How to make a peltier cooled cloud chamber. No Dry Ice!

Part 3: Assembly.

These pages provide instructions on how to build the peltier cooled cloud chamber shown on my Youtube channel.

Part 1: Shopping list and some design considerations

Part 2: The electronics

Part 3: Assembly

Part 4: How to use

Part 5: Pictures, Videos and some explanations


Now the materials are sourced, and you have calculated all the dimensions of your system, time to start creating your cloud chamber.

Preparing the components for assembly:

The spacer block that defines the distance between the heatsink and cold plate sets the thickness of the closed cell foam used to insulate them.

Total thickness of foam and the support plate will be 1...2mm thicker than the block height + 2 peltier cells (7mm). This extra 1...2mm thickness allows the cold plate to slightly compress the foam holding it in place and removing air pockets from between the layers. This foam insulation can be made from multiple layers of thinner insulation if 2 or 3 layers of suitably thick foam is not available.

  Peltier cell and insulation
Cut your foam pieces to be a snug fit around the spacer block and an easy fit around the peltier cells and approx' 1mm all larger than the cold plate. eg: If cold plate is 80 x 60 the foam pieces should be about 81*61, about 0.5mm protruding from each edge. See pic' on right. Any foam (debris, dust etc) that gets trapped between the peltiers (even tiny fragments) will cause mechanical stress on the cell and create a local hotspot. The slight oversize of the foam will create a snug...almost tight fit between the foam and polystyrene base insulation.

Cut small channels in the foam for the wires from the peltiers to exit. Ensure that the wires exit the peltiers straight - do not bend the wires near their cell connections, they could easily break or damage the cell. Also cut holes for the nylon fixing screws to pass through.

  peltier cell insulation
Cold plate   Ensure that the 'best' side of the cold plate is in contact with the peltiers. The flatter and smoother the contact surface the better.

4 holes, 3.5mm and countersunk, are drilled into the cold plate. These are for the nylon screws to fix the cold plate to the spring bars. Put the holes as near the corners as practical and make the spring plates to match.

This pic' is before the plate was sprayed with 'calkboard' black (flat, mat) paint.

Note: Apply several thin layers of spray paint as instructed on the can but leave the paint to cure in a warm place for SEVERAL days. The plate gets immersed in pure alcohol at very low temperatures and any residual solvent in the paint can react and soften the coating. Do not wipe the painted surface or touch with anything hard when cold - the paint becomes VERY brittle.

Cloud chamber support plate   The mounting plate. 3.5mm perspex. Square hole cut in the centre for the spacer block, 4 holes cut to pass the 4 nylon mounting screws though to the spring bars. These holes are larger than the m3 screws as I had to use extension pillars from the spring plates.

This is a failed disk (centre hub cracked) from a wimshurst machine I built.

The support plate needs to be large enough to cover your watertank (or at least most of it - a small gap is useful for adding ice or topping up) and be sufficiently stiff to support the system stably.


Top of Page.

  The dome. A simple perspex 'pudding' bowl. Get a bowl (or similar) that is 'water clear' approx' 15cm diameter and 10cm or so high, try and avoid ones with large thick rims.

The size is not critical but it needs to easily clear the corners of the cold plate to leave some space for insulation. The air next to the dome wall is heated the most by conduction through that wall. So having 1...2cm gap from the corner of the plate to the dome wall is desirable.

A perspex 'box' will work but is harder to source.

Glass domes look great, but glass is a much better thermal conductor than perspex and will impact the minimum temperatures that can be achieved.

Cut a disk from the 20mm (or so) polystyrene sheet to fit loosely in the bottom of the dome - about 1.5mm gap all round to fit a strip of 2...3mm foam.

A hole for the cold plate and foam insulation is cut in the centre of the polystyrene disk. This hole needs to be approx' 0.5mm larger all round than the cold plate. The foam filler below the cold plates should be a snug...almost tight fit when the polystyrene base is fitted. Cut grooves for the peltier wires to exit.

