Shin-Etsu MicroSi Thermal Grease Test

Like it says – Karl Williams

I ran the tests of this paste over the course of 2 three week periods. I allowed two weeks for the paste to settle, and then ran the main test, with the processor loaded for one week with Folding@Home (for Team 32, of course :)).

The package came in the mail and I opened it. My first reaction to the applicator tube was “Why is somebody sending me a heroin syringe?” My initial reaction to the grease itself was that Joe was right, and this stuff is like cold butter!

I expected a real challenge applying it correctly to the die, but I was pleasantly surprised, in that it wasn’t as hard as I thought it would be. It took a little bit of extra effort to get it to go where I wanted, but it wasn’t like spreading cold butter.

I’m not sure if it has a metal base for thermal transfer, like Arctic Silver/Arctic Alumina, so I wasn’t about to rub it between my fingers or anything to see what it felt like. Maybe I’ll do that if I can find a website that actually has what is in this paste. The main Shin-Etsu site doesn’t have that information – it just lists it as thermal interface material.

The test equipment is my own personal rig, which pretty much never goes down. The specs are as follows:

• AMD Athlon XP 1800+ at stock speed/stock voltage
• Asus A7V266-E RAID
• Alpha PAL 8045 w/ a Sunon 80mm Fan (unfortunately, I have no specs on it. It has a strange model number, and I can’t find specs anywhere. It’s plenty powerful, however)
• Enermax Whisper 431w Power Supply
• 1 Intake 80mm
• 1 Exhaust 80mm

The rest is kind of negligible. I’ll also be running my RAM (768 MB – 1 stick 256 MB PC2100 Micron, 1 stick 512 MB PC2700 AZENRAM) at CL2.5 for testing, due to stability concerns. I don’t believe that this will affect the testing at all. Both sticks are capable of CL2 at 133 FSB, but I’d rather eliminate any potential problems.

I took 3 measurements per day. All temperatures are from a Compu-nurse with the thermal probe touching the CPU die. First measurement was taken at about 1:00 PM, second measurement around 5:00 PM, third measurement around 11:00 PM.

This was due to several factors: 1) My sleep schedule is very screwed up due to work and 2) It will give my room time to normalize temperature with the rest of the house. My room is not very well insulated, so the temperature likes to fluctuate a lot.

My initial measurements (first run):

Ambient Room Temperature: 26 C
Temperature 3″ from CPU fan: 26 C

Both of those temperatures were taken when the machine had been off for approximately 12 hours. The ambient room temperature will fluctuate, but not much. It’s pretty constant when the door to my room is open.

CPU Idle Temperature (allowing for complete idle): 43.5 C

I allowed the system to boot, and then let the idle temperature settle. That’s the number that it settled at. This took approximately 40 minutes to happen.

My initial measurements (Second run):

Ambient Room Temperature: 21.5 C
Temperature 3″ from CPU fan: 21.5 C

Both of those temperatures were taken when the machine had been off for approximately 12 hours. The ambient room temperature will fluctuate, but not much. It’s pretty constant when the door to my room is open.

CPU Idle Temperature (allowing for complete idle): 40.5 C

I allowed the system to boot, and then let the idle temperature settle. That’s the number that it settled at. This took approximately 40 minutes to happen

After testing fairly extensively, I’ve come to the conclusion that this paste is better than the Arctic Silver II I was using, but not something to fall all over yourself to pick up. It’s not terribly hard to spread, but it is more difficult than something like Arctic Silver. It’s thicker and takes a bit more to get where you want it. It’s also pretty hard to clean up after if it gets everywhere.

I wouldn’t recommend it over Arctic Silver III, but it might be alright for RAMsinks; I don’t think it has any metal in it, which would eliminate shorting and possible fried cards.

Karl Williams