Another machine-vision success story

Today's telecommunications industry is so profitable that it can afford to automate manufacturing processes using machine-vision systems.

Th Acfb95

Today's telecommunications industry is so profitable that it can afford to automate manufacturing processes using machine-vision systems.

Th Acfb95
Click here to enlarge image

During a visit to a major telecommunications company, the company revealed that in its efforts to meet surging market demands it was having problems making and testing its OEM telecommunications products and finding experienced production, test, and quality-control engineers. Further highlighting the success of the OEM telecommunications market are the earnings of JDS Uniphase (Nepean, Ont., Canada). Its sales for the first quarter ended Sept. 30, 2000, were $786 million. These figures were 23% greater than its sales of $641 million for the quarter ended June 30, 2000, and 171% greater than its sales of $290 million for the quarter ended September 30, 1999.

Stephen Anderson, editor in chief of our sister publication Laser Focus World, spotlighted the emerging technical trends in the telecommunications industry at this year's National Fiber Optic Engineering Conference in Denver, CO. He said, "Planar technology [with its roots in the silicon-chip arena] is allowing multiple optical components to be integrated into single devices such as arrayed waveguide gratings. And, any overall approach to large-scale optical integration must also include alternative materials such as indium phosphide. The point, of course, is to add bandwidth and speed to optical networks."

New source of revenue?
It seems apparent that lying untapped is a possible huge source of revenue for machine-vision companies: delivering products and systems that help automate the production process of amplifiers, modulators, transmitters, and receivers for telecommunication-component manufacturers. For example, Cognex Corp. (Natick, MA) recently declared a plan for tapping into the telecommunications industry boom. It formed an engineering team to develop machine-vision solutions specifically for the fiberoptic sector.

Headed by Cognex cofounder and senior vice president of engineering Marilyn Matz, the Cognex fiberoptics team is focusing on manufacturers of fiberoptic components who want to automate the assembly of optical amplifiers, laser diodes, and pump lasers, for example. Much of this manufacturing is currently performed manually, and, as a result, yields are low and production rates are not keeping up with market demands.

According to Justin Testa, Cognex senior vice president of marketing, several optical-component manufacturers are already using the company's machine-vision systems to automatically align components during fiberoptic assembly. Cognex is currently collaborating with Automation Engineering Inc. (AEI; Woburn, MA) to integrate Cognex's PatMax pattern-recognition technology with AEI's FlexAuto open-architecture software platform for integrating motion control and machine vision. This alliance is offering automated solutions for fiberoptic-component alignment, assembly, and test; improvements in product time to market; and rapid increases in production capacities, according to AIE's president Andre By.

There are several reasons why this approach can succeed. First, the telecommunications industry is so profitable that it can afford to automate manufacturing using machine-vision systems. Second, because hiring talented test professionals is so difficult, given today's economy, the return on investment of such machine-vision systems is easily justified.

Cognex field application engineers already have experience in deploying systems for integrated-circuit and printed-circuit-board inspection. And, although tailoring such systems to inspect optical amplifiers, laser diodes, and pump lasers requires meticulous work, it is based on knowledge similar to that gained from customers of automated semiconductor-manufacturing equipment.

Kudos to Cognex for recognizing another successful application for machine vision.

by Andy Wilson
EDITOR
andyw@pennwell.com

More in Boards & Software