Electronic damping of MEMS devices using a look-up table

[0002] This application is a continuation-in-part of copending U.S. patent application Ser. No. 09/783,730, filed Feb. 13, 2001 by Kruglick, et al., entitled METHOD AND APPARATUS FOR ELECTRONIC DAMPING OF COMPLEX DYNAMIC SYSTEMS, herein incorporated by reference in its entirety. This application is related to U.S. patent application Ser. No. 09896021, by Kruglick, et al., entitled METHOD FOR DETERMINING AND IMPLEMENTING ELECTRICAL DAMPING COEFFICIENTS, filed herewith, herein incorporated by reference in its entirety.

[0003] Optical switching plays an important role in telecommunication networks, optical instrumentation, and optical signal processing systems. Optical switches can be used to turn the light output of an optical fiber on or off with respect to an output fiber, or, alternatively, to redirect the light to various different fibers, all under electronic control. Optical switches that provide switchable cross connects between an array of input fibers and an array of output fibers are often referred to as “optical cross-connects.” Optical cross-connects are a fundamental building block in the development of an all-optical communications network.

[0004] There are many different types of optical switches. One general class of optical switches may be referred to as “bulk optomechanical switches” or simply “optomechanical switches.” Such switches employ physical motion of one, or more, optical elements to perform optical switching. An optomechanical switch can be implemented either in a free-space approach or in a waveguide (e.g., optical fiber) approach. The free-space approach is more scalable compared to the waveguide approach.

[0005] In optomechanical switches employing the free space approach, optical signals are switched between different fibers by a number of different methods. Typically, these methods utilize selective reflection of the optical signal off of a reflective material, such as a mirror, into a fiber. The optical signal passes through free space from an input fiber to reach the mirror, and after reflection, passes through free space to an output fiber. The optical signals are typically collimated in order to minimize coupling loss of the optical signal between an input and output fiber.

[0006] Micro-Electro-Mechanical Systems or MEMS are electrical-mechanical structures typically sized on a millimeter scale or smaller. These structures are used in a wide variety of applications including for example, sensing, electrical and optical switching, and micron scale (or smaller) machinery such as robotics and motors. MEMS structures can utilize both the mechanical and electrical attributes of material to achieve desired results. Because of their small size, MEMS devices may be fabricated utilizing semiconductor processing methods and other microfabrication techniques such as thin film processing and photolithography. Once fabricated, the MEMS structures are assembled to form MEMS devices.

[0007] MEMS structures have been shown to offer many advantages for building optomechanical switches. Namely, the use of MEMS structures can significantly reduce the size, weight and cost of optomechanical switches. The switching time can also be reduced because of the lower mass of the smaller optomechanical switches.

[0008] Movable MEMS structures are capable of oscillating uncontrollably if they are not damped. Such oscillation is due to MEMS structure design and/or fabrication. For example, very low friction in the hinges of MEMS structures allows them to move easily and repeatedly bounce off of stationary objects such as motion stops. Known methods for damping MEMS structures do not provide quick and efficient damping for all types of structures. Thus, there is a need for a method and/or apparatus that provides quick and efficient damping of MEMS structures.

[0009] In one embodiment, a MEMS apparatus having a MEMS array including a plurality of MEMS devices is provided. In some embodiments, each of the plurality of MEMS devices includes a movable structure and a second structure. In addition, in some embodiments, a plurality of signal sources are coupled to the plurality of MEMS devices so as to be capable of supplying actuation signals for actuating the movable structure to impact the second structure. Further, in some embodiments, at least one processor is coupled to the plurality of signal sources to control the actuation signals, and is configured such that each of the plurality of MEMS devices is provided with a corresponding custom actuation signal.

[0010] FIG. 1 is a top view illustrating a two-dimensional optical switch having MEMS switching cells.

[0011] FIG. 2 is a simplified perspective view illustrating one of the MEMS switching cells shown in FIG. 1 in the transmission position.

[0012] FIG. 3 is a simplified perspective view illustrating the MEMS switching cell shown in FIG. 2 in the reflection position.

[0013] FIG. 4 is a plot illustrating an actuation signal and resulting output signal when the actuation signal is applied to the MEMS switching cell shown in FIG. 2.

[0014] FIG. 5 is a simplified perspective view illustrating the actuator plate of the MEMS switching cell shown in FIG. 2 bouncing off of the motion stop.

[0015] FIG. 6 is a plot illustrating an actuation signal in accordance with an embodiment of the present invention and the resulting output signal when the actuation signal is applied to the MEMS switching cell shown in FIG. 2.

