Best heat sink compound?

[Ed Anderson]

DaveC <me@privacy.net> wrote in news:0001HW.BBF0376D003D7AA4F0080600
@news.individual.net:

White "tooth" paste kind, or clear greasy "hair stuff" kind?

I use thermally conductive glue or arctic silver compound.

 

[Allen Windhorn]

DaveC <me@privacy.net> writes:

White "tooth" paste kind, or clear greasy "hair stuff" kind?

If you need to have electrical conductivity to the heatsink, use the
clear stuff (DAMHIKT).

Regards,
Allen
--
Allen Windhorn (507) 345-2782 FAX (507) 345-2805
Kato Engineering (Though I do not speak for Kato)
P.O. Box 8447, N. Mankato, MN 56002
Allen.Windhorn@LSUSA.com

 

[w_tom]

Better heatsinks are not perfectly flat. They are so minor
tapered so that maximum pressure appears at that little spot
where most heat transverses 'semiconductor to heatsink'
junction.

Because the spot is so small, then apply thermal compound
only at center of a CPU. Almost no heat is transferred at the
outer half of a heatsink surface. Therefore the center is
also where greater pressure is applied.

As for thermal compound, demonstrated that even toothpaste
was a better thermal conductor than Arctic Silver III - in
short term test:
http://www.dansdata.com/goop.htm

As Floyd Davidson so properly posted, once numbers are
applied to all that hype, then perspective says thermal
compound is more determined by dollars than effectiveness.
Some thermal compounds offer so little improvement in thermal
conductivity that they even forget to provide the thermal
conductivity numbers in "°C / watt / inch squared" or "°C inch
/ Watt". If Arctic Silver was that superior, then AS would be
preaching those numbers everywhere up front. Where are the
numbers? Why do they forget to put that so essential number
on dispenser?

John Larkin wrote:

On Mon, 1 Dec 2003 0:10:21 -0800, DaveC <me@privacy.net> wrote:
White "tooth" paste kind, or clear greasy "hair stuff" kind?

The white thermal compound stuff is better. For serious heat sinking,

both surfaces must be flat; *very* flat. Most extruded heat
sinks aren't unless they are machined after extrusion.

don't use an insulator. If you must, use 0.5 mil
hard anodize on the heat sink.

*don't* use a silicone sil-pad or phase-change stuff.
They are both awful thermally.

Apply high, uniform mounting pressure. Belleville washers
and clamps are good.

Make sure the heatsink baseplate is thick enough to spread
the heat laterally; otherwise there will be a hot spot under
the device and all this effort is mostly wasted. Use a copper
heat spreader if needed.

If this is for a CPU, forget the whole thing.

John

 

[Tim Auton]

w_tom <w_tom1@hotmail.com> wrote:

Better heatsinks are not perfectly flat. They are so minor
tapered so that maximum pressure appears at that little spot
where most heat transverses 'semiconductor to heatsink'
junction.

Because the spot is so small, then apply thermal compound
only at center of a CPU. Almost no heat is transferred at the
outer half of a heatsink surface. Therefore the center is
also where greater pressure is applied.

I can see the logic, but you're talking about some damn fine
manufacturing tolerances there. Anybody actually sell heatsinks like
that?

As for thermal compound, demonstrated that even toothpaste
was a better thermal conductor than Arctic Silver III - in
short term test:
http://www.dansdata.com/goop.htm

"in short term test" being the operative term. Once it's dry it's
useless. I'm sure spit would give adequate results short-term but
that's hardly relevant to the real-world.

As Floyd Davidson so properly posted, once numbers are
applied to all that hype, then perspective says thermal
compound is more determined by dollars than effectiveness.
Some thermal compounds offer so little improvement in thermal
conductivity that they even forget to provide the thermal
conductivity numbers in "°C / watt / inch squared" or "°C inch
/ Watt". If Arctic Silver was that superior, then AS would be
preaching those numbers everywhere up front. Where are the
numbers? Why do they forget to put that so essential number
on dispenser?

