IC Diamond 7 Carat Thermal Compound vs Arctic Silver 5
After reading a few articles and tests regarding the new IC Diamond 7 Carat Thermal Compound (IC7) (here, here and here), I had to try it out myself. I've been using AS5 for a long time, and until now, I guess I haven't really appreciated how easy AS5 is to work with. IC7 consistency, compared to the rather oily AS5, is more like working with model putty. It reminds me of the ShinEtsu I used years ago. The manufacturer’s instructions states that you should apply a thin coat of IC7 on the IHS, but that is easier said than done. It tends to peel off leaving holes in the coat as you apply it. The place I bought it from suggests you should heat up the syringe before using it, but problem is, as soon as you place it on the IHS it cools down again. So I also had to heat up the CPU/IHS (using a hot halogen lamp), and I'm guessing the IHS temp was around 45-50°C when I finally was successful applying it. The instructions also states you should allow the coat to cure for 10-15min before mounting the cooler, which is ok, but it increases the risk of dust particles sticking to the coat, just something to keep in mind.
I tested the compounds on a Q6600, running at 3.3GHz and 1.36Vcore, open case and water cooled (bowed ApogeeGT, Eheim 1048 and a heatercore with a single 120mm 1300rpm fan). I only tested full load (prime95 v25.3 SmallFFT test). Ambient room temp measured at radiator intake, and it was within 0.1°C between the two tests. CPU temps stabilize within 5min in this setup, so I only ran Prime for ~15min.
AS5:
IC7:
Not bad, a 2°C (3.6F) improvement on all 4 cores
Too bad it's so hard to apply. Also the IC7 syringe, with the amount I used for this test, probably won't last for more than 2-3 mounts. Though I'm sure it depends on how good you are at applying the coat, as I made a bit of a mess. I'm probably going to restrict my IC7 usage to quad core CPU's, and more permanent builds, as this is too much work and also a bit more expensive than AS5.
I also wrote down the temps under FAH SMP, for those that are interested:
AS5, Core1/2/3/4: 58/54/48/51
IC7, Core 1/2/3/4: 55/52/47/49
I tested the compounds on a Q6600, running at 3.3GHz and 1.36Vcore, open case and water cooled (bowed ApogeeGT, Eheim 1048 and a heatercore with a single 120mm 1300rpm fan). I only tested full load (prime95 v25.3 SmallFFT test). Ambient room temp measured at radiator intake, and it was within 0.1°C between the two tests. CPU temps stabilize within 5min in this setup, so I only ran Prime for ~15min.
AS5:
IC7:
Not bad, a 2°C (3.6F) improvement on all 4 cores
Too bad it's so hard to apply. Also the IC7 syringe, with the amount I used for this test, probably won't last for more than 2-3 mounts. Though I'm sure it depends on how good you are at applying the coat, as I made a bit of a mess. I'm probably going to restrict my IC7 usage to quad core CPU's, and more permanent builds, as this is too much work and also a bit more expensive than AS5.I also wrote down the temps under FAH SMP, for those that are interested:
AS5, Core1/2/3/4: 58/54/48/51
IC7, Core 1/2/3/4: 55/52/47/49
0
Comments
Judging by the motherboard temperature the room was cooler during the second test.
edit: Also did you wait 200 hours before you took those measurements as arctic silver 5 requires for optimal efficiency. (Temps are typically 2-5C lower after you do, If you don't AS5 is less effective then AS3).
edit: Also on the preclean did you use something with at least 70% alcohol.
Thanks for sharing your findings, Isevald!
Yeah, that’s what I tried first. And it certainly helps, but it takes some time getting the coat on, and as the CPU/IHS is cold (room temp), it cools down quickly (almost instantly, it's thermally conductive after all) and starts to peel and tear off again. But I guess I will feel better about it after using it a few times and getting some practice. I will try heating it up even more the next time, as I have a feeling I used a bit too much last time (others seems to get even better results). But that's kinda my point too, with AS5 you almost can't miss, it gets pretty much perfect every time, and for that to happen you only need clean surfaces and a small blob in the centre of the IHS (no need to spread it out).
Yes, the AS5 had been on for more than a week, the IC7 for 2-3 hours (the IC7 manufacturer states 2 hours minimum cure time for peak performance)
The sensors are mislabled in Everest, I have no idea what the one labled "Motherboard" is for, but I suspect it's some function of CPU temp (or close to it, maybe a sensor placed in or near the CPU socket). And as you can see, the other temps (hdd and aux) were spot on.
The system had been running for more than a week (probably 2 weeks) with AS5, I ran the test before switching to IC7.
The ambient temps were 22.3°C for the AS5 test, and 22.4°C for the IC7 test (measured with a digital thermometer, with the probe placed directly in front of the radiator, at the air intake).
I use Artic Clean to remove thermal compounds
BTW, Overclockers.com has this thread linked for IC7 in their December 22, 2007 article.