You will need the 'fuel':  Isopropyl alcohol is very effective producing good clear results, but the cloud chamber will work with almost any low boiling point solvent. Heed the warning on the left when using flammable solvents. This chamber has produced results using water as the fuel.

For testing I use an Americium 241 (Am241) 'puck'. This can readily be removed from an old smoke detector, many online videos demonstrate. I have embedded mine in a piece of foam so it's easier to handle.

A piece of granite works and there are many radioactive minerals available online (Ebay etc) that will work well in a cloud chamber. Got an old luminescent watch/clock dial - that paint is lively!

Preparation.

Setup your system on a strong bench/table - the water tank will be very heavy when filled(1Kg per litre) with sufficient room on the bench/worktop/table for the power supply and sundry items (keep them well clear of the water tank).

I have run this with saltwater brine (250g salt per litre of water, slushy but just about liquid at -15C) cooled heatsink to below -10C, but it's messy - have to rinse everything in fresh water or get salt crystallising everywhere. Coldplate went below -58C!

Place the cloud chamber assembly onto the watertank, best if the heatsink is clear of the edge so that water can exit from between the fins easily.

Setup the pump (extension tube if necessary) so that the water flows up into the fins directly onto the base metal plate between them. It should 'cling' to the fins and then exit above the waterline if the base is above the waterline. See sketch

Level the cloud chamber, use a spirit level if possible. The coldplate needs to be level to reduce convection currents within the chamber and keep the coolest air over the coldplate. Careful not to spill the water!

If your using a temperature probe Set the probe tip onto the coldplate - not onto the nylon screw heads. A thin film of fuel will help with the probe thermal contact with the coldplate. The temperature difference between the corners of the plate and the centre is negligible.

If your using an unmodified probe then you will need a 5mm(or so) hole cut through the polystyrene base and black card top - to pass the probe through. Pass several cm of the wire into the chamber so it's cooled and does not conduct heat to the (relatively huge) probe and affect the temperature reading. Seal the 5mm hole with foam strip/tube.

Test source. Put a small cut in the side of the foam block the source is embedded in. Fit the cut onto the fence to hold the foam block just off the surface of the coldplate. If it touches the fuel will soak the surface of the foam by capillary action - a thin layer of fuel will prevent most alphas from passing through, so a dry source is desirable.

Position the source as close to the coldplate as possible without the foam touching. The sensitive zone above the coldplate is only a few mm thick.

The exact amount is guesswork but add 2...3 small teaspoons of fuel to the felt. Place the dome on a flat surface to keep the fumes in whilst you prepare the plate.

Drip, evenly over the coldplate, about half a teaspoon of the fuel (don't do this bit if your using water as fuel, you'll just get a insect sized skating rink). The exact amount is guesswork, you'll see when you have the right amount, it will spread over the coldplate in about a minute or so - top to bottom pic' about 1 minute.

This layer of fuel stops droplets of condensate forming on the coldplate and creates an even, 'mirrored' surface at low viewing angles. This 'mirror' film of fuel helps the viewing by enhancing the contrast.

An airtight seal is not necessary, just firmly centre and position the dome.

If you have built an anti-condensation fan position it to clear your viewing 'port'. See pic.

Plug in the lamp power, fans, ion scrubber and peltiers.

Turn on the power supply. No BANG, good -  The coldplate will rapidly cool.

If your plate does not get down to below -26C and your water tank is cooler than +20C, and your not in a 'tropical heat wave' with high humidity then you have problems with something. Remove the fans (if fitted) and illumination and just run the cloud chamber. If it still does not get cold: Why won't it work?

If the mist is consistently flowing in one direction adjust the levelling a mm or two at a time. If even extreme adjustment (1cm or so) does not even out the 'flow' check that the dome is on straight, the felt pad centred and there are no gaps around the polystyrene base foam strip to dome mating surface.

The cloud chamber works with a 'flow', but its more effective if you can get the flow to be even, not in one general direction. See sketch.

Solder reel acts as the heat source to keep the fuel from overcooling, doesn't take long below -40C.

 Here coldplate is at -43.1C, water (heatsink) at +14.6C. Temp difference 57.7C, calculated max' temp' difference is 60C, so fairly close - guesswork on the cooling load is probably the error. See explanations

Using this and salt water brine I have achieved my lowest coldplate temperature: -58.5C (assuming the probe is reasonably accurate near its limit).

Poly' bucket for the brine all frosted over - the aquarium may have shattered at the initial brine temperature of -10...-15C. 'Hot' temperature probe is fixed to heatsink. No fans, and the viewing 'port' closed/covered most of the time.

Inside of the 'cosy' was all frosted after the run.

The level of background radiation is what amazed me. The sensitive zone of the cloud chamber is only 60 * 80 * 5mm to see so much activity in such a small volume!

Time to experiment - different sources, different fuels, how low can you go? My record coldplate low is -58.5C as of 20/12/2013.

If you have a strong alpha source - try placing thin aluminium foil in their path and image both sides - see an alpha hit the foil and 'explode' out the other side with all sorts of debris - rare but spectacular, no images/videos of this yet.