Single page Print

Loading up The Beast
A cornerstone of our power supply testing methodology is a custom-built load generator created by forum regular just brew it! that we like to call The Beast. If you're unfamiliar with its frightening array of resistors, I strongly suggest checking out this page from our first round-up to use The Beast, which details the rig's internals and our testing methodology.

We recalibrated each and every one of the variable-load resistors inside The Beast prior to testing for this round-up. None of the resistors needed more than a minor adjustment, but since we did tweak things some, the results in this article aren't directly comparable to those from our previous PSU round-ups.


Behold The Beast!

We use The Beast to push each PSU to 25, 50, 75, and 100% of its output capacity while measuring DC voltage, AC ripple content, and gathering data that can be used to calculate overall efficiency. This creates a kind of level playing field on which each PSU is pushed to its individual—and advertised—limits. We should note, however, that our testing methods are not identical to those used by the 80 Plus program or by vendors who publish their own efficiency ratings. The Beast is a different breed of load generator than those used by the folks at 80 Plus; it only taps the 3.3, 5, and 12V rails, leaving -12V and standby 5V lines unused. To ease confusion, we'll be referring to efficiency ratings gleaned from our own testing as Beast efficiencies. Those figures shouldn't be compared to efficiency ratings posted by PSU makers.

The Beast is also limited to applying loads in 2A increments, so we borrow a page from The Price is Right and use amperage loads that come as close as possible to our targets without going over. The chart below shows the amperage loads applied to each PSU.

Total loads (Amps)
25% 50% 75% 100%
3.3V 5V 12V 3.3V 5V 12V 3.3V 5V 12V 3.3V 5V 12V
Corsair HX750W 750W 2 2 12 6 6 26 10 10 38 14 14 52
Enermax Revolution85+ 850W 4 4 14 8 8 28 12 12 44 16 16 58
Seasonic X Series 750W 2 2 12 6 6 26 8 8 40 12 12 52
XFX Black Edition 850W 2 4 14 6 8 30 10 12 44 12 16 60

When testing with The Beast, each power supply was hooked up using its primary and auxiliary 12V connectors, and when available, two PCIe power connectors and six 4-pin peripheral connectors. We used a Pico ADC-212 digital oscilloscope to probe the 3.3 and 5V wires on the primary power connector. The 12V lines were probed at the primary power connector and also at one of the PCIe power connectors. In the graphs on the following pages, 12V power from the primary connector will be marked 12V1, while power from the PCIe connector will be marked 12V2.

Rather than calculating efficiency based on static 3.3, 5, and 12V, er, voltages, our calculations take into account the actual DC voltage delivered on each line during testing. This should compensate for any voltage fluctuations that some PSUs exhibit under load.

Our testing methods
Testing was conducted in two parts. First, the PSUs were run in the system detailed below for a series of power draw, temperature, and noise level tests. They were then hooked up to The Beast to test power delivery and overall efficiency.

All tests were run three times, and their results were averaged.

Processor Athlon 64 X2 5000+ 2.6GHz
System bus HyperTransport 16-bit/1GHz
Motherboard Asus M2N32-SLI Deluxe Wireless Edition
Bios revision 0906
North bridge nForce 590 SLI SPP
South bridge nForce 590 SLI MCP
Chipset drivers ForceWare 9.35
Memory size 1GB (2 DIMMs)
Memory type Corsair CM2X512A-5400UL DDR2 SDRAM at 742MHz
CAS latency (CL) 5
RAS to CAS delay (tRCD) 5
RAS precharge (tRP) 5
Cycle time (tRAS) 12
Audio codec Integrated nForce 590 SLI/AD1988B with 5.10.1.4530 drivers
Graphics GeForce 8800 GTS 640MB with ForceWare 162.18 drivers
Hard drives 2 x Western Digital Caviar RE2 400GB SATA
OS Windows XP Professional
OS updates Service Pack 2

We used the following versions of our test applications:

The test systems' Windows desktop was set at 1280x1024 in 32-bit color at an 85Hz screen refresh rate. Vertical refresh sync (vsync) was disabled for all tests.

All the tests and methods we employed are publicly available and reproducible. If you have questions about our methods, hit our forums to talk with us about them.