Heard of CAN buses? Of Smart Connectors? Don't know what they're all about? Here's the story on how conventional wiring harnesses are going the way of the dodo. Nearly all current cars use communication buses to allow the various electronic components to 'talk' with one another. Even the Holden Barina, one of Australia's cheapest new cars, uses two data buses - a high-speed one that allows the engine and safety (airbag, ABS) systems to communicate and a lower speed bus that handles items such and the lighting, windscreen wipers and doorlocks. More complex cars have up to seven in-vehicle multiplexed communication buses. So what actually are these things and what benefits do they provide? GrowthThe amount of electronics in cars has skyrocketed over the last few decades. With electronic control of the engine, brakes and transmission being now nearly universal and with the widespread application of much more sophisticated entertainment systems, there has been a massive increase in the size and complexity of car wiring harnesses. In fact the length of the harness has increased from typically 45 metres in vehicles produced in 1955 to more than 4000 metres in many of today's luxury vehicles. In fact, it has been suggested that car wiring harnesses were fast on their way to becoming the most expensive and complicated components in the whole of the vehicle's electrical system! The use of multiplexed networks have been used to reduce this wiring complexity, resulting in:
On the weight factor alone, Motorola stated in 1998 that replacing wiring harnesses with Local Area Networks in each of the four doors of a BMW reduced that vehicle's weight by 15kg. However, multiplexing has in many cases proved to be more expensive than traditional wiring - so the benefits usually need to be realised in areas other than direct cost. This is especially the case with the most extensive of multiplexing techniques - the use of smart connectors where there may be as many as 200 multiplexed sensors, actuators and lamps. We'll come back to this type of system later - but first, what automotive communication protocols are in place or being developed? ClassificationsThe Society of Automotive Engineers (SAE) has a three-stage classification system for automotive serial data communications:
This class describes a multiplex system where data is communicated over a signal bus between nodes where previously in a conventionally wired vehicle this would have been achieved by individual wires. Bit rates are generally less than 10 kbits per second and support event-driven message transmission. Examples of protocols being used include:
(HVAC stands for heating, air-conditioning and ventilation) An example of this type of networking can be seen in the Lexus LS400. From the 1994 model, Lexus replaced its conventional wiring with a multiplexed system using Toyota's BEAN protocol. The lower part of the diagram shows the previous model's conventional wiring, with the upper diagram shows the multiplexed system. Note the reduction in the required number of wires.
This refers to systems where data is transferred between nodes in order that redundant sensors and other system elements are able to be deleted. An example of this is vehicle speed data that might be communicated from the engine electronic control unit to the dashboard control unit, allowing the use of only one speed sensor for both the engine ECU and speedometer. The speed in this class is between 10 kbits per second and approximately 125 kbits per second. Examples of protocols being used include:
This class of communication uses high data rate signals, typically associated with real-time control systems such as ABS and engine control. Its bit-rate is between 125 kbits per second and 1 Mbits per second, with unshielded twisted pair the medium of choice, except above 500 kbits per second where optical fibre is used.
However, it's now suggested that an additional four classes of networks be identified, with the following being added:
Note that this doesn't necessarily suggest that seven different protocols are required; some of these networks need to be separate for safety rather than technical communication reasons.
All cars sold in the US are required by law to have an Onboard Diagnostics (OBD) port to allow the checking of real-time and logged engine management data. This is so that the emissions equipment functioning can be validated on an on-going basis. (Note that such is the all-pervading influence of US legislation, many cars sold around the world also have this data port as standard.) All OBD-equipped cars can communicate the legally required minimum data, but most manufacturers also use this same port to communicate a much more extensive range of brand-specific parameters. The protocol standard for emissions diagnostics is rapidly becoming Keyword 2000.
This category can be further subdivided:
Aftermarket manufacturers of entertainment and computer equipment are watching these mobile media communications standards with great interest - given that many original equipment car entertainment systems are now completely integrated with other car control functions, any consumer electronics upgrades within the life of the vehicle may be dependent on the development of universal communications protocols. Simply, if the aftermarket doesn't keep up with OE communication protocols, their products will be in many cases unable to be fitted to cars.
