Review: Asetek WaterChill Xtreme 12V Pump/Reservoir

Testing

Time to test the cooling (or should that be pumping) performance of the Xtreme 12V Pump. System specs at the ready:

A fairly simple cooling loop, so not much resistance for either pump to deal with. The radiator is ample given that there's only the CPU block attached, so before embarking on the tests I wasn't expecting too big a difference in watercooling performance regardless of the pump used.

I didn't do any overclocking with the stock HSF, but did take the CPU up to 2.55GHz with 4x 255MHz HTT and the memory divider changed to keep DDR400 speeds. I took the voltage up to 1.5, adding more than this didn't help the overclock; I got a bad week Winchester, unfortunately. If you're interested, here's a screenshot from CPUz.

Idle temperatures were taken after leaving the system alone for 15 minutes or so to let temperatures settle down. Load temperatures were obtained using Stress Prime with small FFTs to stress the CPU. Tests were run again for 10 to 15 minutes, although usually I saw temperatures stop rising after 5, I kept the tests running to be sure. When I changed pumps I waited a couple of days before running the tests again to ensure the new thermal compound had bedded in. I ran the Xtreme 12V pump at its default speed, which is 3000rpm according to the software.

Ambient temperature was recorded with the temperature probe on my multimeter and included in the charts because room temperature did fluctuate slightly. CPU temperatures were reported through Motherboard Monitor. Obviously the accuracy of onboard temperature monitoring is always questionable, but in reality the relative change is what's important here.

Let's start with idle temperatures.

Nothing exciting here. The L20 seems to do a better job than the Xtreme 12V at idle temps despite a higher recorded ambient. Still, what's 1ºC?

Now things get a little more interesting. When using stock speeds, both pumps get the same cooling performance out of the circuit. Start putting some more heat into the circuit however, and the Xtreme 12V performs some 5ºC better than the L20.

Given that the radiator is only dissipating heat from one source, I'd expect that if more sources of heat were added, the difference in performance between the L20 and Xtreme 12V would grow. The higher flow rate and greater head on the Xtreme 12V should help it as the resistance in the circuit mounts up.

I was going to perform some tests demonstrating the change in cooling performance as the speed of the Xtreme 12V pump is changed. However, using a fully loaded, overclocked system, changing the pump speed didn't change temperatures at all. Perhaps with a more complex cooling circuit and more heat being put into it, pump speed would become more relevant.