NZXT Kraken Guide: How to Optimize for Silence and Performance

This article was written by Steve Burke from GamersNexus

NZXT Kraken coolers allow their owners to control two aspects of cooling and noise: pump speed and fan speed. By giving both of these controls to the user, it becomes easy to reduce noise levels while also maintaining a good balance of CPU cooling performance. Our testing at GamersNexus has long looked at multiple fan RPMs when reviewing cooler performance, and it’s helped us come to find an ideal spot for Kraken coolers to sit.

First, note a few things:

  • Different coolers will require different fan RPMs for peak noise/cooling efficiency, as the smaller coolers (X42) will need a higher RPM to keep up with the larger coolers (X62).
  • Different CPUs have different needs. Our testing was done on an overclocked 5930K generating about 130W of heat. If your system has a higher TDP CPU, some of these rules will not apply as well.

We’re testing at three primary RPMs: 1050, 1500, and max RPM for the cooler. The X42 is again an exception, as its max RPM is ~1700, so testing both 1700 and 1500 is a bit superfluous.

Starting first with temperature levels, we found the following:

And, just to correlate versus noise, here’s the counter to the argument for just boosting fan RPM alone:

How much noise is tolerable is entirely subjective, but we find that, generally speaking, most people are OK with 40-45dBA of noise from their coolers.

If that’s the case, a configuration of the following would achieve 40dBA while still offering excellent cooling:

The Kraken X62 is the most forgiving in the noise-cooling department. The cooler’s larger size and larger fans enable the fans to spin more slowly and still maintain good temperatures on our 5930K. This is seen even at 1050RPM, above, where the X62 holds at 37C delta T over ambient. The max RPM with max pump speed puts us at ~34.6C dT, which isn’t all that much better and is about ~13dBA higher (and remember: the decibel scale is logarithmic, so ~10dBA is a roughly a perceived doubling of noise to a human). The 3C gain for ~2x perceived noise increase probably isn’t worth it for most people, though it’s nice to have the option. For most, we think the X62 would do well at around 1050-1200RPM while maintaining relative silence. It’s likely that other system components will be more noticeable at this point, like the GPU or PSU fan.

The X52 isn’t far behind, and sits at 40C dT over ambient when at 1500RPM (for 46dBA – not bad), though our 800RPM test doesn’t do as well at 55.6C dT over ambient (but is dead-silent). Again, a 1050RPM value would be acceptable, though ~1200~1300RPM would sit at around 40dBA and offer good thermal performance.

The X42 has to fight a bit harder. At 1050RPM, the X42 sits at 46C on our overclocked 5930K, so we might instead recommend considering ~1300~1400RPM for users on hot processors. That said, our 5930K has a high thermal output; if running the X42 on a cheaper CPU (reasonable) with a lower TDP, it’s possible that you could get away with 1050RPM – it’s still cool enough, after all. Once accounting for higher ambient temperatures or high internal case ambient from other components, though, it’d be better if the fan speed were closer to 1300~1400RPM. This is still relatively quiet, too.

As for pump speed, we’ve found that there’s not too big of an impact to noise with regulating pump speed. Some users may be more sensitive to pump noise than others, in which case we’d recommend trying a ~90% speed first. This will often reduce most potential pump whine. After that, try stepping down in very gradual increments. The “silent” pump speed option does produce different (higher) thermal results than full speed, so we prefer to recommend creating a custom curve that’s a bit more aggressive than the silent curve, but still noise-conscious. This is best done by the user, considering it’s entirely subjective from here on.

Hope that helps!