Heat sink grease

When mounting power transistors and chips to a heat sink they often use
heat shrink grease to help dissipate the heat. Over the years I have
replaced or added this grease which I purchased at Radio Shack. One tube
lasted many years for the occasional times I needed it.

I dont need any right now, but I do not have any more on hand for the
next time i need it.

Anyhow, I understand this grease is pure 100% silicone grease. I've seen
it in both clear and white. I just tuned up my car and replaced the
plugs and wires. The auto parts store guy suggested using something
called "Spark Plug boot grease". It was only $3 and I know all too well
how plug wires tend to get stuck on the plugs and by the time I get them
off, I ripped off the terminal inside the boot. So, I bought and used
this grease. (which I know also repels moisture from getting into the
plug boots).

There is a lot left in the tube and I was just putting it away so I can
find it the next time I need it. I was reading the package and it says
100% pure silicone grease.

Hmmmm, wouldn't this be the same stuff used to mount semiconductors to
heat sinks? It sure does look the same too... If it's the same, I'll
have some for heat sink use too, and in that case, it gets put in with
my electronics repair stuff, rather than in my garage. I doubt it's
affected by cold weather, but keeping it in the house will keep it at
room temp all year long, rather than freezing in the garage in winter.

I thought I'd post this to see if anyone has used this for
semiconductors on heat sink, grease, or is there some reason NOT to use
this stuff for that purpose?

(If this is the same stuff, it's a heck of a lot cheaper to buy it at an
auto parts store).

 

[Trevor Wilson]

On 12/04/2018 3:37 PM, oldschool@tubes.com wrote:

When mounting power transistors and chips to a heat sink they often use
heat shrink grease to help dissipate the heat. Over the years I have
replaced or added this grease which I purchased at Radio Shack. One tube
lasted many years for the occasional times I needed it.

I dont need any right now, but I do not have any more on hand for the
next time i need it.

Anyhow, I understand this grease is pure 100% silicone grease. I've seen
it in both clear and white. I just tuned up my car and replaced the
plugs and wires. The auto parts store guy suggested using something
called "Spark Plug boot grease". It was only $3 and I know all too well
how plug wires tend to get stuck on the plugs and by the time I get them
off, I ripped off the terminal inside the boot. So, I bought and used
this grease. (which I know also repels moisture from getting into the
plug boots).

There is a lot left in the tube and I was just putting it away so I can
find it the next time I need it. I was reading the package and it says
100% pure silicone grease.

Hmmmm, wouldn't this be the same stuff used to mount semiconductors to
heat sinks? It sure does look the same too... If it's the same, I'll
have some for heat sink use too, and in that case, it gets put in with
my electronics repair stuff, rather than in my garage. I doubt it's
affected by cold weather, but keeping it in the house will keep it at
room temp all year long, rather than freezing in the garage in winter.

I thought I'd post this to see if anyone has used this for
semiconductors on heat sink, grease, or is there some reason NOT to use
this stuff for that purpose?

(If this is the same stuff, it's a heck of a lot cheaper to buy it at an
auto parts store).

**It's not the same stuff. Heat sink grease, is not really grease. It is
a thermally conductive powder, suspended in a suitable medium, that
allows it to be spread on a semiconductor. Focus on the words:
"Thermally conductive".

I recall that clear stuff from the early 1970s. It is significantly
inferior to more modern products.

This will do the trick, at a reasonable price:

https://www.digikey.com.au/product-detail/en/wakefield-vette/126-4/345-1011-ND/340309

Don't muck about, buy the correct stuff to do the job.

--
Trevor Wilson
www.rageaudio.com.au

 

On Thu, 12 Apr 2018 15:48:53 +1000, Trevor Wilson
<trevor@SPAMBLOCKrageaudio.com.au> wrote:

On 12/04/2018 3:37 PM, oldschool@tubes.com wrote:
When mounting power transistors and chips to a heat sink they often use
heat shrink grease to help dissipate the heat. Over the years I have
replaced or added this grease which I purchased at Radio Shack. One tube
lasted many years for the occasional times I needed it.

I dont need any right now, but I do not have any more on hand for the
next time i need it.

Anyhow, I understand this grease is pure 100% silicone grease. I've seen
it in both clear and white. I just tuned up my car and replaced the
plugs and wires. The auto parts store guy suggested using something
called "Spark Plug boot grease". It was only $3 and I know all too well
how plug wires tend to get stuck on the plugs and by the time I get them
off, I ripped off the terminal inside the boot. So, I bought and used
this grease. (which I know also repels moisture from getting into the
plug boots).

