So what if thermal compound spreads?

Toggle sidebar Toggle sidebar
C

CrackerJack

Distinguished
Jul 1, 2003
17
0
18,510
Archived from groups: alt.comp.hardware.overclocking.amd,alt.comp.hardware,alt.comp.hardware.pc-homebuilt,uk.comp.homebuilt (More info?)

What exactly is the problem if too much cpu compound is put on the
core and it gets squashed out onto the surrounding area?

Apart from looking messy, is there any real problem with this?
 
Archived from groups: alt.comp.hardware.overclocking.amd,alt.comp.hardware,alt.comp.hardware.pc-homebuilt,uk.comp.homebuilt (More info?)

CrackerJack wrote:

> What exactly is the problem if too much cpu compound is put on the
> core and it gets squashed out onto the surrounding area?
>
> Apart from looking messy, is there any real problem with this?


If you put on so much that the compound squashes out past the edges,
that's too much. It generally won't do any damage, but will be nowhere
near as efficient as if you had used less.


-WD
 
Archived from groups: alt.comp.hardware.overclocking.amd,alt.comp.hardware,alt.comp.hardware.pc-homebuilt,uk.comp.homebuilt (More info?)

Depends, depends on how much and the CPU. On the Intel, much less of a
problem (as long as it's not creeping around on to the underside!). On the
AMD, much more problematic. The AMD core is *exposed*, and thus there are
various "bridges" that surround it. The last thing you want to do is
"short" one of those bridges because you got to sloppy w/ the thermal paste!

But having too much paste indicates another problem. The paste is ONLY
there to fill the microscopic imperfections between the mating surfaces. If
those surfaces were perfect, you wouldn't even need the paste, in fact, it
would *hinder* heat transfer. If you have so much paste on the mating
surfaces that installation causes much of it to squeeze out, it indicates
you have too much paste! Think of it this way, if we could use the paste to
fill ONLY the imperfections, that would be ideal. Every bit of paste that
*interferes* with contact between the CPU and heatsink is working NEGATIVELY
against heat trasnfer. We're only interested in preventing VOIDS between
the mating surfaces. Anything that's NOT filling the voids and is actually
*preventing* surface to surface contact between the CPU and heatsink and
thus *hindering* heat transfer, not helping.

Bottomline: The less the better, ideally zero, but since this is an
imperfect world, we need some, so use as little as possible.

HTH

Jim


"CrackerJack" <binaryblobNOTTHISBIT@hotpop.com> wrote in message
news:94C89175463C6AD265@64.62.191.93...
> What exactly is the problem if too much cpu compound is put on the
> core and it gets squashed out onto the surrounding area?
>
> Apart from looking messy, is there any real problem with this?
 
Archived from groups: alt.comp.hardware.overclocking.amd,alt.comp.hardware,alt.comp.hardware.pc-homebuilt,uk.comp.homebuilt (More info?)

In article <94C89175463C6AD265@64.62.191.93>, "CrackerJack"
binaryblobNOTTHISBIT@hotpop.com says...
> What exactly is the problem if too much cpu compound is put on the
> core and it gets squashed out onto the surrounding area?
>
> Apart from looking messy, is there any real problem with this?
>
Only if it's electrically conductive, or if there's so much that it
actually holds the heatsink off the CPU. But why not just put a
reasonable amount on in the first place?
 
Archived from groups: alt.comp.hardware.overclocking.amd,alt.comp.hardware,alt.comp.hardware.pc-homebuilt,uk.comp.homebuilt (More info?)

On Sat, 10 Apr 2004 23:51:38 GMT, Will Dormann
<wdormann@yahoo.com.invalid> wrote:

>CrackerJack wrote:
>
>> What exactly is the problem if too much cpu compound is put on the
>> core and it gets squashed out onto the surrounding area?
>>
>> Apart from looking messy, is there any real problem with this?
>
>
>If you put on so much that the compound squashes out past the edges,
>that's too much. It generally won't do any damage, but will be nowhere
>near as efficient as if you had used less.
>

Unless you start out with an absolutely horrendous heatsink and don't
apply enough compound (it is practically impossible to apply EXACTLY the
amount needed) it will always squish out to a certain extent. Take off a
heatsink and look at the base, you see the outline of the CPU core due to
the compound squishing out. Even the thinnest layer of compound possible
should squish out a little if the heatsink is properly finished.
 
Archived from groups: alt.comp.hardware.overclocking.amd (More info?)

