This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

AM2634-Q1: root cause of CAN bus-off

Part Number: AM2634-Q1
Other Parts Discussed in Thread: AM2634

Hi, experts,

Customer encounters bus off on car, need you help to find the root cause.

The bus off occurs only on car, it can not be reproduced in lab. It occurs several times on different cars, but takes quite long time to be reproduced on individual car. While trying to reproduce the bus off condition, customer feedbacks that it is more likely to occurs during power off/on the whole vehicle.

When bus off condition occurs, there is error frame on the CAN bus, it is sent from AM2634.

The CAN transceiver is TJA1145, its status is normal during the bus off condition.

Below are registers dumped from AM2634 in bus off condition.

MCAN0_CFG_PSR Register:0x000007E7

MCAN0_CFG_IR Register   :  0x1BC00001

MCAN0_CFG_ECR Register:  0x002000F8

Please help to give some ideas on finding the root cause of bus off.

Thanks.

  • Hello Fan,

    Thanks for the initial details. In addition to the software details provided, will there be any possibility for the CAN bus lines to be monitored from a hardware standpoint to observe the communication on the bus during the issue? Ideally with an oscilloscope to monitor the health of the lines as it's only occurring in the full vehicle and not in lab environment.

    I will forward these details to our software experts for review, but please inform whether there is possibility to monitor the physical bus lines as part of the analysis.

    Best Regards,

    Ralph Jacobi

  • Hello Fan,

    The registers have not provided anything useful to determine root cause for the bus off, so getting captures of the CAN bus will be important for this. Can you advise on how the customer can acquire such?

    Also does the issue only occur during power off / on, or can it occur after the vehicle & CAN bus have been powered on for a while?

    Best Regards,

    Ralph Jacobi

  • Hi, Ralph,

    Understood.

    As the issue can only be reproduced on car, which involves more difficulties on capturing the signals in physical bus lines. Customer engineers are trying their best to get it.

    Thanks.

  • Hi Fan,

    Very much understandable, see if you can find details about whether it only happens on power off / on as if it happens while CAN is powered on for a while that may indicate it could be software related more than hardware.

    Best Regards,

    Ralph Jacobi

  • Hi, Ralph,

    Please check below signal captured during bus-off occurrence, and share your comment.

    Yellow: CAN_H

    Blue: TX

    Purple: RX

  • Hello Fan,

    Thanks for sharing these captures. This gives us a much better idea what is occurring on the bus now. We are reviewing them and another expert will provide some feedback and queries by end of Monday at latest.

    Best Regards,

    Ralph Jacobi

  • Hello Fan,

    After reviewing the information provided there a few points and follow up questions that arise:

    - There is a lot of overshoot and undershoot on CAN_H line (Assuming the same of the CAN_L line). Which is typical, but it seems a bit excessive.

    - Looks like there could also be some slight impedance mismatch on the CAN_TX line causing a bit of ringing.

    - Please verify that all nodes on the bus are using the same baud rate as this is the most common cause of repeated Transmit Errors.

    1) Is it be possible to see the termination circuits/HW implementation between device and transceiver and transceiver and bus?

    2) Is it possible to provide details on the scope/probe used to make these measurements? Probing these signals could likely in itself cause some interference/noise/skew in the system signals depending on the type and quality.

    3) Can an OBD reader be connected to the CAN harness and any warnings/errors reported?

    4)  We noticed that the transceiver in use was designed to work with a CAN Bus containing both CAN and CAN-FD transmissions. What are the system expectations for the AM263x device? (Baud Rate, CAN/CAN-FD, # of Nodes, Cable connection Length)

    Thanks and Regards,

    Zackary Fleenor

  • Hey Fan,

    Just following up on this thread. Were you able to resolve this issue with the CAN interface?

    Best Regards,

    Zackary Fleenor

  • Hi, Fleenor,

    Bus off can be recovered in 60ms, but data in this period is missed. Customer wants to find the root cause of bus off and try to avoid it.  

    Thanks.

  • Hey Fan,

    Please provide feedback on the previous response:
    "

    Hello Fan,

    After reviewing the information provided there a few points and follow up questions that arise:

    - There is a lot of overshoot and undershoot on CAN_H line (Assuming the same of the CAN_L line). Which is typical, but it seems a bit excessive.

    - Looks like there could also be some slight impedance mismatch on the CAN_TX line causing a bit of ringing.

    - Please verify that all nodes on the bus are using the same baud rate as this is the most common cause of repeated Transmit Errors.

    1) Is it be possible to see the termination circuits/HW implementation between device and transceiver and transceiver and bus?

    2) Is it possible to provide details on the scope/probe used to make these measurements? Probing these signals could likely in itself cause some interference/noise/skew in the system signals depending on the type and quality.

    3) Can an OBD reader be connected to the CAN harness and any warnings/errors reported?

    4)  We noticed that the transceiver in use was designed to work with a CAN Bus containing both CAN and CAN-FD transmissions. What are the system expectations for the AM263x device? (Baud Rate, CAN/CAN-FD, # of Nodes, Cable connection Length)

    Thanks and Regards,

    Zackary Fleenor

    "

  • Hi Fan,

    Customer wants to find the root cause of bus off and try to avoid it.  

    As Zack requested, we really need more information here to help with this. So answers to those questions are key next steps to root cause this issue. Without more details we won't be able to help determine root cause. Please help get these crucial answers from the customer.

    Best Regards,

    Ralph Jacobi

  • Hi, Zack and Ralph,

    I am checking with customer on above questions. Thanks. 

  • Hey Fan,

    We are looking forward to any updates here so we can better understand and help to resolve this issue.

    Best Regards,

    Zackary Fleenor

  • Hi, Zack,

    Thanks for your patient.

    I am waiting for customers' feedback on above questions.