Heatsink compounds

Damien Elmes clug at repose.cx
Thu Jun 27 20:28:58 EST 2002

Rasjid Wilcox <rasjidw at openminddev.net> writes:

> On Thu, 27 Jun 2002 2:01 pm, Rodney Peters wrote:
>> I recognize that this discussion thread was nominally closed off, but as an
>> aspiring Athlon owner I'd like to explore a couple of points:
>> The heatsink compound which I use on Socket 7 type CPU is cheapo white
>> stuff, which, I believe, contains Zinc Oxide (non electrically conductive).
>>  I'd be wary of using conductive (silver) compound on Duron/Athlon, since
>> the latter CPU have accessible bridging points on the surface of the
>> package.  I believe that overclockers deliberately apply conductive pastes
>> to selected bridges to alter clock multipliers etc.
> If electrical conductivity was a problem for something touching the core, then 
> the heatsink itself (which is mostly metal) would surely be a problem.  Since 
> it is not, then I don't think the heatsink compound will be.  That being 
> said, I was careful to put the compound on the core only, and not the rest of 
> the CPU.
>> The purpose of heatsink compound is to eliminate a thin film of air between
>> the CPU & heatsink.  That was an issue with the large (not perfectly) flat
>> CPU packages & heatsinks, but the Athlon/Duron have a relatively small,
>> raised contact patch.
>> Maybe AMD have a point - that heatsink compound is not needed - what
>> practice do other Athlon/Duron owners follow ?
> Well, after my review of what I could find on the web, I went with the 'smear' 
> technique - putting a thin layer of heatsink compond on the core.  I cut up 
> the lid of a takeaway container to use as a spatula.  So far everything seems 
> just fine.

The key to remember is the "less is more" principle. Too much heatsink
compound is worse than not having any at all.

For the adventurous, you might even like to try vegemite. It has a better
efficiency than arctic silver, though doesn't last as long..


Damien Elmes

More information about the linux mailing list