Optical switch module

0001The present invention relates to fiber optic communication equipment and in particular to fiber optic switches and fiber optic switch modules.

CROSS REFERENCE TO RELATED APPLICATIONS

0002This application is a companion application to the following related applications filed concurrently with this application: Serial No. <patcit id="pcit0001" dnum="WO11728362A"><text>11/728,362</text></patcit>, MEMS Mirror Array and Control, Serial No. <patcit id="pcit0002" dnum="WO11728345A"><text>11/728345</text></patcit>, Beam Position Sensor, Serial No. <patcit id="pcit0003" dnum="WO11728344A"><text>11/728,344</text></patcit>, Optical Fiber Array Alignment Unit, and Serial No. <patcit id="pcit0004" dnum="WO11728335A"><text>11/728,335</text></patcit>, Optical Switch with Co-Axial Alignment Beam, all of which are incorporated herein by reference.

BACKGROUND OF THE INVENTION

Fiber Optic Communication

0003Over the past few decades, the telecommunications industry has exploded, and the incorporation of optical fiber into this industry is revolutionizing the way information is transmitted. Communication systems which use optical fiber as the transmission media offer some significant advantages over past copper-based systems. These advantages include higher bandwidths and transmission rates, lower transmission losses, and greater signal isolation. There exist in the United States many million miles of optical fibers. Information must be routed through this maze of fibers at the speed of light from millions of transmitters to millions of receivers.

Fiber Optic Multiplexing and De-Multiplexing

0004In a typical fiber optic communication system several fibers may be bundled together with many separate signals combined within the light beam carried in each of the fibers in the bundle. This combination of separate signals into a single beam carried by a single fiber is called multiplexing the signals. Both time division and frequency division multiplexing may be utilized. In typical fiber optic systems each signal carries with it a code so that traffic controls in the system can direct the signal to its proper destination. Light beams comprised of many signals typically travel serially through several or many fibers before reaching its sub-destination or final destination. Individual signals are collected into a single fiber in a process called multiplexing and separated out from other signals in a beam in processes called de-multiplexing. This operation may occur once or several times during the transit of information in the form of these light signals from sender to receiver.

0005<figref idref="f0026">FIG. 13 A</figref> depicts de-multiplexing and <figref idref="f0026">FIG. 13B</figref> depicts multiplexing. <figref idref="f0026">FIG. 13C</figref> shows a prior art static cross connection with two de-multiplexers and two multiplexers showing how signals carried at four separate wavelength ranges on two separate fibers can be switched to two other fibers. Optical filters that transmit a single wavelength range and reflect all other wavelength ranges are often used to separate wavelength ranges in multiplexers and de-multiplexers. A well known filter is the thin film filter as shown in <figref idref="f0027">FIG. 13D</figref>. These filters are typically built up on a glass substrate with thin films of one or more sets of 1/4 wave dielectric reflectors on both sides of a 1/2 wave cavity. <figref idref="f0027">FIG. 13E</figref> shows the result of one, two and three sets of 1/4 wave reflectors and 1/2 wave cavities. <figref idref="f0028">FIG. 13F</figref> shows how these narrow band filters can be used to produce a de-multiplexer. A multiplexer results from switching the directions indicated by the arrows.

Adjusting Fiber Routes

0006Traffic controls can route a particular signal from sender through many fibers to the receiver without changing the way the various optical fibers of the system are connected. However, as particular fiber routes become crowded, the connections between fibers must be modified to reduce the crowdedness or to route the signals more efficiently. This is the job of the fiber optic switch. This operation can be done by changing the actual connections between fibers in a fiber switch unit. Historically, fast switching of optical-beam routes through optical fibers has been accomplished using hybrid optical-electrical-optical switches for detection and conversion of optical signals entering the switch from a first fiber to an electrical signal that is used to produce a new optical signal for transmission over a second optical fiber.

MEMS Mirrors

0007MEMS mirrors are lithographically produced mirrors that are operated with voltage signals applied through integrated circuits produced with similar lithographic techniques. These mirrors typically are very tiny having dimensions measured in millimeters or fractions of millimeters. They are designed with extremely tight tolerances necessary for proper angular alignment of the various reflective elements, and usually require very sophisticated feedback control systems.

Automatic All Optical Cross Connect Switches

0008Recently, a number of optical cross connect switches have become available for switching optical signals directly from one fiber to another, thereby eliminating the need to convert the optical signal to an interim electrical signal. These optical switches incorporate various optical switch elements, such as mirrors, prisms, fiber collimators, and complicated drive mechanisms, to route optical signals through the switch. For some optical switches, MEMS mirrors have been utilized. All optical switches are described in the following patents recently issued which contain features similar to some of the features of the present invention: 7,190,509, Optically Addressed MEMS and 7,177,497, Porous Silicon Filter for Wavelength Multiplexing and De-Multiplexing, both of which are incorporated herein by reference, see also <patcit id="pcit0005" dnum="WO03036364A"><text>WO 03/036364</text></patcit>.

Applications of All Optical Automatic Cross Connect Switches

0009Known uses of all optical cross connect switches include (1) use as the principal component in a automated fiber patch panel, (2) use a component of a reconfigurable optical add drop multiplexer (ROADM) system and (3) use for automatic testing and measurement of optical components and systems.

Automated Fiber Patch Panel

0010Automated fiber patch panels are the components of a fiber optic communication network where communication routes are established and modified. These panels can be computer controlled to maintain network efficiency, to avoid overload and to respond quickly to fault situations.

Reconfigurable Optical Add Drop Multiplexers

0011When individual fibers are carrying many separate signals, the network must provide for adding new signals to the fiber and extracting (dropping) other signals. <figref idref="f0028">FIG. 13G</figref> shows an add-drop unit comprised of multiplexers and de-multiplexers but no switch. This unit would be considered a static unit and requires an operator to reconfigure it. <figref idref="f0029">FIG. 13H</figref> is a similar unit but with an optical switch that can be remotely operated or programmed to operate automatically. This unit includes tunable transponders permitting control of the wavelengths added. <figref idref="f0030">FIG. 13I</figref> shows a ROADM comprising four separate optical switches for switching signals among fibers as well as controlling the adding and dropping of signals for local service.

Test and Measurement

0012Automated all optical cross connect switches can greatly simplify testing of optical components especially components of typical communication networks simultaneously carrying millions of messages.

