JP4565474B2 - Optical switch and optical switch control method - Google Patents

Description

  The present invention relates to an optical switch and a method for controlling the optical switch.

  As one technique for realizing an optical switch, a technique using a micromirror has been proposed (see Non-Patent Document 1). A conventional optical switch using a micromirror is shown in FIG.

  The optical switch shown in FIG. 14 includes an input port 1a, an output port 1b, an input side micromirror array 2a, and an output side micromirror array 2b. The input port 1a and the output port 1b are each composed of a plurality of optical fibers arranged two-dimensionally, and the micromirror arrays 2a and 2b are each composed of a plurality of micromirror devices 3a and 3b arranged two-dimensionally. . In addition, the arrow in FIG. 14 has shown the advancing direction of the light beam.

  An optical signal emitted from a certain input port 1a is reflected by the mirror of the micromirror device 3a of the input side micromirror array 2a corresponding to this input port 1a, and its traveling direction is changed. As will be described later, the mirror of the micromirror device 3a is configured to be rotatable about two axes, and the reflected light of the micromirror device 3a can be directed to any micromirror device 3b of the output side micromirror array 2b. it can. Similarly, the mirror of the micromirror device 3b is also configured to be rotatable about two axes, and the reflected light of the micromirror device 3b is directed to an arbitrary output port 1b by appropriately controlling the tilt angle of the mirror. be able to. Therefore, by appropriately controlling the tilt angles of the mirrors of the input-side micromirror array 2a and the output-side micromirror array 2b, the optical path is switched, and the arbitrary input port 1a and output port 1b arranged two-dimensionally Can be connected. Here, the optical intensity of the optical signal emitted from the output port 1b is monitored by the output light measuring device. Examples of the output light measuring device include a photodiode (PD) and a Tap-PD that guides and monitors part of the optical power in the fiber to the PD.

  The most characteristic components of such an optical switch are micromirror devices 3a and 3b having mirrors. Conventionally, as shown in FIGS. 15 and 16, a micromirror device has a structure in which a mirror substrate 200 on which a mirror is formed and an electrode substrate 300 on which an electrode is formed are arranged in parallel (non-patent document). Reference 1).

  The mirror substrate 200 includes a plate-shaped frame portion 210, a movable frame 220 disposed in the opening of the frame portion 210, and a mirror 230 disposed in the opening of the movable frame 220. The frame part 210, the torsion springs 211a, 211b, 221a, 221b, the movable frame 220 and the mirror 230 are integrally formed of, for example, single crystal silicon. For example, three Ti / Pt / Au layers are formed on the surface of the mirror 230. The pair of torsion springs 211 a and 211 b connect the frame part 210 and the movable frame 220. The movable frame 220 can rotate about the movable frame rotation axis X of FIG. 15 passing through the pair of torsion springs 211a and 211b. Similarly, the pair of torsion springs 221 a and 221 b couple the movable frame 220 and the mirror 230. The mirror 230 can rotate about the mirror rotation axis Y of FIG. 15 passing through the pair of torsion springs 221a and 221b. The movable frame rotation axis X and the mirror rotation axis Y are orthogonal to each other. As a result, the mirror 230 rotates about two orthogonal axes.

  The electrode substrate 300 has a plate-like base portion 310 and a terrace-like protruding portion 320. The base 310 and the protrusion 320 are made of, for example, single crystal silicon. The protrusion 320 includes a second terrace 322 having a truncated pyramid shape formed on the upper surface of the base 310, a first terrace 321 having a truncated pyramid shape formed on the upper surface of the second terrace 322, and the first terrace. And a pivot 330 having a columnar shape formed on the upper surface of 321. Four electrodes 340 a to 340 d are formed on the four corners of the protruding portion 320 and the upper surface of the base portion 310 following the four corners. In addition, a pair of convex portions 360 a and 360 b are provided on the upper surface of the base portion 310 so as to sandwich the protruding portion 320. Further, a wiring 370 is formed on the upper surface of the base 310, and electrodes 340a to 340d are connected to the wiring 370 via lead lines 341a to 341d. An insulating layer 311 made of silicon oxide or the like is formed on the surface of the base 310, and electrodes 340a to 340d, lead lines 341a to 341d, and wirings 370 are formed on the insulating layer 311.

