All testing is done using an SM-8800 active load tester in a 20°C room. Artificial static loads are placed on each rail, then voltages and total wattage is recorded. The loads represent a +5VSB only load, 10%, 20%, 35%, 50%, 65%, 80% and 100% as well as two cross-loads: High +12V and high +3.3V & +5V. The amount of wattage being pulled from the wall is also recorded and efficiency calculated. I also record the power factor of the unit. This particular unit has no -5V rail and we do not load test the rarely used -12V rail, so there are no results for those two rails.

Test 0 is the PSU in it's soft off mode with only a 2A load on the +5VSB. This test will show us the PSU's standby efficiency. Test 1 is the same as test 0, but we've turned the PSU on. Test 2 through 8 are our progressively increased loads (10% to 100%) and test 9 and 10 are our crossloads.

Temperatures are measured at both the intake fan and exhaust grill of the PSU.

First observaion is test 0 which shows very good efficiency while the unit is in standby.

Going across the columns of the table for cold loads, we find that the 3.3V dropped 3% going from 10 to 100% load. +5V dropped almost 2.5%. This isn't too bad, especially considering that this unit uses DC to DC for the +3.3V and +5V. +12V, on the other hand, barely dropped 2%.

If we look at test 3, 5 and 8 (20%, 50% and 100%), we can see that we just missed the 80 Plus Gold criteria (87%, 90%, 87%) with this unit. Instead, we have very respectable silver performance. The numbers are so close, it's barely worth mentioning. Different variables come into play that could cause the efficiency to be slightly different between my unit and the one that 80 Plus tested, such as how the loads are distributed across the rails or how well the connection is between the modular cables and the PSU body or the connection between the connectors and the load tester itself.

One thing to note is a strange electrical noise I heard during test 3. Once I proceeded to test 4, the noise went away. A second sample provided by Corsair exhibited no such noise.

Another thing to note is that the fan does not spin on this unit until temperatures reach a certain point. For me, the fan didn't start spinning until test 5, and this can be seen in the temperature results. As the fan began to spin, it pulled in more of the room temperature air, dropping the intake temperatures, and increased the exhaust temperatures. The fan was barely audible, especially over the roaring noise of the SunMoon's exhaust fan. By test 8, the fan seemed to step up to an audible level, but nothing in the least bit annoying.

Cross load results were very acceptable. Not that such loads are realistic, but because this unit is essentially a +12V PSU with DC to DC modules regulating the +3.3V and +5V, cross loading results are expected to be uneventful.

Let's see what the oscilloscope has to show us...

Oscilloscope Measurements for Corsair AX850
Test #	+3.3V	+5V	+12V	+5VSB
Test 1 (+5VSB Only)
Test 2
Test 3
Test 4
Test 5
Test 6
Test 7
Test 8
Test 9 (CL 1)
Test 10 (CL2)

Very clean voltage here. One would say "uneventful" if it weren't for that +5VSB. What the heck is that thing doing? Each test produced a different waveform even though the load is 2A on all but the crossload tests. And then, during the crossload tests, when there's no load on the +5VSB, there's a different wave form depending on the crossload. With all of the load on the +12V, the +5VSB waveform is down right out of control! Very strange.

Now for the hot load tests....

Hot loads are produced by pumping the hot air produced by the load in the load tester into the intake of the power supply:

Well that's interesting! It seems that some of my efficiency numbers have gone UP! Hmm.... I guess the AX850 needs breaking in just like a high performance racing engine. I still didn't hit 80 Plus Gold, though.

Now that we've wrapped up the load tests, let's open up this beast and have a look inside....