To determine the PSU behavior I decided to use the a high-power system in three modes of operation: stand-by mode (shut-down), idle (OS, no activity) and under full load (full system load). The maximum load is achieved via simultaneous operations from AIDA64 stress test on CPU/FPU/RAM and video. As a means of measuring I used an Energy Monitor, which will register the power consumption for the entire machine.
Testing the efficiency is not necessarily an easy task, especially since I do not have an accurate way of measuring load. Yet by using stock components and bringing the machine at about 689Watt requirements, I was able to at least check the "85Plus" cert they plastered on the box.
Meanwhile, for voltage stability the test is quite simple. I kept the machine in idle and full-load for 6-hours each, and logged voltage values at intervals of 5-seconds in a log-file, which I afterwards evaluated. Using the log, I was able to determine whether fluctuations occured during the operations. To ensure the accuracy of the results, added measurements were made using a multimeter on the ATX connector on the motherboard during operation.
When it comes to efficiency, FSP has listed Hyper 700 as a "Meet 85 Plus" sticker on the front of the box. This isn't the standard Bronze / Silver / Gold / Platinum / Titanium certification system, but instead a requirement by the EU for system builders. In order for a system-builder to use a PSU in the perspective system, said power-supply needs to pass an 82-85-72% @ 230V input check-up.
By using a load of 689Watts, which counts for a total of 98.4% of the power-supply's rated output. At this load, the power supply required 794 watts from the socket to produce the needed power required by the test-bench. As such, the efficiency results suggest an 86.7% rating, which is not quite shabby. Of course, I also have to take into account some error margin, but even with them taken into account, the figure suggests figures close to those on the box.
During the test, the voltage values didn't register any voltage irregularities. For our 6-hour continuous load test, the power-supply was always stable. With the exception of the +12V rail, all power values were close to the ATX standard value.
As I previously mentioned, Hyper 700 uses a 2000RPM, 88.0CFM, 40dBA fan, manufactured by Yate Loon when running at 12V. Under idle, the power supply is audible, but not by much, the sound-meter registered a 28dBA at 1meter away. Under heavy load, the noise output is clearly audible, with the noise output resting at 34dBA.