The Need

0013As the telecommunications industry continues to develop and grow to service more and more customers, the need for large scale, reliable optical switches increases. Consequently, there is a need for an optical cross connect switch which can be readily integrated into existing telecommunications systems, and which can reliably switch optical signals from each fiber of an array of input fibers, to each fiber in an array of output fibers, and which can accomplish this switching quickly, with minimal power loss and at a minimal cost per channel.

SUMMARY OF THE INVENTION

Cross Connect Switch

0014The present invention provides an all optical cross connect switch utilizing MEMS mirrors for cross connecting optical fibers in a first set of optical fibers to optical fibers in a second set of optical fibers. The optical fibers are preferably arranged in rectangular arrays. These arrays include array sizes such as 4 X 8, 16 X 16 and 8 X 16. A preferred embodiment built and tested by Applicants is a modular optical switch in which an input 8 X 16 array of optical fibers from sixteen eight-fiber ribbons are cross-connected into an output 8 X 16 array of optical fibers also from sixteen eight-fiber ribbons.

MEMS Mirror Arrays

0015The cross connect switch of the present invention includes two MEMS mirror arrays. In preferred embodiments each of the MEMS mirrors are driven in two axes by vertical comb drive actuators.

Fiber-Microlens Positioning Array

0016The optical fibers in the first set of optical fibers are precisely positioned in a first fiber-microlens positioning array to define a first set of parallel collimated cross-connect communication beam paths, with each collimated cross-connect communication beam path connecting an optical fiber in the first set of optical fibers with a MEMS mirror in a first MEMS mirror array. The optical fibers in the second set of optical fibers are precisely positioned in a second fiber-microlens positioning array to define a second set of parallel collimated cross-connect communication beam paths, with each collimated cross-connect communication beam path connecting an optical fiber in the second set of optical fibers with a MEMS mirror in a second MEMS mirror array. These parallel collimated beam paths establish a correspondence between the first fiber-microlens positioning array and the first microlens array and a correspondence between the second fiber-microlens positioning array and the second microlens array so that each optical fiber in the first positioning array has its own corresponding microlens in the first microlens array and each optical fiber in the second positioning array has its own corresponding microlens in the second microlens array.

Co-Axial Alignment Beams

0017A first set of alignment beams are added to and aligned co-axially with each of the first set of parallel collimated cross-connect communication beams. A second set of alignment beams are added to and aligned co-axially with each of the second set of parallel collimated cross-connect communication beams.

Dichroic Mirror

0018A dichroic mirror is positioned to reflect communication beams from MEMS mirrors in the first MEMS mirror array and to reflect cross connect communication beams from MEMS mirrors in the second MEMS mirror array and to transmit the first set of alignment beams and the second set of alignment beams.

MEMS Control System

0019A MEMS control system is provided to position the MEMS mirrors in the first and second MEMS mirror arrays to optically connect any of the optical fibers in the first set of optical fibers to any optical fiber in the second set of optical fibers. In preferred embodiments the MEMS control system is adapted to position each MEMS mirror in the first MEMS mirror arrays so as to reflect the cross connect communication beams from its corresponding fiber in the first set of optical fibers off the dichroic mirror and onto to a MEMS mirror in the second set of MEMS mirrors corresponding to a any selected optical fiber in the second set of optical fibers and to position the corresponding MEMS mirror in the second set of MEMS mirrors to direct the communication beam to its corresponding optical fiber in the second set of optical fibers. In preferred embodiments the mirrors are controlled by adjusting voltage potentials applied to the comb drive actuators of the individual mirrors in order to establish desired optical communication paths between optical fibers in the first set of optical fiber and the second set of optical fibers.

Beam Direction Sensors

0020The MEMS control system includes a first beam direction sensor unit positioned to detect each alignment beam in the first set of alignment beams transmitted through the dichroic mirror and a second beam direction sensor unit positioned to detect each alignment beam in the second set of alignment beams transmitted through the dichroic mirror. Each of the beam direction sensor units each includes an alignment beam detection screen and a video camera for viewing the position of the intersections of the alignment beams with the unit's viewing screen. In these embodiments the MEMS control system includes a processor programmed to provide a closed loop adjustments of pairs of MEMS mirrors (one from each of the two MEMS mirror array) in order to determine appropriate voltage potentials to be applied to the comb drive actuators in order to provide each desired optical path between the two sets of optical fibers. Applicants' tests have shown that once the calibration has been performed there is very little drift in the beam paths under normal conditions. However, significant changes in environmental conditions could require a recalibration. In some embodiments the switch could be adapted to automatically re-calibrate itself periodically or at the direction of operating personnel.

V-Groove Fiber-Microlens Positioning Arrays

0021In preferred embodiments the first and second fiber-microlens positioning arrays include a positioning plate with lithographically-defined sub-micron V-groove alignment features for alignment a set of ribbon optical fibers. The fibers are first positioned in the V-grooves, then glued in place after which the ends of the fibers are polished smooth with the exit surface of the positioning plate and matched with a micro lens array to provide the fiber-micro lens positioning array with sub-micron positioning accuracy.

BRIEF DESCRIPTION OF THE DRAWINGS

0022<ul id="ul0001" list-style="none"><li><figref idref="f0001">FIG. 1</figref> shows a prospective view of a preferred embodiment of the present invention with alignment cameras in place.</li><li><figref idref="f0002">FIGS. 1A and 1B</figref> show features of an alignment beam insertion component.</li><li><figref idref="f0003">FIGS. 2A and 2B</figref> respectively show a prospective and a top view of portions of the preferred embodiment of the present invention.</li><li><figref idref="f0004">FIGS. 3A, 3B and 3C</figref> show additional alignment beam features and test results.</li><li><figref idref="f0005">FIG. 4</figref> shows the size of the preferred embodiment in its modular form relative to an adult human hand.</li><li><figref idref="f0005">FIGS. 4A, 4B and 4C</figref> respectively show the preferred module utilized in a RODAM application, a patch panel application and a test and measurement unit.</li><li><figref idref="f0006">FIGS. 5A</figref>, <figref idref="f0007">5B</figref> and <figref idref="f0008">5C</figref> show optical paths in a preferred optical switch.</li><li><figref idref="f0009">FIGS. 6A</figref>, <figref idref="f0010">6B</figref> and <figref idref="f0011">6C</figref> show features of a preferred V-groove fiber-microlens positioning array.</li><li><figref idref="f0012">FIG. 7</figref> shows a MEMS driver control circuit.</li><li><figref idref="f0013">FIG. 8A</figref> shows a MEMS mirror array designed by applicants.</li><li><figref idref="f0014">FIGS. 8B</figref>, <figref idref="f0015">8C</figref> and <figref idref="f0016">8D</figref> show magnified portions of the MEMS mirror array.</li><li><figref idref="f0017 f0018">FIGS. 9A through 9K</figref> demonstrate important lithography steps used to make one of the vertical comb drives for the MEMS mirror in the MEMS mirror array.</li><li><figref idref="f0019">FIG. 10</figref> shows a control system.</li><li><figref idref="f0020">FIG. 11</figref> shows a nationwide fiber optic network.</li><li><figref idref="f0021 f0022 f0023 f0024 f0025">FIGS. 12A through 12J</figref> show concepts for using porous silicon filters for multiplexing and de-multiplexing.</li><li><figref idref="f0026 f0027 f0028 f0029 f0030">FIG. 13A through 13I</figref> show prior art multiplexing techniques.</li></ul>

