Bad Peltier modules

[raptor1956]

I'm working on a project that will use Peltier module for both cooling and heating and have picked up some TEC1-12706 and 12709 modules for testing.  I constructed a test rig that consists of the following:

1.  An aluminum tube (rectangular) that is 4" x 2" with 1/8" walls -- this tube was cut to length at 5" tall. 

2.  An off-the-shelf CPU cooler with integrated fan

3.  4-DS18B20 temp sensors

4.  1/2" insulation surrounding the tube

5.  Teensy 3.6 control

6.  Bench power supplies to provide 12V to the fan and another to provide 4A to the module

7.  490g water

So, the setup has the peltier bonded to one side (outside) of the tube using thermal paste on both sides of the module.  The other side is then bonded to the CPU cooler with fan.  I have one temp sensor ty-wrapped to the fan guard to measure air temp.  I have another temp sensor bonded to the CPU cooler heatsink as close to the peltier as possible.  Another temp sensor is bonded to the aluminum tube on the other side of the peltier and outside of the tube.  The last temp sensor is inside the tube measuring the water temp.

OK, with that out of the way I setup a test by first filling the tube with water that was in a sealed container and at room temp for days.  I turned the CPU cooler fan on and let everything stabilize for a couple hours before starting the test.  At 2PM I turned on the peltier and confirmed it was at 4A -- the voltage was around 9V at the start but eventually ended up about 12.4V by the end of the test although the current was locked at 4A the whole time.

I watched things for a while and noticed that instead of the Al tube and water getting colder they both got warmer.  I continued the test for 5 hours by which time the water temp increased from 24.06C to 50.0C while the Al tube temp increased from 24.19C to 48.94C.  At this point I turned the power to the peltier off while keeping the fan power on.  The temps decreased as you'd expect.

OK, did I wire it backwards or install it backwards ... no I did not.  Just to be sure I repeated the test today but this time I did swap the power leads to the module and restarted the test -- here's what happened.

After 10 minutes the water temp had increased from 24.06C to 31.69C and the Al tube from 24.19C to 33.56C.  After 80 minutes the water was at 93.0C and the tube at 89.19C.  I anticipated the water getting to near boil at 90 minutes, but when I came back to record things I noticed there was water spilled out so the water had already begun to boil.  I shut the peltier power off at 90 minutes and am recording the cool down values.

So, this is not good -- the peltier module I'd perchased to cool is not cooling at all and instead it either heats slowly or real fast.  I'd calculated a heating rate of the first test at about 8.5W-9W whereas the second test, with the leads reversed, at about 37W.

My guess is that either there's something bridging the hot and cold sides and conducting heat to the cold side OR some of the thermocouples are wired backwards.  Either way the QC has to be pretty dreadful.

This was the first module to be tested and I'll need to confirm with another 12709 as well as the 12706 modules I have, but so far my confidence is not high. 


Brian

[Zero999]

It wouldn't surprise me if it's bad. Peltier coolers aren't the most reliable of refrigeration systems. I've seen them go bad before.

Silly question but are you sure it's rated for constant current operating? If so, are you certain, the current was within the specification? Could there be a problem with your measurements or power supply, which resulted in more current, than you wanted.

[raptor1956]

It wouldn't surprise me if it's bad. Peltier coolers aren't the most reliable of refrigeration systems. I've seen them go bad before.

Silly question but are you sure it's rated for constant current operating? If so, are you certain, the current was within the specification? Could there be a problem with your measurements or power supply, which resulted in more current, than you wanted.

I used two bench power supplies so the current readings should be very accurate -- they have been very accurate when checked in the past. 


Brian

[ajb]

What is your heatsink doing, temperature wise?  Do you have a complete datasheet for your TEC, with proper performance graphs?

Overdriving the TEC, either to the point where the COP collapses to uselessness or to the point where you overwhelm the heatsink, would also cause the cold side temperature to rise as you're seeing.

[raptor1956]

What is your heatsink doing, temperature wise?  Do you have a complete datasheet for your TEC, with proper performance graphs?

Overdriving the TEC, either to the point where the COP collapses to uselessness or to the point where you overwhelm the heatsink, would also cause the cold side temperature to rise as you're seeing.

Tomorrow, when this latest test is complete, I'll see if I can post a shot of the spreadsheet I've created that shows all four temps at 10 minute intervals. 

The air temp sensor barely moves but you can see it change a bit as the air conditioner kicks on and off.  The heat sink temp does increase by as much as 10C but also shows slight ups and downs as the air conditioner turns on/off. 

I'm in no way pushing the module beyond its purported design limits so I'm most definitely not over driving it.  The module being tested is the TEC1-12709 type which is rated at 9A but to get beyond 5A you have to up the voltage well above 12VDC.

I have some PCB work to do over the next couple days and when complete I will repeat the test with another 12709 as well as a 12706.  Again, I suspect that some of the 127 TC's in the current module were wired backwards.


Brian