FUJITSU TEN TECHNICAL JOURNAL,No.33,KOBE,JUL.2009 ISSN 0289-3789
2009-Jul No.33 TECHNICAL JOURNAL Front cover description AVN Lite "AVN339M"
This model incorporates a digital terrestrial TV tuner that provides stable reception of high quality Full seg video and audio with four antennas and four tuners. Its main unit has a slot for a B-CAS card that is required to watch and listen to the digital terrestrial TV. It plays commercial-release DVDs and home-made videos recorded in the DVD- VR format. With its compatibility with MP3/WMA files in CD-R/RW, this model gives you access to more videos and audios than our conventional models.
FUJITSU TEN LIMITED PROFILE
Sales Growth (¥100 million) Head Office: 2-28, Gosyo-dori, 1-chome, 3,800 Hyogo-ku, Kobe, 652-8510, Japan 3,641 3,600 Phone: +81-78-671-5081 3,400 FAX: +81-78-671-5325 3,125 3,200 Established: October 25, 1972 2,992 3,000 Capital: ¥5300 million 2,746 2,800 No. of Employees: 3,783 (Consolidated: 12,267) 2,600 Product Line: Infotainment Products 2,400 • Audio, Navigation Products (for automobiles) 2,200 • Audio (for home use) 2,000 1,800 • Mobile Communication radios 1,600 Car electronic devices 1,400 Domestic Offices: Utsunomiya, Tokyo, Toyota 1,200 Kobe Logistics Center 1,000 Nakatsugawa Technical Center 800 Kobe Plant, Nakatsugawa Plant 600 Affiliated Companies: TOCHIGI FUJITSU TEN LIMITED 400 FUJITSU TEN EAST JAPAN SALES LIMITED 200 FUJITSU TEN CHUBU SALES LIMITED 0 FUJITSU TEN WEST JAPAN SALES LIMITED '05 '06 '07 '08 FUJITSU TEN ACTY LIMITED Consolidated sales FUJITSU TEN RESERCH LIMITED FUJITSU TEN STAFF LIMITED FUJITSU TEN TECHNOLOGY LIMITED
Head Office / Main Plant Nakatsugawa Plant Tochigi Fujitsu TEN Limited
FUJITSU TEN TECH. J. No.33(2009) Editorial Board
Editor Tohru OGAWA
Associate Editor Takao KAMAI
Editorial Staff Yuji NISHINO Shinji FUKUDA Takashi EGUSA Kazuhiro YOSHIYAMA Keizo ISHIMURA Takeshi TANAKA Eiji TARUTANI Yoshinori MARUYAMA Tetsuya KITANI Michihiro KITA
Bureau Keiichi YOSHII Toyokazu MATSUI Izumi SASAKURA
• FUJITSU TEN TECHNICAL JOURNAL is published semiannually by FUJITSU TEN LIMITED to report the results of research conducted by FUJITSU TEN LIMITED in car electronics and related fields. • All rights are reserved by FUJITSU TEN LIMITED concerning the publication, and publication after translation into other languages, of articles appearing herein. • Permission to publish these articles may be obtained by contacting the editor.
FUJITSU TEN LIMITED 2-28, Gosyo-dori, 1-chome, Hyogo-ku, Kobe, Japan Phone +81-78-671-5081 http://www.fujitsu-ten.co.jp/ Printed by RYOSAN PRINTING © 2009 FUJITSU TEN LIMITED (Jul. 2009) All brand names and product names are trademarks or registered trademarks of their respective companies.
FUJITSU TEN TECH. J. No.33(2009)
FUJITSU TEN TECHNICAL REPORT
Jul. 2009 No.33
PREFACE Power of Integrating Each Single ...... 1 Hiroshi OGAWA
PAPERS Development of 09CY Audio Platform ...... 3 Akira YOSHIDA Hirokazu MATSUNAGA Teruaki HAYASHI Kiyotaka NAKAGAWA Yoshikazu FUJITA Yasuhiro YAMAKAWA
Development of Time Domain Speaker: TD712zMK2 ...... 11 Kazuhiko HAMADA Kiyosei SHIBATA Yoshikuni MIKI Ryousuke SOU Masahito KAWAI
Development of FUJITSU TEN's First SD Memory Card Navigation System, AVN Lite, and its Linkage Service...... 18 Kenichirou MASE Takeshi FUKADA Kenji UEMATSU Teru SAWADA
AVM System ...... 27 Satoshi SHIGEMATSU Yoshikazu CHUJO Haruo YOSHIDA
NOTE Technology Trend in Remote Engine Starter ...... 35 Norio TSURUTA Akira TAKEE
Approaches to Technology of Thermal Fatigue Life Prediction of Solder Joints ...... 40 Hidehisa SAKAI Katsufumi MORIMUNE Takashi YONEMOTO
INTRODUCTION OF OVERSEAS TECHNICAL CENTER Introduction of FUJITSU TEN SOLUTIONS PHILIPPINES, INC. (FTSP)...... 48 Akira IKEZOE Hisashi FUJII Catalino GAYOSA III Rommel P. TUAZON Philip S. TENORIO Lez T. LEE
PREFACE Power of Integrating Each Single
Director Hiroshi OGAWA
We can see the keywords, "eco, green, and earth-friendly," in the newspaper, TV, and elsewhere every day. Also in the inaugural address of U.S President of this year, he stated his environmental policy, "We will harness the sun and the winds and the soil to fuel our cars and run our factories." Hence, the technology that responds to the environment is necessary for future product develop- ment. Through the whole life cycle of products including material, design, manufacture, logistics, and disposal, we must create high-qual- ity, eco-friendly, and competitive green products. As for our engine control unit (called EFI-ECU) that the AE Group works on, a quarter of a century has passed since its mass production began in 1983, and its cumulative production now exceeds 20 million units. I would like to thank our senior associates and all involved staff for their integrating each efforts. The EFI-ECU is the electronic equipment that precisely calculates optimum amount of fuel injection and ignition timing in accordance with the condition of engine, and cleans the exhaust emission. The EFI-ECU is therefore an environmental contribution product. In 2006, we commercialized the equipment for hybrid vehicle control (called HV-ECU) as another environmental contribution product, following EFI. The HV-ECU offers good fuel consumption, clean exhaust emission, and smooth driving by precisely calculating and controlling the output torque of the engine and driving motor of a hybrid vehicle. The EFI-ECU and the HV-ECU are classified in the category of the power train in the automotive industry. They relate to "run", which is one of the significant functions of a car, so called "run, turn, and stop," so their quality is considered most important. For these products, the integration of right parts, right software, and right monozukuri is specially required. The most important issues for commercialization are that each work process is clear, high-quality work is integrated along the policy of Σ0=0 (1), and preventive action
(1) It is an activity for quality improvement in which each and every person in the Fujitsu Ten group achieves 0 defect individually to result in total 0 defects.
FUJITSU TEN TECH. J. No.33(2009) 1 is front-loaded, and all activities to avoid defects are completed before mass production. In the design and manufacturing site, opti- mization, standardization, and visualization must be done in all processes from input to output, and activities for brushup must be made every day. Comparing the first model of the EFI-ECU with the latest HV- ECU, we can see the variation and evolution of car electronics dur- ing a quarter of a century. At first, an 8-bit microprocessor was used, but now 32-bit microprocessor is mainstream. We have improved the quality level dramatically while the operation speed is increased by 100 times and memory size is increased by 200 times due to the high-speed and high-density current of semiconductors. In line with the increase of proportion of software design in product development, we have established our method by platform concep- tion, development tool, management tool, and so forth for efficient development and design. Furthermore, in the product as a whole, we have continuously achieved a good balance between functional extension and reduction in size and weight by improving every aspect, including the miniaturization of element device to be installed in the product, evolution of high-density mounting technology, and the establishment of slim and highly-efficient production line. On the other hand, the amount of information used in an ECU has been increasing dramatically. In the current ECU, the information required for the control is obtained by intercommunication through multiple networks among ECU's. By linking, cooperation and liaison between control functions can be fully carried out, and the range of working field is further expanded for automotive engineers. We must understand the operation of other devices interacting and coor- dinating with each other, and must repeat discussions with the experts including other companies from the early stage of develop- ment. Not withdrawing into ourselves, we must widen the circle of linkage, fusion, and cooperation in and out of our company. With the sophistication and progress of computerization of cars in the future, we would like to increase as many human resources as possible, who are capable enough to respond and solve any issue, so that we stay ahead of other companies. We, who had overcome the once-in-200-years earthquake disaster, now face once-in-a-century unprecedented economic circumstances, and have been pursuing our plan to survive this crisis and grow. We all will step up our efforts of contributing to engineering devel- opment one by one so as to benefit future generations and advance to the next stage.
FUJITSU TEN TECH. J. NO.33(2009) 2 Development of 09CY Audio Platform
Akira YOSHIDA Hirokazu MATSUNAGA Teruaki HAYASHI Kiyotaka NAKAGAWA Yoshikazu FUJITA Yasuhiro YAMAKAWA
Abstract
With the development of digital technology, the mainstream digital media for use in car AV products has changed: from CD, DVD to HDD. In terms of broadcasting media, various types of digital broadcasting have begun in countries all over the world. While in AM / FM radios, further performance improvement is required. Under these circumstances, from 2005CY we introduced the digital processing platform for the circuit blocks of AM / FM IF and thereafter. However it became relatively expensive for a low-grade CD-tuner, because it had too many various specifications. Therefore, we developed a new platform with improved performance and optimized functions, dedicated to a CD-tuner. In this paper, we set out the function outline, characteristics and main technology involved.
3 FUJITSU TEN TECHNICAL JOURNAL
1. Introduction Digital IF processing tuner Introduction Digital IF processing tuner 1 (1st-generation) (2nd-generation)
FUJITSU TEN TECH. J. NO.33(2009) 4 Development of 09CY Audio Platform
・System / Radio / Audio controller CPU： 32bit, 17.64MHz (decided conforming to multiple control and expandabil- ity for requirements) ・CD controller CPU：8bit, 17.64MHz (decided conforming to downsized scale and low power consumption) ・Radio / Audio / MP3, WMA, AAC decoder DSP： 24bit, 135MHz (decided conforming arithmetic accuracy and arith- metic processing power)
4.2 Configuration of audio processing and func- Fig.2 System Block Diagram of Conventional Product (07 model) tional specifications of I/O terminal Integration of radio / audio LSI and CD LSI Fig. 5 shows the diagram of the audio processing block. Table 1 shows specifications for I/O audio signals. This audio configuration enabled audio processing includ- ing mixing of multiple sound sources.
Fig.3 System Block Diagram of Developed Product (09 model)
4 Device4. Device Development Development 4.1 Device Configuration The basic system of radio / audio / CD control is Fig.4 Device Configuration Diagram composed of two devices below. (Fig. 4 shows this device configuration and skeleton block configuration inside the devices.) a) Radio front-end LSI b) Radio back-end & audio & CD LSI (Integrated LSI of radio / audio / CD) The integrated LSI mentioned in b) is composed of the following five blocks. (1) System / Radio / Audio controller (2) Radio back-end block (3) Audio processing block (4) CD controller (5) CD & MP3 / WMA / AAC decoder block Controller blocks mentioned in (1) and (4) are to be controlled by a built-in CPU. This system enables the microcomputer system to control the blocks by com- mands, and reduce the control load of microcomputer sys- tem. Each block of radio / audio / CD is composed of hardware processing and software processing (DSP). Fig.5 Diagram of Audio Processing Block Each throughput (bit width / operating cycle) of installed CPU / DSP is as follows.
FUJITSU TEN TECH. J. NO.33(2009) 5 FUJITSU TEN TECHNICAL JOURNAL
Table 1 Specifications for I / O Audio Signal with heat resistance performance prior to designing.Also, using the estimated result of this current consumption by I/O Audio processable Type block for the block layout mentioned above, the block lay- I/O system signal (DSP) out in consideration of the dispersion of power (current I2S PORT 5 sys flow) is implemented. Further, building the low-power 2 sys Max. 4 stereo DIR PORT consumption mode by block (or by function unit) led to (exclusion) signals selectable the circuit in which respective functions can work with Digital （Radio Block） 1 sys from 8 sys the minimum current flow. （CD Block） 1 sys c) Effort for fail safe Balanced Building fail safe processing function for each individ- 2 sys Max. 2 analog PORT ual block by the multi-function system enables the
Input signals selectable respective recovery of radio / audio and CD function. Single 1 sys from 3 sys This processing outline is explained in Section 4.4. PORT
Balanced Analog 2 sys Max. 2 stereo 4.4 Fail Safe Function PORT signals selectable As explained in the device configuration above, this Single Stereo monaural device is equipped with various functions and multiple 1 sys from 3 sys PORT CPUs / DSPs. In case of malfunctions such as runaway
Non (Only of CPU / DSP, by building the fail safe processing (detect- 1 sys selected output) ing malfunctions and recovering) by function, the load of I2S PORT Max. 3 stereo signal the microcomputer system is reduced and prompt actions Digital 3 sys against malfunctions are possible. output possible The fail processing details are as follows. Max. 2 stereo signal Output DAC PORT 4 ch ・To build the measure of giving notice when a mal- output possible function is detected by a built-in controller (radio / Analog EVR PORT 4 ch Non (dedicated I/O ) audio system and CD system) in a device For the notice of a case that the runaway of CPU / 4.3 Efforts toward integration (CD built-in) DSP is detected from the two blocks of radio / audio We took the following measures as tasks for function- system and CD system, outputting respective error sig- al integration (Radio / Audio / CD). nals of watchdog timer is established. In this system, a) Countermeasure against noise through the measure with notices by terminal output b) Figuring maximum current flow and effort for low and respective runaway notices, prompt action against power consumption mode runaways and countermeasures by block are possible. c) Effort for fail safe ・To build the measure of detecting malfunctions inside a device by built-in controller in a device a) Countermeasure against noise (CPU), self-recovering and giving a message An integrated system is composed of multiple analog Each built-in CPU processes for detecting malfunctions circuits (IF-ADC / DAC for radio, ADC / DAC / EVR for respectively by radio / audio system and CD system. audio, RF / PLL and others for CD-Servo, and system A device itself can recover from some malfunctions. clock forming network such as system PLL), digital cir- The malfunction results and self-recovering can be cuits (block circuits of different operating frequencies), judged every time by the microcomputer system and multiple terminal pins. Aiming for integration, first, through communication, and besides, the action (recov- we optimized the layout (e.g. isolated position for interfer- ering) by the microcomputer system can be added and ing block) for respective analog circuits and terminals changed. (group unit), and then, took the countermeasure for noise Further, by breaking up the judgments respectively, reduction by the method of optimum layout of power sup- such as by system, radio and audio system in a radio / ply (LSI layout design method) and by enhanced power audio system, and by CD playback and compressed supply methods such as enlarging the capacity between audio playback system in a CD system, in this system, power supplies in the LSI. respective recoveries by malfunction content are possi- b) Figuring maximum current flow and effort for low ble. power consumption mode ・To detect malfunctions to command form (communi- By estimating the current flow increment for each cation) block caused by the integration, we read for consistency This device is controlled by a command base and the
FUJITSU TEN TECH. J. NO.33(2009) 6 Development of 09CY Audio Platform
communication (judging the command contents) is Usable sensitivity: 27.5dBu judged in the device. In this system, communication result (normal / malfunction) by command can be checked by the microcomputer system, and recovery by command communication is possible.
5. Newly Developed Function 5 Newly Developed Function 5.1 Prevention of Noise Converging Point Elevation (AM) Fig.8 One Signal Characteristic of New Model (09PF) with Recent years, for vehicle radio antennas, active anten- Antenna Amplifier for Short Pole Combined nas such as glass antenna and short-pole antenna are mainstream, replacing the conventional passive antennas. 5.2 Prevention of Noise Converging Point In accordance with this change, noise level elevation Elevation (FM) under no-signal circumstances became problematic. We saw the opportunity for detecting the repeating Signal strength was obtained by S-meter based on IF signal envelop drops as holes (circled in red); which AGC signals controlled for detection. However, by are caused by random noises, to detect weak electric field preparing AGC for S-meter detection shown in Fig. 6, without any influence by antenna amplifier. Fig. 9 shows low-pass filter can be placed prior to the secondary AGC the image of the hole detection. (for S-meter DC detection). This enabled preventing the noise level from being elevated under no-signal circum- Hole detection stances. FM IF signal The characteristic of the one AM signal through the circuit with antenna amplifier for short pole combined is Threshold as follows. level Fig. 7 shows the graph of the conventional model, and IF level Fig. 8 shows the graph of the new model. Without anten- na amplifier combined, both of the models output signals with the same characteristics.
AM IF signal Time To AM detection Varia ble band filter Fig.9 Detection Image of FM Weak Electric Field
S-meter (DC) signal Threshold B S-meter (DC) signal
Fig.6 Block Circuit of DC Part S-meter Threshold A
Usable sensitivity: 34dBu Hole detection value
Time Hole detection value and detection threshold Fig.10 Relation between Hole Detection Amount and Threshold
Fig.7 One Signal Characteristic of Conventional Model (07PF) with Antenna Amplifier for Short Pole Combined
FUJITSU TEN TECH. J. NO.33(2009) 7 FUJITSU TEN TECHNICAL JOURNAL
As shown in Fig. 10, by setting two threshold levels Hole detection start to the hole detection values, the mute gain is to be con- trolled by following the algorism in Fig. 11. a) When the measured S-DC value is above the threshold Measurement of hole level range (the value is variable by parameter), hole detection value detection is unnecessary. b) When the hole detection value is above the threshold A (the value is variable by parameter), the softmute processing conforming to the hole detection is to be Whether the measured implemented. S-DC is in the range for hole detection or not c) When the hole detection value is under the threshold A, the hole detection result is not to be reflected to the S-mute. The softmute processing conforming to the normal S-DC is to be implemented. Whether the hole By processing as above, we succeeded in preventing detection value meets the threshold condition or not noise level elevation under no-signal circumstances.
5.3 CD sequencer Here is the explanation of the CD part. In this newly Calculation of Normal softmute developed system, to reduce the capacity of software and softmute gain processing based on S-DC to accelerate the processing speed by reducing the load of software, various controls that were processed by soft- ware are now to be processed inside the LSI by Addition to softmute gain sequencer (CPU) packed in a ROM. In the result, we suc- based on S-DC ceeded in reducing CD control program capacity, inte- grating the dedicated CD control microcomputer into product control microcomputer, and operating CD play- Hole detection end back function and RDS function concurrently. Besides, the sequencer packed in a ROM enabled the speed-up of Fig.11 Soft-mute Processing following Hole Detection processing and reduced the time required from disc insertion to playback start.