[0016] FIG. 7 is a plot further illustrating the actuation signal shown in FIG. 6.

[0017] FIG. 8A is a representative plot illustrating the torque and velocity that result from the actuation signal shown in FIG. 4.

[0018] FIG. 8B is a representative plot illustrating the torque and velocity that result from the actuation signal shown in FIG. 6.

[0019] FIGS. 9A, 9B, 9C, 9D, 9E and 9F are plots illustrating actuation signals in accordance with alternative embodiments of the present invention.

[0020] FIG. 10 is a plot illustrating an actuation signal in accordance with another embodiment of the present invention and the resulting output signal when the actuation signal is applied to the MEMS switching cell shown in FIG. 2.

[0021] FIG. 11 is a block diagram illustrating an exemplary system for generating the actuation signal shown in FIG. 6 in accordance with one embodiment of the present invention.

[0022] FIGS. 12A and 12B are simplified perspective views illustrating the use of an actuation signal in accordance with another embodiment of the present invention with a complex dynamic system that is actuated using electromagnetic attraction.

[0023] FIGS. 13A and 13B are simplified perspective views illustrating the use of an actuation signal in accordance with another embodiment of the present invention with a MEMS switching cell that is actuated using electromagnetic attraction.

[0024] FIG. 14 is a block diagram illustrating an exemplary system for constructing and supplying the actuation signal to a MEMS array in accordance with an embodiment of the present invention.

[0025] FIG. 15 is a plot illustrating the effects of varying the offset and duration of the divot in the actuation signal of FIG. 6.

[0026] FIG. 16 illustrates a method for determining a damping characteristic for each of the switches in an array in accordance with an implementation of the present invention.

[0027] FIG. 17 is a block diagram illustrating an exemplary system for constructing and supplying the actuation signal to a MEMS array in accordance with an embodiment of the present invention.

[0028] FIG. 18 is a block diagram illustrating an exemplary system for constructing and supplying the actuation signal to a MEMS array in accordance with an embodiment of the present invention.

[0029] This application hereby incorporates by reference in its entirety, U.S. patent application Ser. No. 09/783,730, by Kruglick, et al., entitled METHOD AND APPARATUS FOR ELECTRONIC DAMPING OF COMPLEX DYNAMIC SYSTEMS, and hereby incorporates by reference in its entirety U.S. patent application Ser. No. 09,896,021, by Kruglick, et al., entitled METHOD FOR DETERMINING AND IMPLEMENTING ELECTRICAL DAMPING COEFFICIENTS, filed herewith.

[0030] The following description is not to be taken in a limiting sense, but is made for the purpose of describing one or more embodiments of the invention. The scope of the invention should be determined with reference to the claims.

[0031] Referring to FIG. 1, there is illustrated an optical crossbar switch 100, or simply, an optical switch 100. The optical switch 100 is a two-dimensional (2D) optical switch that is capable of providing switchable cross connects between an array of input fibers 103 and an array of output fibers 121. Specifically, the optical switch 100 provides an array of free-space optical connections between the input and output fibers 103, 121. This enables each of a plurality of optical input channels to be directed to a desired optical output channel. Each of the optical input channels may also be dropped via the dropped channel fibers 123, or other channels may be added via the add channel fibers 111 to replace certain input channels. That is, the optical switch is capable of performing optical switching as well as wavelength-selective add/drop filtering.

[0032] A cable 102 containing a plurality of input fibers 103 is incident upon a demultiplexer 104, which separates the optical beam carried by each fiber into a number of input channels. The input channels are provided, via a collimator array 106, to an array 108 of optomechanical switching cells. Similarly, a cable 110 containing a plurality of add channel fibers 111 is incident upon a demultiplexer 112 followed by a collimator array 114. The optomechanical switching cells can be configured to direct particular input channels to desired output channels, as well as to implement the channel add/drop functionality mentioned above. Separate multiplexers 116, 118 are provided for multiplexing the output channels and the “dropped” channels onto the fibers 121, 123 of separate cables 120, 122, respectively.

[0033] The array 108 of optomechanical switching cells are connected to a substrate 124. The optomechanical switching cells are preferably fabricated in accordance with Micro-Electro-Mechanical Systems (MEMS) technology. Each of the switching cells includes a mirror and an actuator. The lenses of the collimator arrays and the mirrors of each MEMS switching cell are aligned so that each switching cell can selectively intercept a beam output from one of the collimator arrays.