Cut a disk from the black card the same size as the polystyrene disk and fix, with PVA wood glue, to the top of the disk, see pic' on right. Leave compressed (card downwards on a flat surface) for at least 24 hours to allow the PVA adhesive to set. This card surface must be as flat as possible. It will bond to polystyrene with PVA glue but not strongly. Cut a hole in the card matching the coldplate hole in the polystyrene.

Cut a strip(s) of the thin 2...3mm foam sheet and fix to the outside of the polystyrene disk edge again with PVA glue, see pic' on right. This provides a flexible surface for the dome to seal to, the foam compensating for small errors in the polystyrene disk. This also protects the polystyrene from damage when the dome is removed and replaced.

I found that PVA wood glue fixed the foam to the polystyrene reasonably well.



The fence is just a strip of thin clear polycarbonate.

Add 5mm (height of the fence) to the thickness of the cold plate (mine is 6mm thick, therefore 11mm). Cut a strip of the polycarb' 11mm wide and long enough to wrap around the cold plate with about 20...40mm overlap. (Plate 60*80 = strip 300mm with 20mm overlap)

Carefully bend the strip so that it fits snugly around the cold plate with the overlap on one of the short ends.

Shown fitted to the coldplate in pic' on right.


Spring bars.

The design of these will depend upon your heatsink. If you can find a heatsink with ledges or space for bars to go straight through: pic right of a heatsink with 'small' ledges. Then straight bars, with cutouts if necessary will fit.

If your heatsink is 'all fins' then cut 2 spring bars so that the ends can be folded to provide lands for the screw fixings. See sketch right.

These spring bars should be made of steel, however aluminium will work - it's just not very 'springy'.

When tightening the coldpate screws:

Steel spring plate:  Tighten until screws feel tight, steel (if thin) will bend very slightly.

Aluminium 'non-springy' plate: Tighten the coldplate screws until the aluminium (if thin, less than 1mm) just starts to 'give'. If the screws are not at least showing reasonable resistance (tightness) when the plate starts to give the aluminium too soft or too thin/narrow.


Lower foam protector. Cut a disk of the 2...3mm foam the same size as the polystyrene disk. Cut a hole in the centre for the aluminium block. Cut 4 holes to align with the nylon fixing screws. Cut 2/4 holes for the heatsink to mounting plate screws.

Pic' on right shows disk before holes for the nylon screws were cut. 'O' is for orientation. I only have 2 screws to fix the mounting plate to the heatsink.

This adds one layer of insulation and also provided space for the heatsink...plate mounting screw heads.

Top of Page.


Assembling the cloud chamber:

Fix the heatsink (flat side up) to a worktop/bench/vice so as to provide a stable platform for the build.

.Ensure the area is clean/dust free. All mating surfaces (heatsinks, spacers, mounting plate) must be spotless to prevent bad contact between the various components.   'Run through this procedure and dry' fit all the components together separately to ensure that they fit properly.

Check that the nylon screws are long enough to fix the assembly. Use threaded spacers if sufficiently long screws cannot be found.

Have some thermal paste available - only a small amount will be required - and a plastic/wood spatula to spread the paste. A metal one will damage the aluminium surfaces. Do not use the highly popular 'pea' or 'blob' methods shown online / youtube - these are NOT manufacturers recommended practice and you will end up with thick (and therefore unstable) and uneven thickness resulting in less effective thermal connections.

If a spacer block is used apply a very small amount of thermal paste to the side that will mate with the heatsink. Spread this out with a soft spatula or similar (a blade will scratch the surface!) Place the block on the correct spot on the heatsink and, applying light but increasing pressure, work the block in a small circular pattern. You will feel the change in resistance as the layer of paste thins out and excess is squeezed out at the edges. We are aiming for the thinnest possible layer of paste whilst not having any dry spots.

This is a process worth practicing. When the final assembly is under way it's better not to guess when it 'feels right'.

Short video on right demonstrates this process but with an old PC processor.


