#rotor style Northbridge cooler
 

just doodling in a CAD program thought this might work ok. Need a copper top though and soldered barbs because the inlet and outlet would be too close for normal threaded barbs.

interesting design & I like the top, but since its only the northbridge, why make it more complicated than it needs to be? As far as I know the rotor design isn't really an impingement type anyway. just move the center nozzle and have flow go from corner to corner. but hey, if you've got access to the materials and a mill, go for it. ya lucky bum.

http://3rotor.homelinux.com/images/c...P04663_sml.JPG

as you can see.... not much space is a 40x40 mm block. :)

40 mm is very tight especially with 11.8 mm holes for threaded barbs.

It wasn't really for impingement just a concept really for getting a more even flow. I was just playing with the software but I might make one in a month or so when I have a bit more time. I might just make it corner to corner though

Does anyone know if it's so important to watercool the northbridge ? I though passive aircooling was sufficient...

with the more popular boards at least- nforce2, canter, spring etc- passive cooling is only sufficient at stock, and slight overclocks- active convection increases the potential for overclocks and voltage increases
watercooling them furthers the potential for voltage increases and stability

btw- if it isn't for the sake of low pressure impingement, don't bother, I think you'll get more 'even' flow with the barbs in a side to side configuration
also note that pin fin doesn't do impingement well anyway

Re: Northbridge WCing

Yeah holy shit these Canterwoods/springdale's roast! I have an IC7 and an IS7, both of them come with heatsinks circa early 90's pentium 1 - with a high peed noisy fan :(. So I ripped the fan off hoping to put a big (1.5" tall) old AMD K2 heatsink without a fan on there - but it was getting so hot I had to put a slow 80mm on it.

They roast.