Waterblock thickness

[KLowD9x]

I heard in an earlier post that aluminium and copper thicker than 1/32nd of an inch becomes an insulator...I was looking for some waterblocks and they all appear to be more than 1/32... Wouldnt they cool better if the block was thinner? also another idea I had was to put fins in the waterblock, wouldnt that cool better too?

[.. AE 86 ..]

fins on the waterblock wouldn't help a lot

cux most heat is carried away in the water and the waterblock wouldn't get even warm...

[KLowD9x]

but, whatabout making the waterblock thinner?

[It_The_Cow]

At 1/32 of an inch, there wouldn't be much water that would be going through unless the block was completely hollow and water was dispersed all within. It would be more difficult to manufacture, and it would still be a pee-shooter

The logical answer to adding fins is yes, but it may not be. A very well designed block without fins can be better than a poorly designed one with fins, so it would take some research and plenty of models

[KLowD9x]

im not talking about the whole block, im just talking about the bottom.

[It_The_Cow]

Ohhh. In that case, this (http://overclockers.com/articles305) may clarify things a bit. The'y re for cold plates, but they can still apply to waterblock bases. It would seem better to have the water as close to the die as possible, but the heat dispersion could be more benficial. If they're still at it, www.spodesabode.com should be working on direct-die water cooling project, which is somewhat related to what you're aiming at

[KLowD9x]

thanx

[ptcg]

I made my waterblock out of .025" thick copper sheet folded into a box & soldered at the seams... worked just fine... held 40C and under at 100%load (1200Mhz Athlon @ 1350Mhz/1.85V)

I didn't have the best pump or radiator setup either.. I'm sure it would have held cooler with some high quality gear.

Oh.. Had fins with a fan on top too (as a backup for pump failure)

[The Overclocker]

i think it depends on the type of metal, aluminum heatsinks should have a large base to absorb more heat, copper should have thinner fins to dissapate more heat

[r0ckstarbob]

I had that question not too long ago. Dave Smith from the AMDmB forums straightened me right up. He's a Thermodynamic Engineer who designs cooling processes for paper mills in Michigan. This is the gist of what i learned on the topic.

Water cannot absorb heat at an acceptable rate to make this a worth while idea. Copper is used as a medium to regulate the heat transfer from your CPU to your coolant. It regulates it to levels where the water will be able to keep up and provide the most benefit. there also isn't enough surface area on the Die itself for the water to do a sufficent job of pulling the heat off of it. Copper also helps increase the surface area in which the transfer of heat can be accomplished.

example: whats easier to cool down - one square centimeter putting out 100 degrees of temperature per second

or

10 square centimeters putting out 10 degrees of heat per second?

see what i mean?

otherwise, they'd be making waterblocks with the thnnest pieces of copper they could and cold plates would be a thing of the past. if the goal was to merely reduce the amount of resistance between the item creating the heat and the heat removing medium, then we actually wouldn't need heat sinks at all seeing as we're relying on the air to cool it off in an aircooled system. we'd just blow fans directly onto our CPU's

i'd definately recommend against directly watercooling your TEC or your CPU. bad things are bound to happen.

if you have further questions, try reading this
http://overclockers.com/articles305/

peazzout

RSB

[BillA]

KLowD
I must admit to being continously intrigued by the idea of increasing the surface area on the topside of the wb bp

take a look at this (true monster)

http://www.forumoc.com/showthread.php?s=&threadid=22306

there is an older Intel paper on the net describing testing that shows an internally finned wb to be superior to a labyrinth

how this would compare directly to a die area impingment wb I cannot say
but I should know in a month or so

virtually no good data on wb testing

be cool