I just went to the AS website, they certainly don't make a secret of
the figures you want, they're right there after the usual marketing
blurb:

Thermal Conductance:

350,000W/m2 °C (0.001 inch layer)

Thermal Resistance:
<0.0045°C-in2/Watt (0.001 inch layer)

http://www.arcticsilver.com/as5.htm

Anyway, for the market AS are going for (low volume home computer
users) an extra Ł2 or Ł3 for a tube of compound that will last you
through your next 10 or 20 CPUs isn't much of an expense. I personally
have had better results using Arctic Silver 3 than using generic white
paste, not wildly better, but detectable.


Tim
--
And the beast shall be made legion. Its numbers shall be increased a
thousand thousand fold. The din of a million keyboards like unto a great
storm shall cover the earth, and the followers of Mammon shall tremble.
- The Book of Mozilla, 3:31

 

[Michael A. Terrell]

Michael Black wrote:

"Michael A. Terrell" (mike.terrell@earthlink.net) writes:
Guy Macon wrote:

DaveC <me@privacy.net> says...

White "tooth" paste kind, or clear greasy "hair stuff" kind?

Neither.

http://www.ksbrainstorms.com/index.php?pagename=Arctic_Silver_5_Review
http://www.xbox-connection.com/hostedsites/hotlinesrc//reviews/arcticsilver5.php
http://www.overklokking.no/annet/arctic_silver_5/eindex.html
http://overclockersclub.com/reviews/articsilvercermaiquereview.php
http://www.modsynergy.com/Review%20109.htm

--
Guy Macon, Electronics Engineer & Project Manager for hire.
Remember Doc Brown from the _Back to the Future_ movies? Do you
have an "impossible" engineering project that only someone like
Doc Brown can solve? My resume is at http://www.guymacon.com/

Make sure you use all the other over hyped crap so it doesn't get
lonely.

Considering the question, and the large cross-posting, I kind of
expected a computer newsgroup to be included. And the threads over
there about heatskink compound can be somewhat amusing. They range
from "Do I really need heatsink compound" to debate over the brand name
of the compound.

It strikes me that it's right up there with the audio consumers
and gold plated cables and ten dollar capacitors.

Michael

I wonder if they buy the stuff from Mr. Eustace Haney?




































The con man on Green acres. ;-)
--
24 days!


Michael A. Terrell
Central Florida

 

[Costas Vlachos]

"John Larkin" <jjlarkin@highlandSNIPtechTHISnologyPLEASE.com> wrote in
message news:efrmsv4d8du7v2fe1k4621vsc7h0811im8@4ax.com...

The 'thin' rubber films are typically 5-10 mils thick, which is huge
in this game. They are rotten thermal conductors, even if you believe
their published specs. They dominate thermal resistance except for the
wimpiest of heat sinks.

John

Agreed, I don't like them either... I'd avoid using them for any serious
heatsinking work. But for standard CPU usage they seem to be adequate. If
the guy isn't going to overclock anyway, then why bother at all? But yes,
they are to be avoided for serious work. What I use for power BJTs/MOSFETs
etc., is good quality thermal paste and clips to ensure even pressure. And I
try to insulate the heatsink instead of the device(s) if possible.

Costas

 

[Guy Macon]

Michael A. Terrell <mike.terrell@earthlink.net> says...

I wonder if they buy the stuff from Mr. Eustace Haney?

I believe they buy it from Inspector Gadget.

 

[DarkMatter]

On Mon, 1 Dec 2003 14:56:14 +0000 (UTC), "Miles Golding"
<mjgoldingNO@SPAMbtopenworld.com> Gave us:

This question is timely for me. I'm about to build my own (first time) and
have bought an Intel P4 2.4 /heatsink pack. Having researched a fair bit
before starting the job, I discovered thermal paste. Is it really necessary?
I don't intend to overclock, so will the thin film already attached to the
base of the heatsink do the job adequately?

Yes, it will. Quite adequately, in fact.

 

[DarkMatter]

On Mon, 01 Dec 2003 06:37:23 -0900, Floyd Davidson <floyd@barrow.com>
Gave us:

First, if you saw that much difference, then it is *obvious* that
the initial installation with OEM heat sink compound was flawed.
10C is *way* too much!

Exactly. 2 or 3 degrees is the most one will see, at best.

Hardly worth the price or the benefit.

