Testing was done using a multi meter, Intel Hardware monitor, and my sound pressure level meter. I tested on the following platform
The E-Power Tiger 550w Silent Engine showed excellent power regulation on the 3.3v rail. Under full load with Stanford’s Folding at Home and 3Dmark05 running there was only a 3/100 of a volt difference between both no load and idle loads. When compared against a Sparkle 400w power supply the Epower looks good already.
Moving on to the 5v rail, we see quite a bit more veryation, this is lily due only to the lack of any resistance or load on the 5v rail. Under idle load and full load the E-Power again exhibits quality voltage regulation, equal to that of the Sparkle unit.
The E-Power Tiger 550w power supply lacks two 12v rails as suggested on the box, label and website, thus there is only one 12v shown. The Sparkle is a pre EPS unit and only has a single 12v rail as well. When compared the Epower actual shows slightly better voltage regulation on the 12v rail. It is also apparent based on the higher 5v rail and lower 12v rail under no load that both use a feedback regulation. This means that if either your 5v load or your 12v load get to great or too small it can cause the other voltage to be out of spec. At first glance this seems like a bad thing, but this is a very standard regulation method used in top of the line power supplies.
The final graph shows the same quality of regulation on the -12v and -5v rails. The consistency of the Epower honestly is impressive, its better than I was expecting.