diamond thermal grease?

[Tech Geek Deluxe]

BTW ... I'm expecting the diamond to be far hotter to the touch all the way up the pipes ... than the copper will be ... even though the diamond will be transferring heat to it.

[Elysium]

Keep us updated! Sounds like a really good project... Also some pics would be appreciated...

[gvblake22]

I think some diagrams or pictures would really help illustrate what you're going for here. Almost sounds to me like you are attempting to make a self-sustained cooler that acts similar to a TEC in that it very quickly moves heat from one spot to another. Obviously it won't be as effective as a TEC (no sub-ambient temperatures) but is this kindof what you're going for?

[Tech Geek Deluxe]

Well "%#"*{$ !!!!" ... I carved off several capacitors off of a New Windsor 5200+ ... in an impatient attempt at removing the internal IHS ... bad thing is eating the $115 for the CPU and the restocking fee for the MOBO ... Yup I sent the MOBO back and decided to do it right.

Good thing ... I ordered C2Q Q6600 [G0 stepping] ... and a ABIT IP35 Pro to finish the Project.

I won't be carving off the IHS on it!

Good thing ... AMD is using a very high content silver solder or epoxy on their Windsor 5200+ cores. So removing the IHS isn't something I will do in the future. It was way too easy to slice off a capacitor.
So ... "the Project" has been modified to be an easier project. [With a better Mobo and Cpu ... especially for overclocking]

I'll still do the diamond skin and thermal grease ... but not the diamond IHS.

The Goal will be to keep the core at below 50C 100% load in a Case Ambient Temp of 32C and room Temp of 23C ... at a 24/7 stable 3.2Ghz

I may do a before and after depending on when the diamond powder arrives. If it is before the New Mobo and CPU ... then I will go strait to a permanent build.

[fstroupe]

EDITED POST....didn't notice that there was four pages to this thread. Too bad...I had a nice rant going about the testing contraption and testing itself.

Don't you think you might be better off with a highly lapped heatsink and heat spreader...work up to 2000 grit paper then polished with something even finer...and then use no thermal compound? Or an extremely fine diamond dust that is tiny enough to fill what microscopic voids that are left? Cause two surfaces touching have better thermal contact than the surfaces with a dust-filled goop between them, diamond or otherwise. Though, that probably wouldn't work because you couldn't insure that some of the dust wouldn't be left on the two surfaces...which would defeat the purpose.

As was alluded to earlier, adding the odd shaped diamond dust will probably defeat the purpose of the flat silver flakes in the AS5.

[Tech Geek Deluxe]

Don't you think you might be better off with a highly lapped heatsink and heat spreader

Yeah ... I have been down that road and typically get 3-4 degrees Celsius decreased temps.

I don't think I will lap this one however. The reason is that while lapping is awesome ... the transfer media of the .25 micron diamonds is far smaller than even the thin edge side of the silver in Arctic silver ... even if I were to lap perfectly the surfaces of each surface down to 8000 grit ... the hills and valleys left at that level of perfect lapping would be greater than 1/2 micron ... E.G. enough for the potential of 4 diamond particles to span filling in each surface imperfection on each surface.

Soooo .... I will be slapping on an 80% diamond spread on the horribly rough surfaces of CPU Sink and Heatsink and providing a transfer path of heat siphoning "diamond epoxy skin" all the way up the sink and along the heat pipes.

There will be thousands and thousands of heat paths of diamonds between the CPU and heatsink and the spead of heatloss on the diamond side of the IHS will far exceed the IHS heat absorption, transfer and retention from the die side of it.

If the ""diamond skin" sucessfully carries that farther and faster from the CPU die, IHS, and heatsink base and core ... well ... I am hoping with proper air circ and ventilation ... I will be able to get that 10/15 temp drop below the average others are seeing. [Heck ... I may just do a before and after ... but I'm still hurting from carving up a perfectly good cpu. Plus I forgot to mention I roasted it with a torch to get the solder to pop.]

I only did the AMD and decided to play IHS doctor to squeeze the most out of it. Well I killed it ... and now am just going to safely squeeze some extra clock power from a Core 2 Quad.

I know I can lap it and get perhaps 3-5 degrees C better cooling ... have done so a number of times ... but I got the Heatsink free so I spent $24 on diamond 1/4 micron powder that should now last through the initial "diamond skin siphon" project and maybe two more diamond thermal grease projects when the quads drop below $125

If the diamond skin project produces a 10-15 degree difference ... then I will likely make it a staple for every build ...

I will be re-greasing with the diamond thermal grease ... a core 2 2.4ghz ... and a dual core Pentium D 3.4ghz ... between these two ... I will have a satisfactory reference for simple expected gains from the diamond thermal grease versus "diamond thermal grease plus "diamond skin."

I expect to do a test on idle and load for both ... before the change ... immediately after the change and 12-24 hours burn in ... where I typically tighten the heatsink ... just a tad more to mesh the surfaces better.

I had initially hoped to fab a diamond IHS ... so heat transfer would not be slowed by the cheapo IHS from the CPU die ... but that cost $138 [Dead CPU, Mobo restocking, and shipping] to decide it wasn't worth the risk

[Tech Geek Deluxe]

Well ... I finally made up a batch of Diamond Thermal grease. It was pretty hard to get over 50% diamond in it by volume. Had to use a pretty runny base grease to get more diamond volume. I suspect I am about 50%+ diamond in the compound and 5% silver and the rest basic white silicone and misc. It is thick however ... but spreadable. Used about 10 Carrots of .25 100,000 mesh diamond powder and have a gob the size of an Almond. (Thats enough paste for more system CPU's than I'll go through in 5 years.) Thats about $10 retail worth of diamond powder. FYI

I have 3 systems that will need paste right away. I''l post before and after results on two of them ... if they can outdo my AC5 by 3 degrees or better cooling. Immediate ... short and long term results.

[gvblake22]

Any pics of this concoction?

[Elysium]

Originally Posted by gvblake22

Any pics of this concoction?

I was wondering the same thing... Is it shiny?

[Tyreal]

I'm so excited. I just can't hide it.