Commissioning and testing
Open heart operations are always tricky, as water and electricity are not really good friends. That's why we conducted the first density test with an external power supply. You never know. In addition, we wanted to test how this behaves with the two pumps.
In principle, one of the two pumps is sufficient to guarantee the necessary flow, but some hurdles and long distances have to be overcome, which makes the pump seem very strenuous, because we cannot operate it throttled. That, in turn, is something you can hear.
Since we supply the two pumps via a second system, we can regulate the voltage almost continuously and thus control the speeds and, of course, read them out. Operation has proved to be optimal at seven to eight volts, ensuring that both pumps run at the same speeds.
The close observation of the water and the use of a suitable body sound microphone have confirmed that there are no measurable interactions that could negatively affect this cascade. Once the train has set in motion, the pass is constant and almost swirl-free. Microbubbles were not an issue either. Only then did we commission the original system and first test ourselves step by step for maximum performance.
In the non-overclocked state, the water temperature after all components in front of the large radiator was at a soothing 35°C when graphics cards and CPU were running in the stress test. This is not a bad value in itself, because there is still enough air to go up.
After that, both graphics cards were overclocked to 2 GHz and the CPU to 4.3 GHz. The measured power consumption was in the total in the Torture Run just under 700 watts, of which an estimated 620 to 630 watts landed in the water as waste heat and had to be dissipated to the environment via the radiators. We have already written about the part that is delivered directly to the air.
We used the table opening without a glass plate, but with our special foil airtight closed so that the infrared measurements come as close as possible to the real use. The measured values are barely above 50°C at the top, and in most places they are even significantly lower.
With slightly less than 37°C in the lead-up to the first graphics cards and approx. 40°C after the CPU you can live comfortably and wait for the next Sahara summer. After all, these are already absolute worst-case values, which are far above the normally achievable temperatures.
The cooling performance is therefore sufficient and the required fan speeds are still low enough not to be perceived as annoying. However, it is not entirely silent here either. This is of course also due to the Riing fans used by Thermaltalke, which are PWM-controlled, but also in this version do not belong to the quiet representatives of their kind.
The originally planned Phobya fans, which correspond to the bionic eLoops of Noiseblocker, we unfortunately had to discard, because the slightly thicker radiators in this constellation simply need more static pressure, which the bionic fans do not need. can offer. It was quieter, but the water was already at about. 44°C, which was just too much for our taste.
The whole life is a single compromise and so it comforts us in the end at least that the new fans can at least score visually. It's not really loud, but it's not completely inaudible either.
Kommentieren