democratize access to car decoder rings
Go to file
Shane Smiskol 132b1a66b0
Bump pre-commit hooks, more coverage (similar to openpilot) (#1078)
* pytest defaults

* move some stuff to pyproject

* bump codespell

* more like openpilot
2024-08-16 22:24:44 -07:00
.github/workflows new docker image structure (#1070) 2024-07-31 21:14:59 -07:00
opendbc Bump pre-commit hooks, more coverage (similar to openpilot) (#1078) 2024-08-16 22:24:44 -07:00
site_scons/site_tools cython dependency scanner 2021-01-11 14:16:02 -08:00
.gitignore move to nested opendbc/ (#1069) 2024-07-31 22:46:43 -07:00
.pre-commit-config.yaml Bump pre-commit hooks, more coverage (similar to openpilot) (#1078) 2024-08-16 22:24:44 -07:00
Dockerfile move to nested opendbc/ (#1069) 2024-07-31 22:46:43 -07:00
LICENSE add license 2021-04-30 14:25:37 -07:00
README.md update README link to cabana 2022-11-21 00:08:22 -08:00
SConstruct CANParser: remove dependence on cereal (#934) 2024-07-31 00:10:20 -07:00
pyproject.toml Bump pre-commit hooks, more coverage (similar to openpilot) (#1078) 2024-08-16 22:24:44 -07:00
requirements.txt Convert tests to pytests (#1060) 2024-07-09 21:56:39 +00:00

README.md

DBC file basics

A DBC file encodes, in a humanly readable way, the information needed to understand a vehicle's CAN bus traffic. A vehicle might have multiple CAN buses and every CAN bus is represented by its own dbc file. Wondering what's the DBC file format? Here and Here a couple of good overviews.

How to start reverse engineering cars

opendbc is integrated with cabana.

Use panda to connect your car to a computer.

How to use reverse engineered DBC

To create custom CAN simulations or send reverse engineered signals back to the car you can use CANdevStudio project.

DBC file preprocessor

DBC files for different models of the same brand have a lot of overlap. Therefore, we wrote a preprocessor to create DBC files from a brand DBC file and a model specific DBC file. The source DBC files can be found in the generator folder. After changing one of the files run the generator.py script to regenerate the output files. These output files will be placed in the root of the opendbc repository and are suffixed by _generated.

Good practices for contributing to opendbc

  • Comments: the best way to store comments is to add them directly to the DBC files. For example:

    CM_ SG_ 490 LONG_ACCEL "wheel speed derivative, noisy and zero snapping";
    

    is a comment that refers to signal LONG_ACCEL in message 490. Using comments is highly recommended, especially for doubts and uncertainties. cabana can easily display/add/edit comments to signals and messages.

  • Units: when applicable, it's recommended to convert signals into physical units, by using a proper signal factor. Using a SI unit is preferred, unless a non-SI unit rounds the signal factor much better. For example:

    SG_ VEHICLE_SPEED : 7|15@0+ (0.00278,0) [0|70] "m/s" PCM
    

    is better than:

    SG_ VEHICLE_SPEED : 7|15@0+ (0.00620,0) [0|115] "mph" PCM
    

    However, the cleanest option is really:

    SG_ VEHICLE_SPEED : 7|15@0+ (0.01,0) [0|250] "kph" PCM
    
  • Signal size: always use the smallest amount of bits possible. For example, let's say I'm reverse engineering the gas pedal position and I've determined that it's in a 3 bytes message. For 0% pedal position I read a message value of 0x00 0x00 0x00, while for 100% of pedal position I read 0x64 0x00 0x00: clearly, the gas pedal position is within the first byte of the message and I might be tempted to define the signal GAS_POS as:

    SG_ GAS_POS : 7|8@0+ (1,0) [0|100] "%" PCM
    

    However, I can't be sure that the very first bit of the message is referred to the pedal position: I haven't seen it changing! Therefore, a safer way of defining the signal is:

    SG_ GAS_POS : 6|7@0+ (1,0) [0|100] "%" PCM
    

    which leaves the first bit unallocated. This prevents from very erroneous reading of the gas pedal position, in case the first bit is indeed used for something else.