M
Max Coppin
I need to re-attach the heatsink to my processor, should I use a thermal pad
or paste? What are the pros / cons?
or paste? What are the pros / cons?
S
snoopy
Max Coppin said:
paste period end of discussion. pads are only good for... well, never
mind... ;-) YMMV
S
Stacey
Max said:
Paste runs cooler, pads are easy for a dumbass to install i.e. stick one to
the bottom of the supplied HS and the end user can't forget to install it.
D
David Maynard
Stacey said:
It's also less messy, more rugged, reliably repeatable, and readily machine
applied in mass production.
Mention using thermal compound in most modern assembly facilities and
you're likely to be run out of the place on a rail.
D
dino
pick up some Arctic Silver and apply ...it works way better than the
silicone paste or pad
silicone paste or pad
S
snoopy
David Maynard said:
Right pads lend themselves nicely to mass production. (more rugged? i've
never been able to break that paste!)
But is it better for the purpose of heat transmission?
P
Paul
David Maynard said:
That is because engineers in mass production situations, want
what they hope will be a zero maintenance solution. Paste/grease
needs to be reapplied regularly, any time a rising CPU temp suggests
the paste/grease is no longer filling the gap between CPU and HSF.
Thermal "pumping" or drying degrade just about any paste/grease,
so at some point, the CPU/HSF interface has to be redone.
There are some "gooey" solutions (look like silicon rubber) that
can be injected into a heatsink assembly, using an injection port
and an observation port, but the performance of that kind of solution
is worse than a pad.
For an Athlon, the combination of a bare die (i.e. limited contact
area) and high power dissipation, really limit what will work to
keep the processor cool. While AMD doesn't approve of AS3, home
builders find it works just fine. (You may want to read up on
AMD warranty issues if you use AS3 or something similar. If
returning a processor, make sure the processor is cleaned first.
Don't be sloppy with the paste/grease.)
Paul
D
David Maynard
snoopy said:
To 'break' it is to do something that causes it to not serve the purpose
and thermal compound that gets smeared off during handling, as one example,
or wasn't applied properly in the first place, doesn't do its job. It's
function is 'broke'.
Depends on how you define 'better'. If you mean what is the best that can
be achieved under ideal conditions, and ignoring long term effects, then
thermal compound is probably 'better'. If you mean being able to count on
all of the applications, and the devices it's applied to, working properly
without costly failures/rework then a pad is better.
T
Triffid
Max said:
The choice depends on the application.
Pads are designed to provide an adequate thermal interface regardless of
installer competency. IOW, guaranteed & foolproof. They also last the
life of the processor installation, but can be difficult to disassemble
if required.
Paste has the potential to provide somewhat improved thermal transfer
performance if properly applied, which is largely irrelevant unless you
overclock or otherwise operate under stressful conditions. Paste
performance can degrade over time as it dries out, but disassembly to
replace it is not difficult.
Thermal paste should be applied such that it only fills voids which
would exist if no thermal interface were used - direct contact is
preferable. It's not easy to get this right in practice, hence the
manufacturer's preference for the predictability of pads.
E
Ed
I've seen as much as a 10C difference when applying paste to my AMD
Barton, man these things are touchy! Built a few (slower) with the AMD
boxed for friends and the temps are all about the same, these ppl don't
OC and could care less what their CPU temps are, as long as the PC
doesn't crash they are smiling.
Did a few AMD64s with paste and found out you don't want to cover the
whole top of those, just the center area.
Cheers,
Ed
R
Rob Stow
Ed said:
With Opterons, I only see about a 3'C difference between
pad and paste when using the stock heat sinks and fans
that come with the processors. For an Opty 240 I see
about 42'C with the paste (AS3) and 45'C with the pad.
For an Opty 246, add 6'C to both numbers.
I covered the whole top (very thinly) of the Opties I did.
I'll try it your way sometime and see if it makes a difference.
A
Andrew J
At 40C the pad turns to paste, so it is a non issue. 99% of all the
talk about paste being better is just hype. The best improvement I've
ever seen using paste over a pad is 2C. Hardly even worth the effort.
BTW a pad has never shorted out a CPU but paste does all the time.
That's why AMD doesn't want anyone using paste and if they find out
you did your 3 year warrany is void.
