Some standard counter results

This is bound to get mixed up with things in Electronics, check both. Physics-specific stuff here, mostly.

Some standard counter results

Postby Doug Coulter » Sun Aug 07, 2011 8:00 pm

Well, the standard counter project is well along, only a few more things on the TODO list there (minor improvements and such) left. I got some results testing it that might make it a little easier to understand the nature of radiation and it's nature of both being random and quantum, so I thought I'd present those here. Here's a plot of a few minutes counting background (as usual, click to make it big).
Screenshot-4.png
Background from standard counter
On the left we see a plot of some output, on the right the code, a terminal, and the standard counter software that shows the output from the hardware in real time as well. The red on the plot is a by second plot of the counts recorded in the previous second. The green plot is a simple moving average of the past ten of those. No fancy math. Both are converted into CPM from CPS simply by multipling by 60 and 6, respectively, in integer math (avoiding the need to have floating point for the CPM10 output for one thing).

You will note in the read plot that you can only have an integer number of counts in a second -- that's the vertial quantization you see there. You can only have 0,1,2,3,4,5 and so on. Also, counts happen when they happen, when something that's quite random finally tunnels, a comic ray hits after a zillion years of flying across the universe whatever -- it's pretty random. There is some diurnal variation in the cosmics, but it's too slow to see on this short run. This is due to whatever direction they happen to be coming from interacting with the earth's variable magnetic field (in shape and size and strength). Nothing we can do about that except go down into a mine, or build a very very think lead shield to at least shield us from the secondary showers. So, even the green line, which is ten seconds worth of data per point -- isn't very smooth at all in this plot. I could (and have) put a low pass filter on that, but that would be cheating, and it would vastly increase the time taken to get a measurement to any accuracy if you used that output for that. In all the signal processing theory I know, you just have to wait for enough data to have a certain amount of precision if that data is random. There's no real way you can cheat that one. Or is there? There's certainly a way to reduce this entire plot to one pretty accurate number, so lets do that for fun. We note the last visible line on the text display has a time, and a total count. So we can take that total count, divide it by total time, and bingo. We have 7:30, or 450 seconds. During that time we had 259 total counts (not a huge number for high accuracy). That works out to .575555... on my calculator, in counts per second. Times 60 (for CPM) we get: 34.53 CPM, a little higher than I got when I tried this earlier with special attention to no sources being within 20 feet or so, but not out of cosmic ray variability either.

That variablity, mostly from cosmics, sets a lower limit to how small a radiation you can measure with confidence to a certain precision in a certain time. In a good system, which this is, it sets that limit - these counter tubes have extremely low self-counting. So, like a receiver with a quiet front end -- all the noise is input the antenna, not the hardware here. I don't think this is going to be a problem with any reasonable activation sample, but of course there's going to be little point counting silver for hours and subtracting a background count taken over hours, if the silver all decays much faster than that. Here, activations between 500 and 1500 counts (on another pancake not the same) are the norm, so I don't see this noise level in the few CPM as being a serious limit for measuring integrated neutron output from an experiment. I will, however, do just that and of course, post the results up here for all to see. Perhaps I can take counts from both sides of a silver sample with two counters at the same time, to establish cross-calibration from what was my standard, to these. I expect that since these show much higher beta sensitivity than the one I'd been using, that they will give similar numbers despite smaller area, but that's a guess - We can measure and we'll know for certain soon.

Screenshot-3.png
A longer run with various sources and backgound in between them.


Both of these plots are log scale which gives the same height (of the variations) on the screen for the same SNR. The first thing to notice is that on this timescale and set of averages, background is still pretty noisy a signal -- but when I put something hot on there, it really smooths out - just as you'd expect when you have more data input. You can even see which source I put up there if you know how they read on this thing (and I do, this is my lab set of standards). The sources, in order are: A big radium based spark gap tube for military radar, very loud. Then the cal source we intend to ship with these, a thorium lamp mantle. Then a lot of background, then a small Cs-137 spark gap tube. After that put on a small sample of ammonium di-uranate, which is a little bit glued into a 3/8" hole washer, between two microscope cover slips. I think that probably stops most of the alpha, but not much of the beta. Last is that same source sitting on the aluminum beta-stop cover, showing us still seeing the gammas but no longer as many of the betas from this source.
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.
User avatar
Doug Coulter
 
Posts: 3515
Joined: Wed Jul 14, 2010 7:05 pm
Location: Floyd county, VA, USA

Return to Metrology

Who is online

Users browsing this forum: No registered users and 6 guests

cron