  Fit the mounting plate onto the heatsink and fix with 2 or 4 machine screws. These should be 'cheese' head or similar and may need a spreader washer if brittle (perspex!) materials are used.

Careful not to 'lift' the aluminium block (if used) whilst settling down the plate. The block can move sideways (a little) without problems.

In this image the 'spring bars' are fitted and have the 10mm nylon threaded standoffs to 'extend' the 25mm screws. The heatasink I used has excellent flat areas for the bars.

  Place over the aluminium block the 2...3mm foam disk (I used 2).

Place over the block the first piece of foam insulation (or more if you are building up from thinner layers). The top of the block needs to be clear so we can 'work' the peltier cell into it.

  Apply a very small amount of thermal paste to the hot side of one peltier cell. Spread this out with a soft spatula or similar. Place the peltier cell onto the top of the heatsink/block assembly and, applying light even pressure, work the cell in a small circular pattern - see Thermal paste video. You will feel the change in resistance as the layer of paste thins out as excess is squeezed out at the edges. We are aiming for the thinnest possible layer of paste whilst not having any dry spots.

Repeat this process for the second peltier cell, working it into the top of the first cell. Ensure that you have the wires exiting on the side you have cut the channels in the foam insulation.

  Loosely fit the spring bars into the heatsink - hold in place with blutack if necessary.

Fit the final piece(s) of foam insulation over the peltier cells. Ensure that no stray fibres, particles etc are on the peltier cell.

Apply a very small amount of thermal paste to the top peltier cell. Spread this out with a soft spatula or similar. Place the cold plate onto the peltier stack, applying light even pressure, work the plate in a small circular pattern - see thermal paste video. The amount of pressure will be larger that the earlier thermal paste processes since you will also be slightly compressing the foam insulation - however it's still possible to feel when the paste has spread sufficiently. Once the paste is evenly spread keep steady pressure on the cold plate to hold it in place until the nylon screws are fitted and 'finger' tightened.


  Loosly fit the 4 nylon fixing screws through the stack, they should protrude far enough to be bolted to the spring plates. I fitted 4 x 10mm threaded nylon pillars to the spring plate - my 25mm screws were too short.

Gently and evenly finger tighten the 4 screws. The foam must be compressed whilst these screws are fitted, if the plate lifts up at all from the peltiers it MUST be removed and the thermal paste process repeated.

Once evenly finger tight, tighten each screw in a diagonal pattern no more than an eighth of a turn at a time. The plate MUST compress the peltier stack evenly. Tighten the screws in this fashion until the spring plates begins to bend - see note here - or the screws are getting tight.

The tighter the compression the more effective the thermal contacts, nylon screws are a compromise as they limit the pressure to a few Kg, but their thermal properties outway that limitation.

If the cut channels for the peltier wires are a bit oversized cut thin strips of foam approx' the same size as the channels. 'Feed them in' over/around/past the wires to close off the gaps. Try and 'feed' the strips deep into the foam as close to the peltiers as possible, without stressing the connection wires.

The stack is now completed.

Top of Page.

  Connect the 4 peltier wires to the terminal block/connector of your choice. Ensure that you know which wires are for which cell. Putting 12v onto the 'cold' peltier and 5v onto the hot peltier will result in an interesting, if not fatal, thermal condition in the stack.



Gently fit the polystyrene base insulation over the stack. Ideally the coldplate will be about 0.5 to 1mm above the surface of the base. If much higher add layers of 2...3mm foam under the polystyrene - cut a cold plate sized hole in the foam sheet - no need to dismantle the whole stack.

If the coldplate is below the level of the polystyrene base - review your measurements! There are 3 options: First option: Remove the bottom layer of 2...3mm insulation - cut it around the foam coldplate insulation, do not pull it out as this will change the pressures within the peltier stack and may loosen the insulation foam. Second option: Dismantle the peltier stack and use a larger spacer block. Third option: Use thinner polystyrene sheet for the base insulation - do not go below 15mm or thermal performance will be degraded. 


