View Full Version : "Stackable" CPU Waterblock - been done/tried?

12-26-2008, 10:59 AM
I've been mulling over an idea I've had for a while. What I would like to try is to build a waterblock that has vertically stackable copper bases. Kinds of like SLI for waterblocks. You could have the water circulate from the bottom block up through each additional block. I do think each additional block would conduct more and more heat away from the CPU so it would work in theory but I don't know if it would fall under the category of 'not worth it due to diminishing returns' - kind of alot like SLI or Xfire now that I say that. Has anyone seen or tried such a thing? If not, I think I am going to try it.

12-26-2008, 11:44 AM
I'm not sure I understand...

Got any sketches??

12-26-2008, 11:54 AM
I don't think it would really add anything. All of the heat that is transferred from the first base to the water would then have to be transferred from the water to another base and then to the water at an earlier point in the loop. I may not be understanding what you are saying so correct me if I am wrong.

12-26-2008, 11:59 AM
I think I do.....if I understood correctly what you are proposing I can safely say it will probably work poorly. This is by no means a way to discourage you. On the contrary!

What I think would happen is:

a) flow kill....the path the water would have to follow would be long and restrictive

b) in the event that in each "stage" water would flow easily then basically you are looking at doing the exact opposite of what watercooling is about. The whole point is to use water as a transfer medium and copper as a "thermal interface". From this you can conclude that water must flow as quickly as possible from inlet to outlet whilst maximizing heat absorption. The interface material should provide maximum surface area to share the heat with the medium and still be thin enough to not work against the thermal properties of flowing water.

Let us imagine you made a large block with a "deep" inlet and a "dep" outlet carved inside of it. Baically a 10cm tall 4x4cm base block. To begin with the copper you just addded would not be absorbing anyheat from the cpu cores. The thermal gradient would prove the added copper to be completely useless.

Infact almost all manufacturers are working on achieving even thinner bases and denser microchannel arrays.

But by all means try it out!

12-26-2008, 04:42 PM
The heat has to move through more stuff, thus worse thermal conductivity. Well not exactly, but you catch my drift... the ideal block would be a 0.01 cm block of diamond (900 - 2320 versus copper which has around 380 W/(m·K) ) that causes a lot of turbulence, not a big block with a lo of restriction.

But we like DIY more than ideal examples as above !

12-26-2008, 04:54 PM
2 half cpu blocks side by side with each its own radiator and pump would be a better idea.