Bill Fain has kindly arranged to obtain 9 new Russian pancake geiger tubes for us. I am wonder if there is interest among the fusion oriented members in my building some standardized, cross calibrated geiger counters for us, so we can all compare things like silver activation using known sensitivity equipment. Since this is very early stage (they haven't arrived yet) there is some flexibility in the design at this point, and of course a ton of extra features would cost somewhat more to build up. Here is what I am thinking about as a straw man, so anyone interested should reply on this thread with comments -- if you want more or less features, and so on.
Basic idea:
Geiger tube, power supply, microprocessor, LCD display, rs-232, runs off wall wart.
This should give the basic needs. Depending on how easy it is to refactor some code from our fancier multi-geiger project, the code might contain variable measurement intervals, setups that are saved across power downs, and background collection and subtraction. The basic mode I've found most useful here is 10 second counts, normalized to a minute (eg multiplied by 6 for display of cpm), and no background stuff. Our background is so darn variable with cosmics here that subtracting it is kind of a waste of time, since our lab is also very quiet as well. So I never use it myself, I just take an extra count before and after, and kind of visually subtract the average. When counting something like silver or indium, it's kind of obvious if a cosmic ray burst spoiled one count, which is a pretty good reason to do short, 10 second counts, actually. Also, with silver, the decay rate is on the quick side, especially at first (metastable states) and you miss the beauty of that with a longer count interval. On the other side going too short (1 second say) is just too short and might have fenceposting issues when the counter is read and reset in software, so 10 seconds has been the all the time mode here.
Things I could add are battery input -- it won't be too much a pig, but may need 7 volts or so (5 cells) to work correctly, the rs232 converter needs real 5 volts, though the CPU (PIC 18lf2325 doesn't. I find rs-232 very worth it here, as almost any PC has some kind of software that will slurp that into a file for you. Typically, we give a timestamp, the count, interval, and any other data on the interval -- all in human readable form, and easy to parse in other software as well.
Another thing I might be able to add is a couple of a/d channels if I don't run out of pic pins for that. One will be used to sense and regulate the high voltage for the geiger tube via the pic's pwm output, but I might be able to get 4 if the pins are not all used up (the LCD takes 8 for a two line display run in 4 bit mode). When I do that, what I actually do is take data at around 50 samples/second, and sum them all up for reporting once per second -- you get some noise averaging and maybe another bit or two accuracy that way. The basic a/d in this pic is 12 bits, we did this trick with a 10 bit one, and due to the inevitable noises, got 11-13 bit resolution out of it doing it that way. The noise in this case acts like "dithering" for those who speak signal processing. I've been putting that out raw, but I suppose it is possible to input a scaling factor to make it volts. Our basic UI uses a rotary/push knob, so putting in floating point numbers with that is problematic, though it could be accomplished at programming time, or perhaps with rs232 input from a PC.
You tell me. The basic thing I have in mind would have to go for about $225, plus or minus say $30 (US) for me to break even, more if people want a board made by a pro, instead of what I do there (that adds a setup fee from the boardhouse).
Any interest, or did we just get a bunch of nice geiger tubes for ourself alone?