The system can now be tested for operation.

It is wise to check the peltiers before applying full power for any more than a few seconds. With a current meter in circuit apply 5 volts to each peltier in turn for just a second or two. Ensure the current is about 1.8...2 amps in each cell.

If at 5 volts both cells currents look good, apply 12 volts to the lower (hot) cell. Ensure the current is in the 4...5 amp range for a 1206. The coldplate should noticeably cool down even in just a few seconds

If the hot cell and cold cell currents are in range, connect the hot cell to 12 volt supply, the upper (cold) cell to a 5 volt supply and switch the power on for a moment (no more than a few seconds). The coldplate should be noticeably cold and the heatsink warm. If so - it's working.

Note:  Do not run the stack for more than a few seconds without watercooling the heatsink. Over 60 Watts are being dissipated and the heatsink will warm up very quickly.


Fit the fence.

The fence should slide snugly between the coldplate and the polystyrene insulation. Gently work it in so that the fence is an even 5mm or so above the coldplate all round. The overlap must be at one short end to improve viewing/lighting.

It should slide down until it meets the foam insulation below the coldplate. Take care not to damage the fine temp' probe wires if installed. 

  Water cooling.

Any suitable container will suffice. What is suitable will depend upon the dimensions of your base plate. As described here, this system running with 2 x 1206 peltiers at 12v and 5v heats water at about 1C per Litre per minute. So 15 litres will heat up at the rate of 4C per hour.

A flat sided container is preferred since you will be able to mount the water pump directly to the side. A small/medium aquarium is ideal, although a basic bucket, large paint container, almost anything watertight and suitable dimensions could be used.

Fit the plastic 15...20mm tube over the exit pipe of the pump and arrange it so that the water flows directly through the fins under the peltier cells and washes along the flat surface between the fins, see here. and pic' on left - water forms a 'bulge' where the heatsink would be, 'washing' along the heatsink base plate between the fins.


  Fuel supply.

Cloud chambers will work with almost any low boiling point solvent, I have even had this one working with water, although the results are indistinct and the coldplate soon forms into a frost garden, see here.

The fuel supply is stored in a felt pad held at the top of the chamber by magnets (one on the steel 'star' the other outside the container).

Cut a disk of felt about the same size as the flat(ish) base of your bowl (dome), or about 5cm. I made a 'star' (perforated steel, old front cover panel from a PC) to ensure the felt is pinned to the top of the dome- see pic - this helps with heat transfer and slows the cooling of the fuel supply. If the felt/fuel overcool they stop producing sufficient vapour for clear operation of the system.

I often use a reel of solder as a 'gentle' heat source to stop the fuel overcooling. If the felt was 'loose' of the top of the dome then the heat transfer to the fuel would be very limited and the system stop working very quickly at low running temperatures.

If you are only building a basic unit (no optional ion scrubber, fans, lamp power etc) then you are ready to start running your cloud chamber.



Top of Page.

  Ion scrubber, optional. A pair of wire loops that fit over the dome.

Make insulated wire loops from the solid cores of some 'ring main' cable. Larger loop fits easily over the base of the dome, the smaller loop is about the size of the flat(ish) base of the bowl (dome). The small loop is held in position over the dome by extending the wire from the loop over and bent down straight to the insulated connections. the loop terminations are crimped/soldered together then insulated with minimum 2 layers.

The connections of the solid conductors to the double insulated mains cable must be individually insulated then sleeved together - heatshrink tubing is excellent for the job. Ensure that NO BARE METAL of the conductors is visible anywhere. These will be at high dc voltages (350 volts). The 2 x 470K / 1M resistors fitted to the power supply limit the short circuit current to a few hundred microamps so it should not be more than a minor jolt - but why risk it - INSULATE well.

I used a power supply extension plug to connect the cable to the high voltage output described in the 'Electronics' section. Removing all pins from plug and socket and reconnecting and re-inserting only the outer two to the high voltage lines.


  LED Light shroud.

