Thermal Cement - where to buy

Archived from groups: alt.comp.hardware.pc-homebuilt (More info?)

"Mike Smith" <mike.smith@doesnotwork.com> wrote in message
news:4lmlc11okt0lr7ean48virj5kdd914aqpj@4ax.com...
> Howdy folks
>
> One of the fans on my vertically mounted motherboard has come unstuck,
> and is plonked on the reverse of the graphics board. Fortunately no
> major damage, but I need to glue/cement it back in place.
>
> I've looked around online and found plenty of thermal paste, but I
> need something to hold it in place, as well as conduct. Any
> recommended sources/products? (Ideally UK)
>
> The fan (someone who knows more than me) is evidetly there to help
> cool the onboard motherboard chips ie its not a processor fan, and its
> not just there for general cooling.
>
> Any help appreciated?
>

I'd use a dab of epoxy. Some mobos use a double sided adhesive film. SVC.com
and others were/are selling an Arctic Silver branded epoxy specifically for
attaching heat sinks.

Archived from groups: alt.comp.hardware.pc-homebuilt (More info?)

Mike Smith wrote:

> One of the fans on my vertically mounted motherboard
> has come unstuck, and is plonked on the reverse of
> the graphics board. Fortunately no major damage, but
> I need to glue/cement it back in place.

> The fan (someone who knows more than me) is evidetly
> there to help cool the onboard motherboard chips ie
> its not a processor fan

It would be highly unusual for just the fan to come off because every
such fan I've seen was attached very securely to its heatsink with
clips or screws. But if you mean a heatsink came off, then you can
simply use regular epoxy or silicone rubber glue (RTV) since the chip
you're referring to isn't one of the hotter ones. Just clean off any
existing glue or tape very thoroughly from the mating surfaces, and
press and swirl the heatsink to squeeze out excess glue. Let the glue
cure sufficiently before putting the device back into use; that would
be 30-60 minutes for RTV or regular epoxy, 1-5 minutes for fast cure
epoxy (must be mixed much more carefully or may never cure). I prefer
RTV since it can be sliced through if you ever decide to remove the
heatsink later.

Ignore any bad advice about applying thermal paste to the middle of the
chip and using a drop of super glue or epoxy in each corner of the chip
package. Also ignore hysterical types who insist you need to use a
special thermal glue since that's true only for CPUs and some video
chips.

Archived from groups: alt.comp.hardware.pc-homebuilt (More info?)

"Mike Smith" <mike.smith@doesnotwork.com> wrote in message news:4lmlc11okt0lr7ean48virj5kdd914aqpj@4ax.com...
> Howdy folks
>
> One of the fans on my vertically mounted motherboard has come unstuck,
> and is plonked on the reverse of the graphics board. Fortunately no
> major damage, but I need to glue/cement it back in place.
>
> I've looked around online and found plenty of thermal paste, but I
> need something to hold it in place, as well as conduct. Any
> recommended sources/products? (Ideally UK)

http://www.arcticsilver.com/arctic_silver_thermal_adhes...

-- Bob Day
http://bobday.vze.com

Archived from groups: alt.comp.hardware.pc-homebuilt (More info?)

On 5 Jul 2005 16:10:36 -0700, do_not_spam_me@my-deja.com wrote:

>Mike Smith wrote:
>
>> One of the fans on my vertically mounted motherboard
>> has come unstuck, and is plonked on the reverse of
>> the graphics board. Fortunately no major damage, but
>> I need to glue/cement it back in place.
>
>> The fan (someone who knows more than me) is evidetly
>> there to help cool the onboard motherboard chips ie
>> its not a processor fan
>
>It would be highly unusual for just the fan to come off because every
>such fan I've seen was attached very securely to its heatsink with
>clips or screws. But if you mean a heatsink came off, then you can
>simply use regular epoxy or silicone rubber glue (RTV) since the chip
>you're referring to isn't one of the hotter ones. Just clean off any
>existing glue or tape very thoroughly from the mating surfaces, and
>press and swirl the heatsink to squeeze out excess glue. Let the glue
>cure sufficiently before putting the device back into use; that would
>be 30-60 minutes for RTV or regular epoxy, 1-5 minutes for fast cure
>epoxy (must be mixed much more carefully or may never cure). I prefer
>RTV since it can be sliced through if you ever decide to remove the
>heatsink later.
>
>Ignore any bad advice about applying thermal paste to the middle of the
>chip and using a drop of super glue or epoxy in each corner of the chip
>package. Also ignore hysterical types who insist you need to use a
>special thermal glue since that's true only for CPUs and some video
>chips.