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

First Preferred Embodiment

0023<figref idref="f0001">FIG. 1</figref> shows at 2 a perspective view of features of a first preferred embodiment of the present invention. This first preferred embodiment is an optical switch module 2. Its size and general shape is described by the drawing in <figref idref="f0005">FIG. 4</figref> as compared to a adult human hand. This optical switch module 2 is designed for switching optical communication beams carried by the fibers of a first 128 (8 X 16) optical fiber bundle such as bundle 4 as shown in <figref idref="f0001">FIG. 1</figref> into the fibers of a second 128 optical fiber bundle such as bundle 6. The beams in any fiber of bundle 4 can be switched into any fiber of bundle 6. The switch is symmetrical and can be operated in either direction so that any fiber of bundle 6 can be switched into any fiber of bundle 4.

0024We sometimes refer to the fibers in bundle 4 as being the input fibers and the fibers in bundle 6 being the output fibers, recognizing that the switch works just as well with the fibers in bundle 6 being the input fibers and the fibers in bundle 4 being the output fibers. Also, in some cases a single fiber link could have communication beams flowing in both directions at the same time. In addition in some configurations links between fibers in bundle 4 and bundle 6 could be 4 to 6 for some fibers and 6 to 4 in other fibers. In this specification and in the claims, Application refer to the light beams carried by the optical fibers as "communication beams" primarily to distinguish them from the "alignment beams" but there may be applications of the present invention where the beams carried by the optical fibers may not be generally considered as communication beams in the normal sense because they are not specifically carrying information from one place to the other. Nevertheless, for purpose of this specification and the claims, all light beams carried by the optical fibers in bundles 4 and 6 are to be considered as communication beams.

0025This module is designed for easy integration into a line card for insertion into a standard communication panel in a reconfigurable add-drop application as shown in <figref idref="f0005">FIG. 4A</figref>. The module can be utilized in a patch panel as shown in <figref idref="f0005">FIG. 4B</figref> and can be utilized in a large nation-wide communication system such as the one discussed in the section of this description entitled "Application in a Nation Scale Fiber Optic Network" in which each fiber is carrying information at a large number of separate frequencies. The switch is also useful as a tool for component testing and system monitoring by installation in a test and measurement panel as shown in <figref idref="f0005">FIG. 4C</figref>.

0026Important components of the preferred embodiment shown in <figref idref="f0002">FIGS. 1A, 1B</figref>, <figref idref="f0003">2A and 2B</figref> are alignment beam units 8 and 10, dichroic beam splitter 24 and alignment screen and MEMS mirror arrays 20 and 22. Camera units 16 and 18 are shown in <figref idref="f0006">FIG 5A</figref>.

Fiber-Microlens Positioning Array

0027The optical fibers in the first set of optical fibers are precisely positioned in a first fiber-microlens positioning array unit 8 as shown in <figref idref="f0003">FIG. 2A and 2B</figref> to define a first set of parallel collimated cross-connect communication beam paths, with each collimated cross-connect communication beam path connecting an optical fiber in the first set of optical fibers with a MEMS mirror in a first MEMS mirror array. The optical fibers in the second set of optical fibers are precisely positioned in a second fiber-microlens positioning array 9 to define a second set of parallel collimated cross-connect communication beam paths, with each collimated cross-connect communication beam path connecting an optical fiber in the second set of optical fibers with a MEMS mirror in a second MEMS mirror array.

0028<figref idref="f0009">FIGS. 6A</figref>, <figref idref="f0010">6B</figref> and <figref idref="f0011">6C</figref> show features of a preferred fiber-microlens positioning array units. In the preferred switch each of the two fiber-microlens positioning arrays and each of the two microlens arrays define matching 8 X 16 matrices that Applicants have labeled 1-1 beginning at the top left as shown in <figref idref="f0015">FIG. 6C</figref> to 16-8 at the bottom right. Some of the corresponding matrices numbers are shown on the MEMS mirror drawing in <figref idref="f0014">FIG. 8B</figref>. These parallel collimated beam paths establish a correspondence between the first fiber-microlens positioning array and the first microlens array which are similarily labeled and a correspondence between the second fiber-microlens positioning array and the second microlens array so that each optical fiber in the first positioning array has its own corresponding microlens in the first microlens array and each optical fiber in the second positioning array has its own corresponding microlens in the second microlens array. In preferred embodiments the first and second fiber-microlens positioning array units are identical.

V-Groove Position Plate

0029Details of these preferred fiber-microlens positioning arrays are described in <figref idref="f0009">FIGS. 6A</figref>, <figref idref="f0010">6B</figref> and <figref idref="f0011">6C</figref>. These units are designed to easily and precisely position optical fibers in sixteen standard eight fiber ribbons with fibers having 125 micron cores. In <figref idref="f0010">FIG. 6B</figref> the ribbons are shown at 4A, individual fibers are shown at 4B and the bundle of fibers are shown at 4. Positioning plate 7A is comprised of a plate comprised of silicon that has been fabricated using lithographic techniques. Plate 7A is mounted on mounting plate 7B which contains mounting slots for mounting the fiber-microlens array unit in the optical switch module. Elongated horizontal slots 8 are provided for the eight fibers in each or the 16 ribbons. Precision V-groves slots 8A are cut in the bottom of elongated slots 8 These slots have 250 micron wide vertical sides 8B with a 45-degree V section 8C at the bottom of the groove, all as shown in <figref idref="f0009">FIG. 6A</figref>. The slots are on 2 millimeter centers in the horizontal direction and the slots are on 1 millimeter centers in the vertical direction. The V-groove slots in each row are off-set from the slots in their closest neighbor row or rows as shown in <figref idref="f0010">FIG. 6B</figref>, so the array cross-section dimension is 15 millimeters X 15 millimeters. In these preferred embodiments the individual fibers are inserted into the slots and are precisely located in the V-grooves and glued in place. After the fibers are solidly located the ends of the fibers are cut and then the protruding and excess glue are polished away to obtain a precisely perpendicular exit surface for each fiber.

