Ursa II. Power supply 1834.0 volts, continuously monitored as well as regulated. I used a negative supply with the anode lightly loaded and put into the input of the URSA unit, so no variation from possible drift in the dynode chain resistors in that current (other than if they did drift, the gain of the tube might have changed, but the current was constant and the voltage stayed within .5v the entire time). I had experimented with various supply voltages and gains in the URSA to get the skinniest lines at the higher energy Cs peak. Some pics and words here:
viewtopic.php?f=11&t=337 Yes, I know it's a terrible setup for real measuring...but nothing there should cause drift.
I only had two point cal with Cs-137 for the drift testing, and with the big tube, the lower peak was just out of tube noise with this huge crystal, it took some fiddling with the threshold and other parameters to make it clear and obvious out of the sort of "1/E background", but I got that to happen on both of the big guys we have here.
It's interesting that I also have several other NaI's but the only other type that seems worth it really is one I have with a very tiny crystal, thin window, built in preamp. It really gets that low energy stuff clean and neat, but of course fails miserably on the higher energy stuff that comptons out of the tiny xtal. The midsize ones I have, about 1" sq by about 3.5" deep seem like a compromise that gets everything wrong. For that matter, one we have 1/2" sq by ??? deep seems as good or better than the 1" guys. I've not yet hooked up one of my homebrew BGO's or plastics to the URSA, I suppose I should. Even with the lousier resolution the amount of drift I'm seeing should show and tell me which end of the signal chain it's coming from.
I'd like to kill off that drift so I can do neutron activations to measure other things with some confidence (minerals, soil samples)...gotta get the test gear around 10x better than the thing you're measuring, isn't that the good old rule? And it's not like a random gamma spectrum is all that easy to decode - it's at least as bad as a mass spectrum with "illegal" chemistry that can exist long enough to register a mass line funny stuff like COH, H3 and H3O all show up in those, stuff that "can't exist" in normal conditions. I don't mention it much but one of the fusors here is plenty hot enough to activate "you name it" as far as something as sensitive as this is concerned. Dunno if I have any "record output" but it's in the millions neuts/second somewhere. Still don't have anything but a relative calibration on that off Richard's data. Comparing silver activations, his 20-something minute run at 2009 HEAS did silver to 497 cpm, a good 5 min timed run here does the same size sample to the 1500's -- we are both using that nice pancake counter we got from you for counting those, but we plan to substitute our new standard counter as a reference for that, it will be a little less sensitive, but the tube we got for it (a Russian 3 element tube, claimed to be beta optimized) seems much more relatively beta sensitive than the Geo detector -- so it counts less background but seems to count about the same on a beta emitter in tests here. Yet it only counts 3/5 as much on a pure gamma source. Of course, that means they won't read the same ratio on all possible sources -- which was the point behind coming up with a standard counter design in the first place - so we could get good cross-lab comparisons and calibrations.
Posting as just me, not as the forum owner. Everything I say is "in my opinion" and YMMV -- which should go for everyone without saying.