TI AUTOMOTIVE · interfaces
LIN or CAN FD: start with message deadlines, harness and sleep policy
Use TLIN1029-Q1 and TI CAN candidates to plan body subnets, separating 20 kbps LIN transmission, 100 kbps programming reception, controller roles and gateway acceptance.
Sources reviewed 2026-11-06 · NuAxi technical editorial · source facts and engineering interpretation distinguished
Key answer
Evaluate LIN first for low-rate actuator and sensor subnets that can be scheduled by a commander. Evaluate CAN/CAN FD when higher throughput, priority-based bus arbitration or integration with an existing CAN network is needed. TLIN1029-Q1 is a LIN physical-layer device intended for 12 V applications, transmitting up to 20 kbps. Its receiver's 100 kbps programming capability is not a 100 kbps LIN-network rating. Neither transceiver alone implements a cross-protocol gateway.
1. Define the traffic contract before choosing the transceiver
For each signal, specify data volume, update period, maximum acceptable latency, fault response and who wakes a sleeping node. Periodic seat-position, window-switch or HVAC-flap updates can be LIN-subnet candidates, but a body-electronics label is not enough: if waiting for a schedule cycle violates the deadline, revise the schedule, split the subnet or choose another network architecture.
LIN transmission opportunities are scheduled by the commander, whereas CAN controllers perform bus arbitration. Both require software to manage messages, timeouts and fault responses. A faster physical layer does not automatically solve the worst-case waiting time of low-priority traffic. Compare normal traffic, diagnostics and fault retries for the same message set rather than comparing only 20 kbps with a multi-Mbps headline.
2. Compare LIN and CAN/CAN FD under the same system constraints
LIN's single signal wire and CAN's differential pair are different electrical architectures; their termination and protection networks are not interchangeable. The table combines selection reasons with validation responsibilities. It does not mandate a protocol for an application class or treat transceiver protection as complete system fault tolerance.
Scroll wide tables horizontally; column headers remain visible within the table.
| Decision dimension | LIN: TLIN1029-Q1 | CAN/CAN FD: TI CAN candidates |
|---|---|---|
| Bus connection | Single signal wire; commander and responder termination differ | CANH/CANL differential pair; termination and stubs follow the harness topology |
| Transmission timing | Commander schedule; budget worst-case polling and diagnostic occupancy | Controller arbitration; budget priority blocking, load and retries |
| Rate boundary | Transmit up to 20 kbps; 100 kbps refers only to programming reception | Select by device and controller; TCAN1044A's 8 Mbps applies to simpler networks |
| Host responsibility | MCU LIN/UART peripheral plus protocol software; the transceiver does not generate schedules | A CAN controller is required; a host without one can evaluate TCAN4550 over SPI |
| Supply and wake | Intended for 12 V applications, 4–36 V supply; LIN-bus or EN wake | Standby, INH power control or WUF selective wake depend on the exact device |
| Useful first evaluation | Whether a low-rate periodic subnet meets deadlines with the full schedule and diagnostics | Existing-network compatibility, arbitration deadlines and FD data-phase margin |
3. Three common misreadings of TLIN1029-Q1
First, 20 kbps and 100 kbps apply to different directions and uses. The LIN transmitter supports network operation up to 20 kbps; receiver support up to 100 kbps enables faster in-line programming. It does not upgrade the entire bus, every node or normal protocol scheduling to 100 kbps. A programming fixture still needs a compatible transmitter, electrical waveform, host serial configuration and target receiver.
Second, a 4–36 V supply range does not establish drop-in suitability for any 24 V LIN network. TI explicitly positions TLIN1029-Q1 for 12 V applications. Check bus thresholds, termination, ground offsets and vehicle transient requirements rather than only the VSUP ceiling. Third, RXD is open drain and requires an appropriate logic-level pull-up. The LIN-pin termination resistor/diode network and the RXD logic pull-up serve different purposes.
4. A subnet and gateway decision that can be calculated
Consider a zone controller collecting several door or seat nodes and reporting upstream over CAN. LIN can serve periodic downstream collection and CAN the upstream messages, but the zone controller must explicitly manage freshness, queuing, timeouts and state transitions. Wiring two transceivers' TXD/RXD together does not create a gateway. If a downstream response is lost, the upstream message should identify invalid or stale data instead of presenting the old value as a new sample.