There is a lot left in the tube and I was just putting it away so I can
find it the next time I need it. I was reading the package and it says
100% pure silicone grease.

Hmmmm, wouldn't this be the same stuff used to mount semiconductors to
heat sinks? It sure does look the same too... If it's the same, I'll
have some for heat sink use too, and in that case, it gets put in with
my electronics repair stuff, rather than in my garage. I doubt it's
affected by cold weather, but keeping it in the house will keep it at
room temp all year long, rather than freezing in the garage in winter.

I thought I'd post this to see if anyone has used this for
semiconductors on heat sink, grease, or is there some reason NOT to use
this stuff for that purpose?

(If this is the same stuff, it's a heck of a lot cheaper to buy it at an
auto parts store).


**It's not the same stuff. Heat sink grease, is not really grease. It is
a thermally conductive powder, suspended in a suitable medium, that
allows it to be spread on a semiconductor. Focus on the words:
"Thermally conductive".

I recall that clear stuff from the early 1970s. It is significantly
inferior to more modern products.

This will do the trick, at a reasonable price:

https://www.digikey.com.au/product-detail/en/wakefield-vette/126-4/345-1011-ND/340309

Don't muck about, buy the correct stuff to do the job.

I guess that answers this question.

Thanks for the speedy reply.

I kind of thought that there could be a special composition involved,
but I had to ask to find out.

Thanks again.

 

I like Dow 340. Also, do you know how to apply it ? You put a daub or a bead near or between the mounting screw(s) and tighten to squeeze out the excess. Spreading it can cause air pockets which defeats the purpose. This method prevents that.

You probably won't find a website that tells this method, but I know of what I speak. Just like you won't find on the web WHY to use 100 % ethylene glycol for your engine coolant, but I can make me case and it is hard to argue against. Anyone wants to dispute it make your case and then my case will bury you. Guaranteed.

You get sort of a feel for how big of a daub or bead to apply. Ideally it should stick slightly out the sides of the device, that assures full coverage. You don't want a huge amount of excess. Also in some cases it is a good idea to push the device on to the heatsink by hand with as much force as possible. Screws strip in aluminum and some of them have a spring type hold down.

The clear stuff is no good. Even the best stuff is actually a thermal insulator but it beats air. the white stuff has microscopic particles in suspension and they do most of the heat conduction. the clear stuff doesn't have it. Don't use it.

 

[jfeng@my-deja.com]

On Thursday, April 12, 2018 at 6:06:05 AM UTC-7, tabb...@gmail.com wrote:

If you're stuck for some, toothpaste is surprisingly good. I tried it on a
high diss AMD 8 core CPU, it only ran a couple of degrees hotter than the
proper stuff.
If you need electrical insulation, petroleum jelly is probably better. The main objective is to fill the air gaps.

 

On Thursday, April 12, 2018 at 7:51:56 AM UTC-5, pf...@aol.com wrote:

I like these:

https://www.digikey.com/product-detail/en/bergquist/SP900S-0.009-00-02/BER171-ND/307793


Peter Wieck
Melrose Park, PA

Are those the gray rubbery ones ? I like them, no goop required. One place I worked told us to use goop with them anyway, they were kind of anal about things.

 

On Thursday, 12 April 2018 06:38:19 UTC+1, olds...@tubes.com wrote:

When mounting power transistors and chips to a heat sink they often use
heat shrink grease to help dissipate the heat. Over the years I have
replaced or added this grease which I purchased at Radio Shack. One tube
lasted many years for the occasional times I needed it.

I dont need any right now, but I do not have any more on hand for the
next time i need it.

Anyhow, I understand this grease is pure 100% silicone grease. I've seen
it in both clear and white. I just tuned up my car and replaced the
plugs and wires. The auto parts store guy suggested using something
called "Spark Plug boot grease". It was only $3 and I know all too well
how plug wires tend to get stuck on the plugs and by the time I get them
off, I ripped off the terminal inside the boot. So, I bought and used
this grease. (which I know also repels moisture from getting into the
plug boots).

There is a lot left in the tube and I was just putting it away so I can
find it the next time I need it. I was reading the package and it says
100% pure silicone grease.

Hmmmm, wouldn't this be the same stuff used to mount semiconductors to
heat sinks? It sure does look the same too... If it's the same, I'll
have some for heat sink use too, and in that case, it gets put in with
my electronics repair stuff, rather than in my garage. I doubt it's
affected by cold weather, but keeping it in the house will keep it at
room temp all year long, rather than freezing in the garage in winter.

