Thermal pad or Thermal paste?

Tom's an idiot. Any more questions?

-
larrymoencurly stood up at show-n-tell, in
, and said:

w_tom wrote in message
...

Author provides numbers with his results for heatsink bare
verses heatsink with thermal compounds:
http://www.dansdata.com/goop.htm

His copper heatsink also had concentric ridges which were not
removed. And yet thermal compound resulted in only single
digit temperature improvements.

If thermal compound doesn't matter much, why do all CPU manufacturers
include something like it with their retail boxed CPUs and heatsinks?

The retail boxed AMD XP1800+ I bought last month cmae with a layer of
phase change material on its heatsink, and my old 300 MHz Slot 1 Intel
Celeron heatsink, which was installed at the factory, had grey grease
on it. Both heatsinks were fairly flat, although I didn't measure
this.

--
Strontium

"It's no surprise, to me. I am my own worst enemy. `Cause every
now, and then, I kick the livin' **** `outta me." - Lit

Can you guarantee that one will mount every CPU properly? I
too would recommend hobbyists use thermal compound. The stuff
is so cheap. So many computer assemblers make mistakes. Best
to just tell them to always use thermal compound. However CPU
manufacturers have provided some CPU heatsink assemblies
without thermal compound when customer assembly was not
required.

Too many computer assemblers completely avoid numbers such
as the essential "degree C per watt" parameter. Better to
tell them to install thermal compound to help compensate for a
poorly surfaced or not even machined heatsinks. To
compensate for heatsinks erroneously selected on hype, price,
or a useless parameter such as CFM.

Just because cheap thermal compound is used does not prove
that thermal compound is required.

Furthermore, a recent Intel paper demonstrated little
difference between thermal compound and bare 'heatsink to CPU'
for high power semiconductors. In that same paper, Intel also
demonstrated advantages of phase change material. But those
numbers are beyond the scope of where this discussion has
proceeded. This discussion demonstrates how to experiment -
to learn heatsink interface quality by running without and
then without thermal compound. This discussion debunks myths
about thermal compound being 'essential' by providing both
underlying theory and experimental examples. This discussion
demonstrates why specifications are so important in thermal
analysis and which specifications are important.

A "degree C per watt" specification is more important than
thermal compound. Specifications associated with that "degree
C per watt" parameter demonstrate pros and cons of thermal
pads, thermal compound, and heatsink quality. Nothing from
Intel or AMD 'instructions for assemblers' proves that thermal
compound is essential. If thermal compound was essential,
then one would have easily provided those numbers days ago.

larrymoencurly wrote:
If thermal compound doesn't matter much, why do all CPU manufacturers
include something like it with their retail boxed CPUs and heatsinks?

The retail boxed AMD XP1800+ I bought last month cmae with a layer of
phase change material on its heatsink, and my old 300 MHz Slot 1 Intel
Celeron heatsink, which was installed at the factory, had grey grease
on it. Both heatsinks were fairly flat, although I didn't measure
this.

Without specifications, that post is just another junk
science proclamation. But then too many use such junk science
logic which also makes Arctic Silver so profitable for its
manufacturer.

Strontium wrote:
Tom's an idiot. Any more questions?

w_tom wrote in message ...

If thermal compound doesn't matter much, why do all CPU
manufacturers include something like it with their retail
boxed CPUs and heatsinks?

my old 300 MHz Slot 1 Intel Celeron heatsink, which was installed
at the factory, had grey grease on it. [Heatsink was] fairly
flat, although I didn't measure this.

Can you guarantee that one will mount every CPU properly? I
too would recommend hobbyists use thermal compound. The stuff
is so cheap. So many computer assemblers make mistakes.

I don't think that it's even possible to mount a Pentium4 heatsink
improperly because of the widely-spaced supports, and with an Athlon
improper mounting either causes the center part to be cracked or the
CPU to burn up right away because of a big gap with the heatsink. But
what about the example I mentioned previously, where the Slot 1
Celeron came with a factory-installed heatsink with thermal compound
on it?