SAE TECHNICAL PAPER SERIES
2001-01-0072
LIN Bus and its Potential for use in Distributed Multiplex Applications John V. V. DeNuto, Stephen Ewbank, Francis Kleja, Christopher A. Lupini and Robert A. Perisho, Jr. Delphi Automotive Systems
Reprinted From: In-Vehicle Networks 2001 (SP–1594)
SAE 2001 World Congress Detroit, Michigan March 5-8, 2001 400 Commonweal Commonwealth th Drive, Drive, Warrendal Warrendale, e, PA 15096-0001 15096-0001 U.S.A. U.S.A.
Tel: (724) 776-4841 776-4841 Fax: (724) 776-5760 776-5760
The appearance of this ISSN code at the bottom of this page indicates SAE’s consent that copies of the paper may be made for personal or internal use of specific clients. This consent is given on the condition, however, that the copier pay a $7.00 per article copy fee through the Copyright Clearance Center, Inc. Operations Center, 222 Rosewood Drive, Danvers, MA 01923 for copying beyond that permitted by Sections 107 or 108 of the U.S. Copyright Law. This consent does not extend to other kinds of copying such as copying for general distribution, for advertising or promotional purposes, for creating new collective works, or for resale. SAE routinely stocks printed papers for a period of three years following date of publication. Direct your orders to SAE Customer Sales and Satisfaction Department. Quantity reprint rates can be obtained from the Customer Sales and Satisfaction Department. To request permission to reprint a technical paper or permission to use copyrighted SAE publications in other works, contact the SAE Publications Group.
All SAE papers, standards, and selected books are abstracted and indexed in the Global Mobility Database
No part of this publication may be reproduced in any form, in an electronic retrieval system or otherwise, without the prior written permission of the publisher. ISSN 0148-7191 Copyright 2001 Society of Automotive Engineers, Inc.
Positions and opinions advanced in this paper are those of the author(s) and not necessarily those of SAE. The author is solely responsible for the content of the paper. A process is available by which discussions will be printed with the paper if it is published in SAE Transactions. For permission to publish this paper in full or in part, contact the SAE Publications Group. Persons wishing to submit papers to be considered for presentation or publication through SAE should send the manuscript or a 300 word abstract of a proposed manuscript to: Secretary, Engineering Meetings Board, SAE.
Printed in USA
2001-01-0072
LIN Bus and its Potential for Use in Distributed Multiplex Applications John V. DeNuto, Stephen Ewbank, Francis Kleja, Christopher A. Lupini and Robert A. Perisho Jr Delphi Automotive Systems
Copyright © 2001 Society of Automotive Engineers, Inc.
ABSTRACT The increasing features and complexity of today's automotive architectures are becoming increasingly difficult to manage. Each new innovation typically requires additional mechanical actuators and associated electrical controllers. The sheer number of black boxes and wiring are being limited not by features or cost but by the inability to physically assemble them into a vehicle. A new architecture is required which will support the ability to add new features but also enable the Vehicle Assembly Plants to easily assemble and test each subsystem. One such architecture is a distributed multiplex arrangement that reduces the number of wires while enabling flexibility and expandability. Previous versions have had to deal with issues such as noise immunity at high switching currents. The LIN Bus with its low cost and rail-to-rail capability may be the key enabling technology to make the multiplexed architecture a reality.
