Special issue MOST. One technology many applications. Fachmedium für Entwicklungen in der Kfz-Elektronik und Telematik

Transkript

1 Elektronik automotive March 2010 Special issue MOST New car models using MOST technology Interview Jens Kötz: MOST at Audi Integration of mobile device using MOST S2 / March ,00 Fachmedium für Entwicklungen in der Kfz-Elektronik und Telematik Special issue MOST One technology many applications

2 IIII Editorial MOST technology in use: Not just for multimedia Very different bus systems are used to develop electrical/electronic architectures for today s vehicles. Depending on their purpose you find CAN, LIN, FlexRay and, of course, mediaoriented systems transport or MOST. In more than one sense MOST is something special. When it was first developed at the end of the 90s, the automotive landscape was still very I The members of the MOST Cooperation divide into associated partners, suppliers or system integrators and core partners. The latter Audi, BMW, Daimler, Harman/Becker and SMSC form the steering committee. much impacted by proprietary developments. But industry already started to think driven by increasing cost pressures about the feasibility of crossbrand standardization where products did not differ through their individual features. At the same time there was this new, adventurous terrain of multimedia. And nobody could envisage what amounts of data there would be to transmit today, what functionalities to implement and what devices to interconnect. Of course the MOST Cooperation founded in 1998 did not get off to the kind of start where everything ran as smoothly as could be wished. Technical problems in the first series applications, for example mastering optical data transmission, and political squabbles when it came to supplying components were food for the critics, who within a few years were already predicting the end of MOST technology. Nevertheless or perhaps precisely because of these initial obstacles MOST was able to establish itself worldwide as an automotive networking standard. In the meantime it has also found much favor on the Asian market. In MOST150 the standard has finally achieved maturity. Its transmission rate offers sufficient reserves for multiple HD data streams, and investigations are already looking into the use of MOST technology in safety-critical applications. So it is quite possible that we will see MOST used to network the control units of different driver assistant systems. Until then I wish you informative reading of this special issue of Elektronik automotive for the MOST Forum Stephan Janouch Chefredakteur Elektronik automotive Schreiben Sie mir Ihre Meinung: sjanouch@elektronik.de Elektronik automotive special edition MOST 3

3 Content Plug in ready This paper proposes a smooth solution for the integration of consumer devices into the vehicle across a layer-3 network. It describes the signaling to establish an audio streaming between a consumer device, e.g. an iphone, and the vehicle s infotainment system and finally presents a use case. >> 41 Definitely safe This paper presents an analysis which discusses the possibility to implement safety related applications using the MOST150 protocol respecting current and forthcoming safety standards, especially the current draft of the ISO standard for safety in the automotive field. A safety layer provides safe communication over a MOST network using safety codes. >> 32 Influence on the Bandwidth This article shows that the bandwidth of a polymer optical fiber (POF) cannot be regarded as an autonomous parameter to quantify its high frequency performance. Thus the question will be answered, what are the additionally required details to be linked to a bandwidth value. >> 14 Editorial 3 MOST technology in use: Not just for multimedia Preface 6 MOST history: From Germany into the whole wide world MOST news 8 MMI 3G: Specially for the Asian markets 8 Application area: MOST150 is ready for more 8 SAE Convergence: Continues triumphant advance 10 MOST Interconnectivity Asia: Current MOST150 developments will be presented 10 New car models with integrated MOST technology Interview 12 Let there be light A premium producer implements optical transmission technology Physical layer 14 Influence on the bandwidth Dependencies of bandwidth of polymer optical fiber for MOST systems 17 Use in development and production Stress test tool for MOST150 physical layer 20 Green light for MOST GaN high-speed LEDs and their application in a POF-based communication module 22 Automated determination of worst-case scenarios For MOST optical physical layer specification point 3 Modelling + simulation 25 System behaviour in initialisation phase MOST control channel modelling 29 Timing master faults Analysis and cure of timing master faults 32 Definitely safe MOST150 for safety related applications 35 Verification and performance analysis Different MHP configurations using virtual prototypes 4 Elektronik automotive special edition MOST

4 Beyond the horizon The age of distributed networked infotainment systems began with the 1998 Mercedes-Benz S-Class and the introduction of the D2B Optical infotainment network. Since then, the evolution of infotainment networking continues. Today, MOST is the de-facto standard for automotive infotainment networking. The MOST Cooperation comprises of 77 member companies. >> 49 System architecture 38 Window to the CE world Using UPnP in MOST-based infotainment systems 41 Plug in ready UPnP as a proposal for a MOST/IP gateway 44 Universally applicable Using MOST as general communication protocol 46 Audio framework High-performance audio streaming framework for MOST 49 Beyond the horizon MOST as networking infrastructure for different automotive domains 52 True value Evaluation of MOST150 technology MOST products 57 News 59 Imprint 59 Index of advertisers

5 Preface IIII MOST history: From Germany into the whole wide world The MOST Cooperation the organization through which the leading automotive multimedia network Media Oriented Systems Transport (MOST) is standardized looks back on twelve years of successfully establishing MOST as the leading standard for automotive infotainment networking. The third generation of the infotainment backbone with faster data rates of 150 Mbit/s and various new features has been standardized, while robustness, quality, and efficiency have been further optimized. The adoption of MOST technology worldwide is going rapidly, with focus on the Asian market. So far, a grand total of 19 vehicle models manufactured by Asian automakers have MOST built in. In total, MOST is integrated in over 90 car models that are available worldwide. Recently, the multimedia network has also made inroads into the first mass-produced models of the mid- and low-range sector. Please let me start by summarizing the story of the automotive multimedia network. Over twelve years ago, the carmakers BMW and Daimler, together with Becker Automotive Systems (today Harman International) and Oasis Silicon Systems (today SMSC), started cooperating on defining and designing the Media Oriented Systems Transport technology. With their clear vision, the companies saw the need for a common infotainment network standard instead of proprietary solutions. The activities around MOST aim at concrete and usable results for volume production rather than only defining a theoretical standard. In 1998, the companies founded the MOST Cooperation, with Audi joining shortly thereafter. They quickly developed the first specification, since they had to meet deadlines for mass production. In 2001, only three years after, BMW introduced the 7 series as the first MOST car. The following year, 13 more models implemented the MOST infotainment backbone. Now, twelve years later, 16 carmakers and 61 premier suppliers contribute to the success of the MOST technology. Today, MOST technology is prepared for a green future and ready for implementation in hybrid and electric cars. The plastic optical fiber chosen ten years ago significantly reduces the weight of the wiring harness. Particularly for electric mobility, every saving is crucial because all the electrical energy has to be stored in rechargeable batteries before commencing travel. Another issue is electric smog. Because of its physical properties, an optical transmission does not emit electromagnetic fields. The vehicle models that have implemented MOST to date now include the first hybrid vehicles. With respect to environmental sustainability and electric mobility, MOST is state of the art technology and constitutes the modern platform for energy saving infotainment systems. The first and second generations of MOST are already on the road all around the world, and the third generation is heading towards series production. The first generation, MOST25, operates at 25 Mbit/s. A detailed Data Link Layer Specification for MOST25 has already been made available by SMSC and Harman/Becker, and can be requested by members of the MOST Cooperation. With the MOST Specification Rev. 2.5 for MOST50 the second generation the MOST Cooperation doubled the bandwidth for automotive infotainment solutions from 25 to 50 Mbit/s. The MOST Specification of Electrical Physical Layer Rev. 1.1 is an additional key specification that enables data transmission over an unshielded twisted pair (UTP) of copper wires while meeting the stringent automotive electromagnetic compatibility requirements. Several car manufacturers in Asia are now using MOST50 UTP connections and doubledensity coding to increase bandwidth to 50 Mbit/s. MOST150 is the third generation, offering a bandwidth of 150 Mbit/s. With transmission over the legacy POF/LED optical physical layer, MOST150 offers a smooth migration from MOST25 and MOST50, allowing carmakers to continue to use POFs and LEDs as light sources. But also electrical physical layers are possible. In addition to higher bandwidth, MOST150 features an isochronous transport mechanism to support extensive video applications, as well as an ethernet channel for efficient transport of IP-based packet data. This channel carries legacy ethernet packets (according to IEEE 802.3) so standard TCP/IP stacks can be transmitted without change. Thus, the latest generation of MOST provides the automotive-ready physical layer for ethernet in the car. Consequently, MOST will be open to a broad variety of IP based applications. It even allows adjustment of the bandwidth of conventional streaming connections on one hand and IP communication on the other, according to the corresponding requirements. In addition, MOST150 still supports the well known asynchronous channel to ensure backward compatibility of MOST25 applications. With MOST150, audio and video signals can be transported with high bandwidth efficiency and without any overhead for addressing, collision 6 Elektronik automotive special edition MOST

