2016. július 11., hétfő

Building a radiation detector from scratch 3 - the first detector

The detector will consist of a tin can, two transistors and a DVM. I will also show how to improve and stabilize the basic circuit by adding an extra transistor and an identical "dummy" circuit to cancel leakage current and temperature dependence.
The circuit will be same as the one on this video: https://www.youtube.com/watch?v=96bybrgI6V0

An ion chamber is "just" a very sensitive resistance (or conductivity) measuring device. It measures the resistance of a gas, in our case, thin air.

Air is not a terribly conductive medium. It's resistance is somewhere between 1.3*1016 and 3.3*1016 ohmmeter (http://chemistry.about.com/od/moleculescompounds/a/Table-Of-Electrical-Resistivity-And-Conductivity.htm). 10 volts only causes 8*10-14 to 3*10-14 Amperes, or 8-3 femtoamperes. We need a device capable of measuring such tiny currents. Actually since we'd like to detect radiation, we will be dealing with much higher currents, in the picoampere (10-12) range. With hot samples, currents can go up to tens of nanoamperes (n*10*10-9). (http://www.teralab.co.uk/Experiments/Ion_Chamber/Ion_Chamber_Page1.htm)

A DVM, or digital voltmeter can measure milliamperes, or 10-3 Amperes. We still need to get about nine orders of magnitude gain.

We can quicly get an other 4 by cheating, and using the DVM's voltage measuring function to actually measure tiny currents. The DVM has an intrinsic internal resistance, often 10 megaohms, or 107 ohms. If the DVM can read millivolts, that means it will read 1mV if we send 10-10 amperes through it. That's not terribly smaller than our target of 10-12 - 10-15 amperes!

We will jump the gap with two appropriately connected transistors. Meet the Darlington pair.

To build a Darlington pair, connect two transistors together by their collectors, and connect one emitter to the other's base. The circuit with the two connected collectors, the unconnected emitter and the unconnected base leads forms a device similar to a single transistor, only it's DC current amplification constant equals to the two transistors' same parameters multiplied.

To build a good Darlington pair for our purposes, I recommend using cheap, generic bipolar transistors with large "hfe" or "beta" values. The BC548-C device for example is a good candidate, it has a hfe around 800. That is, a typical BC548-C transitor will allow a collector-emitter current 800 times larger to pass if you sent a small base-emitter current through it. Connect two of these puppies into a Darlington chain, and you get 160000, or one hundred and sixty thousand times the current out than the current in.

Fix the darlington pair near the hole on the bottom of a tin can. Solder the collectors to the can. Pass the base lead through the hole. You probably want to solder a stiff wire a few cm long to the base lead. It should be at the center of the tin can NOT TOUCHING IT AT ANY TIME.

Connect a 9V battery's positive lead to the can. Between the emitter lead and the negative battery lead should go your DVM in voltage measurement mode.

There you go, your first ion chamber!

After hooking it up, wait a minute or so to let the DVM reading to stabilize. Carefully move a piece of radioactive material near the tin can. The radioactive radiation will generate ions that will be driven towards either the can or the naked Darlington base. A tiny-tiny current will flow. The transistors will amplify this current, which will pass through the DVM. The high internal resistance of the DVM will cause the still-small current to cause a significant voltage value to be shown. Congratulations, you've just measured some radiation. Cool.

Could it be even better?

You could connect 2-3 9V batteries in series, to improve performance. You should get sharper, larger readings with more batteries. Don't add too much though, after maybe 5 batteries you risk shocking yourself!

Drift and zero reading

Transistors are nasty devices. First of all, they are not linear, meaning they do much more than just multiply their base current with a constant (although that's a good approximation for a number of applications). Their parameters are also dependent on their temperature. This simple detector doubles as a temperature sensor. Which is not good at all. Readings will jump around wildly even if you touch the them for a few seconds. Background radiation levels are also hard to establish and/or compare, since the transistors have a significant leakage current, they allow a small current to flow across their C-E leads even if no current is flowing across their base.

The leakage current AND the temperature dependence be greatly decreased by comparing the current in two identical circuits, one that measures radiation, the other measures nothing. When the temperature changes, the current also changes in both circuits in a very similar manner. The only difference is the radiation, and that can be measured by subtracting the leakage current.

The next circuit will work like this:

  • Build two very similar amplifiers, that are thermally coupled, meaning they are very close together, connected by a medium that can transfer heat well, so their temperature will hardly differ.
  • Make the current from each amplifier flow through a pair of identical resistors connected to ground.
  • Measure the voltage between the hot (non-grounded) lead of the resistors
  • The change in difference will be proportional to the change of conductivity in the ion chamber.
The resulting circuit will be surprisingly stable. With an extra part or two, it can also be zeroed, e.g. it can be adjusted to show (near) zero at normal background. It is sensitive enough to be used with fairly low activity samples such as thoriated tungsten welding rods, rainwater, and to absolutely go nuts around decent specimens of radioactive minerals.

This device can be used to indicate and compare the radioactivity of samples, and even to determine the type of radiation emitted by the them. Specimen with very short half lives can be monitored to see the exponential drop in activity (like in the case of fresh rainwater).

It is perfect for home and school experiments, or for hunting radioactive material.

Unfortunately the device is probably pretty much non-linear, and most likely would take a complicated extra microcontroller circuit that would correct this error, and a lengthy process of calibration. However, once done, the controller could take the temperature, humidity and driving voltage into account to make a decent guess and bring this device close to a real radiation measurement equipment.

There are other places of improvement too. The voltage applied across the ion chamber could be raised by more than an order of magnitude, to something like 400V. This change would make the device faster (respond with sharper pulses), and more sensitive, since ions and their electrons would be ripped apart by the large voltage potential, preventing recombination, and pulling them quickly to the electrode and the wall.

I'm considering to do both changes to the current design.