 

[DarkMatter]

On Mon, 01 Dec 2003 15:44:04 +0000, Guy Macon
<http://www.guymacon.com> Gave us:

Point well taken. The OEM heat sink compound was applied at a
Compaq factory in 1996, and I replaced it in 2003, so age is
almost certainly a factor.

It's a quad CPU unit from over 7 years ago? HAHAHA!

The CPUs of today exhibit more heat than those old suckers do.

Bet my Dually beats it on the benchmark too!

 

[DarkMatter]

On Mon, 01 Dec 2003 08:11:23 -0800, Jim Weir <jim@rst-engr.com> Gave
us:

BTW, the "thin rubber fiber" that some other folks are talking about are from a
company called Bergquist if I'm not mistaken. Expensive, but damned good
thermal conductivity.

They are impregnated with the silicates for gap filling and actually
have better dissipation numbers than pastes do.

 

[DarkMatter]

On Mon, 1 Dec 2003 8:12:04 -0800, DaveC <me@privacy.net> Gave us:

On Mon, 1 Dec 2003 7:37:23 -0800, Floyd Davidson wrote
(in message <87u14kblcc.fld@barrow.com&gt:

NAPA Item# 765-2569
Copper Anti-Seize Lubriant

and you recommend this as a head-sink compound for semiconductors?

You have obviously not been around very long.

The discussion is centering around a silver based media, so if
conductivity is what has your bloomers in a bunch, it is not an issue.

The material he mentions isn't a lube, it's a spray used to seal
high pressure metal gaskets, like a head gasket.

It works absolutely flawlessly.

 

[DarkMatter]

On 1 Dec 2003 16:34:34 GMT, et472@FreeNet.Carleton.CA (Michael Black)
Gave us:

It strikes me that it's right up there with the audio consumers
and gold plated cables and ten dollar capacitors.

No. There are actually measurable differences in this case.

The thing is whether those minuscule differences are worth the extra
cost.

To me... they are not. I know that my *normal* heat sink compound
is just fine for the job.

Hey... look at toilet paper. To me, the industrial/institutional
variety is the best value, as it has the most paper on the roll. and
is quite adequate in the "softness" realm. Consumer TP is designed to
be used up quickly on a loosely filled roll, that has little on it,
marked with "Extra Fluffy" or some such stupidity. Consumers are not
the best analysts of quality, unless challenged. THEN they wake the
fuck up... maybe.

All it is is another way for them to sell us LESS for MORE.

The heat sink cream is better... about one or two percent better.
Hardly worth the cost.

If I OC my PC, it will have an active cooling method employed on it.

 

[DarkMatter]

On Mon, 01 Dec 2003 08:35:25 -0800, John Larkin
<jjlarkin@highlandSNIPtechTHISnologyPLEASE.com> Gave us:

*don't* use a silicone sil-pad or phase-change stuff.
They are both awful thermally.

Yer an idiot.

 

[DarkMatter]

On Mon, 01 Dec 2003 18:22:19 +0000, Tim Auton <tim.auton@uton.[group
sex without the y on the end]> Gave us:

I can see the logic, but you're talking about some damn fine
manufacturing tolerances there. Anybody actually sell heatsinks like
that?

Almost all laterally extruded heat sinks have the tilt on the device
mating surface.

Vertical extrusion types do not, and stamped types do not.

 

[DarkMatter]

On Mon, 1 Dec 2003 18:24:28 +0000 (UTC), "Costas Vlachos"
<c-X-vlachos@hot-X-mail.com> Gave us:

Agreed, I don't like them either... I'd avoid using them for any serious
heatsinking work.

They ARE in use, across the industry... for "Serious heat sinking
work". Doh!

But for standard CPU usage they seem to be adequate. If
the guy isn't going to overclock anyway, then why bother at all? But yes,
they are to be avoided for serious work. What I use for power BJTs/MOSFETs
etc., is good quality thermal paste and clips to ensure even pressure. And I
try to insulate the heatsink instead of the device(s) if possible.

Hahaha... The industry uses them, on a regular basis. Large and
small package form factors, large and small wattages.

You are both fucking nuts.

 

[John Larkin]

On Mon, 01 Dec 2003 18:23:15 -0800, DarkMatter
<DarkMatter@thebarattheendoftheuniverse.org> wrote:

On Mon, 1 Dec 2003 18:24:28 +0000 (UTC), "Costas Vlachos"
c-X-vlachos@hot-X-mail.com> Gave us:


Agreed, I don't like them either... I'd avoid using them for any serious
heatsinking work.