I thought I'd post this to see if anyone has used this for
semiconductors on heat sink, grease, or is there some reason NOT to use
this stuff for that purpose?

(If this is the same stuff, it's a heck of a lot cheaper to buy it at an
auto parts store).

heatsink compound is grese loaded with zinc oxide powder. The ZO provides the thermal conductivity.

If you're stuck for some, toothpaste is surprisingly good. I tried it on a high diss AMD 8 core CPU, it only ran a couple of degrees hotter than the proper stuff.


NT

 

[pfjw@aol.com]

I like these:

https://www.digikey.com/product-detail/en/bergquist/SP900S-0.009-00-02/BER171-ND/307793


Peter Wieck
Melrose Park, PA

 

[John-Del]

On Thursday, April 12, 2018 at 3:30:24 AM UTC-4, jurb...@gmail.com wrote:

I like Dow 340.

I've got several tubes of Dow 340. Very heavy per fluid oz; lots of solids. Expensive stuff but worth it.

 

On Thursday, 12 April 2018 14:35:16 UTC+1, jf...@my-deja.com wrote:

On Thursday, April 12, 2018 at 6:06:05 AM UTC-7, tabby wrote:
If you're stuck for some, toothpaste is surprisingly good. I tried it on a
high diss AMD 8 core CPU, it only ran a couple of degrees hotter than the
proper stuff.
If you need electrical insulation, petroleum jelly is probably better. The main objective is to fill the air gaps.

Petroleum jelly melts & runs.


NT

 

[Mike Coon]

In article <2890347f-7c06-4b80-8025-d11f8ba8991b@googlegroups.com>,
jurb6006@gmail.com says...

On Thursday, April 12, 2018 at 7:51:56 AM UTC-5, pf...@aol.com wrote:
I like these:

https://www.digikey.com/product-detail/en/bergquist/SP900S-0.009-00-02/BER171-ND/307793


Peter Wieck
Melrose Park, PA

Are those the gray rubbery ones ? I like them, no goop required. One place I worked told us to use goop with them anyway, they were kind of anal about things.

The web page says they are pink; probably why Peter likes them!

Mike.

 

[Lee]

On 12/04/2018 08:30, jurb6006@gmail.com wrote:

> You probably won't find a website that tells this method, but I know of what I speak. Just like you won't find on the web WHY to use 100 % ethylene glycol for your engine coolant, but I can make me case and it is hard to argue against. Anyone wants to dispute it make your case and then my case will bury you. Guaranteed.

I remember when we had this argument at work after a certain automotive
TV program claimed the same.
We finally settled for 60% mix being better because the higher SHC over
100% ethylene glycol was preferred in a heat transfer system.
Knowing how marginal some automotive cooling systems are, I'd plump for
the 60% mix personally, even if there are other benefits to using 100%.

This was a few years back though, maybe newer coolants behave differently?

 

[pfjw@aol.com]

They are a very subtle shade of Dove Grey....

Peter Wieck
Melrose Park, PA

 

[Jeff Liebermann]

On Thu, 12 Apr 2018 00:37:18 -0500, oldschool@tubes.com wrote:

I thought I'd post this to see if anyone has used this for
semiconductors on heat sink, grease, or is there some reason NOT to use
this stuff for that purpose?

The best way to use thermal goo is to have the heat sink and device
touch each other with direct metal to metal contact, and with whatever
thermal goo you select filling in only the gaps. Making a thermal
sandwich with a thick layer of thermal goo which prevents direct
contact doesn't work very well. Idea is to grind flat and mirror
polish the heat sink and whatever is getting hot to get more metal to
metal contact, and use very little thermal goo. Some CPU's and video
chips have a mirror finish.

However, if you want to roll your own, diamonds are your best friend.
Here's a list of thermal conductivity of various compounds and
concoctions.
W/m*K
Diamond 1000
h-BN 600 (boron nitride)
c-BN 740 (boron nitride)
Silver 406
Copper 385
Gold 314
AlN 285 (aluminum nitride ceramic)
Aluminum 205
Graphite 200
Carbon 150
SiC 120 (silicon carbide)
Brass 109
ZnO 50 (zinc oxide)
Al2O3 25 (aluminum oxide ceramic)
TiO2 10 (titanium dioxide)

The common white thermal grease is zinc oxide, aluminum oxide, or
both. The expensive stuff adds boron nitride. The clear stuff is
just the grease, which is totally useless for thermal conduction but
might be a handy carrier if you want you mix your own using various
powders.

