100% Passive Water Cooling idea
Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee
-
- Friend of SPCR
- Posts: 502
- Joined: Sat Oct 23, 2004 2:02 pm
100% Passive Water Cooling idea
Hey, earlier today I was thinking about water cooling and had an idea to make the entire thing passive. Simply construct a water block in such a way that the output tube would be oriented at a 45 degree to whatever is considered to be up (so the motherboard can be mounted horizontally or vertically) and let convection currents pump the water instead of an actual pump. Then have the water come to and from the reservoir. Maybe it would be possible to create a valve mechanism in the tubing like in human veins to prevent backflow.
What do you think? Would it work or would the CPU fry before the water gets hot enough?
What do you think? Would it work or would the CPU fry before the water gets hot enough?
-
- *Lifetime Patron*
- Posts: 475
- Joined: Fri Sep 26, 2003 6:28 pm
- Location: CT, USA
i've seen this before... i think even on this site. try a search, maybe under thermo-buoyant flow or something like that.
yes it works, but obviously not as well as with a pump (and forget about having it on cpu + nb + hot vga). you basically need to design carefully and minimize the head of your system (use short lengths of wide tubing etc.). but if you have a lower power cpu, it might be worth a shot.
yes it works, but obviously not as well as with a pump (and forget about having it on cpu + nb + hot vga). you basically need to design carefully and minimize the head of your system (use short lengths of wide tubing etc.). but if you have a lower power cpu, it might be worth a shot.
-
- Friend of SPCR
- Posts: 502
- Joined: Sat Oct 23, 2004 2:02 pm
-
- Posts: 271
- Joined: Thu Jun 24, 2004 3:19 pm
The setup you speak of isnt realistic. #1 thermosyphon type setups suck. And ontop of that, the ones that were made had very small loops. As in the ID of the return line was like 2mm and the "condenser" portion was only an inch away from the base/block. None were meant to be passive. Heatpipes pretty much phased these things out even though I dont know of them ever being used.Shining Arcanine wrote:I have an Intel 2.4GHz Pentium 4 Northwood with an 800MHz FSB. Do you think that is low power enough to be cooled effectively by this type of water cooling setup? Do you know of anyone who sells efficient ready made water blocks for this kind of setup?
#2 you'd also have to worry about a passive way to radiate the heat. Doesnt matter how you are moving the fluid or transfering the heat to the fluid, you still need to radiate the heat somehow. Reserator, reserator clone, or base board heaters would be a good place to look.
Trying not to need a pump is going to be very hard. If you can come up with a solution that moves heat great distances (relatively) utilizing no energy you'd be able to make a very large chunk of change. By all means try. But if you arent some brilliant engineer chances are you are going to be banging your head against a wall. It's been asked many times before for a very long time by a very large population. Computers arent the only things that need cooling
This guy asked pretty much the same question:
http://www.ocforums.com/showthread.php?t=392919
Not only has it already been done, but tubes that do exactly this are available on the market right now. They're called heatpipes, and use phase-change to move heat energy far more effectively than you could without. You've still got some thermal resistance and have to dump the heat somewhere, but they're pretty much ideal for your application - far smaller, lighter and better at transferring heat than what you have in mind.jamesavery22 wrote:If you can come up with a solution that moves heat great distances (relatively) utilizing no energy you'd be able to make a very large chunk of change. By all means try. But if you arent some brilliant engineer chances are you are going to be banging your head against a wall. It's been asked many times before for a very long time by a very large population. Computers arent the only things that need cooling
-
- Posts: 271
- Joined: Thu Jun 24, 2004 3:19 pm
Should have bolded "great distances." As in more than a few feet. A standard heatpipe is limited to a couple of feet and its capacity is greatly reduced. Look at that zalman passive case thingy. Takes 6 heatpipes to carry the heat less than a foot away. Course part of that design is to "spread" the heat around.pdf27 wrote:Not only has it already been done, but tubes that do exactly this are available on the market right now. They're called heatpipes, and use phase-change to move heat energy far more effectively than you could without. You've still got some thermal resistance and have to dump the heat somewhere, but they're pretty much ideal for your application - far smaller, lighter and better at transferring heat than what you have in mind.jamesavery22 wrote:If you can come up with a solution that moves heat great distances (relatively) utilizing no energy you'd be able to make a very large chunk of change. By all means try. But if you arent some brilliant engineer chances are you are going to be banging your head against a wall. It's been asked many times before for a very long time by a very large population. Computers arent the only things that need cooling
Actually what you have proposed is very realistic. It is a thermosyphon.jamesavery22 wrote:The setup you speak of isnt realistic. #1 thermosyphon type setups suck. And ontop of that, the ones that were made had very small loops. As in the ID of the return line was like 2mm and the "condenser" portion was only an inch away from the base/block. None were meant to be passive. Heatpipes pretty much phased these things out even though I dont know of them ever being used.
#2 you'd also have to worry about a passive way to radiate the heat. Doesnt matter how you are moving the fluid or transfering the heat to the fluid, you still need to radiate the heat somehow. Reserator, reserator clone, or base board heaters would be a good place to look.
Heatpipes are a similar technology (i.e. using the latent heat of vaporiztion of a liquid) as the mechanism for cooling, however, the driving force for mass transfer in a heat pipe are capillary forces.
In a themosyhon this force results from the gravatationally induced buoyancy.
The pratical implication of this is that heatpipes have the advantage that they are able to operate in any orientation. Thermosyphons cannot - they must be oriented so that the heat sink is below the condenser.
The benefit of a thermosyphon is that they are at least an order of magnitude better heat conductors than heat pipes.
In both cases the heat can be conducted quite far from the source - I have read demonstrations that show thermosyphons being used to move heat 10 or 15 cm from the source.
One of the keys in both HPs and TSs is to use a working fluid that (1) changes phase at the appropriate temperature; (2) has a large latent heat of vaporization.