2009-06-07
Pancake Geiger Müller Tubes and a new Power Supply
I found some very nice ~ 50 mm pancake Geiger Müller tubes on eBay. They were about $60 AUD landed from the Ukraine, so naturally I couldn't help myself, I bought a pair.
They are fairly modern devices, built in the 90s with a Neon/Bromine fill and an operating voltage of about 400-500 volts. Their end-window is a bit thicker and more alpha absorbing than the one in my old high school tube (likely for the required mechanical robustness to not implode), but their overall sensitivity is much better because they are physically larger.
With the end window being quite transparent you can easily see the discharges in the tube caused by the ionisation radiation. While not particularly localised (the plasma sheath expands rapidly by secondary ionisations in the avalanche) it is a bit like a spark chamber, showing you roughly where the particle passed through the detector. I quite enjoyed observing the flashes in the dark, it is a bit like a spinthariscope, only with much brighter diffuse pink flashes. I tried to capture the effect on video, but wasn't very successful.
Initially I used the circuit I built for the previous tube, but I decided to build a smaller portable version I could use a "survey meter" - there are some Coleman lamp mantles at Big-W I want to determine are Thoriated or not... A new inverter was constructed on a scrap of PCB material. I tried a different 10 mH inductor from the junkbox this time and found it worked better than the unit I had used previously. This let me use fewer multiplier stages for the supply, the circuit ending up virtually identical to Charles Wenzel's, except I used a protective gas-gap tube (from Rockby's clearance sales) for the regulation string and picked off the voltage at the last multiplier stage. This gives my unit excellent regulation and I lucked out that the tube gives almost exactly 500 volts in the circuit. (Open-loop the device delivers 783 volts.) The 3-pin transient protection tube is designed to be placed across telecom lines and features a 350 volt breakdown from each line to earth. It just so happens the breakdown from line to line is about 500 volts - perfect for my purposes. I originally grabbed them for a crazy idea of using them in Marx generators as a substitute to spark gaps or avalanche devices, but I am yet to try that.
To measure the very high impedance output of the supply I threw together a simple instrument using a 1 GΩ resistor and a commercial 200 mV FSD LCD digital display module. The module came from Rockby and has an input impedance that exceeds 200 MΩ, but I am shunting it with 10 kΩ so in relative terms it is quite insignificant. The 1 GΩ resistor I got from Farnell, and cost a ridiculous $8.40. It is rated to 5 kV, but I've tested it to beyond 15 kV. I arranged the meter decimal point so it reads directly in kV, forming a 20 kV FSD meter, but I wouldn't trust the single 1 GΩ resistor at 20 kV. I've used it to measure 15 kV sources however. I killed one of my favourite multimeters trying the same thing at 30 kV, so I am hesitant to push my luck much further until I build a similar device with good insulation and protection devices across the divided output.
Anyway, the resulting radiation measurement instrument is not calibrated in any way and offers just a piezo clicker for registering tube counts. I intend to add a count-out jack and probably interface it with a microcontroller driven LCD. The tubes came with sufficient information to roughly calibrate the device (about 200 counts per micro-Röntegen for Cobalt-60 gammas).
It fits the bill for a compact activity sensing unit I can take down the mall and check out the Coleman mantles with. Here is a video of the unit clicking away to the beta emissions of the small Potassium 40 content of natural Potassium Chloride.
The sensitivity of the tube is quite excellent. Without integration the emissions from the KCl are easily missed when using the old end-window tube. With the pancake tube they are quite obvious to the ear alone. The supply will work just fine with the older tube which is handy, and I've re-plugged it with a BNC to facilitate quick swapping.
12 comments.