Using CAN bus with NUCLEO boards (Demo for the CANnucleo library).

Dependencies:   CANnucleo mbed-dev

Dependents:   BMS_2 Can_sniffer_BMS_GER Can_sniffer_bms ECU_1

Using CAN bus with NUCLEO boards

Demo for the CANnucleo library


Because CAN support has been finally implemented into the mbed library also for the STM boards there is no need to use the CANnucleo library anymore (however you may if you want). See the CAN_Hello example which is trying to demonstrate the mbed built-in CAN API using NUCLEO boards.

Two low cost STM32F103C8T6 boards are connected to the same CAN bus via transceivers (MCP2551 or TJA1040, or etc.). CAN transceivers are not part of NUCLEO boards, therefore must be added by you. Remember also that CAN bus (even a short one) must be terminated with 120 Ohm resitors at both ends.


Zoom in



/media/uploads/hudakz/20150724_080148.jpg Zoom in

The mbed boards in this example are transmitting CAN messages carrying two data items:

uint8_t   counter;  // one byte
float     voltage;  // four bytes

So in this case the total length of payload data is five bytes (must not exceed eight bytes).
For our convenience, the "<<" (append) operator is used to add data to the CAN message.
The usage of "<<" and ">>" operators is similar to the C++ io-streams operators. We can append data one at a time

txMsg << counter;
txMsg << voltage;

or combine all into one expression.

txMsg << counter << voltage;

The actual data length of a CAN message is automatically updated when using "<<" or ">>" operators.
After successful transmission the CAN message is printed to the serial terminal of the connected PC. So we can check the details (ID, type, format, length and raw data). If something goes wrong during transmission a "Transmission error" message is printed to the serial terminal.

On arrival of a CAN message it's also printed to the serial terminal of the connected PC. So we can see the details (ID, type, format, length and raw data). Then its ID is checked. If there is a match with the ID of awaited message then data is extracted from the CAN message (in the same sequence as it was appended before transmitting) using the ">>" (extract) operator one at a time

rxMsg >> counter;
rxMsg >> voltage;

or all in one shot

rxMsg >> counter >> voltage;


Before compiling the project, in the mbed-dev library open the device.h file associated with the selected target board and add #undef DEVICE_CAN as follows:



#define DEVICE_ID_LENGTH       24


#include "objects.h"


NOTE: Failing to do so will result in compilation errors.

The same source code is used for both boards, but:

  • For board #1 compile the example without any change to main.cpp
  • For board #2 comment out the line #define BOARD1 1 before compiling

Once binaries have been downloaded to the boards, reset board #1.


The code published here was written for the official NUCLEO boards. When using STM32F103C8T6 boards, shown in the picture above (LED1 is connected to pin PC_13 and, via a resistor, to +3.3V),

  • Import the mbed-STM32F103C8T6 library into your project.
  • Include (uncomment) the line #define TARGET_STM32F103C8T6 1
  • Select NUCLEO-F103RB as target platform for the online compiler.

CAN bus related information

Download repository: zip gz

Files at revision 30:65c8d1465863

Name Size Actions
CANnucleo.lib 58 Revisions Annotate
main.cpp 6286 Revisions Annotate
mbed-dev.lib 64 Revisions Annotate