Tooth paste is the most common "alternative" thermal goo, although sun
screen which contains zinc oxide and perhaps titanium oxide should
also work.

If you can used and control a conductive thermal goo, try "liquid
metal", and alloy of bismuth and indium:
<https://www.gamersnexus.net/guides/3064-intels-thermal-problem-pt1-liquid-metal-vs-thermal-paste-benchmarks-7900x>
<http://www.thermal-grizzly.com/en/products/26-conductonaut-en>

Good luck.

--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

 

[Jeff Liebermann]

On Thu, 12 Apr 2018 00:30:20 -0700 (PDT), jurb6006@gmail.com wrote:

>I like Dow 340.

I have several tubes of the stuff. It works, but I prefer Arctic
Silver 5 for CPU overclocking and RF power devices. Actually, I
prefer lapping and polishing the device and heatsink, but that's not
always an option. A small tube of Arctic Silver 5 lasts me about 20
i7 size CPU's or about $0.25/per CPU because I use very little.

Also, do you know how to apply it? You put a daub or a bead
near or between the mounting screw(s) and tighten to squeeze out
the excess. Spreading it can cause air pockets which defeats the
purpose. This method prevents that.

That's not very compatible with my dictum "The less thermal goo you
use, the better it works". You want metal to metal as much as
possible with the thermal goo just filling in the cracks and gouges.

You probably won't find a website that tells this method, but I
know of what I speak. Just like you won't find on the web WHY
to use 100 % ethylene glycol for your engine coolant, but I
can make me case and it is hard to argue against. Anyone wants
to dispute it make your case and then my case will bury you.
Guaranteed.

I've posted my test results in sci.electronics.design. I'll see if I
can find the article. You might find it of interest.

You get sort of a feel for how big of a daub or bead to apply.
Ideally it should stick slightly out the sides of the device,
that assures full coverage. You don't want a huge amount of
excess. Also in some cases it is a good idea to push the device
on to the heatsink by hand with as much force as possible.
Screws strip in aluminum and some of them have a spring type
hold down.

Sigh.

The clear stuff is no good. Even the best stuff is actually a
thermal insulator but it beats air. the white stuff has microscopic
particles in suspension and they do most of the heat conduction.
the clear stuff doesn't have it. Don't use it.

At least we agree on that part.

--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

 

[Jeff Liebermann]

On Thu, 12 Apr 2018 10:16:55 -0700, Jeff Liebermann <jeffl@cruzio.com>
wrote:

I've posted my test results in sci.electronics.design. I'll see if I
can find the article. You might find it of interest.

This should work:
<https://groups.google.com/d/msg/sci.electronics.design/s_TJxYnypVk/1oSFPQdCBwAJ>

A bit on tooth paste:
<https://groups.google.com/d/msg/sci.electronics.design/s_TJxYnypVk/545qJWVLBwAJ>

Using gold leaf:
<https://groups.google.com/d/msg/sci.electronics.design/s_TJxYnypVk/yCgpwwtQBwAJ>

Plenty more in that thread:

I found a more complete list of thermal conductivity:

W/m*K
Diamond 1000
c-BN 740 (Cubic Boron Nitride)
h-BN 600 (Hexagonal Boron Nitride)
Silver 406
Copper 385
Gold 314
AlN 285 (aluminum nitride ceramic)
Aluminum 205
Graphite 200
Carbon 150
SiC 120
Brass 109
Indium 86
ZnO 50 (zinc oxide)
Al2O3 25 (aluminum oxide ceramic)
Pastes 4.0
SilPad 2000 3.5
Circuit Works 1.84
Dow Corning 340 0.67

Note that most white thermal goo at the low end of the thermal
conductivity list with Dow Corning 340 in the really awful category.
Don't believe me? See:
<http://www4.dowcorning.com/applications/search/products/details.aspx?prod=01015443&type=PROD>
and look in the box under "properties".
Thermal Conductivity = 0.67 Watts per meter K


--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

 

[John-Del]

On Thursday, April 12, 2018 at 1:17:00 PM UTC-4, Jeff Liebermann wrote:

You get sort of a feel for how big of a daub or bead to apply.
Ideally it should stick slightly out the sides of the device,
that assures full coverage. You don't want a huge amount of
excess. Also in some cases it is a good idea to push the device
on to the heatsink by hand with as much force as possible.
Screws strip in aluminum and some of them have a spring type
hold down.

