The Optoelectronic Technology Consortium formed by General Electric Co, AT&T Co, Honeywell Inc and IBM Corp (CI No 1,976) is aiming to open up the presently static photonics and optics markets with the development of prototype Gallium Arsenide integrated circuits for optical interconnections within and between computer and communications equipment. The 30-month research project aims to produce a monolithic Gallium Arsenide device with 32 laser light sources that can transmit parallel data at 16 Gigabits per second. A separate Silicon-based integrated circuit is being developed for the transmitter control circuitry, and with it a companion receiver integrated circuit that will include 32 sensors, decoders and amplifiers all on the same chip. AT&T Bell Laboratories director Philip Anthony predicts that the new 32 channel conn-ectors will increase performance but not costs. They could open up new markets such as fibre-to-the-home systems, fibre-optic computer connection and high-speed, high data optical circuitry in electronic systems too.

Complete changes

Complete changes in the design of computers could then result, according to David Lewis, the Director of optoelectronic research at GE, in turn transforming the connectors into mainline components. GE is anticipating widescale use of optoelectronics devices in sonar, radar and electronic warfare systems. Fibre optic connections could also have applications for massively parallel super-compters. The consortium, which is combining investment of its own with an $8m US government grant, does not plan to add any further corporate members to its number. But it will be setting up a Users Application Group to help guide design and to receive technical progress reports. The four have divided the responsibilities as follows: GE will manage the project and will work on link des ign, modelling and testing; AT&T will produce surface-emitting lasers and fibre-array interconnections; Honeywell will develop modula tion technology and optical sour ces; and IBM will concentrate on optical receiver arrays, link mod elling and network analysis. Initially they will try two methods for producing the 32-laser circuit – building a 32 surface-emitting laser array or combining a single laser with mult-iple optical modulators. They are hopeful that proven 0.75 to 1 micron linewidth fabrication processeses can be used to produce the high level of integration required for the device.