As an engineering calculation, eight payload bytes at 20 kbps require 8 × 8 ÷ 20000 = 3.2 ms for the payload bits alone. This excludes serial start/stop bits, header, checksum, gaps and schedule waiting, so 3.2 ms cannot be used as the complete deadline for an eight-byte transaction. Build the full schedule cycle, then add diagnostics and exception handling. If it exceeds the deadline, consider splitting the subnet, reducing refresh rates or migrating traffic to CAN rather than using the receiver's programming rate as a shortcut.
If the zone controller already has a suitable CAN FD controller, TCAN1042-Q1 or TCAN1044A-Q1 can be upstream physical-layer candidates. If it lacks a controller port, evaluate TCAN4550-Q1, but include SPI servicing and buffering in the gateway's end-to-end budget. Measure from the downstream event to valid data reaching the upstream application, not merely the duration of one CAN transmission.
5. Map each wake source to the power domains it enables
TLIN1029-Q1 supports LIN-bus and local EN wake from low-power operation, but wake detection does not mean the application is immediately ready. Check whether the MCU retains power, how RXD reaches a wake interrupt, the clock/peripheral restart time and whether the first scheduled exchange arrives before the node is ready. A sleep-current budget must include the whole node, not just the transceiver's typical sleep current.
On the CAN side, TCAN1043A-Q1 provides INH to enable external power. If unrelated network traffic must leave the ECU asleep, evaluate TCAN1145-Q1's WUF partial-networking capability instead. INH power gating and selective frame matching are different mechanisms. The gateway must define whether downstream LIN wake enables upstream CAN and whether upstream wake starts every LIN subnet, with a state machine that prevents either side from repeatedly waking the other.
6. Practical checklist before freezing the selection
Use the same input messages and fault conditions for both protocol evaluations so the choice is defensible. Record normal traffic, diagnostics, power recovery and wake behavior separately. A successful exchange proves neither schedule deadlines nor electrical margin on the vehicle harness.
- List each message's bytes, period, priority/slot, maximum latency and stale-data behavior, including diagnostics and retries.
- Check LIN termination by commander/responder role and CAN differential termination by topology; review ground returns and harness connections.
- Evaluate normal TLIN1029 transmission at up to 20 kbps; create a separate fixture and waveform test for any 100 kbps programming-receive case.
- Confirm the 12 V application conditions, 4–36 V operating range, RXD pull-up voltage and supply-transient protection; withstand ratings are not operating conditions.
- Verify system behavior under a stuck-dominant bus, unpowered nodes, ground offsets and communication timeout, not only whether the transceiver resets.
- Measure wake-to-first-valid-application-data time and whole-node sleep current for LIN, CAN and both gateway wake directions.
- If TCAN4550 expands CAN, include SPI service delays under host load and verify consistency between downstream sample times and upstream timestamps.
Frequently asked questions
Can TLIN1029-Q1's 100 kbps receiver speed up a normal LIN network?
Not on that basis. Its transmit specification is up to 20 kbps, while 100 kbps reception targets faster in-line programming. The transmitter, other nodes and protocol schedule are not upgraded by that capability; validate programming as a separate operating case.
Does a 36 V supply limit make TLIN1029-Q1 a drop-in choice for 24 V LIN?
No. TI intends it for 12 V applications; 4–36 V is the device's supply operating range. A 24 V network still requires checking bus thresholds, termination, ground offsets and transients. The supply ceiling alone does not establish suitability.
Can UART software be retained by simply replacing a LIN transceiver with a CAN transceiver?
That is not a protocol migration. CAN needs a controller for framing, arbitration and error handling, and its harness and termination differ. A host without a CAN controller can evaluate TCAN4550-Q1, but still needs a SPI driver, message mapping and error handling.
Why might a low-rate actuator still use CAN instead of LIN?
Payload volume is only one factor. Existing interfaces, worst-case response time, diagnostic occupancy, controller resources and wake policy can determine the choice. Calculate LIN schedule waiting and CAN arbitration waiting for the full message set; neither protocol meets a deadline merely because of its peak rate.
Related parts and complete OPNs
Continue reading
NuAxi is independent of TI. This guide organizes official evidence and engineering checks; it is not substitution approval, system certification, procurement assurance or completed design validation. Review the exact OPN, datasheet revision and project requirements.