CPU temps

[Tony Alfrey]

Leon Goldstein wrote:
> Steve Jardine wrote:
> 
>> <snip>
>>
>> Has anyone uses a Peltier cooler? I have seen them in portable coolers 
>> for cars, but they seem terribly inefficient.
>>
> As a CPU cooler, Peltier heatsinks are too efficient; condensation can 
> form at the CPU-heatsink contact.  Since the Peltier is basically 
> refrigerating the CPU, the heatsink fins are also much hotter than with 
> a normal cooler, and this extra heat must be dissipated and removed from 
> the case.
> 
> 
> 
  and to elaborate, thermoelectric (Peltier) coolers require some work 
to analyze

http://www.spin1.com/TechNote1.pdf

but perhaps a good way to think about the problem is in terms of power, 
not temperature.  One dissipates power by moving air at low temperature 
across a surface at high temperature.  By sticking a thermoelectric 
cooler in the path between CPU and heat sink, additional heat power is 
deposited into the heat sink (the thermoelectric cooler requires power 
to operate).  Therefore, a thermoelectric cooler can yield a lower CPU 
temperature, but at the expense of *considerably* larger cooling fins. 
A thermoelectric cooler is useful when there is some specific need to 
pull the temperature *below* room temperature (not possible with a 
passive system except for David's swamp cooler) and when the heat flow 
(power) need not be very high, as in refrigerating a beer.  So that's 
why Vu's article on big fancy aftermarket CPU coolers shows none with TE 
coolers.  They all passively move heat from a thing with small surface 
area (CPU) to a thing with big surface area (the fins) in contact with 
the ambient air.


-- 
Tony Alfrey
tonyalfrey at earthlink.net
"I'd Rather Be Sailing"