6 MOST news IIII detection/recovery or broadcast. Multiple high-definition (HD) video streams and multiple streams of multi-channel surround sound with premium quality of service can be transmitted, while simultaneously moving high loads of packet data around. The latest public demonstrations include high-definition video over MOST, ethernet tunneling over MOST, MOST consumer connectivity solutions and MOST compliance solutions. MOST150 series adoption is in process, proving that this latest MOST technology is ready for implementation. Various carmakers have already started with series projects. With the first carmakers committing to the MOST150 network in selected vehicles from 2011 onwards, the suppliers of MOST devices, tools and software solutions are focusing on optimizing their MOST150 product portfolio to ease implementation in future car models. Key targets are cost, efficiency and low power consumption while meeting MOST compliance requirements. MOST solutions and tools are already available to enable successful infotainment implementations at the OEM and supplier level. The built-in signal routing and enhanced power management features are important building blocks for successful MOST system design. MOST150 meets the requirements in all major application fields. Among them are the traditional fields entertainment and information but also new areas like mobile connectivity, connected services, and driver assistance. MOST150 offers the technology and the audio/video capability for next generation automotive infotainment functions and devices such as head units, rear seat entertainment, amplifiers, TV tuners and video displays. But it is also ready for new features like internet access devices, ipods, cameras or video processing units. Dr. Christian Thiel MOST Cooperation Administrator Aisin AW is the tier 1 supplier for the Audi MOST infotainment system MMI 3G in the Asian markets. Integration of high level Asian navigation as well as the adaptation and localization of the European platform for China, Japan and Korea posed a high level of sophistication to the role of the tier-1 system integrator. Aisin AW engineers made substantial modifications to the MMI system Application area: MOST150 is ready for more Expanding the areas of automotive use cases the MOST Cooperation will focus on the broadening variety of applications supported by the MOST150 network. Themed MOST150 Does It, five major fields of in-car use cases are defined: MOST150 meets the requirements not only for the traditional areas of entertainment and information, but MMI 3G: Specially for the Asian markets and developed specific procedures and tools for simulation and testing of the MOST network. As a first in the Chinese market Aisin AW implemented the RDS-TMC traffic information service. The reward of this challenging project is a fully localized HMI (Human Machine Interface) and successful introduction of MMI 3G systems to Audi carlines A4, A6, Q5 and Q7. sj also for the new domains of mobile connectivity, connected services, and driver assistance. MOST150 offers the technology and the audio/video capability for next generation automotive infotainment features and devices such as head units, rear-seat entertainment, amplifiers, TVtuners and video displays. In addition, it is ready for new fea- SAE Convergence: Continues triumphant advance The MOST Cooperation will be exhibiting at the SAE Convergence 2010 show in Detroit, Michigan, United States. MOST50 and MOST150 demonstrations will be shown at booth 627 on October 19th to 20th, 2010, in the Cobo Center. MOST is rapidly expanding around the world. In addition to the success of MOST in the European and Asian car markets, American carmakers have tures like internet access devices, ipods, cameras, video processing units, and many more. At present, this latest MOST150 technology is ready for implementation, with various car makers already working on first series projects. MOST technology will be exceeding the limits of infotainment. In current and future vehicles, driver assistance functions are starting to complete and extend the feature set of traditional infotainment systems. Along with information features such as navigation systems, traffic information, and function warnings, the number of vehicles with driver assistance features like camera systems, distance controls, or lane departure warnings will be rapidly increasing. With MOST150, automotive network technology is ready to take this essential step beyond infotainment: Necessary improvements are being investigated, and recent studies prove that MOST is able to support safety-critical applications and is available as the future network for driver assistance systems. sj started to seriously consider the MOST technology. The presentation of the latest MOST technology at SAE Convergence will therefore be an ideal exposure to the American market. The presence at this key automotive electronics event will give US carmakers and suppliers the opportunity to stay up-to-date with latest MOST highlights and achievements. sj 8 Elektronik automotive special edition MOST

