Cheap Sensitive Radiation Detector
This project started out as a gag project, a "Nuclear War Detector". The idea was to make a device that would flash a little red light if a nuclear attack were underway or if there were some sort of massive radiation spill right outside the house. It would be like a rain detector built into an umbrella, perfectly functional but totally useless. The sensitivity would be on the order of those "survey meters" you often see on eBay with scales from 0.5 to 500 r/hour. If the meter ever moves on one of these, you have a problem! They should probably be calibrated in "hours left to live"! I decided that a small ion chamber with a Darlington transistor current amplifier would serve as the detector.
But when I hooked up a Darlington with the base directly connected to the sense wire, there was almost no collector current. I expected some leakage with a floating base and a gain in the tens-of-thousands. I don't know if all NPN Darlingtons are as good as these MPSW45A but the leakage current was amazingly low and the gain seemed very high, perhaps 30,000 with only a few tens of picoamperes base current. (I tested the gain with a 100,00 megohm test resistor connected to a variable supply.)
Suddenly, it seemed possible to use these ordinary parts to make a really sensitive detector. I added another transistor as shown below:
Who needs bias resistors! The can is a 4" diameter tin peanut can with a hole in the bottom for the wire antenna and aluminum foil over the open end. I quickly learned that a resistor in the base of the 2N4403 (10 k) is a good idea to prevent damage if the wrong points get shorted. The performance of this circuit was amazing, easily seeing a Coleman lantern mantle! So why not add another Darlington? It seemed ridiculous but here is what I built:
I actually used 9 V but I recommend a somewhat higher voltage to get sufficient potential in the chamber. The resistors were added to protect against inadvertent shorts that can quickly zap a transistor or a current meter; in normal operation they don't do much.
This circuit actually works well and, after giving it 5 or 10 minutes to settle down, it could detect the lantern mantle from several inches away. The circuit is temperature sensitive and the meter will move up scale with a slight increase in room temperature. I decided to add temperature compensation by constructing an identical circuit but without the sense wire on the base and take the output across the two circuits' outputs:
It looks a little involved, but it is really quite simple to build. The circuit was built into the same peanut can used in one of the JFET projects above and all of the parts were wired on the terminals of the 8-pin header. The observant will notice that I actually used 2.4 k and 5.6 k resistors, but the values are not important. I should have used a bypass capacitor across the battery, maybe a 10 uF. The pickup wire is directly connected to the base of the transistor and goes through a hole drilled in the bottom of the peanut can. The circuit is pretty sensitive to electric fields so it is a good idea to have a circuit cover like this. A mint can could serve.
Give the circuit several minutes to settle after applying power and the meter reading should drop to a small fraction of full scale. If the meter goes negative, switch the pickup wire to the other transistor base and reverse the meter leads. If the voltage on the 2.2k resistors isn't low, maybe under a volt, try cleaning everything with a solvent and thoroughly drying it. Once the reading is low and steady, hold a radioactive source like a lantern mantle near the foil window and the meter should quickly climb. The meter can be a digital voltmeter with a 1 volt scale or a 100 uA current meter. The meter shown below just happens to have a radiation-related scale and the reading near 2.2 is due to the lantern mantle.
This is a simple detector considering its sensitivity! The adventurous experimenter might try other transistors, perhaps a Darlington made from MPSA18 or other types or maybe a transconductance op-amp like the CA3080 operating open-loop.
Schematics on theme: