In my haste to make it to the airport, I neglected to take a picture of a sample mount, but the amount of TIM required is very little. It’s a pretty thick and sticky paste, but it’s not hard to work with at all. I used a dot in the center method for all three mounts and thought the results were “consistent enough.”
Which brings me to my next point…due to time constraints, I was only able to do 8 mounts total for this supplemental review. I decided to split them up in favor of MX-3, 5-3. It’s not a slight against Ceramique, I wish I had time to do 5 full mounts, but it is what it is. The data is still fairly solid and is a good approximate representation of how it compares to other TIMs.
In yet another moment of incompetence of this supplemental write-up….MX-3’s data has its own flaw. And it’s actually stems from a product flaw/oddity–MX-3 performs its best when very large quantities are used. It’s unlike any other paste I’ve used before in that I got better and better results the more I applied, even if I wasn’t seeing any visible difference once unmounting the block from the IHS (even IC Diamond stops performing better once you use enough to cover the IHS). From the first mount to the fifth mount, I increased bead sizes from ~6mm to ~10mm and each mount was better than the last. If Arctic Cooling had bothered to have any product instructions indicating this, this could have been avoided. If I had more time to test, it also could have been avoided. Either way, the data shows a strong correlation between increased TIM usage and improved temperatures and that’s an interesting point on its own. But it doesn’t show the true performance of MX-3 due to the ‘bad’ mounts when I was improperly using it. I have included a conservative performance projection for MX-3 when it’s properly used–in the individual mount chart, that’s the “expected” line and that is what all the asterisks are about on the tables and charts.
Now finally some results! First up, the big graph with all my data presented as conveniently as possible.
Note: “Concluding Temp” is the average of all temperatures once the TIM stays below .1% above final temperature. In the case of Indigo Xtreme, MX-2, MX-3, Ceramique, and Shin-Etsu X23-7783D, it is indicative of performance once broken-in/cured. AS5 and ICD are still curing at the end of twelve hours and I would expect their eventual temperatures to be lower after a full curing session.
Now that we have looked at the plotted results, let’s isolate the data into groupings of an individual TIMs and look at the specific mount data.
Arctic Silver Ceramique:
Ceramique has a pretty large break-in and is very similar to AS5’s; must be an Arctic Silver thing. In the first hour (especially the first few minutes) it performs horribly and then slowly eases downward into better temperatures. In my opinion, at the end of twelve hours, it is very close to finishing it’s break-in period. The ‘bump’ at 60 minutes is due to the moving average calculation no longer including the really bad temperatures from the first few minutes of load. The reason why only three mounts are done was covered above.
Arctic Cooling MX-3:
MX-3, unlike MX-2, actually stands up to the “non-curing” claim and barely has any variation over the 12 hours. The “expected” curve and asterisks and performance explanations are all covered above and is pretty important to read.
When the dust settles and you look at the data carefully, MX-3 is a pretty exciting paste. I expect it to perform really, really well, albeit slightly worse than the Shin-Etsu X23-7783D; but I expect it to be more readily available and have an even less noticeable cure time, which are two pluses. Arctic Cooling really should post some instructions regarding mounting with it though–clearly standard procedures are not sufficient as using an absurdly large 1CM bead actually netted the best performance. And the second largest bead (~9mm) netted the second best performance; in fact, there’s strong correlation between increased TIM usage and increased performance. Go figure.
As for Ceramique, it also notches a win in my book. But it’s only a win over its sibling: AS5. It has extremely similar performance to AS5 but is less expensive, easier to clean, easier to work with, non-capacitive and non-conductive, and has a slightly faster cure. AS5 and Ceramique really aren’t in the same league as the other pastes (let alone Indigo Xtreme), but I’m sure they’ll remain as popular as ever.
All in all, neither MX-3 nor Ceramique can dethrone either of the two performance winners from the main review: Indigo Xtreme and Shin-Etsu X23-7783D. I expect the Shin-Etsu to remain the stronger performer compared to MX-3, even when MX-3 is used to its full potential…and we can’t forget about Indigo Xtreme, which is nearly an order of magnitude ahead of both of them in terms of overall thermal resistance.