How about it? Experiments of the third kind , take 999999.

On a sunny day (Fri, 23 Mar 2012 13:38:42 +0000) it happened John Devereux wrote in :

OK, 'funded' Ha, what a joke. Now lets talk some numbers, you have read the Russian paper[1] too of course. I worked a bit on my test setup, and now can easily get about a volt signal into an ADC with a volt reference. A 10 bit ADC. So that makes 1024 steps, or 1/10 of a percent sensitivity. Now lets allow for some drift due to whatever in the setup, even if we play thermostat, and then we are still good to detect 1% changes. If it is less than 1% then I am not interested in the effect, as then it has no practical value for me, I am not CERN (Crazy Expensive Return Nothing), I am doing it so if the effects exist so I can use them. There must a be, always, a practical application of what you do. Else it becomes like string theory, or the CERN particle zoo, where all you have is an A4 piece of paper with a theory of almost everything, except what you need to advance, oh and that needs more funding. So 1% or a bit better will do, and I do not give a f*ck about what anybody thinks it SHOULD be, unless they get of their rear end and do an experiment themselves. Then I will listen.

This is a fun project and any kid of any age can do it, with just pocket money. Also old kids like me.

I challenge anyone here with the numbers.

Now sure if you want to donate a million $ to support the experiment, then I will accept those and go to a nice warm island with hulahoop girls and cool drinks and dream about building stuff. I would not, really, go live in a tunnel on the border of Switzerland and H.. sorry some other country. :-)

On a sunny day (Fri, 23 Mar 2012 15:06:56 GMT) it happened Jan Panteltje wrote in :

[1] I am 99% sure the effect they did see is due to earth magnetic field. they wen to the north pole with some expedition and there the effect was zero. Sign on the wall.

the past.

year

proportional

=20

Not to mention in design of experiment, you want to obtain the best reasonably achievable signal to noise ratio, thus half lives on the order of kiloyears. Rather than swamp the desired signal under an unduly short half life.

?-)

quartz tube.

cell,

45°C.

more power,

compensation I think,

shielding I think,

I remember you saying you have some PMT's, if the tritium could replace the phosphor in a PMT, then it would just be an electron multiplier tube getting electrons from the tritium and would be more sensitive than having to measure the light off a phosphor, should be able to detect each electron emitted from the tritium atoms?

cheers, Jamie

The variation is

On a sunny day (Fri, 23 Mar 2012 20:43:37 -0700) it happened Jamie M wrote in :

The big problem with PMTs is that they are so sensitive to magnetic fields. Even the earth magnetic field, and especially at low anode voltages, as then the electrons have low speed and are easily deflected. I think magnetic influences can explain the seasonal changes in the Russian experiment with Plutonium, more current the electric wiring in the winter (light on longer) than in the summer, it could even be somebody walking in with a steel watch that is slightly magnetic, I have tried walking about with a magnet with the tritium light taped to the front of the PMT, and measuring the voltage drop over the PMT anode resistor, 10% changes at 1 meter are normal. So that rules out, for an experiment that has to run so long, PMTs. I was very happy and actually surprised that the tritium light does not seem to react to a powerful magnet right next to it., One systematic error less to worry about.

On a sunny day (Sat, 24 Mar 2012 09:37:18 +0000) it happened Martin Brown wrote in :

Well, the experiment is now with tritium, and the result, if any, will be added to the pool of knowledge of the human beans[1].

[1] LOL

with metal shielding I think,

When i see a post like this it makes my soul hurt. There is no phosphor in a PMT. Just a sensitive photocathode, dynodes, and ending with an anode, where the output signal occurs.

?-0

metal shielding I think,

You'd make a good school teacher!

cheers, Jamie

with metal shielding I think,

replace

phosphor

I may try that yet.

?-)

One thing springs to mind: perhaps eliminate much front-end drift by making this an AC system; employ a spinning mirror (or moving shutter) in the beam path. Crank the gain high and add a synchronous detector only at the end. If drift can be reduced enough, you might detect tritium changes on a scale of days or hours. With a mechanical shutter you could also include a separate time slice for "calibration LED" light sent through the same system. Any changes in the tritium channel which don't also appear in the LED channel would be verified as not simple drift-artifacts.

((((((((((((((((((((((( ( ( (o) ) ) ))))))))))))))))))))))) William J. Beaty Research Engineer beaty a chem washington edu UW Chem Dept, Bagley Hall RM74 billb a eskimo com Box 351700, Seattle, WA 98195-1700 ph206-762-3818

Hmm, but LEDs aren't the most ideal sources. They age when left on (not sure about when off), and they have lots of nonlinear (current and temperature dependent) effects.

Could use a tungsten light, only powered when it's required if battery operated. But even tungsten evaporates and filament resistance and temperature change; it could be calibrated to the same operating temperature using an independent method, e.g. two wavelength calibration. When detectors are in the correct ratio (assuming similar detector drifts), bulb voltage is correct and the detector recalibrated.

