Arctic-Cooling Super Silent 4Pro L - Testing the 4Pro L
(Page 4 of 5 )
I compared the 4Pro L on all three fan speeds to the boxed Intel cooler. I waited two full days after installing the 4Pro L before performing tests to allow the thermal paste to cure. Both coolers were tested on the same CPU and same motherboard, in the same closed case. (P4b 2.4Ghz, Abit IS7, Cooler Master WaveMaster). The side of the case was previously modified to fit an 80mm blowhole over the CPU area. The same 80mm fan was obviously used in testing the two coolers.
There has been a great deal of skepticism behind the validity of data pulled from onboard sensors. However, there are different methods by which a motherboard may obtain CPU temperature, some are better than others. Some boards use an actual thermal probe to read temperatures. This probe is located at the center of the CPU socket on the motherboard. The problem with this type of temperature detection however, is that the position of the probe may change when reseating the CPU which greatly alters the consistency of the data the probe gives. Further, the probe is not supposed to touch the underside of the CPU; it is in fact only measuring the temperature of the air trapped between the CPU and the motherboard. Thus there is credible reason to not trust temperatures pulled from this type of onboard sensor.
The Abit IS7 used in today's testing however, uses a very different method of finding CPU temperature. It is in fact calculated, not measured. The onboard I/O chip, "probes a voltage given by the CPU thermal diode, and then BIOS / hardware monitor reads this value and uses some mathematical way to calculate the approximate temperature." (This is not the same CPU thermal diode that shuts the CPU down automatically when it reaches 135 degrees Celsius.) Basically, so long as the same motherboard and same CPU is used in testing, data will in theory, be consistent. It may not necessarily be accurate, but there are so many variables in the environment that testing is being conducted in that accuracy of a thermal reading is really of little significance to the reader/end user. Precision in data is the absolute most important factor when comparing coolers.
*DevH Writer, DMOS, adds the following: "The reason why data pulled from THIS type of onboard diode can be inaccurate is because the P4 internal probe that is being probed is located at the edge of the die, in the cache area. The accurate one, which is used to determine when to trip the thermal shutoff, is in the area of the functional units."
Temperatures and fan speeds were recorded from winbond Hardware Monitor. Data was taken at 20-30 minute intervals. Prime95 was used to simulate load. Room temperature was maintained at 22 degrees Celsius.
Clearly, the increased surface area of the 4Pro L heat sink and its 80mm fan puts its cooling performance ahead of the boxed Intel cooler, even with the fan at its slowest setting. The medium fan setting provides the best cooling/RPM ratio, and is able to run quietly while still giving a significant performance increase over stock cooling. Lastly, the fan speed data shows the variable fan speed switch is working correctly.
The next step in evaluating the 4Pro L was to examine its noise level. Subjectively, I felt that that the 4Pro L on its medium setting was similar to the boxed Intel cooler, the low setting was quieter, and the high setting was louder. I decided to record five seconds of the 4Pro L running in its three different modes as well as the boxed cooler running by default. I used the microphone on a Plantronics .Audio 90 headset, placing it against the center of the fans to do the recording. Click the links below and listen to what the 4Pro L sounds like:
- Boxed Intel Cooler
- 4Pro L - Low
- 4Pro L - Medium
- 4Pro L - High
The boxed Intel cooler emits a very high pitched sound, but is about just as loud as the 4Pro L medium. And while the 4Pro L high's performance certainly does not justify its loudness, the low setting is almost inaudible.
Next: Conclusion >>
More PC Cooling Articles
More By KaoMAN
| Recommended by Dev Hardware |
|---|
|
/ 25 
