Big thank you to CryoPC for providing the following test rig:
CPU: Core i5 750 @ 4GHZ
Motherboard: Asus P7P55D-E
GPU: Sapphire 5850 1GB
RAM: Corsair Dominator 4GB PC3-12800 DDR3
Storage: Samsung F3 500GB 16MB
CASE: NZXT Hades
Here at XSReviews, we don’t have the money for a full test bench when it comes to PSUs. Load testers cost thousands of pounds which unfortunately we just don’t have. Because of this, we don’t pretend to test the rail stability or efficiency of a PSU, as we simply can’t do it effectively.
However, what we can do is test the Ripple/Noise quality of each rail using an oscilloscope. The one used was USB Instruments Stingray DS1M12 using HP 3060 probes and also in line with ATX specifications, I used 10uF and 0.1uF capacitors to simulate added load.
Noise and Ripple
The last PSU review I did included a pretty good explanation of Noise/Ripple, so if you don’t mind I’ll just reproduce that here instead of rewriting it:
“So what is noise and ripple? This is the description of minute voltage fluctuations on each of the main rails in any PSU. In a nutshell, the good power supplies have very little, the awful ones fluctuate a lot. According to ATX Specifications, there shouldn’t be more than 120mv in peak to peak voltage fluctuation on the 12v rails, and no more than 50mv on the 5v and 3.3v rails. These are the theoretical maximums, with good PSUs hitting somewhere under half, so 60mv and 25mv respectively.
If noise and ripple are present in abundance, you’re most likely going to find it at full load. Not posessing a load tester, the best I was able to do was to load up Furmark and S&M CPU stress and max out the test rig. This ended up drawing 375w from the mains. This will likely equate to nearly 10% less used by the PSU because no power supply is 100% efficient. All we have to go on here is the 80 PLUS gold rating of the PSU, which under 50% load conditions should be around 92%.
All images are of the average fluctuations. Not posessing a hot box to ensure that the unit is tested under stressfully heated conditions, air intakes and exhausts were partially covered. The internal temperature of the PSU was then measured using a probe, with results taken when it reached 40 degrees centigrade.”
Since I used the .01v scale, each line on the graph represents 0.01v, or 10mv.
Numbers on the right of each image to take note of are either the middle or top (they are identical) peak to peak frequency.
While the 3.3v and 5v rails on the 800w unit were absolutely fine, well below half the maximum fluctuation, the 12v rail wasn’t as strong. Sure it was still well within maximum specifications, but it wasn’t as stable as it could have been.
With this PSU the 3.3v and 5v rails were not quite as stable as those of the 800w, though the 12v was much better.
As usual, this review makes no claims on the efficiency of the units tested as it’s simply something we can’t track without a load tester. However, as for Noise/Ripple tests, I’m happy to say that both the units passed admirably. The 800w is perhaps more stable than that of its larger brother, but in either case these are safe units to use.
The only real downside I can comment on is the price. I couldn’t find these anywhere for sale in the UK apart from one site that had both the 800w and 1000w creeping up on the £200 mark which puts these in the most expensive category of units.
Noise/Ripple on all rails was well within specifications on all rails under mid-level load
High quality packaging and product