The cold-to-hot resistance ratio could be used, although this would be subject to material drift properties -- I wouldn't expect the tungsten metal itself to change much over time, but impurities could change it faster (e.g., oxidation due to leakage?). Wonder what the dynamics of that are.

Tim

-- Deep Friar: a very philosophical monk. Website:

On a sunny day (Fri, 13 Apr 2012 17:26:34 -0700 (PDT)) it happened Bill Beaty wrote in :

Thanks, this is a very good idea, and I considered it. The experimental setup has however evolved into using 2 photo diodes, one blacked out.

This does not rule out any future mechanical shutter ideas, and I even later considered using a pair of old Nvidia 3D goggles as LCD shutter, as that perhaps does not wear out so much, and has no vibration (microphonics on ceramic caps is what I am worried about now), and no wear over time.

Thank you very much for the suggestion :-)

eaty

DOH, I did finally read the rest of the thread!

Build two? I'm always tempted on small projects to copy everything as I go, and in this case a pair of loggers would have several obvious uses.

Also, might you have collected a bibliography? I've heard of the solar flare anomaly, and people ranting about C14 dating.

Here's this:

8&sciodt=3D1,48&hl=3Den &sciodt=3D1,48&hl=3Den

((((((((((((((((((((((( ( ( (o) ) ) ))))))))))))))))))))))) William J. Beaty Research Engineer beaty, chem washington edu UW Chem Dept, Bagley Hall RM74 billb, eskimocom Box 351700, Seattle, WA 98195-1700 ph 206-762-3818

On a sunny day (Thu, 19 Apr 2012 12:54:30 -0700 (PDT)) it happened Bill Beaty wrote in :

Yes, great, indeed it was sort of a hot topic because some of that stuff is used for calibration. I looked up the links, but to get the full papers you have to buy them. But I have read from other sources about it. I did some test runs just now, and I get 1 part in 1000 accuracy no problem it seems. I had a small temperature controller problem, it turned out the diode I was using as temperature sensor that was glued against the hot plate that holds the electronics had come lose (2 component glue failing??). I replaced it by a small NPN with a mounting hole and use the collector-base junction as temp sensor, mounted with a screw. Now it is running, hopefully tomorrow or in the weekend I will finish the soft for the main board and can go for a one week run or so (I hope). No idea if I will see *changes* in one week, it is more to get silly problems out.

...or be working at a university library terminal.

I found a good one from this year, w/review of many papers:

I've seen that before: slightly greasy surface sometimes acts as "mold release" for making epoxy castings. Need sandpaper, or scrub under running water.

((((((((((((((((((((((( ( ( (o) ) ) ))))))))))))))))))))))) William J. Beaty Research Engineer beaty, chem washington edu UW Chem Dept, Bagley Hall RM74 billb, eskimocom Box 351700, Seattle, WA 98195-1700 ph 206-762-3818

On a sunny day (Fri, 20 Apr 2012 12:08:34 -0700 (PDT)) it happened Bill Beaty wrote in :

Interesting paper, and many seasonal changes. I am still thinking environmental influences in some of those. I guess you have to be an electronics guy and not a physicist to be aware of how many things can influence your equipment. It would be great fun to see a seasonal change with this. It ran all night, 24 data points, and I just set it up for an other all night run, now with the latest firmware. I added anything I could think of..

Yes, in this case the epoxy was a bit soft, I used those 2 injection tube type dispensers, so the mixing ratio was OK. But it was very old (10 years) glue, dunno if that makes a difference. I had to scrape the rest of and mounted that transistor with some heat paste.

Just longwave IR can do odd things. For close control of thermal effects, metal enclosures are needed, since some visually-opaque plastics are transparent to thermal radiation. If in doubt, wave a

1KW electric heater around to prove there's no unexpected optical paths to create temperature change.

On a sunny day (Fri, 20 Apr 2012 17:48:49 -0700 (PDT)) it happened Bill Beaty wrote in :

Right, one of the next things I will try is the fridge test (to -5C), and the oven test (up to 40C). Both on battery and mains power.

It is funny how you can overlook things, this morning I realized I forgot something very important: A one year alarm!

Imagine, and I know how one forgets, somebody asking in September next year: "Hey how about that experiment you were doing>" And "oops, should have checked that in July!" So a one year (365 days really) beeper, need to add that, if you let it running I think now it will flip over when the EEPROMS are full and start at data point zero again [1]... (These EEPROMS according to the data sheet just roll over the address counter). You would still have more than a year data, but why wait so long...

There are more little thing like that, I uses a green LED in the box as 2.14 V reference for something, it really needs to be painted black, etc etc.

Yes it is all in an alum box too, decoupling caps go even on the RS232 in and out. Cellphone interference test already happened.

[1] I am now saving 6 bytes data per one hour sample (3 per EEPROM), so an address run over will not start at zero.