2009-04-18
Ionisation Chamber Radiation Detector
This is something I've been meaning to try for quite a while. Radioactivity has become a recent area of study, so this unit was thrown together. I half expected it not to work, back of the envelope calculations suggested I'd need a very large number of ions produced per second to produce the base current required... but it works quite well.
The circuit is basically Charles Wenzel's nuclear war detector. I didn't have Darlington devices in a single package, so I just used discrete transistors to implement the NPN pair. I used an MPSA18 for the device connected directly to the chamber and generic transistors for everything else. The "switching" device for the piezo buzzer is a 2N7000. The chamber is an eclipse mint tin, the collection electrode is a circle of tinned copper wire insulated from the chamber wall with a glass bead. The connection is made to the MPSA18 base in mid air.
As radiation detectors go, it is insanely insensitive. If this thing ever goes of when you aren't provoking it artificially then you are in rather serious trouble. The 37 kBq Americium 241 source I extracted from a smoke detector will activate it from about 1" if you leave the lid of the tin open to let the Alpha particles in. Alphas are very effective at generating ions. As a comparison, at the same distance from my Geiger counter's Mylar window I only get about 180 CPM (and about 22 kCPM point-blank).
It is also sensitive to ions produced by combustion. Lighting a butane torch or a match near the chamber will trigger the alarm. You can also blow ions into the chamber from a gas flame or near the Americium source with your breath or by wafting, which will also trigger the detector. Naturally touching the circuit high-Z points will also set it off. Putting the Americium source right up to the transistors or heating them with a hot-air gun or contact with the soldering iron will also cause an indication.
The battery is fairly flat (measures 6.8 volts) and came from smoke detector that the Am-241 source was removed from. The circuit performs a little better with a fresh battery, but the main improvement is the loudness of the piezo and speed of recovery after radiation/leakage is removed, not its absolute sensitivity.
I was surprised I could get away with building the circuit as unshielded as it is. I suspect not being mains powered and being fairly compact helps. You can hear a little mains interference on the 'edges' of its turn on and off points, where the MOSFET is in its linear region and amplifies the noise seen across its base resistor. It is quite mild however and needs careful modulation of the alpha flux to maintain it outside of saturation or cut-off.
All in all, it is a pretty cool toy. Quite amazing really that conventional transistors have sufficiently low leakages to make this practical. I want to build a real ion chamber now! I'm having trouble finding the high-value resistors required. I did find 1 GΩ resistors at Farnell for some insane price and have ordered a pair so I can build a device capable of measuring larger resistances - hopefully I can build my own 100 GΩ - 10 TΩ resistors.
5 comments.
Attachments
| title | type | size |
|---|---|---|
| Toy Ion Chamber Circuit Source | application/postscript | 12.278 kbytes |