[0034] FIGS. 2 and 3 are simplified diagrams illustrating an exemplary MEMS switching cell 200 that may be included in the array 108 of optomechanical switching cells. A micromirror 202 is mounted to an actuator plate 204. The actuator plate 204 is hingedly connected to the substrate 124 with a torsion hinge 206, which by way of example may be of the type described in U.S. patent application Ser. No. 09/697,762, filed Oct. 25, 2000, entitled “MEMS Optical Switch with Torsional Hinge and Method of Fabrication Thereof”, by inventor Li Fan, and identified by the entire contents of which are hereby expressly incorporated by reference into the present application as if fully set forth herein. The micromirror 202 is shown in a vertical position with the mirror surface being perpendicular to the substrate 124. Another torsion hinge (not shown) may be used to permit the micromirror 202 to pivot relative to the actuator plate 204.

[0035] FIG. 2 illustrates the switching cell 200 in its transmission state, and FIG. 3 illustrates the switching cell 200 in its reflection state. In the reflection state, the actuator plate 204 rests on a motion stop 208 (or other landing structure) mounted on the substrate 124. By way of example, the motion stop 208 may comprise a jack stop of the type described in U.S. patent application Ser. No. 09/697,767, filed Oct. 25, 2000, entitled “MEMS Microstructure Positioner and Method of Fabrication Thereof”, by inventor Li Fan, and identified by the entire contents of which are hereby expressly incorporated by reference into the present application as if fully set forth herein. During operation, an optical beam 210 is incident at an approximately 45° angle from the normal of the micromirror 202. By pivoting the actuator plate 204 about the torsion hinge 206, the micromirror 202 is moved in and out of the path of the optical beam 210, switching the output of the optical beam 210 between a reflection direction 212 and the transmission direction (210 of FIG. 2).

[0036] The actuator plate 204, an actuator electrode 214 mounted on the substrate 124, and the gap therebetween form an electrostatic actuator. In one embodiment, the switching cell 200 is activated by a circuit that selectively develops an electrostatic force between the actuator plate 204 and the actuator electrode 214. This force causes the actuator plate 204 and the micromirror 202 mounted thereto to pivot in angular position relative to the substrate 124, i.e., in the direction indicated by arrow 216. In other words, when a voltage bias, symbolized by voltage source 218, is applied between the actuator plate 204 and the actuator electrode 214 (or substrate 124), the actuator plate 204 is caused to draw close and contact the motion stop 208. Thus, a voltage bias is applied between the actuator plate 204 and the actuator electrode 214 to cause the switching process by electrostatic attraction. Alternate actuation methods/forces may also be used, such as for example, electromagnetic, thermal expansion, and piezoelectric. Such alternate actuation methods/forces will be discussed below.

[0037] Because of the presence of the motion stop 208, the MEMS switching cell 200 forms a complex dynamic system for purposes of analyzing the angular movement of the actuator plate 204. A complex dynamic system may also be referred to as a nonlinear system. The MEMS switching cell 200 forms a complex dynamic system because the motion stop 208 causes the actuator plate 204 to move in a vibrating manner when the actuator plate 204 impacts the motion stop 208.

[0038] In order to bring the actuator plate 204 to rest on the motion stop 208 as quickly as possible, damping is used. The damping of moving structures in complex dynamic systems is difficult, and such systems do not lend themselves to conventional methods of damping. In order to illustrate this difficulty, reference is made to FIG. 4, which illustrates a conventional actuation signal 230 and the resulting output signal 232 for the MEMS switching cell 200. In accordance with known principles, the actuation signal 230 can be made to provide effective damping in many non-complex dynamic systems that do not have motion stops. FIG. 4 illustrates the result of application of the actuation signal 230 to the switching cell 200, which is a complex dynamic system.

[0039] The actuation signal 230 is applied to the electrostatic actuator, i.e., between the actuator plate 204 and the actuator electrode 214, to create a force that pulls the actuator plate 204 towards the motion stop 208. The resulting output signal 232 represents the optical power of the output optical beam 212 that is reflected off of the micromirror 202.

[0040] As illustrated, shortly after the actuation signal 230 is activated, the output signal 232 starts oscillating and then eventually settles in its on state, i.e., full optical power of the reflected output optical beam 212. The output signal 232 oscillates because the downward force imparted on the actuator plate 204 by the actuation signal 230 causes the actuator plate 204 to bounce off of the motion stop 208 several times. This bouncing, depicted in FIG. 5, causes the micromirror 202 to move through and within the optical path several times, which switches the output optical beam 212 between the reflection and the transmission states several times creating the oscillating output signal 232. In this example, it takes 11.36 milliseconds (ms) for the output signal 232 to reach 90% settlement. As evidenced by this rather lengthy settlement time, the actuation signal 230 does not provide quick and efficient damping in complex dynamic systems that have motion stops.

