Not every signal in a car needs CAN's bandwidth or cost. LIN is the cheaper network for the ones that don't.
A window switch, a mirror adjustment motor, a rain sensor, or a seat position module doesn't need CAN's speed, robustness, or cost. LIN (Local Interconnect Network) is a deliberately simple, low-cost single-wire bus for exactly this class of device - typically used within one sub-system (a door, a seat) that then connects to the vehicle's main CAN network through a gateway ECU, rather than every LIN device talking on CAN directly.
| CAN | LIN | |
|---|---|---|
| Wires | 2 (differential pair) | 1 (plus ground/power) |
| Speed | up to 1 Mbit/s (classic) | up to 20 kbit/s |
| Arbitration | Multi-master, priority-based | Single master, polled |
| Typical cost per node | Higher | Lower - simpler transceiver silicon |
This is LIN's biggest structural difference from CAN. There's no arbitration contest because there's no contest to have: one master node (usually the gateway ECU for that sub-system) decides what happens and when, sending out a "header" that names which slave node should respond next. That slave then sends the "response" - the actual data. Nothing on a LIN bus transmits unprompted, which is what keeps the protocol and the slave-node silicon simple enough to be cheap.
Because everything is scheduled by the master, a LIN network's entire message timing is deterministic and defined up front in a LIN Description File (LDF) - LIN's equivalent of a DBC file, but also describing the schedule (which message is polled when), not just the signal layout. Capturing and decoding LIN traffic means listening in on the same header/response exchanges as a passive observer, using the same signal-scaling concepts as CAN (see DBC Files & Signal Decoding) once you have the message layout.
LIN capture sits alongside CAN/CAN FD as part of ONAC's broader vehicle network systems work - particularly relevant for body-electronics and NVH investigations where a LIN-controlled component (a motor, a sensor) is a suspect.