Thermoelectric Modules

[alan2here]

The thermoelectric modules about 1/3rd down this page.

http://www.unitednuclear.com/new.htm

 

Roughly how cold does the chip get given how much power.

 

For example a button cell, an 'aaa' cell, a powerful lithum battery about the size of a 'd' cell?

 

Also could a resonable amount of power be created by placing one of thease in the temperature gradient between two materals of naturally different temperatures? I saw somone online doing something with a hollow materal full of a gas at a pressure to get the optimum of the materal being cold to the touch.

Edited May 2, 2009 by alan2here

[swansont]

Usually they'll have power specs for a delta-T. You can stack them to get a larger temperature difference.

 

Yes, you can generate electricity with a temperature difference — it's the principle behind radioisotope thermal generators.

[alan2here]

At what rate of cooling for what power in turms of what type of battery or power source I might know about does delta-T mean. What is a delta-T. I tried to Google it with no luck.

 

If the temperature difference between two different surfaces that work only by absorbing ambient temperature from the air creates a typical consumer electronics battery current and 0.01v then 800 might be needed to power something, they appear very small however cost may be signifegent at that quantity making them sutable for solar style arrays but not for consumer electronics. However a constant 2V each means that only 4 are needed for the same power requirement.

Edited May 3, 2009 by alan2here

[alan2here]

bump

[Ndi]

What you have there is called a TEC.

 

It is based (sorta) on the Peltier effect, and it moves heat at a certain rate. They are quite inefficient, and have several drawbacks, such as weld point failing with temperature and failure if the hot side is not maintained cool.

 

These are used in portable cooling systems because they are compact and solid state, but are woefully inefficient versus the traditional heat pumps. A compressor heat pump can beat 100% "effective" efficiency (measured as heat brought in versus consumption) under favorable conditions, whereas TECs wallow in mediocrity.

 

Most of the juicy stuff on these you'll find on computer sites, not physics sites, because with computers you have little heat to move (100-200W), and plenty power (maybe a few hundred watts to spare) but the cold side is critical, as is MTBF. Solid state is way, way better. Just look for "TEC" cooling and you get specs, ratings, consumption, everything.

 

Just try to avoid cases where the hot side is not cool enough. Even the web sites smell of smoke.

 

P.S. Oh, you wondered about how cold/hot you get from a battery. Quick rules to help you:

 

a) Any decent TEC will eat 12, 18 or 24 volts. It works by potential difference. Batteries suck.

 

b) Rule of thumb is - To move some 100W of heat at a constant rate (how hot the sides get is another story), you need 50W. As a result, the hot side needs to lose 150W.

 

c) Under hold-in-hand conditions, a 50W TEC has a cold side of about -20C, hot side too hot to touch, needs fan. If you put your hand on the cold side, it drops to -7 or so. Temperature depends on the heat load, and when closing the rated heat transfer rate the power the TEC eats starts to add up, the cold side becomes hot and it self-destructs quite rapidly. If shorted, that's good for the batteries.

 

d) A D battery is 1.5V, maybe 7Ah. That's 11 Watts, with a 50 required of a small -puny-TEC. That would hive you 13 minutes. NOTE that a TEC needs 12V, so it will probably bust, you need a DC-to-DC SMPSU, that takes its toll in efficiency.

 

e) A high-perf Li-Po has 3.7V nominal, 4Ah, so you'll need 3 of them in-series to get voltage. That gives you 12V, 4A, a total of 48Wh, or 48 minutes of coldness. However, note Li-Po can't be safely charged in-series, so you need a multi-branch charger.

 

f) On failure, the TEC shorts sometimes. Li-Po cascades on heating, you need trip sensors.

 

g) Car batteries with fuses work. A SLA has 70-100 AH at 12V, almost a KWh. That could work (and does in portable car fridges and ice boxes)

 

I wouldn't. Whatever it is you want.

[swansont]

At what rate of cooling for what power in turms of what type of battery or power source I might know about does delta-T mean. What is a delta-T. I tried to Google it with no luck.

 

[math]\Delta T[/math] = temperature difference. Units are often specified with given maximum temperature difference they can achieve

[alan2here]

Thanks.

 

There is lots of data like that on pages like this but little information.

http://www.ferrotec.co.uk/products/thermal/modules/singleStage/

 

So if it's got 10 VMax then I will need to generate 10 volts to get -0.5* the \Delta T on one side and +0.5* the \Delta T on the other relative to the tempreture of the enviroment?

Edited June 11, 2009 by alan2here

[Ndi]

In that page I see

 

(Model Number)

I/V - Maximum voltage and the maximum intensity (amps) when fully loaded (W=I*V here)

ΔT- Maximum temperature difference _in ideal conditions_

Qc Max - Maximum heat transfer at peak efficiency (when cold=hot)

 

The way I read it, they all get 72C cold-to-hot (IDEAL, please take note. You never get this because the hot side is rarely efficiently cooled and the cold side WILL go up the second you try to move heat, PLUS the hot side radiates in real life, seal is not perfect and the whole shamoogie is metal).

 

The more power it eats, the more heat it can transfer. You need to give it less heat that the rated transfer _at the rated efficiency_, meaning that a 100W pump will move 100W when the cold equals hot (max efficiency), and to move 100 you need 150. As the delta increases, efficiency drops. At 72, it's zero (thus the maximum temp).

 

That's why most PC shops give it a rated moving power, meaning a 200W 30C TEC will keep 30C below ambient (22C does -8C) if you feed it less that 200W in heat (quad core at some 4.8 GHz?). Not very scientific, but mighty clear. It will also suck some 300W from the PSU, forging an upgrade.

 

Also, if you look for a temperature in degrees, you need to know the exact quantity of heat you give it. If you use the cold side to chill an airtight, sealed and insulated fridge, then given enough time and proper ventilation you get almost 72 degrees below hot side (NOT ambient unless you are in a hurricane).

 

In the heat outside, there will be 30 C, and the hot side will probably stay hot to the touch but not too much so if vented properly. If you can touch it without recoiling in pain it's under 60C (golden rule of an IT specialist ). That would give you -15 theoretical, but most likely below zero, or even a plus 3-4. Thus the fridge applications. (the closer it gets to -15 the worse it performs and each drop of heat that oozes through the insulation takes longer and longer to pump so it stabilizes).

 

If not insulated, well, then it moves the ambient heat, meaning the ventilated part will lose to the air as the cold will gain. If you don't fan the 50W-ish, you cook it. The more you fan, the more it will go.

 

Oh, and then there's stacking.

[alan2here]

lol, and I was thinking of this as a simple thing for my simple lack of electronics knowlege. Could make things like a self cooling thermos flask with a rechargeable button cell, a thermal unit and two pieces of wire. Power something with the warming effect of your body heat or alternativly cool you down with a few of the thinner ones in some clothing. I was imagining that if two were connected together with the wires that linked them crossed over then cooling one would cool the other and vice versa creating a sort of heat see-saw to synchronise the temperature between two places.