Now that I know what connector goes to what label, I know how I'm going to set up the load tester.

For those who don't know how I go about testing power supplies; I use a SunMoon SM8800 power supply ATE (automated test equipment.) Essentially, it's a machine made specifically for the testing of ATX12V power supplies.  It can put up to 10 different static loads on up to 10 different rails at a time.  I can also ramp loads up and down and do short tests, etc. 

For the sake of a review, I just pre-program a few different loads into the tester and run each one for about a half an hour.  For the ST75F, I've programmed seven different loads ranging from 145.4W to 761.9W.  The seventh, and final test, is a crossload test with the +12V rails maxed out, and a minimal load on the 3.3V and 5V rails.

For more information about my testing methodology, read this.

So I plugged the 8-pin EPS+12V of the power supply into the 8-pin of the ATE for the +12V1 and +12V2 tests. Since the load on the 8-pin is split across two +12V rails, I pretty muched doubled up the load I put on that connector.

The +12V3 load is split up between the main ATX connector and one of the PCI-e connectors.  The +12V4 load is split up between the two PCI-e connectors plugged into the +12V4 interface of the ST75F.

Intake and exhaust temperatures are measured with a temperature probe at the intake fan and exhaust vent of the power supply.

Going from test 1 to 5, we see excellent voltage regulation.  During test 6, we see things break down a bit, but this can be for multiple reasons.  One of which being that we're actually a tad over spec at 761.9W.  The other is that resistance building up in the wire leads and modular connections of the power supply are causing voltages to drop.

To give you an idea of how much the SilverStone ST75F didn't like being loaded up to 761.9W, I shot this video.  You'll note how the voltage readings of each rail seem to bounce around considerably.

In retrospect, check out this video of the ST75F during test 7.  Only the 3.3V and 5V rails have been reduced from the previous test.  The 12V rails are still juiced up putting out 648W of power, yet the bounce in the voltage has been reduced significantly.  In fact, it's almost non-existent.  This tells us the problem we had seen in the video for test 6 is not likely caused by the voltage output rectifiers, because otherwise the bounce would be present regardless of the total power output of the PSU and consistant with the actual load on each rail.

Efficiency is pretty decent, but nothing really special now that we're seeing more and more power supplies that are consistantly over 80% efficient.  Only when the PSU was around 50% loaded did it give me efficiency over 80%.

The fan was very quiet and could barely be heard over the fan of the load tester.

Now were going to run the "hot" load tests. To perform the "hot load tests" I put a side panel on the case the PSU is mounted in and run a hose and an 80MM intake fan into the TAC CPU duct. The heat generated by the load tester is pumped into the other end of the hose.

The voltages seem to be unaffected by the high temperatures.  The fan did kick into high speed during test 5 and then to an even higher speed again during test 6, which explains the drop in intake temperature during those tests, but the fan was never really "too loud."