motherboard monitor

G

Guest

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

Hi,
I was just browsing the Motherboard Monitor 5 settings, when I spotted the
following crazy option: 'low temperature warning'.
This sounds weird to me, I always thought 'the colder the better!' Can
anyone explain this???
 
G

Guest

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

0ne_Up wrote:
> Hi,
> I was just browsing the Motherboard Monitor 5 settings, when I
> spotted the following crazy option: 'low temperature warning'.
> This sounds weird to me, I always thought 'the colder the better!' Can
> anyone explain this???

IIRC, if you get the chip really cold (-50 deg C or so) the conductivity of
the die (though not the interconnet of course) has dropped down
significantly compared to when it's at room temperature. This results in
instability as the transistors can no longer switch fast enough. The problem
first occurs in rarely used parts of the chips (which don't produce so much
heat, so are cooler than the "die" temperature). People running recharged
phase-change systems often have a "minimum" temperature set on the
compressor control, and have reported instability if the chips get too cold
(though more voltage usually fixes things :) ).

A bigger problem is due to mechanical issues. The colder you get the CPU,
the more brittle it gets, and hence the more susceptible to hairline cracks
from vibration (from compressors, etc). This especially applies to solder
joints, though these are fairly minimal on the CPU itself. This obviously
isn't an immediate effect, but it does shorten the lifespan of the device
unless significant care is taken to avoid vibrations. Also, the chip is made
of many different materials each with different temperature coefficients of
expansion. So as the chip is cooled, different parts of the chip change size
at different rates, resulting in increased mechanical stress on the chip. I
don't think this latter one is very bad, though I'm open to corrections :)

Finally, you get problems with things around the CPU if you run it cold
enough. I'm not sure what motherboard manufacturers use nowadays, but most
electrolytic capacitors (the big ones) used to have water-based dielectrics
in them. Cool them down sufficiently and the dielectrics would freeze, with
a similar effect to leaving a bottle of water in the freezer too long. This
may not apply to modern boards. Finally, if your board is getting quite
cold, then you increase the risk of solder failure as mentioned above.

--
Michael Brown
www.emboss.co.nz : OOS/RSI software and more :)
Add michael@ to emboss.co.nz - My inbox is always open
 
G

Guest

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

And someone would cool his system to -200c with Liquid Nitrogen...
"Michael Brown" <see@signature.below> schreef in bericht
news:eek:rqad.12164$JQ4.761871@news.xtra.co.nz...
> 0ne_Up wrote:
>> Hi,
>> I was just browsing the Motherboard Monitor 5 settings, when I
>> spotted the following crazy option: 'low temperature warning'.
>> This sounds weird to me, I always thought 'the colder the better!' Can
>> anyone explain this???
>
> IIRC, if you get the chip really cold (-50 deg C or so) the conductivity
> of
> the die (though not the interconnet of course) has dropped down
> significantly compared to when it's at room temperature. This results in
> instability as the transistors can no longer switch fast enough. The
> problem
> first occurs in rarely used parts of the chips (which don't produce so
> much
> heat, so are cooler than the "die" temperature). People running recharged
> phase-change systems often have a "minimum" temperature set on the
> compressor control, and have reported instability if the chips get too
> cold
> (though more voltage usually fixes things :) ).
>
> A bigger problem is due to mechanical issues. The colder you get the CPU,
> the more brittle it gets, and hence the more susceptible to hairline
> cracks
> from vibration (from compressors, etc). This especially applies to solder
> joints, though these are fairly minimal on the CPU itself. This obviously
> isn't an immediate effect, but it does shorten the lifespan of the device
> unless significant care is taken to avoid vibrations. Also, the chip is
> made
> of many different materials each with different temperature coefficients
> of
> expansion. So as the chip is cooled, different parts of the chip change
> size
> at different rates, resulting in increased mechanical stress on the chip.
> I
> don't think this latter one is very bad, though I'm open to corrections :)
>
> Finally, you get problems with things around the CPU if you run it cold
> enough. I'm not sure what motherboard manufacturers use nowadays, but most
> electrolytic capacitors (the big ones) used to have water-based
> dielectrics
> in them. Cool them down sufficiently and the dielectrics would freeze,
> with
> a similar effect to leaving a bottle of water in the freezer too long.
> This
> may not apply to modern boards. Finally, if your board is getting quite
> cold, then you increase the risk of solder failure as mentioned above.
>
> --
> Michael Brown
> www.emboss.co.nz : OOS/RSI software and more :)
> Add michael@ to emboss.co.nz - My inbox is always open
>
>
 
G

Guest

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

0ne_Up wrote:
> And someone would cool his system to -200c with Liquid Nitrogen...

