Is there such a thing as Too Cold???

[BorisDaSpider]

The reason I ask is I just added a 85 Watt Pelitier to my home made water cooled system,And Im at -40C Right now just cruzing the internet. My system is Way overclocked(PIII700e@1050)
But WOW is this thing running cold under load it gets to -5C is this good??too cold??

[ElectricMan]

I don't think you could ever be too cold, I just added a 85 watt TEC to my water system also, I learned how fast condensation form's . that is all I would worry about,, COLD GOOD , WATER VERY BAD!

[BorisDaSpider]

I know what ya mean,, I did Plenty of homework on it I insulated the crap out of it,, I can still smell the silicone sealant on my gasket!!
and by the way if you do get silicone in your slot1 its a real pain in the ass to get out!

[dasouthernocer]

Well there is a sucha thing as too cold but you're not gonnna get there with just a peltier :). 0K would stop the electrons from orbiting the prontons and neutrons and that wouldn't be good :).

-dasouthernocer

[schmidty]

First - you don't have to worry about getting to 0 ºK since the coldest it gets anywhere is about 2-3 ºK - so the electrons won't stop. :-) And if it was to get that cold - there would be major problems prior to the electrons stopping (like the materials shrinking and connections breaking - thermal expansion / contraction)

I think the only thing that i would worry about (in getting too cold - not counting H20) would be a point that the substrate (or any other parts that the die is on, heck even the die its self) becomes brittle.

I'm not a materials chemist, so I can't recite any temps on where abouts this might happen. But I'm going to hegde a six pack, that a pelt won't get that low.

cheers

[BorisDaSpider]

LOL Ok Ill keep a eye on my electorns,if they stop moving ill reboot
Thanks

[qpetershen]

semiconductor performance increases as temperature approaches absolute zero, that is -273ºC or 0 Kelvin, this is due to the fact that there is less lattice scattering/disruption due to presence of thermal energy.. i.e. electrons move much faster...so your aim is to get your processor as cool as possible.. i know some ppl here use LN2 to get their processors down to -182ºC.. that's as cold as you can practically get your processor to run without using exotic cryogenic temperatures

as a rough rule of thumb.. figure a 5% speed increase for every 10-20ºC drop in processor temperature.. i've seen this in my school's labs... this means a 933mhz p3@ 0ºC will be THEORETICALLY faster than a similar 933mhz p3@ 35ºC.. plus the processor at 0ºC has more o/c headroom... IMHO the difference is more pronounced the lower u go below 0ºC .. so the moral?.. u can never be too cold =)

p.s. kurt is right about the brittleness part.. if you're running below 0ºC make sure u dont suddenly lose your fans/waterpump u can crack the die with the sudden heat rush

[Big Lar]

When I used to run 2 Ice 71 Pelts on my P3650 ( back when I had that cpu), I got the core down to -48c, problem was with the insulation, The cpu actually froze on the backside! I re-insulated, and my next dilema was the Slot1 contacts freezing..... That was back before I used Dielectric grease.... So, in a way, yes there is such a thing as too cold...... I now run my P3 800 at 1066, at -5c or so, and can adjust the temp of the cpu, by how cold I run my fluid mixture thru the block...... I've had it down to -23c, but I didn't see any significant speed increase, hence I run at -5c I forgot to mention that my temps now are with 1 Ice 71...