Peltiers

Who needs that big chunky freezer of yours?

Peltier devices are literally heat pumps, which have two sides; a hot side, and a cold side. When 12 - 15V is applied (usually), heat magically gets pumped from the cold side to the hot side via junctions. So the cold side gets cold, and the hot side obviously gets very hot. Self explanatory really... The whole point is, to cool stuff down to below freezing. I'm not going to elaborate into too much depth on how they work, because that's what google is for. Peltier devices have different power ratings, corresponding to how fast they can cool something down. They also have this magic number, Delta T, which is the maximum thermal difference in temperature between both sides.

So I managed to pick up a standard 80W unit, 65 degrees Delta T, for $11. I guess I could have scrounged around a bit on ebay, but nah. Okay the deal with peltier devices is that they SUCK until you have a mega heatsink and fan. First few times I tried with a wimpy heatsink and the heatsink got pretty damn hot, and the "cold" side stayed at room temperature (if not slightly higher).

Oh, I should say right now, that the figure 65 degrees Delta T means that if the hot side is at 50 degrees (compensating for some degree of heating), then the MINIMUM temperature you'll ever get is -15 degrees. So the bigger heatsink you have, the colder (relatively) the hot side will be, thus the cold side will be even colder. In reality this is impossible, but you can get close...

After much scrounging around, this mega heatsink that came out of my old computer finally made something happen. Well okay, i did get some water condensing on the peltier with a crap heatsink, but that wasn't good enough for me.

I set this up on a drowsy sunday morning, hence notice the rubber bands. That's a high speed fan mounted to the heatsink, and a big battery to the left. A 10 cent coin on the peltier ready for freezing...And it works! Surprisingly well!

Here is some frost and chunks of ice forming on the coin half a minute after power up. It cools relatively slowly, as can be seen from those frozen water droplets on the coin.

Here's another go at it, again with ice chunks forming where water had condensed onto the block.

I decided it was cooling way too slowly. I mean.. half a minute, who on earth would wait that long. I had this idea of dumping the fan, and just dropping the heatsink into ice cold water. Yeah, cheating, I know. But hey. This occurred within 10 seconds of power up, and see how quickly it's frozen over. No random chunks of ice anywhere, just nice even frost. Notice how the rubber bands are frozen too. Oh, its a bigger coin too if you haven't yet noticed. One of our newer 50 cent ones.

Very self explanatory photo I think... just to show the depth of the ice/frost coating.

And an extreme close up of the ice crystals. Frozen so fast that they don't even have time to form a smooth surface. Kind of like up there in the clouds, you know, up there?

Okay, new peltier setup! The previous setup proved to be a pain in the a** for getting rid of the massive amounts of heat driven from the peltier. Charles Darwin University threw away this beautiful huge heatsink, so I got it for free, the only price I had to pay was the labour involved in dismantling all the crap that it had mounted to it before (see all the holes?).

The fan below is a high velocity one given to me by Steve and Shaun, thanks guys!

Any water on the surface of the peltier freezes within two or three seconds upon turning it on. Here is a small button cell battery being frozen to death, and a sultana having the worst time of its life.

A closeup of the ice and frost covering the surface of the button cell battery. The sides are covered in ice which is the result of water trickling down just before the whole cell drops to below zero temperatures and freezes over.

Frost starts to form out of control everywhere on the peltier after a while.

Another shot from a slightly different angle showing the silver heatsink paste used to provide thermal conductivity between the hot side of the peltier (bottom), and the surface of the heatsink. This stuff is very expensive and is normally only used for CPU mounting. I only use it because I can get it cheap and don't have any other types of heatsink paste anyway...

Each time the peltier is turned off, the ice slab melts dramatically fast, and water rushes everywhere...even in the peltier itself, which would provide a short circuit if it was turned on again. So I have to put it in the sun to dry everytime after I turn it off... I should really get around to sealing the peltier element with silicone or something.

The recently acquired 1-Wire data loggers were put to use to measure the temperatures and humidities reached by the peltier device, and comparing the peltier element either attached to a heatsink plus fan cooling, or just a bare heatsink.

The setup was as above, with the data logger pressed down against the surface of the peltier by means of a weight (separated via plastic insulation to avoid thermal conductivity between the weight and the data logger). To attempt to minimise random error, the setup was not disturbed during logging. The peltier was switched on for 6 minutes and 42 seconds for each trial. On the graph, the switch on time corresponds to time = 70.

From the graph above it is immediately obvious that the difference between using fan cooling is minimal, although this could of course been due to the angle of attack of the wind against the heatsink. The peltier managed one degree cooler with the fan assisted heatsink, which isn't very erm, well satisfying really. Notice how the temperature begins to climb slowly after the minimum temperature is reached, as the heatsink starts to become warm and useless.

Note: Peltier switched on at 70 seconds, switched off at 470 seconds

Lowest temperature (heatsink only): -4.960 °C

Lowest temperature (heatsink and fan): -5.977 °C

Note: Peltier switched on at 70 seconds, switched off at 470 seconds

This graph shows relative humidity of the surface of the peltier, and is somewhat related to the temperature, as the lower the temperature, the more water is going to condense onto the peltier element. It is of no surprise then, that the average humidity of the fan assisted heatsink setup is slightly higher than the heatsink-only setup.

After a long break and following the acquisition of a new 3 speed 150W industrial fan, the possibilities of revisiting the peltier project were immediately obvious. The new setup consists of the heatsink mounted directly on top of the high powered fan, and the peltier element is insulated by means of a foam box in order to minimise the effects of the high velocity winds affecting output temperature.

Note: Peltier switched on at 0 seconds

The results were impressive, with all three fan speeds achieving peltier temperatures of less than -12°C. The inverse relationship between wind speed and temperature is clear, but I suspect there would be less of an effect (perhaps up to a point) as the wind speed is increased even further.

Lowest temperature (Fan speed 1): -12.213 °C

Lowest temperature (Fan speed 2): -13.349 °C

Lowest temperature (Fan speed 3): -14.233 °C

Also note that there was never a distinct "minimum" temperature reached, unlike the previous setup where temperature would drop to a minimum then start to rise as the heatsink warmed slowly. Presumably the fan has such cooling power that it is able to maintain the heatsink at a constant temperature. Tests on this later showed the heatsink temperature to rise slowly and stabilise at around 35°C over time.

I believe with proper insulation and heat transferral, the minimum temperature can be lowered even further. Future setups may incorporate water cooling, but at the moment I am still trying to maximise results using only air assisted cooling.

This space is intentionally blank