• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Most effective cooling

D

Daovonnaex

Golden Member
As part of my effort to smash the world overclocking record, I need to know this. My thought was a heat pipe with LHe in it that pipes the heat away (preferrably outside the case or into flourinert). However, there may be a better solution. Could anyone suggest a better one?
 
Liquid nitrogen or any other liquified gas are expensive and get used up quickly. It's also very dangerous if used in an enclosed space: the volume of gas produced from most liquified gases when they evaporate is roughly 800 times that of the liquid volume, and most of the gases used for cooling are either asphixiants (N2, He or Ar) or toxic (CO2).

Also using these very cold liquified substances to try and cool a secondary liquid is a waste of time, as it will just freeze it solid.

A more realistic approach is to try and use some kind of cyclic phase change. A phase change from liquid to gas requires alot of energy. An easy way to see this is to put some water in a terracotta or porous pot. The water will soak through the pot and evaporate (phase change) on the outer surfaces. The water needs energy to evaporate so it takes it from the surface of the pot. This cools the pot and the water inside (over a period of time). If you try it you should see a 5 degree drop with, say, a 1 litre pot. If you can then retreive the evaporated substance and condense it, by taking energy away from it, you can the recycle the liquid. This makes a very effective heat transfer system.

Another way is to use the Joule-Thompson effect. This uses a similar principle of gas expansion and contraction. Expanding a compressed gas causes it to absorb energy because the molecules have to travel faster to get further apart from eachother and require energy to do so. In reverse if the gas is compressed it heats up, all that kinetic energy is transferred into heat energy as the molcules slow down and are squashed together. This is employed in most cooling appliances like fridges and air conditioners. Heres a demonstration.

There is a evaporator / condensor sytem on the market at the moment called Vapochill.. But cheap it ain't.

Yoshi.
 


<< Liquid nitrogen or any other liquified gas are expensive and get used up quickly. It's also very dangerous if used in an enclosed space: the volume of gas produced from most liquified gases when they evaporate is roughly 800 times that of the liquid volume, and most of the gases used for cooling are either asphixiants (N2, He or Ar) or toxic (CO2).

Also using these very cold liquified substances to try and cool a secondary liquid is a waste of time, as it will just freeze it solid.

A more realistic approach is to try and use some kind of cyclic phase change. A phase change from liquid to gas requires alot of energy. An easy way to see this is to put some water in a terracotta or porous pot. The water will soak through the pot and evaporate (phase change) on the outer surfaces. The water needs energy to evaporate so it takes it from the surface of the pot. This cools the pot and the water inside (over a period of time). If you try it you should see a 5 degree drop with, say, a 1 litre pot. If you can then retreive the evaporated substance and condense it, by taking energy away from it, you can the recycle the liquid. This makes a very effective heat transfer system.

Another way is to use the Joule-Thompson effect. This uses a similar principle of gas expansion and contraction. Expanding a compressed gas causes it to absorb energy because the molecules have to travel faster to get further apart from eachother and require energy to do so. In reverse if the gas is compressed it heats up, all that kinetic energy is transferred into heat energy as the molcules slow down and are squashed together. This is employed in most cooling appliances like fridges and air conditioners. Heres a demonstration.

There is a evaporator / condensor sytem on the market at the moment called Vapochill.. But cheap it ain't.

Yoshi. >>

I'm familiar with both cyclic phase change as well as the Joule-Thompson effect. While realistic, neither provide the performance necessary. My only real option is liquid helium if I'm to smash the world overclocking record and keep the title for awhile. However, I'm certain that there's a more effective method of implementing it than a simple copper box. As I am not an engineer (at least, not yet), I am happy to entertain suggestions.
 
If you were seriously devoted, you could rig up a system consisting of a CRAY liquid immersion chiller filled with 3M FC-70/FC-77 Fluorinert pumped to a Fluorinert-to-liquid N2 heat exchanger whose contents are pumped to a huge (and I mean huge, huge paired with titanically powerful) compressor.

