Sign in with
Sign up | Sign in
Your question

Why so very small amount of thermal compound?

Last response: in Overclocking
Share
Anonymous
a b K Overclocking
June 16, 2004 3:00:02 AM

Archived from groups: alt.comp.hardware.overclocking (More info?)

I have read posts that suggest minimal thermal compound. The Arctic Silver
instructions also suggest a very small amount.
http://www.arcticsilver.com/arctic_silver_instructions....

I understand this advice if the concern is that too much compound may
contaminate the pins and cause electrical problems. But, I don't understand
it if the concern is heat transfer.
I realize that the objective is to only fill the microscopic gaps between
the two surfaces.

After the heat sink is mounted, the force pushing the two surfaces together
is so high that any extra compound will be squeezed out of the sides. The
compound is a fluid after all. There will be no compound between the peaks
(only in the valleys). Is that not true?
Am I missing something?

Thanks,

Navid
Anonymous
a b K Overclocking
June 16, 2004 3:00:03 AM

Archived from groups: alt.comp.hardware.overclocking (More info?)

> After the heat sink is mounted, the force pushing the two surfaces together
> is so high that any extra compound will be squeezed out of the sides. The
> compound is a fluid after all. There will be no compound between the peaks
> (only in the valleys). Is that not true?

Well, the same theory could be applied to water under a car's tires, and
yet cars DO hydroplane.

You are right the force is great, but some small amount will remain.
Anonymous
a b K Overclocking
June 16, 2004 3:00:03 AM

Archived from groups: alt.comp.hardware.overclocking (More info?)

Navid wrote:
> I have read posts that suggest minimal thermal compound. The Arctic Silver
> instructions also suggest a very small amount.
> http://www.arcticsilver.com/arctic_silver_instructions....
>
> I understand this advice if the concern is that too much compound may
> contaminate the pins and cause electrical problems. But, I don't understand
> it if the concern is heat transfer.
> I realize that the objective is to only fill the microscopic gaps between
> the two surfaces.
>
> After the heat sink is mounted, the force pushing the two surfaces together
> is so high that any extra compound will be squeezed out of the sides. The
> compound is a fluid after all. There will be no compound between the peaks
> (only in the valleys). Is that not true?
> Am I missing something?

The force isn't as high as you think and the thermal compound doesn't flow
as easily as you think.

>
> Thanks,
>
> Navid
>
>
>
Related resources
Anonymous
a b K Overclocking
June 16, 2004 6:13:15 AM

Archived from groups: alt.comp.hardware.overclocking (More info?)

> I understand this advice if the concern is that too much compound may
> contaminate the pins and cause electrical problems. But, I don't understand
> it if the concern is heat transfer.

The thermal conductivity of any fluid or suspension, however
well-designed, will never be as good as the thermal conductivity of a
metal. This is due to the way in which heat flows - heat is a
consequence of the random motion of atoms and is transmitted more
efficiently if the atoms are bonded in a regular (dense) structure, with
lots of (relatively) weak bonds to transmit the random vibrations along.

Hence, you want to apply as little as possible, only filling the valleys
between the surfaces (removing air which does not conduct heat well),
but not adding any more layers (of less thermal conductivity than the
heatsink) than is absolutely necessary between the chip and the sink.

The ultimate 'thermal compound' would be liquid metal that gets poured
between the chip and sink and forms a perfect seal upon cooling.
However, this would destroy your chip...

> I realize that the objective is to only fill the microscopic gaps between
> the two surfaces.

This is because the thermal conductivity of your Arctic Silver (or even
generic thermal grease) is greater than that of air, so you remove all
air pockets.

> After the heat sink is mounted, the force pushing the two surfaces together
> is so high that any extra compound will be squeezed out of the sides. The
> compound is a fluid after all. There will be no compound between the peaks
> (only in the valleys). Is that not true?

There will always be a small amount of thermal compound between the
peaks, however little, as it will be nigh-on impossible to remove all of
it purely by pressure. However, the separation between the peaks will be
much less than that between valleys, as the force is indeed rather large.
Anonymous
a b K Overclocking
June 16, 2004 7:47:28 AM

Archived from groups: alt.comp.hardware.overclocking (More info?)

"David Besack" <daveREMOVEbesack@mac.com> wrote in message
news:cao72n$sfq$1@netnews.upenn.edu...
>> After the heat sink is mounted, the force pushing the two surfaces
>> together
>> is so high that any extra compound will be squeezed out of the sides.
>> The
>> compound is a fluid after all. There will be no compound between the
>> peaks
>> (only in the valleys). Is that not true?
>
> Well, the same theory could be applied to water under a car's tires, and
> yet cars DO hydroplane.
>
> You are right the force is great, but some small amount will remain.