[0041] It is desirable to minimize the settlement time of the actuator plate 204 in order to reduce the switching time. Such a reduction in the switching time of the MEMS switching cells increases the performance of the optical switch 100. The settlement time of the actuator plate 204 can be reduced by employing a damping method that provides more effective damping than the damping provided by the actuation signal 230.

[0042] Referring to FIG. 6, there is illustrated an actuation signal 240 in accordance with one embodiment of the present invention. The actuation signal 240 provides highly effective damping of a structure in a complex dynamic system having a positive motion stop. Similar to above, the actuation signal 240 is the signal that is applied between the actuator plate 204 and the actuator electrode 214 to create a force that pulls the actuator plate 204 towards the motion stop 208. The resulting output signal 242 represents the optical power of the output optical beam 212 that is reflected off of the micromirror 202.

[0043] As illustrated, shortly after the actuation signal 240 is activated, the output signal 242 rises above 90% of its maximum level and stays there. The oscillations in the output signal 242 are small and do not fall below the 90% level. In this example, it takes only 3.5 ms for the output signal 242 to reach 90% settlement. Thus, the actuation signal 240 provides a greatly reduced settling time as compared to the settling time achieved with the actuation signal 230 of FIG. 4.

[0044] The actuation signal 240 provides highly effective damping because it reduces the magnitude of bouncing off of the motion stop 208 and reduces the length of time that the actuator plate 204 bounces off of the motion stop 208. These features reduce the settling time of the actuator plate 204. In this way, the actuation signal 240 provides electronic damping of the actuator plate 204.

[0045] Referring to FIG. 7, the actuation signal waveform 240 provides highly effective and simple to implement electronic damping for the actuator plate 204 (of FIGS. 2 and 3) at the mechanical motion stop 208 (of FIGS. 2 and 3). Important characteristics of the signal 240 are a decrease in drive timed with respect to the first impact of the moving actuator plate 204 on the motion stop 208. Specifically, the actuation signal waveform 240 preferably begins with an acceleration phase 244 during which the actuator plate 204 accelerates towards the motion stop 208. In one embodiment, the acceleration phase 244 involves the application of a substantially constant magnitude voltage between the actuator plate 204 and the actuator electrode 214 (of FIGS. 2 and 3). The acceleration phase 244 is followed by a coast phase 246 where the acceleration of the actuator plate 204 is decreased as it approaches the motion stop 208. The coast phase 246 involves decreasing the magnitude of the actuation voltage in a linear manner prior to the actuator plate 204 impacting the motion stop 208. Next, a segue phase 248 increases the downward force on the actuator plate 204 at about the time the actuator plate 204 impacts the motion stop 208. The segue phase 248 involves increasing the magnitude of the actuation voltage in a linear manner. Finally, a hold down phase 250 applies maximum downward force to the actuator plate 204 to hold it against the motion stop 208. The hold down phase 250 involves leveling off the magnitude of the actuation voltage.

[0046] FIG. 8A illustrates the torque and velocity of the actuator plate 204 (of FIGS. 2 and 3) resulting from the actuation signal 230 (of FIG. 4), and FIG. 8B illustrates the torque and velocity of the actuator plate 204 resulting from the actuation signal 240 (of FIG. 6). Torque and the actuation signal voltage have the following approximate electrostatic relationship: $\begin{array}{cc}\mathrm{Torque}\propto \frac{{\mathrm{Voltage}}^2}{{\mathrm{gap}}^2}& \left(1\right)\end{array}$

[0047] where the “gap” is the distance between the actuator plate 204 and actuator electrode 214.

[0048] With the actuation signal 230, the torque and velocity continue to increase right up to the motion stop 208. This causes severe bouncing of the actuator plate 204. In contrast, with the actuation signal 240, the torque and velocity do not continue to increase right up to the motion stop 208. Instead, the torque and velocity decrease somewhat from the values which signal 230 would cause before reaching the motion stop 208, which reduces the strength of the bounce of the actuator plate 204.