I think I'm missing something though. 🙁
 
Really though, whats the point? It's going to cost a lot, and your not even really going to be able to use it. Make sure you get a good chip though, not just an average chip, you might have to look through some.
 
you would have to put a mortgage on your house to afford liquid nitrogen.... besides liquid helium is way colder 😉

it should work if you just use a very fast flowing water cooling system, tom says that you should make the water coolin thing outside of your house. cold water to CPU, piped away as hot water, led outside through some snow or something, back into the system as cold water.
 
Well liquid nitrogen or helium would be the best, but I would like to see more multistage TEC and phase change systems.

Basically, a refrigerator and some thermoeletric coolers. Can't remember exact temps, but I believe I saw one such system
that reached about -50C. You should be able to get lower with proper staging.
 
i've seen 160watt peltiers and such...
those have to get pretty damn cold

not like liquid helium or such... but they might have more powerful ones.. only problem is trying to dissipate the 160watts of heat on the hot side=\
 


<< If you were seriously devoted, you could rig up a system consisting of a CRAY liquid immersion chiller filled with 3M FC-70/FC-77 Fluorinert pumped to a Fluorinert-to-liquid N2 heat exchanger whose contents are pumped to a huge (and I mean huge, huge paired with titanically powerful) compressor.

I think I'm missing something though. 🙁 >>



I thought Flourinert turned solid as it approached LN2 temperatures.
 


<<

<< If you were seriously devoted, you could rig up a system consisting of a CRAY liquid immersion chiller filled with 3M FC-70/FC-77 Fluorinert pumped to a Fluorinert-to-liquid N2 heat exchanger whose contents are pumped to a huge (and I mean huge, huge paired with titanically powerful) compressor.

I think I'm missing something though. 🙁 >>



I thought Flourinert turned solid as it approached LN2 temperatures. >>

It does, unfortunately. My use for flourinert would be to cool parts of the computer that are not around the CPU, as I don't want them to burn, either. Please, I do not want these "practical" solutions, as the main purpose of this is to break the world overclocking record. So essentially, only the most powerful cooling (clearly revolving around liquid helium) will do.
 
If you want an absolute impractical solution, Encoat the entire surfact of the CPU/MB in Diamond with a fractal surface and submerse into Liquid Helium. Diamond has the best heat conductivity of all known non-exotic materials so it will act as a VERY good heatsink.
 


<< If you want an absolute impractical solution, Encoat the entire surfact of the CPU/MB in Diamond with a fractal surface and submerse into Liquid Helium. Diamond has the best heat conductivity of all known non-exotic materials so it will act as a VERY good heatsink. >>



The whole thing would shatter as soon as it touched something as cold as liquid helium. I think you'd need to cool it down slowly to avoid rapid shrinking/expansion of the components.
 


<< If you want an absolute impractical solution, Encoat the entire surfact of the CPU/MB in Diamond with a fractal surface and submerse into Liquid Helium. Diamond has the best heat conductivity of all known non-exotic materials so it will act as a VERY good heatsink. >>

Wouldn't it be more effective to cool the die directly?
 


<< Liquid nitrogen or any other liquified gas are expensive and get used up quickly. It's also very dangerous if used in an enclosed space: the volume of gas produced from most liquified gases when they evaporate is roughly 800 times that of the liquid volume, and most of the gases used for cooling are either asphixiants (N2, He or Ar) or toxic (CO2). >>




😕Ouch, my brain:Q
 
I could be wrong. But most MBs aren't designed to work a lot under -40 degrees. You are subject to damage both the CPU and MB if you can't cool them down "evenly".
 


<< I could be wrong. But most MBs aren't designed to work a lot under -40 degrees. You are subject to damage both the CPU and MB if you can't cool them down "evenly". >>

You're very correct, and neither are CPUs. After a few months of LHe cooling, I'm sure that the CPU would cease to function.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top