Installation and Use
Aside from the motherboard and CPU connectors, permanently attached to the unit are three SATA connectors, three device connectors, and one of each of the PCI Express connectors. There was sufficient connectivity provided that I didn’t even have to use any modular cables.
I used OCCT Perestroika in conjunction with Speedfan for benchmarking and monitoring on Windows XP Pro 64-bit after installing into my standard testing rig with an ASUS M3A32-MVP motherboard, Phenom II 940, 8 GB of DDR2 RAM, three SATA hard drives, a Creative X-Fi gamer, and a Foxconn-made nVidia 8800 GTX.
The 3.3V rail idled near 3.2V and rippled 1% during the load test. It dipped to approximately 3.17 V during testing. 3.135V is the lower end of the plus or minus 5% threshold we like ’round these parts, so continued use would probably merit monitoring that rail for further dips.
The 5V rail idled near 4.95V and rippled 1.09% during the load test. It was always well within the +- 5% range, unlike the 3.3V rail.
The 12V rail idled at approximately 11.96V and rippled a mere 0.53% during the load test. There’s an interesting pattern in the output graph-discuss in the forums what you think it may be.
All of these rails undervolted. All were within 5% of spec, but they likely still merit observation for decreased performance. I may have been able to lighten the 12V load by using one of the modular PCI-Express cables on an extra rail.
Normally, I include output graphs from a review of a similar product, but I felt that none of my recently reviewed PSUs where close enough matches to this one in output and stability.
The unit didn’t produce an abnormal amount of noise outside the standard humm, based on my low-tech ear-near-the-computer test. The ambient noise of the Thermaltake DuOrb VGA cooler is undoubtedly louder than just about any PSU, even at full load.