MCU 'mcu' shutdown: Timer too close during BED_MESH_CALIBRATE

[Sineos]

@tettabeta

why do you say “I have never seen a log like this”? What is unusual?

See the paste in my above comment. I have never seen such a pattern before

@tettabeta

is it possible that a (much) higher clock MCU leads to the host to be overloaded, even with the very same step rates?

If this would be the case the error would be Rescheduled timer in the past

@tettabeta

• can you confirm from the log that this is actually a “MCU timer too late” (i.e., received commands are in the past or too close to current time) and there are no prior issues?
• is there anything else that can cause “MCU timer too late”, apart what is listed in the knowledge base, that we can investigate?

What do you want me to confirm? See the log.
This type of error occurs when the host schedules a command that is (from the MCU perspective) already in the past. Thats why the typical reason is an overloaded / too slow host.

In addition you have a

mcu 'mcu': got {'#receive_time': 3963.727281821, u'next_clock': 1147228928, u'oid': 33, u'value': 30703, '#name': u'analog_in_state', '#sent_time': 3963.681299046}
Resetting prediction variance 3963.733: freq=400000000 diff=4129047 stddev=400000.000

which indicates the host software was unable to make accurate predictions of the micro-controller clock. Usually only a warning but in combination with a shutdown points to some serious issues like failing hardware.

Personally I would expect that in your case the timing is rather messed up on the MCU side but as stated before, I’m neither a hard- nor software developer.