Sigh.

Not sure where you disagree Jeff. Not speaking for the other Jeff (jurb), but I agree with him. If you get too much compound out of the sides, you've wasted it. If none appears you may not have put enough.

The idea is to squeeze as much out as is possible ensuring that only enough to fill gaps is left, on that we all agree. Forcing the device down by external pressure helps make sure any excess is forced out without relying on the devices own hardware to accomplish.

The center daub verses the even spread is debatable either way. RCA back in the late 60s told us to use the center daub and let it spread out on it's own. Their reason was to avoid air pockets. On small devices like transistors, that's what I do. On larger devices, I don't think enough pressure can be exerted to ensure a large area can force enough compound out of the joint, so I tend to carefully spread as thin a layer as I can accomplish.

 

On Thu, 12 Apr 2018 10:31:09 -0700, Jeff Liebermann <jeffl@cruzio.com>
wrote:

On Thu, 12 Apr 2018 10:16:55 -0700, Jeff Liebermann <jeffl@cruzio.com
wrote:

I've posted my test results in sci.electronics.design. I'll see if I
can find the article. You might find it of interest.

This should work:
https://groups.google.com/d/msg/sci.electronics.design/s_TJxYnypVk/1oSFPQdCBwAJ

A bit on tooth paste:
https://groups.google.com/d/msg/sci.electronics.design/s_TJxYnypVk/545qJWVLBwAJ

Using gold leaf:
https://groups.google.com/d/msg/sci.electronics.design/s_TJxYnypVk/yCgpwwtQBwAJ

Plenty more in that thread:

I found a more complete list of thermal conductivity:

W/m*K
Diamond 1000
c-BN 740 (Cubic Boron Nitride)
h-BN 600 (Hexagonal Boron Nitride)
Silver 406
Copper 385
Gold 314
AlN 285 (aluminum nitride ceramic)
Aluminum 205
Graphite 200
Carbon 150
SiC 120
Brass 109
Indium 86
ZnO 50 (zinc oxide)
Al2O3 25 (aluminum oxide ceramic)
Pastes 4.0
SilPad 2000 3.5
Circuit Works 1.84
Dow Corning 340 0.67

Note that most white thermal goo at the low end of the thermal
conductivity list with Dow Corning 340 in the really awful category.
Don't believe me? See:
http://www4.dowcorning.com/applications/search/products/details.aspx?prod=01015443&type=PROD
and look in the box under "properties".
Thermal Conductivity = 0.67 Watts per meter K

I use some very fine particle size diamond lapping compounds in my
shop. Maybe I should consider using some for heat sink compound next
time I want to overclock the crap out of something. In fact, it could
first be used for lapping the parts and then just left in place.
Eric

 

[Jeff Liebermann]

On Thu, 12 Apr 2018 11:59:48 -0700, etpm@whidbey.com wrote:

I use some very fine particle size diamond lapping compounds in my
shop.

What's the grain diameter? If it's larger than the depth of the
cracks and crevasses in the aluminum extruded heat sink, then you're
taking a step backwards. The bulk of the heat is passed by metal to
metal contact.

Maybe I should consider using some for heat sink compound next
time I want to overclock the crap out of something. In fact, it could
first be used for lapping the parts and then just left in place.

As I vaguely recall, the typical flatness specs for extruded heat
sinks is something like 0.007 in/in. So, if the heat sink is 2 inches
wide, it can be warped 0.014" and still be considered usable. Ugh. It
will take an awfully large amount of heat sink goo to fill a 0.014"
gap.

The lack of flatness also causes problems when one tries to using such
a heat sink as a lapping plate. If the CPU can is softer than the
aluminum, it will convert the formerly flat can into a warped version
of the not very flat heat sink. I did stuff like this in the distant
past and found a really flat lapping plat to be a necessity.

If you want to see how bad it can get, take a black felt tip pen and
"paint" the top of the CPU or the heatsink black. Find a flat surface
(glass is good). Drop a sheet of fine emery cloth (2000 grit) on the
flat plate. Move the heat sink very slightly over the sand paper.
Inspect the heatsink. Where the heat sink was raised, the sandpaper
will remove the ink. Where there was a depression, the paint will
remain in place. Extra credit for repeating the test with a hot or
cold heat sink. Thermal expansion in an extrusion isn't uniform and
you will see bending in the heatsink.