7 MOST news IIII detection/recovery or broadcast. Multiple high-definition (HD) video streams and multiple streams of multi-channel surround sound with premium quality of service can be transmitted, while simultaneously moving high loads of packet data around. The latest public demonstrations include high-definition video over MOST, ethernet tunneling over MOST, MOST consumer connectivity solutions and MOST compliance solutions. MOST150 series adoption is in process, proving that this latest MOST technology is ready for implementation. Various carmakers have already started with series projects. With the first carmakers committing to the MOST150 network in selected vehicles from 2011 onwards, the suppliers of MOST devices, tools and software solutions are focusing on optimizing their MOST150 product portfolio to ease implementation in future car models. Key targets are cost, efficiency and low power consumption while meeting MOST compliance requirements. MOST solutions and tools are already available to enable successful infotainment implementations at the OEM and supplier level. The built-in signal routing and enhanced power management features are important building blocks for successful MOST system design. MOST150 meets the requirements in all major application fields. Among them are the traditional fields entertainment and information but also new areas like mobile connectivity, connected services, and driver assistance. MOST150 offers the technology and the audio/video capability for next generation automotive infotainment functions and devices such as head units, rear seat entertainment, amplifiers, TV tuners and video displays. But it is also ready for new features like internet access devices, ipods, cameras or video processing units. Dr. Christian Thiel MOST Cooperation Administrator Aisin AW is the tier 1 supplier for the Audi MOST infotainment system MMI 3G in the Asian markets. Integration of high level Asian navigation as well as the adaptation and localization of the European platform for China, Japan and Korea posed a high level of sophistication to the role of the tier-1 system integrator. Aisin AW engineers made substantial modifications to the MMI system Application area: MOST150 is ready for more Expanding the areas of automotive use cases the MOST Cooperation will focus on the broadening variety of applications supported by the MOST150 network. Themed MOST150 Does It, five major fields of in-car use cases are defined: MOST150 meets the requirements not only for the traditional areas of entertainment and information, but MMI 3G: Specially for the Asian markets and developed specific procedures and tools for simulation and testing of the MOST network. As a first in the Chinese market Aisin AW implemented the RDS-TMC traffic information service. The reward of this challenging project is a fully localized HMI (Human Machine Interface) and successful introduction of MMI 3G systems to Audi carlines A4, A6, Q5 and Q7. sj also for the new domains of mobile connectivity, connected services, and driver assistance. MOST150 offers the technology and the audio/video capability for next generation automotive infotainment features and devices such as head units, rear-seat entertainment, amplifiers, TVtuners and video displays. In addition, it is ready for new fea- SAE Convergence: Continues triumphant advance The MOST Cooperation will be exhibiting at the SAE Convergence 2010 show in Detroit, Michigan, United States. MOST50 and MOST150 demonstrations will be shown at booth 627 on October 19th to 20th, 2010, in the Cobo Center. MOST is rapidly expanding around the world. In addition to the success of MOST in the European and Asian car markets, American carmakers have tures like internet access devices, ipods, cameras, video processing units, and many more. At present, this latest MOST150 technology is ready for implementation, with various car makers already working on first series projects. MOST technology will be exceeding the limits of infotainment. In current and future vehicles, driver assistance functions are starting to complete and extend the feature set of traditional infotainment systems. Along with information features such as navigation systems, traffic information, and function warnings, the number of vehicles with driver assistance features like camera systems, distance controls, or lane departure warnings will be rapidly increasing. With MOST150, automotive network technology is ready to take this essential step beyond infotainment: Necessary improvements are being investigated, and recent studies prove that MOST is able to support safety-critical applications and is available as the future network for driver assistance systems. sj started to seriously consider the MOST technology. The presentation of the latest MOST technology at SAE Convergence will therefore be an ideal exposure to the American market. The presence at this key automotive electronics event will give US carmakers and suppliers the opportunity to stay up-to-date with latest MOST highlights and achievements. sj 8 Elektronik automotive special edition MOST

8 MOST news IIII MOST Interconnectivity Asia: Current MOST150 developments will be presented On November 18th, 2010, the MOST Cooperation will host the 11th MOST Interconnectivity Conference Asia in Seoul, South Korea. Representatives of the MOST Cooperation will present an update on recent achievements of the organization and the MOST standardization roadmap to numerous attendees of Asian carmakers and their suppliers. In its eleventh year of presenting MOST to the Asian automotive industry, nineteen of the more than 90 vehicle models that rely on the MOST infotainment technology models are manufactured by Asian automakers. In the accompanying exhibition various demonstrations will show MOST50 and MOST150 solutions. sj New car models with integrated MOST technology I Figure 1. Aston Martin DBS I Figure 6. Lexus GX 470 I Figure 11. Toyota Land Cruiser Prado I Figure 2. Aston Martin Rapide I Figure 7. Lexus LS I Figure 12. Toyota Mark X I Figure 3. Audi A8 I Figure 8. MB E-Class Coupé I Figure 13. Toyota SAI I Figure 4. BMW 5 Series I Figure 9. MB SLS AMG I Figure 14. VW Touareg I Figure 5. Land Rover Discovery 4 I Figure 10. Rolls Royce Ghost I Figure 15. Porsche Cayenne 10 Elektronik automotive special edition MOST

9 Interview IIII MOST at Audi Let there be light A premium producer implements optical transmission technology Optical data transmission in automobiles was long regarded as a difficult proposition. Nevertheless, MOST has succeeded in establishing itself as an infotainment bus system. In an interview with Jens Kötz, responsible for electrical/electronic design, networking, energy systems and system safety at Audi, we hear where MOST is being used at present and where it is going in the future. Jens, where is Audi currently using MOST? Jens Kötz: Well, Audi first introduced the MOST infotainment network in 2002 in its A8 model. Since then the technology has gained a firm place as a communication network. In the Volkswagen group the MOST bus is currently being implemented in what s called the modular longitudinal power train or MLB, and as part of a topdown approach it s appearing in the Audi series A4, A5, Q5, A6, Q7 and A8. Series use in VW models has started with the Touareg. MOST is also scheduled to go into further series that are nearing SOP like the Audi A1, Audi A7 and a Bentley model. What s your experience with MOST been like up to now? Kötz: MOST is a technology created by infotainment developers for infotainment applications, in other words precisely tailored to the requirements of this domain. The many years of involvement of series production developers guarantees a broad and stable de facto networking standard for premium infotainment systems. The reservations that existed at the beginning about fiber optics in the automobile turned out to be unfounded. The processes are firmly in place, and the quality figures are of a very high level, despite the short innovation cycles common in the infotainment sector. Where does MOST go so far and no further? Are there any alternatives for certain purposes? Kötz: MOST shows to full advantage in infotainment applications like audio streaming. But MOST25 as a speed grade will soon reach its limit because of the increasing bandwidth demand we re going to see in premium infotainment systems. Speed grade MOST150 then offers interesting possibilities for solutions as well as performance reserves. So you re very positive in your assessment of MOST150? Kötz: The appearance of new infotainment functionality like video transmission, distributed HMI concepts and the integration of consumer electronics create a greater need for transmission bandwidth and communication services. MOST150 offers unique advantages as a speed grade through con- 12 Elektronik automotive special edition MOST

