Site: Sumitomo Electric Industries, Ltd.
1, Taya-cho, Sakae-ku
Yokohama, 244 Japan Date Visited: November 17, 1994
Report Author: S. Esener
Sumitomo Electric was founded in 1897 for manufacturing electric wires. Since then, Sumitomo Electric has expanded its wire and cable operations to include power transmission and communication cables as well as optical fiber cables. Sumitomo is clearly the world leader in the electric wire and cable industry. Sumitomo Electric has also expanded into other areas, including steel wires, sintered alloys, and new materials. In 1993, the company reported a consolidated net sales of about $10.7 billion. Of this, 49% were generated from electrical wire and cables (including fiber-optic cables), 5.4% from steel wires, 8.9% from sintered alloy products, 7.8% from brake products, 4.5% from hybrid products, 10.6% from cable accessories and engineering, and 13.9% from new business. Hybrid products include coating of aluminum with sumiflon, infrared materials, and flexible printed circuits. New business includes fiber-optic products (e.g., fiber-optic modules), system products (e.g., traffic control systems), and new materials (e.g., III-V substrates and heat conductive materials such as artificial diamond).
R&D expenditures account for 3.7% of Sumitomo's net sales. Corporate R&D is carried out at four different locations: Osaka, Itami, Yokohama, and Harima. Emphasis is put on creative technology - technology with the potential of bringing about change in support of new business activities. These four locations hosts nine corporate R&D groups:
It should be noted that the Yokohama Research Laboratories are within the Yokohama Works facilities that include manufacturing plants for optical fiber cables, electric wires and cables, communication devices, and hard metal products. The close proximity of the plants and the R&D facilities plays a strong role in the interaction. The R&D effort at Sumitomo is very close to manufacturing. Yokohama Works began operation in 1961, and the R&D laboratories were established there in 1987. The site area is 439,000 sq. meters and employs 2,600 people. Four hundred researchers and technical assistants, 250 of whom are engineers, work in the R&D lab.
Sumitomo Electric holds 50% of the GaAs substrate market, providing high-quality 4 in. wafers. The JTEC panel learned that 6 in. wafers are in advanced R&D, with presently only the center 4 in. usable. Defects in these 6 in. wafers seem to form at the periphery.
On the InP side, Sumitomo Electric produces 3 in. wafers that are commercially available.
Both thin-film and bulk artificial diamond are produced as heat sinks for laser diodes. Ultrahigh-strength aluminum is produced for aerospace and automobile industry by rapid cooling of Al.
Sumitomo foresees optoelectronics as one of the key technologies for its future and invests significant R&D efforts into this topic. The efforts concentrate in four areas: optical fiber, fiber-optic modules, OE devices, and optomechatronics. Out of these four areas, the discussion during the JTEC visit focused on fiber-optic and OE devices.
The OE R&D labs at Yokohama have 10 PhDs, 70 Engineers (BS), and 50 technical assistants. Their major activity is in development; very few people are involved in research. During the recession, twenty researchers were shifted to the system business.
Super Sumilink is an LED-based optical LAN using multimode fibers operating at 1.3 microns wavelength. It provides 120 Mbit/s for 0.8 W power dissipation and sells for $100/transceiver. Sumitomo sells about 10,000 units a month. It is an integrated transceiver with MIC receptacle with single in-line pin configuration with a plastic mold package providing small size and water proofing.
This product went through a typical development cycle to hit the market: The concept was developed in the R&D lab six years before the JTEC visit and transferred to the Yokohama facility by moving the technical person. Sumitomo managers find it critical to match the capabilities of the production plant to the new concept from the beginning. The concept must be designed at research state for mass production within 3 to 6 months. For this particular product, a major issue was the assembly with automatic machines. Customized machines manufactured outside Sumitomo, for example, at Tokyo Electron Labs, were used. But the Sumitomo development team was given the specifications and was in very close contact. (It was mentioned that many small equipment manufacturers exist in Japan. In addition, NEC, which is also part of the Sumitomo keiretsu, has an affiliate company for equipment construction). Now Sumitomo has 30% of the U.S. market with this Sumilink product.
The low-end LAN products were introduced first. In these products, to achieve higher bandwidths the LEDs are replaced by laser diodes, and the receiver is a higher-sensitivity GaAs receiver. At the time of the JTEC visit, Sumitomo was selling 1,000/month at a cost of $1,000/pair. The company could ramp up production to 5,000/month. A 2.4 Gbit/s system was being developed that necessitated new production machines, expected to be ready within about a year after the JTEC visit, that require submicron accuracy.
Since GaAs transistors have higher transconductance, GaAs receivers exhibit higher sensitivity and speed than silicon. For example, at 622 MHz, GaAs receivers require 2.3 dB less signal power than silicon receivers for the same BER. Sumitomo has also GaAs drivers to drive the lasers. It prefers the GaAs solution, since chip costs are negligible compared to assembly and testing costs. The company uses a proprietary screening process by aging to select the transmitters for higher reliability
This effort is carried out in conjunction with NTT's fiber-to-the-home project. Pumping lasers, Er-doped fibers, and fiber couplers are developed at Sumitomo. The company produces 1.48 micron lasers and continues to work on the reliability of .98 micron ones.
The Japanese CATV market is just opening up, and many projects are underway to establish CATV. Depending on MPEG acceptance, Sumitomo researchers feel that analog systems may be favored. They are developing devices to this end, such as narrow-spectrum DFB lasers with good linearity (60 dB). They use Peltier cooling and single-mode fibers. Without distortion precompensation, they feel they can reach 60 channels. The cost of the systems presently are on the order of $10,000. The selection process for the laser chip as well as its distortion degradation during assembly are critical issues to be resolved. Collaboration with NTT in Japan, ANTEC CATV in the United States, and Northern Telecom in Canada is underway.
The idea was conceived at the antenna business section. Its application areas include the use of cellular communication in subways, parking, and so forth. This application requires less bandwidth than CATV applications, but higher frequency of operation (1.9 Ghz). Optical technology offers the potential of transmitting directly at the carrier frequency when compared to coax that presently uses IF. The idea is being tested in the Osaka area.
Sumitomo researchers indicated that they have no significant basic research programs. They conduct some basic system research on multimedia systems with NTT in Japan. In addition, Sumitomo participates in MITI-funded research programs such as the following:
A total of four people work on MITI projects.
The JTEC team's hosts demonstrated several of their accomplishments, including various electrical cables, optical fibers, optical transceivers, artificial blood vessels, traffic systems (result of the Japanese Navigation Project), various tooling and substrates, and specialty materials.