Water vs Air CPU Cooling

I have been using my watercooling system, described in detail in article “Simple Watercooling System”, for more than six months now and it is enough to say that I am more than satisfied with it. Three months ago, I upgraded to ABIT/Duron system and have never used it with a conventional heatsink and fan cooling. It is obvious that the Duron 700 @ 1000 CPU is generating more heat than my old Celeron 466 @ 582, therefore needing more efficient cooling. The old cooling system is doing a great job, even better than I had expected.

I have been reading reviews and articles on various heatsinks for a while and found myself wondering if the watercooling is really performing well enough. According to some tests, there are heatsinks that are able to keep an overclocked Athlon or Duron CPU, with Vcore set to 1.85V, under 40°C! I found this very strange, even hard to be believed.

I am fully aware that different motherboards can produce various temperature test results, but still, I thought of making a small test comparison. A friend of mine uses a well-known FOP32 HSF for his Duron 650 on a Soltek motherboard. I asked him to bring it over to my place for some testing and he did so yesterday. I am bringing you the test results immediately.

First of all, a few words about the system and test environment:

The main components are an ABIT KT7A motherboard with AMD Duron 700 CPU, overclocked to 980 MHz. The VCore is set to 1.85V and the room temperature was 25°C at the beginning of the test and 26°C at the end of it. The case was opened during the testing. The thermistor on ABIT motherboard is located in the middle of the Socket, under the CPU:


I twisted it a little bit upwards to ensure a good contact with the processor and put a piece of sponge under to isolate it from the surrounding air. I know that this does not guarantee accurate readings, but I hope that it contributes the regularity a little.

The first test was made with FOP32 in place, with a small amount of thermal paste smeared over the CPU core. The lowest temperature reading achieved with processor at idle was 32°C:


Then I started CPU Stability Test 6.0 CPU warming and after nine minutes, the temperature reached 52°C!


It looked like it will not go any higher than that, but the computer became unstable – an attempt to start another program brought a notorious “blue screen” to the monitor. It was obvious that this heatsink was not good enough for overclocking this high.

After switching to watercooling, I let the CPU temperature stabilize at idle and VIA Hardware Monitor reading was 30°C:


It took about fifteen minutes for the temperature to stabilize after CPU warming was turned on. After 55 minutes of testing, the temperature was still at 41°C:


There was no need to let the test run longer than that. The difference was not only in 11°C lower temperature than with FOP32, but in total stability of the computer as well.

It is more than obvious that the water cooling is superior to air cooling. I am aware that this is not a way to measure the CPU temperature accurately, but it is definitely good enough to show the difference between the two cooling systems.

Petar Lavarevski – Yugoslavia