Thermal conductivity and Peltier Gen

Thread Starter

pityocamptes

Joined Jul 25, 2012
82
Would it be better to have a high thermal conductive material between the "hot" of the peltier gen and the heat source, or between the peltier "cold" side and the heatsink? Which in theory would yield the highest output and efficiency? Thanks!
 

MrChips

Joined Oct 2, 2009
30,823
Both. The efficiency is dependent on the temperature difference between the hot and cold sides and how efficiently you can move the heat from the hot to the cold side.
 

Thread Starter

pityocamptes

Joined Jul 25, 2012
82
Both. The efficiency is dependent on the temperature difference between the hot and cold sides and how efficiently you can move the heat from the hot to the cold side.

So I was thinking a copper shim (pc shim) sandwiched between the hot side and an aluminum base. The cold side would just attach to an aluminum heatsink with fins. Should I sandwich the peltier gen between TWO copper shims?
 

luvv

Joined May 26, 2011
191
I'm not sure what your goals are but instead of trying to upgrade thermal conductivity,you should look into adding thermal mass.

Beastly chunks of heatsink will act as buffers against any sudden changes in temp,and that will reduce changing current needs.

I've got a T.E.C unit from a refrigerator,I'll up some pics later so you can get a idea of the heatsink size ratios for a constant on application.
 

MrChips

Joined Oct 2, 2009
30,823
Use the analogy of electrical resistance. Anything you put in the path of current flow will add resistance. If you use one layer of copper shims, you now have two layers of paste plus one layer of copper. One layer of paste is still the lowest thermal resistance.
 

Thread Starter

pityocamptes

Joined Jul 25, 2012
82
I'm not sure what your goals are but instead of trying to upgrade thermal conductivity,you should look into adding thermal mass.

Beastly chunks of heatsink will act as buffers against any sudden changes in temp,and that will reduce changing current needs.

I've got a T.E.C unit from a refrigerator,I'll up some pics later so you can get a idea of the heatsink size ratios for a constant on application.
Heatsink is pretty big.
 

MrChips

Joined Oct 2, 2009
30,823
Depends on what you are doing. Adding thermal mass doesn't quite cut it. You need to increase heat flow. I built a T.E. cooler using cold water to remove the heat.
 

Thread Starter

pityocamptes

Joined Jul 25, 2012
82
Use the analogy of electrical resistance. Anything you put in the path of current flow will add resistance. If you use one layer of copper shims, you now have two layers of paste plus one layer of copper. One layer of paste is still the lowest thermal resistance.

OK, thanks! ... however would it be considered "resistance" when in fact copper has a higher thermal conductivity than aluminum? I know they are dissimilar metals with alum being around 200 and copper in the 400 range. After thinking about it, you are probably correct.
 

Thread Starter

pityocamptes

Joined Jul 25, 2012
82
Depends on what you are doing. Adding thermal mass doesn't quite cut it. You need to increase heat flow. I built a T.E. cooler using cold water to remove the heat.

A cooler may be out of the specs for my project. Heat diff will be around 600 F (hot) to ambient air temp (cool) with heatsink.
 

THE_RB

Joined Feb 11, 2008
5,438
The hot side moves much more energy than the cold side, maybe a few times more. So that's where you need the best thermal conduction.

600'F ABOVE ambient temp sounds very high! Have you checked the max ratings of the peltier device?
 

Thread Starter

pityocamptes

Joined Jul 25, 2012
82
The hot side moves much more energy than the cold side, maybe a few times more. So that's where you need the best thermal conduction.

600'F ABOVE ambient temp sounds very high! Have you checked the max ratings of the peltier device?
Yes, max op temp is 620F (they are high temp units) for the hot side. So with an aluminum base, I wouldn't get more heat transfer by sanwiching a copper CPU shim between the alum base and unit since copper conducts higher temps than alum? I understand that adding too much material could cause a "resistance" but I was thinking that the shims, which are used on CPU's for heat transfer between the chip and the heatsink would be plausible... thanks.

I got this quote from a computer forum:

Copper has a higher thermal conductivity, and therefore is superior to aluminum in processor cooling.
Copper is better at conducting heat than aluminum, but aluminum is able to radiate the heat into the air better than copper because of its lower density.
 

MrChips

Joined Oct 2, 2009
30,823
You are dealing with two separate issues.

The first problem is getting the heat to the heat sink. Any additional interface between the heat source and the sink is going to reduce heat transfer. Hence shims are a bad idea.

The second problem is getting the heat away from the heat sink. This is where thermal mass and conduction/convection currents play a crucial role. Increased water or air circulation is the most effective.
 

Thread Starter

pityocamptes

Joined Jul 25, 2012
82
You are dealing with two separate issues.

The first problem is getting the heat to the heat sink. Any additional interface between the heat source and the sink is going to reduce heat transfer. Hence shims are a bad idea.

The second problem is getting the heat away from the heat sink. This is where thermal mass and conduction/convection currents play a crucial role. Increased water or air circulation is the most effective.

Thanks. So, the shim is out. - but this is what I was planning to do, just to make sure we are on the same page:

1/4" aluminum base
silicon grease
1mm copper shim
silicon grease
hot side of peltier unit
silicon grease
20mm aluminum finned heat sink

As far as the aluminum finned heat sink, I could possibly add a 3v fan to the circuit. What would be the best way to mount a computer fan to a finned heat sink? Also, would it be better to pull the air off the heatsink via the fan, or push air from the fan against the heat sink? Thanks again!
 
Top