Optical: systems and elements – Optical modulator – Light wave temporal modulation
Reexamination Certificate
2001-01-17
2004-08-31
Schwartz, Jordan M. (Department: 2873)
Optical: systems and elements
Optical modulator
Light wave temporal modulation
C359S291000, C359S221200
Reexamination Certificate
active
06785038
ABSTRACT:
STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT
Not applicable.
REFERENCE TO MICROFICHE APPENDIX
Not applicable.
BACKGROUND OF THE INVENTION
The invention relates generally to optical switching arrays, and more specifically to an optical cross-connect of micro-electro-mechanical system cells, each cell having a magnetic actuator and integrated mirror.
The use of optical signal transmission is rapidly growing in the telecommunications (“telecom”) industry. In particular, optical transmission techniques are being used for local data transfer (“metro”), as well as point-to-point “long-haul” transmissions. Similarly, the number of optical signals, or channels, carried on an optic fiber is growing. The implementation of wave-division multiplexing (“WDM”) has allowed the number of channels carried on a fiber to increase from one to over 16, with further expansions planned.
Thus, the need for optical switching technology is expanding. In the case of WDM technology, an optical channel is removed from a multi-channel fiber with an optical bandpass filter, diffraction grating, or other wavelength-selective device, and routed to one of perhaps several destinations. For example, a channel on a long-haul transmission line might be routed between a local user for a period of time and then switched back onto the long-haul transmission line (“re-inserted”). This is known as 1×2 switching because a single input is switched between one of two possible outputs. As the complexity of optical networks grows, the complexity of the desired switching matrices also grows.
Switching matrices are being developed for several different optical network applications. “Small fabric” applications have been developed using 1×2, 2×2, and 1×8 type switches. However, there is a need for “medium fabric” applications that can provide 8×8 up to and beyond 32×32 type switching arrays, and even for “large fabric” switching arrays that can handle 1024×1024 or more switching applications. The switching arrays that allow any input to be connected to any output are generally called “cross-connects”, but in some applications there may be limited switching of some ports.
Unfortunately, attempting to merely scale the techniques developed for small fabric applications may not meet system requirements, such as switching speed, switching array space limitations, and power limitations. In particular, it is often desirable to upgrade an optical network to handle more traffic by adding additional channels onto the installed fiber base, and that the switching arrays be able to fit into the existing “footprint” allowed for the switching matrix. In many cases, the footprint is actually a 3-dimensional restriction. Similar restrictions might apply to the available power, or allowable power dissipation.
Various techniques have been developed to address the problems arising in the development of more complicated switching arrays. Several approaches have adapted photolithographic methods developed primarily for the field of semiconductor processing to the fabrication of optical switching arrays. In one approach, micro-electro-mechanical systems (“MEMS”) are used to create a very small motive device (motor), such as an electrostatic comb drive, electrostatic scratch drive, magnetic drive, thermal drive, or the like, attached to an optical switching element, typically a mirror. The mirror is usually either fabricated in the major plane of the process wafer and rotated to become perpendicular to a switchable light signal, or is fabricated perpendicular to the major plane of the wafer. In the first instance, establishing and maintaining verticality of the mirror is very important to ensure that the light signal is reflected to the desired output port. In the second instance, fabricating a mirror-smooth surface on a vertical plane of the wafer can be difficult, as can be depositing a reflective metal layer on that surface.
Another approach uses cantilevered beams that bend away from the major plane of the substrate in response to a magnetic field. An electrostatic field can be used to hold the cantilevered beam, which can have a mirror on its end, in a switch position when the magnetic field is removed. Again, maintaining the verticality of the mirror during switch operation or after many switch cycles is important. If many cantilevered mirrors are used in an optical switching array, the magnetic field might be applied globally to the entire array, thus field uniformity issues might arise.
However, MEMS technology has the advantage of using photolithography to establish a precise relationship between MEMS cells on a wafer, and between elements within a cell. Several MEMS approaches involve fabricating one portion of the entire switching array on a first chip, fabricating a second portion of the entire switching array on a second chip, and then aligning and assembling the two chips to result in the switching array. While this approach appears convenient in that all 1,024 switch elements in a 32×32 array, for example, might be assembled at once, a defect in the alignment and assembly process can ruin the entire assembly. Similarly, each optical input signal needs to be aligned to each of its associated output paths. Even with MEMS technology, such simultaneous alignment is difficult.
Furthermore, additional cells are usually provided on the chip(s) to account for bad cells. For example, a 40×40 array might be built if a 32×32 array is desired, in case some of the cells in the included 32×32 array are bad. The switching array can be assembled to avoid the bad cell by using one of the 8 “extra” paths. However, this approach takes up additional space on the wafer, affecting yields, and doesn't work if more extra paths are required than what has been designed for. Furthermore, the control logic required to switch the semi-custom switching array is different from a standard array. This lack of standard switching logic can contribute to signal routing problems, both upon installation or replacement of a switching array.
Thus, a need exists for optical switching arrays that provide complex, efficient optical switching functions in a small footprint with high switching speeds and low power requirements.
BRIEF SUMMARY OF THE INVENTION
An optical cross connect is fabricated from MEMS cells edge-mounted to mounting substrate serving as a miniature optical bench. Each cell is individually addressable and raises a mirror at least 400 microns above the upper edge of the cell, allowing individual addressing of cells and switching of free space optical beams for high isolation between channels. The mirrors are rotated about an axis essentially parallel to the major surface of the mounting substrate. Optical input fibers and collimators are mounted on fiber mounting blocks bonded to the mounting substrate. In a preferred embodiment, the mounting substrate, fiber mounting blocks, and MEMS cells are made of silicon to match thermal expansion.
The individual drive mechanisms on each cell allow the configuration of the cross connect to be determined without disrupting any optical path by measuring the impedance of the drive circuit. In a particular embodiment, the drive mechanism is a magnetic drive that moves a magnetic slug or tab affixed to the mirror structure between magnetic poles on the base of the cell to latch the mirror either an extended or retracted position. In one embodiment, such a drive is insensitive to the polarity of the control signal, or pulse. In a further embodiment, the duration of the pulse is selected to first accelerate, and then decelerate the magnetic tab and hence mirror. This reduces bouncing and improves settling time.
Providing individual drives also results in low power requirements, as generally only 2N cells need to be switched to completely reconfigure the cross connect. In many operations reconfiguration is achieved by switching even fewer cells. Thus, the cross connect does not require switching of a high-power global magnetic or electronic
Foster John S.
Hichwa Bryant P.
Martin Richard T.
Rubel Paul J.
Stocker John W.
Optical Coating Laboratory, Inc.
Schwartz Jordan M.
Stultz Jessica
LandOfFree
Optical cross-connect with magnetic micro-electro-mechanical... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Optical cross-connect with magnetic micro-electro-mechanical..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Optical cross-connect with magnetic micro-electro-mechanical... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3272197