diamond thermal grease?

[Tech Geek Deluxe]

Hehe ... It is Charcoal grey ... remember the powder is industrial diamond and is tinier in particle size than the actual sharp edge of a Razor Blades cutting edge. So the carbon shows ... but no refraction.

I'll do some pics as I get the 3 builds ready. Mobo comes in tomorrow ... it is the last part I am waiting for.

I haven't made up my mind on the "Diamond skin" however.

I can make a simple epoxy diamond skin ... it will look grey and kinda like a kid did it.

Or I could use a "Jappaning" compound I created for tool treatment and bake it on ... the whole bottom core and heatpipes and the outer fin edges would look piano black and super "shiny" ... it would look like a professional manufactured "diamond skin" and make me feel like a kid.

The Jappaning compound has a 60% carbon content and would be a super hard finish. Hard enough where you can't scratch it with a fingernail. [Made from Ashpaltum, Humic acid powder, Rosin, Linseed oil, Tung Oil and Turpentine.]

Any comments?

[Elysium]

Jappaning gets my vote. Same thermal characteristics as the diamond stuff, right?

[Tech Geek Deluxe]

No ... the jappaning is high carbon ... but with the diamond as carrier media ... it will provide heat transfer to the air and core of the sink and the heat pipes. Adhesion will be excellent [Baked on] ... so I figure It to be a better solution over epoxy resin.

I kinda wish I would have also used carbon black when I made the mix ... it would even have better properties ...

[gvblake22]

Couldn't you start with the simple "diamond skin" and see how it does and then do the "Jappaning" process and re-test? That would give you an interesting before and after comparison...

[Tech Geek Deluxe]

Problem is ... I bought only 25 carrots of diamond.

Enough for my original project.

1.) Diamond replacement IHS
2.) Diamond Thermal Grease
3.) Diamond "thermal siphon skin"

Well ... I wasted an AMD 5200+ trying to remove the IHS ... So I went IP35Pro Abit and a 2.4GHZ Quad.

So I figured to make a good amt of diamond thermal Grease and test it on a core 2 and a pentium d dual core ... then also use it on my new Quad.

The Diamond IHS would have taken up 7 carrots ... I used that for the "extra" diamond thermal grease. My original test was to try and get a huge margin of temp reduction by reducing the stock and inefficient IHS heat bottleneck.

Now I am simply after definable degrees of improvement with the Grease and a large heat reduction with the skin.

Plus I wanted to make sure I had plenty for future GPU or Northbridge or other needs.

This leaves me about 12-15 Carrots for the skin.

Only a one time shot deal.

I hope the "jappaning" process works right. [Has thermal properties that won't insulate the diamond powders properties.

Epoxy resin would for sure ... jappaning can sag some ... though after the diamond powder is added ... I think it will be pretty thick.

[Elysium]

Dude, I really need some pics to visualize all this...

[Tech Geek Deluxe]

Well ... I opted for the "Jappaning" method.

I will do pics ... and post them after I have done my diamond thermal grease tests.

Diamond skin is irrelevant if the thermal grease doesn't show significant results. [Unless I show like 15 degrees C below the average out there ... then I will post pics of the diamond skin.]

Btw the jappaning worked very well. The high carbon rosin mix made adhesion perfect and left only what would make the diamonds adhere like a solid skin ... the rest of the mix ran off ... leaving perhaps an 80-85% diamond skin and most of the rest as carbon solids.

I won't know till I take a closer look ... right now my heat sink is at 390 F ... and in the oven.

I will post pics contingent on definable results. If the results are stupid ... then I have allready bored you and wasted time. If the results are really kewl ... then I will have at least some stats [and pics] to support the effort.

[Elysium]

Can't wait man!

[Tech Geek Deluxe]

Preliminary results are in.

Not good. There was no real change in the temps of the Pentium D

Temps were the same as with Arctic silver 5.

The temp recovery was considerably faster ... but I have no way to measure it.

Unless the temps improve with burn in ... like the arctic silver did and was ... I had 50 hours of burn in ... well ... then I'm pretty much done.

The skin idea ... well ... My Quad runs ridiculously cool at a solid ultrastable overclock from 2.4ghz to 3ghz. Could easily go to 3.5 stable ... but "ultrastable" suits me better.

There is no way of telling it had anything to do with diamond thermal grease or diamond skin however. I took pics but since I can't prove anything except I worked hard at it and spent additional money I didn't need to ... well ... perhaps another day.

I had hoped that as soon as I turned on my PC ... that I could feel the heat all the way up the heat pipes that had diamond skin ... but the whole thing runs so cool at 3ghz ... there simply isn't any hot areas at all.

I will post my specs ... speed, temps, voltages and you can surmise if I am lower or average in temps with similar builds. I suspect I am running about average.

The one thing I have noticed with the diamond skin and thermal grease ... the recovery from the heat on full load ... down to low temps is superfast ... [I used all of the last of my diamond thermal grease on the Pentium D and ran a thick skin over the whole base of the core of the heatsink. It was running 50 Idle and 61 Full load with lapped surfaces and Arctic silver 5 ... the sink was a mid priced copper Rosewill RCX-Z3 92mm Ball CPU cooler ... decent ... but not near the performance of a Typhoon or Gemini2

Temps before burn in with the diamond thermal grease and 1/2" skin up the base ... well ... the temps are exactly the same. The ambient was 7 degrees warmer however ... and there wasn't any burn in ... I don't think it will make any difference tho ...

I will play with it tomorrow when my ambient room temp is the same. I suspect I may get 1 degree after burn in and better ambient. Certainly not enough to repeat the experiment.

I have one PC left ... a 2.4ghz Core 2 duo with Lapped surfaces ... arctic 5 and a Typhoon on it ... I must have got a bad chip ... because it doesn't hold an ultastable clock over 2.8ghz. I can clock it to 3.6ghz ... but it runs too hot and isn't ultra stable.

It runs flawlessly at 2.6 ... solid at 2.8 ... but sputters a bit starting at 3ghz. At about 3.4ghz ... I can expect a blue screen once a month ... at 3.6 ghz ... a blue screen once a week.

One blue screen a year is about all I can stomach per PC.

I have found that my best results of overclocks are on a new install ... If my overclock makes the windows install hickup in any way ... then I tune it down so I have a perfect install. Found that if I do that ... though it is a hassle ... I have a perfect baseline for stability.

[fstroupe]

The ambient was 7 degrees warmer however ...

If you have decent case ventilation, it is shocking how much difference even 2F or 3F ambient room temp makes towards cooling the CPU. Not being a thermal engineer, I have no idea why.

Just last weekend during a CPU cooler review, lowering the room temp 2F lowered the CPU temp by 3C.