|Mushkin HP-580AP Modular 580W PSU|
|Reviews - Featured Reviews: Power|
|Written by Olin Coles|
|Monday, 04 December 2006|
Page 5 of 7
Load Testing Methodology
Power supplies are not like most other computer components, where it is a simple matter of comparing the item to others in the same category. Power supply's are comprised of several different variables, such as maximum output, voltage regulation, and current ripple. Each of these variables must be analyzed with unique tools, which makes it a difficult product to properly review.
I am well aware of what the more knowledgeable readers think about these power supply reviews: impractical and often useless. That's about to change. In the past, Benchmark Reviews has been guilty of the same thing nearly all of the other sites are guilty of: testing with a digital multimeter alone. So beginning now with this review, we are taking a corrected approach to testing PSU's and hope to offer the best analysis possible.
Benchmark Reviews has researched the equipment necessary to complete the most thorough power supply unit review possible. We spent the time learning what it takes because the best possible PSU tests are what we want to give to our readers. But after discovering the prices on a programmable output DC power supply system, variable range load testing units, and a DC power analyzer (Oscilloscope), we felt that more than ten-thousand dollars worth of equipment would be far too cost prohibitive for testing a hundred-dollar power supply. So without compromising too much, we made a slightly less-expensive investment into a good quality Oscilloscope in order to test DC voltage regulation and AC power ripple. Together, the Oscilloscope and digital multimeter will offer readers the best measurement of power stability any review site could reasonably offer.
Our testing process is comprised of measuring the AC current ripple, and the DC voltage regulation. There are several key steps, all of which allow us to measure and record our readings using the identical methods for every test we perform. Consistent testing methods are key to obtaining comparable results.
At the start of every test, the Velleman PCSU-1000 Oscilloscope is calibrated to the PC-Lab2000SE software. After calibration is complete, the voltage on the 12V rails are measured and recorded with the Extech 450 digital multimeter to ensure comparable margin across all rails. Once the Velleman 60MHz probe has been grounded and attached to the 12V lead, our test system is powered on and left at the Windows logon screen for ten minutes.
Once this lightly-loaded idle warm up period is complete, the Velleman PCSU-1000 Oscilloscope was allowed to run for one minute measuring the AC power ripple. Then once ripple is recorded, the 12V DC voltage regulation is recorded after another minute.
After the results have been recorded from the light idle load, our test system then receives heavy load by utilizing the following tools: two console versions of the http://folding.stanford.edu/client operate and task each CPU core to 100%, hard disks are stressed by benchmarking each with HD Tach RW, system memory (RAM) is given a stress test with Lavalys EVEREST, and ATITool scans for artifacts which forced the video card into high-power 3D mode. After ten minutes of heavy load the power supply is again measured, and the AC power ripple and 12V DC voltage regulation results are recorded.