Related Articles

So what do we use for our voltage testing? Nothing other than a pair of ATI X1900XT graphics cards running in CrossFire mode. CrossFire is a current hungry beast that low quality power supplies, even ones rated as high as 650W, have problems with. If said supply cannot deliver enough juice, then that supply will go into its over drive mode and output more Wattage than it is rated for and potentially damaging the PSU.  So CrossFire is a good test for a power supply.

For testing, I tossed those two ATI X1900XT cards into our standard video test bench system to come up with the following configuration.

• Intel Core 2 Duo E6600 Conroe @ 3.0GHz (333 MHz * 9) processor
• Asus P5B Deluxe / Wifi-Ap motherboard 
• 2 * ATI X1900XT 256 MB video cards in CrossFire mode
• 2 * 1024 MB Corsair's XMS2-5400UL DDR2-667 5-5-5-15/2T
• Seagate Barracuda 7200.9 SATA hard drive
• Samsung WriteMaster DVD burner
• Windows XP Professional SP2

And used the following drivers to get it all running.

• Nvidia ForceWare 93.71
• ATI Catalyst 7.1
• DirectX 9.0c
• DotnetFix 2.0

So it's all set up and ready for stressing and testing. But who will we test against?

• CoolerMaster RS-850-EMBA 
• OCZ GameXStream 700W
• Thermaltake ToughPower 750W Cable Management

We have some pretty recent supplies in here. But one added not for the testing procedures. Whenever I test a power supply with a modular cable design, I always test both the connections on the cables and the connections on the side panel of the supply. I do this because every electrical connection in a circuit creates resistance. Granted, the resistance in the cases of modular power supplies is so minimal that it won't really have any affect upon performance. These little voltage differences between the connectors and the side panel bus bar are just interesting to see. So I grabbed our Mastercraft 052-0060-2 volt meter and was off to test. We'll be checking the 3.3 Volt, 5 Volt, and 12 Volt rails, both at idle and load.

To load things up, we simultaneously run 3DMark06 (the Deep Freeze test because of the intense HDR going on) and two instances of CPU burn-in churning away in the background. What did we get?

Now granted that the Corsair HX620W supply doesn't meet the exact ATX power specs of 3.3, 5.0, and 12.0 volts exactly (I have yet to see a power supply that does), the differences between idle and load are so insignificant that they don't make any difference whatsoever, if even worth mentioning. Needless to say, I am impressed. But what about those other supplies, how do they fare?

So we did see a few jumps here and there, up and down with respect to voltage levels between idle and load. But the two winners here are definately the Corsair HX620W and the CoolerMaster RS850EMBA power supplies. Perhaps sound level testing will be the tie breaker.

For that we turn to our Omega HHSL1 sound meter. For this test I use a room at home that gets down to around 28 to 30 dBA ambient in the middle of the night.

Okay, we do have that slightly lower ambient sound level reading for the HX620W, but that isn't going to overly distort out results. That 1-2 dBA difference in ambient sound is so insignificant that the human ear would have a difficult, if not near impossible time discerning the difference.

But as I was the one taking these sound measurements (at 3am on a very quiet night) I can fully attest that the Corsair HX620W is one of the quietest power supplies I have come across in some time.

With testing out of the way, I think now we may be able to form a very definite conclusion on the Corsair HX620W power supply.