Testing (continued):
The charts below analyze the five power supplies over a thirty minute period while the system was run from the idle to the load condition described. As seen on the chart, a total of sixteen readings were taken at two minute intervals, with the first two readings in the idle condition, and the next fourteen readings in the load condition. The times are approximate, as load readings were not taken if 3DMark was in between load tests, and because only one rail could be read at a time. For example, I would connect to the 12V rail, wait for it to stabilize, take a reading, and then repeat the process on the 5V and 3.3V rail, making sure to get back to the 12V rail in about two minutes. The transition between graphics tests was generally just over a minute, and really didn't delay any readings.
Let's take a look at the results from each rail, starting with the 12V rail. The Lian Li unit started out very close to specification at 12.06V and the regulation was fairly good since it hit a maximum of 12.20V under load. The deviation from idle to load was 0.14V or 1.16%. The 12V rail results for this unit were well within specification, but the change from idle to load was the most dynamic.
On the 5V rail the readings for the Lian Li unit started just under specification at 4.98V, and under a load went as low as 4.92V. The deviation from idle to load was 0.06V or 1.2%. Regulation was once again good, but the values did start out the lowest of all units compared.
The 3.3V rail on the Lian Li unit didn't flinch and was right on the money. The display never registered anything other than 3.33V regardless of system load, and you can't ask for much better than that.
