We always use the same test method to measure the cooling performance of cases. We have come up with various test configurations and tried them to have finally reached a test platform that is fairly simple, but effective and consistent.
As is well known, every computer contains heat sources that cause the temperature in the case to rise when the case does not feature adequate ventilation options. The largest sources of heat in a powerful PC are the processor and the graphics card. Depending on the type, a processor can easily produce 90 to 100 watt of heat and with new high-end processors this can even go up to 130 watt. When overclocking a processor, this can increase even further.
The second major source of heat is the graphics card. A modern high-end card can easily use 250 watt, while models that feature two GPUs can use even more than that. A reasonable processor and graphics card combined can therefore easily use 300 to 350 watt of energy, of which nearly everything is converted to heat. Add to that the power draw of the motherboard, hard disk as well as the optical drive and the total quickly approaches 400 watt.
To be able to consistently produce 400 watt heat generation we have chosen to work with two controllable heat sources of 200 watt each. The heat sources of choice are two professional 200W incandescent bulbs. With an efficiency of less than 3%, a 200W incandescent bulb produces about 195 watt of heat effectively, and in a very consistent fashion. Our test configuration consists of a defective motherboard on which we have mounted two sockets in such a way that the first occupies a position roughly where the processor would be and the second at the location where the graphics card would usually be positioned.
By turning one or both lamps on we can simulate a system that produces nearly 200 watt and 400 watt of heat respectively. On the prepared motherboard we have also mounted a temperature sensor which allows us to measure the temperature within the case. After letting the temperature settle to a stable equilibrium, we measure this value at a load of 200 watt to simulate a mid-range system and 400 watt to simulate a high-end system. Both scenarios are carried out with the included case fans at 12 volt and at 7 volt. All measured temperatures are normalized at an ambient temperature of 20 degrees.
In the photo below you can see our test platform in use:
We test the noise levels in a soundproof box, in which we can detect all noise above 17 dB(A). Just to be clear: standard ambient noise is about 30 dB(A). In the case we install a Scythe Mugen 2 CPU-cooler, a Scythe Gouriki 700W power supply and an old Hitachi 80GB 7200 rpm hard disk drive. We test each case without the case fans running, in order to test the sound dampening ability of the case itself. Additionally, we test the noise levels when the case fans are running at low speed (7V) and high speed (12V).