Archives


- Beowulf
- Beowulf Announce
- Scyld-users
- Beowulf on Debian

Heat pipes? - copper

Many of your questions may have already been answered in earlier discussions or in the FAQ. The search results page will indicate current discussions as well as past list serves, articles, and papers.

Search

Bari Ari bari at onelabs.com
Mon Jun 4 21:06:30 PDT 2001


alvin at Mail.Linux-Consulting.com wrote:

> 
>> Copper is a better conductor of heat than aluminum and a bit pricier... 
>> why not just add more surface area to your aluminum heat sinks or 
>> increase the amount of forced convection across the surface for the same 
>> effect at lower cost?
> 
> 
> extruded aluminum is terrible heat conductor but very cheap to make
> and good enough for slower p3-cpu

Terrible? 100W/mK - 150W/mK is terrible?

> 
> an aluminum block that is cut is "lot better" heat conductor as its
> molecules are relatively intact ... according to those heatsink
> manufacturers .. just a few cents more to cut than to extrude ...
> but when making millions of heatsinks units...guess it adds up...
> 
Increasing the surface area of a heatsink or increasing the amount of 
forced convection across the surface area is how the thermal resistance 
of a heatsink is reduced.

>>>>> - get a liquid cooled heatsink - copper pipes with antifreeze
>>>>> ( like a radiator ) --  seems like a fun project
>>>>> 
>>>>> - heat pipes are expensive ?? ..as is peltier cooling
>>>>> 
>>>> 
>> The least expensive approach to cooling P4s and Athlons in 1Us is by 
>> just using aluminum extrusions for heat sinks that form the top of the 
>> enclosures and using forced convection across the the units. We're using 
>> some profiles now that are only 0.08 - 0.11 degC/W at 100 CFM. These 
>> would allow the use of a P4 in an ambient of up to 68 deg C.... but I 
>> wouldn't want to be in that room :-)
> 
> 
> isnt 68deg still on the high side ??? ....

The 68 deg C ambient is the air temperature of the environment that the 
CPU is in. Using a heat sink with a thermal resistance of 0.08 - 0.11 
degC/W allows us to keep the P4 case temp below the 75 deg C maximum 
spec. at an ambient air temp of up to 68 deg C. So there is only a 7 deg 
C rise above ambient.

> 
> some of ofour cpu temp measurements was around 30-35 range...hitting
> up to  40 or so when runnign heavy loads during the day...
> 	- the bigger the heatsink got... the hotter the cpu ran too
> 	cause the itty bitty fans was not good enough ...

A bigger heatsink doesn't always mean better. Surface area and the the 
profile is what counts.

What was your ambient air temp for these mesurements? 20 deg C? -10C? At 
an average room temp of 20 deg C the CPU case temp would only rise to 27 
deg C while running at the full 62W max. continuously with the heatsink 
I mentioned above.

Bari







More information about the Beowulf mailing list