Microcomputer for product control Microcomputer for product control User operating part User operating part
CD control library CD control library Function mode Function mode control group control group
CD-TEXT Command to CD shock acquisition File analysis Compressed Playback proof control control control audio control control group be issued directly to LSI
Trace FAST Auto- TOC Pause Search Servo control tuning input group
LSI control (inherent to LSI) Driver groupLSI control (inherent to LSI) Driver group
Conventional CD control LSI 09CY LSI CD sequencer CD-TEXT CD shock acquisition File analysis Compressed Playback proof control control control audio control control group
Trace FAST Auto- TOC Pause Search Servo control tuning input group Conventional model (07CY) 09CY Fig.12 Block Diagram of CD Control Firmware
FUJITSU TEN TECH. J. NO.33(2009) 8 Development of 09CY Audio Platform
Table 2 Required Time from Insertion to Replay Start (Single CD) 3dB Gain up Conventional model (s) 09PF (s) 09PF CDDA Disc 10.8 8.6 MP-3 Disc 13.7 9.7
As above, the CD sequencer was built in the LSI. Volt (mV) Playback was However, the various know-hows in performance and stopped at 30 h function for an in-vehicle CD deck are incorporated in the Conventional 12dB Gain up conventional software controlling part. model We have been developing the CD sequencer so as to Hour (h) incorporate the following functions based on our consecu- Fig.14 Comparative Efficacy with Dust Test tive ideal with repeating DR (Design Review) with LSI manufacturers from the beginning of the development. There are various discs in the market, which differ in ・Adding functions so as to make a retry suitable for reflectivity. Our new measure to make gain adjustment the remaining amount of memory for shockproof by 3 dB step method is effective in playback of these function. Preventing noise generation and sound discs in their appropriate conditions, and the new model skips even under bad environment with vibrations. provides higher performance in playback than the con- Taking measure that users are unaware of if any of ventional model. sound skips occur. ・Incorporating function for playback of CCCD (copy 5.5 Quality Improvement by Error Code control CD) and nonstandard CD In-vehicle CD players are used under bad environ- ・Adding function for fail safe ment with outside vibrations and of high / low tempera- ・Enhancing function for retry ture. So, software in CD players incorporates high-power ・Function for output of error code information fail safe that provides recovery function from error status ・Adding AGC function during playback that users are unaware of. With the fail safe function, users can enjoy music as though at home. But in some 5.4 Improving CD Read Performance cases, the fail safe makes the evaluation impossible due to It is inevitable for in-vehicle devices to be used under its recovery function. Finding the way to reduce the bad environment with dust, sand, etc. in some countries cases of NTF (No Trouble Found) and low-frequency and areas such as developing countries. We previously defects became our new task. have taken some measures to prevent outside dusts from In this development, to eradicate NTF and low-fre- entering. However, it is difficult to seal up the in-vehicle quency defects, we built a new function into the software, devices because there are opening around the in-vehicle which plays a role in accumulating information into non- devices due to their slot-in mechanisms, and because the volatile memory for detecting abnormal operations includ- devices require spaces for heat release. These circum- ing phenomena that users are unaware of and to identify stances require a high playback capability even with the cause. dusts inevitably entering into lens. The following information is accumulated into non- In 09PF, by tripling the adjustable width of input volatile memory. gain, we achieved over 20% improvement in the dust ・Abnormal state detected by software resistance of the lens. ・Details that fail safe ran ・Information about disc type / disc playback time 09CY Conventional (address) ・Result per disc adjusted by servo system 0dB 0dB RF amplifier to to ・Temperature information 12dB RF amplifier 3dB step This system is beneficial not just to analyze defective to products, but also to analyze products with no defects by 21dB reliability test such as for durability in operation, test with poor quality media collected from countries, and Signal level lowered Signal level lowered rough road driving test by actual vehicle using all error ⇒ 12dB Gain up ⇒ Gain up per 3dB ⇒ Possible to gain up to codes. This system enables detection of such control Max. 21dB problems that do not come to the surface. We succeeded Fig.13 Variable Amount Expansion of Control Gain in solving all problems in the development step.
FUJITSU TEN TECH. J. NO.33(2009) 9 FUJITSU TEN TECHNICAL JOURNAL
6. Effect 7. Conclusion 6 Effect 7 Conclusion As above, with ensuring the higher radio perfor- In this development, we achieved development of mance and CD performance than the conventional plat- platform with its performance and its cost optimized. form, we achieved cost reduction and space reduction As mentioned in the abstract, various types of digital (20% reduction of LSI chip area) shown in Fig. 15. broadcasting have begun in countries and are becoming a This platform is the current main platform for our major media. CD-tuner products, and is mass-produced for respective We are pursuing the development of platform for destinations for OEM customers. other digital media while aiming for higher radio perfor- mance. In conclusion, we appreciate the efforts of people of device manufactures and departments involved in this system development.
Conventional product Developed product
Radio/audio integrated LSI Package：QFP 208pin Size：28mm×28mm
Reduction in space (LSI chip area: ▲20%)
CD LSI Package：QFP 100pin Size：14mm×14mm Radio/audio/CD integrated LSI Package：QFP 208pin Size：28mm×28mm Fig.15 Comparison of Main Printed Boards in CD-tuners Equipped with Developed Product and Conventional Product
Profiles of External Writer
Akira YOSHIDA Entered the company in 1989. Since then, has engaged in the development of LSI for digital signal processing. Currently in the Custom LSI Division Digital Tuner Development Dept.
Profiles of Writers
Hirokazu MATSUNAGA Teruaki HAYASHI Kiyotaka NAKAGAWA Entered the company in 1984. Since Entered the company in 2001. Since Entered the company in 2001. Since then, has engaged in the development then, has engaged in the design / devel- then, has engaged in the development and design of antennas and tuner mod- opment of audio head units. Currently of tuner LSI. Currently in the Radio ules. Currently in the Engineering in the Audio Media Team, Radio Technology Research & Development Department 2, Fujitsu Ten Technology Technology Research & Development Department, Engineering Development Limited. Department, Engineering Development Division, Common Technology Division, Common Technology Group. Group. Yoshikazu FUJITA Yasuhiro YAMAKAWA Entered the company in 1989. Since Entered the company in 1982. Since then, has engaged in the design / then, has engaged in the development development of CD and DVD decks. of LSI / millimeter-wave devices. Currently the Team Leader of BD Since 2007, has engaged in the devel- Team, Deck Mechanism Engineering opment of tuners. Currently the Department, Engineering Division 2, Team Leader of Design Team 2, CI•OEM Group. Engineering Department 2, Fujitsu Ten Technology Limited.
FUJITSU TEN TECH. J. NO.33(2009) 10 Development of Time Domain Speaker: TD712zMK2
Kazuhiko HAMADA Kiyosei SHIBATA Yoshikuni MIKI Ryousuke SOU Masahito KAWAI
Abstract
FUJITSU TEN has been developing and launching speakers based on Time Domain theory since 2000. This speaker series is highly praised in the market all over the world for its novel design and accurate reproduction of input sounds. Since the release of TD712z as a flagship model in 2004, we have recognized the market needs for higher capa- bility of bass reproduction and easier installation, with expectations of developing and releasing a new model of the series. Our new model 712zMK2 has extended the bass reproduction range to the lower-limit of 35Hz, which used to be 40Hz, by reviewing the speaker unit design and enlarging the enclosure capacity while retaining the ɔ12cm- speaker units. Also, this model has succeeded dramatically in improving the angle adjustment structure: supporting the bot- tom of the enclosure by three spikes and requiring only one pulling screw centrally placed. With these improve- ments, this model was released in February 2009 as a user-friendly flagship model with toolless and hand- adjustable structure.
11 FUJITSU TEN TECHNICAL JOURNAL
1 Introduction1. Introduction a speaker. Here are three principal technologies that we have developed to minimize the unwanted vibrations that Since its release in 2001, our home audio systems, are the major factors in reverberation. ECLIPSE TD series, have been highly praised in world (1) Grand anchor audio magazines and elsewhere. They are now widely A conical metal object called a ground anchor is set in used not only by audiophiles but also by worldwide top the back of the speaker unit, and its inertia mass sup- artists, being used regularly in the top studios throughout presses the reaction of a speaker unit. This structure the world. enables a cone to push air accurately. Specially our TD712zMK2, the successor of TD712z (2) Floating structure released in 2004 as a flagship model, has been much This adopted floating structure eliminates mechanical praised since its release in February 2009. contact between the enclosure and the speaker unit, and This paper explains the current status of FUJITSU suppresses the transmission of vibrations of the speaker TEN home audio products, and goes on to explain the unit to the enclosure. Therefore, the enclosure-specific development background, development concept and vibrations are reduced and this leads to the minimization development results. of the speaker-specific unwanted sound. (3) Eggshell enclosure 22. CurrentCurrent Status Status of ECLIPSEof ECLIPSE Home Home Audio Audio Speakers Speakers The eggshell enclosure achieves the suppression of standing waves due to no parallel surface in the enclo- 2.1 Concept Underlying Sound Creation sure. The concept underlying ECLIPSE home audio speak- Besides, the shape makes the baffle surfaces round ers is to reproduce accurate sounds by reproducing accu- with no corners and leads to the suppression of the dif- rate waveforms. fraction waves that are normally generated in the process Most conventional speakers have been creating of sound (spherical wave) expansion. sounds with the aim of high-powered / wide range sounds by focusing on frequency response, emphasizing 3. Eggshell enclosure 1. Grand anchor the ability to reproduce flat sound from low frequencies to high frequencies with minimum distortion. On the other hand, at ECLIPSE, we have been creat- ing accuracy- / nature-conscious sounds based on the time domain theory, which focuses on the accurate repro- duction of the movement of air from sound generation to sound loss.
2. Floating structure Ideal ECLIPSE region TD series Fig.2 Internal Structure of TD712zMK2
2.3 Sound Characteristics of ECLIPSE TD Series The sounds of ECLIPSE TD series which are pursu- Conventional ing the reproduction of the most accurate waveforms pos- Accurate / natural Hi-Fi speaker sible, have the following three major characteristics. (1) Increased sound clarity (Focusing on time domain) (Even minute sounds are heard without masking unwanted sound) High power / wide range (2) Faster and tighter reproduction of sound (Focusing on frequency response) (The rising and falling of sound reproduction are both Fig.1 Concept Map in Sound Creation quick) (3) Improved space reproduction 2.2 Unique Technology Employed for ECLIPSE (The listeners are now less aware of the existence of TD Series speakers, which means that what they hear comes One of the factors causing an adverse effect on the from their surrounding space) accurate reproduction of waveforms is "reverberation" of
FUJITSU TEN TECH. J. NO.33(2009) 12 Development of Time Domain Speaker: TD712zMK2
3 Development3. Development Background 4 Development4. Development of of TD712zMK2TD712zMK2
3.1 Evaluation / Task of Previous Model (TD712z) To achieve the concept mentioned in Section 3.2, we The previous model aimed the ideal region based on considered the following points to develop TD712zMK2. the above-mentioned sound creation concept. However, 1. Lowering limit of bass reproduction we received feedback that the model has less power in 2. Keeping or improving impulse response bass than the conventional Hi-Fi speakers. 3. Simplifying angle adjustment Also, we received many requests for a model with a Here are the details of this development. simpler structure for setting to the appropriate listening points by adjusting the enclosure angle shown in Fig. 3: 4.1 Lowering Limit of Bass Reproduction the previous model required tools for an adjustment at The previous model was developed aiming at the four points. ideal region based on the above-mentioned sound creation concept. However, since the previous model had less power in bass than conventional Hi-Fi speakers, the new model was demanded to have a lower limit of bass repro- duction. Generally speaking, in order to lower the limit of bass reproduction, one method for making the lowest reso- nance frequency (fo) lower is by adopting a larger-mass vibration system of a speaker unit. However, as a trade-off problem, adopting the larger- Points for mass vibration system will make the performances of angle adjustment transient responses worse and especially sound rising worse. Fig.3 Angle Adjustment Function of TD712z In the development of TD712zMK2, instead of adopt- ing the larger-mass vibration system, we decided to take 3.2 Development Concept a measure to reduce stiffness (stiffness in cone's vibra- To solve the tasks mentioned above and to go close to tion). This requires the influence of back pressure inside the ideal region as a flagship model, we decided to change a speaker box to the cone's vibration to be small. So, we the model. The development concepts for the new model adopted an approx. 1.5 times capacity (compared to the are as follows. previous model) with design in mind so as to lower the (1) As a flagship model, to keep the level of or upgrade limit of bass reproduction. With this enlarged enclosure the level from the previous model in accuracy of capacity, the (design) constants of the speaker unit, such waveforms, and to aim at wider range as fo, were obtained by simulating the bass reproduction (2) To improve the installation workability by simplify- with setting the target to approx. 5Hz lowered of bass ing the angle adjustment reproduction limit. (Fig. 5) As for the design, the several models released since 2005 (TD510, TD508Ⅱ, TD307Ⅱ) are highly praised in design because of their organic enclosure forms. Thus, we decided to adopt the design in this model across the Simulation series at the time of model change.
Previous model TD712z
Fig.5 Simulation of Bass Reproduction
The TD712zMK2 developed as above enables the TD510 TD508Ⅱ TD307Ⅱ reproduction of richer and higher quality bass above the lower limit of 35Hz, which used to be 40Hz. The follow- Fig.4 TD Speaker Series Released Since 2005 ing figure (Fig. 6) shows the comparison of frequency responses, indicating our success to lower the limit of bass reproduction.
FUJITSU TEN TECH. J. NO.33(2009) 13 FUJITSU TEN TECHNICAL JOURNAL
TD712zMK2 Conventional model
TD712z (Previous model)
Fig.7 Conventional Magnetic Circuit / Grand Anchor Structure
Fig.6 Comparison of Frequency Response Nonmagnetic spacer (aluminum) 4.2 Improving Impulse Response Time domain speakers are developed with focusing on the accurate reproduction of time waveforms of sounds, in other words, impulse response. The points and the method for improving impulse response in the development of TD712zMK2 are as fol- Nonmagnetic Table 1 lows. ( ) screw (SUS) Table 1 Point and Method for Improving Impulse Response Fig.8 Developed Magnetic Circuit / Grand Anchor Structure Point Method Rising Magnetic circuit enhancement 4.2.2 Reduction of Unwanted Resonance by Enhancing Internal Falling Reduction of unwanted resonance by Loss of Cone enhancing internal loss of cone The cone in the previous models comprised only one Accuracy in Improvement of symmetry property of waveform forward-and-backward amplitude in layer of fiberglass, however, in this development we supporting system affixed cotton: high internal loss material, to the rear of the fiberglass to make a dual structure (approx. 25 times 4.2.1 Magnetic Circuit Enhancement of internal loss). Fig. 9 and Fig. 10 show respective For the grand anchor of this development, we adopt- impulse responses of these two cones. ed the same inexpensive and easily-worked iron material With these figures, the dual layer cone is much supe- as the conventional products. However, when the iron rior in reducing the unwanted resonances. material is installed to the back side of the speaker unit directly as in Fig. 7, magnetic flux leakage occurs, which leads to the decrease of the cone's driving force and exerts a bad influence in the rising characteristics of impulse responses. In this development, we adopted the structure that minimizes the magnetic flux leakage by inserting a nonmagnetic spacer between a speaker unit and a grand anchor. Further, we took some other detailed measures such as setting the speaker unit and the grand anchor apart with an optimized distance, and adopting a nonmagnetic stainless-steel screw to fix the grand anchor. This development resulted in improve- Fig.9 Impulse Response (cone: one layer of fiberglass) ment of magnetic circuit efficiency; approx. 10% up of magnetic flux density compared to the conventional den- sity. (1.1T→1.25T)
FUJITSU TEN TECH. J. NO.33(2009) 14 Development of Time Domain Speaker: TD712zMK2
High symmetry shape 3
Cross-section shape image
Inner periphery Outer periphery
Fig.12 Cross-section Shape of Spider
Simulation of up-and-down amplitude volume
Fig.10 Impulse Response (cone: dual layer of fiberglass and cotton)
4.2.3 Improvement of Symmetry Property of Forward-and- backward Amplitude in Supporting System To reproduce time waveforms of sounds accurately, the symmetry property of forward-and-backward ampli- tude to input signals in the supporting system of speaker cone must be improved. As for the surround, conventional speakers use a gen- eral rolled surround, which has a tendency to move for- Feature ward smoothly but backward jerkily. This simulation is excellent in the symmetry property of up-and-down amplitude up to 14mm each range and the amplitude volume with 1kgf load or less. In this development, we adopted a corrugated sur- Fig.13 Simulation of Spider Amplitude round that is superior in amplitude linearity. Since the appropriate amplitude amount for the corrugated sur- As mentioned above, the optimized shapes for the round has to be ensured, we obtained the optimized surround and the spider were obtained by simulation. shape of the surround by simulation to ensure the sym- Making a test sample with these shapes, we verified the metry property of forward-and-backward amplitude even effect. Fig. 14 and Fig. 15 show the forward-and-back- when there are significant power outputs. (Fig. 11) ward amplitude characteristics of our conventional prod- As for the spider, we also obtained by simulation the ucts and developed products. With these figures, the con- corrugated shape superior in symmetry property of for- ventional products have gaps in amounts between for- ward-and-backward amplitude. (Fig. 12) Fig. 13 is a simu- ward and backward amplitude, and besides, have sur- lation result. It shows that the forward amplitude (in round resonances up and down in a graph line depending gray) overlaps the backward amplitude (in black), and it on the frequency. By contrast, the developed products indicates symmetry property is ensured. have the matched amounts in forward-and-backward amplitude characteristics and have the surround reso- nances modified smoothly in a line.
10.31 mm
Forward amplitude The surround has a symmetry property of back-and-forth amplitude. Fig.14 Amplitude Characteristics of Conventional Product
10.24 mm
Backward amplitude Fig.11 Simulation of Surround Amplitude
Fig.15 Amplitude Characteristics of Developed Product
FUJITSU TEN TECH. J. NO.33(2009) 15 FUJITSU TEN TECHNICAL JOURNAL
4.3 Angle Adjustment Structure of Speaker 4.4 Specifications To realize the high-quality space reproduction, one of TD712zMK2's appearance (Fig. 18) and major specifi- the TD speaker characteristics, speakers must be located cations are as follows. correctly to face the listening point by adjusting their Table 2 Specifications vertical angle (upward and downward) and horizontal direction (right and left). As for the direction, speakers TD712z TD712zMK2 can be positioned easily to face the listening point. (Previous model) However, as for the angle, speakers need to be adjustable Regular Short Regular Short in installation in accordance with the setting location or Speaker unit the circumstances of the room. φ12 φ12 (cm) TD712z (previous model) was adjustable in the verti- cal angle in the range from 0°to +12°, enabling adjust- Impedance (Ω) 6 6 ment in angle in accordance with the height from the Frequency 40 to 20k 35 to 26k floor to the listening point. However, the structure Response (Hz) required operations to adjust three pulling screws and Maximum W347× W347× W347× W347× one pushing screw (four screws in total) by extracting dimensions H988× H600× H1000× H611× and inserting a tool (hex wrench) repeatedly. This com- (mm) D384 D384 D431 D431 plicated angle adjustment had to be simplified. (Fig. 16) Weight (kg) 25 18 25 18
Pulling screw
Spike Pushing screw
Fig.16 Angle Adjustment Structure of TD712z (Previous model)
So, in the development of TD712zMK2, we introduced the structure which sets three spikes above a post so that the speaker box can be fixed by one pulling screw on these spikes. The speaker is fixed on the spikes with the pulling screw that is screwed into the upper side of the post and that is hooked at the screw head to the ditch of the enclosure's bottom. Besides, since it is possi- ble to grab the outer part of the φ60mm disc shape at TD712zMK2（Regular） the periphery of the pulling screw axis, tools are not nec- essary for screwing the speaker. With this structure, we succeeded in reducing the required time for angle adjust- ment to approx. 10s, which used to take couple of min- utes for one screw. (Fig. 17)
TD712zMK2（Short） Spike Pulling screw Fig.18 Appearance
Fig.17 Angle Adjustment Structure of TD712zMK2
FUJITSU TEN TECH. J. NO.33(2009) 16 Development of Time Domain Speaker: TD712zMK2
5 Conclusion5. Conclusion
This paper has described the development for TD712zMK2, the newly remodeled flagship model of this series, while mentioning our product technology based on Time Domain theory. Since the first release in 2001 of ECLIPSE home audios, we have developed and released various models based on a unified concept of underlying sound creation. This year is the ninth year since the first release. Users such as musicians and recording engineers, who sympa- thize with our wish to carry artist's sensibility by repro- ducing accurate sounds, are increasing yearly. This model is a flagship model. We keep trying to develop superior products by improving our technology. We sincerely appreciate the efforts of everyone con- cerned inside and outside the company for this develop- ment.