INTRODUCTION HISTORY OF MULTIPLEXING –The multiplexing of automotive electrical data onto communication buses dates back to the late 1970s (1). This technology is only beginning to reach its stride. It was originally hoped that a single bus protocol could handle the needs of any vehicle. Gradually that expanded to the SAE categorization of Class A, B, and C and the realization that up to three protocols and/or networks may be necessary. Today, it is realized that at least seven in-vehicle protocols may be necessary (2). These include, besides the existing SAE classes - diagnostics, airbag, mobile media, and x-by-wire. Each area needs its own protocol and one or more networks running that protocol. Sometimes this is for safety reasons, such as with airbags or x-by-wire. Another reason for such a “niche”
protocol is specialty use. An example of a niche protocol is for the application of “smart sensor and actuator” technology, sometimes also known as “smart connector”. This relatively new area of automotive electronics involves multiplexing low-level functions such as relays, high/low side drivers, etc. onto a data bus. For now smart connector functions are lumped into the SAE class A category. Perhaps in the future it will become its own class. Indeed, there are advanced vehicle electronic architectures being considered that utilize up to 200 multiplexed sensors, actuators, and lamps. Each one of these devices becomes a node on one or more smart connector buses. The challenge is to make it cost effective. Multiplexing has always had an issue with cost. Replacing some discrete wires and connectors with a data bus interface tends to add cost. There are benefits associated with increased diagnostic capabilities and system expandability, but the bottom line is that the ICs necessary for the multiplex circuit add additional cost – making smart connector multiplexing appear unattractive. It is well known that electronic systems decrease their cost more quickly than traditional wiring. This makes the eventual cost effectiveness of multiplex inevitable, a question of when, not if. Up to now the buses used in the first, basic implementations of smart connectors, have been custom or proprietary (e.g. CCD, BEAN, etc.), too complicated (e.g. TTP/A) or too slow and cumbersome (e.g. ISO 9141) (3). The latest entry into the low-cost smart connector arena is Local Interconnect Network (LIN). With LIN lies the hope of a worldwide smart connector data bus protocol standard (4).
Specific advantages of LIN bus when applied to a distributed multiplex architecture include: • • • • • •
Standardization Low cost silicon 12V single wire interface Self Synchronization without a crystal Guaranteed latency times Speeds up to 20Kbit/s
imperative to obtain the minimum thermal resistance with no increase in part cost. Advanced computer simulations are being employed to obtain the optimum packaging solution. The third challenge is to demonstrate the performance of such technology with technical demonstrations and test data. With shortened product development cycles, OEM’s are requiring this level of development prior to including a new technology into one of their programs. We are meeting these needs with functional demonstrations of distributed multiplex sub-systems in seats and doors and comprehensive testing.
OBJECTIVES A distributed multiplex architecture is currently being developed by Delphi Automotive Systems. This new architecture increases the electronic content while decreasing the conventional wiring content. The multiplex nodes are referred to as Smart Connectors. The goals of this approach are to offer the customer greater flexibility and better packaging at a reduced subsystem life-cycle cost.
Finally, we must overcome our natural resistance to change. The old adage “if it isn’t broke, don’t fix it” is often applied to technical innovations such as distributed multiplex. It takes a technology leader, with a real vision of the future, to see the many benefits that a distributed multiplex system has to offer but are difficult to quantify.
BENEFITS
APPROACH
The potential benefits of architecture are listed below: • •
•
•
•
a
distributed
multiplex
Simplified vehicle assembly Common application on multiple vehicles resulting in higher volumes and lower costs Significantly increased flexibility – the ability to add features to a vehicle in a plug & play manner Improved quality and reliability due to a significant reduction in the number of wires Simplified wiring assembly and reduced mass
Figures 1 through 3 display three comparative architectures applied to a driver’s door of a typical automotive application. In Figure 1, no multiplexing is used. Each device interfaces directly with a centralized body computer and / or bussed electrical center. This architecture possesses the least electronic content, the most wires, and the least flexibility to change.
OSRV Mirror 14
Driver’s Switch Assembly
Door Lock Assembly
9
14
6
Many of these benefits are compounded when the same Smart Connectors are applied to multiple platforms.
2 2
CHALLENGES The primary challenge to implementing a distributed multiplex architecture remains cost. The cost increase in electronics must be offset with a decrease in the cost of the wiring system and other structural cost improvements. A cost competitive solution for a distributed multiplex door sub-system has been proposed by Delphi. In order to keep the electronic cost increase as small as possible, increased volumes, penetrations, and packaging improvements must be realized. Only then can cost parity be achieved. The second challenge management. Specific being applied to Smart minimize package size
lies in the area of thermal new packaging techniques are Connector designs in order to and thermal resistance. It is
Window Motor
Courtesy Lamp
33
Vehicle Body 33
Electronic Module Conventional Connector Wiring
Body Computer / Electrical Center
Figure 1: Driver’s door with conventional (non-multiplex) architecture
OSRV Mirror
Driver’s Switch Assembly
14
Door Lock Assembly
9 6 2
Door Multiplex Module
2
1. 2. 3. 4.
