Peltier Cooler Trouble

This is a comprehensive guide to Peltier cooling ....

http://www.heatsink-guide.com/peltier.htm

I see from that link that peltier coolers actually operate with a bigger temperature difference with a slightly higher heat transfer rate when they operate at a higher temp (60 vs 20C). One thing I changed actually is adding better cooling on the hot side. It also shows that at the optimum voltage for cooling a cpu was about 11 volts to get the lowest temps for whatever reason. Of course the deltaT in that case was significantly different than what we want to do (less than 0C temps) but it shows that a higher voltage doesn't necessarily give you a bigger deltaT.

Thanks for the link. I wonder if my interpretation of that web page is relevant to what is going on with ours.

How does it behave when you reverse the polarity?

Did you use any heatsink grease when assembling it?

I will have to try reversing the polarity. I had initially installed it backwards with the correct polarity (hot/cold side isn't necessarily obvious) and the results were clear.

The original peltier had heatsink grease but the new one does not. Doesn't seem to make much difference as the problem remains with the new one. Surely, it may help but the area is quite large in comparison to for example old CPUs with a die area of ~1cm^2 so it shouldn't make a massive difference.

You do have to make sure that the voltage AND current limitations are not exceeded. The other issue is you cannot reliably parallel two power supplies UNLESS they are designed to be paralleled. PWM (Pulse Width Modulation) is not a bad way of controlling a Peltier cooler. Invest in a commercial controller.

Yeah, I realize it was a lame way of getting a power supply. One could end up doing more work than the other, but the power supplies are identical and they warm up to the same temp which makes me think they are doing about the same work each. It draws the same current whether either one or both are plugged in and delivers the same voltage (16.5v +- 0.1v).

It cooled sufficiently before, I just can't figure out why it's not now. The old one and new barely get cold at all and draw somewhat less current than before. I wonder if it has to anything to do with what I posted above about the link which was given.

Unfortunately I dropped the old peltier and it cracked the ceramic plate so I can't test that one anymore...

Lower than expected current draw you say.

It is well within the realm of possibilities that you scored a peltier device that was DOA.

I agree on what you said regarding the parallel power supplies. Should work no problem.

Are you absolutely sure that the power supplies are working? Can they make a lamp glow?

Tried using a car battery across your peltier yet?

Yeah, I wonder if it just happens to be another one which is damaged lol. I think the power supplies must be working if they supply a voltage and current. I don't think I have anything which can accept a higher current draw than the peltier device that I have.

I did hook it up to my car. Luckily, even though my car is old, it has a new 88 Ah battery so I don't need to start the car very often :) It's possible it got a little bit colder but it was also probably a little bit cooler outside.

They are easily mechanically damaged. If yo exceed the delta T they are toast. They are also very inneficient. Another issue that crops up is condensation. They aren't too happy about it because it may cause corrosion.

Here is a note on the parallel operation of Switchmode Power Supplies. http://www.weidmuller.com/system/files/webfm/downloads/pdfs/literature/PS06_Parallel_Operation_of_Switchmode_Power_Supplies.pdf

That is interesting, but it seems as though they are both operating correctly. I would assume that if they both warm up the same (i.e. one is not much hotter than the other) they must be doing the same work. I will try and check to see what variance there is between the two though. It's possible the old one was damaged by condensation even though it was surrounded with foam.

They are usually made of parallel circuits. It is possible that thermal cycling or excessive mechanical force has broken the contacts. Somebody mentioned corrosion as a valid point that could cut some traces.

Loosing some sections would reduce its cooling capacity and current drawn.

Short-circuiting would also reduce its cooling capacity but increase current drawn.

Good luck.

Peltier devices both pump heat and generate heat. The heat generated (I2R) is proportional to the square of the current whereas the heat pumped is proportional to the current. For this reason your power supply should be filtered to produce steady dc. For example, a 50% duty cycle 0A/2A would pump the same amount of heat as a steady 1A current, but would generate twice as much heat (1/2 x 4R = 2R vice 1R watts for steady 1A current. If your supply is not steady dc, it may complicate measuring current and voltage, especially if you are using digital meters. Borrow a scope to see what's going on.

Note that the Peltier is a thermodynamically reversible device in that a temperature differential will produce a back emf, so your voltage is the sum of the IxR and this back emf. The current should start out higher and reduce to a lower value as this temperature differential is established.

You're right, I didn't think of the duty cycle aspect of power and heat pumped. But I think since they are laptop power supplies, they must be delivering reasonably clean power. Clean enough that even if there was some noise from the PWM varying between 15 and 17 volts averaging 16 it wouldn't hurt such a crude device.

The back EMF is readily apparent as it starts at around 5.5-6 and goes down to 4.5 after maybe 10 seconds or so. So it seems to be operating, just not well.

Maybe not. In the laptop, it has the battery to smooth things out.

Try it again with some heat sink grease then.

It has sufficient cooling. The hot side heat sink is only about 5°C warmer than ambient

I assume you meant insufficient cooling. The heat sink should be hotter.

My experience with them says that they don't like ripple in the power supply. Try a regulated power supply to see if it works that way. If not, you probably have some leakage in the cell.

It has a quite large heat sink which was designed to cool just passively with no fans and with two 12v fans operating at 16 volts it stays pretty darn cool. I think it would get very hot without the fans. It should easily be able to dissipate 100+ watts absolutely no problem.

I'll have to check on the ripple somehow. Unfortunately I'm still in Europe where all my things in Ohio are quite far away :p I guess it wouldn't be too surprising if something happened to these power supplies because the old peltier would turn the cold heat sink (4 inches square by 1 inch or so) into a block of ice in a few hours.

I would carefully examine how the device is mounted. My last job used these to cool reagent compartments and the devices were somewhat loosely mounted to allow for heat expansion. Its possible that if the device is too rigidly mounted, expansion may be breaking the internal contacts on some of the individual elements, which would cause the sort of symptom you describe, degraded performance rather than total failure.

Yeah, it has a plastic holder which keeps it in place between the hot side heat sink and an aluminum block. There are two long screws which keep it together which I was pretty careful to make sure were evenly tight.

But, I'm not sure why it's working now. I tried a new power supply and even though the voltage was more like 15.5 instead of 16.5 volts, the current was much higher and in a matter of minutes the cold side heat sink was covered in condensation. The hot side is quite a bit warmer now (36 instead of 30C) which makes sense given the extra energy transferred/consumed.

Then I tried the old power supplies and it ALSO had a higher current and at least in an unscientific observation, seemed to get about as cold. My guess is there was an intermittent problem with the old power supplies. They had some ripple because a few days ago a volt meter set on AC would read 1 volt where of course it should read zero lol. I guess it was an unscientific way to see how clean DC is if the multimeter is capable of reading a wide range of frequencies (this one I know from the past could read 1Hz-20kHz no problem with my audio experience). Now both power supplies read zero as they should :p Also, once it was obvious it was cooling better we put some thermal paste on it as well. That certainly helped as well.

So, I guess I learned that thermoelectrics don't function well without clean DC. I just assumed as long as the voltage didn't go negative and the average voltage/current was within specs it would function. But, I was very wrong!

Thank you everyone for posting! I learned a lot of interesting things. That's why I like this forum so much.

-Nick

Very good point!