The Test System and Thermal Results
Testing for the CPU coolers is done with the use of RealTemp to ascertain temperatures, Intel Burn Test to deliver the load to the CPU and CPU-Z to verify the CPU speed and the voltage being used in Windows. All of the testing is done with an ambient temperature of 24.5-25°C and humidity is maintained to 35% sometimes less.
For the "stock" runs, it's more of a plug and play setup where the PWM of the motherboard is in control of the fans speeds for both the idle and load results. Speed Step is active and the processor idles at 1600 MHz and loads at 3500 MHz for the stock settings. I also set the memory to run at 1600 MHz for stock. As for the overclocked runs, I load the CPU at 4.5 GHz and idle results are obtained with 7.5V to the fans while the load run is set to deliver 12V to the fans. This allows me to gauge the lowest and highest fan ratings for my charts.
When I booted the PC and allowed it to rest for a bit, I was pleased to see RealTemp showing that the NH-L9i is able to keep the CPU at 26 degrees in its idle state. That is something the PWM on most tower coolers don't accomplish.
With the stock clocks running on the 2600K, I loaded up the CPU with a bit of IBT, and I found the NH-L9i is more than capable of keeping most CPUs cool. In tighter spaces it will need assistance from chassis fans, but even with the TDP limit, I could overclock this CPU on stock volts.
Now in my conversations with Noctua, they wanted me to be sure to follow the TDP guidelines and not abuse the cooler too much. Even so I went ahead and tried the overclocked profile I always use just to see what would happen. With the increase at stock the cooler only jumped two degrees. Once I let IBT lose, there was a thermal shutdown of the CPU, but the NH-L9i hung in there for close to 20 minutes before it just couldn't cope any longer. This is why I say there is some room to play around, but you need to tend to chassis airflow to make this happen.