How efficient will this CPU waterblock be in dissipating heat ?

[xmanli]

Hi,
I have been thinking scientifically about the heat dissipation efficiency of the cpu waterblocks currently available in the market. I was not satisfied cause scientifically, I found that water is not a better heat conductor than silver/copper/aluminium. Where as the heat sink are better conductors, but why can't they cool a cpu better than water cooling.

My guess was:
- Water has a special feature. It is movable. It comes to the CPU while cool, takes away the heat with it, and moves away quickly allowing other colder water to reach the cpu. If silver/copper/aluminium had the same movable/flowable feature as water, then instead of using water, people would have used those instead.
- Even though Metal is better conductor, they could just take away more heat from the CPU faster than water, but when they get as hot as the CPU, their limit is reach. Then air is blown to/away from the heat sink, but this is very in-efficient.

So I thought wouldn't it be a great idea, to actually combine them both as both of them have their own benefits. Guess waht I came up with, A Huge heat sink within a water-block.

I drew my idea which in below link. Of course, the heat-pipes are not enough, it would be better to have at least 12 heat-pipes, 6 on each sides.

Any comments please, before someone actually starts to make this a reality.

xmanli

http://www.pictureshoster.com/viewer.php?file=k1vgww5sskzvnynjykd.jpg

 

[rubix_1011]

You basically are describing a slushbox, only in most cases, it is chilled water run over a watercooling radiator to cool the internal coolant lower than ambient. In this case, you are cooling an air radiator with water, but you would be far more effective using water to absorb and remove the initial heat from the CPU, not air or heatpipes.

 

[jrhii]

This may be more efficient then conventional waterblocks, but you logic is slightly wrong.

Yes, metal is better at conducting the heat than water, but that isn't the problem so much as air is worse at conducting heat than water, and the heat needs to transfer from the metal fins to air.

What this idea is is that you are using water to conduct the heat away from the heatsink, which should perform better than a block because it has more surface area, but you would need to be sure that most of the heat block has good water flow, and reduce static areas of water.

 

[RJR]

You may want to do a little more research on cpu blocks. The business end of the blocks are solid copper with some form of area enhancement like groves, spikes, etc. to maximize the surface area that the water passes over to help dissipate the heat load as quickly and efficiently as possible. Trying to take the heat load farther away before trying to remove it will be counter productive, to say the least.

Here is a good look at some blocks:
http://www.overclock.net/water-cooling/393942-hall-blocks-cpu-updated-often.html

 

[simon12]

Has anyone ever tried pumping water through the heat pipes on a coolmaster hyper 212+ or similar? This would get both air and water cooling in 1 and its much cheaper than a water block. I think it would be very hard to attach the water in and out pipes though.

 

[rubix_1011]

You'd have to completely redesign the block like that, and once you did, you'd have almost the same thing as a regular CPU waterblock, just with a bunch of smaller water channels instead of one larger one. It's all about flow, and more is better in most cases. If the block is correctly designed to have a dense heat collection surface (copper block) and also provide large amounts of surface area for the water to run through (the channels) then the flow takes care of the rest.

 

[xmanli]

I guess my design is not efficient enough. As said above, it seems to lack a good flow of water and a large surface area for heat dissipation.

Anyone have a better design idea ?

I am seriously thinking to build a water-block that could at least perform 25% better in lowering CPU temperature than current water-blocks in the market.

 

[xmanli]

Best answer selected by xmanli.