I've never seen someone who actually did it with direct-on-die liquid
nitrogen, and I wouldn't be surprised if the thermal shock of pouring liquid
nitrogen onto the die would shatter it. I've seen a non-pyrex glass beaker
shattered by liquid nitrogen (as a demonstation that you should check before
you pour), so it's certainly possible. All the LN2 articles I've seen use a
copper plate (or more often even a copper heatsink), so the die itself
doesn't get anywhere near -200 deg C. It's also mitigated by running very
high voltages as I mentioned, which most of these sort of people do.

[...]

--
Michael Brown
www.emboss.co.nz : OOS/RSI software and more :)
Add michael@ to emboss.co.nz - My inbox is always open
 
G

Guest

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

you're right, all guys use a copper plate, but still the actual CPU
temperature can be -140.
some people even achieved 200% overclocks with ln2...
"Michael Brown" <see@signature.below> schreef in bericht
news:Azuad.12280$JQ4.765975@news.xtra.co.nz...
> 0ne_Up wrote:
>> And someone would cool his system to -200c with Liquid Nitrogen...
>
> I've never seen someone who actually did it with direct-on-die liquid
> nitrogen, and I wouldn't be surprised if the thermal shock of pouring
> liquid
> nitrogen onto the die would shatter it. I've seen a non-pyrex glass beaker
> shattered by liquid nitrogen (as a demonstation that you should check
> before
> you pour), so it's certainly possible. All the LN2 articles I've seen use
> a
> copper plate (or more often even a copper heatsink), so the die itself
> doesn't get anywhere near -200 deg C. It's also mitigated by running very
> high voltages as I mentioned, which most of these sort of people do.
>
> [...]
>
> --
> Michael Brown
> www.emboss.co.nz : OOS/RSI software and more :)
> Add michael@ to emboss.co.nz - My inbox is always open
>
>
 

Clob

Distinguished
Sep 7, 2003
1,317
0
19,280
Archived from groups: alt.comp.hardware.overclocking (More info?)

How long would the liquid nitrogen last before it has to be replinished?


"0ne_Up" <0neup@tiscali.nl> wrote in message
news:416a788c$0$44072$5fc3050@dreader2.news.tiscali.nl...
> you're right, all guys use a copper plate, but still the actual CPU
> temperature can be -140.
> some people even achieved 200% overclocks with ln2...
> "Michael Brown" <see@signature.below> schreef in bericht
> news:Azuad.12280$JQ4.765975@news.xtra.co.nz...
>> 0ne_Up wrote:
>>> And someone would cool his system to -200c with Liquid Nitrogen...
>>
>> I've never seen someone who actually did it with direct-on-die liquid
>> nitrogen, and I wouldn't be surprised if the thermal shock of pouring
>> liquid
>> nitrogen onto the die would shatter it. I've seen a non-pyrex glass
>> beaker
>> shattered by liquid nitrogen (as a demonstation that you should check
>> before
>> you pour), so it's certainly possible. All the LN2 articles I've seen use
>> a
>> copper plate (or more often even a copper heatsink), so the die itself
>> doesn't get anywhere near -200 deg C. It's also mitigated by running very
>> high voltages as I mentioned, which most of these sort of people do.
>>
>> [...]
>>
>> --
>> Michael Brown
>> www.emboss.co.nz : OOS/RSI software and more :)
>> Add michael@ to emboss.co.nz - My inbox is always open
>>
>>
>
>
 
G

Guest

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

Clob wrote:
> How long would the liquid nitrogen last before it has to be
> replinished?

The latent heat of vaporisation of nitrogen is ~200 kJ/kg, so if your CPU is
pumping out 100W then you'd need ~0.5g/sec of liquid nitrogen. However, due
to losses from the top of the liquid, heat absorbed through the sides of the
container, and the probability that the CPU would be putting out more than
100W, I'd put it at closer to 1 gram a second for the typical liquid
nitrogen cooled setup. Density of liquid nitrogen is ~0.8g/mL so you'd go
through about 1.25 mL/s, or about 75mL per minute, give or take a large
margin of error :)

[...]

--
Michael Brown
www.emboss.co.nz : OOS/RSI software and more :)
Add michael@ to emboss.co.nz - My inbox is always open