  The mirror substrate 200 and the electrode substrate 300 as described above are formed by joining the lower surface of the frame portion 210 and the upper surfaces of the convex portions 360a and 360b so that the mirror 230 and the electrodes 340a to 340d are opposed to each other. A micromirror device as shown in FIG. 16 is configured. In such a micromirror device, the mirror 230 is grounded, a positive drive voltage is applied to the electrodes 340 a to 340 d, and an asymmetric potential difference is generated between each of the electrodes 340 a to 340 d and the mirror 230. Can be attracted by electrostatic attraction, and the mirror 230 can be rotated in an arbitrary direction. Here, assuming that the tilt angle of the mirror 230 is θx, θy, and the applied voltages to the electrodes are V1, V2, V3, V4, the applied voltages are expressed by the following equations (1) to (4), for example.

V1 = Vo + Vx (1)
V2 = Vo + Vy (2)
V3 = Vo−Vx (3)
V4 = Vo−Vy (4)

  Here, Vo is called a bias voltage, and has an effect of improving the linearity from the voltage applied to the electrode to the mirror tilt angle. Vx and Vy are operation amounts corresponding respectively to the tilt angles θx and θy of the mirror, and the mirror 230 can be tilted in any direction by operating Vx and Vy.

T. Yamamoto, et al., `` A three-dimensional MEMS optical switching module having 100 input and 100 output ports '', Photonics Technology Letters, IEEE, Volume 15, Issue: 10

  In the optical switch as described above, the position error between the input / output port and the mirror and the mirror tilt angle change due to environmental changes such as ambient temperature and humidity, and external vibration, etc. The deviation from the corner gradually increases, and a drift may occur in which the power loss of the output light varies with time. An optical switch used in a general optical network system has a large effect on the entire optical network system when the optical switch itself causes fluctuations in loss. It is necessary to keep it within the allowable value.

  However, if the amount of drift per unit time is large, the optical connection strength may exceed the allowable loss if no measures are taken to suppress this drift. For this reason, in an optical switch, stabilization control is performed to obtain a stable optical connection strength by monitoring the light intensity of output light. Specifically, this stabilization control is performed by the following procedure. First, a control device (not shown) that controls the tilt angle of the mirror 230 supplies a drive voltage that periodically changes to the micromirror devices 3a and 3b, and perturbs (vibrates) the mirror 230 while outputting the output port. The light intensity of the output light is measured by an output light measuring device (not shown) provided on the output end side of 1b. Next, the drive voltage perturbation pattern and the light intensity value of the output light are stored on the storage device of the control device, and the drive voltage that obtains the maximum optical connection strength is optimized while comparing the maximum values in the perturbation pattern. Obtained as drive voltage. For example, the optimum drive voltage is obtained by a so-called hill-climbing method in which perturbation is performed with a voltage width ± ΔV set from the initial output voltage and the maximum value is searched while comparing the output intensity at that time. Finally, applying the obtained optimum drive voltage to the mirror sequentially is repeated at regular time intervals. The optical connection strength has been stabilized by such a method.

  In such so-called optimum value search, the micromirror device 3a is perturbed, and after detecting and correcting the error from the power fluctuation at this time, the micromirror device 3b is perturbed, and the error from the power fluctuation at this time is detected. In general, detection and correction are performed. As described above, since error detection and correction are sequentially performed in each of the micromirror devices 3a and 3b, the conventional optimum value search requires a long time for detection and correction of the optimum value. In addition, since the error of each mirror 230 affects the error detection of the other mirror 230, there is a problem that the detection accuracy of the optimum value is deteriorated.

  Accordingly, an object of the present invention is to shorten the time required to detect and correct the optimum value. Another purpose is to improve the optimum value detection system.