Micro-Lens Array

0030In these preferred embodiments an 8 X 16 microlens array 9, also fabricated with precision using lithographic techniques is provided with the microlens array precisely positioned to correspond to the position of the fibers in the V-grove position plate. Each lens has a diameter of about 1.1 millimeter and a focal length of 3.7 millimeter. Precision spacers not shown are used to position the microlens array so that the lenses are positioned about 3.7 mm from the ends of the fibers in the V-groove position plate. When positioning microlens array 9 the final precise location is chosen so that the waist of communication beams passing through the switch from one fiber bundle to the other is located as close as reasonably feasible to dichroic mirror 24 as shown in <figref idref="f0001 f0006 f0007 f0008">FIGS. 1 and FIGS. 5A, 5B and 5C</figref>.

Co-Axial Alignment Beams

0031Preferred embodiments of the present invention include provisions for providing alignment beams aligned co-axially with each of the beams exiting each fiber in both of the fiber bundles 4 and 16. A particular technique for providing these alignment beams is shown in <figref idref="f0001">FIGS. 1</figref>, <figref idref="f0002">1A and 1B</figref>. Alignment techniques are described by reference to <figref idref="f0006">FIGS. 5A</figref>, <figref idref="f0007">5B</figref> and <figref idref="f0008">5C</figref>. As shown in <figref idref="f0001">FIG. 1</figref> this preferred embodiment includes alignment unit 10 for aligning communication beams from fiber bundle 4 and alignment unit 12 for aligning communication beams from fiber bundle 6. Alignment unit 12 is also shown in <figref idref="f0002">FIG. 1A</figref> and a cut away drawing of it is shown in <figref idref="f0002">FIG. 1B</figref>. Unit 12 includes an injection unit 12A with a "point" visible light source shown at 12F such as a vertical cavity surface emitting laser (VCSEL) operating in the near infrared at 850 nm and having a divergence of about 30 degrees. The beam from this source is collimated with collimator optics 12B to produce a collimated beam with a cross sectional dimension of about 16 millimeter X 16 millimeters. This collimated beam is separated into 128 separate beams with mask 12C having 128 0.6 mm diameter apertures that are positioned to align the 128 separate parallel beams with the communication beams from fiber bundle 6. The actual alignment occurs at dichroic beam splitter 12D that passes the communication beam and reflects the alignment beams.

MEMS Mirror Beam Control

0032In preferred embodiments communication beams are directed within the switch module with two MEMS mirror arrays 20 and 22. The two arrays are identical. MEMS mirror array 20 is shown in <figref idref="f0013">FIG. 8A</figref>. The position of the mirrors correspond precisely to the positions of the fibers and the microlenses in fiber-microlens array unit 8 and 10. An expanded drawing of the top left corner of <figref idref="f0013">FIG. 8A</figref> is shown in <figref idref="f0014">FIG. 8B</figref> and a further expanded drawing of one of the mirrors (mirror 3-1) is shown in <figref idref="f0015">FIG. 8C</figref>. <figref idref="f0016">FIG. 8D</figref> shows an even further expansion of one of the vertical comb drives, drive 60, of one of the <figref idref="f0015">FIG 8C</figref> mirrors. Techniques for producing MEMS mirrors are well known and well reported in the prior art. The MEMS mirrors of the present invention include special features adapted to produce the excellent performance of preferred embodiments of the present invention. As shown in <figref idref="f0015">FIG. 8C</figref> a gold reflecting surface 80 of each MEMS mirror can be pivoted about two axes, a first axis 82 about which pivots inner frame 84 and a second axis 86 about which reflecting surface 80 pivots. Comb fingers attached to the axels are positioned below "stationary" fingers (shown white in <figref idref="f0013 f0014 f0015">FIGS. 8A, 8B and 8C</figref>) attached to the frames. Electrical potentials of 0 to 200 volts causes the lower fingers to be pulled up between the "stationary" fingers causing the mirror surface to pivot about it respective axes.

Fabrication Techniques for Comb Drives

0033A special preferred technique for producing the comb drives for the MEMS mirrors shown in the figures is described with respect to <figref idref="f0017 f0018">FIGS. 9A through 9K</figref>. As indicated in <figref idref="f0017">FIG. 9A</figref> the process starts with a wafer of two 25 micron thick layers 30 and 32 of single crystal silicon on a 380 micron thick single crystal silicon handle layer 34. The three layers of silicon are separated by a 1.5 micron buried oxide insulator 36 and a 0.5 micron buried oxide insulator layers 38. The silicon layers are all doped to reduce their electrical resistivity to the range of about 10 to 20 milli-ohm centimeters. A SiO2 layer 40 is added as shown in <figref idref="f0017">FIG. 9B</figref> which is patterned using photo resist and a contact mask to produce the pattern 42 shown in <figref idref="f0017">FIG. 9C</figref>. Next as shown in <figref idref="f0017">FIG. 9D</figref> a resist mask 44 is applied covering part of oxide pattern 42 as shown in 9D. This modification of mask pattern 42 is a very important step since it and some following steps will assure that the extremely thin comb elements in this comb drive will be properly aligned. For as shown in <figref idref="f0017">FIGS. 9E</figref>, the mask laid on in <figref idref="f0017">FIG. 9D</figref> is used to etch away part of the oxide pattern laid down in the <figref idref="f0017">FIG. 9C</figref> step. Next as shown in <figref idref="f0017">FIG. 9F</figref> a deep reactive ion etch is used to etch silicon layer 30 using the same mask pattern 44. After this etching step has completely removed the silicon between the wafer surface and oxide layer 36 the photo resist applied in the <figref idref="f0017">FIG. 9D</figref> step is removed as shown in <figref idref="f0018">FIG. 9G</figref>. A blanket oxide etch is applied to remove all exposed buried oxide while leaving enough of the modified mask 42 to protect the silicon beneath it form the subsequent silicon etching step shown in <figref idref="f0018">FIG. 9J</figref>. Next backside etches are applied to remove most of handle 34 and oxide layer 38 under silicon layer 32 as shown in <figref idref="f0018">FIG. 9I</figref>. Then as shown in <figref idref="f0018">FIG. 9J</figref> the wafer is temporally mounted on a handle wafer and the unprotected portions of silicon layers 30 and 32 are etched away. And an oxide etch is applied to remove the oxide layers form the tops of the comb fingers to leave the comb fingers as shown in <figref idref="f0018">FIG. 9K</figref>. The tall fingers 50 are a part of the frame of the MEMS mirror and will be a fixed part of the MEMS frame or the central frame element of the MEMS mirror. The short fingers will be move with the mirror element as potential is applied as discussed later. The tall fingers are shown in white in <figref idref="f0016">FIG. 8D</figref> at 50 and the short moving fingers are shown in black at 52 in <figref idref="f0016">FIG. 8D</figref> outlined in a thin white line. These are the fingers that rotate with the gold mirror surface 80 of the MEMS mirror as shown in <figref idref="f0015">FIG. 8C</figref>.