The lighting of a cloud chamber to get the best viewing results is critical. Many online designs show the lamp at 90 degrees to the viewing angle - this works but is by no means the best angle. The best views are achieved when the light and the viewing angle are almost coincident - see diagram below and Videos, Pictures and explanations


However to achieve so close an angle of view to the light source the light must be masked to form a reasonably confined beam - see pics' left. This is why a LED lamp with a line of LED's is better than a lamp with a circular cluster.

Make a card tube that fits over the lamp you use. Create a mask with a rectangular hole that can be slid up and down the tube. The exact dimensions of the tube and mask will need fine tuning to get the best illumination.

Position the tube and lamp at the edge of the polystyrene base and using foam blocks or similar adjust the height of the lamp so that the beam just 'skims' the surface of the card on top of the polystyrene. Adjust the mask so that the beam just covers the coldplate at its widest and 'skims' over the surface. Any dust or debris will be clearly visible when the light is skimming the coldplate surface, see image below.

Adjustment of this light takes time - but it is truly worth it. The difference in good and poor lighting can be seen here. Pic below clearly shows dust and debris on the plate all the way across.

Top of Page.


I fitted fans to the light and viewing areas since I wanted to mainly image and video the chamber, demisting the windows every few minutes is a pain!

Viewing window fan. Using the black card make a box that fits around one fan. Fix the fan into one end - ensuring the airflow is into the box. Place inside a deflector panel to push the airflow towards the dome surface. Cut a 'viewing window' into another piece of card and fix to the dome side of the fan box so that the airflow out of the fan washes across the 'window'. See pic' on left. Improved de-misting is achieved by fitting a 'fence' around the window to confine the ariflow as seen in pic.

Lamp fan. Take the lamp shroud made above and remove a section of the top the same length as the fan, obstructing half the fan as seen in the pic's is not a problem: Minimal, but sufficient, airflow is the aim. Fit the fan to cover this hole, airflow into the shroud. The air will then flow out of the end of the shroud and (if the shroud is mounted close to the dome) wash directly onto the dome only in the area of the light entry point. The position of this fan limits the range of positions the sliding light mask can take. Ensure that you shape a mask that works when it is 'behind' the fan.

It may seem OTT to keep the airflow from washing over more of the dome but the outside air is a significant (if not the most significant) source of heat leakage into the chamber. The less air that moves the less heat is transferred. See Pictures, Videos and explanations for more info'.

That's it, its built and ready to be tested for real.


Part 1: Shopping list and some design considerations

Part 2: The electronics

Part 3: Assembly

Part 4: How to use

Part 5: Pictures, Videos and some explanations



How to make a CD wimshust machine - A 9.3 minute video providing full instructions on how to make a wimshurst machine from CD blanks and household items.

CD wimshurst machine 1 - wimshurst machine, cobbled together in an afternoon from scraps and a couple of CD's.

CD winshurst machine 2 - Short video of the wimshurst machine made in the 'How to build a CD wimshurst machine' showing max sparks of 15mm and overloads.

32cm wimshurts machine - A short video of a wimshurst machine with 32cm disks, this produced over 130mm sparks and is a very reliable wimshurst machine.

Cloud chamber 1 - 1 short video of an Alpha source (Am241) and background radiation in a homemade peltier cooled cloud chamber without dry ice.

Cloud chamber with Water as the 'fuel' - a short video demonstrating water as active vapour, very quickly frosted up! How to make a cloud chamber without dry ice.

Poorly Lit homemade cloud chamber- contrasting good and bad lighting in my homemade peltier cloud chamber! No dry ice.

Granite in a cloud chamber - no dry ice - some activity from a piece of granite. No dry ice, peltier cooled.

Background radiation in a peltier cloud chamber - no active sources anywhere near, just background radiation. Peltier cooled, homemade cloud chamber not using dry ice.

How to make a cloud chamber - a 5 minute video on how to build a simple cloud chamber in less than an hour. How to build a cloud chamber without dry ice, peltier cooled.