Standardized measurement for fusor comparisons

[Doug Coulter]

One issue with this "distributed development" (similar to open source software) we fusor people are doing is how can we compare results run to run, and importantly, across labs?
Most of us stress and strain to have any detectors at all, don't know the calibration factors of them to get to "absolute" numbers (dodgy at best), and very few of us have the exact same kind of thing to use for cross-lab comparisons. So I'm going to propose something here that is fairly inexpensive and hopefully "good enough" to let us compare cross lab and run to run, so we can even know if someone or some technique is doing better. Else it's like trying to learn violin whilst being tone-deaf and unaware of your errors -- you're not going to get too far!

Various things have been proposed, from expensive 3He tubes to expensive BTI's to unreliable B10 tubes (prone to noise pickup), you name it. Trouble is, even if we all had the nice 3He tubes, what are the chances we all get them the same sensitivity, size, equivalent moderators, and so forth? All those are sources of error.

I'm going to propose we standardize on silver activation, as it is very reliable, and isn't fooled by the things that can make tubes inaccurate, in conjunction with timed runs. Coupled with a specific design of moderator, this should get us all within around 10% or so of the same numbers for a given neutron output. That's not bad, pro labs often have more discrepancy than that when making more-advanced measurements (spectra, isotropy, cross section vs energy etc).

We managed to get some nice 2" pancake geiger sensors from Geo when he was selling them (maybe he still is), which we like a lot as your basic geiger counter head. These are thin-wall, and will see alphas as well as the rest, and even came with a calibration sheet for X rays, which most geigers aren't very sensitive to. These are pretty good. Using that as a basis for other measurements seems good. A friend just found some Russian pancakes on ebay, so we'll try one of those too and compare to see if we can use something like this as a standard for not too much money and have some consistency.

Based on the 2" window, we use sheets of silver foil, in our case 10 mils thick (probably not real important) to activate with our fusor, and 2" square. We got the silver from a jewelry supply house, not too terribly expensive. This is placed in our "neutron oven" which I will picture and describe. We use HDPE UMHW as our moderator, as it's about the best there is for this, and we want fairly high sensitivity. When we find we've gotten silver insanely hot to the point of it being less useful, we can always switch to indium, which is relatively "numb" but easier to measure as it has a longer half life after activation, at least once you get it well out of your background.

Here, our counter is set to count in 10 second intervals and multiply the answers by 6 to get counts per minute. This seems the best interval to use to get some time resolution (silver decays FAST) and decent averaging, but it's for sure a tradeoff, and we often take long runs on things like background to get a better long term average, as radiation, especially cosmic, is bursty in nature and pretty unpredictable.

Here we see long term averages on background of 42 cpm, but it can vary from 6-120 at any time, sometimes will stay "hot" for minutes on end due to magnetic weather and solar weather interactions with the cosmics. So any scheme that wants to measure some other radiation has to get the count rate up enough to swamp that out fairly well, and silver does that at the near million neutron per second rates just fine -- we see 700-1500 cpm activation on the first interval here when we are at those kinds of outputs, so the variable background doesn't add too much noise to the results.

So the main variable once we have geiger counters that agree is the neutron oven. Carl Willis calculated the best thickness between the silver and the source as being about 1.5" of HDPE, so that's what we use on that side. On the other side, we have about 3-4" more HDPE behind the silver to help reflect neutrons back into it, as the moderation process works that way -- random direction changes of the neutrons. We have the 'input side" more or less touching the tank wall, or 3.5" from the grid center. Changes in spacing are fairly easy to take into account mathematically for those using a different size apparatus or spacing, and are not too severe if you stay fairly close. I basically cut up a 4" diameter HDPE rod so the bottom matches the tank curvature, made a cut through where it was 1.5" thick at the thinnest point due to that curvature, and put the rest on top of the activation sample. I found setting it right on the tank not to be so good -- the tank gets hot enough to melt it, so I glued some carbon rods on the bottom that space it up about a quarter inch to let some air get under there for cooling. The whole thing is placed in a hole in the lead shielding I have all over the rest of the gear to keep the X rays down in the lab (which is a good idea in general, square law only gets you so far).

Neutron Oven parts

Above are the parts of the neutron oven, with the bottom upside down so you can see how I cut it to fit my tank, and the carbon rods I used as spacers so it would stop melting.

Oven with samples

Here is the oven with the base right side up and the activation samples sitting where they sit during the process. The odd shaped one started as a pea sized piece of indium I pounded with a hammer. The other is a 2" sq by .01" thick piece of 99.9+ fine silver.

On the fusor

And here it is ready to go, on the fusor. My cylindrical grid is in the sidearm the oven sits atop, in a hole in the shielding so it gets everything full blast. The HV for the main grid comes in from the back/right in this view. As you can see, I have reasonably decent lead coverage, even though I sit well on the other side. X rays do scatter, and my wife sits above this to watch TV on the floor above...The lead pretty much stops all the power supply energy X rays, but not the neutrons or the real hot gammas from fusion. Obviously there is more I could do with shielding here.

