Another long winded post. I'm sorry.
So after doing some research, I can pretty safely say that the inventgeek results are BS. I'm not saying that diamond thermal grease doesn't have a valid place among the other thermal greases. There are just too many illogical steps in their process:
1. They used Polydimethylsiloxane and Polytetrafluoroethylene as there suspension fluid. Problem - Polytetrafluoroethylene is a solid. It is has no use within the thermal grease. In fact, silicone greases that come with PTFE in it have it in the form of non-uniform particles with sizes up to 1 mm. This would negate the use of micron size diamond particles.
2. They mix their diamond powder and grease by hand. Problem - that would introduce lots of air into the grease in the form of micro-macro bubbles. That would severely reduce the effectiveness of their grease. In fact, once the grease heats up, the bubbles would expand and push more of the TIM out and reduce performance even more. You would HAVE to have a degassing step in order to have a decent thermal grease.
3. Measuring the temperature of your heat sink as some sort of measure of performance. Problem - your heat sink temperature is directly related to the wattage output of your cpu and the temperature of the ambient air. It is no way related to the temperature of your cpu. This diagram will show you what I mean. Assume you have a cpu producing 60 watts of heat at thermal equilibrium:
60W -> cpu -> TIM -> Heat sink -> Air -> 60W
i.e. 60 watts in, 60 watts out. no heat stored or created or destroyed at equilibrium.
The amount of energy that the heat sink dissipates to the air is dependent on the temperature difference between the heat sink and the air and the amount of air pumped through.
60W = (some coefficient based on heat sink design) x (T_heat sink - T_air) * (volume flow-rate of air)
The 60 W is fixed by the cpu output, T_air is fixed, and the volume flow-rate of the air is fixed by the fan. That means the equilibrium temperature of your heat sink is set no matter what TIM you use.
The better the TIM, then the closer your cpu surface will reach to the heat sink temperature. If you have terrible TIM, your CPU temp will have to be really high in order to transfer that same 60 watts through your TIM.
As you can see, nobody that knows what they are doing would take temperature readings of the heat sink.
Having said all that, there is no reason why you can't use a diamond suspension to cool your cpu. Here is what you need:
1. Diamond powder of the smallest particle size you can get.
2. Pure silicone grease or any other viscous liquid with high thermal conductivity.
3. A vacuum chamber to remove any air from the mixture.
I have my 100,000 mesh diamond powder already. I bought some silicone grease from autozone today.
Now I'm in the process of constructing my own vacuum chamber. I have an empty fire extinguisher that I'm going to cut in half as my chamber. I have a food sealer vacuum pump. I just need a vacuum gauge, some tubing, and plastic plate and I will be set to go.
As soon as I get that setup, I'm going to make a few batches: a "perfect" batch, one without degassing, one without degassing and includes PTFE.
Then I'll setup a test using an electric resistance heater and a heat sink and test how the batches perform compared to AS5, ceramique, alumina, pure silicone grease, and industrial thermal grease.
And yes, I am unemployed at the moment.
Edited by freddyman - 8/22/09 at 3:32pm