Window Motor
Courtesy Lamp
Vehicle Body Body Computer / Electrical Center
central door module are distributed to four discrete Smart Connectors using the LIN protocol. This approach uses the most electronics and the least wiring of the three approaches presented. The four Smart Connectors defined in this system control the:
Electronic Module Conventional Connector Power Ground CAN data Other wiring
Figure 2: Driver’s Door with centralized multiplex architecture
Window Motor OutSide RearView (OSRV) Mirror Assembly Driver’s Door Switch Door Lock Assembly
This approach offers the greatest flexibility to change and the greatest potential for inter-platform application. Four doors in a given vehicle can be executed using the four Smart Connectors detailed above. This four-door approach is shown in Figure 4. Master
Figure 2 depicts the same door loads controlled using a centralized door multiplex module. This type of architecture normally utilizes the CAN data bus, allowing the door module to exist as a peer to the vehicle ’s body computer. Using this approach reduces the number of discrete wires but still limits flexibility. A change in any load can require a change to the entire door module. This technique usually results in different modules being applied to the driver’s, passenger’s, and rear doors as well as different modules on nearly every platform within an OEM.
Mirror
Mirror
Dr Switch Lock Switch
Switch
Lamp
Switch Lamp
Switches Lamp
Lamp
Window Motor
Window Motor
Window Motor
Window Motor
Door Lock
Door Lock
Door Lock
Door Lock
Driver’s Door
Left Rear Door
Right Rear Door
Passenger’s Door
Figure 4: Four-door vehicle controlled using 4 types of Smart Connectors
OSRV Mirror 14
Driver’s Switch Assembly
Door Lock Assembly
Window Motor Courtesy Lamp Vehicle Body Body Computer / Electrical Center
Electronic Module Smart Connector IDC Interface Conventional Connector Power Ground LIN data
Figure 3: Driver’s Door with distributed multiplex architecture using “Smart Connectors”
Figure 3 presents the same loads controlled using a distributed multiplex architecture employing Smart Connectors. In this approach, the functions of the
Content changes in a vehicle can be executed more easily using this approach. A load can be added to a vehicle by plugging a Smart Connector to the bus and modifying the software in the master. This essentially makes the sub-system plug and play. Since the Smart connectors share a common three-wire bus structure, adding a Smart Connector need not change the wiring content. The distributed architecture also offers the possibility of automated wiring harness assembly through the use of Insulation Displacement Crimp (IDC) terminals A comparison of the total wiring content can be seen in Figure 5. As seen in the figure, the distributed multiplex architecture uses the least number of discrete wires and enables automated assembly of the three-wire data bus. Of the remaining wires, 14 out of the 21 (66%) are associated with the mirror pigtail.
90
Cut Leads (Body) Cut Leads (Automated assembly)
80 70
Cut Leads (Door) Mirror Pig Tail
60
Hi Current Ground Hi Current Power Switch InputCkt 1
50 40
Passenger Window Switch
30
Switch InputCkt 2 Switch Ground
20 10 Backlighting
0
Power
Conventional
Centralized Multiplex
Distributed Multiplex Lock Switch
LIN bus Switch InputCkt 3 Switch InputCkt 4 Ground
Figure 5: Comparison of wiring in conventional, multiplexed, and distributed Smart Connector architectures
Courtesy Lamp Output Courtesy Lamp Courtesy Lamp Ground Motor Ckt 1 Window Motor Motor Ckt 2
The four Smart Connectors described above are all currently under development using Delphi’s Advanced Development Process (ADP). The project recently completed a comprehensive requirements review. A more detailed description of each Smart Connector follows:
Figure 6: Window Motor Smart Connector I/O block diagram
The first Smart Connector described is the window motor Smart Connector. An input / output block diagram of this Smart Connector can be seen in Figure 6. This Smart Connector is used in each door of a vehicle. It controls the window motor and several other miscellaneous door functions. These miscellaneous functions include the courtesy lamp and window and lock switch inputs. This functionality is included in the window motor Smart Connector because dedicated Smart Connectors controlling these functions were estimated to increase the sub-system’s total cost.
