Natedog: a rule tumb (read in one of those links i posteed before; cannot remember where (maybe at Diamonex or P1 Diamond inc. site's) is to use a heat spreader as thick as the core you are coolinng and 5 times as big

I think Joe and BB2K were confused here. Graphite and "common" carbon-fiber or carbon-foam based heat spreaders are anisotropic materials, that indeed transfer heat mostly in one direction.
Diamond, synthetic or not, is completely anisotropic. Its cristalline structure is well known, and identical in both natural and synthetic diamonds. Heat conductivity and spreading abilities are incredible, and the highest one can get on Earth nowaday.
A "localized" hot spot would be a sign of a low conductivity, which is not the case obviously. A diamond plate on a CPU core would have a more uniform temp distribution than an equivalent copper plate.

So basically it would work.

This is totally overkill maybe, but people, remember this is not the point here. Let's keep constructive comments.

Shaping the diamond piece is the tough point here. If Natedog manages to make it i'll be eager to see performance figures.

well I think Im going to scratch the diamond concept as a 12mmx12mm 1mmthick CVD cost $500!!! from www.diamonex.com
I was hoping for more like $50. $500 OUCH If anyone knows where to get one for cheaper LMK.

you would need a min of 3mm thick also, at 1mm it would be so fragile it would shatter just trying to mount it on a core. Thats what stopped my investigation into it. When it was basicly determined that the diamond couldnt take the small stress point in the center like that.

Price/practicality is always a point IMO. At $500 for a 1mm plate who is going to buy it and who is going to make it. For a 3mm plate (still think it will break) is what? $600-$700 maybe over a $1,000? Then the cost of making it? Ouch. Better off with somekind of active cooling and a Copper block. I thought the price was known or I would have pointed it out in my last post. If it was only $50 I would have done it already just for bragging rights of having a diamond WB even if it didn't work. :evilaugh:

If someone wants to afford this and give it a shot have at it though. Otherwise Silver is still an option and hasn't been done to much yet. Only can think of a couple playing with it. You can get a 5oz 99.5% pure bar for what? $25? What could be milled into a nice insert of somekind where most of the water contact would still be on the silver so the heat transfer between metals wouldn't be much of an issue. I have no clue how to mill silver or how hard it is to work with. One of these days I may try it just to see for myself. Can't be harder than diamonds though.

hehe its sad when you could get a full vapochill system for the price of a cold plate :)

A rich geek.
Consider some ppl buy Ferraris or McLarens just for fun. So what's $500, or $1000 for them ? Not even the price of a tire for their car. (try to get a 275/45/18...)
Besides why are you in watercooling at all. It's expensive as well. I do it because i can.

I heard it was rather soft. So it would be quite easy to mill.

Water cooling expensive? Not for me it isn't. I make my own blocks with my mill and put the rest together myself. Have yet to pay over $100 for a water cooling system. I have about $100 in the 15 or so project blocks I made but that is just a hobbie and one that will not extend into the $1,000 range for ONE block that may or may not work. Good god there are better ways. If someone had that kind of money to blow it would be on something other than this I would hope. Still if they want to do it have at it! I just don't see to many people running around these forums with $1,000 burning a hole in their pocket. ;) If they did I doubt they would be here in the first place as they could hire someone to do the engineering.

Copper is soft to and is a major bitch to mill. Softness is meaningless, it is the density that matters. ;)