10 MOST at Audi IIII Interview sistent further development of the existing transmission technology. Infotainment applications place very different demands on the network infrastructure, and MOST150 gives you all the communication mechanisms over one interface, an all-in-one approach: Data rate is six times that of MOST25. Synchronous streaming to transmit audio data. Packet data transmission for navigation maps. Isochronous streaming for video transmission. Ethernet channel to tie in consumer electronics and online applications. Control channel to transmit control commands. Large potential for synergy in terms of process and investment, for example node costs identical to MOST25, plus high reusability of application software. These unbeatable advantages form a basis for deciding to introduce MOST150 throughout the Volkswagen group. To back up first-time use at Audi we ve prototyped all our planned applications on the network side based on two reference projects. Our findings, what we ve learnt from this is that MOST150 is ready to take off. And when, to be concrete, are we going to see MOST150 from Audi? Kötz: First-time use of MOST150 will be with the SOP of the new Audi A3 series. After that the technology will roll out successively through each of the series in the Volkswagen group. That means all vehicles of the modular transverse matrix MQB and the modular longitudinal matrix MLB. What then is the current status of MOST150 implementation at Audi? Kötz: Basic validation of the technology for going into series has been successfully completed. And at the moment we re developing, translating MOST150 into an SOP status for the Audi A3. How big is Audi s very own share in MOST150? Kötz: Network management and basic MOST functionality set up 100 percent on the MOST specification. In contrast to MOST25 technology we ve worked with the MOST Cooperation in specifying communication for ring break diagnosis. Through this standardization we re expecting a further betterment in quality. Analogous to MOST25 the function blocks in particular feature Audi-specific enhancements. Where are the interfaces between Audi and VW on MOST150? Kötz: As part of a module strategy we re developing the VW group s cross-brand and cross-series modular infotainment platform MIB. MOST150 is the group s dedicated networking technology for the extra featured standard, high and premium equipment variants. Here Audi is responsible for basic MOST150 development in the About Jens Kötz, Dipl.-Ing. Following technical college Jens Kötz started his career as a trainee communications electronics technician with Deutsche Telekom in Kempten. After that he studied electrical energy and automation technology at Konstanz university of applied science. In 1999 he joined Audi AG in the course of his dissertation on the subject Structure and analysis of an optical bus node (MOST). One year later he was given a position in Audi predevelopment and worked on the introduction of new networking concepts for the A3 and A8 models. Volkswagen group. Group-wide development makes it essential to cooperate closely with our VW colleagues. In this way specification and validation of the technology were jointly driven and harmonized. How do you rate MOST150 compared to Ethernet for instance? Kötz: MOST150 currently presents the best potential for synergy, is more versatile in application, for us it s the next logical step in technology. Compared to Ethernet, MOST150 is a synchronous bus with an extra embedded Ethernet channel. Time-synchronized transmission of audio and video signals can be implemented at attractive cost. You would first have to modify Ethernet software and networking technology for these applications. Given such constraints there are neither functional nor commercial benefits in moving over to Ethernet. So, to begin with, Ethernet will make its way into other automotive domains. Solutions in vehicle diagnostics access DoIP are being discussed for example. Can you imagine seeing MOST150 in non-infotainment? Kötz: The technical advantages being as they are, MOST technology could I Jens Kötz, responsible for electrical/electronic design, networking, energy systems, system safety In 2002 he moved to series development, focusing on the implementation of a networking architecture with central gateway. Parallel to this he expanded electrical/electronic architecture development at Audi. In mid-2005 he became manager of E/E architecture development/gateway. The focus here was on the definition of a modular E/E architecture within the modular longitudinal power train for the A4 family and FlexRay development for the new A8. On January 1, 2009 Jens Kötz was appointed director of the division electrical/electronic design, networking, energy systems, system safety. well be used for other automotive applications like transmitting image or sensor data in driver assistant systems. A number of technical aspects have to be evaluated here when it comes to topology and reliability of transmission. We re looking at these questions at the moment as part of examining the concept. Jens Kötz was interviewed by Stephan Janouch/sj Elektronik automotive special edition MOST 13

11 Physical layer IIII POF Influence on the bandwidth Adifficulty of polymer optical fibers (POF) is that they suffer significantly from intermodal dispersion. This means that if an optical pulse is launched into such a fiber, the optical power in the pulse is distributed over all (or most) of the modes of the fiber. Each of the modes that propagate in the POF travels at a slightly different speed. As a consequence of that, the modes in a given optical pulse arrive at the fiber end at slightly different times, thus causing the pulse to spread out in time with additionally increased rise and fall times as it travels along the fiber. This effect reduces the data rate transmission capabilities, i.e. decreases the transmission bandwidth. In consideration of the fact, that the high frequency performance and consequently the information-carrying capacity of the POF do not only depend on its own parameters like numerical aperture and its length, the outstanding question is, which other operating and environmental factors influence the bandwidth value and how much. Additionally, from the perspective of the system designer a certain value of the bandwidth of a POF can only be exploited entirely if the underlying transfer function is known. The term bandwidth has a different meaning depending on its application in the optical or electrical domain. The definition of the optical bandwidth is normally done in terms of the 3dB-frequency, which is the modulation frequency at which the optical power has fallen to one-half the value of the zero frequency modulation (dc-value). Thus, 3 db optical means a 50 percent optical power reduction. However, the electrical bandwidth is related to that frequency at which the Norm. Magnitude 0 db Dependencies of bandwidth of polymer optical fiber for MOST systems This article shows that the bandwidth of a polymer optical fiber (POF) cannot be regarded as an autonomous parameter to quantify its high frequency performance. Thus the question will be answered, what are the additionally required details to be linked to a bandwidth value. By Prof. Dr. Jürgen Krapp received electrical power has dropped to one-half of its dc-value. Under the assumption that the transfer function has a definite characteristic the relation of the electrical to the optical bandwidth can be calculated. In case of a Gaussian shaped transfer function the electrical bandwidth B 3dB ele is related to the optical bandwidth B 3dB opt by B = 2 B 3dB opt 3dB ele (1) A system s temporal performance is often specified by either pulse duration or rise time. Which one of these parameters is appropriate depends on the application. Obviously, pulse width, rise time and bandwidth are related quantities. Mathematically, the step response of Gaussian Transfer Function for 8 m POF; Launch AN = 0,5 B -3 db ele = 182 MHz Filter Function Measured Transfer Function MHz 800 Frequency (measured bandwidth = 187 MHz) I Figure 1. Measured transfer function of Mitsubishi Rayon POF and its Gaussian approximation. a linear system can be obtained by integrating its impulse response. Symmetrical pulse responses without tails or ringing can be approximated by a Gaussian shape. As a consequence of that, they can be regarded as the response of a Gaussian low pass filter. The step responses of Gaussian systems have a rise time t (10 90 %) that is only ten percent longer than the FWHM of pulse response and the rise time t (20 80 %) is approximately 71 percent of the FWHM pulse response. For this case additionally the relation to the electrical bandwidth is Frequency domain measurements yield information on amplitude-versus-frequency response and phat B 3dB =. / ele (2) In the essence that means, that besides clarifying the type of bandwidth (optical or electrical) the characteristic of the system s transfer function has to be known in order to convert the bandwidth value to the suitable temporal parameter. Investigation of POF transfer characteristic 14 Elektronik automotive special edition MOST