The second Smart Connector defined is the OSRV Mirror Smart Connector. An input / output block diagram of this Smart Connector can be seen in Figure 7. This Smart Connector controls the mirror tilt motors, mirror fold motor, mirror heat, an exterior courtesy lamp, and reads sensors for mirror position and outside temperature. This Smart Connector would be used in each front vehicle door.
The fourth Smart Connector defined is the door lock Smart Connector. An input / output block diagram of this Smart Connector can be seen in Figure 9. This Smart Connector drives the door lock and super-lock motors. It can also monitor the position of two digital input switches. It will be used in each vehicle door.
OSRV Mirror Assembly Lamp Feed Puddle Lamp Ground Ground Heater Heater Feed Fold Motor Ckt1 Fold Motor Fold Motor Ckt2
Power
Vertical Motor LIN Bus
Vertical Motor
Door Lock Assembly Switch Input Ckt 1
Ground Motor Common Horizontal Motor
Switch Input Ckt 2 Horizontal Motor
Power
Vertical Position Input Horizontal Position Input
Switch Ground Ckt
5V Feed
LIN Bus
Door Lock Motor Ckt Ground
5V Return
Door Lock Motor
Thermistor Temperature Input
Motor Common Ckt. Motor 2 Motor 2 Ckt.
Figure 7: OSRV Mirror Smart Connector I/O block diagram
Figure 9: Door Lock Smart Connector I/O Block Diagram
The third Smart Connector defined is the Driver ’s Switch Smart Connector. An input / output block diagram of this Smart Connector can be seen in Figure 8. This Smart Connector is designed to read window switch inputs for four windows, read the position of the mirror select / fold switch, read the position of the mirror joystick, read the position of the rear window lockout switch, and drive dimmable backlighting.
Driver Door Switch Switch Ground Ckt1 Backlighting Backlighting Driver’s Window
Passenger’s Window
Mirror Joystick
Switch Input 1
Switch Input 2
Power LIN Bus
Switch Input 5 Ground
Mirror Select / Fold LR Window
RR Window
Switch Input 6 Switch Input 3
Switch Input 4
Window Lockout Switch Input 7
Figure 8: Driver’s Switch Smart Connector I/O Block Diagram
PLANS AND TESTING Extensive test plans have been developed to demonstrate the feasibility of the LIN bus for production applications. Since the bus protocol has just recently been defined (and is still subject to revision), there is little field experience to indicate how the bus will perform under real world conditions. Not only must the protocol be shown to be workable, but manufacturer's claims regarding integrated circuit functionality and development systems capabilities must be verified. Areas of focus include EMC, data bandwidth, latency, and bus operation under high current load switching. In addition to bench testing of individual components, an actual 11-node production intent system will be built and tested. All of this is being done under Delphi's ADP process, and will result in a system capable of being deployed on OEM vehicles throughout the world.
EMC EMC will be an important issue for LIN, and requirements have been gathered from several different vehicle manufacturers. In keeping with its goal of being a low cost solution, the LIN bus uses a single unshielded wire as its transmission medium. Present day single wire systems run at speeds to about 10K bits/second, and use wave shaping on the signal to reduce radiation. But how will the LIN bus do? Bench testing has been performed to acquire LIN bus waveforms for use in mathematical modeling, and to measure radiated emission under worst case conditions -- highest bus traffic at the highest speeds (5). The LIN bus protocol specifies permissible data rates of up to 20K bits/second, faster than today's single wire systems. The 10-node test system will run as close as possible to this speed to provide another check on EMC performance. If 20K bits/second turns out to be unrealistic, it will be necessary to determine what the actual upper limit is. Speeds much lower will result in added cost, as more buses will be required to handle communications.
sub-system. Worst-case analysis, as well as in-vehicle testing, will be used to demonstrate acceptable latency performance. HIGH CURRENT LOADS AND BUS DISRUPTIONS Earlier prototype systems suffered from communications disruptions when high current loads were switched off and on. Load currents (figure 10) in the three-wire smart connector configuration can result in a temporary (large) voltage shift in the ground wire.
