Python BrushLess Motor Control

---

Note: This package was mostly unmaintained for a while. The master branch is now ported to Python 3 but not much testing has been done yet, so not everything may work out of the box. For the old catkin/Python2-version see the v1.0.0 tag.

---

The blmc package contains an API for the CAN interface of the motor control microcontrollers of the Open Dynamic Robot Initiative as well as the OptoForce sensor.

The scripts in the scripts directory use this API for several tasks like evaluating the CAN interface, displaying motor parameters or running simple velocity/position controllers.

This assumes that the CAN-firmware from udriver_firmware is running on the board.

Note: This package is mostly meant for quick tests of the hardware. It is not optimised for speed, so for actual control applications where a high control rate is important, we recommend to use the C++ API instead, see the blmc_drivers package.

Requirements

• python-can
  Note: When using on Windows with PCAN, make sure you have version >= 1.5.1
• numpy
• LaunchPad with two motors, running the dual_motor_torque_ctrl program and connected to the computer via CAN.

Installation

Install from source using pip:

git clone https://github.com/open-dynamic-robot-initiative/python_blmc.git
cd python_blmc
pip install .

Further steps:

1. You need to set up a configuration file for python-can to specify which driver/prot to use. See the python-can documentation
2. Some of the demo scripts need a prior calibration the offset between zero and stop position of the leg. Run the script calibrate.py (see below) to do this.

About the Scripts

There are a bunch of scripts that can be used for fast analysing and testing of a motor. At the same time these scripts serve as examples on how to use the CAN API in the blmc Python module (at least for now, these examples are the only documentation...).

For short description of the single scripts, see Example Scripts.

Calibrate / Initialize the Leg Position

Since the motor encoders only provide relative information, the absolute position of the leg at start is unknown. For the demos, a simple initialization step is implemented, where the user is asked to manually move the joints of the leg in positive direction until the limit stop. Based on this, the zero position (where the leg is fully stretched) is determined.

For this to work, the offset between limit stop and zero position has to be known. To determine it, the script leg_calibrate.py is used. Simply run the script and follow its instructions. The result is written to the file "calibration_data.p" which is then loaded by the other scripts.

A demo of how the initialization is done in the code is found in leg_init_position.py. This script can also be used to verify that the calibration is correct.