In order to solve the above-described problems, an optical switch according to the present invention includes at least one input port for inputting input light, at least one output port for outputting output light, an X axis, and the X axis. A first operation amount corresponding to the tilt angle of the first mirror, comprising: a first mirror supported so as to be rotatable with respect to the orthogonal Y axis; and a first electrode disposed opposite to the mirror. A first mirror device that deflects input light by applying a driving voltage according to the above to the first electrode and tilting the mirror, and is rotatable about the X axis and the Y axis orthogonal to the X axis a second mirror that is supported, and a second electrode which is the mirror disposed opposite to the driving voltage corresponding to the second operation amount corresponding to the tilt angle of the second mirror to the second electrode Applying the first mirror to tilt the mirror Generating a second mirror device and the second initial value is the initial value of the first initial value and the second manipulated variable is the initial value of the first manipulated variable to be outputted to the output port to deflect the light reflected by the An initial value generation unit and an operation amount that periodically changes around the first initial value are set as the first operation amount, and the first mirror is perturbed at the first perturbation frequency by the set operation amount. And a perturbation unit that sets an operation amount that periodically changes around the second initial value as a second operation amount, and perturbs the second mirror at the second perturbation frequency by the set operation amount ; when perturb the first and second mirror, a detector for detecting the light intensity of the output light the input light is output from the output port from said input port, the light intensity of the output light is optimum respectively corresponding to each of the tilt angles of the first and second mirror An error calculator for calculating the first and second error is an error of the first and second operation amount against each optimum operation amount, the first initial value based on the first error, a second A correction unit that corrects the second initial value based on each error using a predetermined time response waveform and updates the second initial value, and the perturbation unit includes a first perturbation frequency, a second perturbation frequency, and And the first mirror is perturbed so that the reflected light beam draws a substantially conical locus whose apex is located on the first mirror, and the second mirror is perturbed by the reflected light beam. The first and second manipulated variables are set so as to perturb so as to draw a substantially conical locus whose apex is located on the second mirror, and the perturbation of the first mirror and the second mirror are The error calculator is synchronized with the perturbations of the first and second mirrors. The light intensity of the output light is detected, the light intensity of the detected output light is subjected to frequency analysis, and the amplitude and phase of the first light intensity component relating to the first perturbation frequency mixed with the light intensity of the output light. The amplitude and phase of the second light intensity component related to the second perturbation frequency are obtained, the first error is calculated based on the obtained amplitude and phase of the first light intensity component, and the obtained first The second error is calculated based on the amplitude and phase of the second light intensity component .

Moreover, perturbation portion includes a first perturbation frequency, the ratio of the second perturbation frequency, m: k (m and k is an integer), and said first perturbation frequency of said second perturbation frequency n The first and second manipulated variables may be set so that (n is an integer) is not doubled or is not 1 / n .
The error calculation unit includes a first perturbation period specified by a first perturbation frequency, the light intensity of the output light in a time zone which is a least common multiple of the second perturbation period specified by the second perturbation frequency Based on the above, the first and second errors may be calculated. Further, the error calculation unit is configured to calculate the first and second values based on the average value of the product-sum calculation between the light intensity of the output light in the time zone and each of the first and second manipulated variables generated by the perturbation unit . The second error may be calculated.
In here, the error calculation unit calculates the first first error phase delay component of the perturbation frequency is subtracted from the first light intensity component phase specified by the dynamic characteristics of the first mirror, The second error may be calculated by subtracting the phase delay component at the second perturbation frequency specified by the dynamic characteristics of the second mirror from the phase of the second light intensity component .

Further, the correction unit includes a first calculated by the error calculation section, first by using the correction value obtained by multiplying a predetermined constant on each of the second error, even if the second initial value to update each Good .
Also, further comprising a waveform storage unit for storing a coefficient sequence corresponding to the time response waveform correcting unit, a correction value sequence obtained by multiplying the further coefficient sequence to the correction value may be calculated. Here, the time response waveform may be a step response waveform in which the resonance frequency component in the tilt direction of the first and second mirrors is attenuated.

The optical switch control method according to the present invention includes at least one input port for inputting input light, at least one output port for outputting output light, an X axis, and a Y axis orthogonal to the X axis. a first mirror that is rotatably supported by Te, and a first electrode which is the mirror disposed opposite to the driving voltage corresponding to the first operation amount corresponding to the tilt angle of the first mirror A first mirror device that deflects input light by applying to the first electrode and tilting the mirror; and a second mirror device that is rotatably supported with respect to the X axis and the Y axis perpendicular to the X axis . A mirror, and a second electrode disposed opposite to the mirror, and tilting the mirror by applying a driving voltage corresponding to a second operation amount corresponding to the tilt angle of the second mirror to the second electrode. By deflecting the light reflected by the first mirror device A second mirror device for outputting the allowed output port, a control method of an optical switch and a detector for detecting the light intensity of the output light the input light is output from the output port from the input port, the A first step of generating a first initial value that is an initial value of the first manipulated variable and a second initial value that is an initial value of the second manipulated variable, and a periodic change centered on the first initial value , An operation amount for perturbing the first mirror so that the reflected light beam draws a substantially conical locus whose apex is located on the first mirror is set as the first operation amount, and the set operation amount The first mirror is perturbed at the first perturbation frequency and is periodically changed around the second initial value, and the second mirror is substantially changed so that the reflected light beam has a vertex on the second mirror. The operation amount to be perturbed to draw a conical locus is set as the second operation amount. , Perturbed by the different second perturbation frequency and a second mirror first perturbation frequency by the operation amount that is the set, and the perturbation of the first mirror, starts in synchronization with the perturbation of the second mirror And an error of the first and second manipulated variables with respect to the optimum manipulated variables corresponding to the respective tilt angles of the first and second mirrors at which the light intensity of the output light is optimum. The light intensity of the output light is detected in synchronism with the perturbation of the first and second mirrors, respectively, and the detected light intensity of the output light is frequency-analyzed to obtain the light intensity of the output light. The amplitude and phase of the first light intensity component related to the first perturbation frequency and the amplitude and phase of the second light intensity component related to the second perturbation frequency mixed together are obtained, and the obtained first light Calculate the first error based on the amplitude and phase of the intensity component And a third step of calculating a second error based on the amplitude of the determined second light intensity component phase and a first initial value based on the first error, based on the second error the second initial value each, respectively and a fourth step of updating each corrected using the predetermined time response waveform, and repeating the second to fourth steps.

