CHICAGO Five European automakers will take a major step this week toward creating an electrical standard that would enable them to eliminate hundreds of feet of wiring from the average car. Audi, BMW, DaimlerChrysler, Volvo and Volkswagen will team with Motorola and Volcano Communications Technologies to form a consortium that will lend support to the standardization of a new low-cost, low-speed network bus. They will announce the formation of the consortium Monday (March 6) at the Society of Automotive Engineers World Congress in Detroit.
The new network bus, known as Local Interconnect Network (LIN), will be used for simple on-off devices such as car seats, door locks, sun roofs, rain sensors, HVAC flaps, cruise control, windshield wipers and mirrors, as well as a host of other simple applications.
It's not yet known, however, whether American and Asian automakers will support the proposed standard. Apparent by their absence from the consortium were GM, Ford, Toyota, Nissan, Honda and semiconductor manufacturers other than Motorola.
Consortium members said they sent copies of the specifications to more than 120 companies worldwide, including the other major automakers. "Those companies have received specs," said Hans-Christian von der Wense, senior software design engineer for Motorola's Transportation Systems Group in Munich, Germany. "But European automakers want to stay in control of the steering committee. They want to make sure LIN works and fits what they really need."
Members of the consortium believe that the LIN bus will cut costs for automakers and simplify engineering for suppliers. LIN reportedly offers several cost advantages over controller-area network (CAN) buses, which are now heavily used in automotive applications.
The cost advantages of LIN include one signal wire instead of two; elimination of on-board voltage regulators and control modules; simpler electrical interfaces and receivers; and use of less-costly microcontrollers. Each LIN node would cost approximately $1 less per node than would a CAN bus, backers estimated.
"CAN nodes are quite expensive-more expensive than they need to be for simple applications," von der Wense said. "The auto industry needs a bus that can complement CAN in low-speed and low-performance applications."
Many automakers now use CAN buses to enable safety-critical systems such as antilock brakes, engine controllers, road-sensing suspensions and traction controllers to talk to one another.
CAN buses, however, have been too costly for wide acceptance in noncritical applications, such as door locks and power windows.
That's why many vehicles still use discrete point-to-point wiring to connect door locks, power windows and other simple devices to microcontrollers. The result is that many cars contain hundreds of feet of wiring in thick bundles within the doors and chassis.
In the search for a less costly bus for simple on-off applications, automakers in recent years have turned to the SAE J-1850 bus. J-1850, however, has failed to become an industry-wide standard. Most automakers have created their own proprietary versions of it. Toyota, for example, uses the Body Electronic Area Network bus. Ford, Chrysler and GM are said to have their own flavors of J-1850 as well. The result is that suppliers must rewrite drivers and redesign interfaces for each proprietary version of J-1850.
"It's well-known in the automotive industry that J-1850 is a semifailed standard," according to Scott Andrews, project general manager for Toyota's Electronics Engineering Division (San Jose, Calif.). "From a supplier standpoint, there's not a lot of benefit to using it. So if this LIN is an attempt to fix J-1850, then it's a good thing."
Proponents of the new LIN bus said that it could accomplish many of the same tasks as J-1850, but at lower cost. Operating at speeds of up to 20 kbits/second, it offers sufficient performance to send signals back and forth between, for example, the steering wheel and cruise-control system or radio. It does not, however, offer the kind of bandwidth needed for multimedia applications such as cell phones, navigation systems and CD players.
Most automakers now plan to use a CAN-based bus known as IDB-C for those high-bandwidth applications, since the bus operates at speeds of up to 250 kbits/s. "LIN is not going to replace CAN buses," said Andre Oberschachtsiek, manager of Volkswagen of America's Electronics Research Laboratory (Sunnyvale, Calif.). "It's for applications where it doesn't matter if the information comes 50 milliseconds sooner or later."
Each LIN bus will consist of a maximum of 16 nodes-one master and 15 slaves. Because the bus incorporates a synchronization mechanism for the slave nodes, no crystal resonator or clock source will be needed. That means that the microcontroller circuit board can be simpler and less costly. Power for the single-wire bus comes from the car's 12-volt battery.
Consortium members said that implementation of the LIN bus will be simple. A single chip will be placed in stepper motors or other types of motors at nodes along the LIN bus. The chip will then attach to three wires connected to the LIN bus.
The LIN bus uses the universal asynchronous receiver transmitter protocol, which was at one time commonly used for PC modem communication. Because many microcontrollers already use UART protocols, creators of the LIN standard believe that implementation will be simpler. "We wanted to use a protocol that was already in use," von der Wense of Motorola said. "We didn't want to have to reinvent anything." UART is typically not used in CAN-based buses, however, because it is too slow.
For communication and security, creators of the LIN standard expect to employ a gateway between the automakers' proprietary CAN buses and the LIN bus. The gateway would probably be contained in a microcontroller, maybe in a door module.
The key to the success of the LIN bus, however, may go beyond the technology. The force of the consortium, backed by Motorola and five of Europe's major automakers, may be enough to make car manufacturers comply with the standard. "We defined LIN's physical layer very narrowly," von der Wense said. "That way the car companies have to comply with the physical layer and configuration. We only want one flavor of this bus."
LIN's first applications are expected in the 2001 model year. Most of those initial applications will incorporate from three to ten LIN nodes, probably in door modules. Ultimately, however, consortium members said they expect the number of nodes to grow to about 20 per vehicle within five to ten years, reaching a worldwide volume of approximately 1.2 billion LIN nodes per year by the end of the decade.