How cool is too cool?

[Reid]

I have been experimenting on various homemade cooling systems and ideas. One question I have not been able to find an answer for is, what are the lower limits of cpu function in regards to temperature? 10 degrees C? sub zero degrees C?

 

[linnemeyerhere]

Not sure how cool is cool cause I'm still trying to figure out what is hip? Sorry I just have to post this.
I figure if the mad scientist cool the CPU's with liquid nitrogen while massive over clocking they can operate way cool!

 

[scrooge]

I have been experimenting on various homemade cooling systems and ideas. One question I have not been able to find an answer for is, what are the lower limits of cpu function in regards to temperature? 10 degrees C? sub zero degrees C?

have you come up with anything yet?
don't keep us hanglng. let see your work. i always like to see some new ideas.

scrooge

 

[profdlp]

I can't really answer your question (sorry), but I know one of the concerns with extremely low CPU temps has to do with condensation.

See the second section regarding "Dewpoint" here:

Relative Humidity

 

[A Guy]

I believe that CPU's will be fine at temps like -40c, probably to -100c, I've seen a test where they got to -191c. Of course, that isn't constant over long periods. I'd think you'd be looking at the silicon breaking down before the electrical components. In any case, you'd never approach those temps with any "normal" coolers. A Guy

 

[bobtran]

My pc is water cooled and consistently displays 35C-45C on a Q6600 @ 2.4GHz (or 40C - 50C @ 3.0GHz overclock) with dual 9800GT video cards running SLI and a 1200 watt PWS in the same case.

I have heard of people running in the upper 20C to lower 30'sC.....but I must agree that you have to watch out for condensation if you go too cool otherwise we would all use small AC units to cool our pc's and be done with the whole cooling issue.

 

[karlsnooks]

check the specs on the chip involved. I've been away from such things for too long, otherwise, I could instantly give you an answer. must be old-timer's diseaase.

I have an acquaintance who puts his laptop into the refrigerator to keep the temps down while he is rendering. The lower temps give him about 5 times better output in terms of frames/sec since the cpu doesn't get throttled this way.

Your computer will have some components which will have a higher lower limit. Unless you are really going to extremes, the specs for components is normally on the safe side.

As someone else mentioned, your bigger concern will be condensation.

 

[Reid]

Yes condensation is a concern, but I have found that if you double tube (ie using a refrigerant pipe with freon surrounded by a regular plastic tube), you can keep the PIPE condensation contained and drained away from your system. I havent actually hooked this up to a computer yet but have been running this system on a trial basis to see what kind of issues emerge. This is what actually brought my initial question to light because it gets extremely cold concentrated on the "heatsink" which would be butted up against my cpu.
On another note, what is keeping someone from just housing their reservoir and radiator in a fridge or freezer at around 0 degrees C?

 

[Thorsen]

-190 here i think

5 GHz Project: CPU Cooling With Liquid Nitrogen : Record Attempt: The 5 GHz Project

 

[Reid]

LOL, that is absolutely awesome Thorsen! Thank you. Just do not ever shut it off long enough to thaw out eh?

 

[A Guy]

-190 here i think

5 GHz Project: CPU Cooling With Liquid Nitrogen : Record Attempt: The 5 GHz Project

Yeah that's the one A Guy

 

[Lordbob75]

I have a friend who has phase change cooling setup, and his CPU runs at 0C at load, and below that on a regular basis. There really isn't a bottom end of heat limit (I.E., it can really be run as cold as you want), as long as its above 0K and there isn't so much condensation building up that it freezes or causes shorts.

~Lordbob

 

[carwiz]

Yeah, and that was 5G's back in December of 2003. :shock: (Think P4)
There's some 5Ghz PCs now without the radical cooling. At least I would consider liquid nitrogen radical.

 

[Thorsen]

Here ya go: (6.3Ghz @ -230C)

 

[Thorsen]

not sure what the temp was or how it was controlled, but here is another record:

New CPU Overclocking Record Set, 8.20GHz - Softpedia

HEXUS.net - News :: Overclocker hits 8.19GHz with Intel Celeron CPU, sets new world record : Page - 1/1

as the doc says it wasnt benchmarked that thoroughly so high mark at 8.20 Ghz but

 

[smarteyeball]

Celeron D at 8.2GHz... nice. Utterly pointless, but still nice.

 

[Thorsen]

Its the only way you can beat Minesweeper in less than 4 seconds....

Or couple it with this for epic solitaire runs:

 

[simmo]

Yes condensation is a concern, but I have found that if you double tube (ie using a refrigerant pipe with freon surrounded by a regular plastic tube), you can keep the PIPE condensation contained and drained away from your system. I havent actually hooked this up to a computer yet but have been running this system on a trial basis to see what kind of issues emerge. This is what actually brought my initial question to light because it gets extremely cold concentrated on the "heatsink" which would be butted up against my cpu.
On another note, what is keeping someone from just housing their reservoir and radiator in a fridge or freezer at around 0 degrees C?

Hi mate..
Firstly. You will get condensation whether you dual tube the pipe or not..it will occur between the CPU and Socket and you will get condensation all over pretty much around the Socket if you are going to Sub Zero cool.
I'm assuming by your post you are looking at phase change cooling.
You will need to insulate the entire area around the socket front and back of the board and use grease between the socket/CPU.

As to chilling water for a W/C system. You will have the same issues with socket/CPU condensation. I have considered a similar option with this but haven't built/run this setup. I do W/C tho as I O/C the backside off whatever CPU I have installed at the time.
Only concern I have with chilled water (other than CPU/socket condensation) would be condensation occuring around the socket on the mobo..
I think it would be most likely minimal and case fans (hi cfm) would evaporate that as fast as it occured. Condensation on the back of the board may need to looked at also. However chilled water will, obviously, not be at xtremely low levels like Phase change/LN2/DICE cooling so other than possible condensation in the socket than condensation on the Mobo may not occur at all. Probably be something on the net tho.
Or you could check forums on sites like OCAU or similar..most sites dedicated to O/C will have threads on extreme cooling as they are chasing speed and that comes with Heat and so those guys are using Sub zero cooling methods.

cheers

 

[carwiz]

This may be far-fetched but it is science backed and would work. One way to eliminate the condensation problem is to eliminate the source of condensation and that is moisture. If you could seal the cabinet and run it in a vacuum, moisture would be near non-existent. Water and/or water vapor (even ice) will boil-off at a lower temperature as the vacuum increases. It's impossible to get an absolute vacuum on Earth but it is possible to get around 29.919"Hg. Moisture in that vacuum will boil at about -60F and be carried out through the vacuum pump.

Can't be any more radical than running in mineral oil.

An after thought: You wouldn't even have to go that low in vacuum. Just about any discount A/C vacuum pump will draw down to 1,000 Microns. That's about 29.88"Hg. Water will boil at about 0 degrees. That's plenty if you vacuum the cabinet before starting the cooling system. That would remove the moisture.

 

[LiquidSnak]

There is "normal operating range" of temperatures. I can only assume that performance decreases or the system will stop/stall at a certain low point as well as high.