic studs that are melted and mushed over during the construction process. I didn't remove the plate (a piece of circuit board) to inspect its top surface, but I assume it is just a sheet of unetched copper, not a spiral antenna or anything more complex. The battery wiring is tacked in place with hot-melt glue. The DC power socket is switched to disconnect the internal batteries like any good ext-DC socket should.

the guts of the device

The circuit board is quite high quality. Single layer, with a silkscreen, solder mask, and nice wide tracks where it counts. The power device is heat-sunk and the transformer is clearly a custom made device for the application. The resistors used appear to be 1/8W devices, they are very small. The soldering isn't too bad at all, apparently hand done, but the flux isn't cleaned off (not a huge deal - looks like one of my projects :-).

Tracing the tracks and inspecting the transformer windings with a multimeter reveals a circuit similar to this:

plasma cup inverter circuit

I did not remove the transformer from the circuit to do phasing and ratio testing, the secondary is definitely an tapped winding and is where the feedback comes from. But it may be that the tap is grounded, it is hard to tell because the resistance of the lower half is very low compared to the upper half.

The circuit is clearly a power oscillator using a moderate power LF transistor and a minimal component inductive feedback topology. The output frequency pulls with load, but is around 100-150 kHz. I did write down how much power it pulls at various voltages but I've lost the piece of paper!

You can pull a very small RF arc off the output wire-to-plate solder joint, the energy is very low, not enough to burn you unless you really try hard. There is a considerable RF field around the collector of the transistor (and heat-sink), which can light the glass fairly well on its own. I noticed this before I opened the device, scanning the entire surface of the box with the glass looking for hot-spots indicative of internal wiring or components.

That's about it really. It is a fairly unremarkable LF power oscillator with capacitive coupling to an envelope of Neon gas at low pressure. The only disappointing thing is that the plasma doesn't move around like it does in a plasma ball/lamp, it just stays still and glows Neon orange. The extra loading of a beverage in the cup pretty much confines the plasma to the lower half of the glass unless and external DC supply of slightly higher voltage is used.