0034The reader should note that the tall fingers 50 are divided into two conductive silicon parts 47 and 43 by insulating layer 45. During control of the MEMS mirrors electrical potentials of between 0 and 200 volts are applied between silicon parts 47 of the tall "stationary" fingers and the short moving fingers 52. This produces an attractive electrical force pulling moving fingers 52 up in between silicon parts 47 of the tall "stationary" fingers in order to tilt the reflecting mirror surface.

Control of the MEMS Mirrors

0035A preferred circuit for controlling the MEMS mirrors is shown in <figref idref="f0012">FIG. 7</figref>. Two circuits are needed for each comb drive unit. There are two comb drive units for each of the two axes for each of the 128 mirrors. So this embodiment requires 1024 circuits to control the 128 mirrors. The good news is that the components of the circuits shown in <figref idref="f0012">FIG. 7</figref> are very inexpensive, only a few cents for all of the components. The circuit is provided with a 200 volt bias as shown in the figure. This 200 volt bias charges capacitor C2 only when transistor X19 is in an "off" state. When transistor X19 is "on" there is a discharge path between capacitor C2 and the ground node. Transistor X19 is turned on and off with two kilohertz pulse width modulator indicated on the drawing as "2Khz PWM" which in turn is controlled (in Applicants' prototype unit) by a field programmable gate array or a programmable logic device (both not shown). The electric potential on capacitor C2 determines the position of one of the comb drives of one of the MEMS mirrors. The potential can be any potential between 0 volts and 200 volts. The potential on capacitor 2C depends on the modulation of the 2 Khz pulse width modulator shown in <figref idref="f0012">FIG. 7</figref>. If the width of each of the 2,000 pulses per second is maximum, transistor X19 will be "on" continuously and the voltage on C2 will be approximately zero. If the width of each of the 2,000 pulses per second is minimum, transistor X19 will be "off" continuously and the voltage on C2 will be approximately 200 volts. The pulse width can be varied between minimum and maximum to vary the voltage potential on capacitor C2 from 200 volts to zero volts. The tilt of MEMS mirrors such as MEMS mirror 3-1 depends on the voltage applied to the four vertical comb drive units controlling the position of each mirror.

0036Each of the MEMS mirrors has sufficient range to direct beams as needed to connect any of the fibers in bundle 4 to any of the fibers in bundle 6.

Beam Direction Sensors

0037In preferred embodiments the MEMS control system includes a first beam direction sensor unit 16 as shown in <figref idref="f0001">FIG. 1</figref> positioned to detect each alignment beam in the first set of alignment beams transmitted from injection unit 10 co-aligned with communication beams from fiber bundle 4, reflected from MEMS mirror array unit 20 and transmitted through the dichroic mirror 24. The embodiment also includes a second beam direction sensor unit 18 positioned to detect each alignment beam in the second set of alignment beams transmitted from injection unit 12 co-aligned with communication beams from fiber bundle 6 and transmitted through the dichroic mirror 24. Each of the beam direction sensor units each includes an alignment beam detection screen 60 or 62 and a video camera 64 or 66 for viewing the position of the intersections of the alignment beams with the unit's viewing screen. The screen can be any of a wide variety of screens that produce an image when illuminated with an alignment beam. <figref idref="f0004">FIG. 3B</figref> shows test results produced with a variety of screen candidates. Results are plotted as a function of mirror angle. <figref idref="f0004">FIG. 3A</figref> shows a typical illumination pattern viewed from the side of the screen opposite the beam intersection side. These include frosted glass and various diffusers. A preferred screen is a holographic diffusers supplied by Luminit with offices in Torrence, California. <figref idref="f0004">FIG. 3C</figref> provides an illustration of the beam direction sensor system. The reader should note that the screen in this preferred embodiment will be adapted to detect 128 beams with an 8 X 16 pattern corresponding to the mirror arrays 20 and 22 and the fiber-microlens arrays 8 and 10.

0038In these embodiments the MEMS control system includes a processor programmed to provide a closed loop adjustments of pairs of MEMS mirrors (one from each of the two MEMS mirror array) in order to determine appropriate voltage potentials to be applied to the comb drive actuators in order to provide each desired optical path between the two sets of optical fibers. Applicants' tests have shown that once the calibration has been performed there is very little drift in the beam paths under normal conditions. However, significant changes in environmental conditions could require a recalibration. In some embodiments the switch could be adapted to automatically re-calibrate itself periodically or at the direction of operating personnel.

Directing the Beams

0039<figref idref="f0006">FIGS. 5A</figref>, <figref idref="f0007">5B</figref> and <figref idref="f0008">5C</figref> show some simple illustrations of the operation of the preferred embodiment described in detail above. <figref idref="f0006">FIG. 5A</figref> illustrates the connection of a fiber column 1 of bundle 1 with a fiber in column 8 of bundle 6. The reader should note that camera 64 and 66 detect the co-aligned alignment beams passing through dichroic mirror 24 to provide feedback to the MEMS mirror control to adjust positions of the MEMS mirrors to assure that the beam path is the correct one. <figref idref="f0007">FIG. 5B</figref> demonstrates a connection between the same input fiber and a fiber in one of the columns near the center of the fiber-microlens array and <figref idref="f0008">FIG. 5C</figref> demonstrates a connection between the same input fiber and a fiber in column 1 of the fiber-microlens array. The reader should not the position of the alignment beam in <figref idref="f0007">FIGS. 5B</figref> and <figref idref="f0008">5C</figref>.