Thanks for this

Just to clarify - this pic shows the fan dangling
http://www.bestsf.net/images/100_1432.jpg


and this shows where it had fallen from (the green square above the
graphics card)
http://www.bestsf.net/images/100_1433.jpg


So there were never any screws, but just the two 'hooks' thingies

Archived from groups: alt.comp.hardware.pc-homebuilt (More info?)

On Wed, 06 Jul 2005 08:44:09 +0100, Mike Smith
<mike.smith@doesnotwork.com> had a flock of green cheek conures squawk
out:

>On 5 Jul 2005 16:10:36 -0700, do_not_spam_me@my-deja.com wrote:
>
>>Mike Smith wrote:
>>
>>> One of the fans on my vertically mounted motherboard
>>> has come unstuck, and is plonked on the reverse of
>>> the graphics board. Fortunately no major damage, but
>>> I need to glue/cement it back in place.
>>
>>> The fan (someone who knows more than me) is evidetly
>>> there to help cool the onboard motherboard chips ie
>>> its not a processor fan
>>
>>It would be highly unusual for just the fan to come off because every
>>such fan I've seen was attached very securely to its heatsink with
>>clips or screws. But if you mean a heatsink came off, then you can
>>simply use regular epoxy or silicone rubber glue (RTV) since the chip
>>you're referring to isn't one of the hotter ones. Just clean off any
>>existing glue or tape very thoroughly from the mating surfaces, and
>>press and swirl the heatsink to squeeze out excess glue. Let the glue
>>cure sufficiently before putting the device back into use; that would
>>be 30-60 minutes for RTV or regular epoxy, 1-5 minutes for fast cure
>>epoxy (must be mixed much more carefully or may never cure). I prefer
>>RTV since it can be sliced through if you ever decide to remove the
>>heatsink later.
>>
>>Ignore any bad advice about applying thermal paste to the middle of the
>>chip and using a drop of super glue or epoxy in each corner of the chip
>>package. Also ignore hysterical types who insist you need to use a
>>special thermal glue since that's true only for CPUs and some video
>>chips.
>
>
>Thanks for this
>
>Just to clarify - this pic shows the fan dangling
>http://www.bestsf.net/images/100_1432.jpg
>
>
>and this shows where it had fallen from (the green square above the
>graphics card)
>http://www.bestsf.net/images/100_1433.jpg
>
>
>So there were never any screws, but just the two 'hooks' thingies
>
>
>

Use the wire clip that's still on the fan and clip the ends onto two
opposing wire loops on the motherboard. If one of the loops detached
from the motherboard, just re-solder it back in before reinstalling
the fan.

Stephen


--

Archived from groups: alt.comp.hardware.pc-homebuilt (More info?)

Matt wrote:
> do_not_spam_me@my-deja.com wrote:

> > Ignore any bad advice about applying thermal paste to
> > the middle of the chip and using a drop of super glue
> > or epoxy in each corner of the chip package.

> Hmmm ... that doesn't sound like such a bad idea. Why is it bad?

You don't mix oil and glue, except acrylic glue. Check any
manufacturer's directions.

On the other hand chip temperatures are unlikely to be a problem even
for regular super glue. Still I'd rather use silicone rubber or epoxy
glue.

Archived from groups: alt.comp.hardware.pc-homebuilt (More info?)

Matt wrote:

> I would think that using ordinary epoxy or silicone rubber
> would be worse than simply leaving the chip bare with no
> fan. O' course, I don't know everything.

Why do you think that, and what tests have you run or seen? Epoxy or
RTV would have to be very, very thick to make the chip run hotter than
with no fan or heatsink, and any realistic thickness of of either
material between the chip package and heatsink will improve the
cooling, simply because any solid or liquid conducts heat far, far
better than air does.

