Can Omron Slash Component Costs?

Japan's Omron Corp. wants to get out of making low-tech optical components because of the threat of low-cost Chinese competition.

Guess how it plans to do it?

By slashing the cost of manufacturing more advanced components.

Specifically, Omron has developed a new way of making optical waveguide chips that it claims reduces manufacturing costs by a factor of 10. In other words, it says it can make a component for $10 that would cost other manufacturers $100 to make.

Omron's first cut-price products will be available next spring and will include a 1X4 optical coupler, an optical tap coupler, a variable optical attenuator (VOA), a transceiver/waveguide and a 1X2 switch, according to Hayami Hosokawa, a manager in Omron’s Central R&D Labs.

Initially, Omron will target the domestic Japanese market with products for metro, access and home networking. After that, Omron will begin exports, Hosokawa adds.

The production method Omron is touting scraps the conventional semiconductor process and replaces it with a method the company calls its Stacked Polymer Optical IC Advanced technology, or SPICA.

A typical semiconductor approach to building waveguides typically involves a nine-stage process, including such steps as substrate cladding, coating, adding resist, lithography, developing, ion resist removing, and cladding, says Hosokawa. This is the sort of process used by NTT Electronics Corp. (NEL), Sumitomo Electric Industries Ltd., and NEC Electronics Inc., to name a few, he says. Hosokawa’s boffins use a 1-millimeter thick, 6-inch diameter glass substrate and a stamping-cladding process that he claims will produce waveguides in three steps.

Step one involves coating the glass polymer disk’s top surface with a 20-micron to 30-micron thick acrylic resin. Step two is to flash this with UV for hardening. Then the hammer comes down and stamps out the pattern. So, basically, Omron claims it just adds a lick of paint, a flash of blinding light and then a nickel pattern stamp, and hey presto! there's the waveguide.

“We’re changing the whole approach to the manufacturing process and it’s gonna result in much lower costs. Conventionally, you have to use high-cost semiconductor process equipment. We have a process that has only three stages and works like a name stamp,” claims Hosokawa. (In Japan, people usually don’t sign documents, they use a stamp with their family name in Kanji characters.)

If SPICA sounds too simple to be true, Hosokawa admits that it's taken more than a decade to perfect the process at the company’s two main research labs in Kyoto and Tsukuba Science City, north of Tokyo. The technique’s origins began with patents taken out in the late 80s, which lead to the company developing micro lens arrays and lightwave control devices that got fitted into LEDs and backlights, says Hosokawa. Those businesses will account for about ¥5.7 billion (US$47.2 million) in revenue this year, according to the company. The components will perform competitively, claims Hosokawa. For example, the 1X4 coupler has insertion losses of 7.2 dB at 1.31 and 7.9 dB at 1.55 micron wavelengths, uniformity of 0.2 dB, return loss of 58 dB and temperature dependent loss of 0.3 dB. The company is prepared to reel off data about the other chips to interested parties.

Omron seems to be growing discontented as a second-tier components maker. Its U.S. subsidiary Omron Electronics produces all sort of widgets, such as RF relays and amps, markets in which the company is being squeezed by cheap Chinese competition (see US to China: Do You Copy?). So headquarters in Kyoto is pressing to develop optoelectronics and MEMS technologies to take the company further up the snack chain. Last year the company as a whole lost ¥15.7 billion (US$130 million) on sales of ¥423 billion (US$3.5 billion) in the U.S. and Japan.

The company wants to become numero uno in lightwave control devices -- but it doesn’t say when. Omron’s sales projections envision ¥20 billion ($165 million) in revenue for optical communication devices in 2005, with half of this coming from the new optical waveguides.

The path to this kind of success, however, isn’t yet clear. Omron will concentrate on the optical transceiver market for fiber to the home (FTTH) before moving into metro and long haul, hints Hosokawa. Hosokawa confirms he will be producing samples this year, but the company isn’t sure for whom it will be stamping out product.

The focus on FTTH products suggests that the first components will be low-performance. Omron will need to develop more-integrated, higher-performance products to achieve its goal of escaping the threat of cut-price Chinese competition.

— Paul Kallender, special to Light Reading
http://www.lightreading.com