The HV probe in the foreground is looking at the power to a supplementary grid I have out in the main volume of the tank, as an ion source and pulse inducer for the main rig. Pretty cool, works a charm, and there's a lot more left to discover with that setup.



I am fairly sure this will work for people, but I'm also pretty sure it's not the only way, so if anyone has ideas on how to do it better -- speak up! This one seems to have the fewest things that have to be controlled to make it work as a standard -- you need a standard geiger with standard area of sensitivity, an easy to make neutron oven, and a chunk of pure silver, available most anywhere.

Probably the hardest part for everyone to match up on will be that geiger sensor, and those are at least fairly cheap and fairly available as these things go. Not to mention, useful around the lab for checking sources and personal safety, so people should have one anyway. The UMHW HDPE is a standard item at McMaster, and the machining for this is fairly easy. The lead isn't required, just a good idea.

We are using a homebrew HV supply for the geiger, off a CCFL, and doing the counting with a PIC 18f6520 board I designed and wrote code for. It emits counts on 5 channels every 10 seconds, and a/d readings on 4 channels evrey second on RS-232 for my PC data logging software to slurp up -- and I can see realtime data on the wide screen TV lashed to the computer in the lab.
(anyone who wants design data, pcb layouts, code for that is welcome to it, just ask)

This is a nice setup, we just log data from before power on to after the activation measurements. That way, you can see the power go off, the geiger counter go back to background, and then rise again when I put the silver on it, with fairly precise timing being logged on it all, in case it should take me a longer or shorter time to get there with the silver (it matters).
Timing on silver matters due to a real fast decay metastable state -- in the 12 second range half life, so we usually measure enough of the decay to let that go away, then do computatons with the next slower decay it has (there are a ton of possibilities with silver as the CRC handbook will show). Indium is a lot simpler -- pretty much just one half life, but you need more neutrons to get it to count well out of the noise. I run both at once, and take the indium measurement last, as timing is not so important to that, and we have time information if we want to back out the effects of the delay anyway. That quick mode in the silver is probably more responsible for the variation we see than the fusor itself, and I'm going to write some code to back out that variation using the known rates for silver and the known delay from fusor off to measure start.

I will try and get a movie soon to put here so people can see the whole process. Pretty simple, at least after you see it done once.

[chrismb]

This is not a well-thought out post, just shooting from the hip, but I would tend to think that there is some way to actually self-calibrate a steady neutron source.

I am thinking along the following lines; a neutron source with an emission rate N will give a flux 4F at, say, 500mm and F at 1000mm. If we now have two sets of moderator then, presumably, there is some 'performance function' that can be derived and it is some exponential (?) function of moderator thickness (w), say Ae^[kw] for constants A and k.

We also have a piece of silver that after activation by the [4]F neutrons gives a particular reading on an uncalibrated (but consistent) Geiger counter. So we also have a conversion factor, C, which is F(per sec) = C.[counts]

If we now run 4 measurements, each at the two distances and with each of the two moderators, we end up with 4 measurements and 4 unknowns, A, k, F and C. So presumably we must now be able to work out, in absolute measure, the total neutron count?

(This reminds me a bit of how you can cross-calibrate antennas; because the function of a radiating antenna differs to that of a receiving antenna, you can fire 3 such antennas at each other and end up with as many measurements as unknowns, and can cross-calibrate them all. In this case, we are playing with the fact that one variable is a ^2 term [viz. distance] and the other is exponential [viz. moderator thickness]. This should mean the two functions will cross, in Cartesian space.)

[Doug Coulter]

Something like that does seem possible over the short ranges you propose, with a fast-only neutron detector (for which a moderator acts more or less like a perfect shield, or a very thin one a known attenuator). EG BTI or plastic scintillator or recoil/gas detector. Of those, all but the BTI (expensive and unreliable) respond to other radiation too much to be used as a prime calibration of a fusor output. They are useful for relative measurements, and I do use them for tuning in real time, just not for absolute as you don't know how much of the output is from the other things.

A joke in the fast neutron labs that do "open geometry" experiments to reduce scattering errors is "first, take out all the air in the lab".

With a slow neutron detector it seems you're kind of messed up from the beginning there -- neutrons scatter off everything around, and out of moderators, and so on, with functions that are hard to quantify really well.

For the scheme I proposed, we could, worst case, mail a standard source (old gas lamp mantle, with Th) around so everyone could cross calibrate their geiger counters. All legal and so on, and pretty easy to shield as regards any detectors in the mail systems, as Al foil would do for alphas from that and not freak out an X ray machine with a big black looking piece of lead. Or a smoke detector source of known brand that can be bought anywhere and spaced at a measured distance from the counter would do too, I suppose.

One reason you don't worry too much about neutrons from a far away power plant is (other than that they do shield pretty well) the air itself moderates and captures them (in the water) long before they get to long distances. Once slowed, of course they have a half life even if not captured, so you're pretty safe at distance.

That's not to discount JohnF's statement at all -- a tandem doing beam on target neutron generation is one "heck" of a loud source! So even if the probability of any neutron making it to miles is nearly nil (it is), if you start out with enough of them, there's still a few that will make it.