On Sun, 11 Apr 2004 02:04:43 GMT, kony <spam@spam.com> wrote:


>Unless you start out with an absolutely horrendous heatsink and don't
>apply enough compound (it is practically impossible to apply EXACTLY the
>amount needed) it will always squish out to a certain extent. Take off a
>heatsink and look at the base, you see the outline of the CPU core due to
>the compound squishing out. Even the thinnest layer of compound possible
>should squish out a little if the heatsink is properly finished.

So put some on and wipe it off. What's left is probably more than
enough. What part of "just a little" don't you understand ?
 
Archived from groups: alt.comp.hardware.overclocking.amd,alt.comp.hardware,alt.comp.hardware.pc-homebuilt,uk.comp.homebuilt (More info?)

Conor <conor_turton@hotmail.com> wrote in message news:<MPG.1ae2ae31b90d7b0a98a2a5@news.claranews.com>...
> In article <94C89175463C6AD265@64.62.191.93>,
> binaryblobNOTTHISBIT@hotpop.com says...
> > What exactly is the problem if too much cpu compound is put on the
> > core and it gets squashed out onto the surrounding area?
> >
> > Apart from looking messy, is there any real problem with this?
> >
> Yes as it acts as an insulator. The idea of thermal compound is to fill
> the very small scratches on the faces.

Actually the thermal paste is non-conductive. For maximum heat
transfer from the core to the heatsink, the two metal pieces need to
touch as much as possible. Hence, the addition on the paste fills in
the imperfections.

Other than being a mess, things should work just fine.
 
Archived from groups: alt.comp.hardware.overclocking.amd,alt.comp.hardware,alt.comp.hardware.pc-homebuilt,uk.comp.homebuilt (More info?)

No problem at all .... if you don't mind higher temps !
You need to maximise metal-metal contact.
With a mirror-shine finish (above 1200 grit), you only need invisible
amounts of goop.
Enough to take off the shine, only.

"CrackerJack" <binaryblobNOTTHISBIT@hotpop.com> wrote in message
news:94C89175463C6AD265@64.62.191.93...
> What exactly is the problem if too much cpu compound is put on the
> core and it gets squashed out onto the surrounding area?
>
> Apart from looking messy, is there any real problem with this?
 
Archived from groups: alt.comp.hardware.overclocking.amd,alt.comp.hardware,alt.comp.hardware.pc-homebuilt,uk.comp.homebuilt (More info?)

You don't say why it's inefficient !
Do you mean wasteful or something else constituting inefficient.
A vague and meaningless answer.


"Will Dormann" <wdormann@yahoo.com.invalid> wrote in message
news:eM%dc.2804$T86.1685@fe2.columbus.rr.com...
> CrackerJack wrote:
>
> > What exactly is the problem if too much cpu compound is put on the
> > core and it gets squashed out onto the surrounding area?
> >
> > Apart from looking messy, is there any real problem with this?
>
>
> If you put on so much that the compound squashes out past the edges,
> that's too much. It generally won't do any damage, but will be nowhere
> near as efficient as if you had used less.
>
>
> -WD
 
Archived from groups: alt.comp.hardware.overclocking.amd (More info?)

<headbanger@wall.com> wrote in message
news:s1ph7093ei5rougrqpp2uvuv4u7jd8pamk@4ax.com...
> On Sun, 11 Apr 2004 02:04:43 GMT, kony <spam@spam.com> wrote:
>
>
> >Unless you start out with an absolutely horrendous heatsink and don't
> >apply enough compound (it is practically impossible to apply EXACTLY the
> >amount needed) it will always squish out to a certain extent. Take off a
> >heatsink and look at the base, you see the outline of the CPU core due to
> >the compound squishing out. Even the thinnest layer of compound possible
> >should squish out a little if the heatsink is properly finished.
>
> So put some on and wipe it off.


Or just put on as much as you need, to start with.
Duh !



What's left is probably more than
> enough. What part of "just a little" don't you understand ?
>
>
 
Archived from groups: alt.comp.hardware.overclocking.amd,alt.comp.hardware,alt.comp.hardware.pc-homebuilt,uk.comp.homebuilt (More info?)

