Anyone use Thermoelectric Coolers before

[rainman]

Hi all,

I'm noob to the heating process and would like to know few things about the
Thermoelectric Coolers.



http://www.marlow.com/Products/produ...p?ProductID=98

1) Why is it call "cooler"? It seemed that it is actually used for heating?

2) What can I understand from the specifications from the above link? Does it mean if I supply 14.7VDC the temp is 27 deg c, and 50 deg c when it is 16.4VDC?

3) How do I control for Thermoelectric cooler? I seemed that DAC controller available on the market does not have voltage or current of that range (14.7VDC to 16.4VDC 5.6AMP)

4) For this usage, do I need a thermocoupler for close-loop control?

Thanks and Cheers!

[johnmac]

When you apply a current to it, one side gets cool. The other side gets warm.

This is what's used in food coolers that plug into the cigarette lighter socket of a car.

I think the temperature rating is the amount of drop from ambient.

John

[martinw]

Dear rainman,

I had to build a cooling enclosure a while ago in order to calibrate some naked temperature sensors that had to be tested in air. The enclosure was only about 12 by 8 by 8 inches in volume. We used a Peltier cooling wafer similar to the spec you posted. Google "Peltier", and you will get a whole heap of information.

It is just a semiconductor heat pump. You supply electrical energy , and heat from one side comes out as "coolth" on the other. The problem is that the device is very thin, and that means it tends to work at very low efficiency if you cannot dump the energy on both sides . You have to use alarmingly large heatsinks, otherwise there is a thermal short circuit across the wafer itself.

Tell me what you want to use a similar device for, and I may be able to offer modest advice about controls. No promises, but maybe...

Best wishes,

Martin

[rainman]

Yup John you are right.

It seemed to be quite a useful device, perhaps I can use it on my overclocked CPU.

[rainman]

Originally Posted by martinw Dear rainman, I had to build a cooling enclosure a while ago in order to calibrate some naked temperature sensors that had to be tested in air. The enclosure was only about 12 by 8 by 8 inches in volume. We used a Peltier cooling wafer similar to the spec you posted. Google "Peltier", and you will get a whole heap of information. It is just a semiconductor heat pump. You supply electrical energy , and heat from one side comes out as "coolth" on the other. The problem is that the device is very thin, and that means it tends to work at very low efficiency if you cannot dump the energy on both sides . You have to use alarmingly large heatsinks, otherwise there is a thermal short circuit across the wafer itself. Tell me what you want to use a similar device for, and I may be able to offer modest advice about controls. No promises, but maybe... Best wishes, Martin

Hi Martin,

Thanks for the useful information, I'm trying to use it in a lab project, where it will specimen temperature is controlled and there are also some pumps and motors involved. The specimen is kept within a 100 by 100mm enclosure.

I had also googled Peltier and now I have some understanding on this cooler, but still I have no idea on how to use it.

Looking at the product specs that I use, does it mean the hot side will not go higher than 50 deg c and why are there 2 types of hot side temperature?


Is it controllable in the first place? Do I just supply a 12V power supply with sufficient ampere? lol

Thanks.
Rgds,
Steve

[martinw]

Dear Steve,

Well, I'm just a "Bush Mechanic" but those two graphs of "hot side temperature" are pretty much the same, so you may as well delete one and assume the whole shebang is the same between the two temperatures.

Sorry. but I'm still scratching my head about the lower one as regards V, I ,Q and T.

Best wishes,

Martin

[vger]

Steve,

I pulled apart a machine that used Peltier's to keep a chamber (10" cube) cool. One entire face of the cube was a massive heat sink with thin fins about 2 inches deep spaced about 1/8 inch apart with 3 Peltier wafers about 6 cm square sandwitched to another massive heatsink and fan assembly. It was capable of lowering the chamber temp to just below 0 deg C. Peltiers, as mentioned are solid state semiconductor devices and as such have a limit on how hot they can get and still work. The current draw will vary somewhat with the junction temperature of the device. You should use a current regulated supply or use a thermistor attached to the hot side to control the voltage (and thereby the current) to keep out of thermal run away.

Yes you can control it very easily. Just vary the current flowing through it. With a PID loop controller you can get very accurate levels.

Need to heat instead of cool? Just reverse the polarity of the peltier.
Need to get to a lower temp? Stack the peltiers in thermal series.
Need to generate a little power? Heat one side and cool the other to generate a current flow (very inefficient but works)
Need to reduce the dark current on your CCD telescope camera? Attach a peltier cooler to the back of the CCD chip.

THERE IS NO SUCH THING AS COLD! You can not make cold. There is only heat, more or less of it.

Just remember not to let the junctions get too hot or the device will fail.

Steve

[martinw]

Dear Rainman,

To control the temperature in our box, we ran the Peltier flat out and used a lightbulb as a heater. The lightbulb was controlled by a PID controller with a PRT as the temperature sensor. (It is a good idea to paint the lightbulb matt black to minimise radiative effects. Stirring the air up with an instrument fan is a good idea as well.

Another calibration box used a scrap beer chiller connected to a motorbike radiator running on ethylene glycol. Again, we ran the beer chiller flat out, and controlled the temperature with PID/lightbulb/PRT. We could maintain the set-point to within 0.05 degrees C, no problem. PID controllers are just amazing if you set up the right parameters.

Best wishes,

Martin

[rainman]

Hi vger,

Thank you for your reply too.

I think I will take your advice to buy a PID controller for it....

I'm looking at this, specially for peltier


Do I have to consider the fact that the controller is able to output up to 5.6Amp?

Thanks.
Cheers!

[martinw]

Dear rainman,

Control algorithms built into PID controllers are constantly being refined, and they may have improved enormously since my first efforts.

Forgive me, but here is my take on the matter.

As you know, a Peltier in your application probably requires a very a large aluminium heat sink inside your controlled enclosure. The thermal mass of the heat sink/cooler is probably orders of magnitude greater than that of the thermal mass of the air inside.

This is not an ideal situation. No matter how clever the PID control algorithms in the PID may be to the Peltier cooler , the feedback loop is extremely sluggish timewise. IMMVHO, you will be far better to use a really thermally "lightweight" heater to control the air temperature, rather than the massive Peltier/heatsink itself. Hence the suggestion of a lightbulb.

I could well be wrong.

Best wishes,

Martin

[CalG]

I have a very nice (bit a bit large) TE control box with a large temperature readout (meter) polarity switching etc. It's 12V at maybe 10 Amps. Made by Denton in N.J.

I think it's mostly a boxed Variac and a transformer along with the temp display. I would need to look at it again.

surplus to my needs if it's something you would be interested in, drop me a note.

Cal