[0049] Referring to FIG. 7, important parameters for the actuation signal waveform 240 are: (1) the total time “η” for the acceleration and coast phases 244, 246; and (2) the decrease (or depth) in voltage “d” during the coast phase 246. In one embodiment of the invention, the depth of modulation “d” is chosen to target an approximately 30-50% reduction in drive as the actuator plate 204 touches down on the motion stop 208. By way of example, “d” may be approximately equal to 40-60% of the substantially constant magnitude of the acceleration phase 244. Furthermore, by way of example, the coast phase 246 may involve decreasing the magnitude of the actuation signal from the substantially constant magnitude of the acceleration phase 244 by the value “d” in an amount of time “τ”, where “τ” falls in the range of 1 to 3 ms for the illustrated embodiment.

[0050] The value of “η” is preferably chosen to approximately target the first peak in the output signal 232 in FIG. 4, thus lowering the drive just as touchdown occurs. This way the “divot” formed by the coast and segue phases 246, 248 hits just as the actuator plate 204 (of FIGS. 2 and 3) touches down upon the motion stop 208 (of FIGS. 2 and 3), which minimizes the force at approximately the time of contact. Thus, the acceleration and coast phases 244, 246 comprise a total amount of time approximately equal to the time it takes the undamped actuator plate 204 to reach the motion stop 208.

[0051] Thus, the actuation signal waveform 240 is typically tuned in amplitude or frequency. Namely, the amplitude sets “d” and thus the magnitude of the acceleration and hold down phases 244, 250,and the frequency sets “η” and thus the coast and segue phases 246, 248.

[0052] It was mentioned above that the coast and segue phases 246, 248 preferably involve decreasing and increasing, respectively, the magnitude of the actuation signal in a linear manner. It should be well understood, however, that such linear decreasing and increasing is not a requirement of the present invention. Namely, either or both of the coast and segue phases 246, 248 may alternatively be implemented by decreasing and increasing, respectively, the actuation signal in a nonlinear manner in accordance with the present invention. For example, FIG. 9A illustrates a linear coast phase 262 and a nonlinear segue phase 264, and FIG. 9B illustrates a nonlinear coast phase 270 and a linear segue phase 272. Although the linear scheme provides for ease of implementation, a nonlinear scheme may be employed by the present invention.

[0053] In addition, the acceleration and hold down phases 244, 250 are not required to have the same magnitude. For example, FIG. 9C illustrates the acceleration phase 278 having a greater magnitude than the hold down phase 280, and FIG. 9D illustrates the acceleration phase 286 having a smaller magnitude than the hold down phase 288. Nonlinear coast and segue phases 290, 292 are also shown.

[0054] FIG. 9E illustrates that one or more additional fluctuations 294 may be inserted into the actuation signal between the segue phase 296 and the hold down phase 298 in accordance with the present invention.

[0055] FIG. 9F illustrates that the coast and segue phases 261, 263 may also be implemented as a negative pulse in accordance with the present invention. Namely, the decreased voltage of the coast and segue phases 261, 263 form a pulse, providing in essence a pulsewidth modulated output signal.

[0056] Referring to FIG. 10, there is illustrated an actuation signal 275 in accordance with another embodiment of the present invention. The actuation signal 275 also provides highly effective damping of a structure in a complex dynamic system having a positive motion stop. Similar to above, the actuation signal 275 is applied between the actuator plate 204 (of FIGS. 2 and 3) and the actuator electrode 214 (of FIGS. 2 and 3) to create a force that pulls the actuator plate 204 towards the motion stop 208 (of FIGS. 2 and 3). The resulting output signal 277 represents the optical power of the output optical beam 212 (of FIG. 3) that is reflected off of the micromirror 202 (of FIG. 3).

[0057] The actuation signal 275 includes only two phases, namely, an acceleration phase 279 and a hold down phase 281. Because there are only two phases, the actuation signal 275 may be referred to as a “two-step” actuation signal. In the illustrated embodiment, the acceleration phase 279 involves the application of a substantially constant magnitude voltage between the actuator plate 204 and the actuator electrode 214. Following the acceleration phase 279 is the hold down phase 281, which involves the application of a substantially constant magnitude voltage between the actuator plate 204 and the actuator electrode 214 that is smaller than the magnitude applied during the acceleration phase 279.