Archived from groups: alt.comp.hardware.pc-homebuilt (More info?)

Mike Smith wrote:

> Just to clarify - this pic shows the fan dangling
> http://www.bestsf.net/images/100_1432.jpg
>
> and this shows where it had fallen from (the green square above the
> graphics card)
> http://www.bestsf.net/images/100_1433.jpg

> So there were never any screws, but just the two 'hooks' thingies

I suggest you epoxy the heatsink back to the fan and also get new
mechanical attachments. Electronics parts dealers and sometimes even
hardware stores (True Value, etc., not Lowe's or Home Depot) have such
nylon clips, or you should be able to use nylon nuts and bolts.

Archived from groups: alt.comp.hardware.pc-homebuilt (More info?)

Matt wrote:
> do_not_spam_me@my-deja.com wrote:

>I would think that using ordinary epoxy or silicone rubber
>would be worse than simply leaving the chip bare with no
>fan. O' course, I don't know everything.


> > Why do you think that, and what tests have you run or seen?
> > Epoxy or RTV would have to be very, very thick to make the
> > chip run hotter than with no fan or heatsink, and any
> > realistic thickness of of either material between the chip
> > package and heatsink will improve the cooling, simply
> > because any solid or liquid conducts heat far, far better
> > than air does.
>
> Not saying you're wrong ... but the air is moving, so that the air in
> contact with the surface of the chip is never heated above the ambient
> case temperature.

Then how do you explain the air inside the case being hotter than room
air since it also moves? By your line of reasoning, a chip package
exposed directly to ambient air should stay cooler than one in good
contact with a heatsink.

Again, I ask: What tests have you run or seen?

Archived from groups: alt.comp.hardware.pc-homebuilt (More info?)

Matt wrote:

> do_not_spam_me@my-deja.com wrote:
>
>>
>> Matt wrote:
>>
>>
>>> I would think that using ordinary epoxy or silicone rubber
>>> would be worse than simply leaving the chip bare with no
>>> fan. O' course, I don't know everything.
>>
>>
>>
>> Why do you think that, and what tests have you run or seen? Epoxy or
>> RTV would have to be very, very thick to make the chip run hotter than
>> with no fan or heatsink, and any realistic thickness of of either
>> material between the chip package and heatsink will improve the
>> cooling, simply because any solid or liquid conducts heat far, far
>> better than air does.
>
>
> Not saying you're wrong ... but the air is moving, so that the air in
> contact with the surface of the chip is never heated above the ambient
> case temperature.

Heat flows from hot to cold so if the ambient is 'warmer' than the chip
it's going to be heating the IC. Clearly, this is not the reason one runs
air over it.

In fact, the chip will always be warmer than the ambient case air since it
must be warmer to cause heat to flow (from hot to cold) and when the heat
flows it warms the immediate air which will then be warmer than the case
ambient from whence it came before being heated by the warmer IC.

> On the other hand, I haven't seen a northbridge
> without a heatsink ...
>

Archived from groups: alt.comp.hardware.pc-homebuilt (More info?)

Matt wrote:
> do_not_spam_me@my-deja.com wrote:

>>Epoxy or RTV would have to be very, very thick to make the
>>chip run hotter than with no fan or heatsink, and any
>>realistic thickness of of either material between the chip
>>package and heatsink will improve the cooling, simply
>>because any solid or liquid conducts heat far, far better
>>than air does.

>>By your line of reasoning, a chip package exposed directly
>>to ambient air should stay cooler than one in good contact
>>with a heatsink.

> I was a little careless in my statement. Yes, the air
> right at the surface is hotter than the average case air
> temperature. My point should have been that the bare
> package is cooled by convection, so that there is a flow
> of the cooling material (air). My intuition was that
> such convection would be more effective than conduction
> through a glass-like or rubber-like solid.

Again, that's correct for thick layers of those solids but not for the
thin layers used between chip packages and heatsinks.

Just measuring chip or chip package temperature against ambient air
temperature is a good way to test a heatsink or the material used
between them.