Profiles of Writers
Kazuhiko HAMADA Kiyosei SHIBATA Yoshikuni MIKI Entered the company in 1986. Since Entered the company in 1998. Since Entered the company in 1997. Since then, has engaged in the development of then, has engaged in the development then, has engaged in the development mechanisms for car audio systems, devel- of OEM car speakers, sound systems, of millimeter-wave radar system, and opment of music software, and then and then car / home commercial then development of speakers for development of speakers for home-use speakers. Currently in the Acoustic home-use since 2001. Currently in since 2001. Currently in the Acoustic Engineering Department, Engineering the Acoustic Engineering Department, Research & Development Department, Division 1, CI•OEM Group. Engineering Division 1, CI•OEM Engineering Division 1, CI•OEM Group. Group. Ryousuke SOU Masahito KAWAI Entered the company in 2005. Since Entered the company in 1984. Since then, has engaged in the development then, has engaged in the development of car / home commercial speakers. design of audio systems. Currently Currently in the Acoustic Engineering the Team Leader of the QCD Department, Engineering Division 1, Planning Department, Engineering CI•OEM Group. Division 1, CI•OEM Group.
FUJITSU TEN TECH. J. NO.33(2009) 17 Development of FUJITSU TEN's First SD Memory Card Navigation System, AVN Lite, and its Linkage Service
Kenichirou MASE Takeshi FUKADA Kenji UEMATSU Teru SAWADA
Abstract
Since 1997, when FUJITSU TEN released an all-in-one unit consisting of Audio, Visual, and Navigation ahead of the industry, "AVN" has become synonymous with in-dash and all-in-one navigation system. On the other hand, recently, on-dash navigation system (PND: Portable Navigation Device) that provides only basic func- tion of navigation system at a low price has been penetrating the market and the market needs for low-end models have been growing. With the motto, "secure," "easy," and "fun," FUJITSU TEN therefore planned and developed "AVN Lite" that offers the necessary and sufficient functions at a reasonable price including the path search and audio assist which are the basic functions of the car navigation system, storing the map data in the SD card, and also including CD, One-Seg, etc. as AV functions, and we released the "AVN Lite" in October 2008. The navigation system using the method that stores the map data in the SD card is the first product of this kind for FUJITSU TEN. We introduce the problems and solutions for adoption of the SD card as the storage media, and various services linked to this product in this article.
18 Development of FUJITSU TEN's First SD Memory Card Navigation System, AVN Lite, and its Linkage Service
1 Introduction1. Introduction ・Carefully select and arrange four basic buttons of "pre- sent location," volume, etc. at the bottom of the naviga- "AVN," released in 1997 as ECLIPSE brand, has tion device. become the standard configuration of a car navigation system due to the emergence of other manufacturers and its adoption by automotive manufacturers. Since then, we have advanced the development seek- ing the high added value consistently with the technologi- cal evolution, but on the other hand, many users require the car navigation system that provides its primary func- tions at a reasonable price as typified by PND. Based on these needs from the market, we have planned and developed the reasonably-priced "AVNLite" Fig.2 Double-screen Display that incorporates the necessary and sufficient functions including path search and audio assist which are the 3) Fun basic function of the car navigation system, and also ・With 7-inch wide screen, and receiving One-seg broad- including CD and One-seg and others as AV functions casting with realistic sensation in a car. with the motto, "secure," "easy," and "fun" in order to have ・Play favorite music anytime and anywhere with CD the AVN used by a wide range of people. playback and iPod connection. We introduce the development concept, technology, ・Obtain the hot place and vacancy information of park- and implementation method in this article. ing with KEITAI-Link service using a mobile phone.
2. IntentionIntention of of Development Development and and Summary Summary of ProductProduct Also, since the beginning, we have made product 2 planning considering the use for other than the mass After comparing the features of a navigation system merchandiser market. We incorporated the functions for with full specifications as typified by "AVN" with those of the rental car (AVN118MRE) and corporations PND, we considered the functions required for new prod- (AVN118MCO) into the product specifications and devel- ucts category, "AVNLite," and formulated the product oped WEB services and PC applications linked to the concept with keywords, "secure," "easy," "fun." product.
3. Development of New Services 3 Development of New Services This product intends to differentiate the functions for each destination to deliver by WEB services and PC applications linked to AVNLite. Especially for rental car (AVN118MRE) and corporations (AVN118MCO), dedicat- ed PC applications are sold separately, and searching data and menu indication of AVNLite can be customized (Fig. 3). The dedicated PC application (APR118MRE) for Fig.1 AVN Lite AVNLite for car rental companies allows individual cus- tomization specialized to the needs of car rental compa- 1) Secure nies, including the individual addition/edit/deletion of the ・Acquire the vehicle position with vehicle speed sensor data of service offices of car rental companies, rewriting and gyroscope as well as GPS. Allow safe driving on the opening image, and so forth. Also the dedicated PC an unknown road. application (APR118MCO) for AVNLite for corporations ・Support parking and putting a car into the garage with allows display of the list of customers to visit on the navi- connection of rear view camera sold separately. gation screen by creating the data of customers to visit ・Give audio assist for the distance and intersection name on PC and writing it on SD card, in order to support the at which to turn. visiting activities as typified by route sales with naviga- 2) Easy tion system. ・Give easy-to-see double-screen display and display each In this article, we explain the various service and operation screen of navigation and AV. applications linked to AVNLite individually. ・Widely indicate the button to be often used. Ensure and easy destination search.
FUJITSU TEN TECH. J. NO.33(2009) 19 FUJITSU TEN TECHNICAL JOURNAL
①Commercial model (AVN118M)
Service for commercial model ・Paid map delivery
②Rental car model (AVN118MRE) ・Data of service offices Service for ・Opening image rental car model ・Data of terms of service ・Data of emergency phone number ③Corporation model (AVN118MCO)
Service for ・Support for visiting corporation model ・Support for urgent customer
Fig.3 Linkage Service for Each Destination Fig.4 MyPage
3.1 Service for Commercial Model (AVN118M) 3.2 Service for Rental Car Model (AVN118MRE) We have offered the information providing service AVN118MRE is a special model providing various (MyPage) using WEB for users of commercial model functions specialized in utilization of navigation system in (AVN118M) since October 2008. a rental car. This model allows customizing the functions AVN118M users can obtain their own user ID and by using the dedicated application (APR118MRE: sold password to log into MyPage by completing the user reg- separately). The customizable functions with the dedicat- istration in the page of "users support" in Fujitsu Ten's ed application are as follow: website after purchasing the product. Once users obtain ①Editing data of service offices their own user ID and password, they can log into ②Editing data of facilities MyPage and receive various service (Fig. 4). ③Editing data of notification In MyPage, the users can view the list of their own ④Editing data of terms of service products and information for each product. And the ⑤Editing opening image users of AVN118M can use "compensation version map In this article, we introduce these functions respec- data download service" in MyPage. In AVN118M, vari- tively. ous data necessary for navigation including map data, ①Editing data of service offices POI (Point of Interest and facility information) etc. is The car rental companies strongly required the sys- stored in SD card. The SD card is a writable media, dif- tem not to include other car rental companies in the fering from the existing map media, so it has a risk of search target when customers search the location of car damage or losing data. Even if perchance the data stored rental company with a navigation system. This model in the SD card is damaged, the qualified users registered satisfied the requirement by creating data with dedicated in MyPage can download the same data as the data application in advance. In advance, a category "XX Car before being damaged from the compensation version Rental Company" using the application is created, and in map data download service, and can restore the data. this category, the data of service office of XX Car Rental The data stored in the SD card is composed of vari- Company is stored. Only the service offices of XX Car ous format file such as map, POI, audio etc. In this ser- Rental Company will display if the category "XX Car vice, these multiple data are combined into one data, and Rental Company" is selected on the navigation screen. this data is compressed to shorten the downloading time This function is also welcomed because the company can and to restore at one time. edit the data flexibly by itself if a service office is newly This service allows a rapid response to the problem of established or consolidated. data loss concerned when the SD card is used as a stor- ②Editing data of facilities age media of navigation system. As in the editing function of service office data men- We will extend MyPage as a WEB service providing tioned above, the car rental companies can upload the "convenience" and "secure" to our users. local and seasonal information not originally stored in the navigation system by creating the original data of sight- seeing areas or events. Especially in most service offices in Hokkaido or Okinawa where the utilization ratio for sightseeing is high, the companies create the original sightseeing brochures and list the address or map code of destinations on these brochures. The rental car users
FUJITSU TEN TECH. J. NO.33(2009) 20 Development of FUJITSU TEN's First SD Memory Card Navigation System, AVN Lite, and its Linkage Service
operate the navigation system and set the destinations porations. This model can customize its functions using with these brochures. In this case, there are following the dedicated application (APR118MCO: sold separately). problems: for safety reasons, it is not desirable that the To support the sales operations mostly, AVN118MCO users operate the navigation system while looking at the incorporates the function that loads the data of customers brochure, and operation of address search or map code to visit (created at the office in advance) into the naviga- search becomes complicated. However, use of this edit- tion system and sets the destinations easily, along with a ing function can provide the latest information on the function that transfers the data of urgent customer to navigation screen, so the users of rental car can receive visit from the office to the mobile phone of sales person "convenience" and "security." who sends the data to the navigation system using e-mail of the mobile phone easily. A dedicated application must Screen of setting support tool be installed in PC of the office to utilize these functions. Here are the main functions of this application.
Support for creating Notification of urgent customer notification e-mail of urgent customer
Support for creating list of customers to visit KEITAI-Link Functions in offices RAKURAKU visiting Support with functions in offices Support with functions in offices support tool (tool) for visiting by commercial (tool) for notification of urgent vehicle and navigation system customer to sales person Only insert SD card Only send notification before visiting e-mail to navigation system Screen of search from Fig.6 Service for Corporations additional data Fig.5 Application Screen for Rental Car ①Support for creating list of customer to visit The destinations can be set easily based on the cus- ③Editing data of notification tomers list on the navigation screen by editing the data of Most car rental companies give information (emer- customers to visit with PC of the office in advance and gency phone number for occurrence of accident, or cau- storing the data into the SD card. In general, most cus- tions for use of rental car) that they should deliver to tomers lists are compiled as a database. This application their users with printed paper, or affix a caution seal in can load the data of customers to visit in CSV format the car. This model can display such notification on the based on the database of the customers list. navigation screen with special button, and can customize The location information of customers to visit is set the message to be displayed using the application. by converting their addresses to latitude/longitude. ④Editing data of terms of service However, if sales person moves by car, the location of lati- The rental car users confirm the terms of service tude/longitude that is set based on the address of cus- with papers in advance. This model displays the terms of tomer to visit is not always the destination to visit. For service on the navigation screen when the navigation sys- example, in the delivery business, the place of unloading tem is started and encourages users to reconfirm them. may be the place of delivery designated by the customer, The terms of service can be changed using the applica- or the parking area may be located away from the cus- tion. tomer's office. In such cases, the center location on the ⑤Editing opening image map specified by latitude/longitude, converted from The car rental companies strongly require that the address, is changed to the given locations mentioned logo of their own companies is indicated on the opening above to allow guiding to the most appropriate place for image when the navigation system is started. This appli- each customer. cation prepares the opening image in advance and can In addition, the data of customers to visit stored into change the indication freely, checking the indication the SD card includes the customer's information, so the image of the navigation screen using a PC screen. data is written into the SD card after it is encrypted on the application. Therefore, even if the SD card is lost, the 3.3 Service for Corporation Model (AVN118MCO) information leakage can be prevented because the con- AVN118MCO is a special model incorporating the tents of the SD card cannot be accessed. functions specialized in use of navigation system for cor-
FUJITSU TEN TECH. J. NO.33(2009) 21 FUJITSU TEN TECHNICAL JOURNAL
(2) Deterioration of the SD memory caused by file Addition/editing of Batch addition of customers to visit access customers to visit In this article, we introduce the solutions. Customer to visit 4.1 Improvement of read speed of map data
CSV file Encrypted file The navigation software of AVNLite is based on the External loading of list specification that loads the map data from the existing Confirmation/editing of Protection of data including name, kana, by encryption address, phone number, etc location of customer CD or DVD. CD and DVD are a read-only media and not to visit on map rewritable, so the storage location for each data is fixed Fig.7 Application Screen for Corporations when the media is created. Therefore, the data is direct- ly accessed to load the map data at high speed and to dis- ②Support for creating notification e-mail of urgent cus- play the map without the file system. Also for map for- tomers mat, the data reference location on CD or DVD is written For urgent customers to visit, this application can as an absolute storage location on these media, using such send an e-mail of their data to the mobile phone of a sales characteristics. person. CD and DVD (without file system) When the data of urgent customers to visit is created 1. Obtain the initial address of each at the office using this application, the code, which data from management area of file includes the latitude/longitude information of customers, system. to link to the mobile phone application using the KEITAI- 2. Based on that, access the map data. Link service, which is adopted in our commercial model, (Map data position is automatically produced. =file initial address Then an e-mail text including the produced code is +position from top-of-file sent to the mobile phone of the sales person, using an e- Fig.8 Access to CD and DVD mail function of client PC. When the mobile phone receives the notification e- However, if the map data is stored in the rewritable mail of urgent customers produced by this application, SD card, a delay in reading occurs since the file system the dedicated application to link to the navigation system must be interposed to access when the map data is is started by clicking the link, "send to navigation sys- loaded. tem," shown in the e-mail text (if download is not done in This is because the SD card is a writable and erasable advance, transferred to the page for download). media, and the whole media is controlled for each block of Next, the screen page of AVN118MCO is transferred a certain size by file system to use the media effectively. to the screen page of "destination reception screen" of When the file is written, unused block is searched and a KEITAI-Link so that the urgent customer to visit shown certain number of blocks are selected depending on the in the e-mail text can be set as a destination easily, using file size to be written. The selected blocks do not need the infrared function of the mobile phone. to be adjacent to each other, and it does not matter even In the past, the destination of urgent customer to visit if the blocks are positioned discretely since the file sys- was directed by an e-mail or telephone. However it was tem controls the access of file. The media can be used difficult to direct the specific location accurately, so the effectively by writing the file in the nonadjacent block. system for information transmission of precise location For example, when the file with size of ten blocks is was required. This function solves such existing prob- written, ten empty blocks are required. If the file can- lems by using our KEITAI-Link service developed for not be written in the nonadjacent block, the file with size our commercial model, and contributes to efficient corpo- of ten blocks can be written only when ten empty blocks rate business. are aligned sequentially, even if the whole media has capacity. Terminal4. Terminal Development Development However, this control method has a disadvantage of 4 occurrence of file fragment. If writing/erasing of file with AVNLite displays the navigation map by loading the various size is repeated, the written file is divided finely map data from the SD card. The main problems for and recorded. In this case, frequent access to the man- adoption of the SD card as the storage media are as fol- agement area of file system is required, so this causes the lows: deterioration of access speed. This is one of the reasons (1) Speed deterioration of reading map data caused by that the speed of HDD used for a long time in Windows file system and others slows down. Therefore, periodical use of a defragmentation tool is recommended.
FUJITSU TEN TECH. J. NO.33(2009) 22 Development of FUJITSU TEN's First SD Memory Card Navigation System, AVN Lite, and its Linkage Service
1. File "A", "B", and "C" are written. SD card of map data to guarantee the writing without AAABBCCC fragmentation as data storage specification. We received their great cooperation, and would like to take this oppor- tunity to thank them. (3) Operation of SD card of map data created by users AVNLite adopts the function of digital rights manage- 2. File "B" is deleted. (Blocks become empty.) ment (DRM) to avoid illegal copying of map data. The SD AAA CCC card of map data illegally copied by users is not authenti- cated by DRM before AVNLite loads the map data. Therefore, even if the SD card of map data includes the fragmentation by illegal operation of users, AVNLite does not load the map data to prevent defects. 3. File "D" is written. (Written discretely) (4) Map data update by compensation version map data AAADDCCCDD download service In the compensation version map data download ser- vice mentioned in 3.1, the users can download the map data from WEB server and update the map data stored Fig.9 Fragment in the SD card. We introduce the solutions for fragmen- tation with this service in 5.2. This problem of defragmentation occurs in AVNLite too, adopting the SD card as a storage media of map data. 4.2 Guarantee of SD card's life Flash memory generally used for the SD card has the SD card (with file system) limitation of number of writings, and the more the num- 1. Obtain the initial address of each data ber of writings, the more errors. from management area of file system. AVNLite does not have function of data writing into SD card 2. Obtain the data position following the SD card, so no problem occurs. However, the flash the link on file system (Link is provided one for one block. memory has limitations of number of readings as well. If the data, which is positioned 500 The more number of readings, the more errors. blocks ahead from the top-of-file, is Therefore, the system to reduce the number of data read- accessed, 500 accesses to file system ings of the SD card is required. is required.) Fig.10 Access to SD Card (1) Reduction of access to file system management area (FAT) We considered the method to access data without file In data access processing to the SD card, FAT is system, assuming the map data file continuity as with CD accessed most often, and the number of accesses to FAT and DVD. On that premise, only one access to file sys- needs to be reduced. As mentioned in 4.1, AVNLite tem is required on start-up. However, this cannot be achieved the access without file system, and it requires assumed in the normal SD card, so we considered as fol- only one access to FAT when it is started. After that, lows: the whole map data can be accessed freely by using the copy of FAT cached in the internal memory. This sys- (1) System of file access tem reduced the number of accesses to FAT. When the map data is referred with the SD card or (2) Optimization of holding method of internal memory HDD, it is impossible to specify the reference location in management area of map data with the absolute storage location of media. At first, the The map data also has the management area, and file is specified, and then the position of top-of-file is speci- based on this data, individual map data is accessed. fied. However, there are multiple management areas, and Therefore, we incorporated the system that converts some of them are not necessary. Therefore, assuming the the data reference inside the file into the absolute location usage scene of AVNLite, we set priorities based on the inside the media without the file system on the assump- use frequency of each management data and reduced the tion that the SD card of map data has no fragmentation. number of accesses. (2) Acquisition of SD card of map data without frag- mentation To meet the condition that the SD card of map data has no fragmentation, we needed to ask our supplier of
FUJITSU TEN TECH. J. NO.33(2009) 23 FUJITSU TEN TECHNICAL JOURNAL
Table 1 Improvement of the SD Access Number
SceneAnnual Frequency Destination setting Neighboring The number of The number of mileage facility accesses before accesses after searching countermeasure countermeasure (for one travel) (for one travel) Commuting 15,600 km 5days/week 2 times (company, home) None 375 12 2 times (Gas station, Leisure 2days/week 5 times (Address, phone number) 2,094 15 convenience stre) Business 39,000 km 5days/week 12 times (Address, phone number) None 3,034 28
(3) Optimization of cache medium speed [effective speed: around 4Mbps]) that is There is a method, called cache, to reduce the num- currently mainstream. ber of accesses to the external media. The data loaded We therefore considered the most appropriate com- once is held in the memory inside the AVNLite, and from pression format for map file and shortened the download the second time, the data held in the internal memory is time (within two hours). accessed so as to reduce the number of accesses to the external media. However, the capacity of internal memo- 5.1.1 Selection of most appropriate compression format ry is limited, so how long the data is held needs to be The representative algorithms used for the file com- decided in order of priority based on the data characteris- pression are shown in Table 2. tics to be loaded. We performed such optimization. Table 2 List of Compression Technology Using three methods mentioned above, we reduced the number of accesses to the SD card. Further we mea- Compression Outline Developer sured the number of accesses in three usage scenes of format World standard com- the AVNLite (Table 1) and confirmed the improvement ZIP Phillip W. Kats pression format effect. Windows standard Microsoft CAB compression format Corporation
Application5. Application Development Development Compression format Jean-Loup Gailly 5 GZ used in UNIX and Mark Adler
We explain the main technology (map file compres- Compression format BZ2 Julian Seward sion, map file copy, POI data encryption, and map code generally used in UNIX conversion) that is adopted in the development of map Standard compression LZH Haruyasu Yoshizaki format in 1990s in Japan data download service for after market (hereinafter, Compression format referred to as map DL) and dedicated application for DGCA Shinichi Tsuruta strong in text data rental car model and corporations model (hereinafter Compression format referred to as support application). RAR Eugene Roshal strong in multimedia
Log into STEP １ STEP ２ STEP ３ STEP ４ Compression format capable Finish MyPage Map DL File SD formatting 7z Igor Pavlov decompression SD writing of high compression than ZIP Browser Special PC application (map installer) We compressed the actual map (LOADING.KWI) and Fig.11 Flow of Map DL Service for Mass Merchandiser Market compared the compression characteristics.