They ARE in use, across the industry... for "Serious heat sinking
work". Doh!

But for standard CPU usage they seem to be adequate. If
the guy isn't going to overclock anyway, then why bother at all? But yes,
they are to be avoided for serious work. What I use for power BJTs/MOSFETs
etc., is good quality thermal paste and clips to ensure even pressure. And I
try to insulate the heatsink instead of the device(s) if possible.

Hahaha... The industry uses them, on a regular basis. Large and
small package form factors, large and small wattages.

You are both fucking nuts.

DorkMatter,

Just do the math.

John

 

[John Larkin]

On Mon, 01 Dec 2003 18:10:54 -0800, DarkMatter
<DarkMatter@thebarattheendoftheuniverse.org> wrote:

On Mon, 01 Dec 2003 08:35:25 -0800, John Larkin
jjlarkin@highlandSNIPtechTHISnologyPLEASE.com> Gave us:


*don't* use a silicone sil-pad or phase-change stuff.
They are both awful thermally.

Yer an idiot.

DimBulb,

Do the math.

John

 

[Floyd Davidson]

DarkMatter <DarkMatter@thebarattheendoftheuniverse.org> wrote:

On Mon, 1 Dec 2003 8:12:04 -0800, DaveC <me@privacy.net> Gave us:

On Mon, 1 Dec 2003 7:37:23 -0800, Floyd Davidson wrote
(in message <87u14kblcc.fld@barrow.com&gt:

NAPA Item# 765-2569
Copper Anti-Seize Lubriant

and you recommend this as a head-sink compound for semiconductors?


You have obviously not been around very long.

The discussion is centering around a silver based media, so if
conductivity is what has your bloomers in a bunch, it is not an issue.

The material he mentions isn't a lube, it's a spray used to seal
high pressure metal gaskets, like a head gasket.

Actually this stuff is a lube. It comes in a little container
with a brush on the inside of the cap. It is kind of an oily
looking stuff, with a lot of what appears to be copper flakes
suspended in it. (And it is essential to mix it up very good
before applying it, or all the copper will be on the bottom of
the container.)

It is *not* used on head gaskets, but rather on head bolts. The
idea is to smear them with this stuff in hopes that 10 years
later you might be able to unscrew them all without breaking
half of them.

It works absolutely flawlessly.

It does indeed.

--
Floyd L. Davidson <http://web.newsguy.com/floyd_davidson>
Ukpeagvik (Barrow, Alaska) floyd@barrow.com

 

[Cameron Dorrough]

"Floyd Davidson" <floyd@barrow.com> wrote in message
news:87y8tv7vo2.fld@barrow.com...

DarkMatter <DarkMatter@thebarattheendoftheuniverse.org> wrote:
On Mon, 1 Dec 2003 8:12:04 -0800, DaveC <me@privacy.net> Gave us:

On Mon, 1 Dec 2003 7:37:23 -0800, Floyd Davidson wrote
(in message <87u14kblcc.fld@barrow.com&gt:

NAPA Item# 765-2569
Copper Anti-Seize Lubriant

and you recommend this as a head-sink compound for semiconductors?


You have obviously not been around very long.

The discussion is centering around a silver based media, so if
conductivity is what has your bloomers in a bunch, it is not an issue.

The material he mentions isn't a lube, it's a spray used to seal
high pressure metal gaskets, like a head gasket.

Actually this stuff is a lube. It comes in a little container
with a brush on the inside of the cap. It is kind of an oily
looking stuff, with a lot of what appears to be copper flakes
suspended in it. (And it is essential to mix it up very good
before applying it, or all the copper will be on the bottom of
the container.)

It is *not* used on head gaskets, but rather on head bolts. The
idea is to smear them with this stuff in hopes that 10 years
later you might be able to unscrew them all without breaking
half of them.

... it's also used on large battery banks to anti-seize the
bus-bar-to-battery terminal bolts and to protect the connections from
corrosion.

It works absolutely flawlessly.

It does indeed.

Most certainly - though it's not cheap.

JM2CW,
Cameron