Application in a Nation Scale Fiber Optic Network

0040<patcit id="pcit0006" dnum="US677590A" dnum-type="L"><text>U.S. patent application Ser. No. 10/677,590</text></patcit> that has been incorporated herein by reference describes a nation-scale high speed communication network that utilizes all optical switches. The switches described herein would work very well in a network like the ones described in that application. In the preferred embodiment described in the '590 application, the center wavelength of the beams in each fibers is about 1.57 micron (corresponding to 193.1 THz) with a usable bandwidth of 15,000 GHz (between 186,000 GHz to 201,000 GHz). As many as 300 separate communication channels (at 50 GHz each) may be carried in the beam in each fiber. Each of the 300 separate channels (at 50 GHz each) can be divided up into still smaller frequency ranges such as six sub-frequency ranges with 4 GHz spacings. This would permit 1200 signals to be transmitted simultaneously in each fiber, so with 256 fibers per bundle, we could theoretically transmit more than 300,000 separate signals simultaneously per fiber bundle. These signals must be inputted separately at transmit locations to produce the combined beams and separated out from all other signals to be received by each signal recipient. This is referred to as frequency division multiplexing Time division multiplexing permits additional multiplication of the number of separate communications that can be handled during any particular time period.

0041<figref idref="f0020">FIG. 11</figref> is a modified version of <figref idref="f0001">FIG. 1</figref> from the '590 patent application. Here we consider a nation-scale network with end users assigned to one of 250 area codes with roughly equal numbers of end users in each. For instance, area code #1 has been assigned to San Diego, #40 to Seattle, #200 to Washington D.C., and #240 to some subset of international users. The proposed network can have about 400,000 User Nodes per Area Code. As seen in <figref idref="f0001">FIG. 1</figref>, optical cross-connect switches 2K associated with each area code are located at mesh nodes 4K tied together in a mesh network which allows switching of optical signals from any particular area code to any other area code. The particular mesh network would make maximum use of intercity fiber trunk lines which have already been installed.

0042In this preferred embodiment all long haul communication is through optical fibers operating in the wavelength range centered at about 1.57 micron (corresponding to about 193.1 THz). The network is designed to operate at frequencies between 186 THz to 201 THz for a total bandwidth of 15,000 GHz. At 50 GHz spacings, this provides 300 "color" channels per optical fiber. Four separate fibers provide a total of 1200 communication channels.

0043There are in this embodiment 1200 separate wideband communications channels (each with 15 GHz of useable optical bandwidth) leaving and entering each area code. We call these wideband channels FiberColors, and they are distributed as 300 different DWDM wavelengths (standard 50 GHz spacing) on 4 separate fibers. The preferred optical network operates in the C and L bands at a center frequency at 1570 nm (193.1 THz). Thus there are eight lit fibers between an Area Code and its corresponding switch, four for outgoing traffic and four for incoming traffic as shown at 6K on <figref idref="f0020">FIG. 11</figref>.

0044If we divide 1200 FiberColors by the number of area codes (300) we have an average of 4.8 FiberColors per area code. However, the 1200 outgoing FiberColors from any particular area code (say San Diego) are allocated based on usage demand to the 250 area codes with a total bandwidth of about 5 THz per fiber. For instance, the FiberColors for traffic from San Diego might be allocated at a particular time as follows: ten for traffic to Washington, six for traffic to Seattle, one for traffic to Atlanta, etc., until all 1200 FiberColors are accounted for. It is expected that the actual allocation will be automatically adjusted periodically as demand shifts with time of day and day of the week. Therefore, at any particular time, the switches must be configured so that each FiberColor from each origination area code is guided through the network to its destination area code without interference. (That is, the same fiber cannot be used simultaneously for two FiberColors operating at the same wavelength). It was not immediately obvious that this could be done, but applicants have developed an algorithm for accomplishing this task which appears to be robust and to converge in a practically short time. We call this algorithm a Magic Square Algorithm, because the underlying matrices of FiberColors which need to be allocated have rows and columns which add up to the same number. This problem of allocating FiberColors along with its solution is discussed in detail in the section of this specification entitled "Magic Square Software." The solution of this problem is a key technical innovation, as it enables the deployment of a nation-scale all-optical network with a relatively small number of channels without the disadvantages of having to convert any optical signal to an electrical signal or to another DWDM wavelength between the data source and the data destination area codes.

0045Operation of the switches 2K at the Mesh Nodes 4K is discussed in more detail in the '509 patent application." In preferred embodiments, all DWDM wavelengths are de-multiplexed before optical switching, and then re-multiplexed after switching. No wavelength separation is required at a resolution finer than the standard 50 GHz DWDM spacing, so that standard components can be used. (Finer channel resolution only occurs within the source and destination area codes). Customized switches which combine wavelength separation with the optical switching may also be possible. Optical Amplifiers (such as erbium doped fiber amplifiers) are used throughout the network as necessary to maintain appropriate optical signal strength.

Other Control Techniques

0046Referring now to <figref idref="f0019">FIG. 10</figref>, a block diagram of the control system of the Optical Cross Connect Switch of the present invention is shown and generally designated 400. Control system 400 includes a computer 402 containing a real time computer 404, a telecommunications interface 406, and a digital storage device 408. Computer 402 is a system capable of making the computations required to implement a closed-loop feedback control system. It may be comprised of analog or digital electronics, or may be implemented with optical computations units. In a preferred embodiment, the computer consists of digital electronics with at least one component capable of computation, and with at least three digital interfaces. The first interface would be capable of receiving the digitized optical feedback signals, the second interface would be capable of transmitting command signals to the analog electronic driver required for actuation of beam directors 16 and 18. And the third interface would be capable of receiving the network configuration command from an external source and transmitting the state of the optical switch. Other interfaces may be required for certain implementations.

0047In a preferred embodiment, the digital computation electronics could consist of one or more general purpose processors, such as a commercial available digital signal processor or other central processing unit, or might be on e or more application specific integrated circuits design specifically for this task. The digital interfaces could consist of any one of a large variety of parallel or serial links and may conform to some industry standards, or may be custom for a particular implementation.

0048Telecommunication interface 406 provides an electronic interface between computer 402 and a telecommunication exchange via interconnect 410. In a typical environment incorporating the switch of the present invention, interconnect 410 will receive switching information, including the input fibers and the output fibers, which are to be optically coupled. A standard format for receiving this information may be established by a particular telecommunications network, but it is to be appreciated that regardless of the particular protocol, this information will contain particular switch configurations which may be implemented by the present invention.

0049Digital storage device 408 may include both temporary and permanent digital memory media. For example, digital storage device 408 may include random access memory for manipulation of data, and programmable read only memory for storage of programmed computer sequence steps, and may include tables of offset values.

