ergoCub 1.3 S/N:002 – Intermittent issues with head board `head-eb21-j3`

[SimoneMic]

Robot Name 🤖

ergoCub 1.3 S/N:002

Request/Failure description

After several attempts the robot started correctly, but many times the following error happened at startup:

[ERROR] theNVmanager::Impl::ask() had a timeout for BOARD head-eb21-j3 IP 10.0.1.21 and nv ID32 = 0x00000002 -> IND = 0, TAG = eoprot_tag_mn_comm_status_managementprotocolversion

Blocking thus the calibration.
The solution was to restart the motors each time and re-launching the calibration.

Detailed context

At first the robot wasn't starting with errors in the like:

[ERROR] theNVmanager::Impl::ask() had a timeout for BOARD head-eb21-j3 IP 10.0.1.21 and nv ID32 = 0x00000002 -> IND = 0, TAG = eoprot_tag_mn_comm_status_managementprotocolversion

and

[ERROR] theNVmanager::Impl::wait() had a timeout for BOARD head-eb21-j3 IP 10.0.1.21 and nv ID32 = 0x00030001 -> IND = 0, TAG = eoprot_tag_mn_service_status_commandresult
[ERROR] theNVmanager::Impl::command() fails a wait() from IP 10.0.1.21 for nv ID32 = 0x00030001 -> IND = 0, TAG = eoprot_tag_mn_service_status_commandresult
[ERROR] EthResource::serviceCommand() failed an acked activation request to BOARD head-eb21-j3 with IP 10.0.1.21 after 3 attempts each with waiting timeout of 0.5 seconds

After several attempts, and trying to disable the head, the robot started "normally" with the head in the following state:

The pitch joint is in: No fault detected fault state

Finally the robot started normally, at the 5th attempt.

Additional context

Probably it's a wiring issue (intermittent contact)

How does it affect you?

No response

[S-Dafarra]

I had a similar issue, and I noticed that from FirmwareUpdater the eb20 board was appearing in maintenance mode. I forced it to application and I rebooted the motors. After that the robot started.

[S-Dafarra]

Maybe it is not an electrical issue, since the same was happening also on the wrist of the SN000 and the board is again an AMC cc @valegagge @marcoaccame @MSECode

[MSECode]

Hi,
yes it might be that the board did not boot correctly. The amc code has been modified quite a lot lately so it might be good to monitor those behaviors.

[valegagge]

From the logic, it seems that the board at address 10.0.1.21 is an MC4plus.
The log you provided indicates that the board is not responding, which causes the yarprobotinterface to fail.

Please also keep in mind that the head joints are controlled by the AMC board, while the MC4plus only drives the camera tilt.
If the MC4plus is rebooting, the head joints cannot complete calibration, since the AMC board is connected in the daisy chain through the MC4plus.
In that case, communication to and from the AMC is interrupted.

That said, I believe the error related to the head pitch may have a different root cause.

---

To investigate further, I have a few questions:

@SimoneMic

1. Could you tell me the firmware version of the MC4plus at address 10.0.1.21?
2. If you disable the head, do the lower arms start correctly?
3. What is the firmware version of the boards at addresses 10.0.1.22, 10.0.1.23, 10.0.1.24, and 10.0.1.25?

---

Thanks in advance!