Thermoelectric/Peltier Devices

Properly insulated it should work...Part of the equation here is heat load, how rapidly the heat is transferred from ambient to small surface..and is it generating heat itself

http://www.thermoelectric.com/2002/tt/index.html

http://en.wikipedia.org/wiki/Thermoelectric_cooling

TE coolers are used in astro-cameras used by astronomers. Typically these actually only yield a 20C drop below ambient. To get a drop of 70C you might need a 2 or 3-stage TE cooler. Keep in mind that in your set-up you may not achieve the performance the manufacturer's data suggests. Here's an example of one company's 2-stage TEC.

Check the melting point of Peltier devices - most I have seen use low melt materials

We used Peltier devices in an attempt to keep the InGaAs Near and Short Wave Infrared arrays at optimum temps. Ambient was not nearly at hot as your environment and we were happy to get the array temp cool by just a few degrees.

Peltier technology development has to go a lot further before it an deal with your environment.

I side with Usbport on this. It ain't gonna work. Your expectations will not be met.

L.J.

I agree with Laughing Jaguar and USBport - a temperature difference of 70degC -it will not work even not when usig two stage designs. and 120 degC is too much for most of the Peltierelements. Would it be a possibilty to go the PC / or notebook heatpipe systems they are cheap and simple to use or the watercooling systems for CPUs then you have a real chance

You say 'hot environment', the peltier will only work if you can effectively remove the heat it generates in operation. I tend to agree with the other comments here, we manage to achieve temperatures of -40°C with 3 stage peltiers but only in a small, well insulated area with chilled water extracting the peltier heat through heat exchanger blocks.

Have a look at some of the spec's here

http://www.customthermoelectric.com/tecs_imax.html

Although the maximum differential temp is 70°C, even at 60°C the amount of heat you can extract is almost zero.

So to use TECs you would almost certainly have to go to a stacked solution. That's a bit like launching a rocket: most of the fuel used is to get the other fuel into the place you want it.

How "big" is your ambient environment: could you pipe some water in and out.

They are the ones I am looking at, specifically the mini-TEC. I have a call into them, so I'll see what they say.

It's always good to have feedback from those with experience. Sometimes "application engineers" at companies just know how to read the spec. sheet.

A little more detail. I have a sensor chip roughly .25" square, 35 mg. It must operate at <100C, but the lower the better. The ambient temperature is 120C. The chip cannot be insulated, it will be exposed to the ambient temperature.

There is no way to cool with gas/liquid. If there was, I would not need the Peltier device.

The sensor chip has a coating. Since I assume that thermal mass is a big issue, I may be able to coat the Peltier directly. This would bring thermal mass to be cooled down to less than 1 mg.

I thought thermoelectric coolers were a long shot for this application, but the TEC specs. seems tantalizingly close.

The thing about peltiers is that they are only as effective as the method of heat removal used on the hot side. So you need to have some kind of heat dissipation method in place before it will work. So You *might* be able to get it to function if you can install a water cooling system to the hot side of the peltier and pipe cold water from a chiller to the peltier.

While lots of people are saying 120C ambient is too hot for a peltier, thats assuming heat dissipation is using the ambient environment. If you have a water block installed on the hot side of the peltier and you can keep the water block at say, 60C, then that becomes the peltiers ambient temp, and it will cool your sensor to 20 degrees below that.

Where peliers really shine is at low temps near freezing where water cooling just can't take you further. From your description, if you can pump in cool water from somewhere outside your ambient area, I would think that a water cooler is really all you need. Since you are operating near the phase change temp for water - thats is going to be the most effective coolant I would think.

The application engineer seemed very knowledgeable, and gave some good suggestions. He felt strongly that the application is possible under the right (perfect) conditions.

So, a bit more encouraging than most posts here.

It sounds like a long shot, but encouraging enough to test.

A 70 degree differential is possible with a multistage cooler. Your problem is that you will be rejecting heat at a higher temperature than coolers permit. About 100 degrees C is the hottest that these coolers can be operated at.

I have encountered much worse. I installed a laser triangulation unit and had to cool a front surface mirror. I mounted a water cooler directly to the back of a Peltier ,which was attached to the mirror, unit and sent the water to a chiller. The mirror did not distort at temperatures of 1600 degrees F. with a short term exposure.