12 POF IIII Physical layer se-versus-frequency response. These data are usually more useful for system designers than time domain pulse measurements, especially if equalization techniques have to be performed on the detected signal at the receiver. The baseband frequency response is found from the ratio of the sine wave at the output and the input of the fiber. There are three strong advantages of the measurement technique in the frequency domain: The fiber transfer function is found directly and can be used to evaluate the response to any arbitrary input signal by Fourier-transforming the data. Photodetector linearity over a wide range is not required as in the time domain because of the small signal modulation about a constant average light level. The output signal can be received within an extreme narrow measurement bandwidth that reduces the influence of receiver noise significantly and makes the detection of little optical powers possible, e.g. at the end of long fibers. For any linear system the transfer function can be expressed by the central moments M m of its impulse response as (3) where A is area, t C is the central time, i.e. t C corresponds to the pulse delay, σ 2 = M 2 is the variance of the impulse response. Figure 1 shows the measured transfer function of Mitsubishi Rayon Heat Resistant POF of 8 m length using a launch aperture of AN = 0.5. The figure reveals that the POFs transfer function can be approximated at least up to the bandwidth by a Gaussian function. This is also valid for other fiber lengths and other launch apertures. For a Gaussian transfer function all odd central moments (M 3, M 5,... ) of the Gaussian pulse response are zero for symmetry reasons. The even central moments of a Gaussian pulse are given by m M = σ ( m 3) ( m 1) (4) m Thus, because of 4 M 4 = 3σ (5) etc., all terms in the exponent of Eq. (3) of third order and above are zero for a Gaussian transfer function. Consequently, if we disregard fiber attenuation (i.e. A = 1) and delay time (i.e. t C = 0), a Gaussian transfer function given as f Hf ( ) = exp (6) is a useful approximation for the polymer optical fibers transfer function. In Eq. (6) B 3dB is the electrical 3dB-bandwidth being used instead of the rms-width σ of the impulse response. B 3dB can either be determined by measurement or approximated by the relation Eq. (7) given in the next chapter. Influence of fiber length launch aperture ( ) ± ω ω ω H( ω) A exp jωtc σ 2 4 = + j M3 + M4 3σ For a polymer optical fiber the deterioration of bandwidth originates from modal dispersion. Material dispersion effects can be neglected. The analysis shows, that the bandwidth is not only a function of the length of the fiber but also depends on the type of optical source used. In real systems bandwidth will decrease less rapidly after a certain initial length because of mode coupling and differential mode loss. In the fiber the coupling of energy from one mode to another arises from structural imperfections, fiber diameter and refraction index variations, and cabling-induced microbends. B db 2 y = B -3 db 800 MHz The mode coupling has the tendency to average out the different propagation delays associated with the modes, and thus reducing intermodal dispersion. The result of this phenomenon is that, after a certain coupling length L C, the bandwidth will change from L 1 dependence to L 0.5 dependence. Because one dependency approaches the other with increasing fiber length L very slowly the bandwidth-length dependence can be regarded according to the empirical relation B = B L 3dB 0 q (7) The parameter q indicates the a- mount of mode mixing in the fiber and ranges between q = 0,5 for maximum mode mixing (if after the fiber coupling length L C the dynamic mode equilibrium is reached) and q = 1 in case without mode mixing. Figure 2 shows the measured bandwidth versus fiber length for launch apertures between 0.18 and The bandwidth-length dependency is approximated for each launch aperture by a potential regression according to Eq. (7). All regressions estimate the measured data within ±5 % accuracy. The Electrical 3 db-bandwidth for various Launch Apertures (T = 25 C) y = 8336x -1,22 y = 5288x -1,10 y = 4887x -1,11 y = 3243x -1,00 y = 2590x -0,95 y = 1967x -0,90 y = 1468x -0,87 y = 1260x -0,88 y = 1177x -0,87 y = 997x -0, m 20 x = POF-Length 0,18 0,24 0,27 0,30 0,32 potential (AN = 0,18) potential (AN = 0,24) potential (AN = 0,27) potential (AN = 0,30) potential (AN = 0,32) 0,35 0,40 0,45 0,50 0,56 potential (AN = 0,35) potential (AN = 0,40) potential (AN = 0,45) potential (AN = 0,50) potential (AN = 0,56) I Figure 2. Bandwidth versus POF length, various launch apertures (Mitsubishi Rayon). Elektronik automotive special edition MOST 15

13 Physical layer IIII POF B-3 db 200 MHz C 25 C 105 C I Figure 3. Bandwidth dependence on temperature (Mitsubishi Rayon POF). measurements reveal that the coupling length L C also depends on the launching condition because the value of q in the exponent increases for decreasing launch aperture. The conclusion is also that the majority of measurements correspond to lengths less than the coupling length of the fiber if q is close to 1. B -3 db 200 MHz Electrical Bandwidth for Launch Aperture AN = 0, C 200 Temperature I Figure 4. Electrical bandwidth and excess loss versus temperature (Mitsubishi Rayon POF: L = 15 m, launching AN = 0.50). B-3 db 250 MHz B -3 db [MHz] Excess loss [db] Maximum Temperature Temperature Test: heating and cooling down I Figure 5. Irreversible change due to high temperature exposure; bandwidth versus temperature (Mitsubishi Rayon POF: L = 15 m, launch AN = 0.56). 20 Length in m MHz 79 Influence of temperature All results of this chapter refer to Mitsubishi Rayon Heat Resistant POF. Figure 3 illustrates the influence of temperature on bandwidth for a launch aperture of In the essence, heating this POF up to 105 C does not cause an important limitation of transmission bandwidth. 2,5 175 C The effect of temperature increase to 2,0 105 C is not significant and cooling 1,5 the fiber to 40 C 1,0 again even slightly increases bandwidth. 0,5 0 Figure 4 exposes the temperature excess loss by heating T [ C] MHz limit of 175 C beyond which the signal transmission will be interrupted. Heating the fiber to 175 C is accompanied by an excess loss of approximately 2,5 db above the normal fiber attenuation. Figure 5 documents a temperature test of a 15 m long fiber heated up to 170 C (red colored bars) and consecutively cooled down again to 24 C (blue colored bars). The heating of the fiber above 105 up to 170 C results in a bandwidth augmentation which remains irreversible if the fiber is cooled down again to room temperature. This goes along with an excess loss of approximately 3.4 db in addition to the normal fiber attenuation. The fundamental characteristic of the Gaussian shaped transfer function usable for the approximation remains. A Gaussian characteristic is an easy manageable approximation for the POFs transfer function. For each fiber length L its bandwidth can be determined by a potential function B 0 L q considering that its parameters B 0 and q depend on the launch aperture of the light source. Regarding Mitsubishi Rayon Heat Resistant POF environmental temperature within a range of 40 to 105 C has only a minor effect on transmission bandwidth. Beyond of 105 C bandwidth increases irreversibly but the excess loss also. Thus for a given fiber length and launch aperture, realistic analyses can be performed in order to define the limits of performance of the system more clearly. sj Prof. Dr. Jürgen Krapp achieved the degree of Doctor of Science at the University of Saarland (Germany) in Since 1994 he teaches at the University of Applied Science in Aalen in the department of Optoelectronics and in the Master Course of Photonics primarily specialising in the field of optical communication systems. juergen.krapp@htw-aalen.de 16 Elektronik automotive special edition MOST

