Noise Testing and Conclusion
Sound pressure readings were made with a noise floor of 34db in the room. All measurements have a +/- 1.5db margin of error, which is the limitation of my sound level meter. The meter was placed 12" from the front of the test bench for all results.
First we'll look at the Haswell system's 4770K:
It’s worth pointing out the pump noise value on the chart. Pump noise is not often reported when discussing cooler noise, but is the differentiating factor when comparing air to liquid solutions. The noise of the fans often has a different character than a pump, and pump noise can be very noticeable, even inside of a case. The Cooler Master Seidon 240M, which had a very strong showing in these tests, also had the loudest pump of the group. It might look trivial, but the 0.8db increase from the Seidon 240M over the next loudest pump was actually quite noticeable, and the character of the noise was different enough from the fans to make it hard to tune out.
Now the bigger challenge – the Ivy Bridge-E system's 6-core, 130W 4930K:
The H105 generated quite a bit of fan noise at this review's most demanding overclock with 52db, which made the H100’s result quite impressive in comparison given its steady 42.2db at the “high” setting. (The H100 was also the quietest across the board when set to “low” fan speed, though the thermals here with the overclocked loads were not competitive.) The H75 exhibited identical pump and fan noise to the H105 at idle, not surprising considering their similarities, but the H105 was actually much louder at load with the same fans. My impression is that the airflow through the radiator in the push configuration is responsible for this disparity, as the H75 is using push-pull with its two SP120L’s.
The Cooler Master Seidon 240M reached volume levels reminiscent of a GPU blower at its max fan speeds, and even so the pump was quite noticeable as it has a distinct lower frequency. The Noctua NH-D14 needed maximum fan speeds to cope with the overclocked 4930K, but it was easily the quietest solution tested at stock speeds.
The Corsair H105 is a solid performer with tremendous upside. It was able to compete with every other cooling solution at lower fan speeds and noise, but during the most grueling tests it really showed some exceptional cooling performance. The thicker radiator design Corsair has implemented here really pays off, and the H105 is capable of keeping some very high processor overclocks nice and cool.
The tradeoff here is higher noise levels, but this was only noticeable with overclocked loads. For a high-performance system with a processor such as the 4770K or 4930K used in this review, the H105 would provide quiet performance at stock speeds, while leaving a lot of thermal room for overclocking if desired.
The H105 is well-built and easy to install. The SP120L fans are excellent performers, and even at high RPM’s the fan noise retains a slightly softer sound signature that is not unpleasant. The pump noise was noticable only at lower fan speeds, and even then would be reduced considerably depending on the enclosure. Overall, the only detractor to the cooler is price, as the $119 investment might be tough given the excellent performance of a cooler like Corsair’s $79 H75.
Bottom line, if you’re looking for a very high performing solution, particularly to cool overclocked CPU loads, the H105 is going to be hard to beat. Recommended!