BUS LATENCY AND BANDWIDTH In evaluating LIN performance, it is not enough to simply verify that the bandwidth is adequate for all the messages planned for the system. When a user generates a request through the system (by pushing a button, for example), he or she expects an immediate response to the command. A latency requirement has been established for Delphi's system to address this issue. Meeting this requirement has complicated the message strategy for the 11-node test system, since a single LIN bus is used for all communications. It would take over 100 ms for all messages to be transmitted if they were all processed sequentially. A single master node initiates all communications in the LIN bus system. Slave nodes cannot initiate communications themselves, and must wait to be polled by the master before their data can be output. Typically, a schedule table is set up within the master, and this determines when each message will be output on the bus. A benefit of this method is that latency and bus utilization are very predictable. The problem is that using a static schedule table will result in unacceptable delays for the system. Using conditional branches in the schedule table can solve this (6). Test and evaluation of these techniques is commencing using a bus development tool known as LINspector. This will be used to simulate a master node, and will be programmed to change the scheduled table based on data received from the slave nodes. Conditional branches result in some loss of predictability, but should reduce system delays to levels acceptable to vehicle users. Full development of the LIN protocol must also address variability of master nodes and slave nodes designed and built by different suppliers combined into a
Figure 10: Ground shift at a window motor during inrush. LIN data communications are also plotted. Note the ground shift in the slave node’s response.
This is due to long wire lengths and the sharing of a single ground wire by multiple smart connectors. These voltage drops can corrupt the data transmission in the commonly used single wire buses. The LIN physical layer uses a rail-to-rail (ground to battery) swing, which promises increased immunity to this problem. LIN requires a transmitting node to transmit within 20% of the local power and ground voltages. Receiving nodes are designed to interpret messages that are within 40% of its local battery and ground. This provides a significant margin for ground shift when a remote slave node responds to a poll by the master during a high current event. This is shown in Figure 10. As long as the node’s transmission, which is required to be within 20% of its local ground reference, is received within the master ’s 40% tolerance, the message will be valid. The 11-node test system will be used to verify the proper operation of the bus under realistic conditions. The system will be wired in a production intent configuration, using production intent loads. A tool such as LINspector will be used monitor the bus while various loads are energized. Electrical measurements will be taken to
determine noise margins. Some of the smart connectors could have additional power and ground inputs, which may be used to supply current to the high power loads. This may be necessary in some vehicle applications.
SUMMARY: WHAT HAVE WE LEARNED? The development of Smart Connectors requires that we: • • • •
Define requirements to meet customer needs Benchmark current systems Assess the advantages and disadvantages Develop new concepts to capitalize on the strengths and resolve weaknesses
Historically multiplexed systems have suffered from various weaknesses. As these problems are resolved, the systems become viable. Some common drawbacks include:
• • • • •
Custom non-standard protocols Complicated protocols Slow speed Lack of Development Tools Cost
To comprehend the potential commercial viability of such a system, a thorough understanding of the underlying power and signal distribution system must be present. The majority of cost savings that can be realized in a distributed multiplex system are a result of a simplification of the wiring sub-system. Further cost savings and reliability improvements can be achieved through assembly automation. Process equipment that will automatically assemble Smart Connectors with IDC terminals onto a 3-wire bus is under development.
