• State TI Thinks Resolved
  • Replies 3 replies
  • Answers 1 answer
  • Subscribers 34 subscribers
  • Views 113 views
  • Users 0 members are here
Support feedback
Options
Options
Related

TCAN1046A-Q1: 1 to 2 canbus Transceiver

Jeff YANG
Mastermind 9355 points

Part Number: TCAN1046A-Q1
Other Parts Discussed in Thread: TCAN1042-Q1, TCAN4550-Q1

Hi,

Hi TI Team,

I am currently working on a design where the MCU only has a single CAN controller (one set of TXD/RXD), but we need to split it into two independent CAN bus networks at the physical layer to support a star/tree topology.

Since TI doesn't seem to have a dedicated single-chip 1-to-2 CAN repeater IC with built-in routing logic, I am considering the following hardware implementation:

I plan to use a Dual CAN Transceiver (like the TCAN1046A-Q1) or two single-channel transceivers (like TCAN1042-Q1), and pair them with simple Logic ICs on the digital side. The specific connections would be:

  1. RX path: Connect the RXD pins of both transceivers into an AND gate, and feed the output to the MCU's RX pin.

  2. TX path: Fan out the MCU's TX signal simultaneously to the TXD pins of both transceivers.

This seems like a cost-effective way to achieve a 1-to-2 physical layer expansion. I would like to consult with you on the following points:

  1. Feasibility: Is this logic-gate-based repeater implementation feasible and robust in practical applications?

  2. Potential Caveats: Are there any critical issues I should be aware of? Specifically regarding loop delay (propagation delay through the logic gates + transceivers), arbitration handling, or error frame collisions if the two buses encounter different states simultaneously.

  3. Alternative Solutions: Does TI have any other recommended approaches or reference designs for this specific 1-to-2 CAN expansion requirement?

Thanks in advance for your support!

Jeff

  • 0
    TI__Guru* 80132 points

    Hi Jeff,

    The proposed logic-gate approach is electrically feasible for a simple CAN TX fan-out /RX aggregation. However, it should not be considered a true 1-to-2 CAN repeater if the two CAN networks must operate independently.

    • TX path: MCU TXD can drive both transceivers simultaneously, so the same transmitted frame appears on both CAN buses.
    • RX path: ANDing the two RXD signals works logically because CAN dominant is LOW; however, if the two buses have different traffic simultaneously, the MCU receives a bitwise combination that may not correspond to either valid CAN frame.
    • Arbitration / ACK / errors: The single CAN controller mat not be able to distinguish which bus caused a dominant bit, arbitration loss, ACK, or error frame. This can lead to incorrect arbitration / error handling when the buses operate independently.
    • Propagation delay: Yes, dditional transceiver and logic-gate delay reduces timing margin, particularly at higher CAN FD data rates. Hence, total delay should be the specified delay + the delay from everything else in the path. 
    • Recommendation: For genuinely independent CAN networks, use two independent CAN controllers, e.g. the MCU’s native CAN controller for one bus plus a TCAN4550-Q1 for example connected through SPI for the second bus. Firmware can then handle the required A<->B message routing.
    • If the requirement is only to duplicate MCU traffic onto two physical segments and simultaneous independent traffic is not expected, the dual-transceiver + logic-gate approach may be feasible with appropriate timing and system-level validation.

    Hence, the proposed circuit is better described as a CAN TX fan-out / RX aggregation solution, rather than a transparent 1-to-2 CAN repeater or independent CAN bridge, thanks. 

    Best Regards,

    Michael. 

  • 0 in reply to Michael Ikwuyum
    Mastermind 9355 points

    Hi Michael,

    Got it, thank you!

    Jeff

  • 0 in reply to Jeff YANG
    TI__Guru* 80132 points

    Hi Jeff,

    You are welcome. Could you please help resolve if no additional questions? Thanks.

    Best Regards,

    Michael.