X-bit labs

Fan, Cup and Copper Pipes: Zalman CNPS8700 LED and Thermaltake V1 Processor Coolers (page 10)

Category: Coolers

by Sergey Lepilov

[ 08/30/2007 | 12:09 PM ]

Thermal and Acoustic Performance

Before discussing the results I want to tell you why I replaced the quad-four CPU with a dual-core one. It was due to the Thermaltake V1 that proved to be unable to cope even with our slightly overclocked Core 2 Quad (running at 3200MHz, 1.45V). In less than four minutes after the start of the test OCCT would report an error and abort (click here for details). I did not reduce the CPU frequency or voltage because the monitoring graph, which showed a straight line going at an angle of 45 degrees up, indicated that the problem was not about a too-high CPU frequency. Moreover, I had already learned that even cheaper coolers, let alone top-end ones, could cool the Intel Core 2 Quad at 3300MHz.

Of course I checked out the contact between the Thermaltake V1’s base and the CPU heat-spreader, but it was perfectly normal. Then I tested the cooler in two positions: with the heat pipes oriented vertically and horizontally. I got the same result: the error occurred with a 10-15 seconds difference in time. I even replaced the mainboard but this had no effect on the CPU temperature under OCCT. I suspected a defect of the particular sample of the cooler and requested another sample from Thermaltake but it showed exactly the same results. Quite frustrated I tried to check out the cooler on Intel’s dual-core model and was surprised to see it do no worse than the Zalman CNPS9700 LED! After all the tests, reinstallations and replacements I did with the Thermaltake V1 on my Intel Core 2 Quad, I think this cooler does not suit for cooling overclocked quad-core processors although this CPU is on the compatibility list.

That’s why I tested all the coolers on an Intel Core 2 Duo E6400. The highest frequency was determined by the weakest cooler in this review (in a closed system case). It was 3500MHz at a core voltage of 1.5V. The test results are listed below.

Breaking no records, the Zalman CNPS8700 LED and the Thermaltake V1 keep the CPU temperature at about the same level under peak load and are both slightly weaker than the more expensive Zalman CNPS9700 LED. Note that the Thermaltake V1 does not depend much on the fan speed when tested on an open testbed. Equipped with two fans, the Thermalright Ultra-120 eXtreme is beyond competition as was to be expected.

The following table shows the temperature of both CPU cores, the reading of the mainboard’s CPU sensor, and the reading of the mainboard’s sensor proper.

Note that the Zalman CNPS8700 LED copes best with cooling the near-socket space. This stands to reason since it is the single cooler in this test that directs its airflow towards the mainboard rather than along it.

Next I found out the highest CPU frequency I could achieve with each of the coolers. This test was performed on an open testbed. Here are the results:

These results make me confess that the Thermaltake V1 is more efficient than the Zalman CNPS8700 LED and even than the CNPS9700 LED. Note that the stable CPU clock rate was only 24MHz higher with the Thermaltake V1 when it worked at the max speed, but its noise grew considerably.