[0058] As illustrated, shortly after the beginning of the acceleration phase 279, the output signal 277 starts a wide oscillation. This wide oscillation is a result of the actuator plate 204 accelerating toward the motion stop 208 and then bouncing off of it with fairly large magnitude bounces. The hold down phase 281 causes the magnitude of the bounces to decrease and the output signal 277 to eventually settle in its on state, i.e., full optical power of the reflected output optical beam 212. In this example, the acceleration phase 279 continues for 3 ms, and the magnitude of the hold down phase 281 is equal to 40% of the magnitude of the acceleration phase 279. With these values it takes 8.1 ms for the output signal 277 to reach 90% settlement. Thus, the actuation signal 275 provides a reduced settling time as compared to the settling time achieved with the actuation signal 230 of FIG. 4.

[0059] Referring to FIG. 11, there is illustrated an exemplary high voltage system 300 for generating the actuation signal waveform 240 in accordance with one embodiment of the present invention. An input voltage is received at V_input. The input voltage is typically a conventional digital signal falling in the range of 0-5 Volts. The actuation signal waveform 240 is output at V₂₄₀.

[0060] A microprocessor 302 monitors V_input. When V_inputgoes high, the microprocessor 302 controls a high voltage source 304 to produce a high voltage at output V₂₄₀. Furthermore, the microprocessor 302 controls the high voltage source 304 to produce the acceleration, coast and segue phases 244, 246, 248 in the output of the high voltage source 304. The “divot” formed by the coast and segue phases 246, 248 is included in the high voltage signal produced at V₂₄₀. The high voltage source 304 can be based on a standard drive voltage that, by way of example, may be equal to the magnitude of the hold down phase 250 of the actuation signal 240.

[0061] The above discussion focused on the use of electrostatic attraction to cause the switching process in the switching cell 200 (of FIGS. 2 and 3). Namely, the actuation signal 240 (of FIG. 6) is illustrated as a voltage in the figures. It should be well understood, however, that the actuation signals described herein, i.e., the actuation signal 240 and the actuations signals illustrated in FIGS. 9A, 9B, 9C, 9D, 9E, 9F and 10, may be used with many different types of actuation methods/forces in accordance with the present invention. For example, the actuation signals described herein may be used with electromagnetic attraction, thermal expansion, or piezoelectric attraction. When used with these alternate actuation methods/forces, the actuation signals described herein provide highly effective damping of structures in complex dynamic systems.

[0062] FIGS. 12A and 12B illustrate the manner in which the actuation signals of the present invention can be used to provide highly effective damping in a complex dynamic system 400 that is actuated using electromagnetic attraction. Specifically, the complex dynamic system 400 includes a movable structure 402 that is connected to a hinge 404. The movable structure 402 comprises a ferrous or paramagnetic material and is stopped by a motion stop 406. An actuation signal 408 in accordance with the present invention is applied to an electromagnet 410. In this embodiment, the actuation signal 408 is a current I₄₀₈that causes the electromagnet 410 to create a force that pulls the moveable structure 402 towards the motion stop 406. The actuation signal 408, which may have the waveform shape of any of the actuation signals described above, provides highly effective damping for the same reasons discussed above. Namely, the actuation signal 408 provides a decrease in drive timed with respect to the first impact of the moveable structure 408 on the motion stop 406.

[0063] FIGS. 13A and 13B illustrate a MEMS switching cell 430 that is actuated using electromagnetic attraction. A micromirror 432 is mounted to an actuator plate 434. The actuator plate 434 comprises a ferrous or paramagnetic material and is hingedly connected to the substrate 436 with a torsion hinge 438. FIG. 13A illustrates the switching cell 430 in its transmission state where the optical beam 440 is not reflected. FIG. 13B illustrates the switching cell 430 in its reflection state where the optical beam 440 is switched to a reflection direction 442. In the reflection state, the actuator plate 434 rests on a motion stop 444 (or other landing structure) mounted on the substrate 436.

[0064] The actuator plate 434, an electromagnet 446 mounted on the substrate 436, and the gap therebetween form an electromagnetic actuator. The switching cell 430 is activated by providing an actuation signal 448 to the electromagnet 446. In this embodiment, the actuation signal 448 is a current I₄₄₈. The current I₄₄₈causes the electromagnet 446 to create a force that causes the actuator plate 434 and the micromirror 432 mounted thereto to pivot in angular position relative to the substrate 436, i.e., in the direction indicated by arrow 450. This force causes the actuator plate 434 to draw close and contact the motion stop 444. Thus, by applying the actuation signal 448 to the electromagnet 446, the switching process is effectuated by electromagnetic attraction.

