Ion chamber

caustic meatloaf":2qodvh4t said:

I'm guessing it's not too high of a voltage since the 7721 is for some highly precise amplification, but I don't think it would be too much of a stretch to add on a 2nd op-amp stage to the output and use that to drive an LED or something similarly low voltage - that way you could see the LED pulse when you have detection spikes.

Click to expand...

Ya, I was thinking about that. It seems to me that it doesn't detect individual tracks from normal sources, but once in a while, a 'big one' hits (cosmic ray muon? Entire cosmic ray shower?), and it detects that. If I can detect muons rapidly enough (I should be able to detect the ions in flight capacitively even before they hit electrodes), I may be able to make a spark chamber - with two ion chambers, coincidence detector, and high voltage pulse generator.

edit: re open space, I opened to take a photo. I close the cans together when using it. I soldered a wire to keep both cans connected, a bit of issue is that the second can also works as ion chamber, but I don't think it makes any problem.

edit: for future build I may just use one can containing both ion chamber and electronics. I'll also try using some stiff metal for inner electrode, to make it less sensitive to vibration, along with low pass filter. I'll also keep other end of electrode's support insulator at the nearly same voltage as electrode, it seems to me that stuff i find online is putting insulator straight between their +50v or so ion chamber body, and the electrode, and no wonder the insulator needs to be very awesomely insulating.

Got radium watch for testing:

http://youtu.be/y97a5fFMqQI

This is just awesome. I'm going to build a properly sealed detector and make it look nice.

Wudan Master":cdrttiq0 said:

next step NivenBro

seriously though that's really cool.

Have you tried a smoke alarm?

Click to expand...

Haven't got any... though those are very easy to detect, I'd think.

The things I found out so far: it works fine with just 2.5 volts of chamber bias voltage, but responds with a lag. I need to calculate the ion drift time for 2.5 volts...

My current plan is to: seal up the amplifier and resistor in a little metal box, with perhaps moisture absorber, with inputs and outputs and a Teflon-insulated guarded terminal (guarded means I have a shield around which I keep at close voltage to the input, minimizing the leakage over the terminal). Ideally I want a long surface on the teflon terminal, ala -|=|=|=|= . This will make it operate well regardless of moisture level (albeit i havent had moisture problems so far).

This will be inside a circuitry can, in which I will also have the batteries and scale and perhaps range switch. The ion chamber can will attach to the circuitry can using a clamp, and it will have different lids - metal mesh lid, foil lid, and a lid with sample holder.

If I don't drop the idea I'll make a micro-controller version, using some way lower grade amplifier and a reed switch to integrate over N seconds and sample, in a very temperature independent way (if this idea works). I can time the pulse to measure it, so I won't even need separate ADC. Then I can also add a pressure sensor. Other possible perversions include adding a bluetooth module and using it together with smartphone.

I think it really is a better approach than Geiger counter for personal meters. The Geiger counter was awesome idea back when you couldn't amplify the current otherwise. Now it is easier to amplify the tiny current than to make high voltage, maintain a low pressure gas mix (the helium leaks out), etc. And this thing, it can detect potassium fertilizer (with some integration time). This makes it fit for detecting e.g. radioactive contamination in food, something which does not work very well with Geiger counters. On top of that, this, if well sealed, can measure cumulative dose while being powered down entirely for almost all of the time. I think you could make ion chamber detector that samples every 30 seconds, has immediate 'high dose' alarm, and runs for years non-stop off a watch battery, and can be charged by shaking it. That, with a flat response curve no matter if it is 13 KeV tritium or some >1 MeV beta. edit: wait, there is such detector, it's used to detect smoke lol. And detecting lower doses of radiation only requires even less power.

Me and my girlfriend built another one into a project box.

http://imgur.com/a/NPBvB


Most of the components are on the other side of that board. Tested it with a small piece of thorite and an old radium watch.

I've changed the circuit since taking that photo - the board looks a lot busier now, with a voltage regulator (I had a problem due to battery voltage fluctuating by a few millivolts, so I added a series regulator), a couple filtering capacitors, and using small connectors for everything so that it is easier to open the box. I got an old Soviet 50 microampere meter for about 1.2 Euro at the local flea market, heh. Wish the LMP7721 was a micropower chip - then this could work for many months on single charge. I use two tiny 12v batteries for the chamber bias. I'll probably add some step up converter instead because as the battery temperature changes, it's voltage changes a little, and even few millivolts per second screw up the readings.

One other modification that I still have to do is to make an inner chamber that is not mechanically coupled to the lid. The lid is too thin, and pressing on it changes chamber capacitance, messing up the readings (like a condenser microphone). It takes several hours for it to settle after construction - my understanding is that you end up depositing charges on the dielectrics, which take time to dissipate, and while they dissipate they screw up the readings (as they are capacitively coupled to the sense wire).

It's quite sensitive, and seems to respond to background radiation. Cosmic ray showers make the needle jump up. Unfortunately I don't have any contacts where I live for calibrating it properly. I can check linearity with the old radium watch and the thorite sample, though.