Table 3 Experimental Result of Each Compression Technology Map application Customized application for corporations [Shobunsha] Com- Com- Before After Com- Compres- Application Encryption module pression pression compres- compres- pression sion (C#) (C++) perfor- format sion sion late software mance (Megabyte) (Megabyte) (%) .NET Framework High 7z 57.32 24.99 44% Lhaz
OS(Windows XP/Vista) RAR 57.32 27.62 48% WinRAR DGCA 57.32 28.73 50% Explzh Fig.12 Application Configuration for Corporations BZ2 57.32 29.91 52% Lhaz LZH 57.32 30.85 54% Explzh 5.1 Map file compression CAB 57.32 31.34 55% Lhaz The file size of map data (version 2008) for commer- ZIP 57.32 31.36 55% Explzh cial model is 3.5GB or more, and it takes two hours or Low GZ 57.32 31.42 55% Explzh more to download it via the communication line (ADSL /
FUJITSU TEN TECH. J. NO.33(2009) 24 Development of FUJITSU TEN's First SD Memory Card Navigation System, AVN Lite, and its Linkage Service
Table 4 Result of Commercial Map (version 2008) Compression Therefore, in the support application (RAKURAKU Compres- Before After Compres- Compres- visiting support tool), the module to encrypt the POI data sion compression compression sion late sion is invoked from the application via .NET Framework. format (Gigabyte) (Gigabyte) (%) software 7z 3.55 1.87 53% Explzh Customized application for corporations Table 5 Observation Environment Application Encryption module User environment (C#) (C++) OS Windows Vista CPU Intel Celeron M 523（933MHz） .NET Framework Memory 2GB ADSL/medium speed Fig.14 Image of Invoking Module for Encryption Communication line （effective speed： 4Mbps） Thus by modularizing the encryption process, the Table 6 Observation Result application is not affected, even if the encryption system is changed. Down- File decom- SD format- SD writing load pression ting Total ◆ Category ◆CSV file Specify Tokyo Datum for latitude/longitude. Creation of POI 63 min 13 min, 20 sec 7 min, 10 sec 7 min, 50 sec 1.5 hrs No Category Icon category Name Kana Prefecture・・・ Category Customer ○○Pharmaceutical Co.,Ltd ・・・ Tokyo ・・・ Customer to visit A 1 to visit A A.bmp Customer △△Pharmaceutical Co.,Ltd ・・・ Kanagawa・・・ Customer to visit B 2 to visit B B.bmp □□Pharmaceutical Co.,Ltd ・・・ Saitama・・・ Customer to visit C We found that the download time of map data that is 3 Customer C.bmp to visit C ●●Pharmaceutical Co.,Ltd ・・・ Tokyo ・・・ Customer to visit A compressed by "7z" is reduced to 1.5 hours, and achieved ▲▲Pharmaceutical Co.,Ltd ・・・ Kanagawa・・・ Customer to visit B Name ・・・ Category Customer to POI output ○○Pharmaceutical Co.,Ltd ・・・ visit A Data loading the initial targeted value. Customer to processing ●●Pharmaceutical Co.,Ltd ・・・ visit A ◆POI data Name ・・・ Category Customer to △△Pharmaceutical Co.,Ltd ・・・ visit B Customer to Data encryption SD writing 5.2 Map file copy ▲▲Pharmaceutical Co.,Ltd ・・・ visit B Name ・・・ Category (EISERV.POI) Customer to As mentioned in 4.2, AVNLite does not refer to FAT □□Pharmaceutical Co.,Ltd ・・・ visit C POI data is grouped by category and is indicated in when loading map file from the SD card. the order of creating on navigation system (AVN). Therefore, the map file needs to be allocated in the Fig.15 Flow of POI Data Creation continuous area physically by each individual file. However, in general PC (OS: Windows), the allocation in continuous area is not guaranteed due to the file cache 5.4 Map code conversion when the multiple files are copied concurrently. The preemptive notification e-mail service for corpo- Hence, we adopted defragmentation method in the ration model requires the system to convert the position application for map file restoration (map installer) to real- information (latitude/longitude) to the map code to link to locate the fragmented file in the continuous area and to the existing service (KEITAI-Link). guarantee the continuity of individual file.
SD SD Open server RAKURAKU Latitude/ Latitude/ visiting longitude longitude support tool Internet Reallocation Map code
Invoking existing service
Fig.13 Image of Defragmentation
By this map installer, the map data that the car navi- Fig.16 Flow of Preemptive Notification Email Service gation system can read from the SD card correctly is cre- ated, and map data download service with WEB is Hence, we established the interface (assist) to invoke achieved. the map code conversion library on Fujitsu Ten's open server and add it to the existing WEB application to real- 5.3 POI data encryption ize this function. POI data includes the customer information (address, phone number, etc.), so the data is encrypted to prevent the information leakage when the data is lost.
FUJITSU TEN TECH. J. NO.33(2009) 25 FUJITSU TEN TECHNICAL JOURNAL
"assist" Web application Struts 6. Conclusion ：Order of communication 6 Conclusion
ActionServlet for "assist" Completing the development, we released AVNLite in Framework of Struts October 2008. It is with great pleasure that we can Action for "assist" Map code conversion library Client report that the AVNLite is highly appreciated in the mar- ket. It is tribute to all parties concerned inside/outside Message, log definition, etc. 、 、 the company, who supported its rapid development. I JSP for "assist" would like to take this opportunity to thank them. In the
Fig.17 Application Configuration of Public Server future, we will make efforts to continue providing "conve- nience" and "secure" to our customers by enriching our We set the access limitation in this interface (assist) to WEB service and strengthening the functions of linkage prevent the access of terminal other than mobile phone. applications. In addition, we added the authentication key to input parameter to identify the invoking from the preemptive notification e-mail for corporation model, and to strength- en the security.
Profiles of Writers
Kenichirou MASE Takeshi FUKADA Kenji UEMATSU Entered the company in 1986. Since Entered the company in 1995. Since Entered the company in 2005. Since then, has engaged in the software then, has engaged in the browser 2008, has engaged in the WEB appli- development for in-vehicle devices. development for in-vehicle devices cation development by way of the Currently in the .Software Engineering and in the software development for software development for in-vehicle Department 1, Engineering Division, E-iSERV. Currently in the .Software devices. Currently in the .Software Software Engineering Group. Business Department, Engineering Business Department, Engineering Division, Software Engineering Division, Software Engineering Group. Group. Teru SAWADA Entered the company in 1988. Since 2001, has engaged in the product planning operation by way of the product engineering development. Currently the Team Leader of the Software Business Department, Engineering Division, Software Engineering Group.
FUJITSU TEN TECH. J. NO.33(2009) 26 AVM System
Satoshi SHIGEMATSU Yoshikazu CHUJO Haruo YOSHIDA
Abstract
Based on the mobile communication technology used for radios installed in commercial vehicles, in 1981 FUJITSU TEN developed "distributed-transmission AVM system," which locates vehicles by signposts. Since then, we have been successful in making vehicle-dispatching work more efficient for taxi companies by developing a vehicle locating system by GPS and then by connecting it to taxi companies' customer management, taxi auto- matic dispatch, and, most recently, CTI. Our latest model realizes an unmanned vehicle dispatch system by automation of the dispatching work or "completely-automatic vehicle dispatch, using an interactive voice response system and internet order- receiving system. This paper reports on the outline and characteristics such as "completely-automatic vehicle dispatch" of our lat- est AVM system.
27 FUJITSU TEN TECHNICAL JOURNAL
1 Introduction1. Introduction picking up a customer soonest who requests a taxi on the telephone or the Internet. AVM (Automatic Vehicle Monitoring) system refers With reference to Fig. 1 "Explanatory Illustration of to the system that enables computers in a control center Function," the function and procedure of the AVM sys- to monitor and control the location, vehicle dynamics, and tem is explained here, dividing it into several steps from other operation service situations of their vehicles. The order-receiving and search for the optimum vehicle to AVM systems that FUJITSU TEN has developed are dispatching instructions. The individual numbers show- "taxi dispatch systems," which are intended to be used by ing the steps in the explanation below correspond to the taxi companies. ones shown in Fig.1. In 1981, we developed the "distributed-transmission ①When receiving a request for a taxi from a customer AVM system," which locates vehicles by signposts, based on the telephone, the customer's telephone number is on mobile communication technology used in in-vehicle gained by using the caller ID service. radio for business use. Starting with that, we have since ②Based on the telephone number, the customer's infor- developed and supplied cutting-edge AVM systems: in mation [name, location (longitude and latitude), address, 1994 we launched "GPS-AVM system," which located vehi- building name, route, etc.] is searched from the accumu- cles using the GPS; We achieved an automatic vehicle dis- lated customer database. patching system by linking it to our customer's control sys- ③Locations of vehicles (by GPS or navigation) and cur- tem in 1996; and we released, in 1997, "GPS-AVM system rent situation of their availability (of being occupied, link to the CTI," which combines the caller ID function. vacant, or in waiting from taximeter or input to AVM Moreover, we commercialized digital radios for business terminal) are collected in advance from data communi- use for the first time in the industry in 2003. Since then, cations between taxi vehicles and the control center. we have had the largest market share in the industry. The AVM system automatically searches for and This paper elaborates on the outline and characteris- selects the optimum vehicle according to that data in tics of our latest AVM system. addition to the customer data in ②. ④The server transmits dispatch instruction data, ②, 2 Outline2. Outline of of our our AVMAVM System System (address, building name, customer's name, route, notes, etc.) to the vehicle selected in ③. 2.1 Functions ⑤The dispatch instruction data is transmitted to the The taxi AVM system is a vehicle control system selected vehicle from the radio at the base station. that selects and dispatches the most efficient vehicle for
Order received Automatic search Transmission of customer info. Dispatch instructions in in CTI system for optimum vehicle as digital data over radio text＋synthesized voice＋on map
Control center
Caller ID service Automatic Automatic Automatic search for search for Instructions to Customer customer data optimum vehicle pick up customer Base station
Location Location Send me a taxi, of guest please. of vehicle Location information
Customer Vehicle location database database
Dispatch instruction data
※The vehicles can be located all the time through voluntary data transmission, which enables swift search for the optimum vehicle to pick up the customer. Text＋Synthesized voice＋[Map]
Mr. Suzuki In front of the main gate of FUJITSU TEN. Navigation Dispatch It's address isis….…. [optional]
Fig.1 Outline of Functions
FUJITSU TEN TECH. J. NO.33(2009) 28 AVM System
⑥The selected vehicle conveys to its driver the received ・Base station equipment: radio unit installed in the dispatch instructions on a display, in a synthesized base station voice, and/or on the route map. ・Mobile station equipment: radio and ECU installed in ⑦The vehicle that receives the instruction goes to pick vehicle up the customer and the dispatch process is completed The following section explains the composition taking when the vehicle arrives at the destination. Fig. 2 as an example. In this example, the system is composed of a server in 2.2 Product Lineup the center and two PCs at our client. Those PCs include The product lineup of our AVM system includes telephone function by a dedicated board, and they are Type-A for large businesses and Type-S for small busi- connected to a telephone exchange, which enable receiv- nesses. There are two types of terminals (ECU) for ing orders on the telephone. mobile stations: one is high-performance handy terminal In the actual situation, the system is combined with and the other is simple six- button terminal available only ledger sheets, printer for various ledger sheets, a backup for Type-S. server, an uninterruptible power source, MOs for data Table 1 shows our product lineup. back-up, and other equipment to improve the reliability of the system. 2.3 Composition The equipment at the base station comprises a radio 2.3.1 Key Components unit including a radio and an antenna, radio table for The AVM system comprises the following three com- operation of calls, etc. This example shows the simplest ponents, form where a radio for the base station is installed in the ・Center equipment: server, PC, etc. installed in the control center. Depending on the size and the location of control center the taxi company, multiple radios for the base station are
Table 1 Product Lineup Type－A Type－S Base Client server PC Shared by 10,000 vehicles 100 vehicles Operation of Separated from vehicle dispatch Integrated with vehicle dispatch order-receiving Example of Screen for vehicle dispatch Screen of Map display Automatic dispatch of the vehicle selected by the The control center selects and dispatches a computer after the customer information is vehicle, checking the locations of vehicles on the checked. map.
Terminal for Handy terminal Six-button terminal Mobile station (for Type-A and (only for Type-s) Type-S)
FUJITSU TEN TECH. J. NO.33(2009) 29 FUJITSU TEN TECHNICAL JOURNAL
Exchange Open network Control center Server Base station PC PC Mobile station
LAN Radio table Radio unit
Fig.2 Image Drawing of AVM System Components installed at multiple places within its business area. The "Full automatic vehicle dispatch" automates the order- case where multiple radios are installed at multiple places receiving operation by the interactive voice response is explained in Section 3.2.1 "IP Connection of Radios at (IVR) system and the Internet. As a result, taxi users do Base Station." not have to talk to operators to request a taxi to go to A radio, ECU, and handy terminal are installed in their destinations. each vehicle as a mobile station. An emergency switch, 3.1.1 Interactive Voice Response (IVR) System navigation and drive recorder can be connected to the On the IVR system, a computer answers telephone system as options. calls instead of operators, which automates the order- 2.3.2 Improvement of Reliability receiving operation. In order to improve reliability of the system accord- (1) Function ing to the scale of the system and requests from each In the automatic order-receiving, a computer can han- taxi company, fail safe systems are available such as dle orders of "one taxi" or of "multiple taxis." back-up server, hardware duplexing, automatic switch- A taxi company can select the order-receiving opera- over to back-up radio at the base station, etc. tion process of "one taxi" or "multiple taxis" for each tele- phone line (if the taxi company has multiple telephone 3 Characteristics3. Characteristics of of System System lines, they can select and set the process of "one taxi" or "multiple taxi" for each line), and the functions included in "Full automatic vehicle dispatch" and "use of IP net- each process shown in Table 2 can be selected according work" are among the keywords of the latest AVM sys- to the situation of taxi company. tem. Table 2 Available Setting Function of IVR "Full automatic vehicle dispatch" means that full auto- matic operation from receiving an order from a customer One taxi Multiple taxis Notification of license number who needs a taxi to guiding the taxi to the destination. It Selectable Selectable (only for order receiving) enables a taxi to be dispatched in an automatic and effi- Transfer when the number is cient manner. Selectable Selectable unavailable or unregistered "Use of IP network" means that IP network, such as Transfer to operator after the Selectable Selectable Internet network, is used for the digital radio to be con- predetermined minutes nected to the server in the center. This has enhanced Transfer to operator of ― Selectable affinity between data over LAN and open networks, and customer operation it established the base for wide area communication and Operation of confirmation ― Selectable share of equipment among base stations. Largest number ― Selectable This section explains proven examples of "full auto- matic vehicle dispatch" and "use of IP network" in the (2) Characteristics AVM system and their mechanisms. The IVR system automated the order-receiving opera- tion. However, it did not automate all patterns but only 3.1 Full Automatic Vehicle Dispatch certain patterns. The conventional AVM system only included the The following are the conditions for the automatic functions to select a candidate vehicle from the informa- order-receiving. tion about where the customer is located, to automatically ・Only one pick-up place for each telephone number send the pick-up instructions to the candidate vehicle, (=identify the pick-up place by the telephone number) and to guide the vehicle to the customer based on the ・The customer information is registered in the center pick-up instruction data. Orders from customers were database and the taxi company can identify the cus- received by operators. tomer by the caller ID function
FUJITSU TEN TECH. J. NO.33(2009) 30 AVM System
These conditions are intended to eliminate complicat- ⑥The mobile station that receives the dispatch specifica- ed operation, a disadvantage of the IVR. Normally, a cus- tions goes to pick up the customer following the dis- tomer needs to handle all troublesome operation because played text and navigation. the IVR required the customer to input all information These operations acheive full automatic vehicle dis- with push buttons on the telephone. With our new tech- patch without an operator in the center. nology, this disadvantage was eliminated by limiting the The programs that run on the IVR-connected PC in patterns that customers request taxis. We designed a this system produce dispatch specifications from signals simple flow such that only one push of a button is enough from the IVR system, although the specifications are usual- to arrange a taxi after the customer calls the taxi compa- ly produced by an operator when he/she receives an order ny with the IVR system. on the telephone. As a result, the base of the AVM sys- Moreover, in the case of a taxi company, another call tem did not have to be changed to add the IVR function. from the same telephone number within a certain num- 3.1.2 Internet Order-receiving System ber of minutes after receiving an order often turns out to In order to improve operability, the IVR system is be an inquiry or complaint. Therefore, when the same available only in some limited situations. Accordingly, the customer calls again within those specified minutes, the registered pick-up location is limited to only one for each phone call is transferred to an operator. We improved customer. However, in the actual order-receiving, an the operability of the IVR to increase its use by cus- increasing number of customers call for taxis from outside tomers. on the mobile phones that are popular now. Those (3) Flow outline phones include functions for use both as a phone and the Fig. 3 shows the composition of a system using IVR. terminal for the Internet. We use the Internet function of the mobile phones to realize the automatic order-receiving. ① ③ (1) Function Telephone IVR PC linked The Internet order-receiving is available from PCs or ② exchange Server to IVR mobile phones through the Internet network. The follow- ④ ing items can be completed on the Internet order-receiv- ing system. AVM AVM ・Registration and change of user information Client Server User information such as name, password, etc. can be registered and changed with user ID that is an e-mail ⑤ address. ・Request for a taxi (right now or reservation for a ：Telephone line requested time)
：LAN ⑥ It is possible to pick up the customer not only right Fig.3 Composition of IVR System now but also at a requested time ・Request for picking-up at the designated place The outline of the flow from order-receiving to pick- Since a customer can request a taxi from outside, up of the customer is explained below, referring to Fig. 3 he/she can designate a place to be picked by registering "Composition of IVR System." The individual numbers the place in advance or can select from landmarks. showing the steps in the explanation below correspond to ・Check the order processing situation the ones shown in Fig.3. It is possible to check the reservation and location of Flow outline: the taxi on the way on a map by the tie-up with the map ①The IVR server recognizes the customer ID (telephone website. number) and sends his/her ID number to the PC that ・Confirmation is connected to the IVR system when a customer calls. An e-mail is sent to the designated e-mail address to ②The IVR-connected PC searches the customer data on confirm the order, notify the taxi's license number, and the AVM server based on the customer ID and trans- announce the taxi's arrival. mits it to the IVR server. (2) Characteristics ③When the IVR server receives the confirmation from First, for "customers who requests a taxi from outside," the customer, the server transmits it to the IVR-con- Internet ordering from a mobile phone is available through nected PC. the Internet network. It is available from mobile phones of ④When the IVR-connected PC receives the confirmation, all three carriers in Japan (docomo/au/Softbank). it creates dispatch specifications based on the customer Operation to enter letters on mobile phones is more difficult data, and transmits it to the AVM server. than on PC. Therefore, mobile phone model ID (unique ⑤The AVM server transmits instructions to the vehicle number for each mobile phone) is recorded so as to allow in the neighborhood of the customer based on the dis- the customer to log in without user authentication (entering patch specifications. ID and password) from the second time and thereafter.