OK.
So, the pressure is not enough to get rid of all the compound between the
peaks, and a small layer remains.
Thanks for correcting me.
June 17, 2004 5:17:45 AM

Archived from groups: alt.comp.hardware.overclocking (More info?)

"Navid" <nospam@nospam.invalid> wrote in message
news:SbLzc.52262$tH1.1954660@twister.southeast.rr.com...
> I have read posts that suggest minimal thermal compound. The Arctic
Silver
> instructions also suggest a very small amount.
> http://www.arcticsilver.com/arctic_silver_instructions....
>
> I understand this advice if the concern is that too much compound may
> contaminate the pins and cause electrical problems. But, I don't
understand
> it if the concern is heat transfer.
> I realize that the objective is to only fill the microscopic gaps between
> the two surfaces.
>
> After the heat sink is mounted, the force pushing the two surfaces
together
> is so high that any extra compound will be squeezed out of the sides. The
> compound is a fluid after all. There will be no compound between the
peaks
> (only in the valleys). Is that not true?
> Am I missing something?

A layer will always remain, impeding heat transfer.

That said, you can glop on as much compound as you like, and the pressure of
the heatsink will push out enough to always give you half-decent
temperatures.

Unless you're a mad overclocker who wants every single megahertz and every
single 0.1 degree, you shouldn't worry about it. I certainly don't, and I
overclock, AND I've never had heat issues.
Anonymous
a b K Overclocking
June 17, 2004 5:54:10 AM

Archived from groups: alt.comp.hardware.overclocking (More info?)

On Thu, 17 Jun 2004 01:17:45 -0400, "Mercury" <nobody@nospam.null>
wrote:

>
>"Navid" <nospam@nospam.invalid> wrote in message
>news:SbLzc.52262$tH1.1954660@twister.southeast.rr.com...
>> I have read posts that suggest minimal thermal compound. The Arctic
>Silver
>> instructions also suggest a very small amount.
>> http://www.arcticsilver.com/arctic_silver_instructions....

>
>A layer will always remain, impeding heat transfer.
>
>That said, you can glop on as much compound as you like, and the pressure of
>the heatsink will push out enough to always give you half-decent
>temperatures.
>
>Unless you're a mad overclocker who wants every single megahertz and every
>single 0.1 degree, you shouldn't worry about it. I certainly don't, and I
>overclock, AND I've never had heat issues.
>


A wonderful approach.

The moron's guide to unending stupidity.

...of course, those with IQ's over 40 understand it's easier to just
use less.

but... the morons will always be with us.

Let's have a hand for the moron. And let me say it's always been such
a pleasure working with people like you.
Anonymous
a b K Overclocking
June 17, 2004 2:45:06 PM

Archived from groups: alt.comp.hardware.overclocking (More info?)

> A layer will always remain, impeding heat transfer.

True, but it won't impede heat transfer as much as air will - that's why
you apply the stuff in the first place. Adding too much could cause the
stuff to glop out and bridge 2 pins. Some thermal compounds are
conductive, so you have just b0rked your system by doing that. AS is
slightly capacitive, so it could do nasty things to system stability if
it were to flow across 2 pins, and others are electrically neutral (like
their Ceramique compound) so you could smear your whole board in it and
suffer no ill effects.

Plus your syringe of AS will last longer if you use the tiny amounts
prescribed.
Anonymous
a b K Overclocking
June 17, 2004 2:45:07 PM

Archived from groups: alt.comp.hardware.overclocking (More info?)

Lech Staniewicz wrote:

>> A layer will always remain, impeding heat transfer.
>
>
> True, but it won't impede heat transfer as much as air will

Not necessarily true. With the proper amount there is contact with only the
voids filled but with too much there is a layer of the thermal compound
between the two surfaces, separating them. Where, before, with 'dry, you
had contact with some air, too much has thermal compound between all of it
and that layer can be worse than the smaller amount of air in the dry mating.

See here:

http://power.ece.uiuc.edu/Balog/images/How%20to%20keep%...

The principles are the same for processor heatsinks.