[0065] In accordance with the present invention, the actuation signal 448 may comprise the waveform shape of any of the actuation signals described above, i.e., the actuation signal 240 and the actuation signals illustrated in FIGS. 9A, 9B, 9C, 9D, 9E, 9F and 10. The use of such waveform shapes for the current I₄₄₈provides highly effective damping of the actuator plate 434. Highly effective damping is achieved because the current I₄₄₈provides a decrease in drive timed with respect to the first impact of the actuator plate 434 on the motion stop 444.

[0066] The entire contents of U.S. patent application Ser. No. 09/063,644, filed Apr. 20, 1998, entitled “Micromachined Optomechanical Switches”, and U.S. patent application Ser. No. 09/483,276, filed Jan. 13, 2000, entitled “Micromachined Optomechanical Switching Cell with Parallel Plate Actuator and On-Chip Power Monitoring”, are hereby expressly incorporated by reference into the present application as if fully set forth herein.

[0067] Turning to FIG. 14, as discovered by the present inventors, controlling system damping can be further complicated in a MEMS array 500. This is because each of the switches 500a₀, 500a₁, 500a₂, etc., in the array 500 can have different damping characteristics. This may be due to device non-uniformities, which can result from the fabrication process. For example, switches near the periphery of the array may have different damping characteristics than switches near a central portion of the array. Such a situation further adds to the complexity of providing effective damping for the system.

[0068] Shown in FIG. 14, a single high voltage source 504 may be utilized to provide actuation signals to all of the switches 500a₀, 500a₁, 500a₂, etc., in the array 500. In such a system, a single actuation signal can be selected to take into account damping non-uniformities occurring in the switch array 500.

[0069] To determine a single actuation signal for operating the array, an actuation signal having a decreased drive portion is selected as discussed above with reference to FIGS. 6, 7, and 9A-9F. In the implementation of FIG. 14, during operation of the switch array 500, the high voltage source 504 is applied not only to a selected switch for switching, as well as to any and all switches already energized. Thus, in this example, the actuation signal should be selected so as to inhibit already energized switches from moving, or otherwise changing their switch state. Hence, in the MEMS optical switch example discussed above with reference to FIG. 5, the depth of modulation “d” is chosen to reduce the drive of the actuator plate or actuator arm 204, but not to allow release of any energized switches in the array 500.

[0070] Turning to FIG. 15, the selected actuation signal 240 is applied to each of the switches of the array 500 to determine a range of values of a damping coefficient γ for each switch that provide settling times in conformance with a predetermined specification. The damping coefficient γ sets the time η (shown in FIG. 7). In addition, the damping coefficient γ is used to define the relationship between an offset of the coast phase 246 with respect to beginning of the acceleration phase 244, and the combined duration of the coast and seque phases 246 and 248 (both shown in FIG. 7). Thus, the damping coefficient γ can be utilized to proportionally vary the offset and the combined duration as illustrated in FIG. 15. Increasing the damping coefficient γ decreases the time η (shown in FIG. 7). FIG. 15 depicts how the damping coefficient γ can be used to proportionally change the offset and duration of the coast and seque phases (shown as divots 247a-h) of the selected actuation signal 240 of FIG. 7. In FIG. 15, multiple divots 247a-hare shown in the selected actuation signal 240 in response to multiple selected damping coefficients γ. Thus, In this way, a range of values for the damping coefficients γ of the selected waveform can be expediently determined for each of the switches in the array 500.

[0071] Turning to FIG. 16, in one implementation for example, an initial value for the damping coefficient γ_ican be determined by using a test station to scan for an initial value of the damping coefficient γ_ithat provides a settling time in conformance with a predetermined specification as illustrated at box 1610. The damping coefficient γ may then be successively increased from its initial value γ_ito determine a maximum value of the damping coefficient γ_maxthat produces a settling time in conformance with the predetermined specification, illustrated in boxes 1620 and 1630. The maximum damping coefficient γ_manis then recorded 1640. The minimum value of the damping coefficient γ_minis determined, such as by returning to the initial damping coefficient γ_iand successively decreasing the damping coefficient γ to ascertain settling times in conformance with the predetermined specification, illustrated by boxes 1650, 1660, and 1670. The minimum damping coefficient γ_minis recorded 1680.

[0072] Typically, this is carried out at an ambient pressure similar to the device operating conditions. As such, it can be carried out after hermetically sealing. of the package containing the MEMS chip.

