Cybenetics adds load range breakdowns to PSU evaluations
★ Add us as a preferred source on GoogleTL;DR
Cybenetics is supplementing power supply ratings with efficiency, power factor, and noise data broken down into five 20% load ranges and a High-Load Index. These new metrics are informational and do not change a unit's ETA or LAMBDA grade.
Cybenetics is adding load range breakdowns to its power supply evaluations. For units with a measured certification sweep, the company will now report efficiency, power factor, and noise levels across five 20% load intervals: 0-<20%, 20-<40%, 40-<60%, 60-<80%, and 80-100%. Cybenetics is also adding a High-Load Index that covers all measurement points at 70% or more of a unit's rated output.
These metrics are based on the measured data series of more than 1,450 load combinations per unit. Cybenetics excludes test 0 and any interpolated chart data from these calculations. For each range, the report includes the number of measurement points, average efficiency, average power factor, average noise, the loudest individual point, and the percentage of points where the fan was stopped.
Decibel calculations utilize an energy-based averaging method. To determine noise, Cybenetics holds each load point for approximately five seconds and logs the fan speed, which is then converted to a dB(A) value using a table specific to the power supply. The reported noise floor is 5.0 dB(A).
Where fixed-load measurements are available, Cybenetics also provides a table comparing fan speed, noise, and intake temperature at 10, 20, 50, and 100% load. This compares performance at normal ambient temperatures (30-32 °C) against high ambient temperatures (40-47 °C).
These breakdown results are informational only. Such data supplement rather than substitute current figures without altering efficiency or noise grades. Redundant server module evaluations begin their intervals at two percent, presenting efficiency and power factor statistics prior to noise records.
Via Hardware Busters ↗ · Also from Igor's Lab ↗
Spot a mistake? Tell us