14 Development + test IIII Physical layer (Foto: Audi AG) Use in development and production Stress test tool for MOST150 physical layer Performing a physical layer test for control units is simplified by using the Physical Layer Stress Test Tool. This covers essential functions such as generating a worst case pattern, signal conditioning, a pattern comparator and test procedure control. The high grade of transparency and efficiency for carrying out and analyzing physical layer test is also provided. Therefore, the foundations exist for the start of the compliance verification process for MOST control units. From Georg Janker With the introduction of the MOST150 compliance verification process for the physical layer, the MOST Cooperation has developed a test procedure, which reduces the number of components to be used. The signal generation and verification can essentially be covered with a single device: the Physical Layer Stress Test Tool (PhLSTT; figure 1). The measurement setup is complemented only by the usage of standard components, e.g. oscilloscope, light power meter, attenuator or mode mixer. From the beginning, it was possible to create a standard successfully, that allows the highest possible uniformity for the verification of MOST150 control units in the physical layer best preconditions for identical test processes and test results worldwide. The Physical Layer Stress Test Tool is a mandatory part of the compliance verification and therefore it also plays a key role. This opens up the following areas of application for the PhLSTT: performing of the compliance process in the accredited test houses. accompanying development verification at the control unit manufacturers. quality assurance at the end-of-line in production. verification at the system integrators. A physical layer test can be simplistically described by two definitions of tasks (figure 2): Applying worst case Elektronik automotive special edition MOST 17

15 Physical layer IIII Development + test scenarios for the control unit and verifying the signal produced by the DUT. Applying worst case scenarios for DUT The main task of the PhLSTT is to generate worst case patterns, which are applied to the control unit to be tested (so-called Device Under Test, DUT). By using a comparator, the output signal of the DUT is compared with the original pattern and then checked for errors. PHYSICAL LAYER STRESS TEST TOOL worst case pattern verifying DUT's signal quality I Figure 1. The Physical Layer Stress Test Tool (PhLSTT). On the one hand, it can thus be determined if the DUT correctly interprets the bit sequence of the worst case pattern and, on the other hand, can also correctly output it again without violating the compliance standard. The worst case pattern scenario is combined with other stress conditions (e.g. temperature, supply voltage, light power) during compliance test execution. Signal quality check Verification of adherence to the signal quality under all possible conditions is the primary objective of the compliance verification. For this reason, it is also important to check whether the worst case signals are not only correctly received and interpreted, but also whether the control unit adheres at output (SP2 in figure 1) to the compliance requirements, also under these conditions (figure 3). For this purpose, in the measurement setup, in addition to the Physical Layer Stress Test Tool, an oscilloscope is combined in order, e.g. to check the eye diagram. Pattern generator and signal conditioner The pattern generator offers the possibility to apply a pattern at different =? data recovery worst case signal SP2 outputs of the PhLSTT. The pattern is generated by a high precision clock and optimized with a signal conditioner additionally concerning to the pulse width. This reduces the transferred jitter to a minimum. Furthermore, the signal conditioner enables a deterministic adaptation of the duty cycle adjustment in high resolution. The pattern can be picked off via different outputs on the PhLSTT alternatively: A commercially available FOT unit, a LVDS output Temperature Chamber SP3 DUT Master/ Slave SP2 operating with an external EOC or with the integrated High-Quality-EOC. The pattern defined by the MOST Cooperation for the compliance process is integrated in the unit by default. However, individual patterns can also be loaded via a communications interface. High-quality EOC A special feature of the integrated EOC is the high signal quality at maximum light power. The architecture of the EOC is also aligned to the pattern generator and signal conditioner, which enables a very accurate, reproducible signal. As a result, a reference signal can be generated, which can be modified by means of additional components (such as e.g. optical attenuator or mode mixer), to fit the test scenario. An characteristic of the EOC is operation at more than +1 dbm, that enables the important test for compliance with maximum optical output power at the input of the DUT. Stress Pattern Generator V Opt. Tx El. Opt. Rx El. PHYSICAL LAYER STRESS TEST TOOL SP2 Optical Attenuation SP3 Oscilloscope I Figure 2. Physical layer testing scenario. Mode Mixer Meas. OEC Pattern comparator In addition to generation of the signal, it is important to check whether the pattern sent is also correctly recognized by the DUT (device under test) during the test execution. In order to be able to carry out this check, the DUT is set in a special bypass mode; the Retimed Triggered Bypass Mode. This mode forwards the received pattern in unmodified form. In this way, the pattern at the output of the DUT can be compared with the pattern applied from the PhLSTT. For this case a pattern comparator is implemented in the PhLSTT. The pattern comparator can be switched to the integrated FOT unit as well as to an electrical input, which enables the connection of an external OEC. Various scenarios can thereby be realized such as e.g. the combination of an SP2 measurement and an oscilloscope. Analysis of the pattern comparator relates to two characteristics: The bit errors occurred and sudden phase shifts in the received signal. Both incidents are counted by the PhLSTT and reported after completion of the test. During the test, bit or phase errors that occurred can be more precisely located and analyzed via trigger outputs of the PhLSTT. Automation und test procedure The several steps, that are required for performing the test: Switching of the DUT to retimed triggered bypass mode: So that the DUT does not modify the signal in any way, it must be set to a so-called retimed triggered bypass mode. This takes place with the method Physical- LayerTest, which is called in the function block Enhanced Testability in the DUT. Switching of SP3 SP2 DUT Master/ Slave Temperature Chamber V the PhLSTT to stress mode: In order to generate the pattern, the PhLSTT must be set from the normal MOST mode to pattern mode. 18 Elektronik automotive special edition MOST

16 Development + test IIII Physical layer Repeated sending of the pattern for a specific period of time: In the test phase, the pattern is cyclically sent for a predetermined period of time. A trigger output marks the start of the pattern in order to enable analyses with an oscilloscope during the test procedure. Pattern comparator operation: The pattern comparator is activated when it is ensured that the DUT and the PhLSTT are in the stress mode. From now on, all errors occurring both in the DUT and in the pattern comparator are counted. Switching to MOST mode: In order that the results can be compared after Tester RING NIC MOST FRAME XXX I Figure 3. Typical measurement setup. the test, both DUT and PhLSTT have to switch back to normal MOST mode after a previously configured period of time. Readout of the results: The measured values of the unlock indicator and the error counter in the DUT are queried via the method PhysicalLayer- TestResult provided by the function block Enhanced Testability T 1-3 T TEST T T STRESS PATTERN XXX T MOST FRAME The entire test procedure is carried out by the PhLSTT independently and automatically. The parameterization, the start of the test procedure and the results analysis are controlled via a serial interface. An example program is provided with the PhLSTT. Therefore, performing a compliance test is possible immediately after completion of the test setup. bg Georg Janker is CTO of Ruetz System Solutions GmbH and since 1997 he s engaged in development of infotainment systems based on MOST with focus on system definition and validation of networks based on MOST, development of MOST Compliance Standards, MOST Compliance Test House and development of the test and simulation tool Testerlyzer, TTsuite MOST and PhLSTT. Elektronik automotive special edition MOST 19