FUJITSU TEN TECH. J. NO.33(2009) 31 FUJITSU TEN TECHNICAL JOURNAL
Customer J-COM Control center (Automatic dispatch)
ＧＷ Control center Temporary registration Web server Re-enter PC Notification of Information for temporary registration Database temporary registration Temporary data ①Tel No. ②Name (Kanji) Notification of registration for ③Name (Kana) Mobile ④Pick-up place phone online order is completed Registration to Notification that database registration is completed ・Customer・Customer name Available ・Pick-up・Pick-up places Registration (multiple) (multiple) ・Tel・Tel No. ・E-mail・E-mail address Request for re-registration Notification that In the case of registration is not possible insufficient data or wrong information
Fig.4 Flow from Receiving Order on The Internet to Taxi Arrival
Second, the Internet ordering system has an advan- (3) Flow tage that a customer can designate and enter a pick-up Fig. 5 shows the flow of the Internet order-receiving place. However, it may be a disadvantage for the taxi system. company that dispatches a taxi because they cannot The system shares the composition and the flow with judge whether or not the entered place is correct. the IVR system. By installing a web connecting server Therefore, we decided to make this service available only between web application on the Internet side and the after the registered place is confirmed between the cus- AVM system, the IVR function for the Internet order- tomer and the taxi company in advance. (Refer to Fig. 4 receiving could be added without any change of the base Flow from Receiving Order on The Internet to Taxi application of the AVM, like the IVR. Arrival.) ID・Order-receiving The pre-registration contributes not only to decreas- Dispatch specs. ing misunderstanding of the pick-up place between the F/W Web Server AVM customer and the taxi company but also to reducing Server linked to Server web entering operation by customers because he/she can User select a place from the list of pre-registered places. Data Third, we enabled the customer to check the situation Fig.5 Composition of Internet Order-Receiving system of the taxi on the way to him/her. The order confirma- tion e-mail includes an URL, which jumps the customer to (4) Security the map company website that shows the latest location Because this system uses private information on the of the taxi heading for the customer on the screen of Internet, the following security measures, although they his/her PC or mobile phone. The coded customer ID and are common technologies, are taken. the information of the vehicle to be dispatched is embed- ①Cryptographic communications in SSL ded in the order confirmation e-mail. When the customer SSL is a cryptographic protocol for information com- accesses the specified URL, the web application inquires munications over the Internet. It encrypts data for com- the latest taxi vehicle information to the AVM system. munications. The AVM system transmits the latest location of the ②Encryption of information vehicle to the web application, the application jumps to URL is assigned to each taxi customers. Therefore, the website of the map company to show the map with the customer information is encrypted by our original the coordinates. With this function, we realized that the coding not to easily identify those customers. inquiry from the customer can be answered automatically ③Network establishment without asking an operator. As shown in the example of the system composition,
FUJITSU TEN TECH. J. NO.33(2009) 32 AVM System
250000
200000
150000 Number of dispatch by IVR 100000 Number of dispatch 50000
0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct 2006 2007 2008
Fig.6 Shift of Number of Vehicle Dispatch the system is divided into segments and only the web 3.2.1 IP Connection of Radios at Base Station server can be seen from the outside. The firewall is set As explained in Section 2, in the case where a radio up to block unauthorized access to the web server. cannot cover all the business area of a taxi company, the ④Database unification second radio is installed in a remote area (referred to as All data including private information such as cus- the advanced base) in order to cover all the necessary tomer information is brought together in the AVM server area. Fig. 7 shows an example of the structure of two so as not to be accessible. base stations. 3.1.3 Result and Challenge of Full Automatic System Fig. 6 shows the actual dispatching result of a taxi Main base area Advanced base area company that has used the IVR system. Since November Main base 2006 when the company introduced the IVR system, the Advanced base number of dispatches has greatly increased. The rate of the full automatic dispatch is about 10 percent. It can be Exchange Leased line concluded that one of the reasons for the large increase is Fig.7 Composition Example of Two Base Stations because operators can have extra time to take other orders due to the full automatic dispatch (approx. 120% The server automatically determines the main station more than the previous year). or the advanced station for communication with a vehicle The challenge of the full automatic dispatch system is based on the location of the vehicle. how to convince users to use it. Conventionally, the control center is connected with Conventionally, those who use such system were oper- the advanced base station by an analog line and modem. ators of taxi companies. The problems related to the Recently, as the IP network infrastructure is more specifications and operating methods were solved by dis- extensive, the connection between the control center and cussion with those taxi companies. However, in the case the advanced base are compatible with the IP where the of the IVR system, those who use the system are taxi digital line or optical communication line are used. Fig. 8 users. Therefore, we need to understand the policies and shows the structures of the analog and digital lines. sales promotion efforts of each taxi company. In the While the conventional analog communications beginning of the system's introduction, there were some required two lines, one for text data and the other for cases where users did not use the system (which means audio data, the IP communications needs only one line no sales increase). However, advertisement and campaign because both audio and text data are transmitted by activities to users about the automatic order-receiving sys- using the IP, which is the advantage of the IP communi- tem by the taxi companies have gradually increased sales. cations. Moreover, the digital line costs less than the ana- Moreover, some functions have been added as taxi log line in the case of long-distance communications. companies carry out promotion activities such as cash- Therefore, it makes a long-distance connection feasible, back to those who use the full automatic dispatch system. for example, connection between Tokyo and Osaka. Due to our efforts linked to those activities, there are The changes mentioned above leads to an idea that some cases where the full automatic dispatch accounts the remote communications between the control center for 20% of the total. and the base station can be used not only to "supplement the radio area by installing the advanced base" but also to 3.2 Use of IP Network "integrate the dispatch functions at main business offices In addition to the Internet order-receiving, the IP net- of large nationwide companies." In fact, some companies work is promoted for control of the base station and con- started a system where vehicles needed in Tokyo are dis- nection among business offices. patched from the dispatch center in Kyoto.
FUJITSU TEN TECH. J. NO.33(2009) 33 FUJITSU TEN TECHNICAL JOURNAL
Analog leased line connection Center Analog leased line Server
Modem Modem Radio unit at Router Optical modem base station Radio unit Remote control for analog line Optical network Digital leased line connection (IP connection)
Digital leased line Business Office Optical Business Office Optical modem modem Router Router Router Router
Radio Fig.9 Connection among Business Offices VoIP VoIP unit Fig.8 Analog Connection and IP Connection Conclusion4. Conclusion 3.2.2 Link between Business Offices 4 Another example of the IP communication network is We explained the latest technology and characteris- a link between business offices. tics of our new AVM system. Contrary to the recent integration of dispatch func- All radios used in taxis need to be digitalized by 2016. tions, there has been some demand for deconcentration of The taxi industry was deregulated and taxi companies dispatch terminals to individual business offices to dis- are fiercely competing with each other to survive more patch vehicles from each of those offices. The link than ever. Some of them try to find a way to survive by between business offices meets the demand. It enables offering a service different from others with the introduc- the dispatch from multiple bases by connecting a compa- tion of the AVM system. In these circumstances, we ny server with the one of its client by the optical net- need to support them by providing solutions for smooth work. and safe taxi operation with the system, in addition to Fig. 9 shows the structural example of the connection automation and efficient dispatch of taxi vehicles. among business offices. Furthermore, the better connection with computers leads to the ongoing discussion about expansion or sup- plement of service area by sharing a base station among multiple companies or about prediction of traffic conges- tion from the information obtained by the system. We hope to develop the AVM system further as a more sophisticated dispatch system and also as a system that can serve various other purposes by using information.
Profiles of Writers
Satoshi SHIGEMATSU Yoshikazu CHUJO Haruo YOSHIDA Entered the company in 1998. Since Entered the company in 1982. Since Entered in the company in 1997. then, has engaged in GPS-AVM sys- then, has engaged in GPS-AVM sys- Since then, has engaged in GPS- tem (center application) development. tem development by way of develop- AVM system (center application) Currently, in the Information & ment of automotive radio devices. development. Currently, the Team Communication System Engineering Currently, in the Information & Leader of the Information & Department, CI Aftermarket Group. Communication System Engineering Communication System Engineering Department, CI Aftermarket Group. Department, CI Aftermarket Group.
FUJITSU TEN TECH. J. NO.33(2009) 34 NOTE Technology Trend in Remote Engine Starter
Norio TSURUTA Akira TAKEE
1 Introduction1. Introduction Generally in the aftermarket remote engine starter system, a dedicated remote control for engine start caus- A remote engine starter is designed to provide com- es the engine to start from a distance, but in our system, fortable room temperature inside a car whether that is in the remote control of wireless car door lock system can the winter time of cold areas such as Hokkaido or the start the engine. In other words, users are free from the summer time of extreme heat or desert areas such as inconvenience of having two remote controls, and this is Arizona in North America, by operating the air condition- the characteristics of our system. er through activating the car engine before a driver gets The door lock remote control functions within a dis- into the car. tance of approx. 20 m to 50 m, and naturally, our remote Additional benefit is that an exhaust muffler, warmed engine starter is designed to function within the distance up by idling, has a better catalytic effect, and this leads to that a driver can recognize his / her car visually (the a reduction of toxic exhaust gases. range where the wireless door lock system can function). This paper will introduce the outline of our remote Thus, our remote engine starter is suitable for use mainly engine starter system and its market trend. at residential parking or a parking location close to an office. 2 Outline2. Outline of of Our Our Remote Remote Engine Engine StarterStarter SystemSystem 2.2 Engine Start Procedure 2.1 Characteristics of Our System Our remote engine starter requires three-consecutive- Fig. 1 shows the configuration of our remote engine pushes on the lock button of a wireless door lock remote starter system heading for North America, which is com- control for engine start. When a user operates three-con- posed of our remote engine starter ECU and gateway secutive-push (short / short / long push) on the door lock ECU connected to vehicle bus. Communications between button of a remote control, an in-vehicle wireless ECU in-vehicle wireless door lock system and engine control receives the door lock signal and sends the three-consecu- system enable an engine start from a distance. tive-push signals to the remote engine starter ECU
① 3 consecutive signals for door lock Vehicle bus
② Immobilizer deactivation MSG Gateway ＥＣＵ Wireless Immobilizer Engine ECU ECU ＥＣＵ Accessory bus
Remote engine starter Wireless door lock ECU remote control Communication control
Immobilizer control ③ Starter motor driving output
Engine start control
Fig.1 Configuration of FUJITSU TEN Remote Engine Starter System
35 FUJITSU TEN TECHNICAL JOURNAL
through a vehicle bus. The remote engine starter ECU 2.3.3 Warm-up Time that received the door lock signal sends an encrypted The remote engine starter has a function to stop the message to deactivate the immobilizer in order to make engine automatically after a certain period of idling. This the car get ready for engine start. After the immobilizer automatic stop function is designed in consideration of is deactivated, the remote engine starter ECU drives the environment so as not to keep an engine in an idling state starter motor. for a long period. The remote engine starter has a retry function, which Considering user's safety, we provide a function to is to repeat an engine start when the engine can not start stop the engine before a user gets poisoned by the car- while the car satisfies the conditions for engine start after bon monoxide when exhaust fumes fill the space due to the control of the engine starter. starting the engine in an enclosed parking space. With measuring the concentration of carbon monoxide at our 2.3 Safety of Remote Engine Starter System test site, we examined the time from starting an engine 2.3.1 Fail-safe Function to reaching 50ppm of concentration that is the level to This remote engine starter can function only with an cause symptoms of poisoning. The result shows that the automatic transmission car (hereinafter referred to as an concentration exceeded 50ppm in approx. 10 minutes at automatic car). In the automatic car, because of the the measurement points around the car after an engine transmission structure, the signal of the shift position in started. Applying this result, our remote engine starter parking range (P position) can be used as an input signal is designed to stop an engine automatically in 10 minutes for engine starter. In the fail-safe function, the signal is in consideration of both warm-up effect and safety time used to enhance the safety at the time of engine start in up to reaching to the risky level of carbon monoxide con- the method by judging this signal as the one output from centration. a "stopped-state car." With the automatic car, the remote engine starter can 2.4 Preventing Vehicle Theft operate the engine only of the stopped-state car by judg- This remote engine starter requires a function to pre- ing this input signal as above. On the other hand, with a vent car theft because this system functions to start the manual transmission car (hereinafter referred to as a engine of an unmanned car. manual car), it is impossible to recognize its shift lever 2.4.1 Permission of Engine Start state in a neutral position. Considering the safety of this As described in the engine start procedure (Section product making the engine to start from a distance, we 2.2), the remote engine starter makes a car get ready for decided not to install the starter system on a manual car. starting its engine by communicating with an immobiliz- In addition, this system has another control function er. When the encrypted code is matched between the to prevent unexpected problems caused by suspicious remote engine starter and the immobilizer, the immobiliz- acting around / inside a car through monitoring various er gets ready for starting its engine. car signals such as door / hood open-close signals, brake The system prevents these car thefts which have a signals and engine rotation speed as well as the shift modus operandi of affixing a product other than the lever position signal in order to stop its engine. remote engine starter permitted by this car system and 2.3.2 Prevention of Error in Remote Control Operation making the engine to start illegally. This is the charac- In our remote engine starter, since a user pushes the teristic of a genuine system. button of the wireless door lock remote control of a car to 2.4.2 Security Function start its engine, the starter must have a function to deal The unmanned car during a warm-up through the with unintended operations by a user. engine start by using a key is in danger of a car theft. For example, to prevent the case that the engine On the other hand, since the remote engine starter coop- starts unexpectedly due to the lock button in a pocket erates with a door lock control and enables the car being pushed accidentally by a user, we studied the engine to be warmed up with its doors locked, it is cate- appropriate sequence for pushing the lock button several gorized as a security product. times. The more complicated the sequence becomes, the In general, when a key is inserted in an ignition key more inconvenient the operation becomes. Thus, to cylinder, doors cannot be locked due to the function of decide the appropriate sequence, we asked general public key-in-lock prevention. However, the remote engine users to operate the remote control and studied the starter cooperates with a door lock control, and during appropriate number of times, interval and period of a warming up an engine, doors are locked. push. Then, we obtained the best sequence for operation Our remote engine starter, cooperating with a car "to start engine only when a user intends preventing security function, turns the car security system into an errors in operation." alert state when receiving wireless door lock signals for engine start. When the security system detects a suspi-
FUJITSU TEN TECH. J. NO.33(2009) 36 Technology Trend in Remote Engine Starter
cious intrusion during warming up the engine, the system special remote control with an engine start button, anten- stops the engine. This is also one of the genuine system na being attached to a window glass and an ECU. The characteristics. mainstream frequency band for a remote control is 300MHz / 400MHz, which conform to FCC (American 3 Trend3. Trend of Worldof World Remote Remote Engine Starter Starter radio law) and IC (Canadian radio law). The remote control can operate in the range approx. Here is the explanation of the trend of remote engine 100m to 300m around a car, for example in the large starter systems in each region of the world. parking lot of a shopping mall or a hotel. Nowadays, in accordance with FCC revision, the output power up to 3.1 North American Market 1W in 900MHz band, which is opened for a spread spec- 3.1.1 Temperature in North America trum system, is now available. Thus, some new aftermar- Since it is cold in winter with low average tempera- ket systems provide a long distance operation; 1 mile tures in wide areas of North America, demands for (approx. 1.6km) according to their catalogs, and the remote engine starters are strong. (Fig. 2) demands for longer distance operation in remote engine Especially in Canada, areas where the average tem- starters are increasing year by year. peratures in winter are -20°C or less are widespread, [Genuine system] and the demands for remote engine starters are stronger As for automakers' trends, starting with the Malibu than in America. 2003 model in which GM provided the first remote engine starter as a genuine line option, each automaker has [North American Temperature data (Average temperature in January)] expanded the adoption of remote engine starters as gen- uine options. The GM genuine system is convenient for use with no －20℃ or less need to have another remote control because its wireless door lock remote control is equipped with the remote ◆◆WinnipegWinnipeg －10℃ or less ◆◆TorontoToronto engine starter button. Canada-USA Boarder ◆◆NewNew York Other automakers provide remote engine starters as dealer options, but in some cases, the systems sold as genuine systems are just the ones that aftermarket sys- tems are simply changed in their appearance designs. The efforts to develop remote engine starter systems vary according to automakers. The mainstream genuine system in North America is ※Meteorological Agency website: excerpt from world weather statistics a one-way communication system with approx. 100m to Fig.2 North American Temperature Data (January 2008) 200m operation distance, but these years, GM and Honda have improved its performance in a car condition display 3.1.2 System Trend and others, and have released a two-way communication [Aftermarket System] system with a remote control with a liquid crystal display Since remote engine starters have been released as as a genuine system. aftermarket systems in North America since the relative- [Immobilizer deactivation method] ly early years in the late 1980's, the starters now have a The immobilizer deactivation methods are different lot of visibility. Naturally, North American users have a between the genuine system and the aftermarket system. good knowledge of remote engine starters, and at the The genuine system enables an engine to start by same time, have a variety of demands to the system. communicating with an immobilizer as explained in Aftermarket manufacturers provide various systems to Section 2.4.1, but the aftermarket system requires a spe- attend to various users' demands. They provide various cial "immobilizer bypass kit" to be installed. This is systems in functions and cost, for example, two-way com- because the aftermarket system cannot communicate munication system (high functionality and high price) that with an immobilizer to cooperate with a car for control, users can check the car conditions on a remote control by differing from the genuine system. using two-way communication between in-vehicle engine For the deactivation, to read the ID code of the immo- starter ECU and a remote control with a liquid crystal bilizer, the aftermarket system requires a connection display, and one-way communication system (low func- between a spare key and an ignition key cylinder by tionality and low price) with a remote control only to con- attaching to a coil antenna. When the remote engine trol the engine start. starter sends a code for engine start, the system reads an An aftermarket system is basically composed of a ID code from the spare key and sends the ID code to the
FUJITSU TEN TECH. J. NO.33(2009) 37 FUJITSU TEN TECHNICAL JOURNAL
ignition key cylinder, which enables the immobilizer to be ers, as well as by domestic manufacturers, are high in vis- deactivated artificially. The aftermarket system is inferi- ibility. or to the genuine system in safety and reliability because The aftermarket system requires the "immobilizer the spare key has to be kept inserted in a car, and at deactivation adapter" of an aftermarket option to be some attached positions of the coil antenna, the immobiliz- installed to the car with an immobilizer in the same man- er cannot be deactivated. ner as in North America, but the aftermarket system is impossible to be set to a push-start car that differs in an immobilizer deactivation system. [Genuine system] Unlike North American automakers, many domestic Ignition key cylinder automakers including light car automakers provide remote engine starters as their dealer options. In the genuine system, as well as the aftermarket sys- tem, products with the two-way communication system of the specified low power type are provided. But the Coil antenna remote controls of the genuine system, as their features, are designed in appearance to be matched to the cars in cooperation with automakers. For example with Toyota Bypass adapter of products, the designs of controls are different between immobilizer Toyota brands and Lexus brands. (Fig. 4) The designs (for spare key insertion) also follow their car design concepts in waterproof prop- Fig.3 Immobilizer Deactivation Method of Aftermarket System erty, durability and so forth. In Japan, the remote controls for wireless car door To solve these problems, some aftermarket manufac- lock operation use weak radio waves, and the range of turers adopt the system that reads communication sig- approx. 3m to 10m is available. Thus, the use environ- nals from a car connector and stores the immobilizer sig- ment of remote controls for engine start is different from nals on a remote engine starter directly. As above, the that in North America, we have not yet provided car aftermarket systems are improving these days. manufactures with the remote controls for engine start like North American types, as in-built system. 3.2 Japanese Market 3.2.1 Cold Areas in Japan In Japan, remote engine starters are essential for peo- ple in cold areas such as Hokkaido and Tohoku. In Hokkaido, approx. 70% of users buy a remote engine starter at the time of buying a car, and thus, automakers must prepare the remote engine starters as lineups. 3.2.2 System Trend [Aftermarket system] Remote control of Remote control of In Japan, the special remote controls of specified low Toyota brand Lexus brand power type are used in many systems. In the specified low power type, output voltage and antenna gain perfor- mance of a remote control can be set higher than those for the weak radio wave type. Thus, this system is avail- able even under such a residential environment as high- rise condominium or parking at several kilometers away from home. Engine start remote control of other automakers Domestic aftermarket manufacturers provide various Fig.4 Remote Controls of Japanese Automakers product lineups, and users buy and install these products normally at car equipment retailers or dealers. Japanese users tend to favor an advanced system, and the main- 3.3 Other areas stream is a two-way communication system using a In Europe, reflecting the strong consciousness of envi- remote control of specified low power type equipped with ronmental regulations, most countries have rules for a liquid crystal display and a LED. Among Japanese idling. Thus, it is prohibited to start an engine under the users, expensive systems made by overseas manufactur- unmanned situation (where the use of a remote engine
FUJITSU TEN TECH. J. NO.33(2009) 38 Technology Trend in Remote Engine Starter
starter is suitable), and even in low-temperature areas the 4.3 Engine Starter System in Future market of the remote engine starters is still small. The remote engine starter functions away from a car, As for the areas other than cold areas, the engine but in the future, the environment is being ready for this starters are in demand in Middle East area. In contradic- system to function at further distance than the present tion to the cold areas, Middle East is a desert area and remote control operation range as a car-outward commu- the highest temperature is over 40℃ in a daytime. So, nication technology is improving. Through reviewing the the remote engine starter is used for cooling a car inside. conventional concept of safety, it is possible to develop a more convenient system. Besides, in accordance with the 4 Future4. Future EffortEffort development of hybrid cars and battery cars, the future role of the remote engine starter will be changing. To 4.1 Tasks in Our Remote Engine Starter System make the environment inside a car comfortable, we Since the release of the remote engine starter as a should not stick to the concept of the conventional genuine system for Toyota in 2005, we have investigated method for engine start. We need to keep addressing the users' requests to our remote engine starters by having development so as to provide users' comforts and added questionnaires to the users who bought our product or values in accordance with advanced car functions. by holding hearings at dealers. Through the investiga- tion, we received demands for more advanced functions such as two-way communication system and cooperation with smart entry system, as well as for larger available range of remote controls and advanced operation meth- ods. We have set our future tasks to develop products conforming to these various users' demands.
