-
Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!
How cold can computers go.
- Thread starter B-Hizzle
- Start date
As cold as you can get them. During processor OC'ing competitions they use pure Liquid Hydrogen to chill the processors to below 0 temps.
Its not how much can the parts handle, its how low you can get them to go! I would highly doubt that there is something as "too cold" in terms of performance PC cooling.
Its not how much can the parts handle, its how low you can get them to go! I would highly doubt that there is something as "too cold" in terms of performance PC cooling.
- Joined
- Dec 13, 2002
- Location
- Southern California
I suspect things with moving parts, like mechanical hdds, would not like to go much below 0. Things like the cpu, ram, and graphics can usually go much much colder, but will sometimes coldbug, and then you have to warm them up to get them to post again
- Joined
- Nov 12, 2002
- Location
- Rootstown, OH
LN2 goes -196C - most phenom II AMD CPUs can go that cold, most 3770K's can too, many other CPUs can also go that cold. LHe goes -269C.
Many other CPUs have cold bugs however, some don't like to be colder than -20C or so. Many don't like to go more than about -100C.
What temperature different CPUs like varies. The same goes for GPUs. Some motherboards are cold sensitive as well, while some models (older ones) liked LN2 on the northbridge to run high FSB. Then memory, a lot of memory doesn't like to be too cold, not more than -40C often times - this varies immensely though too depending on the sample.
Other than that, I wouldn't run any other parts subzero - I can't think of anything else that would benefit, and the condensation potential would be a real problem.
When talking subzero, the first thing I do when overclocking is test for coldbugs, then keep it warmer than the CB temperature.
Many other CPUs have cold bugs however, some don't like to be colder than -20C or so. Many don't like to go more than about -100C.
What temperature different CPUs like varies. The same goes for GPUs. Some motherboards are cold sensitive as well, while some models (older ones) liked LN2 on the northbridge to run high FSB. Then memory, a lot of memory doesn't like to be too cold, not more than -40C often times - this varies immensely though too depending on the sample.
Other than that, I wouldn't run any other parts subzero - I can't think of anything else that would benefit, and the condensation potential would be a real problem.
When talking subzero, the first thing I do when overclocking is test for coldbugs, then keep it warmer than the CB temperature.
- Joined
- May 10, 2009
All?
Some Geforce 7xxx cards coldbug at +36c. Put an aftermarket cooler on them and they freak out because they aren't warm enough.
The motherboard itself has a variety of bits that are only rated to 0c on it, most of them will go lower.
HDDs (mechanical) do not like subzero and often will suffer permanent death if operated there.
So the official answer to your question is, 40c. That's a safe temp for everything.
None of the parts used in a (normal...) computer will turn into superconductors, regardless of the temperature.
Some Geforce 7xxx cards coldbug at +36c. Put an aftermarket cooler on them and they freak out because they aren't warm enough.
The motherboard itself has a variety of bits that are only rated to 0c on it, most of them will go lower.
HDDs (mechanical) do not like subzero and often will suffer permanent death if operated there.
So the official answer to your question is, 40c. That's a safe temp for everything.
None of the parts used in a (normal...) computer will turn into superconductors, regardless of the temperature.
- Joined
- Jan 10, 2012
- Location
- Idaho
I doubt anyone actually knows the answer to this, but surely there must be a point at which the expansion and contraction of dissimilar materials comes into play, is there not? When you're dealing with "wires" in the nano-meter range I would think that eventually the gold in one part of the IC would no longer touch the silver, or copper, or other gold contacts in the rest of the chip....or something along those lines. Has this ever been addressed to anyone's knowledge?
- Joined
- Nov 12, 2002
- Location
- Rootstown, OH
I think as everything changes temp at about the same rate, it isn't an issue. I have had boards and chips that have been frozen dozens of times and still run like the day they were fabbed.
But you are right in your first comment. I don't really know anything other than my experience.
But you are right in your first comment. I don't really know anything other than my experience.
- Joined
- Jan 10, 2012
- Location
- Idaho
I did a little searching and couldn't really find anything specific to extreme cooling of a processor. However, a quick search revealed that a mm long wire made of gold will shrink .0000142mm for every one degree Celcius drop in temperature. Copper shrinks .0000166mm and 50/50 solder shrinks .000024mm. So, taking a processor down 120 degrees C (possible with LN2 from what I understand) that same 1mm long gold wire still only shrinks .0017mm. ALMOST 2 one-thousandths of a mm.
Is that a big deal on the nano-scale? Who knows?! I haven't heard of many processors failing because of that reason so I guess it must not be too big of a factor.
Is that a big deal on the nano-scale? Who knows?! I haven't heard of many processors failing because of that reason so I guess it must not be too big of a factor.
- Joined
- Sep 15, 2007
- Location
- Small town Emlenton, PA
The changes on silicon (CPU itself) are much much less. And it's a solid object. So no, not a big deal.
