Power supply testing methodology

• For testing power supplies I am using a homemade tester that consists of multiple power resistors and amperometers.
• This allows me to control with pin point accuracy the exact load at each rail
• One more reason to test psu's with this method is that high power power supplies(800+ watts) are very difficult to be tested with a normal system.With the tester I am using I can achieve loads of up to 1800 watts.
• Readings are taken with a high sensitivity digital multimeter, and if required a voltage drop diagram is made.
• In order for a test to be considered successful the psu has to be able to perform it for 2 hours.
• Noise levels are recorded from 10cm away from the back of the psu(air exhaust) 

TESTING Part 1st (Real testing)

At this part the ClacialPower GP-AX950AA will be tested in real conditions, in my pc that was built for this reason.

Testbed

• CPU:Intel core i7-950 @ stock speed
• Mobo:EVGA X58 Sli
• VGA:Dual MSI GTX 260@ Sli
• Memory:Adata 6Gb kit
• HDD: 4x Seagate 500Gb

All reading were taken with a high sensitivity digital multimeter.In order to achive max load several instances of Orthos were running,with Furmark to stress the vga cards.Below we can see the results.

The ClacialPower GP-AX950AA remained very silent at idle and the fan was completely inaudible, but at load the noise was a bit higher and fan was spinning at higher speed.This is normal because a 950w psu is producing lots of heat, but despite the high fan speed the noise was still at very low levels.What amazed me was the stability, there was a really slight voltage drop.This indicates a high quality psu.But it will tested better at the synthetic testing which is more difficult for the power supplies.

TESTING Part 2nd(Synthetic testing)

At this part my homemade psu tester is used.This allows me to know the exact load of each rail.This is very usefull because I can measure the efficiency and also see if the psu can really provide the specified wattage.

Please note that synthetic testing is much more difficult for the power supply, and this means that it has more voltage drop than with the same load at a real system. 

Here are the results:

Part 1 - 30% load

Part 2 - 50% load

Part 3 - 80% load

Part 4 - 100% load

Part 5 - Individual line loads(each line is tested alone with the maximum specified load)

Efficiency

Efficiency is calculated by measuring the consumption of the power supply for known loads.

Temperatures

Here exhaust air temperature was measured.  

ClacialPower GP-AX950AA Sound pressure level

Generated sound was monitored 10cm away of the exhaust of the psu at a sound insulated box in order to eliminate any external sound or echo.Readings are taken using a high sensitivity digital SPL meter.(Values under 17dBA are considered to be inaudible)