PSU Load Testing

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 supplies 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 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.

Test System

Motherboard: Gigabyte GA-P35C-DS3R (Intel P35 chipset) with F4G BIOS
Processor: Intel E6600 Core 2 Duo 2.2GHz operating at multiple speeds
Memory: Winchip DDR2 1200MHz PC2-9600 2GB RAM Kit
Cooling: Thermaltake MaxOrb CL-P0369 CPU Cooler
Video: Foxconn GeForce 8800 GTS 640MB overclocked to 600MHz Core / 900MHz RAM
Audio: HT Omega CLARO 7.1 C-Media 8788 Oxygen Chipset
Hard Disk 1: Western Digital WD360GD 10,000 RPM SATA
Optical Drive: Toshiba Samsung SH-S203 SATA 20x DVD R/W
Enclosure: Zalman Z-Machine GT1000 Black Mid-Tower ATX Gaming Case
PSU: ePower Technology EP-1200P10 xScale 1200W PSU
Operating System: Windows XP Professional SP-2 (optimized to 16 processes at idle)

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 v3.03 Interface Software
• Extech 450 Auto ranging Digital Multimeter

Power Supply Unit 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. 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.

Voltage Regulation & Ripple Test Results

In the test results below, it will be necessary to explain what you are viewing. In each image the AC power ripple is represented by the yellow trace line making up the waveform. While every personal computer power supply unit available to the retail market has some degree of measurable AC power ripple, it is most important that measurable AC ripple is very minor and does not create a large peak to peak voltage (Vpp) distortion. Stable, well-regulated power is critical to system stability and hardware longevity.

AC Ripple Waveform at Light Idle Load

The waveform image above shows the AC power ripple under light idle load, which is measured inside of a 1ms recording frame. Ideally, a lower peak-to-peak voltage (Vpp) and AC current ripple is desired. Waveform data recorded during the light idle load measurement is represented in the chart below:

AC Ripple Waveform Data at Light Idle Load

The waveform data above describes the actual measurements at light idle load. The maximum peak-to-peak AC voltage distortion was 359.4 mV which is higher than the average, and the AC RMS ripple was 14.23 mV under light idle load which is considerably better than the average results we have collected under light idle load.

Here are the average AC RMS ripple measurements under light idle load for power supplies we have recently tested:

The waveform image below shows the AC power ripple under heavy load. The maximum peak-to-peak AC voltage distortion was 393.8 mV which is higher than average, and the AC RMS ripple was 32.7 mV under heavy load which is among the average results we have collected under heavy load.

AC Ripple Waveform at Heavy Load

Under heavy load, the Antec TruePower Quattro 850W PSU TPQ-850 measured an average AC RMS ripple twice as high as the current ripple at light idle load, which is average. The waveform data recorded during the heavy load measurement is represented in the chart below:

AC Ripple Waveform Data at Heavy Load

Because the Antec TruePower Quattro 850W PSU TPQ-850 offered above-average results in the idle readings, I was unsure of how the unit would perform under load. Ultimately it did fine, performing within specification and well within the average of results. Here are the average AC RMS ripple measurements under heavy load for other power supplies we have recently tested:

DC Voltage Regulation Waveform Data at Light Idle and Heavy Load

Many of the power supplies we test here at Benchmark Reviews have DC voltage regulation measuring above the 12V baseline more often than below. For obvious reasons, the diversion from the 12V baseline with increased voltage is more preferred than below it since dropping too far under 12V will cause a system to power off or recycle. Plainly put, no power supply is ever perfectly centered at 12V DC; instead they most often maintain a DC mean voltage higher than they are regulated at.

<< Previous Page - Next Page >>