Paul said:
If that Radio Shack stuff you are referring to, is the white zinc paste,
it is horrible stuff. In years past, I used it on heatsinks for an audio
amp, and the paste separates into a clear fluid and a white residue.
It left the heatsink high and dry, after only a short period. The
reason this happens, is there is no chemical change in the product
with time and heat, and the formulation is such that the chemicals
separate. This is the last product I would be searching for, to
prevent burning up an Athlon. You would be better off using toothpaste!
I hadn't observed this separation but then I haven't used it for quite a
while and, even when I did, I tended to mix it, as a sort of 'thinner'
(probably better to say lubricant), with another thermal compound.
The AS3 product apparently changes a bit over the first several days
of operation with a warm CPU. I think the idea is, it thickens when
heated, so that once the die has settled into position, the material
stays put. The only problem with this approach, is the stuff changes
enough that it approaches a "dry" consistency over a period of
months, and that is when it needs to be changed out.
I rather thought the 'settling' they speak of was the particles in it
seeking a more compact organization.
Kind of difficult to know if it's really 'dry' when, if properly applied,
it's not more than a few thou thick even 'at the bumps'. My 'fingertip wet
detector' just isn't that good. And I've not run across any need to change
it out over a period of months.
Having said that, I can't say, for sure, how long any particular CPU I have
has been in the 'one' place before being moved, or having it's heatsink
moved, for one reason or the other. Checking purchase records though, and
subtracting out a generous estimate for how long I experimented with
various heatsink ducting ideas, I'm pretty sure my XP1800 was in
undisturbed service for over a year before I decided to change heatsinks on
it (decided I wanted that heatsink on a new XP2400+ OC'd to 3400+). Of
course, 'making it a year' wouldn't be a very good design spec for someone
like Dell.
The purpose of any interface material, is to fill the air gaps, as
air is a relatively good insulator. That means, you want a film that
is thick enough to take the place of any air, and not any more than
that. When you use sil pads or phase change material, you have to
select the thickness based on the spec for surface flatness for the
two assemblies you are joining. The thermal designers I've worked
with use thicker materials than I would as a hobbyist.
Thermal designers have to make sure the specified pad works over the entire
range of tolerances and not just for 'the one you have'. And there may be
other reasons you are not aware of (e.g. die cushioning).
Btw, phase change materials flow with applied heat so the thickness changes
from the initially observed dimensions.
And manufacturing people will do whatever you want, if you coax
them nicely and push the right buttons. (Hint: Pretend to consult
with them
Like most people, they will only increase their own burden if you have good
reason for them to do so. Which, of course, is the art of negotiation:
finding a mutually beneficial compromise (which may mean enumerating
benefits the other side might not be initially aware of). But in the
scenario I presented, using a messy, problematic, thermal compound instead
of a clean and easy to apply thermal pad, both of which offer a valid
technical solution, doesn't offer them anything while using the pad is no
significant disadvantage to the designing engineer.
If a production manager accepted thermal compound under those conditions
then he isn't doing his job and neither would the project manager, who
would be accepting a host of other potential problems for no valid reason.