> - that's why
> you apply the stuff in the first place. Adding too much could cause the
> stuff to glop out and bridge 2 pins. Some thermal compounds are
> conductive, so you have just b0rked your system by doing that. AS is
> slightly capacitive, so it could do nasty things to system stability if
> it were to flow across 2 pins, and others are electrically neutral (like
> their Ceramique compound) so you could smear your whole board in it and
> suffer no ill effects.
>
> Plus your syringe of AS will last longer if you use the tiny amounts
> prescribed.
June 17, 2004 9:01:20 PM

Archived from groups: alt.comp.hardware.overclocking (More info?)

<then@again.com> wrote in message
news:uam2d0t4skmrdcr7k6qdp9er9ha6bctak8@4ax.com...
>
> Let's have a hand for the moron.
>

Maybe instead of calling me a moron you could explain what it is I said that
you consider wrong.

Obviously glopping on heatsink compound is wasteful and runs the risk of
causing electrical problems.

My point was that one shouldn't worry about getting the
impossibly-ultra-thin layer that is mentined in the AS3 instructions.

And it's not "easier to use less". I tried the "half-grain of rice", and I
don't have the patience or dexterity to work with it. I put on enough to get
a nice even layer using a credit card. Like buttering toast.

If you don't like it, do it your own way, and don't me a moron.
Anonymous
a b K Overclocking
June 18, 2004 5:01:30 AM

Archived from groups: alt.comp.hardware.overclocking (More info?)

Navid wrote:

> I have read posts that suggest minimal thermal compound. The Arctic Silver
> instructions also suggest a very small amount.
> http://www.arcticsilver.com/arctic_silver_instructions....
>
> I understand this advice if the concern is that too much compound may
> contaminate the pins and cause electrical problems. But, I don't understand
> it if the concern is heat transfer.
> I realize that the objective is to only fill the microscopic gaps between
> the two surfaces.
>
> After the heat sink is mounted, the force pushing the two surfaces together
> is so high that any extra compound will be squeezed out of the sides. The
> compound is a fluid after all. There will be no compound between the peaks
> (only in the valleys). Is that not true?
> Am I missing something?
>
> Thanks,
>
> Navid

I've been following this thread and the responses have been long on
(mostly accurate) theory but short on practice. You can *feel* when
you've got it right!

- apply a dob of compound to the CPU core or heatspreader, spread it
thin with a blade (I prefer a boxcutter blade, but a credit card works),
remove any excess on non-contact areas using a cotton bud
- place the heatsink on top, apply medium pressure and work it in a
circular motion until you can feel the peaks grating against each other
- remove heatsink to inspect coverage, it will be slightly difficult to
unstick if you have it right, much easier if you have too little or too
much compound
- again remove any excess on non-contact areas with the cotton bud, and
use it to add a little compound to any bare contact areas
- repeat until you feel grating but find no excess or bare contact areas
upon heaksink removal
- work heaksink in a circular motion again briefly before final clamping

P2B
Anonymous
a b K Overclocking
June 18, 2004 3:51:46 PM

Archived from groups: alt.comp.hardware.overclocking (More info?)

"P2B" <p2b@sympatico.ca> wrote in message
news:1GuAc.50665$7H1.1718287@news20.bellglobal.com...
>
>
> Navid wrote:
>
>> I have read posts that suggest minimal thermal compound. The Arctic
>> Silver
>> instructions also suggest a very small amount.
>> http://www.arcticsilver.com/arctic_silver_instructions....
>>
>> I understand this advice if the concern is that too much compound may
>> contaminate the pins and cause electrical problems. But, I don't
>> understand
>> it if the concern is heat transfer.
>> I realize that the objective is to only fill the microscopic gaps between
>> the two surfaces.
>>
>> After the heat sink is mounted, the force pushing the two surfaces
>> together
>> is so high that any extra compound will be squeezed out of the sides.
>> The
>> compound is a fluid after all. There will be no compound between the
>> peaks
>> (only in the valleys). Is that not true?
>> Am I missing something?
>>
>> Thanks,
>>
>> Navid
>
> I've been following this thread and the responses have been long on
> (mostly accurate) theory but short on practice. You can *feel* when you've
> got it right!
>
> - apply a dob of compound to the CPU core or heatspreader, spread it thin
> with a blade (I prefer a boxcutter blade, but a credit card works), remove
> any excess on non-contact areas using a cotton bud
> - place the heatsink on top, apply medium pressure and work it in a
> circular motion until you can feel the peaks grating against each other
> - remove heatsink to inspect coverage, it will be slightly difficult to
> unstick if you have it right, much easier if you have too little or too
> much compound
> - again remove any excess on non-contact areas with the cotton bud, and
> use it to add a little compound to any bare contact areas
> - repeat until you feel grating but find no excess or bare contact areas
> upon heaksink removal
> - work heaksink in a circular motion again briefly before final clamping
>
> P2B


Thanks for everyone's replies.

