A test showing heatsink fan rpm variations for different motherboards.–Joe
SUMMARY: Depending on the motherboard brand and power supply, you might see differences of about 4% in fan rpms for the same heatsink.
One thing that I have consistently noticed while using different motherboards to test heatsinks is that the Iwill KK266 and KK266+ give better results than the ABIT KT7 and Iwill BD133. Obviously, different motherboards have different secondary heatpath effects, so that explains some of the variation.
But not all. What I noticed recently is the voltage differences among the three motherboards that I routinely use for testing. A belated light went off – this HAS to effect voltage at the fan header, which means the same fan is going to run at different speeds with exactly the same setup.
I took a 60 mm Everflow fan, #F126025BH, 0.17 amps, and tapped into its 12 volt lines so I could monitor voltage with a multimeter. I then ran the fan in free air while recording power supply +12v, fan header voltages and rpms in BIOS; I used a Glaciator for the heatsink and recorded its rpms also. The Power Supply is an Antec #PP352X rated at 352 watts.
Power Supply Voltage
Fan Header Voltages
12.22, 12.25, 12.25, 12.27
Everflow 4682, Glaciator 5851
12.16, 12.18, 12.19, 12.14
Everflow 4623, Glaciator 5818
12.05, 12.10, 12.07, 12.04
Everflow 4615, Glaciator 5748
The difference from the lowest to highest power supply voltage is 5%; this translates to an increase in the Everflow’s rpms of 1.4% and the Glaciator’s of 1.8%. Now this doesn’t look like a lot, but with a cheapo 250 watt power supply delivering 10% less, the difference in fan rpms could be as much as 4 ½%. Under load, it could even be more.
On top of all this is the production variation in fan rpms, usually +/- 10%. It is entirely possible that a heatsink with a fan rated at 6000 rpm could deliver something like 5200 rpm in one system and 6800 rpm in another, due to the “luck of the draw” and power supply variations. Obviously, cooling performance will be affected.
Some motherboards will deliver better heatsink performance due to higher voltages at the fan headers. In addition, it is not inconceivable that such motherboards are inherently more stable due to higher voltages delivered to its component parts.
I would not be surprised that overvolting a motherboard to something like 14 volts is possible and could lead to better peformance, although the probability of blowing it up is more likely – not something I would do.