17 Physical layer IIII LEDs Green light for MOST GaN high-speed LEDs and their application in a POF-based communication module Toyoda Gosei has developed green Gallium Nitride LEDs for data links using POFs. Their frequency response characteristics and eye diagram measurements show that green LEDs have the capability of high-speed data transmission with up to 500 Mbit/s. Temperature dependencies of the dominant wavelength, emission intensity and forward voltage are small enough for usage in high-speed POF based data links. By Yukitoshi Inui, Naoyuki Okita and Chris Mesnager As an application Toyoda Gosei developed a bidirectional single POF communication module, whose optical circuit consists of a WDM polymer which was fabricated using light-induced self-written waveguide techniques, using green and a conventional red LEDs. In POF data link, the red 650 nm LED has been used as a general light source. Figure 1 shows that the standard step-index (SI) PMMA POF has some minimum loss windows in the visible wavelength region. Temperature measurements of the red and green LEDs during operation show, that the green LED has fewer dependencies of the dominant wavelength; operating between 0 and 85 C the wavelength fluctuates than 3 nm (figure 2). The red light source has a minimum attenuation near 650 nm, yet in a very narrow range. Therefore, the optical power loss will be increased over temperature due to shifts of the peak wavelength. As a countermeasure a resonant cavity structure LED designed for high-speed operation at 650 nm must be chosen. On the other hand, because there is a largely low attenuation in the area from 480 to 530 nm, green LEDs have an advantage in terms of temperature stabilization. Therefore, the green LED is able to enhance the power budget and communicate over long distances. The GaN LED has previously been developed for applications as indicators or displays. The green GaN LED can become the new light source in the field of optical communication. Therefore Toyoda Gosei started to develop a green high-speed LED which can be used at data transmission rates of up to 500 Mbit/s. Toyoda Gosei possesses a light-induced self-written optical waveguide technique that has been developed in cooperation with Toyota Central R&D Labs. Inc. Resin is photo-polymerized using blue light radiation from a fiber. A straight waveguide can be produced due to a self-trapping mechanism. Using this process a bidirectional transceiver using a single POF and two wavelengths (green and red LED) was produced, which means that there is no need for the alignment of optical waveguides and elements such as a light source or photo diode. High-speed GaN LED The green high-speed GaN LED has been developed by aligning LED tip structure and fabrication. For evaluating the transmission characteristics of the developed green LED, a high frequency driver circuit was developed. The random pattern (1.0 V pp ) from the pulse pattern generator (MP1632C Anritsu) was fed into the transceiver at a data rate of 150/400/500 Mbit/s. The optical power through the POF (V L51501 G&G, 1 m length) was detected by a avalanche photo diode and converted to a electrical 10 3 db/km nm 525 Green LED 660 nm 555 Red LED Attenuation ,, Wavelength nm Wavelength nm nm Wavelength Green & Blue LED Conventional Red LED C 100 Temperature C 100 Temperature I Figure 1. Attenuation of a step-index PMMA POF. I Figure 2. Temperature dependencies of green and red LEDs. 20 Elektronik automotive special edition MOST

18 LEDs IIII Physical layer signal. The waveform obtained by an oscilloscope measurement (SDA 5000A LeCroy) under operation at 500 Mbit/s is shown in figure 3. The signal contains a slight jitter, but the large aperture suggests that this transceiver can transmit successfully at data rates of up to 500 Mbit/s. LISW optical waveguide Using the LISW technique a LISW optical waveguide is produced out of resin, which is photo-polymerized by blue laser light coming out of a optical fiber. The resin begins curing from the I Figure 3. Eye diagram of a transmitter using a green LED at a data rate of 500 Mbit/s. core tip where the light is most intense. Because the refractive index of the hardened portion is getting higher than of uncured resin, the core is formed by a self-trapping mechanism. The core grows straight along the optical axis while keeping a uniform diameter. Furthermore, if in advance a half mirror is inserted into the optical axis a branching waveguide can be obtained by a simple process. Applying this fabrication method, it low cost optical communication devices can be produced which do not require a costly alignment. Proposal for a module using a green GaN LED I Figure 5. Schematic view of a bidirectional optical module and picture of a prototype of the transceiver. I Fiugure 4. Prototype of an optical communication module using a green LED. Figure 4 shows the prototype of an optical communication module using a green LED instead of red one; so highspeed communication performance can be achieved at a minimal cost. For example, when some perimeter surveillance cameras are installed on a vehicle, the amount of data that has to be transmitted will be higher than 440 Mbit/s (picture resolution , color 16 bit RGB, 30 fps, 4 channels, uncompressed). Using a green LED it will be possible to transmit this amount of data. Figure 5 shows a schematic diagram of a bidirectional optical module and a prototype of transceiver. The transceiver consists of the LISW waveguide, a WDM filter, LEDs with two wavelengths (500 and 650 nm) and a photo diode. The optical LISW circuit is mounted in a transparent plastic case. Using a branching LISW waveguide and a WDM filter, bidirectional communication using a single POF can be achieved.the process of bidirectional data communication using this module works like this: The light signals emitted by a green LED are launched into the waveguide. The green signals are reflected by the filter and arrive at the other module through an optical fiber connected by in-line connectors. Then the signals are reflected by the filter again, so that the beams reach a photo diode mounted on the other module. The WDM filter is designed to reflect green light and to transmit red light. The MOST is currently based on a ring topology. Therefore, when a camera is used in a car for perimeter surveillance, it can receive the data but it can not loop back a control signal to the camera without an additional connection. As a technical solution, the modules developed by Toyoda Gosei enable a bidirectional transmission using a single POF. sj Yukitoshi Inui joined Research & Development Division of Toyoda Gosei in Since 2000 he is engaged in the development of the polymer optical waveguide. Currently he serves as team leader for the fiber optical transceiver development. Naoyuki Okita works as engineer at the Research & Development Center of Toyoda Gosei Co., Ltd. Since 2008 he is dealing with the circuit design of fiber optical transceivers. Chris Mesnager joined Toyoda Gosei Europe in 2004 as European Sales & Marketing Manager for the Toyoda Gosei Optoelectronics Division. Currently he is European General Manager. Elektronik automotive special edition MOST 21