The main point that I thought was made was that the pressure between the two
surfaces is not enough to spread the compound thin enough. That is why we
should apply a very small amount.

There is still a disconnect, in my understanding, between this and the
Arctic Silver 5 instructions for Pentium 4. The instruction is to place a
small amount of arctic silver 5 "without spreading it" and place the sink
with a wiggle and mount.

Now, if the pressure is not enough to spread the compound, then, how is this
technique going to make the peaks touch? Is it because the compound amount
is small, the contact area is small, hence, the pressure is higher (smaller
area, same force -> higher pressure)?

But, that makes no sense either. The purpose of this is to spread the
compound, which will increase the contact area. Larger area, and we are
back to the low pressure problem!

I don't mean to get into a senseless debate here. I would like to hear
opinions to help me understand. Thanks again.

Navid
Anonymous
a b K Overclocking
June 18, 2004 11:36:11 PM

Archived from groups: alt.comp.hardware.overclocking (More info?)

Navid wrote:

> "P2B" <p2b@sympatico.ca> wrote in message
> news:1GuAc.50665$7H1.1718287@news20.bellglobal.com...
>
>>
>>Navid wrote:
>>
>>
>>>I have read posts that suggest minimal thermal compound. The Arctic
>>>Silver
>>>instructions also suggest a very small amount.
>>>http://www.arcticsilver.com/arctic_silver_instructions....
>>>
>>>I understand this advice if the concern is that too much compound may
>>>contaminate the pins and cause electrical problems. But, I don't
>>>understand
>>>it if the concern is heat transfer.
>>>I realize that the objective is to only fill the microscopic gaps between
>>>the two surfaces.
>>>
>>>After the heat sink is mounted, the force pushing the two surfaces
>>>together
>>>is so high that any extra compound will be squeezed out of the sides.
>>>The
>>>compound is a fluid after all. There will be no compound between the
>>>peaks
>>>(only in the valleys). Is that not true?
>>>Am I missing something?
>>>
>>>Thanks,
>>>
>>>Navid
>>
>>I've been following this thread and the responses have been long on
>>(mostly accurate) theory but short on practice. You can *feel* when you've
>>got it right!
>>
>>- apply a dob of compound to the CPU core or heatspreader, spread it thin
>>with a blade (I prefer a boxcutter blade, but a credit card works), remove
>>any excess on non-contact areas using a cotton bud
>>- place the heatsink on top, apply medium pressure and work it in a
>>circular motion until you can feel the peaks grating against each other
>>- remove heatsink to inspect coverage, it will be slightly difficult to
>>unstick if you have it right, much easier if you have too little or too
>>much compound
>>- again remove any excess on non-contact areas with the cotton bud, and
>>use it to add a little compound to any bare contact areas
>>- repeat until you feel grating but find no excess or bare contact areas
>>upon heaksink removal
>>- work heaksink in a circular motion again briefly before final clamping
>>
>>P2B
>
>
>
> Thanks for everyone's replies.
>
> The main point that I thought was made was that the pressure between the two
> surfaces is not enough to spread the compound thin enough. That is why we
> should apply a very small amount.
>
> There is still a disconnect, in my understanding, between this and the
> Arctic Silver 5 instructions for Pentium 4. The instruction is to place a
> small amount of arctic silver 5 "without spreading it" and place the sink
> with a wiggle and mount.
>
> Now, if the pressure is not enough to spread the compound, then, how is this
> technique going to make the peaks touch? Is it because the compound amount
> is small, the contact area is small, hence, the pressure is higher (smaller
> area, same force -> higher pressure)?

It's the shear stresses of the wiggling.

>
> But, that makes no sense either. The purpose of this is to spread the
> compound, which will increase the contact area. Larger area, and we are
> back to the low pressure problem!
>
> I don't mean to get into a senseless debate here. I would like to hear
> opinions to help me understand. Thanks again.
>
> Navid
>
>
Anonymous
a b K Overclocking
June 19, 2004 1:23:16 AM

Archived from groups: alt.comp.hardware.overclocking (More info?)

Well, why not use a different thermal compound, and not worry about the
Arctic Silver instructions? The application of thermal compound, the
condition of the surfaces, the parallelness of the two surfaces have a
greater affect on cooling than the thermal compound itself. For that
matter, unsalted butter is nearly as good as Arctic Silver or any other
thermal compound for heat transfer (except that it eventually goes rancid,
attracts ants, and leaks away.)