0050Computer 402 is electrically connected to digital interface 414 via electrical connection 412. Digital interface 414 contains high voltage amplifiers, and digital to analog converters that convert digital information from computer 402 to the analog signals necessary to control the mirror elements. Digital interface 414 also transmits and receives any necessary digital data between computer 402 and beam directors 420. (The reader should not that for simplicity <figref idref="f0013 f0014 f0015 f0016">FIG. 8</figref> shows only one mirror and one sensor and should recognize that each cross connection beam is controlled by the positioning of two mirrors and beam directions are monitored by two sensors. However, a preferred technique for maintaining beam position is to adjust only one mirror at a time. It will often be necessary to tweak one mirror in a beam path then tweak the other one several times before perfect control is established. This can all be programmed to occur automatically or as directed by an outside control device or operator.

0051The pivot controls for a single MEMS mirror receive the electronic signals from interface 414 and drive the MEMS mirror its two particular rotational positions in order to direct cross-connect beam in its desired directions. In order to ensure that the MEMS mirrors are properly positioned, the optical sensors measure the position of the alignment beam and provide optical feedback as described above. Analog interface 426 contains analog signal conditioning components, including analog amplifiers and analog to digital converters, which receive the analog signals from optical sensor 422 and generate digital signals for transmission along electrical connection 428 to computer 402. Computer 402 receives the electronic information from sensor 422 regarding the position of the alignment beam, and compares this position to the position contained in the memory 408 to determine whether the beam director elements 420 in beam directors 16 and 18 are properly positioned. If there is a difference between the position of the alignment beam measured by sensor 422 and the position data contained in memory 408, computer 402 adjusts the electronic signals sent to digital interface 414 to modify the rotational position of beam director elements 420 and reposition the alignment beam within the sensor. The position of the alignment beam is then once again measured by optical sensor 422, and the adjustment to the rotational positions of the beam directing elements is repeated if necessary. By properly positioning the alignment beam in this manner, the proper position of the communication beam is achieved without any interference with or measurement of the communication beam itself.

Example of Operation of the Invention

0052In operation, the preferred switch of the present invention transmits an optical signal from an input fiber in bundle 4 to an output fiber in bundle 6. The operation of a preferred embodiment of the present invention is perhaps best understood with reference to <figref idref="f0002">FIGS. 1A and 1B</figref>.

0053Reconfiguration of the input-output mapping of fibers, the switch fabric, is accomplished as follows. Upon receipt of a reconfiguration command, the beam steering mirrors of the affected channels immediately perform an open loop step, moving from their current position to a new position appropriate for completing the commanded reconfiguration. During the open loop step, control feedback is terminated on the affected channels. When the beam steering elements are near their new positions, the alignment beam falls on the portion of the sensor correspond to the new output fiber indicated by the configuration command. At this point, closed loop servo control is re-initiated and the new connection is finalized. During the open loop step, the alignment beam for all the other repositioning channels may be switched off in order to eliminate any contamination of the servo feedback signal of non-switched channels by the guidance beams of the switching channels.

0054In a preferred embodiment, the servo loop may operate only on the mirror elements in the second mirror array. In alternative embodiments, the servo loop may operate on the mirror elements in the first mirror array, the second mirror array, or both the first and second mirror arrays. In one embodiment, the calibration of each of the mirror elements in the first array is sufficiently accurate that it is possible to position these elements with an open loop signal such that substantially all of the optical energy of the alignment and communications beams corresponding to each element will fall on the intended target mirror element in the second beam director array. The open loop pointing of the beam directing mirror elements is calibrated at manufacture, and periodically throughout the lifetime of the device, insuring that the open loop pointing accuracy is high. Also, the open loop pointing accuracy of the mirrors in the first array needs to be no better than a few percent of the full stroke since small errors in position of the first elements are, in effect compensated by the closed loop servo control system operating on elements in the second mirror array. This initial open loop pointing, in combination of the feedback control of the second beam directing elements accurately positions the communication beam onto the center of the output fiber.

Multiplexing and De-Multiplexing

0055An important application of the switch described above is as a part of a frequency multiplexing or de-multiplexing operations. As explained above multiplexing typically involves adding signals at specific frequencies to a trunk line optical fiber carrying other frequencies and de-multiplexing is the opposite. In each case switches are needed to direct the resulting optical signals into the desired direction, typically separate optical fibers. Several multiplexing and de multiplexing techniques are available as discussed in the background section including the use of thin film filters. However, Applicants preferred multiplexing/de-multiplexing technique is one that they have developed that provides substantial advantages over prior art techniques. This technique utilizes porous silicon optical filters. This technique produces results very similar to thin film filters, but the porous silicon filters can be produced much quicker and with much less expense. A short summary description of these porous silicon optical filters is provides below and pictured in <figref idref="f0021">FIGS. 12A and 12B</figref>.

Porous Silicon Optical Filters

0056<figref idref="f0021">FIG. 12A</figref> is a depiction of a portion of a silicon wafer with six layers of varying index of refraction etched in its surface with an electric current supported acid etch porous silicon technology. Porous silicon (PSi) technology is an emerging technology that has many potential applications. A silicon wafer is submerged in hydrofluoric acid (HF) and a current is passed through it. The HF reacts with the silicon and etches nano-pores into the surface. The diameter of the pores is determined by 3 parameters: the current, the HF concentration, and the dopant level of the silicon. Preferred HF concentrations are in the range of about 25 percent to about 50 percent. Preferred silicon doping is about 2.5 X 10¹⁶ ions/cm³ to about 2.5 X 10¹⁷ ions/cm³. The etching always occurs at the interface between the HF solution and the silicon substrate at the bottom of the pores. This makes it possible to etch deep into the silicon and form a thick layer of PSi. The diameter of the pores can be changed during the etching process by varying the current. A larger current increases the diameter of the pores, and a smaller current decreases their diameter. In this manner the porosity of the silicon can be varied as a function of depth. Larger porosity (larger diameter pores) makes the silicon less dense which decreases the index of refraction of the PSi layer. And smaller porosity increases the index. Hence, the index of refraction of the PSi can be varied. This ability to vary the index as a function of depth enables the formation of optical filters. <figref idref="f0021">FIG. 12B</figref> is a drawing showing the general shape of the pours (produced in a silicon substrate 48) greatly magnified in a surface simulating 12 layers. Silicon has an index of refraction of about 3.5 and air has an index of refraction of about 1.00. Since the light we are concerned with has wavelengths much larger than the pores, the light responds to each simulated layer as if it had an index of refraction equal to the weighted average of the air and the silicon volume making up each layer. For example, a low index of refraction is indicated at 51 and a high index of refraction is indicated at 53. <figref idref="f0022">FIG. 12C</figref> shows a graph of current and time utilized to produce the pattern shown in <figref idref="f0021">FIG. 12B</figref> and <figref idref="f0022">FIG. 12D</figref> shows the reflectance from the surface of light at wavelengths of 820 nm to 850 nm. This filter was made with 24 layers as indicated in <figref idref="f0022">FIG. 12C</figref>. The filter can be made narrower by additional layers. <figref idref="f0023">FIG. 12 E</figref> shows the results with 29 layers. <figref idref="f0023">FIG. 12F</figref> is a reflection graph of a porous silicon designed for the wavelength range regularly used for fiber optic communication.

