Checked what you said, PMBus is correctly connected, and fans 1 to 5 are connected to System Fan headers 1 to 5, respectively. Only System Fan 6 header is left unconnected.
Last time, I updated the f/w with package S2400SC_BIOS01060002_ME020105107_BMC0116r4010_FRUSDR0108.zip, which installed System BIOS - 01.06.0002, ME Firmware - 02.01.05.107, BMC Firmware - 01.16.r4010, FRUSDR - 1.08. I chose to update both FRU and SDR data (option 3), entered chassi S/N and P/N, then rebooted to my previously-installed Windows 7 Ultimate 64bit.
Upon accessing ASC, it informs me that P1 and P2 VRD HOT have been tripped, but that seems clearly mistaken as the processors heat sinks are cool to touch. It reports no memory and fan errors, however fan speeds are all to the 8000-rpm range. Processors' temperature do not exceed 35ºC.
I noticed that both PSUs rear fan are also running at full throttle. Is it some software or firmware update that I'm missing, or maybe do I have a corrupted BMC?
Well, additional information. BIOS is configured as follows:
Under System Acoustic and Performance Configuration:
Set Throttle Mode = AUTO
Altitude = 300m or less (we're at sea level)
Set Fan Profile = ACOUSTIC
PWM Offset = 0
Quiet Fan Idle Mode = DISABLED
Under Power and Performance:
CPU Power & Performance Policy = POWER
Lately I've downloaded the UEFI SEL Viewer, and I found out that something really uncommon is happening to the redundant fan array, and I suspect that's why the chassis fans are always at 100%, but still doesn't explain why the PSUs fan are also at 100%. It seems that Fan 4 keeps connecting and disconnecting ad eternum, raising SEL event 0x41 (Fan 4 Present) upon disconnection, which in turn raises event 0xC (Fan Redundancy), only to a few milliseconds later have another event 0x41 raised, and finally another event 0xC resuming the fan redundancy operation.
Here's a link to a saved log: Dropbox - fan4-5.sel - Simplify your life
After that, I reviewed and re-checked all fan and PSU connections, everything is firmly attached and connected. System fans are also individually spinning perfectly.
Any further tips?
Swapped fans with other "good" fans like 1 and 2, and the issue persists. The fans themselves appear to be working correctly. I'm wondering if the chassi hot-swap fan backplate isn't the culprit. I have assembled another system exactly equal to this last week and haven't had any of these issues.
Is there any way I can test the board to make sure at least it is OK? If I run FRUSDR with just fans 1-2-3 connected it will detect only these fans, will it then consider "normal" usage if only these 3 fans are connected, or it will keep boosting the fans because the full set of 5 is not completely present?
Well, swapping the fan headers didn't do any good. Physical fan #4 connected to fan header 1 works flawlessy, and any other physical fan connected to fan headers 4 or 5 keeps getting connected and disconnected immediately.
Ultimately, I had to RMA the board. Since then, no fan disconnection issues so far.