The control method of the optical switch, the third and fourth step, may be continuously give without stopping the periodic perturbation to the second Mirror and our first Mirror.

Here, the sum of the time of the second step and the time of the third and fourth steps is the first perturbation period specified by the first perturbation frequency and the first perturbation frequency specified by the second perturbation frequency . You may make it become the integral multiple of the time slot | zone used as the least common multiple with 2 perturbation periods.

  According to the present invention, the time required for the detection and correction of the optimum value by synchronizing the mirror perturbation of the first mirror device, the mirror perturbation of the second mirror device, and the detection of the light intensity. Can be shortened.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the present embodiment, the same components as those described in the background art section with reference to FIGS. 14 to 16 are denoted by the same names and reference numerals, and the description thereof is omitted as appropriate.

<Configuration of optical switch>
As shown in FIG. 1, the optical switch according to the present embodiment includes an input port 1a, an output port 1b, an input side micromirror device 3a, an output side micromirror device 3b, an output light measuring device 4, and And a control device 5. In such an optical switch, in order to adjust the tilt angle of the mirror 230 of the micromirror devices 3a and 3b so that the light intensity of the output light is optimized, the control device 5 perturbs the mirror 230, and the output light at this time The light intensity is monitored by the output light measurement device 4, and the tilt angle error is calculated and corrected based on the measurement result of the output light measurement device 4 by the control device 5. The tilt angle of each mirror has a one-to-one correspondence with the operation amount output from the control device 5, and the mirror 230 is driven by converting the operation amount into a voltage applied to the electrode of each mirror and applying it.

  As shown in FIG. 2, the control device 5 includes a drive unit 6, a detection unit 7, and a control unit 11.

  The control unit 11 is a functional unit that controls the operation of the entire optical switch, and includes at least an error calculation unit 111, a correction unit 112, an initial value generation unit 113, a perturbation generation unit 114, and a waveform storage unit 115. I have.

  The error calculator 111 calculates the error of each manipulated variable from the light intensity of the output light monitored by the output light measuring device 4 in synchronization with the perturbation of the mirror 230.

  The correction unit 112 corrects and updates the initial operation amount based on the operation amount error calculated by the error calculation unit 111.

  The initial value generation unit 113 sets an operation amount corresponding to the initial tilt angle of the mirror 230.

  The perturbation generation unit 114 sets an operation amount for giving a periodic perturbation around the operation amount generated by the initial value generation unit 113.

  The waveform storage unit 115 stores a waveform that perturbs the mirror 230 set by the perturbation generation unit 114.

  Such a control device 5 includes an arithmetic device such as a CPU, a storage device such as a memory and an HDD (Hard Disk Drive), an input device that detects information input from the outside such as a keyboard, a mouse, a pointing device, a button, and a touch panel, I / F devices and CRT (Cathode Ray Tube), LCD (Liquid Crystal Display), FED (Field) which send and receive various information via communication lines such as the Internet, LAN (Local Area Network), WAN (Wide Area Network) It is comprised from the computer provided with display apparatuses, such as Emission Display) and organic EL (Electro Luminescence), and the program installed in this computer. That is, the hardware resource and the software cooperate to control the hardware resource by a program, thereby realizing the drive unit 6, the detection unit 7, and the control unit 11 described above. The program may be provided in a state where it is recorded on a recording medium such as a flexible disk, a CD-ROM, a DVD-ROM, or a memory card.

<Mirror tilt angle adjustment operation>
Next, with reference to FIG. 3, the adjustment operation of the tilt angle of the mirror 230 in the optical switch according to the present embodiment will be described.

