Analyze problem of AMO readings when a cable is broken/partyially broken or detached

[MSECode]

Here we are going to analyze the second problem mentioned here: https://github.com/icub-tech-iit/study-encoders/issues/22#issuecomment-2464924451, which is the issue of understanding how the AMO encoder is reading the signals if the cable is not detached but just partially broken.
In order to simulate that we just removed one of the data cables from the AMO connector.

[MSECode]

If we removed just one of the data cables of the AMO encoder connector we can see that the joint, if using the AMO encoder, is not able to calibrate as shown by the logs below.
Moreover, as shown by the plot we do not get any errors. As a matter of fact, we just reading zeros for both data and diagnostic.

AMO raw readings

Joint positions

Logs

log-yri-amo-cable-broken.log

Therefore, in case a cable is detached but not the entire connector, we have no information to know this, other than the fact that we are reading all zeros.

cc: @valegagge

[MSECode]

Later, we did another test by cutting one of the signal cables while the joint was moving. The output that we got from the diagnostic is the one of ERROR: eoerror_value_MC_abs_enc_invalid, which means that we actually consider to have invalid data in that condition. In that case the joint get stopped in the last position reached, thus it is moved to HW FAULT condition.
From a data point of view, as shown in the comment above, we reach the same condition, i.e. both the joint positions and the raw position values, as well as the diagnostic, are all zeros after the joint is moved in FAULT condition:

[valegagge]

Hi @MSECode please add a comment.

[MSECode]

During the tests we have observed that if either the AMO encoder cable is broken/detached or partially broken, i.e. one of the signal cables is detached from the connector, the joint is always moved in HW FAULT CONDITION.