Power Supply Load Testing

I am well aware of what the more knowledgeable readers think about these power supply reviews: impractical and often useless. Benchmark Reviews is working hard to change that impression. Recording voltage with a digital multimeter alone offers no value to the test, since power fluctuates a great deal in the span of a single millisecond. Because of this, Benchmark Reviews has taken a corrected approach to testing PSU's and hopes to offer the best analysis possible.

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, energy efficiency, voltage regulation, and current ripple. Each of these variables must be analyzed with unique tools, which makes it properly review a power supply unit without them.

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 power supply units worth $100-$300. 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. Additionally, the CyberPower CP1500AVRLCD 1500VA 900W LCD UPS reports the total power load drawn by the system. Used together, the Oscilloscope. LCD UPS, and digital multimeter will offer readers the best measurement of power stability any review site could reasonably offer.

Test System

• Motherboard: ASUS P5K3 Deluxe (Intel P35 chipset) with version 1001 BIOS
• Processor: Intel E6550 Core 2 Duo 2.33GHz operating at 3.31GHz
• System Memory: Corsair PC3-14400 DDR3 1800MHz 2GB RAM Kit @ 1888MHz
• Cooling: Xigmatek HDT-S1283 Exposed Copper Heatpipe CPU Cooler
• Video: FOXCONN GeForce 8800 GTS 640MB overclocked to 600MHz Core / 900MHz RAM
• Audio: HT Omega Claro Plus+ AD8620BR Op Amp HD Sound Card
• Hard Disk 1: Western Digital WD360GD 10,000 RPM SATA
• Optical Drive: Toshiba Samsung SH-S203 SATA 20x DVD R/W
• Enclosure: Lian Li PC-B25B Black Aluminum Mid-Tower ATX Case
• Operating System: Windows XP Professional SP-2 (optimized to 16 processes at idle)

A/C Power Draw Index

During testing, the test system described above consumed the following power statistics:

Kingwin Mach 1 ABT-800MA1S 800W Modular Power Supply Unit

198W Idle (0% CPU utilization / 2D low power GPU)
306W Fully Loaded (100% CPU utilization / 3D high power GPU)

Test Equipment

• Velleman PCSU-1000 Oscilloscope
  • DC Voltage Regulation Settings: 2V DC Volt/Division, 60MHz 1x, 1ms
  • AC Power Ripple Settings: 50mV Volt/Division, 60MHz 1x, 1ms
• PC-Lab2000SE v3.05 Interface Software
• Extech 450 Auto ranging Digital Multimeter
• CyberPower CP1500AVRLCD 1500VA 900W LCD UPS
• ATITool v.26 Video Card Overclock Utility
• Folding@home v6.00 Beta 1 Console Client

Test Methodology

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. Over time, there have been many changes to the test environment in terms of system hardware and software versions. We understand that consistent in our testing methods is very important in order for the results obtained to be comparable.

At the start of every test, the Velleman PCSU-1000 Oscilloscope is calibrated with 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. These readings are displayed several times per second, so a visual inspection of the most common (average) reading is recorded for our test results.

After the results have been recorded from the light idle load, our test system then receives heavy load by utilizing ATITool to scan for artifacts and two Folding@home console clients. ATITool forces the video card into high-power 3D mode, and the Folding@home clients cause 100% CPU utilization on each core. After many months of research, these two programs have proven to create more power draw than anything else available, including the combined EVEREST System Stability Test. After another minute of heavy load the power supply is again measured, and the AC power ripple and 12V DC voltage regulation results are recorded.

<< Previous Page - Next Page >>