We have several ODrive Pros (0.6.9) and and D6374 motors that are apparently very sensitive during calibration and have trouble passing the motor calibration step. When the state is set to motor calibration (after the appropriate motor properties are established), we can hear the motor energizing (not spinning) and after a second or so we hear a metallic “plink” from the motor. Dumping the errors reveals the “unbalanced phases” error. Only if I really jam the banana plug connectors together on the motor side (even if they appear fully engaged), can I get it to pass. Otherwise, across multiple Odrive Pro boards, and D6374 motors, we cannot pass the motor calibration.
The ODrive is getting 48V DC from a power supply running through the recommended clamp (STP-DRVA-RC-050A).
There is no load on the shaft during calibration.
We have the AMT212B-V-OD encoder mounted on the motor.
I have tried multiple calibration current values, both higher and lower than default, and also increased the resistance_calib_max_voltage value incrementally to no avail.
I have tried different gauge wires both thicker and equal to 16-gauge between the motor and the ODrive. The length of wire between the board and the motor is approximately 20in (50cm).
I have performed successful continuity tests with each phase, and measured the resistances between the phases and all produce similar values.
Now you might say “well if you can get it to pass by really forcing the connectors together, this is solved”, but I have a suspicion that this sensitivity may be the source of another problem that we have been seeing with the “successfully” calibrated motor: If we physically disconnect the motor from the board and reconnect it (while ensuring the same phase configuration before and after the disconnect), we occasionally see a severe reduction in the torque of the motor, or even a reversal in the direction of the motor movement. Even without the disconnect/reconnect, we have seen this motor weakness develop over time. The sneaking suspicion is that these problems are related, and that because that connection is so sensitive, whatever motor calibration was preserved from the initial calibration step is compromised on the reconnect/over time, causing phase-related motor weakness. I know from the FAQ that the usual suspect in the case of a motor getting hot and weak is encoder slippage but we have ensured that the encoder is firmly attached, and furthermore we mostly only see this problem when the motor is disconnected and reconnected via those banana connectors.
In an attempt to ensure good continuity between the motor and the ODrive, we even sacrificed the connectors on one of the motors by clipping off the existing connectors and replacing them with a more “secure” connector - or so we thought. We couldn’t get it to pass the calibration at all with that configuration, with the system throwing the same “Unbalanced Phases” error. EDIT: We see now why that specific configuration failed - there’s a conformal coating on the motor wires which compromised the conductivity.
What might we be doing wrong that is producing this outcome? Any help solving this mystery would be greatly appreciated. The next thing we are trying is to run shorter wires if only to see if that produces consistent successful calibrations.
Thanks ahead of time,