19 Physical layer IIII Specification point 3 Automated determination of worst-case scenarios For MOST optical physical layer specification point 3 In this article an approach to determine and analyze the worst-case scenarios of the input to the specification point 3 (SP3) is presented. Besides the worst-case analysis infomation about degradation mechanisms due to the influence of the plastic optical fiber (POF) is provided. From Andreas Braun, Markus Becht, Dr. Oliver Bringmann and Prof. Dr. Wolfgang Rosenstiel MOST Device MOST Network Interface Controller Tx SP1 EOC Optical networks are immune to interferences from other electrical sources an impact on signal integrity doesn t exist. However, limitations on the bandwidth of the optical transfer medium have an effect on the signal characteristics and need to be considered in specifications for optical links. MOST150-standard uses the identical POF as used in MOST25. While in MOST25 the impact of POF bandwidth limitations was negligible, the higher speed of MOST150 requests a detailed investigation. The bandwidth limiting characteristic of the POF influences the transmitted signal. The fibers bandwidth depends on its length and the numerical aperture (NA) of the POF. A mathematical representation of the POFs transfer function was defined which is useful for prediction of degrading impacts on signal integrity. Additionally, transmitter properties such as launch NA and the pulse shape of the emitted optical signal directly influence the signal passing the fiber. Therefore, it is important to have knowledge of worst-case input scenarios and their effect on signal parameters after the data transmission. SP2 SP3 OEC I Figure 1. Definition of the specification points. SP4 MOST Device MOST Network Interface Controller Rx A profound analysis of extreme transmitter characteristics and their influences on the signal transfer through POF enables to specify the constraints on optical emitters as well as to predict worst-case input conditions for optical receivers. Introduction The MOST150 Optical Physical Layer Specification provides a definition of the optical link by four specification points (SP1 to SP4) marking significant interfaces (figure 1). The optical input parameters for the optical electrical converter (OEC) are specified at SP3. The signal at SP3 is influenced by various sources. First of all, the output signal at the electrical optical converter s (EOC) output SP2 may vary in the given tolerance ranges. This includes variations in pulse shape, extinction ratio, timing distortion, optical output power and launch conditions. The transport medium POF may also influence the data signal. The dominant impact is the optical attenuation caused by the POF and optical interconnects. Another degradation of the signal quality comes with the bandwidth limitation of the POF, leading to degradation of transition times, inter symbol interference (ISI) and duty cycle distortion (DCD). The MOST specification defines a mathematical representation of the POF s transfer function, which was determined based on profound investigations and theoretical considerations. To achieve a deeper understanding of worst-case scenarios at SP3 based on SP2 input conditions a simulation study was started. To find worst-case patterns for SP3 and to learn about degradation mechanisms was one target of this study. For the simulations, the fiber-length and launch NA was fixed as the worstcase settings of the MOST150 automotive optical physical layer. For automated determination of worst-case scenarios for different parameter sets, a program based on Matlab was developed. Matlab is a numerical computing environment and programming language. The software generates separate pulse shapes for rising and falling edges where the full tolerance range given by the MOST physical layer specification is utilized. These edges are combined to a single signal by using valid MOST patterns, coded in DC Adaptive Coding (DCA). Then the impact due to the transmission media is simulated. The use of complete MOST frames guarantees that low frequencies are also included in the calculation. Afterwards the SP3 results are stored for visualization and analysis. These steps are iterated until the desired parameter coverage or cross-coverage is reached. The coverage is determined based on parameters for describing the pulse shapes of falling and rising edges. SP3 input pattern calculation All patterns are generated using normalized amplitude and normalized time- 22 Elektronik automotive special edition MOST

20 Specification point 3 IIII Physical layer scale. The amplitude is based on the steady state high- and low-level of the optical signals (B1, B0). A specific method for achieving these parameters is given in the MOST specification. Amplitude value 1 represents steady state B1; an additional overshoot up to 1.4 is permitted (figure 2). Amplitude 0 represents steady state B0, which doesn t mean zero necessarily. A small bias level is possible; a limitation is given by the parameter extinction ratio. Time scale is defined in Unit Intervals (UI), derived from the network bit rate. Due to the coding scheme, the shortest pulse width is 2 UI. There are further pulse types with pulse length up to 6 UI. The SP3 input pattern calculation is divided into three main steps. These steps are described in detail in the following three sections. Rising and falling edge generation The generated transitions represent pulse shapes of a length of 2 UI. Rising edges start with amplitude 0 and end with amplitude 1, falling edges start with 1 and end with amplitude level 0. A rising and a falling edge is described by four parameters. These parameters are edge, surface, offset and depth for the falling edge or height for the rising edge. Before generating an edge the minimum and maximum values of these parameters can be defined, to generate edges with special properties. Edge: transition time between 20 and 80 % of the normalized amplitude. Surface: accumulated power within the surface over or under the normalized amplitude. Offset: time deviation between transition and zero point on the time scale measured at amplitude level 0.5. Depth, Height: peak amplitude of the generated rising or falling edge pattern. For generating pulse shapes, coordinates at start and end of the curve are fixed, while sample points in between are generated randomly. In order to control the process of generating random coordinates, boxes are defined wherein one random coordinate per trial is generated. Then curves are fitted through the sample points using Matlab Polyfit in which basic polynomial and spline interpolation is possible. The timing resolution of 30 samples/ui was chosen. The method of limiting the degree of freedom by introducing the boxes helps to avoid unrealistic pulse shapes and allows an adaptation to the needs of the chosen fit-algorithm. The random ranges for polynomial interpolation are shown in figure 3. After a calculation run, the pulse shape is checked for consistency with the SP2 requirements and the user defined parameter range. If all constraints are matched, the pulse shape is accepted or else the algorithm runs through 10,000 trials. Combining the rising and falling edge with a pre-defined bit pattern For creating a SP2 pattern, pulse sequences of rising and falling edges are consecutively combined. The data structure is defined by a pattern file providing binary data sequences in UIs. The MOST stress pattern as well as user defined bit pattern can be used. The bit pattern must be offset free and the starting point must be equal to the end point to avoid dirac impulses while using FFT. For every SP2 pattern (figure 4) one particular pulse shape of each, rising and falling edge, is used. Pulses of the pattern which are greater than 2 UI are connected by padding with either 1 or 0 between the rising or falling edges. Though all specifications for SP2 are fulfilled and the additional constraints are matched, the resulting pattern might be idealistic compared to real optical transmitters used in the network. The bandwidth capabilities of optical transmitters is limited, the calculated patterns however might easily exceed this natural limitation. Besides all efforts to create worst-case patterns according to SP2 specification and to avoid unrealistic pulse shapes, the final database will contain scenarios that need to be excluded manually. Calculation of the SP3 pattern Rising Edge (1) 1,6 1,4 Height 1,2 Surface 1,0 0,8 0,6 Edge A 0,4 Offset 0,2 0-0,2-0,4-0,6-0,4-0,2 0 0,2 0,4 0,6 0,8 1,0 1,2 1,4 UI I Figure 2. Pattern parameter. 1,6 1,4 1,2 1,0 0,8 A 0,6 0,4 0,2 0-0,2-0,4 Rising Edge -0,6-0,4-0,2 0 0,2 0,4 0,6 0,8 1,0 1,2 1,4 UI I Figure 3. Polynomial interpolation for a rising edge pattern. A prediction for the scenario at the fiber output is calculated as the convolution of SP2 pattern and the POF s transfer characteristic. The generated SP3 pattern (figure 5) can be visualized and used for further analysis. A data set, containing characteristic parameters for input and output parameters, is stored. This allows loading and also recalculation of the SP3 pattern together with any bit pattern. Simulation and analysis The simulation program can generate multiple randomized rising and falling edges at once. Afterwards different SP2 patterns are generated using all permutations of the rising and falling edges 1,6 1,4 1,2 1,0 0,8 0,6 A 0,4 0,2 0-0,2-0,4 0 I Figure 4. Generated SP2 pattern. SP2 Pattern UI Elektronik automotive special edition MOST 23