Hilbert: he calls his rig the dust buster :)

Yes! I know! I’ll try to hide the cables when/if I get a bigger case. (Naaah!)


Performance and Usage

The cables are sleeved, not in a heat shrunk wrapping though, and are certainly just a bit more stiff than a non-sleeved cable, but nothing to worry about. Shouldn’t give you any problems though. The power plugs are easy to unplug thanks to their grip system.

Let’s go to the important part though - Performance.
Stable and quiet are the only words I can use for it. Using a software tool to measure the voltage on a power supply is not a good thing as it does not give an accurate reading. For example the software that comes with my motherboard would always give a reading of 11.80 volts for the 12V rails. Wouldn’t go up or down from that. But in reality, when measured with a multimeter, the PSU never went below 11.88. That might sound bad for some, but that is with the PC working on 100% in games (Quake4, Half Life2, Half Life: Source ~ hey, still fun) and rthdribl 1.2 (rthdribl is a real time HDR/shader demo tool that is taxing on the graphics hardware). That was also the case when I had rthdribl running along with Steam making a backup of my Steam games, so the hard disk was used a lot as well as the graphics card, and using rthdribl with Nero burning the file Steam made. (Real time HDR image based lighting ~ probably the best tech demo for HDR for the moment).

So, when is a PSU good? When it can handle the load of the PC it is powering and not shut down, while still keeping the voltage level within acceptable limits (5%). The best way to measure the voltage is to simply attach a multimeter to one of the power cables of the power supply unit. It is important to say that the more the PSU is stressed, the more the voltage is to be expected to go down. The unit can be stressed by making the PC work at 100% and by having as many components as possible working away (That means having the fans at full blast, using the CD/DVD/hard disk drives, using as many of the extra cards as possible.)

More specifically, the test I did was to have 3DMark 2006 running, Winamp in the background playing a stream from an online radio, and Nero using 2 optical drives to burn 2CDs at the same time, using the relevant option in Nero. So, my 2 optical drives were being used, the hard disk to send the data for burning, and the graphics card was pretty much maxed out. The tests were done twice. The first time the readings were taken from the cable that feeds the graphics card, the second time from the cable that feeds both of the drives with juice. As I don't really trust the motherboard's readings, a multimeter was used.

On idle, the multimeter showed a reading of 11.91v~11.93 on the 12V line.

On with the testing then. As said before, 3DMark06 was used at the same time as Nero burning on 2 separate writers that were on the same power cable.

The readings from the graphics card's power cable were:

• 11.93v when idling.
• 11.91v with 3DMark06 running, and interestingly, 11.89v~11.90v with the 4th demo of 3DMark, the one with the snow. Only on that. That most probably means that the demo uses some part of the GPU that the others didn't, or perhaps uses the GPU in a different way as the other demos were at 11.91v.
• Burning or reading with the drives didn't seem to affect the power going to the graphics card.

The readings from the optical drive's cable:

• 11.96v idling (Not reading or writting.)
• 11.93v~11.92v while the drives were writting data on the CDs. (Data were around 600MB, burning at 48x speed)
• 11.94v while reading (was verifying the data written)