Installation & Testing:
The four-screw mounting design of the NH-L9i made it a breeze to install. Since my case had a CPU backplate cutout I was able to remove my old cooler and install the NH-L9i without removing a single component from my case. Although the NH-L9i is designed for small cases where CPU cooler clearance is at a minimum, I'll be testing it in my SilverStone TJ08-E mATX case just so I can have some more competition to put it up against (testing it the other way around, in a small case, would mean that most of the CPU coolers I have laying around wouldn't fit in the case to be tested against). After testing is done in the mATX, I'll move it into the FT03-MINI ITX case just to see how it fits.
Don't let size deceive you! This makes for a powerful little system with the GPU is installed too.
There's two main areas I test CPU coolers for; performance and noise level. Performance is a gauge of how well the cooler will cool the CPU in idle and load situations and noise level testing is a simple observational test to see how the cooler's operational noise fares against the competition.
CPU: Intel Core i5-2400
Motherboard: Asus P8Z77-I
Memory: 2x4gb G.SKILL RIPJAW X
Storage: OCZ Vertex 2 SSD, OCZ Vertex 3 SSD, OCZ Agility 3 SSD
DVD+/-RW: Silverstone SOD02
Video: eVGA GeForce 280 GTX
Sound: Line 6 POD XT Pro USB Audio Interface
Power Supply: Silverstone ST55F-G
OS: Microsoft Windows 7 Ultimate x64
To performance test the NH-L9i I'll be recording CPU temperatures after 30 minutes of system idle and again after 30 minutes of system load. The system will be brought to 100% load using AIDA64 Extreme Edition. Before reading my results, remember that the NH-L9i is in a category all it's own and is at least half the size of all the coolers it was tested against. Compact coolers like this are a fairly niche market and I don't have any directly competing models here to test it against, so with the results below you can at least get an idea of where it stands against some performance tower coolers.
During the idle temperature test, the NH-L9i scored surprisingly well in my test setup, with idle CPU temperatures in the lower half of the chart above (although the fluctuation between the lower two thirds of the chart is pretty minor to begin with).
As I expected, during load testing the NH-L9i wasn't able to keep up with the higher end coolers that have been reviewed before it. Temperatures were still well within the safe range for the CPU though, and were achieved under extreme system load conditions that most users likely will not put this cooler through (given that most users will buy it for small form factor or HTPC setups, which usually aren't built for extreme high load computing conditions). But with that being said, I'd still be comfortable gaming on this cooler with my setup because the stress test pushed the entire system harder than an average gaming session would.
Noise Level Testing:
I have nothing but high expectations for the noise levels of Noctua coolers because they've always been some of the quietest to cross my test bench. With the smaller fan on the NH-L9i I was expecting something on the louder end of quiet operation given the small fan size. Smaller fans need to spin much faster than large fans to move the same amount of air, and faster fan speeds equals louder operation.
Lucky for me, I had recently reviewed another CPU cooler with a 92mm cooling fan, the Xigmatek Loki. Although the cooling fan on the NH-L9i could be noticeably heard in my system, it wasn't loud by any means. The system is already optimized for ultra quiet cooling, so saying that I could hear the NH-L9i slightly above the rest of the fans just means that it wasn't the most quiet of all the quiet fans. By no means was the NH-L9i as loud as the Loki - night and day difference there, but also a ~20 CFM airflow difference too. As a silent PC enthusiast, I had absolutely no complaints about the noise levels generated by the NH-L9i. When the system was switched to PWM fan control, or the L.N.A. adapter was installed, the fan had no problem blending in with the rest of the system and not being noticed during regular operation (it would still peep up during heavy load scenarios though).