In article <b24e73dd.0404110201.326cd75e@posting.google.com>, "John"
jkrytus@yahoo.com says...
> Conor <conor_turton@hotmail.com> wrote in message news:<MPG.1ae2ae31b90d7b0a98a2a5@news.claranews.com>...
> > In article <94C89175463C6AD265@64.62.191.93>,
> > binaryblobNOTTHISBIT@hotpop.com says...
> > > What exactly is the problem if too much cpu compound is put on the
> > > core and it gets squashed out onto the surrounding area?
> > >
> > > Apart from looking messy, is there any real problem with this?
> > >
> > Yes as it acts as an insulator. The idea of thermal compound is to fill
> > the very small scratches on the faces.
>
> Actually the thermal paste is non-conductive.

Eh?
 
Archived from groups: alt.comp.hardware.overclocking.amd,alt.comp.hardware,alt.comp.hardware.pc-homebuilt,uk.comp.homebuilt (More info?)

"Rob Morley" <nospam@ntlworld.com> wrote in message
news:MPG.1ae32ca437f9894c98988e@news.individual.net...
> In article <b24e73dd.0404110201.326cd75e@posting.google.com>, "John"
> jkrytus@yahoo.com says...
> > Conor <conor_turton@hotmail.com> wrote in message
news:<MPG.1ae2ae31b90d7b0a98a2a5@news.claranews.com>...
> > > In article <94C89175463C6AD265@64.62.191.93>,
> > > binaryblobNOTTHISBIT@hotpop.com says...
> > > > What exactly is the problem if too much cpu compound is put on the
> > > > core and it gets squashed out onto the surrounding area?
> > > >
> > > > Apart from looking messy, is there any real problem with this?
> > > >
> > > Yes as it acts as an insulator. The idea of thermal compound is to
fill
> > > the very small scratches on the faces.
> >
> > Actually the thermal paste is non-conductive.
>
> Eh?


You thought he meant thermally, we realised he meant electrically.



>
 
Archived from groups: alt.comp.hardware.overclocking.amd,alt.comp.hardware,alt.comp.hardware.pc-homebuilt,uk.comp.homebuilt (More info?)

In article <c5b83m$633$1@sparta.btinternet.com>, "QBall"
qball__@btinternet.com says...
>
> "Rob Morley" <nospam@ntlworld.com> wrote in message
> news:MPG.1ae32ca437f9894c98988e@news.individual.net...
> > In article <b24e73dd.0404110201.326cd75e@posting.google.com>, "John"
> > jkrytus@yahoo.com says...
> > > Conor <conor_turton@hotmail.com> wrote in message
> news:<MPG.1ae2ae31b90d7b0a98a2a5@news.claranews.com>...
> > > > In article <94C89175463C6AD265@64.62.191.93>,
> > > > binaryblobNOTTHISBIT@hotpop.com says...
> > > > > What exactly is the problem if too much cpu compound is put on the
> > > > > core and it gets squashed out onto the surrounding area?
> > > > >
> > > > > Apart from looking messy, is there any real problem with this?
> > > > >
> > > > Yes as it acts as an insulator. The idea of thermal compound is to
> fill
> > > > the very small scratches on the faces.
> > >
> > > Actually the thermal paste is non-conductive.
> >
> > Eh?
>
>
> You thought he meant thermally, we realised he meant electrically.
>
I realised he could have meant either, so asked for elaboration.
Actually some thermal compund is electrically conductive too, so I think
it was a dangerous statement to make.
 
Archived from groups: alt.comp.hardware.overclocking.amd,alt.comp.hardware,alt.comp.hardware.pc-homebuilt,uk.comp.homebuilt (More info?)

"Conor" <conor_turton@hotmail.com> wrote in message
news:MPG.1ae2ae31b90d7b0a98a2a5@news.claranews.com...
> In article <94C89175463C6AD265@64.62.191.93>,
> binaryblobNOTTHISBIT@hotpop.com says...
> > What exactly is the problem if too much cpu compound is put on the
> > core and it gets squashed out onto the surrounding area?
> >
> > Apart from looking messy, is there any real problem with this?
>
> Yes as it acts as an insulator.

But more conductive than air, which is why it is used.

Alex
 
Archived from groups: alt.comp.hardware.overclocking.amd,alt.comp.hardware,alt.comp.hardware.pc-homebuilt,uk.comp.homebuilt (More info?)

"QBall" <qball__@btinternet.com> wrote in message
news:c5b2hr$l9s$1@titan.btinternet.com...
> You don't say why it's inefficient !
> Do you mean wasteful or something else constituting inefficient.
> A vague and meaningless answer.

