MEMS development innovations continue

Th 136920

Industry analysts may be predicting hard times ahead for MEMS-based switch component vendors, but that doesn't mean all innovation in the space has ceased. Despite the closures of such high-profile MEMS manufacturers as OMM, some component vendors are still banking on optical MEMS for high-speed transmission applications.

A few years ago, most vendors were focused on two-dimensional (2-D) MEMS for small and mid-level switches, while their three-dimensional (3-D) counterparts were hyped as the ideal solution for 1,000×1,000-type switches. Even though the monster switches have all but disappeared, a surprising number of vendors are sticking to the 3-D technology.

"The systems that are seeming to find the most success are three-dimensional across the board, from the very small switch arrays to the larger ones," contends Marlene Bourne, senior analyst with In-Stat/MDR (Scottsdale, AZ). "I find that very interesting, but it's not surprising to companies like Calient Networks. They told me several years ago that even the small-end switches would ultimately use a three-dimensional architecture, and that's turning out to be true."

Calient (San Jose, CA) has the largest optical-switch subsystem on the commercial market, the 3-D MEMS-based 256×256-port PX switch. "We're seeing a handful of interesting applications emerging," reports Calient president and chief executive Charles Corbalis. "One of them is an automated fiber patch panel for system test and manufacturing test, where you need to reconfigure the optics on a regular basis."

That said, he hasn't given up on optical switches for high-speed networking applications. "With the cost of 10-Gbit/sec components [declining] and the slow migration toward more and more broadband services, I think we're poised to see the promise of big optical switches for high-speed networking that everyone was talking about two and three years ago," Corbalis says.

A trio of vendors, including Fujitsu, Agiltron, and Xerox, hopes Corbalis is correct. The orders may not be pouring in, but that hasn't stopped these vendors from making the most of their idle time.

Fujitsu (Tokyo) recently announced the development of what it claims is the world's fastest multichannel device; its switching speed clocks in at 1 msec. The company developed the 80-channel, 3-D MEMS switch for its own internal use and plans to incorporate the technology into its optical transmission and crossconnect systems within the next few years.

To combat the challenges of mechanical resonance and insufficient switching speed, Fujitsu's research team developed a notch filter in the MEMS mirror that suppresses the mechanical resonance inherent in optical switching. The filter eliminates only the mirror's resonant frequency from the driving electrical waveform. This technology, coupled with the company's comb-driven rapid-response mirror array, enables faster switching times. Th 136920

Calient says its integrated MEMS mirror array, hermetically sealed and NEBS 3-compliant, has withstood one billion switch cycles without failure.

The switch fabric itself is folded. According to Fujitsu, the input beam is reflected through a roof-type retro-reflector on its way to the output. Compared with flat mirrors, the folded switch fabric halves the optical path, enabling more compact switches. The device measures just 150×400×300 mm. Th 136921

Fujitsu has developed a folded optical-switch-fabric configuration, which halves the length of the optical path, enabling a more compact switch.

Size is also a key consideration for Agiltron (Wilmington, MA), winner of a $3.5-million grant from the Advanced Technology Program of the National Institute of Standards and Technology (NIST) to develop a new type of MEMS-based optical switch. The four-year project will result in a prototype 32×32 optical switch for crossconnect applications.

The folks at Agiltron believe previous MEMS switches suffered from large insertion loss because of the small mirrors they had to employ—a limitation imposed by the actuation mechanisms involved. "They simply do not provide sufficient force to move a mirror of the size required to achieve low loss," notes Jack Salerno, vice president of technology and strategic business at Agiltron.

Agiltron's switch is based on proprietary materials technology that provides what the company calls "solid state motion"; the material expands or contracts in response to an applied electrical field. "We are able to achieve a large amount of motion using little electrical power and without the need for complex actuation mechanisms such as hinges or actuators," explains Salerno.

Joel Kubby, technical manager at Xerox's Wilson Center for Research and Technology (Webster, NY), also received a NIST award, which he used to develop a new technology that integrates an optical MEMS photonic switch with planar lightwave circuits (PLCs) on a single silicon chip. The PLCs or waveguides act like a miniature train-track switch, conducting and directing the light more precisely than earlier mirror-based switches.Th 136922

Xerox's new technology integrates an optical MEMS photonic switch with planar lightwave circuits on a single silicon chip small enough to fit on a fingertip.

Because both the switch and waveguides are made on the same crystal silicon wafer, the Xerox technology avoids the complex alignment issues associated with manually connecting different components with optical fibers, say company representatives. Integration also results in a less costly and smaller device. Xerox contends its switch technology could enable reconfigurable optical add/drop multiplexers (ROADMs) as small as 2×1.5 cm.

Unlike Fujitsu and Agiltron, which developed their technology for internal use, Xerox plans to license its technology to optical-switch vendors. "We've brought it to the proof-of-concept stage," asserts Tony Paine of Xerox's Innovation Business Development Group. "We've shown that it can be done, and we're getting a lot of interest in the technology, the achievement, and the size of the device, which is extremely small."

The question is, will any of this interest turn into a licensing agreement? And just how will these new MEMS technologies fare in a market that may not be ready for them?

"With the downturn, some folks said MEMS technology has lost out on its opportunity, but I don't see anything coming up from behind," reports In-Stat/MDR's Bourne. "I think carriers are still viewing this as the next-generation technology. Hopefully for those companies that are able to continue to get funding and move forward, the products they have to offer when customers are ready to buy them will be that much better than they would have been two years ago."

More in Transport