Urgent - Index pulse counter

r91 · May 14, 2020, 12:12pm

First a question:
By looking into the firmware It seems that the index pulse is only used once to zero the encoder, is this correct? @madcowswe

github.com

madcowswe/ODrive/blob/7148de710abb5701dd2e24ca8e549121870fefa5/Firmware/MotorControl/encoder.cpp#L59


        } else {
            // We can't use the update_offset facility in set_circular_count because
            // we also set the linear count before there is a chance to update. Therefore:
            // Invalidate offset calibration that may have happened before idx search
            is_ready_ = false;
        }
        index_found_ = true;
    }


    // Disable interrupt
    GPIO_unsubscribe(hw_config_.index_port, hw_config_.index_pin);
}


void Encoder::set_idx_subscribe(bool override_enable) {
    if (config_.use_index && (override_enable || !config_.find_idx_on_lockin_only)) {
        GPIO_subscribe(hw_config_.index_port, hw_config_.index_pin, GPIO_PULLDOWN,
                enc_index_cb_wrapper, this);
    } else if (!config_.use_index || config_.find_idx_on_lockin_only) {
        GPIO_unsubscribe(hw_config_.index_port, hw_config_.index_pin);
    }
}

If this is the case, do you have any tips on how to implement the encoder counter so that takes the index pulse into consideration?

In our case, we believe we have some noisy encoder signals, and we experienced that after driving our motor for a while (15 min), it misbehaved, and did not go to the set position (sometimes even went in the other direction). The misbehaviour was gradually increased until it completely stopped. We also noticed that the torque was dropping. So we believe the encoder is driting after some time.
If the index pulse is only zeroed once would explain why we see this misbehavior.

Wetmelon · May 18, 2020, 8:56pm

You’ll probably want to set the circular and linear positions to 0 and some multiple of the cpr, respectively.