5

I have a thin 2x2 mm surface that I'm trying to cool by 5-10 degrees Celsius in 1 second or less. Some very rough back-of-the-envelope calculations reveal that I need to remove ~50 mJ of energy to do this.

I've been thinking of using a Peltier module, but have no real experience with TECs, and therefore don't know exactly what to look for in terms of specs. The speed of the temperature change is my primary consideration. My questions:

  1. Are there any manufacturers that sell very small Peltier modules that would roughly match my specifications, in prototyping quantities?

  2. Are there any further considerations that I've missed?

Brian Carlton
  • 13,252
  • 5
  • 43
  • 64
DNS
  • 203
  • 2
  • 5
  • 1
    I'd think the most significant factor would be the heat capacity (thermal mass) of your surface, plus the junction, plus the cooler. – Nick T Nov 19 '10 at 17:45
  • The cooler thermal mass won't matter, will it? – markrages Nov 19 '10 at 18:35
  • @NickT, He specifies how much energy he must remove, by doing this he is specifying the heat capacity I believe. – Kortuk Nov 20 '10 at 01:40

5 Answers5

3

I found some information by looking around online.

Here I found some mini-TEC (ThermoElectric Coolers). They seem to be able to do Watts, which I think converts to joules per second, with an efficiency of 5% they should be able to do what you need.

You may need to buy a few and test them, or tomorrow I will look on their site more and get more info. I know that this is done, so I am sure you can find a way. They are not cheap though.

They are also quite small. I hope this helps.

Daniel Grillo
  • 7,659
  • 18
  • 51
  • 69
Kortuk
  • 13,362
  • 8
  • 60
  • 85
  • 1
    Those are tiny! It's really too bad about the cost - somewhat outrageous but a small market size I guess. – user391339 Feb 13 '14 at 03:44
3

As well as the other stuff people have said, you may also need to be careful about the lifetime of the modules under such aggressive cycling. Differential thermal expansion rates mean cycling is a cause of 'wear' to Peltier modules, and it's reasonable to think that the faster the ramp-rates are, the worse this wear will be.

You can certainly get small coolers, though - somewhere like here, for example:

http://www.kryothermusa.com/indexf526.html?tid=48

1

Small coolers of this order of size are used inside laser diode modules for fibre-optic comms- not sure how easy or otherwise it would to obtain seperately. Look at datasheets for peltiers to get an idea of heat pumping capacity versus size and power.

mikeselectricstuff
  • 10,655
  • 33
  • 34
0

Just obtained some today. The big problem is that they are a lot more fragile than most. The ones used in Casio projectors are believed to be very efficient, possibly 1.3* better than units found in typical camping refrigerators. You may be able to use one of these as part of a stack.

16006774
  • 1
0

Could you use a hybrid mechanical/electrical solution? You pre-cool a relatively large block of aluminum to a very low temperature. When you want to cool your mystery object, you quickly drop it on top (or maybe you drop the mystery object on the block instead).

I suspect that the limiting factor will be the thermal conductivity of your surface. If you really need to cool just the surface, just clamping on a cold block will do it instantly. To the extent that you need the surface plus a small depth, you just need to be sure that the block is much colder and more conductive than your mystery object.

pingswept
  • 12,581
  • 4
  • 46
  • 64