Reducing thermal insulation in TEC (for better cooling of a small space)

Hello all,

I am working on a project that involves using heat to control the color of a material (thermochromic properties). At room temp, the material is clear, and heating makes it change color. In particular, I am using TEC modules to heat/cool a space that takes up approx. 1x1x1 cm^3. Unfortunately, there is too much thermal insulation which makes cooling quite slow/difficult.

I've looked into some ways to control thermal insulation: pumping air into perforated tubes, closed circuits made of silicon tubes with hot/cold fluid passing through, PGS (pyrolytic graphite sheets which demonstrate excellent thermal conductivity).

I'm wondering if anyone has any different approaches or augmentations to the TEC modules that might be feasible for tackling this problem.

Thanks!

It is hard for me to imagine your setup and to understand why cooling 1 cubic centimeter should present a problem. Perhaps if you posted some photos or a clear drawing, you would get some helpful responses.

I assume your TEC is a Peltier device. Usually these can cool as well as heat. Does the one you're using support cooling? Have you got good thermal connection between the 'waste' side of the device, and the outside world?

I'm not a thermal expert, but can you use a heat sink (something similar to [u]this[/u])?

You reverse the current to activate cooling, right? But, there is still net-heat generated so as peter suggested you can't cool an enclosed space without someplace for the heat to go. A heatsink on the "opposite side", outside of your thermal module, should make the device work better.

jremington:
It is hard for me to imagine your setup and to understand why cooling 1 cubic centimeter should present a problem. Perhaps if you posted some photos or a clear drawing, you would get some helpful responses.

The 1 cubic cm was an estimation; the heating/cooling is intended to be located within the fingertips of a model baby.

PeterH:
I assume your TEC is a Peltier device. Usually these can cool as well as heat. Does the one you're using support cooling? Have you got good thermal connection between the 'waste' side of the device, and the outside world?

Yes it is a cooling Peltier device, and it will be embedded within the fingertips of a model baby, so there is no "waste side," I suppose.

DVDdoug:
I'm not a thermal expert, but can you use a heat sink (something similar to [u]this[/u])?

You reverse the current to activate cooling, right? But, there is still net-heat generated so as peter suggested you can't cool an enclosed space without someplace for the heat to go.

If I am not mistaken, TECs consist of heat sinks to control thermal energy. Actually, they can’t generate cold; they only move heat from one side to the other via a DC current reversal mechanism.

Yes it is a cooling Peltier device, and it will be embedded within the fingertips of a model baby, so there is no "waste side," I suppose.

When you power the Peltier it will have a hot side and a cold side. It isn't totally efficient so if the hot and cold side are in the same space then the nett effect will just be a small amount of heat output. In order to get cooling, you need to connect the hot side to something outside the space you are trying to cool. It will also work better as a heater if you connect the 'cold' side to something outside the space you are trying to heat. If you're stuck for space, perhaps you can flow some water past the 'waste' side and use that to carry the waste heat/cold to a remote heatsink.

PeterH:

Yes it is a cooling Peltier device, and it will be embedded within the fingertips of a model baby, so there is no "waste side," I suppose.

When you power the Peltier it will have a hot side and a cold side. It isn't totally efficient so if the hot and cold side are in the same space then the nett effect will just be a small amount of heat output. In order to get cooling, you need to connect the hot side to something outside the space you are trying to cool. It will also work better as a heater if you connect the 'cold' side to something outside the space you are trying to heat. If you're stuck for space, perhaps you can flow some water past the 'waste' side and use that to carry the waste heat/cold to a remote heatsink.

Unfortunately I will need both heating and cooling to activate/deactivate the thermochromic color changes. Could you explain the mechanism behind "flowing some water past the waste side and using that to carry the waste heat/cold to a remote heatsink"? Apologies for my lack of knowledge; I have little to no experience with practical thermodynamics.

Edit: Nevermind howstuffworks.com explained it (success!)

Thanks!

I'm currently looking into a fluid cooling system. Has anybody had any success with liquid vs. air cooling? It seems that liquid cooling will be quieter and fit into smaller spaces better, so I may proceed in that route. But I believe I'll need a pump, tubing, and a coolant reservoir?

Yes you would need pipes, water, and some way to make the water circulate. That probably implies a pump. For the heat transfer between the water and the peltier's heat exchanger, I guess you would be looking for some small bore copper pipe which could be mechanically connected to the peltier.

Note that if you take this approach, it isn't absolute essential for the peltier to be in the part of the model that you want to be heated/cooled. You could always put the peltier somewhere else and use it to control the temperature of water that circulates through your test section.

Is it plausible to fit a pump and tubes in a model baby?

I am considering an immersion system using dielectric fluid (computers/servers have achieved good cooling using this); that way there is no need for a pump and tubes. But I am wondering how that will affect the heating capabilities that I will need to activate the thermochromic color changes. Or rather, how effectively can dielectric fluids cool--to the point that my TEC will be incapable of heating?

Edit: ^this will still require a pump to transport the heat away.

myom:
Is it plausible to fit a pump and tubes in a model baby?

I don't see any reason why not. Do you?

PeterH:

myom:
Is it plausible to fit a pump and tubes in a model baby?

I don't see any reason why not. Do you?

I suppose not! I didn't imagine that tubes and pumps could get that small.

Would it be safe to follow the model of computer liquid cooling systems in designing this system? Same idea, just within a slightly smaller scale.

myom:
I suppose not! I didn't imagine that tubes and pumps could get that small.

I've got a plain cheap aquarium pump that's about 30mm on a side which only cost a couple of quid, and they come smaller than that. I imagine the price might go up if you're looking for something really tiny, but I don't think you're looking for anything extreme here. The nice thing about these tiny pumps is that they only need a tiny power supply to go with them - mine is powered by the same wallwart that powers the Arduino which controls it.

PeterH:

myom:
I suppose not! I didn't imagine that tubes and pumps could get that small.

I've got a plain cheap aquarium pump that's about 30mm on a side which only cost a couple of quid, and they come smaller than that. I imagine the price might go up if you're looking for something really tiny, but I don't think you're looking for anything extreme here. The nice thing about these tiny pumps is that they only need a tiny power supply to go with them - mine is powered by the same wallwart that powers the Arduino which controls it.

May I ask what you're using the aquarium pump for? Or more specifically, if it's used for cooling, how much/how precise are you able to control temperature with your system?

It's not used for cooling, it's used for irrigation. There's a decent volume flow rate though, ample for what you would need, and you could probably get away with a smaller pump if necessary.