Noctua NH-U9B SE2 Heatsink
Author: Jakob Barnard
Editor: Shawn Knight
Date: 06-30-2010
Provided by: Noctua

Applications used: CPUZ, MaxCPU, and Core Temp.

In testing I am comparing against the stock AMD cooler. The AMD cooler has five heatpipes and a modest sized heatsink. The testing methodology is to check temperature at stock speeds both at idle and then after ten minutes of 100% load. I will then re-run those same tests at a moderate overclock of 3.95 GHz. For the tests I disabled the fan speed control and ran it at 100%. The noise from the stock AMD fan was obnoxious and sounded like the case may take off. At full speed it develops an annoying pitch that wasn't covered up by the other case fans.

Stock cooler at stock CPU speed showed a result of 41⁰C.

I then ran MaxCPU to force 100% load for 10 minutes. After 10 minutes the temperature was 57⁰C. It reached this temperature after just a couple of minutes and maintained it throughout the test.

Next I performed a mild overclock using AMD Overdrive. I increased the multiplier from 16 to 19, the voltage from 1.4 to 1.45, and lastly the HTRef from 200 to 208. During previous attempts at overclocking I have found this to be about the safe point before coming close to instability. Core temp mistakenly shows the frequency as 3,816 MHz, however.

With just a moderate increase in voltage I saw only a 3⁰C increase in temperature during idle.

With max load on the overclock I saw a 5⁰C increase in temperature to 63⁰C. Really not too high, but I try to keep my machine under 60⁰C if possible.

Moving to the Noctua NH-U9B SE2, the first thing to mention about the test is the noise level. I actually cracked the case back open to make sure both of the fans were running. I didn't use the included voltage reducers because they simply weren't needed. The case fans covered any sound the Noctua fans might have made. This is extremely noticeable after using the stock cooler for about a month.

The Noctua at idle was slightly cooler than with the stock cooler, resulting in a temperature of 39⁰C.

After 10 minutes under load, the CPU reached exactly the same temperature as the stock cooler under load, but it took the entire 10 minutes to reach that temperature of 57⁰C. The stock AMD cooler reached this same level in just a few minutes.

Using the same overclock I noted the exact same temperature as the stock cooler at idle: 43⁰C.

I was a little surprised with the results of the Noctua with the CPU at 100% load. I ended up with a 3⁰C higher temperature of 66⁰C. It is important to mention that it took the entire 10 minutes to reach this temperature unlike the stock cooler which achieved it's maximum temperature after a couple of minutes.

Move ahead as we wrap things up with a conclusion.

  Sponsors Links

  Sponsors Links