A LIN (local interconnect network) in-vehicle networking system—mainly used in automotive body and comfort applications for control of subsystems such as air conditioning, door locking, and mirror position—consists of a single master node and several slave nodes, connected with a single communication wire with a communication rate of up to 19,200 bits per second.
For correct LIN operation, each node must have a unique address assigned to it prior to the start of the normal communication mode. Two methodologies exist for setting slave-node addresses. The first is by defining the addresses by the node hardware and implementing them with hard wiring, a special connector, OTP (one-time programmable) programming, or configuration DIP (dual inline package) switches. A second technique is by having the master node assign slave-node addresses during full-system power-up after network installation or maintenance; the individual nodes have no predefined addresses before they are connected to the network. The process of slave-node address assignment at network startup is called auto addressing or SNPD (slave-node position detection).
Multiple nodes on the same LIN bus can perform similar functions while differing just by their LIN addresses. If the network uses predefined addresses, then the replacement of a defective node requires either manual intervention when hardware is added or vendors must have a stock of spare LIN nodes with different addresses. Both of these solutions increase the cost and complexity of LIN network maintenance.
The generic implementation of node-position detection requires a split of the LIN line at each node so that the node physical layer can electrically distinguish between two bus portions branching from the node. This bus split allows a node to determine, relative to itself, “left” or “right” nodes (or “before” and “after” nodes).
There are two main methods of node-position detection: bus line-current measurement and dual-LIN transceivers. The two position-detection routes differ in the way they handle the bus line split. The current measurement approach requires the master and slave nodes to be interconnected as a single linear bus with the master on one end and the bus line passing progressively through all slave nodes without any branching; this limits the network topology. In addition, if a failure occurs, the current-measurement approach does not give any information on the position of the failed node; the master can only signal a failure without specifying its location. A third shortcoming of the current-measurement method is that the small signals detected are susceptible to electromagnetic interference (EMI).
AMIS (AMI Semiconductor) engineers have proposed an alternative architecture for node-position detection that uses the same generic principal as the bus line-current measurement method. The bus line is split at each slave node attached to the bus with two pins: LIN1 and LIN2. The resulting bus portions are not constantly interconnected via measurement resistors but are ordered in a “daisy chain” controlled by switches placed between LIN1 and LIN2. Each slave node can either propagate or block the communication between the bus sections. An example of this implementation is shown above.
After system startup or a reset, all slaves will be in their initial state and all LIN1/LIN2 switches are open. As a consequence, all LIN bus portions remain separated and only the node connected directly to the master will be able to react to incoming LIN message headers from the master. After the master sends an initialization command message, the slave takes an address according to the message. When the address is successfully assigned, the slave closes its switch to the following portion of the bus.
At the end of the first step, Slave 1 has a defined address and the second slave node is ready to communicate with the master to get its address. The process continues until all nodes have gotten their addresses. After the auto-addressing is finished, the system enters normal operation and the partitioned bus behaves as one LIN network.
The use of a daisy-chain node connection lets the master node localize a defective node since the master has immediate feedback on the success of the address assignment. In addition, the node-position detection is based on message exchange with normal signal levels, which means that the auto-addressing robustness is as high as that of the normal mode communication.
Another advantage of the daisy-chain method is its simplicity; only the basic LIN transceiver is used twice in the same node with a purely digital interconnection, while the current measurement requires the design of a relatively precise analog voltage measurement circuit intended to operate under harsh conditions. A dual-LIN node is therefore easier and faster to design.
The daisy-chain method also supports the use of other topologies besides a single linear bus with the master node on one extremity. Correct node-position detection is possible since each dual-LIN node is connected to the branching point of the tree with a different LIN port (LIN1 or LIN2). The auto-addressing can distinguish on which branch the node is connected. Examples of other allowed topologies include loops and nested loops.
In summary, the dual-LIN transceiver method for SNPD offers several advantages over the line-current measurement. It uses normal signal levels and is thus more robust; it increases fault tolerance by detecting the failing node; the bus topology is not limited to a linear bus; and SNPD requires less effort and less design risk, thus reducing LIN-based system design cost.
Pavel Drazdil and Geert Vandensande, AMI Semiconductor, wrote this article for SAE Magazines.