Fans are designed to move air and thereby cool things. So to compare fans, their ability to cool things should be compared. How to test a computer fan’s cooling ability? The most convenient way is to have it blow through a heatsink, preferably one with as little wind resistance as possible. The Megahalems was designed to allow a low-output (i.e. -- quiet) fan to produce a large degree of cooling. One design element that reduces its wind resistance is that the twelve ends of its six heatpipes are aligned in two rows in line with the airflow, reducing the cross section blocking the airflow to two 6mm heatpipes.
The other important aspect of a fan’s performance is how much noise it makes trying to cool things. Placing a sound pressure meter 10cm from the face of the fan will allow a standard noise comparison between fans.
The point of both measurements is not to produce absolute values, but comparative values. The comparisons allow you to choose between various fans. So, for example, the OCCT stress software is better for comparisons than Linpack. The OCCT produces a steady stress level that is not as intense as Linpack. This stable, moderate heating allows for better comparisons between fans.
Although an operator of a server farm would care about a few tenths of an Amp per fan, this is not an important issue for most computer enthusiasts. So the actual current draw of these fans was not measured. As for Voltage, all the fans were plugged into the same socket in the wiring of the same power supply unit to assure identical Voltage. Compared to your PSU, the Voltage will differ, so these measurements are of comparative value only.
Test system
CPU core 0 (hottest) temperatures were reported and logged at the nearest degree centigrade once per second. Ambient temperatures were logged at the nearest 0.1 degree centigrade ten times per minute. Averages were taken of the segment from 10 minutes to 40 minutes, to allow the system to reach equilibrium. Temperature Over Ambient (TOA) is the difference between the mean core temp and the mean ambient temp. For ambient temperatures between 19C and 24C, the TOA is stable from one measurement to another.
Sound pressure level (SPL) on the heatsink was recorded to nearest 0.5 dBA at 10cm, then converted to the equivalent SPL at 1 meter by subtracting 20 dB.
Each fan’s RPM on the heatsink was noted. The fan speed often differed somewhat from its speed in free air.
The first three charts show the TOA and SPL for the comparison fans on the overclocked system, arranged by fan manufacturer and fan name, by TOA and by SPL. The next three charts show a smaller group of fans on the non-overclocked system, where the cooler test provided a way to test the less powerful fans. Finally, to satisfy the curious, there is a chart comparing actual RPM vs TOA for 7-blade and 9-blade fans.
The Megahalems Rev. B was upgraded to Rev. C with parts supplied by Prolimatech.
