by Doug Coulter » Wed Sep 28, 2011 7:33 pm
I check you on all that, Geo.
I explicitly tested for X ray sensitivity, and yes, there is some. Most of my fusor is covered with 3/32" lead, overlapped and soldered, but I left a couple holes for:
1. Seeing X and gamma rays (not all are produced at power supply volts, a spectrum is interesting and I plan to post one soon -- seems fusion generates some ~ 2mev photons -- I 'll get a calibrated number though - that one is a guess from comparing pulse heights on a scope between Cs-137 and the fusor, not calibrated otherwise). Obviously, I'm not going to take my gallon jug Nai down there and expose it to the fusor and activate the I in it - I'll use a crummier one that should still be good enough.
2. Having a place to put my neutron activation oven as close in as possible.
Most people don't shield their fusors, saying sq law is enough. I call bull on that - better to be safe, and I've found that even a hole in the shielding in the back results in plenty of X rays in the front, where I am, even on a geiger counter which is pretty numb to them. They seem to scatter off the air itself, and of course, all the heavy metal machines. I try really hard to keep my exposure well below 1 mrem per run and don't run that often usually, only when there's something new to test. The system is really reliable and repeatable now, so I don't need to run it as much.
Knowing that ZnS has at least some sensitivity to X rays (How could it not? Even UV makes it go!) I explicitly tested this near a hole in my own fusor shielding, as it's going up to HEAS with Bill to look at Richard's unshielded fusor and show off our new Standard Counter project. Yup, problems there, though the X ray pulses weren't as tall as the better neutron ones. I experimentally determined the thinnest lead that stopped the noise while looking at the analog output, which was about .040" thick and made a conformal cap of lead sheet for the whole tube. It still reads about the same on neutrons, but now doesn't change when I move it from behind the thick shield to where there's none. ZnS isn't perfect, but it does respond a lot better to charged particles than to X rays, which most other phosphors I know of don't share. I would otherwise have used a faster phosphor, or scint plastic, but those are optimized the other way - they do X rays better than charged particles from neutron knock-ons. I did mess with that one awhile back, but by the time you stop the X rays well enough to shut up the plastic, you've pretty much either stopped or slowed the neutrons so much you don't have enough energy left to knock protons out of the plastic either.
I could use some mu metal fer sure. I heard if you bend it up too much it has to be re-annealed (in ugh, hydrogen) (Arnold magnetics books) with a very long heating/cooling cycle. I did degauss the 4130 chrome-moly tubing before putting the tube in it, but yes, the spec on that phototube says keep it below 3 gauss or it won't work right. And yes, I've seen things "self magnetize".
There's not a lot of room in there, it would have to be thick to go inside - the ID of the pipe is only about 20 mils over that of the tube/detector and the tape holding the detector on.
I did degauss the tubing before putting in the phototube, of course. Couldn't use Al tubing as it has a capture resonance for fast neutrons right where we're working here -- didn't seem wise.
I don't have data on every material there, but some books mention the Al issue, and also He having a big cross section such that some labs use He filled balloons to filter out fast neutrons(!).
In this case, I decided to use a rather sloppy (but very stable) threshold of the schmitt trigger input of the PIC uP, and just vary the preamp gain to get it about right - threshold is about 4v, and the peak preamp output is about 9v and I set gain to just clip those. This results in missing a good number of counts -- the energy transferred to the phosphor from any one collision is fairly random, producing pulses all the way down to almost zero. I chose gain to make the threshold about 4x the X ray pulse height and various noise in the preamp and phototube. That means I miss some pulses of course, but don't get false hits from anyting other than cosmic rays, and that rate is pretty low ~7-8 cpm. A small gain shift wouldn't make much difference to this, but a big one would. I used tight regulation on the phototube supply and tested that since that's another gain-shift issue, and is a real big effect on multiplier tubes.
Another effect that's worrying is the ZnS:Ag long phosphorescence. This might cause a DC shift that would affect things, as well as take some of it "out of the game" if it's busy doing that, it probably doesn't do the normal fluorescence thing the same. We might be seeing just that in the curves I've already posted up here - the output seems to trend down slightly with time on all runs. I will therefore do one with a count from this in time sync with one of the other neutron detectors and see if it's the sensor - I suspect it is.
My take on neutron metrology (which seems to be generally agreed with) is that none of these real-time things -- boron, 3He, or this thing, can really be counted on to be better than perhaps 10% in otherwise *perfect* conditions. I've noted changes in my other tubes simply due to walking near -- the bag o' water & fat that is "me" adds enough more moderator to the system to change their output! Richard H is fanatic about never moving his, but that's not the only source of errors I've seen, to say the least. Millivolt signals in the presence of serious EMI are fraught with possible problems even in things that seem "tight". The Russian 3he tubes sometimes take off like a relaxation oscillator for a few seconds (the pulses look different on analog, and sound different which is why I use audio monitoring on all these). This is why my own religion is to use silver or Indium activation and count that with a beta sensitive counter when I need to be really sure. For me, the realtime things are mainly useful for real time tuning of the neutron source -- for example does a little more gas in the tank make more? I know before my hand is off the valve. Does more current make more neutrons in scale with the increased power? (answer, no) - for that sort of thing they are great. For computing the actual numerical neutron output from the fusor? No way. Too many ways to fool them.
There are zero ways to fool silver, and heck, it's cheap and easy to do, it just doesn't give you real time feedback for tuning.
In short, there are "issues" it seems with every neutron detector technology there is. The Hornyak has its blessings though - I can and have made them myself - so I am not dependent on a lucky score to have them to sell. And you can always get phototubes. Joe and I may make some boron tubes - we'll be having a F/F meeting after HEAS to discuss things like that. Another approach for fast neutrons (no moderator seems to be a feature here, they're big and heavy) is to look at recoils off hydrogen or helium in a big proportional tube. Those are fairly X ray numb too, but the signal is tiny. Again, the advantage is that they're cheap and we can just make them for other hobbyists.
Geo -- I made you a subforum to list anything you want to sell in...it's all yours, go for it! The more the merrier, and we all benefit, I believe. Look for "Geo's" under Vendors.
Edit, nice little speaker driver circuit you did. The one shot should make it a lot easier to hear. I worry a bit about the peak current in the output transistor, but if you have pulse width control you can control that nicely (speakers have inductance). I am planning to use an lm386 chip here for other reasons. I can put a peak in the response that matches speaker resonance if I like, but I've found that on several things I've got here -- being able to hear the pulse width and height give the user additional information. And in fact, the best "sound improver" is Joe Sousa's pill bottle "tick to tock" converter -- it really works nicely and is passive. The 386 draws very little power, works down to 4v and it's less parts overall. I have also built small stereo amps into rack panels with speakers at the sides as part of every gear rack I have here, with switched inputs and BNC's on the panels, with a "tape out" to go to computer soundcards for recording runs.
I've found with my big expensive 3He tube that during neutron bursts, not only do some of the pulses double or more in height, but also get wider. This is easy to hear since without the fixed pulse stretcher, they get more "bassy" or "thumpy". I've been working with pulsed fusion, so this is nice to have for me, dunno if it would help anyone else. Having done a lot of time as an audiophool, a musician etc, I have pretty good ears and like not having to look to get the info while I'm trying other things. Of course, if your counter is a geiger, there's no point, all the pulses look alike except ones marginally inside the dead time - and you might as well just make them all uniform in that case.
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.