The Phobya CPU waterblock

[mattkosem]

Any thinner and it'd be direct-die cooling. Aah, the old days of glueing plastic caps on Cyrix 6x86s...

--Matt

[chris.y2k.r1]

Any thinner and it'd be direct-die cooling. Aah, the old days of glueing plastic caps on Cyrix 6x86s...

--Matt

Is there such a thing as a direct die CPU block? I was thinking an O ring seal around the IHS of the CPU that fits into a grove in the housing of the block pressed on by the CPU mounting mechanism would work quite well.. Is there such a puppy?

/End HiJack

[aspire.comptech]

Is there such a thing as a direct die CPU block? I was thinking an O ring seal around the IHS of the CPU that fits into a grove in the housing of the block pressed on by the CPU mounting mechanism would work quite well.. Is there such a puppy?

/End HiJack

People ave made them but they are far to risky to ever be sold through retail channels.

Also to get the best heat dissipation you need to mill some kind of fin/pin grid into the ihs as running water over the smooth IHS would in all likelihood perform worse than a good non direct die block.

[chris.y2k.r1]

Also, if applied right you wouldn't need a pin matrix on the IHS itself, just a extreme flat IHS. From everything I've done, that strikes the best balence between direct IHS cooling and usability.

More to come later)

andyc

See that does make sense to me, because you are saying that water directly on the IHS does not conduct heat as well as a heat sync pressed ontop of the IHS with pins... That cannot be correct, it's got to be due to surface area and the lack of with the direct concept. You could compensate so that by bringing the "ceiling" down closer to the IHS after the water enters the area, creating higher pressure in the block.

[chris.y2k.r1]

Well,

I'm not going to entirely tip my design just yet. What I am saying is there are different ways of acheiving better thermal conductivity than what's been produced so far. I'm actually asserting just the opposiste in a round about way

andyc

Actually from what I just checked out, copper is around 800 times better at conducting heat than water so... http://www.engineeringtoolbox.com/th...ity-d_429.html I need to work out some numbers. And, it should not be that hard to get the concept right so it's no more risky than a standard LC system. I'm sure people felt the same way about the risks when the accepted LC systems were first proposed.

Has anyone ever tried adding super fine grains of copper to their loops?? If enough of it was in the loop and did not clog anything that should dramatically increase thermal conductivity and the greatly increase the rate of heat transfer.

[fox3]

I think that would add to the viscocity of the cooling solution and add more work to the pump which might increase the heat dump of the pump.

That is an interesting idea though for better cooling properties of the cooling solution.

[skinnee]

As far as the OP topic goes, I'd love to get one of the blocks if it ever becomes availible in the US. It looks like a great design to me, and I'd love to test it out.

That makes two of us!

[gabe]

Also, if applied right you wouldn't need a pin matrix on the IHS itself, just a extreme flat IHS. From everything I've done, that strikes the best balence between direct IHS cooling and usability.

More to come later)

andyc

[Movieman]

Oh you mcoffey fanboi you!
He does do nice work though and knows his stuff!
Merry Christmas!

[warriorpoet]

[lyl]

What are you sticking your little thumbs up for, you don't have one yet

andyc

Uh ohhh