There's also a problem with diamond dust and grease. While the grease
is great of keeping the diamond dust in place and making it easy to
apply, thermal cycling tends to pump the grease away from hot spots
leaving rather voids. You can sometimes see this when disassembling a
CPU and heatsink combination that has been running for a few years. If
you take it apart immediately after assembly, the thermal goo is
uniformly distributed (as long as the heat sink is reasonably flat).
However a few hundred thermal cycles later, it will have voids.

--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

 

[Jeff Liebermann]

On Thu, 12 Apr 2018 12:05:53 -0700 (PDT), John-Del <ohger1s@gmail.com>
wrote:

On Thursday, April 12, 2018 at 1:17:00 PM UTC-4, Jeff Liebermann wrote:


You get sort of a feel for how big of a daub or bead to apply.
Ideally it should stick slightly out the sides of the device,
that assures full coverage. You don't want a huge amount of
excess. Also in some cases it is a good idea to push the device
on to the heatsink by hand with as much force as possible.
Screws strip in aluminum and some of them have a spring type
hold down.

Sigh.

Not sure where you disagree Jeff.

Sigh 2.0

>Not speaking for the other Jeff (jurb), but I agree with him.

That makes three Jeff's. Nobody ever agrees with me, so you must be
referring to Jeff 1.0.

If you get too much compound out of the sides, you've wasted it.
If none appears you may not have put enough.

The idea is to squeeze as much out as is possible ensuring that
only enough to fill gaps is left, on that we all agree. Forcing
the device down by external pressure helps make sure any excess
is forced out without relying on the devices own hardware to
accomplish.

The center daub verses the even spread is debatable either way.
RCA back in the late 60s told us to use the center daub and let
it spread out on it's own. Their reason was to avoid air pockets.
On small devices like transistors, that's what I do. On larger
devices, I don't think enough pressure can be exerted to ensure
a large area can force enough compound out of the joint, so
I tend to carefully spread as thin a layer as I can accomplish.

As usual, I beg to differ:

1. If you apply enough pressure to make the thermal goo ooooze out
from the sandwich, you will either bend the device (as in a TO-3 or RF
power xsistor) or simply not be able to apply enough pressure with the
mounting screws and springs. I was going to calculate or measure the
pressure applied based on the recommended mounting screw torque for
various heat sinks, but don't have the time. I might have some time
this weekend to sandwich some Dow Corning 340 between two metal slabs
in an arbor press with a load cell and see what it really takes to
squeeze out the thermal goo. As you mention, I also suspect it will
be rather high pressure, far beyond what can be done with small
fasteners. Offhand, I would guess that this method was invented by a
thermal goo salesman in order to inspire users to consume more of his
company's thermal goo.

2. The problem with forming bubbles (voids) in the thermal goo was
originally mentioned in the literature in reference to removing the
heat sink after the thermal goo was applied. That will certainly
create bubbles. Somehow, someone extended that to the initial
application of thermal goo, which is not correct. I've never bothered
to prove this, so if you have time, try a blob of your favorite
thermal goo between two glass plates and look at it under a
microscope. If there are any bubbles, they should be obvious. If you
want, I can do this Friday as I have the necessary equipment:
<http://802.11junk.com/jeffl/pics/microscopes/Olympus%20BH/slides/Olympus%20BH.html>
3. Thermal goo is fairly cheap. However, it does make a mess when
the excess is squeezed out of the heat sink sandwich. I prefer not
cleaning up the mess.

My method of applying thermal go is fairly simple. Find the direction
on the heat sink in which the machining marks and gouges run. Apply a
tiny amount of thermal goo to the heat sink. Use a plastic razor
blade:
<https://www.google.com/search?q=plastic+razor+blade&tbm=isch>
to smear the thermal goo in the direction of ACROSS the machine marks.
The idea is to push the thermal goo into the grooves, crevasses, and
gouges in the heat sink. If the mating part of the sandwich also has
machine marks, do the same thing. The initial blob of thermal goo is
intentionally insufficient to cover the entire heat sink. Add small
blobs and continue to smear until the surface is covered. You will
see quite a bit of metal. That's good as you want metal to metal
contact. If there's any excess, wipe it off with the plastic razor.
When done, clamp it together and DON'T take it apart for "inspection".
If you do, start over by wiping the heat sink clean with alcohol, let
dry, and do the blob thing again. Thermal goo tends to harden as the
carrier evaporates. It will not flow into the cracks easily and will
resist compression. Best to start from scratch.



--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558