Fusor data - main grid the tungsten 8 rod, graphite endcaps one.
D pressure 1.6e2 mbar indicated. The PKR 251 reads factor of 2 high on D (according to the manuf) so converting to microns with that factor included gives 10.66... micons.
Second grid W wire loop (as usual) with 10 ma through 100k ballast for ion source - usually around 5kv (makes no measurable neutrons but allows the lower running pressure).
Power supply nominal 50.7kv, 20 ma current limit, usually about 17ma when running (bounces around, and logged this time)
Run terminated by feedthrough failure. It may work again when everything cools down (again, this has happened before).
Indium in the main neutron oven, silver and gold in the secondary one, silver not counted this time, gold - working on that before it fully decays (software issues URSA/windows).
Here's the plots from 30,000 feet: Geigers and Neutrons, lines only, 10 sec averages only, nearly all of run and count after:
- Plot of SC log file for run and activation count
A couple of things to note. With our rough ad-hoc calibration off our own and Richard Hull's data, we were making about 1.7 million neuts a second at the start of the run. We are seeing this decay for some reasons we know, and some we don't. The ones we know are that this is batch mode, and after some running, we need to put in fresh gas, as it seems baking and outgassing lets in some pollutants that reduce the neutron production. I need to take a mass spectrum after a run soon to see if we're just making hydrocarbons with the graphite parts of the grid and the D.
(All this windows junk seems to decay with time, it's been hard to keep it all running - windows updates break things, have to go and get new verisions of the main software from the uncooperative manufs to fix that, and so on - quite lame, but a nice lock-in FOR THEM since they refuse to do that without more money after a time -- even though you've already paid them 10's of thousands for the hardware.)
We also see a general decline over the time of a long run, which I don't know the cause of -- hot stuff doesn't work as well? We did not see this on Richards data using the same sensors. Since we've logged this, we can get some real accurate numbers by crunching what's in the log files, but this gives you the overview. At the end of the run, you can see the neutrons drop to nothing, the geigers go down when I shut off the main power (so there's a nice reference on what our background is, right in the plot), and after awhile, where I put the In onto the standard counter to see if I got it hot - and I did!
On this log plot, any decay looks like a straight line sloping down. In this case, that's most of the time for the plot of the overall run. With StdPlot software, we can drill down to more detail easily, and will. We can also go into the log files (attached) and find exact seconds, and do things like find the total counts over some arbitrary time interval to get a long term average as statistically good as the data allows.
Guess I ought to put the log files in there. There are two from standard counter, and one from another computer logged counter I run - the main benefit of the latter is it also logs some a/d inputs that show the power supply voltage and current vs time, along with its own pancake counter at 10 sec intervals. I'll have to write some more perl to plot that one, (different file format) and put in the correct conversions from raw A/D numbers to volts and milliamps. However, empirically, when the geiger counts and neutron counts go down in the above plot, what was happening was the supply was current limiting on the additional gas from outgassing (or whatever is going on), and they return to high levels when I pulse the foreline gas solenoid to take out some gas. At a couple points, I took out gas to the point things "went out" and then let in some fresh (in no case all the gas got out), then adjusted for the right operating conditions.
Here's some log files -- the main long one, a shorter one I took later of just the indium, and the one from the old multigeiger setup (has some glitches). That last is roughly 8 seconds earlier than the standard counter logs (need to make that something that is synced). I think once I plot it they will be easy to line up again, since when the geigers say the fusor went out, that's a sharp time market that will match the power supply current going to zero for a little while.
10_16_2011_16_0_25.log- Main log file plotted above
- (213.37 KiB) Downloaded 254 times
- 10_16_2011_17_1_55.log
- Just In, stared ~3-4 min after main log file end
- (31.62 KiB) Downloaded 254 times
- longrun.txt
- Mulit Geiger log (serial port save)
- (128 KiB) Downloaded 252 times
I always monitor all this on the audio and the scope. For fun, here's an example scope screen catpure. Ch1 is a 3He tube, ch2 is a very old B10 tube (farther away) and the third channel is the Hornyak detectos which is what is plotted in the other plots. Nice take on the various sensitivities, and in the case of the 3He, how I need to get back into that preamp and fix up the DC restoration. The B10 preamp is slow (but then, so is the tube response). The baseline noise on the traces comes in because I told the scope to sample fast (2.5 ghz) while the sweep was slow, so I could see even sub-pixel pulses at the slow sweep rate.
- Scope capture
(I'll be editing this to add more plots and whatnot over the next few days).