He means heat transfer will not be as good due to the large gap between the
core and the heatsink (I think large is a good word because much of this is
discuss'd on a "microscopic level"
 
Archived from groups: alt.comp.hardware.overclocking.amd,alt.comp.hardware,alt.comp.hardware.pc-homebuilt,uk.comp.homebuilt (More info?)

Rob Morley <nospam@ntlworld.com> wrote:

>> What exactly is the problem if too much cpu compound is put
>> on the core and it gets squashed out onto the surrounding
>> area?
>>
>> Apart from looking messy, is there any real problem with
>> this?
>>
> Only if it's electrically conductive, or if there's so much
> that it actually holds the heatsink off the CPU. But why not
> just put a reasonable amount on in the first place?


It is hard to judge the exact amount. Rather than put too little
on and run the risk of not conducting heat away from the core, I
usually err on the side of putting too much on.
 
Archived from groups: alt.comp.hardware.overclocking.amd,alt.comp.hardware,alt.comp.hardware.pc-homebuilt,uk.comp.homebuilt (More info?)

"rstlne" <.@text.news.virgin.net> wrote:

>> You don't say why it's inefficient !
>> Do you mean wasteful or something else constituting
>> inefficient. A vague and meaningless answer.
>
> He means heat transfer will not be as good due to the large
> gap between the core and the heatsink (I think large is a good
> word because much of this is discuss'd on a "microscopic
> level"


If the goo squishes out then presumably the layer of goo is no
thicker or thinner than if it didn't squish out?
 
Archived from groups: alt.comp.hardware.overclocking.amd,alt.comp.hardware,alt.comp.hardware.pc-homebuilt,uk.comp.homebuilt (More info?)

jkrytus@yahoo.com (John) wrote:

>> > What exactly is the problem if too much cpu compound is put
>> > on the core and it gets squashed out onto the surrounding
>> > area?
>> >
>> > Apart from looking messy, is there any real problem with
>> > this?
>> >
>> Yes as it acts as an insulator. The idea of thermal compound
>> is to fill the very small scratches on the faces.
>
> Actually the thermal paste is non-conductive. For maximum
> heat transfer from the core to the heatsink, the two metal
> pieces need to touch as much as possible. Hence, the addition
> on the paste fills in the imperfections.
>
> Other than being a mess, things should work just fine.


So something like a pinhead or two should be enough?

And a matchhead would be too much?
 
Archived from groups: alt.comp.hardware.overclocking.amd,alt.comp.hardware,alt.comp.hardware.pc-homebuilt,uk.comp.homebuilt (More info?)

Rob Morley <nospam@ntlworld.com> wrote:

>> > > > Yes as it acts as an insulator. The idea of thermal
>> > > > compound is to fill the very small scratches on the
>> > > > faces.
>> > >
>> > > Actually the thermal paste is non-conductive.
>> >
>> > Eh?
>>
>>
>> You thought he meant thermally, we realised he meant
>> electrically.
>>
> I realised he could have meant either, so asked for
> elaboration. Actually some thermal compund is electrically
> conductive too,


Which one?


> so I think it was a dangerous statement to
> make.
>
 
Archived from groups: alt.comp.hardware.overclocking.amd,alt.comp.hardware,alt.comp.hardware.pc-homebuilt,uk.comp.homebuilt (More info?)

In article <94C89EB51ADD861M2A@208.42.66.156>, "Max"
max.headroom@microsoft.com says...
> Rob Morley <nospam@ntlworld.com> wrote:
>
> >> > > > Yes as it acts as an insulator. The idea of thermal
> >> > > > compound is to fill the very small scratches on the
> >> > > > faces.
> >> > >
> >> > > Actually the thermal paste is non-conductive.
> >> >
> >> > Eh?
> >>
> >>
> >> You thought he meant thermally, we realised he meant
> >> electrically.
> >>
> > I realised he could have meant either, so asked for
> > elaboration. Actually some thermal compund is electrically
> > conductive too,
>
> Which one?
>
Any of the stuff that's filled with metallic particles is potentially
electrically conductive.
 
Archived from groups: alt.comp.hardware.overclocking.amd,alt.comp.hardware,alt.comp.hardware.pc-homebuilt,uk.comp.homebuilt (More info?)

