Application and Testing Methodology
For testing, I’ll be modifying Icrontic’s standard HSF testing methodology slightly. I’d encourage you to read up on it if you have any questions about our testing methods.
The hardware configuration is consistent with our HSF test bench with the exception of Noctua NF-P12 fans being used on the heatsink and case exhaust. Selecting an appropriate heatsink for thermal paste testing is crucial. Because the performance deltas between paste products are so small, a heatsink with a high quality base and very effective mounting system must be used. Noctua’s NH-U12P fit the bill perfectly. Of all the heatsinks we’ve tested to date, it is still my personal favorite. Noctua’s SecuFirm mounting system is simply the best out there and provides a consistently secure mount time and time again.
I will also be maintaining a more consistent ambient temperature for paste testing. For HSF testing, we maintain a temperature of 21°C and do not permit a deviation of more than +/- 0.5°C. For the purposes of this roundup, ambient temperature is maintained at 21°C and not permitted to deviate more than 0.2°C.
Some thermal pastes take upwards of 200 hours to achieve their maximum performance potential, like the popular Arctic Silver 5. Since I am bound by some time constraints for this review, I won’t be testing with Arctic Silver 5. I’ll be sticking with Arctic Ceramique as our baseline thermal interface. It performs very similarly to AS5 and does not require the same very long cure time. Arctic Silver claims that Ceramique will perform only “slightly” better once it has completed its 25-hour “break-in”, whereas AS5 can realize significant improvements over the first 200 hours. Whatever improvement is realized during the first day of use with Ceramique is likely minimal enough to disregard, which is why we have been using this material for all of our HSF tests to date. All of the paste products in this roundup are advertised as requiring “no burn-in,” so I will not be burning in any of these products prior to recording results.
Full load conditions are achieved using Prime95’s large in-place FFTs. The CPU is left for approximately twenty minutes prior to taking a temperature reading. I should also mention that because our monitoring tools can only measure ‘to the degree’ and because paste performance deltas as so small, I’ll be making a slight compromise. I will often see Core Temp jump between two temperatures, making it difficult to determine the actual temperature. In my HSF reviews, I usually make a call. If it is jumping between 45°C and 46°C at a relatively equal interval, I will record the higher of the two temperatures. For paste testing, I will interpret that as a 0.5°C interval. So if I see jumps between 45°C and 46°C at a relatively steady frequency, I will record a temperature of 45.5°C. This is definitely not a perfect measure, but it allows me to distinguish between two different pastes—one at a steady 45°C and another jumping between 45°C and 46°C. Clearly the one holding a steady 45° is the better performer of the two.
To see just how much of an improvement a quality thermal paste can have over a low-quality one, I threw another paste product into the roundup. I purchased a very inexpensive no-name silver compound back in 2003 that I recently found in my parts bin. This paste has most likely outlived its optimal shelf life. That said, I would expect the new, premium grade thermal pastes to outperform this old, no-name thermal paste by a measurable amount. If they are not able to, they are either poor products, or defective. Just how much of a difference it will make is to be determined.
Consistent paste application and HSF installation is critical when testing thermal paste products.
The CPU heatspreader and heatsink base was cleaned with 70% isopropyl alcohol prior to installation.
Although we have been applying thermal paste per Arctic Silver’s recommendations (single small bead or horizontal line depending on the CPU), I was concerned about the reproducibility of the results. Although it is consistent enough for heatsink testing, the small deltas between paste products demand much more careful application.
A liberal amount of paste was applied and then spread very thinly across the center portion of the Q6600 heatspreader (as depicted above). This type of application is generally not recommended by manufacturers as too much paste is used, but it allows for very consistent application. Optimal performance is not as important as consistent results when comparing thermal paste products.
As you can see, the paste is compressed and spread across the entire portion of the heatspreader above the CPU cores. Again, more paste than what is required is used, but the application is very consistent.
As you can see, we get a good imprint on the NH-U12P. Excess paste is pushed off of the heatspreader by the very healthy amount of mounting pressure exerted by the SecuFirm mounting system.
To ensure that I was able to get reproducible mounts using the NH-U12P, I conducted three separate mounting tests and obtained the exact same load temperature every time. I was pleased with the testing methods and feel confident that they are about as accurate as possible without moving to synthetic laboratory measurement.
Articles RSS