Zany home-made liquid cooling idea

[dorker]

I've been thinking of augmenting my CPU sink & fan with a sort of liquid cooling system.
It would involve filling a closed loop of copper tubing with water (or possibly oil) and somehow fixing it to the CPU sink. The copper tubing would most likely have an interior diameter of ~6mm. The liquid should circulate through convection

First of all: is this a safe idea? I have considered oil because it does not conduct electricity in the event of a spill
Second: is it worth it? Is the diameter of the tube too small for the convection to overcome the friction? How does the viscosity of oil affect such a system?

Any thoughts?

 

[belial2k]

1) the properties of oil keep it from being an effective cooling agent.
2) without some kind of block on the cpu you will not make enough contact with just a tube to do any good and provide cooling to the entire chip
3) I don't see how this can be accomplished for less than a traditional system...meaning it would not be as effective and cost you more

 

[cromedome]

If I am not mistaken, on some heat sinks the heat pipes are hollow and have some sort of fluid in them. If you are able to cut the ends of each pipe and flush out the fluid then somehow connect your loop to each pipe and have water+air cooling on the same rig then it might work. It is possible that this might work but I don't think it is worth the time and cost vs performance.

 

[dorker]

Best answer selected by Dorker.

 

[mrmez]

As a project its fine.
If you expect to get better results than a purchased system you will be disappointed.

There are/were a few self contained, single unit (ie: not a separate rad with hoses) systems out there. They were heavy, expensive and did NOT outperform average heatsink and fan designs.

So as a project it will be fine, but remember a few things about physics before you build.

Heat rises. Your cpu (usually) sits vertically. So to utilise convection correctly, your pipes will need 90 degree bends and run upwards. As the hot fluid rises it will cool, and fall back down around the outside of the pipe. Assumptions: 1) Your internal diameter is big enough to allow the fluid to move up through the centre and down around the outside with minimal turbulence. 6mm ID is too small IMHO. 2) You have good case ventilation. Heat rises, right? So all the hot air in your case will gather at the top, right where your cooling pipes are thus reducing their effectiveness.

Your main issue, however is getting the hot fluid to flow easily out of your cpu area and into the pipes to be cooled. The greater the temp differential (ie: how well you can cool the hot fluid), the more this will happen naturally. Design can help this too, but don't ask me how.
Next is cooling. I typical heatsink could easily have 2 square meters of surface area for cooling. Your pipes may just have a few cm. Its just not enough.

Convection alone is just how fluids (air included) move in relation to each other at different temperatures. Its not a magical way of cooling.

 

[Houndsteeth]

Heat pipes (the concept you are describing) is currently being used as the heat transfer mechanism of choice in almost all the high end air cooling HSFs. An enclosed copper tube uses a fluid that has a low boiling temperature (such as any number of ammonia compounds) or water that is under low pressure. As the fluid is heated, it will vaporize and draw heat from its environment. When it reaches the highest point, where it is cooled, the heat drawn during vaporization is released and the fluid condenses, and falls back to the reservoir through capillary action, to be heated and vaporized again.

This has to be an actively cooled system, as a passive system will eventually find equilibrium, and that point of equilibrium could very well be outside the temperature range of the device you are trying to cool. The amount of heat being transferred is not magically lost during the process, but rather transferred into the environment through traditional mechanisms.

Oil makes a very poor medium for this type of cooling as it has a very high vaporization temperature, so the heat pipe does not benefit from the evaporative cooling effect as it does with other fluids.