I've been playing with power supplies for quite a while now and I guess in my old age, I like to sometimes make things interesting. My normal method of testing a PSU is to get the system up and running, get a burning session on a CD-RW running in the background, then cranking up something along the lines of Quake 4 and run a long-winded demo. Once everything is cranked up, I'll measure the voltage levels along the power rails to see how close they remain to specifications. The thought behind this method of testing is to see if the power supply can maintain acceptable power levels while under stress.
While I will continue to use this methodology, I also decided to compare these results with the power levels at system idle just to see how much the unit drops off while the system is under a load. This should give us a little more in depth picture of what the PSU in question is capable of. After all, with power supplies becoming a very important component in modern computers, it pays to know ahead of time what works and what doesn't.
But before we dig into the multimeter readings, lets take a quick look at the test system being used:
DFI LANParty UT nF4 Ultra-D motherboard (nVidia nForce4 Ultra chipset)
AMD Athlon FX-53 processor (Supplied by Newegg.com)
2x 512MB Mushkin "Redline" PC3200 memory (Supplied by Mushkin)
Sapphire X1900XT graphics card (Supplied by Sapphire)
Four case fans + Thermaltake CPU cooler using 80mm x 38mm high performance fan
All right then... the system is running and the tasks are chugging along. The processor is at default speed and voltage is set to 1.5v. The memory is running at 2.9v and all other levels are set to default. The motherboard is an absolute power hog and the video card has sent many lesser power supplies to an early grave. Lets see what this thing can do!
All right, I'm stumped. Impressed, but stumped. All voltage rails, no matter what level of system stress, produced excellent results of over specified ratings. But I cannot figure out how this power supply actually got a higher voltage reading on the 12v rail when the system was pulling. I might buy off on this if I had been using an older AthlonXP based rig, which generally uses more juice from the 5v rail instead of the 12v rail, but not on this hog. I even double checked the results and they were the same. The multimeter is calibrated, so I don't know.
Like I said, I'm impressed with the overall results, but I'm still stumped.
Another item of note during testing is the very steady voltage levels. Many power supplies will produce voltage readings that are constantly moving, even just a bit. All readings taken above were extremely steady with the only jumping being a slight .01v movement in the 12v rail at idle. Very nice indeed!