I have been thinking about using a small heatsink as the basis for a water block, and finally got to putting one together for a test. The idea behind this is simple: The more surface area exposed to water, the better cooling – this is directly analogous to an air-cooled heatsink – the more surface area, the better it cools.
I looked around and found an old 486 aluminum heatsink and decided to use that for this test (pic below). It’s about 2″ around and 3/4″ high – not a lot of surface area but a start.
The approach I used was to cut up some old slot covers and epoxy them to the heatsink, forming a water-tight enclosure. I drilled two holes in the top (the cool water intake in the center) and threaded two 3/8″ brass barbs into the plastic for the inlet/outlet hoses. After about an hour of cutting and epoxying, I tested the heatsink (pic below) to see if it was water-tight; it wasn’t.
I then proceeded to add fiberglass cloth around the top to stem the water – that worked OK. I then had to use a lot of teflon tape on the threads to staunch the water leaking there and was successful enough so that I could test the waterblock.
I used my portable water cooler setup (pic HERE) using two stacked peltiers as the heat source – these put out about 100 watts so the test is pretty severe. I fired everything up, waited for the temp to stabilize, and results were:
Recycled Heatsink: -22.4 C
This is 1 to 2.4 C better than the waterblocks I tested a couple of days ago (HERE).
The point of all this is to demonstrate that the water to surface area ratio is a prime consideration in water block cooling performance, and as effective as straight drilling of holes appears to be, some more esoteric approaches should be considered. The Senfu water block performs quite well for its diminutive size, and I believe its performance is attributable to its high water to surface area ratio.
All this also suggests that adapting air cooled heatsinks to water cooling could yield very interesting results and water block designs. The challenge is how to do it in a cost-effective manner – epoxy might be one approach but I would have to be convinced of its long term viability before I mount it in my system. I would love to get a copper heatsink and solder copper sides to it – I think that could be a real winner.