@Xeroxxx
Bad news, the figures seem to be quite accurate except for one situation where the BMC reports a much higher power draw (server OFF, both PSUs online). I've run some tests with a Peaktech 9040 which isn't calibrated nor does it report its accuracy but the reported and measured values are pretty close. Tested with the following hardware config:
2x E5-2640v3, 148GB (10 DIMMs 2133MHz@1866), LSI3008, Intel 82599 2x10Gb SFP+, Intel Pro1000PT 2x1GbaseT, Mellanox Connect-X3 Pro 2x40Gb QSFP, 6x SATA HDD@7k2rpm, 1x Intel S3500 80GB and 1x Samsung PM1633a 960GB. Network links down except management port, fans @22%, line voltage 231VAC, 2xPSU 750W in load balancing mode.
Hardware state | reported by BMC | measured |
Off, 1xPSU online, BMC run | 18W | 16W |
Off, 2xPSU online, BMC run | 36W | 19W |
POST | 198W | 198W |
OS boot (ESXi 7.0U3c) | 216W | 210W |
OS run (idle, no VMs running) | 192W | 191W |
OS run, 1xPSU online, 1xPSU offline | 198W | 200W |
OS run, 1xPSU online, 1xPSU pulled | 198W | 186W |