4.2 Future Market Growth The remote engine starter market has been growing around North America and Japan, and in the future, it will be growing in developing countries such as BRICs. In fact, the remote engine starters are receiving more interest among people in China and Russia. We need to address the investigation of their market environments and clarify the performances / functions conforming to their market demands.
Profiles of Writers
Norio TSURUTA Akira TAKEE Entered the company in 1988. Since Entered the company in 1987. Since then, has engaged in the development then, has engaged in the development of car electronic devices. Currently of car electronic devices. Currently the Team Leader of the AE the Department General Manager of Engineering Department 2, AE the AE Engineering Department 2, Engineering Division, AE group. AE Engineering Division, AE group.
FUJITSU TEN TECH. J. NO.33(2009) 39 NOTE Approaches to Technology of Thermal Fatigue Life Prediction of Solder Joints
Hidehisa SAKAI Katsufumi MORIMUNE Takashi YONEMOTO
1 Introduction1. Introduction Installed in engine room
Electronic parts for in-vehicle products must endure Installed in cabin for fifteen or twenty years, longer than electronic parts (FEM) Simulation used for other products. Therefore, these electronic parts require longer reliability evaluation to verify and guaran- Installed in engine Installation Heat from the tee their endurance. environment Cabin product itself On the other hand, shorter development cycle has Heat from the Engine room product itself been in strong demand and the evaluation has to be com- Stress of product pleted one time during the development phase. Specially, installed in engine Solder the life of solder joints to various parts must be checked, Electronic part Simulation and it is a challenge to develop a technology of thermal result fatigue life prediction to make the evaluation efficient. Printed wiring board Fatigue life from heat cycle test FUJITSU TEN has been developing power train Difference in coefficient ECUs as in-vehicle products for nearly thirty years and of linear expansion Fatigue life line we have been developing the technology of thermal Fig.1 Outline of Thermal Fatigue Life Prediction of Solder Joints fatigue life prediction for ten years. In those years, tech- nological changes that affect fatigue life of solder joints ⊿ε which is the sum of plastic strain and creep strain occurred. Among them, the technology that enables gained from the simulation; thirdly, the part to be mount- parts to be mounted on all sides was developed from 1997 ed is simulated based on its shape to calculate inelastic to 2000. In 2004, lead-free solder was adopted for environ- amplitude ⊿ε; and finally, the fatigue life of the solder mental conservation. In addition, the BGA was adopted joint is predicted by calculating it back from the gained for small densified ECUs in 2006. The technology of life life curve. prediction has been used in designing various products Specially, inelastic amplitude and fatigue life are and it has produced positive results. known to follow the Coffin-Manson Law and they are Along with case examples, this paper introduces the expressed in the following equation. technology of thermal fatigue life prediction of solder joints, which employs finite element method (FEM) simu- Nf=C×⊿ε^-n lation that FUJITSU TEN has used for life prediction. Wherein, the values of C and n depend on the fatigue 2. CharacteristicsCharacteristics of Technology of ofTechnology Fatigue Life Prediction of Fatigue of Solder Life Joints strength properties of solder joints, and they change from 2 Prediction of Solder Joints one material to another. ⊿ε is inelastic amplitude gained Fig. 1 shows the outline of life prediction of solder from simulation, and Nf is the number of cycle gained joints. from the test. What should be noted here is that many One of the stress(1) on in-vehicle products is tempera- companies report that the above formula is valid in the ture change in the environment where the products are correlation of a breaking cycle and strain amplitude installed, such as cabin, engine room, etc. and the change because they consider that the breaking cycle where a due to the heat produced by the products themselves joint is broken is more important than the cycle where during their use in vehicle. Generally, the heat cycle test the first crack appears although it is widely accepted that is conducted to evaluate the stress. Nf should be the cycle where a crack appears first. Like The stress caused by the temperature change other companies, FUJITSU TEN has established the life appears as plastic strain and creep strain on the solder curve based on the breaking cycle to failure (Fig. 2). joints due to the difference in coefficient of linear expan- Fig. 2 shows an example of the life curve plotted from sion between the printed wiring board and the electronic the correlation between the breaking cycle of eutectic-sol- part. It is well known that the life of solder joints mainly dered chip resistors in the size of 2125, 3216, 4532 and depends on these strains. In order to predict the life of a 4632 from the heat cycle test in the range of -40 to 125℃ solder joint, first, the fatigue life of the solder is verified and their strain amplitudes gained from simulation. This by a heat cycle test; second, the fatigue life curve is plot- figure shows that those points are laid on a straight line ted based on the correlation between inelastic amplitude and follow the Coffin-Manson Law.
40 Approaches to Technology of Thermal Fatigue Life Prediction of Solder Joints
we gained properties of lead-free solder materials and released results(3) of BGA test and simulation to verify the results. Those properties have been used as the base to predict the life of lead-free solder joints of SMD since then. In 2005, we made public the case(4) where we estab- lished selection criteria of parts for the QFP, employing quality engineering and FEM. The case is detailed in Breaking life of chip resistor and simulation result Section 4. We explained what we have studied in the advance analysis study group. We will move onto the advances made in this technology. Fig.2 Fatigue Life Curve Like this, the life of solder joints can be predicted 3.2 Advance of Life Prediction Technology from the strain amplitude ⊿ε calculated by simulating When we began the discussion about the life predic- each part and fatigue life of the solder from the test. The tion technology, simulation tools were not displayed in life of a part can be predicted according to its shape and GUI. Simulation conditions had to be entered in com- temperature range where it is used. mand lines, which made simulation difficult. At that time, computer processing speed was slow, and it took all night 3. ApproachesApproaches and History andof Fatigue History Life Predictionof Fatigue Technology Life to calculate the strain of a solder joint of a chip resistor. 3 Prediction Technology In these circumstances, it became important to simplify 3.1 Approaches to Fatigue Life Prediction Technology models. In many cases, companies adopted a two-step Around the time when FUJITSU TEN began the dis- analysis (Fig. 4) where the entire structure was analyzed cussion on life prediction technology, Mr. Shiratori of in as simple model as possible and then a solder joint to Yokohama National University and others made public be focused upon was picked up and analyzed in detail. the validity(2) of simulation by using Sn-Pb eutectic-sol- dered sample. Toshiba, Hitachi and some other electronic part manufacturers began examination to adopt the tech- nology for their IC packages. Meanwhile, Fujitsu launched in 1998 an advanced analysis study group, for the purpose of establishing life prediction of BGA solder joints, under the leadership of the simulation department of Fujitsu. In 1999, the study 2nd step (Detailed model) group gained its original nonlinear properties of eutectic 1st step solder, and was verifying whether the technology was (Entire model) valid for the 1mm-pitch BGA. FUJITSU TEN joined the Fig.4 Two-Step Analysis study group in 1999. Since then, we have learned the sim- ulation technique and studied application of the simulation. However, for in-vehicle products at that time, minia- Fig. 3 shows what the study group has studied. turization and diversification was mainly required for the chip parts whose structure was simpler than the BGA '97 '00 '01 '04 '05 '06 '07 '08 package. Therefore, FUJITSU TEN focused on 2-D analy- Mounting on sis that needed less computing capability than the two- BGA mounting all sides step analysis and began to employ it for SMD design. Fig. Actual Eutectic solder Lead-free solder mounting 5 shows our progress of analysis technology of solder life. Eutectic solder prediction Base data Lead-free solder prediction Part selection criteria Method Application Analysis of crack Analysis of development of learning development Analysis in 3-D model visualized crack
Simulation to ＱＦＰ Application FUJITSU TEN approach Crack rate to BGA End Gaining of eutectic Analysis in 2-D model Inter- solder properties Integration of quality mediate Actual engineering and FEM Analysis Establishment of BGA Start Number of cycles analysis method Algorithm of crack Gaining of lead-free development solder properties Study Group
Advanced Analysis Presentation at Presentation at academic conference academic conference
Fig.3 Research Theme of the Advanced Analysis Study Group Resistors, capacitors QFP, SOP BGA.CSP Life Better Prediction of We have made presentations at academic conferences prediction accuracy life expectancy twice as collaborative results of the study group. In 2001, Fig.5 Progress of Fatigue Life Analysis Technology
FUJITSU TEN TECH. J. NO.33(2009) 41 FUJITSU TEN TECHNICAL JOURNAL
After that, lead-free solder started to be adopted to Design in conventional Design employing FEM reduce environmental burden. It became understood that method Examination of Examination of the creeping speed of lead-free solder is slower than solder shape solder shape modification No need for 6 types 6 types Modification eutectic solder when the stress is same. Fig. 6 shows the Narrowing down creep properties of the two types of solder. by FEM 2 types Land design Land design 2 types 6 types SMD SMD Lead-free solder evaluation evaluation 1 type 6 types Reliability test Reliability test
Completed Completed
Eutectic solder Gap under Fillet height part Stress (MPa)
Land length Lead-free solder Fillet Fillet Land Land Gap under Gap under Parameters of land shape height 1 height 2 length 1 length 2 part 1 part 2 Eutectic solder Fig.7 Example of Efficient Land Design of Chip Resistor Creeping speed (1/hr) Fig.6 Relationship between Stress and Creep Strain Rate of Solder Conventionally, we examined six solder shapes with different angle, length and amount of solder fillet, created In short, rather than lead-free solder's properties that prototype samples of all the six types, and conducted are hard and less stretchable, the parts soldered with their SMD evaluation and reliability test that required a lead-free solder brought concerns about their reliability long time. Some of the samples were not approved, and although the same parts soldered with eutectic solder in the worst case, all the samples were not approved. In had been used without any problems. In addition, such a case, we had to start the process all over again. responding to the trend of miniaturization of in-vehicle On the other hand, by employing the FEM, we can select products, thin packages such as the SOP and the QFP a few solder shapes that are predicted that their life were adopted. Accordingly, we have improved the analy- should be longer than that of others. That enables not sis in 3-D model to analyze life of parts in complicated only reduction in the number of samples but also elimina- structures. tion of the possibility to have to start the process all over Moreover, recently, we are addressing a more chal- again after the evaluation is completed. lenging prediction that is to predict the development of a crack from its start to end through the intermediate with 4.2 Part Select Map Employing Quality increasing demand of pin assignment based on the predic- Engineering and FEM tion of the order in which balls of the BGA will break and The life of solder joints became a problem, too, for the the prediction of life expectancy of a part with a crack QFP due to miniaturization of packages (hereinafter from its reliability test results to improve the package. referred to as PKG) and narrow pitches (Fig. 8). Moreover, we found that, in the case of chip resistors, 4 Proven4. Proven CasesCases parts soldered with lead-free solder crack and break earli- er than parts with eutectic solder when the size of them We explained our effort to improve life prediction are the same (Fig. 9). technology. From here, we will introduce some cases Part specifications Cross-sectional surface where our technology was applied to actual products. 0.5 pitch Crack rate 14％ PKG size: 7×7mm PKG size 4.1 Efficient Design in FEM 44 pins The life of solder joints of SMD parts mainly depends 0.65 pitch Crack rate 40％ Pitch PKG size: 12×12mm on soldered shape and amount. The shape and amount is 80 pins specified based on land design values which are deter- 0.5 pitch Crack rate 78％ mined according to the SMD land and metal mask shape. PKG size: 14×14mm 100 pins In order to verify the validity of the land design values, a 0.65 pitch Crack rate 0％ long reliability test is required. Therefore, various evalu- PKG size: 14×14mm ations of the land and metal mask were conducted in par- 80 pins allel. As a result, the evaluations took a very long time Fig.8 Examples of Cracks of QFP Solder Joints because many samples had to be evaluated. Given this fact, we narrowed down the samples to be evaluated in advance by employing the FEM to improve the efficiency of the land design. Fig. 7 shows an example of the process.
FUJITSU TEN TECH. J. NO.33(2009) 42 Approaches to Technology of Thermal Fatigue Life Prediction of Solder Joints
＜Lead-free＞ to find acceptable size and height of the PKG, effect para- meters and plotted the part selection map. In Fig. 11, the Lead-free process of the Monte Carlo simulation in Excel is shown.
Eutectic Factor Typical value PKG size PKG height Fatigue life Terminal width Terminal thickness Approval Terminal pitch Random numbers Terminal length Terminal overall length Estimated formula Disapproval Terminal angle Mold material Excel function Estimated equation Approval/Disapproval Parameter PKG size PKG height ＜Eutectic＞ Crack rate (%)
Monte Carlo simulation PKG height in Excel
PKG size Number of cycles → Selection Fig.9 Comparison of Cracks criteria As mentioned above, solder joint reliability emerged PKG height as a problem with the parts soldered with lead-free solder although the same parts had never had a problem when PKG size soldered with eutectic solder. As a result, it became nec- Fig.11 Monte Carlo Simulation in Excel essary to conduct reliability test for solder joints of all parts that we decided to adopt. However, if we had done Fig. 11 shows the results calculated in the estimated that, it would have required huge man-hours and work. equation of solder joint reliability with "○" for approval Therefore, we improved evaluation efficiency by creating and "×" for disapproval. The dividing line shows the a QFP selection criteria map, through simulation, based selection criteria. The QFP selection criteria gained from on which engineering designers can select parts which the process mentioned above serves as a design guideline meet the reliability of solder joints according to the part for engineering designers to design products. shape. Fig. 10 shows the outline of the process of how we created the selection map. 4.3 Use of Crack Development Analysis As mentioned above, we found that the life of parts Extraction of factors soldered with lead-free solder is, in some cases, shorter Creation of than the one of parts with eutectic solder. That required orthogonal table us to estimate the allowance of the life of a part from the
Effects of factors reliability test results in order to improve accuracy of life prediction with little test data. Moreover, for adopting analysis of variance the BGA, we wanted to narrow down the items to be evaluated for validity that is conducted due to unobserv- ・Contribution rate ・Estimated equation able solder joints located inside. For that reason, we (response surface) developed a method for crack development analysis in Estimated formula＝ which solder cracks are visualized and their start, inter-
Calculation of property mediate, and end can be estimated. This section explains values in Excel the methods used by other companies and the application PKG height of our method. Part selection map PKG size 4.3.1 Trend of Crack Development Analysis Technology Fig.10 Process of Creating Part Selection Map Conventionally life of a part was predicted by accu- mulating correlation, as respective database, between the The life of solder joints of the QFP is affected by vari- final breaking cycle found in test per shape of solder ous design factors such as package materials and dimen- joints, package type, and soldered area and their simula- sions of each part. In the process shown in Fig. 10, we tion results, and then predicting the end of life from the determined, through simulation, the level of the effect accumulated results of a part in relatively-similar shape that those factors had on the life of solder joints based on and conditions. Therefore, it was difficult to predict the a quality engineering orthogonal table and extracted the process of cracking because solder joints of various parts effect parameters. Next, we found an estimated equation crack; the shapes of those solder joints changes in the for life of a joint from multiple regression analysis so that course where the cracks develop; and accordingly stress the life can be calculated simply by entering values of fac- distribution continuously changes. tors from the orthogonal table even without simulation. In order to render crack development in the finite ele- Using the estimated equation, we conducted Monte ment method, the initial finite element model needs to be Carlo simulation, where calculation is done several times, changed dynamically as the crack develops. For exam-
FUJITSU TEN TECH. J. NO.33(2009) 43 FUJITSU TEN TECHNICAL JOURNAL
ple, finite elements are deleted or connection between ele- length (50μm) to find how many models are necessary. ments is cancelled because the joint has lost its rigidity The models are prepared as many as the obtained num- and the elements are not connected any longer. ber from the division, and the life of the joint can be cal- Among the methods where a finite element model is culated from inelastic amplitude of each model, using the dynamically changed as a crack develops, there are the Coffin-Manson Law. element deletion method and the damage path simulation For example, in the case of a joint with the length of method where elements are not deleted but its rigidity is 500μm, it is necessary to prepare ten models each of decreased. Fig. 12 shows those methods, taking develop- which has a crack with the length of 50μm. To calculate ment of a crack on a chip as an example. the life of the joint, the life of crack initiation of those ten models is totaled. Development of crack→ However, if a joint is long, this method is inefficient. Therefore, Mr. Igarashi and others propose a method Element where only some models are analyzed, instead of analyz- deletion ing all, and approximation is used for the unanalyzed mod- els based on the analysis results of the analyzed models. Damage The following three are the advantages of this path method. (1) No need for large computational resource Fig.12 Development of Cracks Shown in Both Methods (2) Convergence better than other methods (① and ③) (3) Settable calculation time because the number of mod- In the element deletion method, the elements shown els to be calculated is limited. in red are deleted, and in the damage path method, the area in black loses its rigidity. The areas shown in black ③Damage path method in both methods render cracks. In the methods ① and ②, a crack is rendered by We will explain characteristics of those analysis meth- deleting a part of mesh. However, state of stress similar ods, the element deletion method and the damage path to a crack can be rendered by decreasing rigidity of ele- simulation (Fig. 13). ments, even without deleting them. Mr. Mukai and oth- ers have made public this method of decreasing element ①Dynamic element deletion method rigidity and call it damage path simulation(7)(8). The follow- Mr. Terasaki and others tried a method(5) with solder ing three are the advantages of this method. pumps, using VOXEL elements and micrifying the ele- (1) Relatively independent of mesh size ment mesh size to the actual crack size level. When a (2) No need for large computational resource strain or stress of an element reaches a predetermined (3) No need for creating special mesh value, the element is deleted, which represents the devel- On the other hand, we tested this method and found opment of an actual crack. The following two points are its possible tendency toward less convergence compared considered as the advantage of this method. to the other methods. (1) Accurate prediction of development of a crack with- Evaluation area out special preconditions (Solder pump) Damage path (2) Good convergence compared to the damage path (Crack developing method (discussed later) path) ①Start of crack On the other hand, these two points are the disadvan- tages of the method. (1) Strong dependency on mesh size and even and Cracked element minute mesh being required ②Intermediate (2) Huge model size and much time for calculation required for minute and even mesh
②Static element deletion method ③Final breaking Mr. Igarashi and others propose the method(6) where elements are deleted like the method ① for life prediction for chip resistors. What makes this method different Dynamic element Static element Damage path from the method ① is that multiple models where ele- deletion method deletion method method ments have been already deleted by engineers are pre- Fig.13 Analysis Examples of Those Methods pared in advance. The Coffin-Manson Law, which is used to predict start of a crack, predicts accurately the cycle where a crack starts with a certain length (for example, 50μm). The length of a joint is divided by the standard
FUJITSU TEN TECH. J. NO.33(2009) 44 Approaches to Technology of Thermal Fatigue Life Prediction of Solder Joints
Model scale Element Dependency Number Calculation Analysis Method Estimation of number of element of Test results shape of element in the time accuracy length of 100μm size models Dynamic Longest Not adopted Changes Requires a element VOXEL 100 Yes 1according to Good largescaled deletion crack length equipment Static Short Depend on Not adopted 10 or Changes Requires prediction element Tetrahexa 8 Yes the number Crack rate more according to of development Actual Actual deletion No. of models of models path of cracks Analysis Analysis Long Depend Adopted Damage Tetrahexa 4 Less 1 Changes on mesh Possible to predict path according to development of crack length size cracks Number of cycles Fig.14 Benchmarks of Those Methods Fig.16 Transition of Crack Rate
4.3.2 Our Analysis Method of Crack Development 4.3.3 Case Examples We benchmarked each of those two methods, one of We use the analysis of crack development as the eval- which is the element deletion method where development uation guideline for the BGA or other parts in which a of a crack is rendered by deleting elements and the other crack cannot be located by appearance. Fig. 17 shows is the damage path method where a virtual crack devel- the simulation of a crack development in a BGA package ops by decreasing rigidity of elements, and decided to used for in-vehicle products. Being tolerant to change in adopt the damage path method because model scale less temperature environment in vehicles, the terminals out- depends on the size of a crack and we can use the con- side a package serve as NC to prolong life of terminals ventional model. And then, we developed our own analy- inside. With the conventional analysis method, we could sis method for our products (Fig. 14). predict which solder bump would be broken first, but we In concrete, we established our analysis program that could not predict the second bump and later. Therefore, runs the process from (1) to (3) automatically as user sub- we ground down the package in the sectional area from routine on ABAQUS, a general structure-analysis soft- the outside to evaluate the BGA and other parts. ware (patent application number 2006-12272). (Fig. 15 However, the new analysis method of crack development shows its algorithm.) can locate the bump that breaks second or later and help (1) Adjust the parameters of the Coffin-Manson Law to us find problems in the development phase. proper values according to the actual rate of crack development. (2) Calculate damage state of infinite elements based on the Coffin-Manson Law and decrease rigidity of the 2nd 1st Terminal for elements dynamically based on the calculation. NC releasing heat (3) Regard the elements of which rigidity decreases as an
area where a crack develops, and show them graphi- To predict the life cally. Automatically calculate the length of the ele- of NC part ments and crack rate in each cycle.
