Pro/Forums - View Single Post

Cova · 07-23-2002, 04:00 PM

I think there is a pretty important difference between that attempt myv65, and my idea - on the one you linked to the space directly over the CPU is empty - where the heatsink clip would normally be. In my proposed one, that is where all fins start from. I've tried to optimize the points I think are most important with it, and I can't come up with anything better...

1. As little copper as possible between the water and the core
2. As much surface area as possible between the copper and the water
3. Good copper paths to conduct heat quickly away from the core - without insulating the core in a mass of copper.
4. As little resistance to water-flow as possible

With air cooling the heat spreads throughout the heatsink very fast compared to how much heat is dissipated into the air. With water-cooling, as long as fresh cool water is in supply we don't need that much heatsink material to spread it out.

I almost think the best waterblock would be the one that is perfectly straight, with no regard for surface area - just put maximum flow-rate over a piece of copper virtually hair-thin to keep the CPU dry (as close as you can get to just putting water on the core itself) - but after reading of a few attempts at water right on the core, I think it works better with some type of metal heat spreader.

07-23-2002, 04:00 PM	#53
Cova Cooling Savant Join Date: May 2002 Location: Canada Posts: 247	I think there is a pretty important difference between that attempt myv65, and my idea - on the one you linked to the space directly over the CPU is empty - where the heatsink clip would normally be. In my proposed one, that is where all fins start from. I've tried to optimize the points I think are most important with it, and I can't come up with anything better... 1. As little copper as possible between the water and the core 2. As much surface area as possible between the copper and the water 3. Good copper paths to conduct heat quickly away from the core - without insulating the core in a mass of copper. 4. As little resistance to water-flow as possible With air cooling the heat spreads throughout the heatsink very fast compared to how much heat is dissipated into the air. With water-cooling, as long as fresh cool water is in supply we don't need that much heatsink material to spread it out. I almost think the best waterblock would be the one that is perfectly straight, with no regard for surface area - just put maximum flow-rate over a piece of copper virtually hair-thin to keep the CPU dry (as close as you can get to just putting water on the core itself) - but after reading of a few attempts at water right on the core, I think it works better with some type of metal heat spreader.