[0073] Referring to FIG. 14, the ranges of values that produce settling times in conformance with the predetermined specification for each of the switches in the array 500 are used to select an operating damping coefficient γ_operto be utilized by the microprocessor 502 to control the high voltage source 504. In one implementation, the operating damping coefficient γ_operis selected midway between, the largest value of the minimum damping coefficient γ_minfor all the switches in the array, and the smallest value of the maximum damping coefficient γ_maxfor all the switches in the array 500.

[0074] The microprocessor 502 is programmed to construct a single operating actuation signal having the values of the offset and combined duration corresponding to the operating damping coefficient γ_oper. To construct the operating actuation signal, the microprocessor 502 monitors for commencement of an actuation signal to a switch, such as by monitoring the incoming digital user input supplied to the selector 506. After commencement of an actuation signal has been detected by the microprocessor 502, the microprocessor controls the high voltage signal source 504 to provide a divot, or other selected signal reduction at its output. As discussed above, in the implementation of FIG. 14, the divot or other selected signal reduction is supplied not only to the structure being moved but also to any switches currently energized or otherwise receiving a signal from the high voltage source 504.

[0075] It is also possible to construct a custom operating actuation signal for each switch using the determined settling characteristics for each switch. This may be accomplished by using the stored values of the minimum and maximum damping coefficients γ_minand γ_maxfor each switch. Or, a custom damping coefficient γ_custthat provides the best measured settling time for each switch could be stored and utilized by the microprocessor 502 to construct a custom operating actuation signal for each switch. The custom values may be stored in a look-up table format that is accessed by the microprocessor 602 for selecting the corresponding stored custom damping coefficient for each switch. An individual damping coefficient could be stored in the look-up table with reference to the switch's position in the array, i.e. by column and row. The custom values in the look-up table, which can be the corresponding offset and combined duration, can be utilized to construct a custom operating actuation signal for each switch.

[0076] Turning to FIG. 17, it is also possible to employ several high voltage sources 604a-dto provide actuation signals to the switches of the array 600. In the implementation of FIG. 17, one high voltage source 604aprovides actuation signals for a column 600aof switches 600a₀-a₃. This allows switches in the different columns 600a-dto be switched contemporaneously and/or near contemporaneously without having to wait for a single shared high voltage source 502 to return to an acceleration phase voltage. For example, a switch may be actuated before the divot 246, 248, shown in FIG. 7, which is being applied to a switch in another column has finished.

[0077] Further, because obtaining values from a large look-up table can be microprocessor intensive, utilizing several microprocessors 604a-dallows for smaller look-up tables be used in implementing the custom operating actuation signals. For example, one look-up table associated with processor 602astores the values for the switches in column 600a, and another look-up table associated with processor 602bstores the values for the switches in column 600b, and so on. This can reduce costs by allowing slower less expensive microprocessors to be used. Multiple processors also facilitates contemporaneous and/or near contemporaneous switching of switches in the different columns.

[0078] It is possible to provide separate microprocessor controlled high voltage sources for each of the switches in the array 600. In a conventional optical cross-connect, however, no more that one switch in a column is switched at a time. Thus, for the 4×4 array 600, four high voltage sources are sufficient. Further, as array sizes grow to 8×8, 16×16, and beyond, providing a microprocessor, a DAC, and a high voltage supply for each column ultimately will be more practical than providing a microprocessor, a DAC, and a high voltage supply for each switch.

[0079] Turning to FIG. 18, a single microprocessor 602 may be utilized to construct custom operating actuation signals using multiple high voltage sources 604a-d. As discussed above, a single microprocessor requires a larger look-up table, and therefore typically would require a higher performance microprocessor than with the multiple processor implementation of FIG. 17.

[0080] Referring to FIGS. 14, 17 and 18, although shown as a single block, the selector 506, 606 may be a single component, or it may comprise, or equivalently be, several individual selectors. Thus, a selector 506 or 606 may be a selector that is a single or multiple component device.

[0081] Constructing a custom actuation signal for each switch allows better settling time for the switches in the array. The custom actuation signal has been discovered by the present inventors to provide about 50% better settling times.

[0082] Although in some implementations discussed above the offset of the coast phase with respect to the commencement of the acceleration phase and the combined duration of the coast phase and the seque phase can be varied proportionally, it is also possible to independently vary the offset and/or the combined duration. Further, it is possible to vary one of the waveforms controlling parameters, such as the offset, the combined duration, or the independent durations of the coast or seque phases while an other is held constant.

[0083] While the invention herein disclosed has been described by the specific embodiments and applications thereof, numerous modifications and variations could be made thereto by those skilled in the art without departing from the scope of the invention set forth in the claims.