Start of crack Analysis of elastic-plastic creep 4th Strain Signal terminal amplitude Conventional analysis of Analysis of crack development solder life Calculation in Fig.17 Application Example to BGA Intermediate (Coffin-Manson Law)
(Linear damanage 5. Conclusion theory) 5 Conclusion Decease in Young's modulus (Virtual crack) This paper explained the history and use of FUJITSU TEN analysis method for life of solder joints, pointing out Crack rate ＜○○% SMD of electronic part, lead-free solder, high-density of End parts as the reasons for our effort. End Here, we would like to extend our sincere apprecia- tion to those who gave kind support for us to gain solder Fig.15 Analysis Algorithm properties and limit of parts' life, and verify simulation We gained good correlations as shown in Fig. 16, by results, etc. With these, we contribute to establishment of using the established program. design guideline and efficient evaluation. These days, in-vehicle electronic devices are often installed in the engine compartment, which puts those devices in severe conditions in terms of temperature. Among those devices, some handle high current and the
FUJITSU TEN TECH. J. NO.33(2009) 45 FUJITSU TEN TECHNICAL JOURNAL
number of such devices is increasing. We are now devel- 【Terminology】(9)(10)(11) oping another analysis method because we are concerned ECU (Electronic Control Unit) that reliability of solder joints will become an issue due to Abbreviation of Electronic Control Unit transitional self-heating in ECU in the high temperature Finite Element Method (FEM) environment. The finite element method is a method to analyze We hope to contribute to the development of technol- stress and strain by solving the displacement equation ogy that ensures reliability of miniaturized in-vehicle for a structure deformed by an external force with products by establishing analysis methods employing the structure modeled in an aggregate of small ele- thermic fluid and stress, and developing analysis methods ments. for joints that will continue to be smaller. SMD Abbreviation of Surface Mount Device References The SMD is electronic parts that can be mounted on (1) Yonemoto, Takashi, Shinichi Sugiura, Hiroyuki printed wiring boards with surface mounting technol- Wakabayashi, Takafumi Yasuhara, Takuya Watanabe ogy, such as IC, resistors, capacitors. and Shin Koshiji, Engine Control ECU Development BGA (Ball grid array) for '05 Standard Model, FUJITSU TEN TECHNICAL The BGA is a package with one face covered (or part- REPORT, Vol.46, 2005, Page 8-14 ly covered) with electrode balls (bumps) of solder in a (2) Qiang YU and Masaki Shiratori, Maikurohanda grid pattern. Since electrodes can be provided on the Setsugobu no Kyoudo-shinraiseihyouka to BGA more than the QFP, and leads do not extend Shinraiseisekkei ni Kansuru Shomonndai (Problems outside, the BGA is one of the solutions to producing a related to reliability evaluation of strength and relia- miniature package. However, it is difficult to inspect bility design of micro solder joints), Japan Society of soldering quality by appearance. Mechanical Engineers, RC144 Electronic Packaging SOP (Small Outline Package) niokeru Maikurosetsugo no Shinraisei ni Kansuru The SOP is a small plastic molded package with "gull Kenkyu Bunkakai Hokokusho, 1998, Page 77-154 wing" leads protruding from the two long sides. (3) Kazuhiro Nishimura, Masanori Motegi, Hidehisa Sakai, QFP (Quad Flat Package) Keiko Kawase, Seiki Sakuyama, Kazuyuki Imamura The QFP is a small plastic molded package with "gull and Katsufumi Morimune, Lead Free Solder BGA wing" leads protruding from the all four sides. Fatigue Life Analysis, Journal of Japan Institute of Creep Electronics Packaging, Vol.4 and Vol.5, 2001, Page 416- The creep refers to the property where a strain on a 420 material becomes larger with time as a result of expo- (4) Hidehisa Sakai, Katsufumi Morimune and Toshiko sure to certain stress. The creep is more severe when Saitoh, 212 A study on lead free solder QFP Package the material is subjected to heat. The expression "less Design Parameter Evaluation, The Computational stretchable" is comparison to this property. Mechanics Conference, 2005(18), Page 185-186 GUI (Graphical User Interface) (5) Takeshi Terasaki and Hisashi Tanie, A Method Based The GUI is a user-friendly interface where much on a Crack Propagation Model for Predicting Solder information is provided in graphics to users and most Fatigue Life, Transactions of the Japan Society of basic operation can be done with such a pointing Mechanical Engineers. A, 74(740), 2008, Page .574-582 device as a mouse. (6) Kazuhiro Igarashi, Handa Setsugobu no Hirojyumyou Mesh Suitei Houhou oyobi Sochi (Method and Equipment for The mesh refers to an element in FEM. Fatigue Life Prediction of Solder joints), Patent Tetramesh Number 4179248 The tetramesh refers to the tetrahedral mesh. An (7) Minoru MUKAI, Tomoko Monda, Kenji Hirohata, automatic mesher can easily show even a complicated Hiroyuki Takahashi, Takashi Kawakami and Kuniaki shape in tetrahedral mesh. Compared to the hexago- Takahashi, Damage Path Simulation of Solder Joints, nal mesh, the number of elements is larger and Transactions of the Japan Society of Mechanical accordingly difference is wider. Engineers. A, Vol.72(721), 2006, Page 1364-1369 Hexamesh (8) Mukai Minoru, Damage Path Simulation of Solder The hexamesh refers to the hexagonal mesh. Joints in QFP, The Computational Mechanics Compared to the tetramesh, it is difficult to automati- Conference, 2005(18), Page 183-184 cally show an object in hexagonal mesh. However, (9) CD-adapco JAPAN CAE glossary, CD-adapco JAPAN the number of elements is smaller and difference is http://www.cdaj.co.jp narrower compared to tetramesh. (10)WIKIPEDIA The Free Encyclopedia, Wikimedia VOXEL mesh Foundation, Inc. http://ja.wikipedia.org The voxel mesh refers to the regular hexagonal grid (11)e-Words A Glossary of Computers and Internet, element. An automatic mesher can show an object in Incept Inc. http://e-words.jp the voxel mesh. In the case of a complicated shape, it
FUJITSU TEN TECH. J. NO.33(2009) 46 Approaches to Technology of Thermal Fatigue Life Prediction of Solder Joints
requires huge equipment. The number of elements of voxel mesh is far bigger than the one of tetramesh or hexamesh. As a result, it is accurate. Monte Carlo Method The Monte Carlo Method is a method employing ran- dom numbers in simulation or calculation. As mentioned in the text, we used the RAND func- tion in Excel to calculate variables of characteristic formula in our approach.
Profiles of External Writer
Hidehisa SAKAI Entered in FUJITSU LIMITED in 1990 and engaged in the simulation related to the mounting technology. Since 2007 when TECHNOLOGY CENTER was deconglomerated, has been in FUJITSU ADVANCED TECHNOLOGY HPC CENTER. Currently PROJECT MANAGER of FUJITSU ADVANCED TECHNOLOGY HPC CENTER.
Profiles of Writers
Katsufumi MORIMUKE Takashi YONEMOTO Entered the company in 1992. Since Entered the company in 1987. Since 1999, has engaged in the development 2009, has engaged in the ECU design of mounting technology for car elec- for engine control by way of mounting tronic devices by way of the develop- and structure development of car elec- ment of car electronic devices. tronic devices. Currently Department Currently in the Development Dept. 2, General Manager of Control Solution Research & Technology Development Dept 1, Control Solution Div, AE Div, AE Grp Grp
FUJITSU TEN TECH. J. NO.33(2009) 47 INTRODUCTION OF OVERSEAS TECHNICAL CENTER Introduction of FUJITSU TEN SOLUTIONS PHILIPPINES, INC. (FTSP)
Akira IKEZOE Hisashi FUJII Catalino GAYOSA III Rommel P. TUAZON Philip S. TENORIO Lez T. LEE
FUJITSU TEN SOLUTIONS PHILIPPINES, INC. (FTSP)
Metro Manila
One (1) hour drive from FTSP to FTCP
FUJITSU TEN CORPORATION OF THE PHILIPPINES (FTCP)
Abstract
Fujitsu Ten Solutions Philippines, Inc. (FTSP) has made creative efforts in terms of process, organization, human resource development, etc., in order to overcome the problems in offshore development such as, language, culture and location. Fujitsu Ten Limited (FTL) has entrusted FTSP with the major design projects that enabled FTSP to expand its business from software development including embedded software, information systems, simu- lation tools, etc. to hardware design including audio and AV equipment. Thus, FTSP's contribution in the devel- opment of Fujitsu Ten's software and hardware products is expected to increase significantly. Also, FTSP conducts the suggestion-styled development, cooperating with Fujitsu Ten group of companies globally. In 2007, FTSP achieved maturity Level 3 in CMMI®(*) that software companies in the world use as an indica- tor of development model with a goal of realization of higher quality. We introduce our efforts in this article. (*CMMI is an abbreviation of Capability Maturity Model Integration, and is registered trademarks of Software Engineering Institute (SEI), Carnegie Mellon University.)
48 Introduction of FUJITSU TEN SOLUTIONS PHILIPPINES, INC. (FTSP)
1 Introduction1. Introduction opment and improvement of business quality and produc- tivity. FTSP shows the results by setting up the Labor- 1.1 Background of Establishment Management Council (hereinafter referred to as LMC) Fujitsu Ten Solutions Philippines, Inc. (FTSP) was and improving the labor environment with total coopera- established in 1999 as the first offshore design company tion between labor and management. In 2007, the development in Fujitsu Ten group of companies, It start- Department of Trade and Industry of Philippines recog- ed as a "software business project department" under nized the LMC's activities and gave FTSP the award of Fujitsu Ten Corporation of the Philippines (FTCP) in "Model Company" (six companies won the same award in 1998. The capital of FTSP is composed of 80% of Fujitsu Philippines) in the "category of labor-management cooper- Ten Limited (FTL) and 20% of FTCP, and its office is ation." It is not an exaggeration to say that FTSP has located in the center of the business area of Ortigas become a leading company by this award. In 2008, FTSP Center, Pasig City, Metro Manila. also won the same award in the "category of CSR (Corporate Social Responsibility)."
2.2 Business Introduction We introduce the business of each department. 2.2.1 Car Infotainment Department (CID) The CID develops the embedded software for audio- visual (AV) equipment, and support tools and utilities for design efficiency and automatic testing of products. Fig.1 Main Entrance Reception AV Development Section mainly develops the embed- ded software for Original Equipment Manufacturer (OEM) and commercial car audio products under the 2 Business2. Business Outline Outline and and Its IntroductionIntroduction ECLIPSE brand. The CID works on all software develop- ment of non-base audio products. From 2007, it has par- 2.1 Business Outline ticipated in the base software development. FTSP began with 32 employees who worked on the Multimedia Development Section develops the soft- development of embedded software and simulation tools ware for screen display and operation, called HMI of automotive equipment including electronic control, (Human Machine Interface) of audio equipment with dis- audio and multimedia products. In 2004, we changed play, sharing the development with FTL. Most of the "Software" of the corporate name to "Solutions" predicting development process ranging from creation/editing of our business expansion in the future. In 2005, we moved screen and integration of screen transfer to actual opera- in hardware development and design to combine with tion verification is implemented by simulator established software development. This enables FTSP to do business in PC. The software designed at each department is in hardware design of electrical circuits, mechanical parts, shared immediately, and its operation can be confirmed etc. of audio-visual (AV) equipment. without actual equipment. Therefore, the problems in off- In 2006, we started the information system develop- shore development including the distance and time con- ment business. Since then, we have developed the busi- straint can be overcome and the development cost is ness application software for the global system used in reduced. production sites of Fujitsu Ten group of companies. Tool Section develops the tools to improve the design As of April 2009, FTSP has 375 employees (including efficiency of car AV products. Since 2006, it has collabo- 10 Japanese employees) that comprise the development rated with in-house hardware design department and and management departments (refer to Fig. 2). other subsidiaries of Fujitsu Ten. Quality Support (QS) Section evaluates the software developed in the CID before its release to customers within Fujitsu Ten group of companies. In the evalua- tion, the basic software performance test of products is based on the specifications. The QS section applies the quality engineering principles such as Pair- wise/Orthogonal Array testing, etc. to implement an effi- 21 members 109 members 130 members 61 members 46 members 8 members cient and effective evaluation method. In 2008, the sec- Fig.2 Organizational Structure of FTSP tion began the static verification (Source Code Inspection) of software developed in the CID and FTL. The section In the Philippines, the management principles, which detects the critical defects of software and thus con- came from the combined efforts of labor and manage- tributes to the improvement of the quality of Fujitsu Ten ment keeps the balanced relationship between both par- products. ties. This setup enables the realization of business devel-
FUJITSU TEN TECH. J. NO.33(2009) 49 FUJITSU TEN TECHNICAL JOURNAL
2.2.2 Hardware Design Department (HDD) executing all program in all modules. AED improved the The hardware design of car audio has a long history systemization to analyze the control specification (logical- in the Philippines. In 1995, FTCP started the develop- ly break down to a minimum testing unit) and applies the ment of a cassette tuner for the Philippine market. After testing method specified for each unit (prescribed test that, FTCP undertook the development of commercial sequence including boundary-value inspection and condi- products for places other than the Philippines and the tion estimation). In integration and system testing AED development of OEM audio. In February 2005, hardware promotes the systemization using the same method with design in FTCP was transferred to FTSP to form the unit testing to obtain the stable quality that is not depen- current HDD. dent on the skills of the person in charge of the testing. Currently, the HDD was commissioned by the engi- Currently, the number of members in charge of unit neering departments in FTL to develop the electrical and testing of powertrain and security field is twice as many mechanical design of audio products. HDD focuses on the as that of FTL. This makes AED an indispensable part- development of the non-base models, but also has systems ner for FTL. Also in 2008, AED began the unit and inte- to cover most of the development processes ranging from gration design in the powertrain field to expand its busi- the conceptual design until mass production. ness. HDD aims in the future to be an independent design In the field of development support tool, the AED has and development center in the ASEAN region. This significantly contributed to the improvement of quality includes the expansion of the number of development and productivity of embedded software by promoting the models and audio development that aims for the higher development of management database of CRAMAS, technological level that will become the platform for the CAPAS, and mass production design information, and by "Global Development Expansion Program" of FTL. expanding the range to support the mass production Therefore, the close cooperation among Fujitsu Ten process. group of companies within the ASEAN region is Since 2007, the AED has expanded the process that required. To start this activity, HDD actively pursues the FTSP undertakes, and has developed new tools in the establishment of the "ASEAN Technical Network cooperation with FTL, aiming for "centralization of tool System" and the "ASEAN Local Procurement Promotion development field" as a solution for reducing the tool System." In 2008, outside of the ASEAN region, HDD development cost and ensuring the development person- combined the Europe-specific technology of Fujitsu Ten nel of FTL. Specifically, the AED conducts the spiral (Europe) GmbH (FTEG) and the technology of mass pro- development, using SILS-related tool (Software In the duction model of FTSP and achieved a new prototype Loop Simulation), M-CAPAS (static verification tool of development of audio module with the cooperation of model base development), etc. that would reform the both companies. mass production process, to extract the user's needs and HDD also aggressively conducts early introduction of verify them. new system (CATIA of mechanical 3D-CAD, Mecia of In the future, FTSP will educate analysts who can electrical CAD, etc.) to improve its members' design extract the user's requirements accurately to advance the development skills. HDD introduced "Mecia" to the mass further centralization of development, and will make production design on a trial basis earlier than FTL, coop- efforts to contribute to FTL as a system integrator. erating with Information System Group of FTL. The 2.2.4 Information Systems Development (ISD) usability of "simultaneous parallel design function" is In 1996 and still part of FTCP, ISD has developed and Mecia's feature that was demonstrated by HDD. introduced business operation systems for overseas facto- 2.2.3 Automotive Electronics Department (AED) ries. Presently, in collaborative development with the The AED develops the embedded software for auto- Information System Group of FTL, it plays a vital role in motive electronic control device of the powertrain and introducing this system to factories in China (Tianjin and security control, and develops the development support Wuxi), Thailand, and Mexico as well as in FTCP, in tool including the parser, CAPAS, (Computer Aided achieving the visualization and operational efficiency. Program Analysis System), simulator for system develop- Also ISD strengthens the support after implementa- ment evaluation, CRAMAS, (Computer Aided Multi tion of the system and promptly responds to the system Analysis System), etc. troubles arising at overseas production factories and In the embedded software development, the testing requests for improvement over the phone or by virtual and design works are conducted. The testing accounts network computing. For major upgrade of the system, for 80% of the embedded software development. Errors the members of ISD go to the overseas production facto- must not be built into the design and at the same time ries and give a thorough response ranging from introduc- the overlooked errors must be avoided in the testing. tion preparation to infrastructure set-up and establish- The AED sets the high goal of zero overlooked defects ment of operation. As for the functional improvement, and makes efforts to ensure the quality. the ISD reflects 500 cases of major/minor improvements For example, in the unit testing, AED verifies per year, and improves the system to respond to user's whether the required specification is satisfied or not by requirements accurately and promptly.
