Watercooling a CPU isn't exactly a new idea - websites covering this topic have been online for a long time. However, the watercoolers mentioned on these sites are "do it yourself" solutions - and building such a watercooler  takes a considerable amount of time and skill. Even though some retailers have been selling single parts for watercoolers, such as pumps or "water blocks", no complete, professional watercooling kit (including radiator and everything else you need) has been available until recently.

A few months ago, a Taiwanese manufacturer - Senfu – took the challenge and entered the market with a watercooling kit.  Now, it’s our turn to see how it measures up.

What you get

The Senfu watercooling system consists of three parts: The actual watercooler, to be mounted on the CPU, the water pump, and the radiator. Each part can be purchased separately. Senfu offers two different versions of the radiator - you can chose between a larger dual fan unit and a smaller single fan unit. Reviewed here is the dual fan version.

The performance

How well does the Senfu watercooler perform? Well, let's compare it to the two most efficient Athlon heatsinks, the Global WIN VOS32 and the Alpha P7125. Details about this test are described in the heatsink testing methodology article. The test CPU was an 0.25µ Athlon-650. The VOS32/P7125 results are taken from the February ‘00 Athlon heatsink roundup.

Test under real-world conditions

Cooler 
    Temperature
Senfu watercooler
39
29
Global WIN VOS32
44
32
Alpha P7125
YS Tech fans
45
32

  CPU temperature   Motherboard temp.
All values in °C.

Test with "simulator"

Heatsink 
    Temperature
Senfu watercooler
29.7
Alpha P7125
YS Tech fans
31.8
Global WIN VOS32
32.1

  CPU simulator temperature. All values in °C.

Additional Comments

With a regular heatsink, the CPU temperature usually stabilized after about 15 minutes. This was not the case with the watercooler, since water heats up very slowly. In the test under real world conditions, the CPU temperature continued to rise slightly even after the 20 minute test period. After a longer time, the temperature stabilized at 41°C - very close to the temperature reached with a high end heatsink.

So, just from looking at the CPU temperature, one might say that the cooling performance might not justify the additional cost and possible inconvenience of a watercooler . However, with a watercooler, the heat generated by the CPU isn't dissipated inside the case but, instead, totally removed from the case. As you can see, this results in a lower case temperature. As soon as Peltier elements get involved, this advantage will become more obvious.

The Heat Exchanger

Log in

Don't have an account? Sign up now