Diamond Grease

Anybody who is a true computer hardware enthusiast these days knows that keeping your CPU from overheating while pushing it to its limits is one of the most important concerns. Thermal compounds like arctic silver are the front line of that defense, conducting the heat directly from the CPU to the heat sink. Without the compound, small imperfections in the surface of the heat sink can render the heat sink almost useless and lesser compounds can actually insulate your processor. Thermal compounds that are available for purchase have remained unchanged for some time. So I descided it was time to reinvent my thermal compound needs with some diamond dust!

Update:
To my great amusement we have apparently drawn a lot of attention both positive and critical of our experiments here. It is important to maintain some clarity here in perspective. When we originally published article nearly 3 years ago there was not a single commercial diamond based product available. We were in uncharted territory with our experiments and research. It is possible and we acknowledge that our testing methodologies at the time may have been flawed, however under any and all scrutiny we have never taken the article down or changed our original findings because we feel it would be a disservice to our fans and our selfs to hide or disguise our findings.

We are excited and gratified at the amount of interest and research ongoing with this area of interest and continue to encourage both findings and research be presented in a plane and easy to reproduce method to the public to help further the DIY Enthusiasts ability to make DIY Thermal Compounds a reality.

The most popular type of thermal compound has silver suspended in silicone grease. Research has been done by overclockers.com demonstrating a far superior type of thermal compound using diamond. Silver has a respectable thermal conductivity at: 429W/m K. Diamond on the other hand has a thermal conductivity of 900-2320 W/m K. So worst case scenario we double performance, and best case is roughly a 5x multiplier.

Another nice advantage with diamond is that it is not electrically conductive. Silver based compounds will conduct electricity which makes a sloppy application of the compound a risk for shorting out your expensive hardware.

It appears that several people have come up with this idea independently, but we didn’t want to make that fact stop us from sharing our two cents on the matter as well. There are mumbling’s about commercially available compounds being in the works but we thought, “Why wait when we can have our candy now?” We are going to show you how to make your own diamond compound, and provide you with some of our own results for the improvements it provides in actual cooling performance. We will be comparing Arctic Silver to our home brewed batch of diamond compound. Pending some more testing and refinement we will be giving away a few “experimental samples” in a week or 2.

The Diamonds:
25 Carats (5 grams) of 60,000 mesh diamond powder available for roughly $30 on eBay or as raw material, sometimes used in buffing compounds.

It’s worth mentioning that the purity of the diamond powder is an important factor. While most industrial diamonds are pure, some less than honest peeps will mix in other materials. Check on the purity with your vendor.

The Grease:
Silicon grease compound containing: Polydimethylsiloxane and Polytetraflouroethylene. This is available at some hobby shops or your local chem shop. Make sure that the compound you get does not have a thermal conducting component like zinc oxide as this would provide an insulating value compared to our diamond.

The diamond powder is very abrasive and is a hazard if inhaled. You should mix it in a small container with a closable lid. To make our own mixer we will start by bending a paperclip into a stirring paddle. the depth of the stirrer should be as deep as the container you have.

I found a small ear plug container in my assorted parts bin to use for this but a film canister would work well to, and in hind site I think a small ziplock bag and just squishing it would work real well also. Poke a hole in the lid for the stirrer and insert the bent paperclip .

Safety is king in the project. I thought it was funny i got this little branded reminder on my mixer.

Now we just add our ingredients. At a minimum I recommend using a good dust mask when adding the diamond powder. We are mixing 13 grams total of our compound, 5 diamonds, 4 grams Polydimethylsiloxane, 4 grams Polytetraflouroethylene.

We want to use our little stirring tool here for rough mixing only. We are just trying to knock down the dust this would make. I ended up using a small paint scraper to do the final blending. Make sure to stir and blend very thoroughly. You may feel like the geek equivalent of Bob Ross, mixing your compound in virtually the same way you would mix oil paint for your happy PCs. Your compound should be a gooey consistency. It should be almost runny, and should hold its shape, but you don’t want a clay-like consistency.

We had a few left over empty thermal compound syringes lying around that I decided to fill up. Loading the used syringes is easy; just use a flat head screwdriver to load it up. You can pick these up on the web for about 20 cents each and it’s a nice clean way to store your compound.

After dispensing my diamond compound onto the CPU I used an unused case badge I had lying around to apply an even coat of the grease. One thing I learned from trial and error is that if your compound is too thick you will have problems with the layer being too thick. The actual layer of the compound should be very thin. Before locking the CPU into the motherboard twist it into the compound a bit to insure good coverage and contact.

The Test Rig:
The test machine us running Windows XP 64 bit. We are using one of the hottest (thermally speaking) Pentium D 3 GHz (over-clocked to 3.4 GHz) to really test our thermal conductivity. The heat sink is the Arctic Cooling Freezer Pro 7. This may not be top of the line, but it is in that upper group where a good thermal compound will make a big difference. This heat sink has a copper block (approx 1.5oz) in direct contact with the CPU. Our rig has an Asus p5wd2 premium with 2 gigs of Kingston ram. We are interested in CPU heat so we went with a fairly old video card to put as much work as we can on the CPU. The card we are using is an MSI 6600. We are using 3D Mark and pc mark by Future Mark with the tests set to run completely in software mode.

For our tests we use a combination of the onboard thermal reporting system as well as external inferred thermal testing equipment. Our reported temperatures below were the maximum detected temperature.

Tests Results:
We have results for our first batch (8 grams compound with the 25 Carats of diamond) and our final batch for comparison. Even the thicker compound provided some improvements, but the right consistency does make a big difference.

System Idling System Max load Arctic Silver with a fresh application 42c 57c Arctic Silver with 2 week cure  39c 54c Diamond Grease with a fresh application 29c 38c

Summary / End Results:
While this obviously shows that the diamond compound is effective, the question I have rely had to think about is “is there a sustainable market for this”? With the amount of money that people spend on high end all copper heat sinks alone I believe there is. But a complex part of the problem of mass manufacturing is the purity of the diamond. If I am using a diamond compound I want to get the maximum amount of cooling out of it! I don’t want diamond and quarts mixed for example. Making your own thermal grease is a great option until there are better options, and its a lot cheaper too!

Update:
To my great amusement we have apparently drawn a lot of attention both positive and critical of our experiments here. It is important to maintain some clarity here in perspective. When we originally published article nearly 3 years ago there was not a single commercial diamond based product available. We were in uncharted territory with our experiments and research. It is possible and we acknowledge that our testing methodologies at the time may have been flawed, however under any and all scrutiny we have never taken the article down or changed our original findings because we feel it would be a disservice to our fans and our self’s to hide or disguise our findings.

We are excited and gratified at the amount of interest and research ongoing with this area of interest and continue to encourage both findings and research be presented in a plane and easy to reproduce method to the public to help further the DIY Enthusiasts ability to make DIY Thermal Compounds a reality.