R
Robert Redelmeier
In comp.sys.ibm.pc.hardware.chips Rob Stow said:
My technique for applying thermal grease/paste is to squeeze
out a log of the stuff centered between the long sides of
the dieback. Then very carefully squeeze it down and out
by pressing the heatsink down parallel while rocking slightly.
Practice a few times and check for how well the grease spreads.
Done correctly, this method eliminates air bubbles which are
the big enemies of heat transfer. Both thermal pads and
"trowel the grease flat" can get bubbles.
-- Robert author `cpuburn` http://pages.sbcglobal.net/~redelm
S
Stacey
Andrew said:
Except it's thicker, adds another layer between the CPU and heat sink.
I've seen 6-8C
?? Normal silcone HSG isn't conductive so that would be pretty tough!
A
Andrew J
The most popular ones like AS(silver?) conduct electricity. The
smallest amount left behind voids your AMD warranty which many have
found out the hard way.
http://forums.extremeoverclocking.com/showthread.php?t=86301
D
David Maynard
Andrew said:
First let me say that I've used, and continue to use, arctic silver and it
can be fine if applied properly. However, noting that some of it may 'ooze'
out understates the problem. The stuff will adhere to anything like
gangbusters, with particular affinity for wherever you don't want it, and
one's 'normal' instincts on how to 'wipe it off' generally spreads it all
over the place rather than 'removing' it. Plus, get it on your fingers, an
incredibly easy event, and everything you touch will end up contaminated
with it as well.
It is the epitome of 'messy'.
L
Leythos
Intel includes a syringe of heat-sink paste in their retail kits for
Xeon processors. In general, the least amount used is the best. I
generally put HSP on using a razor knife, and then almost scrape it all
off, leaving only a trace behind. To much paste is always a problem,
most people apply way to much. If the heat sink is machined properly it
a small dab will be more than enough.
I know what you mean about getting it on things - when I started
building circuits a couple decades ago we used the same type of paste
and had the same problems then too - nasty silver or white crap all over
the tips/pads of your fingers
D
David Maynard
Leythos said:
Yes. As I said, if done properly.
And sometimes what seems like a 'good idea' turns out to be a bad idea. For
example, you spread your arctic silver, see a little left over 'dab', and
decide to be 'meticulous' and clean that off, just to be 'safe'. You
diligently take a 'lint free cloth' and delicately 'wipe it up' to discover
you've just taken a small 'dab' and smeared it over the entire surface you
'wiped'.
I know what you mean about getting it on things - when I started
building circuits a couple decades ago we used the same type of paste
and had the same problems then too - nasty silver or white crap all over
the tips/pads of your fingers
Hehe. Yeah. With the 'white crap', though, it's 'just' messy whereas the AS
'mess' can cause electrical problems regardless of their 'non conductive'
claims to the contrary.
S
Stacey
Andrew said:
Actually the "most popular" ones aren't conductive. Now if you said "The
most advertised" or "The most expensive/popular with overclockers who read
websites like they are gospel" maybe I'd go with that.
I've tried AS (someone bought some and brought it to me to use on their
system) vs radio shack HSG and there was no difference in temps to amount
to anything. I can't see why anyone would use electrically conductive HSG
anyway. You're right though if I had to choose between electrically
conductive HSG and a pad, I'd be using a pad! The problem is there are good
HSG's that aren't electrically conductive and do a much better job than a
pad does.
K
KR Williams
Actually the "most popular" ones aren't conductive. Now if you said "The
most advertised" or "The most expensive/popular with overclockers who read
websites like they are gospel" maybe I'd go with that.
I've tried AS (someone bought some and brought it to me to use on their
system) vs radio shack HSG and there was no difference in temps to amount
to anything. I can't see why anyone would use electrically conductive HSG
anyway. You're right though if I had to choose between electrically
conductive HSG and a pad, I'd be using a pad! The problem is there are good
HSG's that aren't electrically conductive and do a much better job than a
pad does.
If you google back a few years (perhaps even five) there were
people selling the idea of *grounding* the HSF to improve the
processor speed. It's amazing what people will buy! If you have
any doubts about people's gullibility, look no further than
audio-phools, or your fav audio retailer. People are ignorant,
and have lots of money with which to be stupid.