Site:Fujikura Sakura Plant
1440 Mutsuzaki, Sakura-shi
Chiba-ken 285, Japan

Date Visited: December 12, 1994

Report Author: P. Shumate



G. Day
P. Shumate


Dr. Koichi Inada
Director and General Manager, Opto-Electronics Laboratory and Fiber-Optic System Division
Dr. Ryozo Yamauchi
General Manager, Optical Fiber Combinations Dept.
Hiroshi Yokosuka
General Manager, Fiber and Cable Accessory Dept.
Mr. Toshikuni Maruoka
General Manager, Transmission Line Department


Fujikura, established in 1885, is a manufacturer of optical fiber cable, accessories, and devices; electric wire and cable; and electric and electronic components and materials. It has four manufacturing plants in Japan, including the Sakura Plant in Chiba (one of the largest cable-manufacturing facilities worldwide), and overseas operations in the United States, Brazil, UK, Thailand, China, Singapore, and Malaysia. It has research laboratories at Sakura (this report), as well as in Tokyo and Shimousa. In Japan, Fujikura employs 4,287 people (1,650 at Sakura) and has annual sales of approximately $250 billion, of which about 4% is invested in R&D. Fujikura representatives said that this is high for companies whose major product is cables (71% of net sales). The Sakura Plant (5.72 million sq. ft. site area; 1.94 million sq. ft. floor space) originally made copper for telecommunications but now has responsibilities for all telecommunications products, including fiber, fiber and metallic cables, and electronics.

The Opto-Electronics Laboratory, which Dr. Koichi Inada leads, was only about six years old at the time of the JTEC visit. During this period, the lab's mission changed from fabricating only fibers and cables to also fabricating systems. (Fujikura was one of the three cable manufacturers to begin joint development of optical fiber with NTT circa 1975.)


Fujikura representatives discussed each of their product responsibilities relevant to the thrust of this study: T. Maruoka discussed fiber cable design, evaluation, and reliability; H. Yokosuka discussed fiber and cable accessories, including connectors, remote fiber test systems, and optical filters; and R. Yamauchi discussed fibers and fiber-related components, including couplers, fiber amplifiers, and future production systems for fiber. JTEC panelists asked about Fujikura's approach to each of the following:

Optical isolators for EDFAs. Fujikura buys isolators; it does not work on them at this time.

1310 nm fiber amplifiers. Fujikura is pessimistic about amplifiers such as PDFFAs, since they are expensive and not compatible with subscriber-loop applications.

Planar waveguide couplers. Fujikura is just now developing silica-on-silicon couplers. When the volume builds to several hundred thousand devices, their cost will be competitive with fiber-based couplers.

Dispersion-shifted fiber. This is 15% of the plant's current (fiber) product line, made for submarine cable and trunking applications. But the panel's hosts believe that the trunking market is past its peak and that the subcable market is at its peak. They do not see an application for DSF in the subscriber loop or in cable TV.

High birefringence fiber. Fujikura presently makes stress-induced high-birefringence fiber that Corning sells in the United States.

Precision pieceparts and fiber geometry. Fujikura says that ferrules with 0.5 (m tolerance are available, but the use of these to minimize the need for precise alignment steps is limited by the fiber itself. While the specifications for fiber o.d. (outside diameter) is +/- 2 microns, the Fujikura factory limit is +/- 1 microns, and typically +/- 0.5 microns. Therefore, until the o.d. tolerance gets down to +/- 0.1 to 0.2 microns, such piecepart tolerances are not warranted. Fujikura representatives feel that available standards will not support such small fiber tolerances, although one member of the JTEC team, Gordon Day, stated that the National Institute of Standards and Technology (NIST) provides a standard for fiber down to +/- 0.1 microns.