Rugate Porous Silicon Filters

0057The types of filter that can be formed in PSi is intriguing. Traditional thin-film filters are made by alternating between thin (quarter-wave thick) layers of a high index material and a low index material. In PSi the index variation oscillates continuously between a high and a low value. This type of filter is called a "Rugate filter". Some interesting effects can be achieved, such as eliminating unwanted side-lobes from a narrow band-pass filter. <figref idref="f0024">FIG. 12G</figref> shows porosity vs depth in the wafer of a Rugate filter made by Applicants and <figref idref="f0024">FIG. 12H</figref> shows the resulting reflection profile as a function of wavelength. The graph shows three very narrow transmission bands.

0058<figref idref="f0025">FIGS. 12I and 12J</figref> show how porous silicon filter unit can be arranged for multiplexing and de-multiplexing. <figref idref="f0025">FIG. 12I</figref> shows four fibers 54 providing an incoming communication beam comprising many signals at wavelength bands λ₁, λ₂ ... λ_n. Porous silicon filters designed to transmit individual wavelength are shown at 56. For example, filter 56A transmits wavelength range λ₁ and reflects all other wavelength bands. Filter 56B transmits another wavelength band and reflects all other wavelength bands. The wavelength bands are focused into individual optical fiber by lens arrays 58. This design as shown in <figref idref="f0025">FIGS. 12I and 12J</figref> uses a series of very narrow-band filters to sequentially separate one wavelength channel at a time from the DWDM beam. At each filter the selected channel is transmitted and the remaining wavelengths are reflected. If visible light is used the silicon substrate may be thinned or removed so that the transmitted light is not completely absorbed. For most optical telecommunications applications silicon is basically transparent to the wavelengths used and the silicon substrate may remain if the doping concentration is not too high. This filtering technique is conceptually simple but may not be the best for DWDM systems with a large number of wavelength channels since the final wavelength must encounter N-1 (N is the total number of wavelength channels in the DWDM beam) filters before it is separated from all other wavelengths. More complicated designs can reduce the total number of filters that any one beam must pass through by having the first several filters transmit more than one wavelength channel. Narrowband filters may then be used to separate the individual wavelength channels from the decimated beams.

0059Many different prior art Rugate filter designs developed for thin film filters can be used with the porous silicon technique to achieve a narrowband filter with the specifications required for DWDM use. With the porous silicon technique Applicants can simulate stacks of thin film filters by producing multiple series of varying index layers that vary sinusoidally. This is equivalent to the fabrication of two or more relatively broadband reflectors on top of each other. This configuration is the Rugate filter equivalent to the multi-cavity filters found in commercial, discrete, thin-film dielectric filters. This type filter can achieve extremely good wavelength resolution without prohibitive length due to its emulation of Fabry-Perot etalon resonant cavities.

0060Stacking the filters on top of each other produces index of refraction profile that is extremely high across the C band except for a few very narrow wavelength regions that correspond to the resonant bands of the combined filter structure. This is where the phase change on transmission through the broadband filter matches the phase change on reflection from the second broadband filter. The exact wavelength and width of these transmission regions can be controlled, within a limited range by the design of the length and index profile of the broadband reflective filters.

Detection of Secret Wavelength

0061The present invention can be easily adapted to detect the presence of a tracker wavelength. These special tracker wavelengths can be added to a communication beam in order to track the beam as it makes its way through a completed optical route. The present invention can easily be programmed to periodically extract beams in each fiber to direct it to a testing unit to test for the tracker wavelength.

Adjusting Beam Intensity

0062Another useful application of the present invention is that it can be easily programmed to adjust the intensity of particular beams. It is important in communication systems that the intensity of various beams in a system be relatively uniform in intensity, especially just prior to amplification. The present invention can easily reduce the intensity of any beam passing through the switch by merely slightly misaligning the beam using units 16 and 18 to provide just the desired amount of misalignment. The more the beam is misaligned the more the beam is attenuated because smaller portions of it are focused into the output fiber.

Other Sizes

0063In the present embodiment, the configuration is an 8 X 16 array designed to be used to connect 8 X 16 fiber optic bundles. However, readers should appreciate that the present invention may be scaled to many other sizes, smaller or larger, without any significant increase in complexity of the manufacturing, alignment, or corresponding control system. Also additional or fewer spare channels may be provided based on experience.

Other Variations

0064While there have been shown what are presently considered to be preferred embodiments of the present invention, it will be apparent to those skilled in the art that various changes and modifications can be made herein without departing from the scope and spirit of the invention. For example, the switch shown in <figref idref="f0002">FIG. 1A</figref> could be scaled up or down as desired. The field programmable gate arrays used to control the mirrors can be replaced with more economical integrated circuits. The MEMS mirrors may be controlled with other variable voltage sources and larger capacitors could be utilized. Many types of automatic controls can be incorporated into the switches or used to control the switches. Filters other than porous silicon filters such as thin film filters could be used for multiplexing and de-multiplexing. Fewer or more or zero spare channels could be provided. The source of the alignment beam may be a resonant cavity light emitting diode. In some embodiments alternate alignment techniques could be used. Many other techniques are available for aligning an alignment beam with the communication beam in each fiber of the switch. There are many alternate methods of arranging the incoming fibers in the input portion of the switch. Small mirror arrays other than MEMS mirror arrays could be substituted for the MEMS mirror arrays. One processor could be used instead of the two shown in <figref idref="f0002">FIGS. 1A and 1B</figref>. There are a large number of applications for the present invention other than the two important ones discussed in detail. For example, the switch could be used in local intra-office or intra-factory communication systems where very high data rate communication is important. Signal grooming features, gain control and amplifiers could be incorporated into the switch. ROADM units of the present invention can be used to support express local and regional communication traffic. The switch of the present invention can be utilized in all of the system architectures including point-to-point, ring (hubbed and meshed). Therefore, the scope of the patent should be determined by the appended claims and not by the examples that have been given.