FUJITSU TEN TECH. J. NO.33(2009) 50 Introduction of FUJITSU TEN SOLUTIONS PHILIPPINES, INC. (FTSP)
The introduction of the high productivity develop- the difference in understandings between them and actu- ment tool (MAGIC: MAGIC Software Enterprises), which al development site. realizes a statement-less environment by original frame- As a result, we officially achieved maturity level 3 in work, enables an extremely-small group (23 developers) in CMMI® (standard process is defined and managed) in all the development of a mission-critical system to carry out software developments of the CID and AED in July 2007. the development, upgrade the infrastructure, assist the In the automotive-related industries, there are few com- introduction, and provide the support. On the quality panies that have achieved the maturity level 3. side, compared with the previous year, the outflow of (Refer to the following website of SEI (Software defects was reduced by 60% or more by process improve- Engineering Institute.) http://sas.sei.cmu.edu/parts.aspx) ment and test process establishment that conform to The establishment of the standard process allows us CMMI® by which the CID and AED got certified in 2007. to sample the data of process and quality data quantita- Since October 2008, as a part of new business devel- tively. We control the process by using these data as a opment, the ISD has undertaken the system improve- feedback in the project monitoring and process improve- ment project of "APROS" that is PDM (Product Data ment. Specifically, we assess the correlation between Management) system of Fujitsu Ten used by the engi- review manhours and number of defects and use it in the neering design departments of FTL. In this project, ISD quality gate pass before proceeding to the next process. realizes the collaborative development with FTL by Although quality of software is being improved yearly "sharing of program source between sites." Also in a to answer the expectation of FTL, it is necessary to FTL-led "next-generation mission-critical system concept," reduce the volatility in quality, analyze the data statisti- ISD is greatly expected to contribute to the overseas pro- cally evaluate, the process, and improve the process duction system. maturity. In the future, the ISD will positively acquire new Also we pursue the creation of an organizational technology including the data linkage between overseas structure that allows predicting the quality of a project, subsidiaries to promote the information utilization in and preventing and avoiding risks. These activities will group view, and will continue to contribute to the infor- lead to the achievement of CMMI® level 5 (state of quali- mation system development of Fujitsu Ten group of com- ty is stable when process is optimized) in the near future. panies. In addition, we obtained the certification of ISO/TS16949:2002 in 2006, as a remote location of FTCP. Efforts3. Efforts in Offshore in Offshore DevelopmentDevelopment 3 3.2 Development of Human Resources As the first company of offshore development in 3.2.1 SUMO Concept Fujitsu Ten group, FTSP greatly contributes not only to Generally, Filipinos have a high motivation for self- reduce the development cost but also to solve the short- realization, so an enriching education system and clear age of development resources through the project devel- career plan will keep and improve their motivation. opment. In this article, we introduce how we have The basic idea of the human resources development worked and developed human resources in offshore. in FTSP is "staff growth = company's target achieve- ment," and to realize the early contribution to the compa- 3.1 Philosophy of Quality Assurance and CMMI® Model ny, thorough trainings are given to the employees as a To succeed in an offshore development, visualization business investment to improve their skills. We clearly and standardization of the development process is impor- demonstrate the medium/long-term career development tant as well as provision of software with stable quality to plan to staff, using the skill matrix and training road map the consigners. This leads to "own-process completion" to promote the continuous career growth. that Fujitsu Ten group pursues. Aiming for SUMO (Skill Up Management Oriented), In 2005, having a philosophy of "the quality of a sys- we pursue the educational system to develop with trinity tem or product is built-in during its development (Fig. 3) of "total education (introducing technology), process," we adopted CMMI® as a global standard and a department education (supplementing technology), self- yardstick among the leading software companies to evalu- learning (sharing technology)." We coach subordinates ate the degree of attainment of our philosophy. using "handed-down education," focusing on in-house We promoted the activity thru Software Engineering trainers to enhance and expand the training effectiveness. Process Group (SEPG) that develops the process, and Simultaneously, this gives a chance to the trainers for Software Quality Assurance (SQA) that monitors the self-fulfillment and leads to an enhanced training content compliance to the standard and audits the deliverable. and highly-motivated training. This activity became prevalent in FTSP by providing As of the end of 2008, in-house training course CMMI® education for each level to make members con- includes 33 courses such as UML, C++, etc. with 40 regis- scious of the importance of the process, and by having tered trainers. In the future, we will enrich the educa- project leaders or higher level people experience SQA for tional curriculum focusing on the development of mid- three months in quality assurance department to bridge level employees.
FUJITSU TEN TECH. J. NO.33(2009) 51 FUJITSU TEN TECHNICAL JOURNAL
Trinity education structure to complete in-house automation for design evaluation of audio products, capi- talizing on FTSP's strengths that can realize the collabo- Total education (introducing technology) ration of hardware and software easily. First of all, we started the development project of a system to automati- cally evaluate the radio performance and sound charac- teristics (curve data) that are the basic functions of audio Self-learning Department education (sharing technology) (supplementing technology) products. The HDD worked on the creation of specifica- tion and actual equipment verification of system, and the CID worked on the development of software to control Fig.3 Educational Concept the measuring equipment and applicable system. This led to the completion of basic system of APES within a 3.2.2 Japanese Language Teaching and In-house Recognition period of one (1) year. Introduction of APES allowed the System of Japanese Language automation of items that have been measured manually, The importance of Japanese language is increasing and improved efficiency by up to 70%. We have expand- with the business expansion of FTSP. ed the functions to achieve the complete automation to Previously, we gave education focusing on daily read- evaluate the design of audio products by 2011. ing and writing to pass the Japanese language proficiency We started to propose this system as a solution for test (JLPT) that is used widely abroad. But now we shift- design efficiency not for cost reduction by offshore devel- ed the main emphasis to "Japanese-language ability to be opment to FUJITSU TEN RESEARCH&DEVELOP- used in practical business" that focuses on the assumed MENT (Tianjin) LTD. (FTRT), and now this system is scene and necessary terms in the operation with FTL, used in the design departments of FTL and design sites and established in-house Japanese language certification in FUJITSU TEN (THAILAND) COMPANY LIMITED system. In Philippines, there is no leading Japanese soft- (FTTL). ware development company that has an original Japanese This proposal led to the discovery of needs for a tool language certification system. to improve efficiency that each site required, and we In 2007, we started the lessons based on our original obtained the business of development of the diagnosis FTSP training material, and we confirm the achievement record data acquisition analysis tool (D-TRACER) for level by the bi-annual certification examination. audio equipment for 2009 model from FUJITSU TEN Moreover, based on the SUMO concept, the senior-level TECHNICAL CENTER, USA, INC. (FTTC). person makes use of his own experience to teach the In addition, the Field Audio System Test (FAST) tool beginner-level persons. In 2008, we introduced this has the function to record the measurement data of video Japanese language certification system in the new and radio simultaneously so that the designers can evalu- employee training, and set it as criteria for employment ate the state of site easily even while they are at develop- regularization. In job requirements, we defined the ment workbench. Thus, FTSP's new business has been Japanese language level and clarified its importance. expanded by the proposal of solutions based on the inde- As of 2008, 88 members (beginner), 16 members pendent development. (basic), and 1 member (intermediate) were certified, and this comprises 31% of all designers. 4.2 Strategic Audio Project for ASEAN One of the significant results of FTSP in 2007 and 4 Solution4. Solution Suggestion Suggestion 2008 was "strategic audio project for ASEAN" that was developed mainly by the HDD. In this project, we orga- 4.1 APES (Automatic Product Evaluation System) nized the project team consisting of the young engineers of electrical circuit design, mechanical design, and soft- ware design, regardless of their departments, and we completed the whole process ranging from development planning to design and prototype manufacturing in FTSP. Also we greatly contributed to the establishment of new networks, including the cooperation with Fujitsu Ten Group of companies in ASEAN region and with local sup- pliers for the local procurement of parts. In this project, we carried out the planning with a concept of "low-cost and compact audio," looking years ahead and with out-of-the-box thinking. As a result, we Fig.4 Outline of APES developed "deck-less audio" that is equipped with the digi- tal interface with the new music media such as USB In 2006, as an improvement activity for hardware (Universal Serial Bus) memory and SD (Secure Digital) design environment, we started to make efforts toward card, thereby removing the CD deck mechanism.
FUJITSU TEN TECH. J. NO.33(2009) 52 Introduction of FUJITSU TEN SOLUTIONS PHILIPPINES, INC. (FTSP)
This "deck-less audio" achieved the downsizing by Like DRBFM (Design Review Based on Failure 58% (volume) and 45% (weight) relative to the existing Mode), the TDD is a method to improve the efficiency of CD tuner, and also achieved the reduction of parts by development by extracting the unclear points of specifica- 45% and of cost by 30%. This figures show the contribu- tions or problems for design to reduce the setback in the tion to the environmental protection as well as to the development process. company's profit. As shown in Fig. 6, we combined the TDD with exist- ing process and applied it to three projects in the devel- opment of mass production support tool.
Existing process TDD Part C Requirement Reception Part B analysis inspection Part A Test case creation Basic design System inspection Code implementation Detailed design Unit inspection Carrying out Unit inspection for each part Code implementation/review (A, B, and C) Coverage verification Changed point Less than Review Fig.5 Audio without Deck Mechanism Repeat until implementation 100% of all functions END
After we informed the overseas subsidiaries of Fig.6 Test Driven Development Fujitsu Ten group as well as FTL about the achievement of "strategic audio project for ASEAN," we made a pre- As a result of this application, we reduced the man- sentation as a concept model at an exhibition for cus- hour of code implementation by 10%, the man-hour of sys- tomers held in ASEAN region and Australia in 2008 and tem inspection by 25%, and total man-hour by 20%. Also we impressed our customers. after we informed FTSP and FTL about this achieve- Now we continue the promotion to commercialize this ment, the introduction of the TDD is increasing as an "strategic audio project for ASEAN" so that it will not effective development method. end up as just a concept model. Furthermore, we have launched an initiative of "next-generation strategic model" Conclusion5. Conclusion to continue our quest for new functions, technology, and 5 market. We are aiming to be a company that can deliver high performance consistently and pursue the thorough quali- 4.3 Efforts toward Test Driven Development ty and productivity by the business concept integrating In the development of a design support tool of the with Fujitsu Ten group of companies. It may be said AED, FTSP tries and verifies the latest programming that "solution" is a "value," but just holding this solution in method and software evaluation method. In 2008, the FTSP would not be valuable. The added value is AED applied the Test Driven Development (hereinafter increased and "value chain" is formed by proposing and referred to as TDD), which recently attracted attention in sharing the solution with the Fujitsu Ten group of com- the IT field to the actual project. panies. Setting this as the business vision, FTSP will In the TDD, the designers create the test case and work toward the establishment of its position as a "global then implement the code to pass the test. Repetition of solution company" that is at the core of the "value chain." this procedure completes the program, so it is called "test driven development."
Profiles of Writers
Akira IKEZOE Hisashi FUJII Catalino GAYOSA III Entered the company in 1982. Since Entered the company in 1991. Since Entered FTCP in 1997. Transferred then, has engaged in the software then, has engaged in the electrical cir- to FTSP in 1999. Since then, has development for powertrain control. cuit design and development of car engaged in the software development Now President of FTSP, appointed audio. Assigned at FTCP in of the development support tool for in July 2005. September 2004, and then transferred automotive multimedia product. to FTSP. Currently the Department Currently the Section Manager of Manager of the HDD. Quality Assurance Section and the CID Quality Support Section. Rommel P. TUAZON Philip S. TENORIO Lez T. LEE Entered FTCP in 1998. Transferred Entered FTCP in 1998. Transferred Entered FTSP in 2005. Since then, to FTSP in 1999. Since then, has to FTSP in 1999. Since then, has has engaged in the electrical circuit engaged in the software development engaged in the software development design and development of car audio. for powertrain control. Currently the for car audio development support Currently assigned at HDD. Section Manager of the AED. tool. Currently the Section Manager of CID Tool Section.
FUJITSU TEN TECH. J. NO.33(2009) 53
FUJITSU TEN CORP. OF AMERICA (FTCA) 天津富士通天電子有限公司 FUJITSU TEN de MEXICO.S.A.deC.V. (FTdM) Los Angeles Operation (LAO) Rushville Indiana Operations (RIO) (TIANJIN FUJITSU TEN ELECTRONICS CO.,LTD.)
Overseas Offices, and Subsidiaries Telephone FAX
北京事務所 郵編 100004 +86-10-6505-3535 +86-10-6505-3534 (Beijing Office) 北京市朝陽区建国門外大街１号 中国国際貿易中心1座2710-2711室 (Room 2710-2711, Building1, China World Trade Center, No.1, Jian Guo Men Wai Avenue, Chao Yang Distric, Beijing, China 100004) FUJITSU TEN CORP. OF AMERICA (FTCA) Michigan Operations (MIO) 47800 Halyard Dr.Plymouth, Michigan 48170, U.S.A. +1-734-414-6650 +1-734-414-6660 Rushville Indiana Operations (RIO) 616 Conrad Harcourt Way, Rushville, Indiana 46173, U.S.A +1-765-938-5555 +1-765-932-3257 Los Angeles Operations (LAO) 19600 South Vermont Avenue, Torrance, California 90502, U.S.A. +1-310-327-2151 +1-310-767-4355 Texas Distribution Center (TDC) 6412 South 36 Street, McAllen Texas 78503, U.S.A +1-956-686-8601 +1-956-686-3481 FUJITSU TEN de MEXICO.S.A. de C.V.(FTdM) Avenida Industrial Del Norte, Manzana 8, Lote 2, Parque Industrial +52-899-921-8700 +52-899-929-2260 Del Norte, Reynosa, Tamaulipas, Mexico FUJITSU TEN ESPAÑA, S.A. (FTESA) Pol. Ind. Guadalhorce, C/Cesar Vallejo, 16, 29004 Malaga, Spain +34-95-213-3000 +34-95-213-3003
FUJITSU TEN (THAILAND) COMPANY LIMITED (FTTL) Bangkok Head Office (BKO) 88 Dr.Gerhard Link Building. 5th Floor, Krungthepkreetha Rd., +66-2-704-4979 +66-2-704-4980 Huamark, Bangkapi, Bangkok 10240, Thailand Rayong Factory (RAF) 253 Moo 11, Rojana Industrial Park, Bankhai-Banbung Rd., +66-3-896-2025 +66-3-896-2031 T. Nongbua, A. Bankhai, Rayong 21120, Thailand 天津富士通天電子有限公司 郵編 300457 +86-22-2532-8520 +86-22-2529-0778 (TIANJIN FUJITSU TEN ELECTRONICS CO.,LTD.) 天津経済技術開発区黄海二街5号 (No.5 Huanghai Second Avenue TEDA, Tianjin, China 300457) 富士通天電子（無錫）有限公司 郵編 214028 +86-510-8866-2288 +86-510-8866-2233 (FUJITSU TEN ELECTRONICS (WUXI) LTD.) (FTEW) 江蘇省無錫市国家高新技術産業開発区新華路19号 (No.19, Xinhua Road, Wuxi National Hi-Tech Industrial Development Zone, Jiangsu, China 214028) FUJITSU TEN CORPORATION OF THE PHILIPPINES 100 South Science Avenue, Laguna Technopark, Sta. Rosa, Laguna, +63-2-793-2900 +63-49-541-0050 (FTCP) Philippines INMOTION AUDIO (AUSTRALIA) PTY. LTD. (IAA) No.5 Southfork Drive, Kilsyth, Victoria 3167, Australia +61-3-9761-7677 +61-3-9761-7520 FUJITSU TEN CANADA INC. (FTCI) 1149 Bellamy Road North, Unit 1, Scarborough, Ontario, M1H 1H7, +1-416-431-9332 +1-416-431-1745 Canada ECLIPSE TD (UK) LIMITED (ETUK) LOE House, 159 Broadhurst Gardens, London NW6 3AU, UK +44-20-7328-4499 +44-20-7624-6384 FUJITSU TEN (EUROPE) GmbH (FTEG) Düsseldolf Office (DUO) Mündelheimer Weg 39, 40472 Düsseldorf, F.R. Germany +49-211-472705 +49-211-4727500 Technical Center Nuremberg (TCN) Südwestpark 23, 90449 Nürnberg, Germany +49-911-30912500 +49-911-30912599 Brussels Office (BRO) Leuvensesteenweg 555/B.6, B-1930 Zaventem, Belgium +32-2-712-8181 +32-2-712-8170 富士通テン韓国株式会社 ソウル特別市鐘路区寿松洞83-1 寿松タワービル16階 +82-2-3787-5377 +82-2-3787-5387 (FUJITSU TEN KOREA LIMITED) (FTKL) (Susong Tower Building 16th floor, 83-1 Susong-Dong, Jongno-gu, Seoul, Korea) 富士通天国際貿易（天津）有限公司 郵編 300074 +86-22-2840-8388 +86-22-2840-8399 (FUJITSU TEN TRADING (TIANJIN) LTD.) (FTTT) 天津市河西区囲堤道125号天信大厦1805 (Tianxin Building 1805, No.125 Weiti road, Hexi district, Tianjin, China 300074) 上海分公司 郵編 200041 +86-21-5292-5225 +86-21-5292-5299 (Shanghai Branch) 上海市南京西路1168号中信泰富広場14階1401室 (Citic Square 1401, No.1168 Nanjing Road West, Shanghai, China 200041) 広州分公司 郵編 510613 +86-20-3877-2926 +86-20-3877-2956 (Guangzhou Branch) 広州市天河北路233号中信広場写字楼1204室 (Room 1204, CITIC Plaza, 233 Tianhe North Road, Guangzhou, China 510613) FUJITSU TEN (AUSTRALIA) PTY.LTD.(FTAL) 89 Cook Street Port Melbourne, Victoria 3207, Australia +61-3-9646-6004 +61-3-9646-8070 FUJITSU TEN TECHNICAL CENTER USA, INC. (FTTC) Head Quarters (HQ) 47800 Halyard Dr.Plymouth, Michigan 48170, U.S.A. +1-734-414-6695 +1-734-414-6660 Los Angeles Operations (LAO) 19600 South Vermont Avenue, Torrance, California 90502, U.S.A. +1-310-327-2151 +1-310-767-4355 Indianapolis Office (IO) 2765 S. Albright Rd. Kokomo, Indiana 46902, U.S.A. +1-765-864-0296 +1-765-864-0317 FUJITSU TEN (SINGAPORE) PTE. LTD. (FTSL) 20 Science Park Road #02-01/03, TeleTech Park, Singapore Science +65-6773-4933 +65-6773-4932 Park2, Singapore 117674 富士通天研究開発（天津）有限公司 郵編 300457 +86-22-5981-5132 +86-22-5981-5134 (FUJITSU TEN RESEARCH & DEVELOPMENT (TIANJIN) 天津経済技術開発区黄海路280号 LTD.) (FTRT) (No.280, Huang Hai road, TEDA, Tianjin, China 300457) FUJITSU TEN SOLUTIONS PHILIPPINES,INC. (FTSP) 8th floor East Tower, Philippine Stock Exchange Centre, Ortigas +63-2-638-5266 +63-2-638-7957 Center, Pasig City, Metro Manila, Philippines
FUJITSU TEN CORP. OF THE PHILIPPINES FUJITSU TEN (AUSTRALIA) FUJITSU TEN (EUROPE) GmbH (FTEG) FUJITSU TEN CANADA INC.(FTCI) (FTCP) PTY.LTD.(FTAL) TECHNICAL JOURNAL http://www.fujitsu-ten.co.jp/
This bulletin was printed with soy ink on recycled paper.