We are proceeding with packaging development of the Smart Connectors specified earlier in this paper. Figure 12 shows a concept for the OSRV Mirror Smart Connector. It mates to the three-wire power, ground, and data bus using IDC terminals. This allows the Smart Connector to be inserted onto a continuous piece of wire and eliminates the need to strip insulation. Thermal evaluation of this Smart Connector is proceeding concurrently with the electronic and packaging development. Figure 13 shows a temperature contour plot of the OSRV Mirror Smart Connector with mirror fold and heat activated. The plot also shows the heat generated by the reverse battery diode, a requirement when using a solid state H-bridge. Concurrent electronic, packaging and thermal analysis offer the greatest opportunity for developing a commercially viable distributed LIN system with rapid speed to market.
The LIN Bus protocol overcomes these drawbacks by promoting a worldwide standard that is simple, fast and low cost. At present, the hardware and development tools are available but still working up the maturity curve. At Delphi, we have completed a comprehensive review of global customer requirements and allocated those requirements into specific product attributes and performance specifications. We have an initial 4-node driver’s door system functioning and undergoing test. We first showed this functional system to the public during Convergence 2000. The system is running at 19.2Kbit/s and meeting our latency requirements. Initial radiated emission testing run at 19.2Kbit/s has shown promising results. Our development team is extending from this 4-node driver’s door system to a full 11-node vehicle architecture.
CONCLUSION The viability of a distributed multiplex LIN bus appears very promising at this time. The development of such a system requires an organization with expertise in electronics, power and signal distribution, and connection systems. Without any one of these three competencies, successful deployment of such a system would be nearly impossible.
Figure 12: Smart Connector concept employing an IDC terminal interface to a three-wire bus.
CONTACT Controlling ASIC
Mirror Heat FET
Reverse Battery Diode
Mirror Fold FET’s
Figure 13: Computer simulation of the heat transfer within a mirror Smart Connector with mirror fold and heat activated.
We are proceeding with the development of a distributed multiplex door subsystem using the LIN communications protocol. A detailed architecture study jointly conducted with a major OEM has shown the cost viability of such a system. Delphi is currently on track to develop this system for production application.
REFERENCES 1. 2. 3.
4. 5. 6.
Bell, “Multiplexing – Past, Present, and Future,” SAE 760178, 1976. Lupini, “Vehicle Multiplex Bus Progression,” SAE 2001-01-0060, 2001 Koptez, Elmenreich, and Mack, “A Comparison of LIN to TTP/A,” Research Report 4/2000 Institut fur Technische Informatik, TU W ien, Austria. LIN Consortium, “LIN Protocol Specification 1.1, http://w.w.w.lin-subbus.org , April 2000. Motorola, MC33399 LIN Interface, Product Proposal Rev 4.0. Specks and Rajnák, “LIN – Protocol, Development Tools, and Software Interfaces for Local Interconnect th Networks in Vehicles,” 9 International Conference on Electronic Systems for Vehicles, Baden-Baden, Oct 2000.
John V. DeNuto is an Engineering Supervisor with Delphi Automotive Systems. He holds an MS degree in Mechanical Engineering from the University of Akron and a BS in Naval Architecture from the United States Naval Academy. He can be reached at (330) 306-1143 or at
[email protected]. Steve Ewbank is a Senior Project Engineer with Delphi Delco Electronics Systems. He holds an MS degree in Electrical Engineering from the University of Kansas. He can be reached at (765) 451-0353 or at
[email protected]. Fran Kleja is a Project Engineer with Delphi Packard Electric Systems. He holds a BSAS degree in Electrical Engineering Technology from Youngstown State University. He can be reached at (330) 306-1124 or at
[email protected]. Chris Lupini is a Senior Project Engineer with Delphi Delco Electronics Systems. He has an MSEE degree from Purdue University and a BSCompE degree from the University of Michigan and is a licensed Professional Engineer. He can be reached at (765) 451-0248 or at
[email protected]. Rob Perisho is a Senior Project Engineer with Delphi Delco Electronic Systems. He holds a BS in Electrical Engineering from Brigham Young University and did his graduate work at San Diego State University. He can be reached at (765) 451-0817 or at
[email protected].