|
Seuss
Error: divide byzero



Registered: 04/27/01
Posts: 23,480
Loc: Caribbean
Last seen: 2 months, 20 days
|
Heat pipe design and water cooling electronics
#8000600 - 02/09/08 03:22 PM (15 years, 11 months ago) |
|
|
I've been doing a lot of research into various methods of thermal cooling for computers. I'm mainly interested in silence, but want decent cooling for overclocking. The solution to this problem is water cooling. Water is used to transfer heat from the electronics to a radiator. The radiator then transfers the heat to the air. A pump is used to circulate the water from the electronics, through the radiator, and back again.
Would it work to ditch the pump and plumb the system with flexible copper tubing filled with the proper amount of a refrigerant forming a heat-pipe? For a computer, the radiator and fans would be on top of the case allowing gravity to return condensation from the radiator back to the hot side.
I haven't worked out the math yet. Is the idea plausible?
-------------------- Just another spore in the wind.
|
supercollider
superconducting



Registered: 10/13/00
Posts: 1,234
Loc: Waxahachie
|
Re: Heat pipe design and water cooling electronics [Re: Seuss]
#8001221 - 02/09/08 05:15 PM (15 years, 11 months ago) |
|
|
Sounds like you're describing heat pipe coolers that have already been on the market for a while. They have sealed copper tubes with water or some other fluid under low pressure inside. It boils at low temperatures to conduct heat along the tube very quickly, and a wick-type construction brings the liquid back to the heat source.
I've never seen a cooler that used more than a few inches of heat pipe though, and that may be because there's a practical limit on the length. Or because heat pipes are expensive. I dunno.
I have seen a graphics card cooler that took up a vacant PCI slot below the card, and had heat pipes running from the GPU out the back of the case to a heatsink out there.
-------------------- Supercollider? I just met her!
|
Seuss
Error: divide byzero



Registered: 04/27/01
Posts: 23,480
Loc: Caribbean
Last seen: 2 months, 20 days
|
Re: Heat pipe design and water cooling electronics [Re: supercollider]
#8003678 - 02/10/08 04:55 AM (15 years, 11 months ago) |
|
|
> 've never seen a cooler that used more than a few inches of heat pipe though
Thats the big difference. I'm looking at piping heat away using a few feet of pipe... and using a standard water radiator as the condensing end of the pipe. I hated thermal, but am probably going to have to break out the book for this one. *sigh*
-------------------- Just another spore in the wind.
|
stefan
work in progress


Registered: 04/11/01
Posts: 8,932
Loc: The Netherlands
Last seen: 3 years, 3 months
|
Re: Heat pipe design and water cooling electronics [Re: Seuss]
#8003686 - 02/10/08 05:06 AM (15 years, 11 months ago) |
|
|
I don't know too much about this, but it sounds like a cool idea, using natural laws to make it circulate.
Just a thought.. I think you would want the pipe on the heat source to be as vertical as possible so the heated water/refridging fluid has the least resistance to flow upwards to be replaced with cool fluid that just came from the radiator.
I think this could work, if the water is moving fast enough. you'll have to figure out the optimal diametre for that, or different ones throughout the pipe to create optimal suction/flow. the you would have a super silent pc 
maybe it's worth it to try it on an old pc, or processor, so that it doens't matter if it'gets ruined by the experiment somehow, that you just overclock to see how well it cooled
|
Seuss
Error: divide byzero



Registered: 04/27/01
Posts: 23,480
Loc: Caribbean
Last seen: 2 months, 20 days
|
Re: Heat pipe design and water cooling electronics [Re: stefan]
#8004634 - 02/10/08 01:16 PM (15 years, 11 months ago) |
|
|
> ust a thought.. I think you would want the pipe on the heat source to be as vertical as possible
It isn't too important. The heat causes a phase change in the refrigerant from liquid to gas. Hot gas rises. (There is also a pressure differential between the hot and cold sides.) There is a small volume of liquid within the heatsink on the hot end.
> if the water is moving fast enough
I would be using something with a much lower boiling point. The thermal transfer comes from the phase change. The hot side heats the liquid to a gas. The hot gas travels up the heat pipe to the radiator. The radiator transfers the heat to the air allowing the gas to condense back to liquid. The liquid drains back down to the hot side heat sink to repeat the process.
> maybe it's worth it to try it on an old pc, or processor,
Nah, I would build a test rig with a controlled heater to proof the concept. I am just trying to avoid working out the math if it isn't possible. Unfortunately, I don't think anybody is going to do it for me, sooo...
-------------------- Just another spore in the wind.
|
|