Plastic optical fiber. Some Japanese companies have recently demonstrated a short-distance (100 m) high-bit-rate (2.5 Gbit/s) link with graded-index POF. Fujikura, however, has reservations about GI-POF because of problems focusing the light from the large core onto small, high-speed photodetectors, and believes silica fiber is the superior technology. Although the plant produces cables and equipment for LANs and automotive applications, both of which could be impacted by POF, Fujikura is not currently working on communications networks for home use.

Cable TV products. Fujikura is looking here as an important future market, and it is developing analog cable TV systems for NTT, which may use them to lease transport facilities to cable operators. The timing of this market is difficult. The plant currently manufactures two-way coaxial cable amplifiers (upstream transmission in the 10 - 50 MHz band).

Interactions with NTT. NTT is a major customer for telecommunications products and system prototypes such as those used for the Kansai fiber-to-the-home (FTTH) trial. Fujikura makes FTTH system equipment, including the ONU at the home for video signals. In the development of this equipment, NTT selects a few (typically three to five) manufacturers through an open-tender process for bids to develop the initial products. Currently, as a result of this open-tender process, AT&T was selected as one of the narrowband ISDN system manufacturers, and Siecor was selected as a supplier of high-count optical fiber cable.

Market studies for NTT-proposed products. Since 80% of the telecommunications market in Japan is controlled by NTT, NTT's proposals for new products are taken seriously as an indication of sufficient market opportunity and sales. NTT does not guarantee any specific volumes, but provides an estimate for the market potential.

Relationship with Alcoa Fujikura in the United States. Alcoa Fujikura operates facilities for cabling optical fiber. The company has been buying U.S.-made fiber.

Automation of product lines. Fujikura develops automation as production quantities rise, based on a detailed, formal payback study for each case. Although the primary motivation is to reduce costs (number of workers), improved yield and quality also are considered. Once approved, a central automation department at the plant works with small groups of scientists and engineers in the labs who are familiar with the technologies and procedures, and who themselves can design automated assembly and testing equipment. (This group may have already designed and built semiautomatic equipment for the new product.) Funding comes from the department receiving the assistance.

Technology transfer. As the JTEC panel has found common, Fujikura transfers technology from the R&D labs to manufacturing sites by transferring design specifications along with up to six or seven key people for from three months to a year. Some but not all of these people return to R&D.

Technical discussions were followed by a brief tour of a system lab where JTEC panelists saw a prototype fiber-to-the-home subscriber system for NTT. It combines ISDN delivery at 1310 nm with video delivery at 1550 nm. The video service combines 11 channels (presently) of analog AM-VSB TV between 90 and 400 MHz with 36 channels of FM video (including HDTV) and 60 channels of QAM compressed digital video between 500 and 2400 MHz. The FM and QAM carriers are in the same channel assignments as satellite DBS systems, so that set-top converters can leverage off of available technology.

The optical network unit at the home is approximately 4 in. x 8 in. x 2 in. It dissipates seven watts, and is available in either one- or four-output versions. By using several stages of amplifiers and optical splitting, one analog DFB laser can provide service to 10,000 homes. The prototype uses dispersion-compensated fiber, but Fujikura representatives feel another approach to dispersion compensation is needed because of the difficulty of handling 25 to 33% of the span length in high-dispersion fiber wound on a spool.


Fujikura, a leading supplier of fiber and fiber cables (as well as metallic cables), is expanding vertically to include all associated assembly equipment (e.g., fusion splicers) and test systems (e.g., remote fiber test systems, hand-held computer-based systems, and OTDRs, which are still in the R&D division). The Sakura plant is further expanding into a full line of fiber-related components including connectors, fiber and planar-waveguide couplers, and fiber amplifiers. Importantly, Fujikura is also developing lightwave systems LANs, cable TV, and fiber-to-the-home types. One-third of the fiber-related business is currently in systems. The plant has an active research program, as evidenced by the many recent papers given to the JTEC team. Fujikura has a structured process for technology transfer and line automation.

Published: February 1996; WTEC Hyper-Librarian