This usage is primarily for airbag systems, with two or more buses used for firing, sensing, etc. Most systems support up to 64 nodes, which can consist of accelerometers, occupant-present sensors, seatbelt pre-tensioners, explosive squibs, etc. Eight different protocols are in development, with speeds ranging from 5 kbits per second - 500 kbits per second.
Requiring very high-speed real time data transmission and a very low fault tolerance, a number of drive-by-wire communication protocols are currently in development. They will be applied to systems such as steer-by-wire, brake-by-wire and throttle-by-wire. Bit rates will be between 1 Mbits per second and 10 Mbits per second with fibre optics necessary due to the high speed. The leading contender is TTP - time-triggered protocol. Audi is one company that is pursuing work in this direction, with BMW another. As can be seen in the above tables, CAN - Controller Area Network - is currently the most popular communications bus protocol. Invented by Bosch in the mid-1980s, CAN enables robust serial communication. It was the first and has remained the most enduring of automotive control networks. CAN is based purely on event controlling - information is entered into a communication system when an event occurs. The priority of the information is controlled by an identification at the start of the message, with the higher the priority, the greater the probability that this message will prevail on the bus. This gives flexibility - it's easy to enter a new message to an existing network, and access to the bus can take place without loss of information - but in larger networks this approach limits speed to 1 Mbit per second. Further, the higher the loading of the bus, the greater the time delay in the transport of the information. TTP, on the other hand, has fixed time slots of variable width. It transmits the data in order and it is precisely predetermined when a piece of information will be on the bus. Audi gives a specific example of the benefits that will be associated with the implementation of TTP buses. In current Audi models, local controllers are used in systems such as Electronic Stability Control, Traction Control and air suspension. The various control systems are only weakly linked by the CAN bus: due to the lack of precision in the system, controllers must be prepared to miss one or more messages! However, if the systems were in real-time communication - by the use of a TTP bus for example - additional functions could be implemented or current functions optimised. These include:
Furthermore, if pure "by wire" systems were to be installed, even greater changes could be undertaken. Distributed ArchitectureAs indicated earlier, the cost of multiplexed systems is usually higher than that of a simpler wiring system. US automotive parts manufacturer, Delphi, has proposed an approach to overcome this cost barrier. Somewhat curiously, it attempts to do so by actually increasing the number of nodes! This distributed multiplex architecture turns each wiring harness connector into a smart design, with the connectors linked by means of a LIN Class-A bus. The following diagrams show the general idea. Traditionally, the wiring of the loads within a front door is done in this way - discrete wires connecting separate switches to each of the loads. No less than 33 wires need to be used to connect the door loads with the main car body; there are 80 wires in the complete harness. The outside rear vision mirror (OSRV), door lock, window motor and in-door courtesy lamp are all controlled by individual switches and power and earth wires. Many cars now use multiplex modules within the doors, connected to the main body computer by means of a CAN bus. By taking this approach, the number of conductors is reduced to 44. The number of discrete wires has been reduced, but the multiplexed module usually needs to be specific to the door in which it sits, reducing scales of economy. Delphi is now proposing the use of smart, multiplexed connectors. While this approach requires the most electronics and the least wiring (only 25 conductors), it offers the greatest flexibility - all the doors can use the same smart connectors and other models in the manufacturer's line-up can also be so equipped. A load can also be easily added to the vehicle by plugging in another smart connector to the bus and then appropriately modifying the software of the master. Of course, just the same approach could also be used with sensors - smart sensors that interface directly with the bus. In order that the cost increase of turning 'dumb' analog sensors into 'smart' digital sensors is kept as low as possible, it's proposed that multiple sensors be grouped within the one package. For example, five sensors could use the same set of bus communication chips. Such a grouping is possible for example where angular rate sensors, high and low g accelerometers, a barometric pressure sensor and a temperature sensor can all be located in the one package. It is suggested that 15 - 20 of the total of 40 - 50 sensors found in a typical car could be grouped in this way, allowing easy and cheap bus communication. ConclusionWith the proliferation of electronic systems in cars it won't be long before a car without at least two internal communication buses will be rare. Further, with the integration of original equipment entertainment systems into the functionality of the car, altering the radio or adding a navigation system is likely to involve connection to at least one of the buses. Car manufacturers worldwide have really decided that this is one bus that they're not going to miss...
Share this Article:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|