--
Phil Weldon, pweldonatmindjumpdotcom
For communication,
replace "at" with the 'at sign'
replace "mindjump" with "mindspring."
replace "dot" with "."

"Navid" <nospam@nospam.invalid> wrote in message
news:mHAAc.57119$2o2.3053324@twister.southeast.rr.com...
>
> "P2B" <p2b@sympatico.ca> wrote in message
> news:1GuAc.50665$7H1.1718287@news20.bellglobal.com...
> >
> >
> > Navid wrote:
> >
> >> I have read posts that suggest minimal thermal compound. The Arctic
> >> Silver
> >> instructions also suggest a very small amount.
> >> http://www.arcticsilver.com/arctic_silver_instructions....
> >>
> >> I understand this advice if the concern is that too much compound may
> >> contaminate the pins and cause electrical problems. But, I don't
> >> understand
> >> it if the concern is heat transfer.
> >> I realize that the objective is to only fill the microscopic gaps
between
> >> the two surfaces.
> >>
> >> After the heat sink is mounted, the force pushing the two surfaces
> >> together
> >> is so high that any extra compound will be squeezed out of the sides.
> >> The
> >> compound is a fluid after all. There will be no compound between the
> >> peaks
> >> (only in the valleys). Is that not true?
> >> Am I missing something?
> >>
> >> Thanks,
> >>
> >> Navid
> >
> > I've been following this thread and the responses have been long on
> > (mostly accurate) theory but short on practice. You can *feel* when
you've
> > got it right!
> >
> > - apply a dob of compound to the CPU core or heatspreader, spread it
thin
> > with a blade (I prefer a boxcutter blade, but a credit card works),
remove
> > any excess on non-contact areas using a cotton bud
> > - place the heatsink on top, apply medium pressure and work it in a
> > circular motion until you can feel the peaks grating against each other
> > - remove heatsink to inspect coverage, it will be slightly difficult to
> > unstick if you have it right, much easier if you have too little or too
> > much compound
> > - again remove any excess on non-contact areas with the cotton bud, and
> > use it to add a little compound to any bare contact areas
> > - repeat until you feel grating but find no excess or bare contact areas
> > upon heaksink removal
> > - work heaksink in a circular motion again briefly before final clamping
> >
> > P2B
>
>
> Thanks for everyone's replies.
>
> The main point that I thought was made was that the pressure between the
two
> surfaces is not enough to spread the compound thin enough. That is why we
> should apply a very small amount.
>
> There is still a disconnect, in my understanding, between this and the
> Arctic Silver 5 instructions for Pentium 4. The instruction is to place a
> small amount of arctic silver 5 "without spreading it" and place the sink
> with a wiggle and mount.
>
> Now, if the pressure is not enough to spread the compound, then, how is
this
> technique going to make the peaks touch? Is it because the compound
amount
> is small, the contact area is small, hence, the pressure is higher
(smaller
> area, same force -> higher pressure)?
>
> But, that makes no sense either. The purpose of this is to spread the
> compound, which will increase the contact area. Larger area, and we are
> back to the low pressure problem!
>
> I don't mean to get into a senseless debate here. I would like to hear
> opinions to help me understand. Thanks again.
>
> Navid
>
>
June 22, 2004 9:06:12 PM

Archived from groups: alt.comp.hardware.overclocking (More info?)

Also of note is some thermal compounds like Ceramique- which I pretty
much use exclusively for system building- is applied differently.
You first smear a small amoutn, then buff on and then off the
heatsink. This leaves an extremely thin layer. then slop on the
BB-size dot (or line on some of the CPUs, like the AMD mobiles, as
they have a longer die) and G-E-N-T-L-Y apply the heatsink. This is
where CPU shims make a difference. For serious overclocking and
water/pelt/phase change, I really encourage a shim. It prevents
shearing of the thermal goop and ensures PRECISE alignment.

Shin-etsu is another great compound and will outperform the Arctic
Silver silver-based compounds.

Having said that, I also have a 12ounce tube (only about 8 ounces now)
of DOW thermal compound I use for work. It also works very well, and
the tube is at least 15 years old.

The bottom line is to apply a very thin even layer and then stick on
your heatsink evenly OR apply as per the manufacturer's instructions.

==============
Posted through www.HowToFixComputers.com/bb - free access to hardware troubleshooting newsgroups.
!