The results of our cooling efficiency tests are given on the following diagram and in the following table:
As we can see, the new Zalman CNPS9900DF looks quite modest against the background of the Thermalright TRUE Spirit 140, which costs only half as much. At the maximum fan speed, TRU Spirit 140 is 3°C ahead in peak CPU temperature, but most importantly, Zalman CNPS9900DF wins only 1°C from the Thermalright cooler, even if the fan speed of the latter has been reduced to 810 RPM. While we gradually lower the fan speed for the Zalman CNPS9900DF fans in 200 RPM increments for the 135 mm fan and in 100 RPM increments for the 120 mm fan, the cooler efficiency lowers smoothly until we hit 1000/800 RPM. However, at the next increment the maximum CPU temperature increases by as much as 5°C, and at the next one – by another 6°C and reaches the critical maximum for the thermal throttling to kick in. In other words, the cooling efficiency of Zalman CNPS9900DF proves highly dependent on the fan speed in the 1000-800/800-650 RPM range, while at higher rotation speed its efficiency doesn’t increase dramatically.
I would like to remind you that we performed our cooling efficiency tests with the CPU overclocked to 4.25 GHz with 1.35 V Vcore instead of the traditional 4.375 GHz and 1.38 V Vcore. We were forced to deviate from our standard overclocking levels, because Zalman CNPS9900DF was unable to cope with this overclocking of our six-core processor even at the maximum fan speeds. We tried several times and checked different cooler positions on the CPU, but nothing helped:
And since Zalman CNPS9900DF failed to pass the “minimum qualification requirements” for our summary diagrams, we will not include it there today. So, let’s move on to the noise tests now.