# Corner of Peltier module is very hot

I've bought a generic TEC1-12706 Peltier module and I'm having trouble making it work.

I wired it to a 12 volts DC, 8-amp power supply, and connected the hot side to a fan & heatsink with thermal grease. Turning it on, I noticed that condensation forms on about 70% of the cold side, but the rest is dry. Most of the cold side feels really cold, but one spot in the dry corner is too hot to touch! An infrared thermometer says the coldest corner (near the negative input wire) is about 2.5°C (36°F), and the hot corner diagonally opposite it (but still on the cold side) averages 46°C (115°F).

Having zero familiarity with these devices, I don't know if such a thing is normal, if I'm doing something wrong, or if the module is faulty. Is there anything I can try before I buy another module and wait weeks for it to come from China?

Picture (sorry for potato quality):

• I don't think that's normal. ;) I'd ask for another one to be sent or find another supplier... or both. While your words sound well reasoned and I'm pretty sure you are thinking correctly, a picture from you and some arrows where you took the measurements would probably clinch it, or not. – jonk Nov 14 '17 at 22:51
• 50W is like a low end overclocked CPU. Do heat sinks extend well beyond part on both sides? $\Delta T=48.5, T_{avg}=24.5$ ...n.b. never use 'F , its not linear. – Tony Stewart EE75 Nov 14 '17 at 22:57
• @Boann Yeah. That thing isn't working. It's toast. Or at the very least, the heat sink on that area isn't thermally coupled to the device. – jonk Nov 14 '17 at 23:52
• According to the datasheet, I think one can measure the hot side temperature by calibrating at room temp with a pulse the current used and compute equivalent resistor, compare with datasheet for 25'C,50'C to see % rise , then like like how any conductor R rises with temp compute V/I to to get Temp rise.. (for future use) I suggest a 50mV current shunt R on the ground side. ( for easy of measurement) – Tony Stewart EE75 Nov 15 '17 at 0:36
• There are explicit torque specs for clamping the heatsink. I dont see any clamps. An adhesive 3m thermal tape is not the way to transfer heat with no clamp. Any surface coplanarity issues or microscopic warp or low or uneven clamp pressure will cause any CPU and Peltier device to have hot spots. 99% silver epoxy under compression is better if no room for clamps. These devices are delicate. read the datasheet ! – Tony Stewart EE75 Nov 15 '17 at 9:56

I suspect bad thermal coupling to the heat-sink.

Having worked on a few of these the tolerances and construction looks cheap. If the thermal coupling is not good they will get hot very quickly. Your numbers would also back that up.

I'd suspect one or the other, or both, is not flat. Also any small contaminant under the thing and it will not sit flat. They are also notoriously hard to hold down without cracking the ceramic.

Also... you mentioned condensation. DO NOT allow any moisture to enter the array. These things need to be heretically sealed to have any kind of life. They are also really easy to kill by over-driving them or through rapid temperature cycling.

There is a decent primer to read here.

• Okay, so, I decided to completely clean the thermal grease off of the module and heatsink, wipe them down with isopropanol, and start again. I made sure the grease covered everything smoothly. Same problem. Same hot spot, in the same sort of area. The device has silicon all around the edge so I assume moisture can not get in. So I still don't know the problem. I don't have anything holding it down though, apart from the sticky thermal goop. – Boann Nov 15 '17 at 11:29
• @Boann, things may not be as flat as you think, perhaps a promontory on the device or the heatsink. Try rotating it and see if the hot spot moves. – Trevor_G Nov 15 '17 at 14:10
• @Boann, interesting. Try turning it over. Does the previous hot spot now get cold? – George Herold Nov 15 '17 at 15:26
• @Trevor Interesting idea. I rotated the module 90° relative to the heat sink and the hot spot remained in the same position relative to the module itself. – Boann Nov 15 '17 at 18:04
• @GeorgeHerold I briefly tested it turned over, and the hot side seems to get hot evenly, near as I can tell. – Boann Nov 15 '17 at 18:07

I suspect you exceeded some critical parameter. Which one is hard to say.

http://www.thermonamic.com/teach_view.asp?id=861 Read before using next time.