We don’t know if this detail was made on purpose or not, but in case of installing the Armada on top of the memory module with the corresponding heat sink, and therefore not using the clip clasps, the fans are free to be turned in such a way that the air can be better distributed between the memory modules.

And speaking about what really interest us the most… Let the testing begin!

We decided to take the memory module to a higher-than-normal work conditions  

We began then with the mounting of the probes.

Two probes were installed on the surface of the heat sink. This was aimed at getting the memory modules’ temperature – what will be measured actually is the temperature of the heat sinks and not that of the memory modules.

Had we stopped there, we would have made a mistake at the moment of measuring the temperatures, since if one side of the probe were in direct contact with the heat sink and the other side were under the air flow produced by the Armada, what we would get is a temperature result that wouldn’t correspond to the heat sink nor the air. Then, to guarantee that the reading belonged to the heat sink without the interference of the air flow, an insulating coat was applied on top of the probes. It is important to take into account that the insulating coat produces a reduction in the exchange surface area of the heat sink that will generate temperatures somewhat higher in case the heat sink’s surface area is totally clear even so, the result will be more reliable than if the probe is cooled in one of the sides with fresh air ;)

To run the tests, we used the usual hardware for temperature testing.

Motherboard                      DFI LanParty NF4 Ultra-D

Processor                           AMD Venice 3000+

Memory modules                OCZ Gold EL VX (2 X 512)

Power supply                     Sunbeamtech NUO 500W

The memory modules were taken to 3.2V with a 220 Mhz frequency and a 1.5-2-2-5 latency – enough to raise the temperature, isn’t it? :D

Like in other reviews, the tests were run with 32M Super PI while the evolution of the temperature was controlled during the entire process, before the beginning of the test until it stabilized to the temperature mark prior to the test. None of the habitual tasks, such as the e-mail client, anti-virus, etc., were interrupted to make the CPU work harder.