Results
The raw data can be seen in the tables below. We tested the fans at 3 different speeds dictated by the voltage limiting accessories/features that come standard with each product.
We have also indicated the subjective noise level for each product at the specified speed/setting. The testing was done in a quiet environment with minimal ambient noise and an open case.
Temperatures were recorded with the ambient temperature and whilst we tried to keep the ambient temperature at a standard 20 degrees, it did vary for some of the tests so we have compared the delta temperatures for the most part in the interests of fairness.
Results When “Pushing” Air Through H100 Radiator
Delta Temperature Graph
Full Temperature Results
Our tests showed some common trends in that as you would expect, the higher the fan speed, the better the temperatures.
The corsair fans were better performing with an overclock but the price for this is the acoustics. The corsair fans are unquestionably louder than the Noctua NF-F12s. We found the H100 stock fans to be loud at full speed and quite noticeable at medium speed. The H100 on low speed wasn’t exactly quiet either – we noticed it when sitting a couple of meters away from the test rig but it is more likely to blend into any background noise. The performance differences between the corsair fans in high and medium settings was minimal with the noise difference being marginal in favour of the medium speed setting. For general use, if you were sticking with the corsair fans on a H100, we would suggest giving the Low speed a go and watching your temps.
The NF-F12 had slightly higher temperature readings when overclocking and using a constant speed but the acoustic difference was quite obvious. With 2 fans at 1500 rpm, you can hear the NF-F12s but the noise isn’t as loud as the H100 stock fans on either Medium or High settings. When you step the NF-F12s down to 1200rpm via the Low Noise Adapter, you can notice the difference as the fan isn’t obvious but it’s still audible. At around the 800-900 rpm mark, the fans are much harder to pick out from ambient room noise. 800-900 rpm is not a stock speed so we didn’t test it but this is where we would be looking if we wanted to lock in a constant speed.
The blue PWM line on the graph is consistently higher than the other results where the fan stayed at a constant speed for the whole test. Due to the variable nature of the fan profile, the CPU was allowed to get warmer by the software fan controller. That said, the fan was almost inaudible at idle and only kicked up when it needed to – for general use and with a mild overclock, this is where we think the NF-F12 shines.
Results When “Pulling” Air Through H100 Radiator
Delta Temperature Graph
Full Temperature Results
Overall, the fans tend to work better when pushing air through the radiator but the performance is still acceptable in a pull configuration. We had the case open and on a bench so case temps were unlikely to influence the results with regard to the intake/exhaust argument.
In Pull configuration, we saw the same trends as push with one exception being the high overclock on the NF-F12 1200rpm test. we had a higher than usual ambient temperature that day and also noticed the delta temperature was on par with the 1500rpm temps. The corsair fans also seemed to perform better with our i7-3770K CPU at stock clocks in pull than the NF-F12.
We didn’t notice any difference in acoustics between running the fans in push or pull operation, the differences were limited to thermal performance only as far as we could tell.