In article <94C89DEF01E1731E75@130.133.1.4>, "Piotr Makley"
pmakley@mail.com says...
> Rob Morley <nospam@ntlworld.com> wrote:
>
> >> What exactly is the problem if too much cpu compound is put
> >> on the core and it gets squashed out onto the surrounding
> >> area?
> >>
> >> Apart from looking messy, is there any real problem with
> >> this?
> >>
> > Only if it's electrically conductive, or if there's so much
> > that it actually holds the heatsink off the CPU. But why not
> > just put a reasonable amount on in the first place?
>
>
> It is hard to judge the exact amount. Rather than put too little
> on and run the risk of not conducting heat away from the core, I
> usually err on the side of putting too much on.
>
If you're not sure then try with a very small amount, remove the
heatsink again and see how it's spread. Then clean and repeat. If you
don't clean it off in between applications there's more chance that you
will trap air in the compound.
 
Archived from groups: alt.comp.hardware.overclocking.amd,alt.comp.hardware,alt.comp.hardware.pc-homebuilt,uk.comp.homebuilt (More info?)

CrackerJack <binaryblobNOTTHISBIT@hotpop.com> wrote:

>What exactly is the problem if too much cpu compound is put on the
>core and it gets squashed out onto the surrounding area?

>Apart from looking messy, is there any real problem with this?

You'll always get a little but if you are squeezing a lot out it
indicates you've applied far too much, meaning the layer of paste
between the processor die/heatspreader and heatsink is too thick.

Remember, the idea is NOT to form a layer between the two surfaces.
The purpose of the paste is to fill the valleys in the contact
surfaces with something which is more efficient at transferring heat
away from the CPU core than the air which would otherwise fill the
gaps. Even the best thermal material is less efficient than direct
contact between the two metal surfaces.

Also, some thermal materials can be slightly electrically conductive,
so if leaks out and bridges gaps in electrical contacts it is possible
that it will result in reliability problems or even permanent damage
to components.

--
>iv< Paul >iv<
 
Archived from groups: alt.comp.hardware.overclocking.amd,alt.comp.hardware,alt.comp.hardware.pc-homebuilt,uk.comp.homebuilt (More info?)

"Rob Morley" <nospam@ntlworld.com> wrote in message
news:MPG.1ae36bb2c4360fac98989b@news.individual.net...
> In article <94C89DEF01E1731E75@130.133.1.4>, "Piotr Makley"
> pmakley@mail.com says...
> > Rob Morley <nospam@ntlworld.com> wrote:
> >
> > >> What exactly is the problem if too much cpu compound is put
> > >> on the core and it gets squashed out onto the surrounding
> > >> area?
> > >>
> > >> Apart from looking messy, is there any real problem with
> > >> this?
> > >>
> > > Only if it's electrically conductive, or if there's so much
> > > that it actually holds the heatsink off the CPU. But why not
> > > just put a reasonable amount on in the first place?
> >
> >
> > It is hard to judge the exact amount. Rather than put too little
> > on and run the risk of not conducting heat away from the core,


You clearly have no understanding of thermal transfer, which is maximised
via metal-metal contact.
You fall into the ingenue's trap of believing more is better, when it fact
less is more.


I
> > usually err on the side of putting too much on.
> >
> If you're not sure then try with a very small amount, remove the
> heatsink again and see how it's spread. Then clean and repeat. If you
> don't clean it off in between applications there's more chance that you
> will trap air in the compound.
 
Archived from groups: alt.comp.hardware.overclocking.amd,alt.comp.hardware,alt.comp.hardware.pc-homebuilt,uk.comp.homebuilt (More info?)

Max wrote:

> Which one?


I do know that artic silver 1 is slightly conductive, but the more worrying
thing is the capacitance if it gets too close to contacts.

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

--
Mark
Iligitimi Non Carborundum!
Twixt hill and high water, N.Wales, UK
 
Archived from groups: alt.comp.hardware.overclocking.amd,alt.comp.hardware,alt.comp.hardware.pc-homebuilt,uk.comp.homebuilt (More info?)

Matt <matt@themattfella.zzzz.com> wrote:

>>>>I realised he could have meant either, so asked for
>>>>elaboration. Actually some thermal compund is electrically
>>>>conductive too,

>> Any of the stuff that's filled with metallic particles is potentially
>> electrically conductive.

>Please name one.

Arctic Silver is capacitive, rather than conductive, but the same
applies in terms of application. Their instructions state:

"While much safer than electrically conductive silver and copper
greases, Arctic Silver should be kept away from electrical traces,
pins, and leads. While it is not electrically conductive, the compound
is slightly capacitive and could potentially cause problems if it
bridges two close-proximity electrical paths.)"

I wouldn't doubt there are other compounds around which are
electrically conductive.

--
>iv< Paul >iv<
 
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom