U.S. cable television operators are aggressively upgrading or rebuilding their systems with fiber-based HFC networks for delivering analog and digital TV. Importantly, these HFC networks can be equipped for two-way services including telephony. In Japan and Europe, cable television is an emerging market. In Japan, the penetration of cable television is only 5%, compared with more than 62% in the United States. Since cable is in a start-up mode in Japan, new facilities (rather than upgrades or rebuilds) are rapidly being constructed, all based on the HFC rather than traditional tree-and-branch network, because of HFC's technical advantages. In addition, the basic one-way analog subcarrier transmission fundamental to HFC is also NTT's choice for video delivery in its FTTH network, as mentioned above; this fact increases the market for linear lasers (or externally modulated continuous wave [CW] lasers) and optical amplifiers. The cable television industry's transition to fiber is recent and will be strong throughout this decade. Several of the companies the panel visited in Japan make cable television equipment as well as fiber and coaxial cable. One of these, Fujikura, manufactures 2-way coaxial amplifiers and coaxial cables. Another supplier, Toshiba, has recently announced its partnership with Time Warner, US West, and Itochu Corporation to invest $400 million to build fiber-based cable TV systems in Japan. A third supplier, Sumitomo, is similarly working with Tele-Communications, Inc. (TCI), with plans for a comparable investment of $500 million.
Most Japanese manufacturers the JTEC panel visited seem to have a stake in fiber-optic cable television or analog video products, at the system, storage, compression, and/or device level. Companies developing linear, high-bandwidth DFB lasers include Fujitsu, Matsushita, Mitsubishi, NEC, Sumitomo, and Toshiba. Hitachi is the notable exception, having chosen to use an external modulator with a continuously operated DFB laser to avoid chirp-related transmission problems.
NEC has traced the low yield (and therefore high cost) of analog lasers to electric field nonlinearities in the cavity arising from nonuniformities in facet reflectivity. To minimize this effect, NEC has successfully used a partially corrugated DBR-like design. Fujitsu, on the other hand, uses the same MQW buried heterostructure for both analog and digital transmission, but "some parameters are different," depending on the application. Sumitomo has developed laser designs capable of 60-channel transmission without predistortion.
Matsushita has optimized a strained-layer MQW DFB laser for high linearity combined with low bias current. This company has recently attained 100-channel AM-VSB operation with a bias current as low as 17 mA above threshold. Using a longer-channel device (400 microns) and operating at a higher power and bias current (60 mA above threshold), Matsushita has reached 151-channel operation, a new record.
In view of the facts that the United States currently has only two manufacturers of high-linearity analog lasers (AT&T and Ortel), and that six or more Japanese companies are in production of these lasers, as well as carrying out research to understand the fundamental limitations and improve the laser designs, the Japanese are ahead in this area. On the other hand, like the Hitachi system, some U.S.-manufactured systems use continuously operated lasers combined with external, linearized modulators (e.g., supplied by Harmonic Lightwaves and Synchronous Communications); thus, a total of four current suppliers plus the aggressive price forecasts of Ortel suggest that the U.S. demand could be satisfied without scaling up capabilities in this area. If the average hybrid fiber/coax node eventually is reduced to serve only 60 homes for highly interactive broadband services, and if one high-performance linear laser is dedicated to each node to augment broadcast channels with narrowcast as well as switched services, then the total U.S. requirement is only on the order of 2 million devices, plus replacements, over the next ten years. (Note that the upstream laser is a lower-performance device and does not raise this number. If telephone companies build a second, similar network to compete with cable, the market could double.)