  First, the initial value generation unit 113 sets an operation amount corresponding to the initial tilt angle of the mirror 230 (step S1). When the operation amount is set by the initial value generation unit 113, the perturbation generation unit 114 sets an operation amount for giving periodic perturbation around the operation amount. Then, a drive voltage based on the operation amount set by the perturbation generation unit 114 by the drive unit 6 is applied to the micromirror devices 3a and 3b, and the respective mirrors 230 are simultaneously perturbed. For the sake of convenience, a case where an optical path is connected between the mirror 230a of the micromirror device 3a and the mirror 230b of the micromirror device 3b will be described below as an example.

  When the mirrors 230a and 230b are perturbed, the error calculator 111 detects the light intensity of the output light monitored by the output light measurement device 4 in synchronization with the perturbation of the mirrors 230a and 230b (step S2).

  When the light intensity of the output light is detected, the error calculator 111 calculates an operation amount error based on the light intensity value (step S3). Here, the perturbations of the mirrors 230a and 230b are synchronized, and are simultaneously perturbing. The output light measuring device 4 also monitors the light intensity of the output light in synchronization with the perturbation of the mirrors 230a and 230b. For example, the light intensity P (t) of the output light obtained when the perturbation of Sinωt is given to the mirrors 230a and 230b has no time difference, and the power fluctuation at the same time as the perturbation is monitored. Details of the operation of the error calculation unit 111 will be described later.

  When the operation amount error is calculated, the correction unit 112 corrects and updates the initial operation amount based on the error (step S4). Details of the operation of the correction unit 112 will be described later.

  When the initial operation amount is updated, when the process is still performed (step S5: NO), the control device 5 returns to the process of step S2. On the other hand, when the process ends (step S5: YES), the control device 5 ends the process.

  As described above, according to the present embodiment, the perturbation of the mirrors 230a and 230b and the monitoring of the light intensity of the output light measuring device 4 are performed in synchronization with each other, thereby reducing the time required for detection and correction of the optimum value. Shortening can be realized.

<Operation of error calculator>
Next, details of the processing operation of the error calculator 111 will be described.

  Since the influence of the perturbation of the mirror 230a and the mirror 230b is mixed with the light intensity of the output light measured by the output light measuring device 4, it is necessary to obtain the operation amount error of each of the mirrors 230a and 230b from the detected light intensity. is there. For this reason, both the perturbations of the mirrors 230a and 230b are perturbed so that the locus of the reflected light beam of the mirror has a conical shape having a vertex on the mirror surface. The reflected light beam trajectory during perturbation of each mirror is shown in FIGS. 4 (a) and 4 (b).

  Focusing on one mirror, as described above, the mirror can be tilted about two axes orthogonal to each other on the mirror surface, and the tilt angles θx and θy about the respective axes are respectively determined by two operation amounts Vx and Vy. Be controlled. When the operation amounts Vx and Vy are represented by the following expressions (5) and (6), the light beam of the reflected light from the mirror 230 draws a conical locus. Here, Vx0 and Vy0 are initial values of the manipulated variables, and perturbation is given around these initial values. Vr is a parameter that determines the radius of the conical perturbation.

Vx = Vx0 + Vr · Cos (2πft) (5)
Vy = Vy0 + Vr · Sin (2πft) (6)

  In order to obtain the operation amount errors of the mirrors 230a and 230b, the perturbation frequencies of the two mirrors 230a and 230b are made different as shown in the following equations (7) to (10). Here, Vx1 and Vy1 are the operation amounts of the mirror 230a, Vx2 and Vy2 are the operation amounts of the mirror 230b, Vx10 and Vy10 are the initial operation amounts of the mirror 230a, Vx20 and Vy20 are the initial operation amounts of the mirror 230b, and Vr1 and Vr2 are respectively F1 is a perturbation frequency of the mirror 230a, and f2 is a perturbation frequency of the mirror 230b.

Vx1 = Vx10 + Vr1 ・ Cos (2πf1t) (7)
Vy1 = Vy10 + Vr1 · Sin (2πf1t) (8)
Vx2 = Vx20 + Vr2 ・ Cos (2πf2t) (9)
Vy2 = Vy20 + Vr2 · Sin (2πf2t) (10)

  If the two perturbation frequencies f1 and f2 are different, it is possible to separate and calculate the operation amount error of the two mirrors, for example, by analyzing the frequency of the optical power response.

  Since the mirror 230 can be modeled by a mass system supported by a torsion spring for tilting movements about each axis, the mirror 230 has a dynamic characteristic called a so-called spring-mass system and having a resonance frequency as shown in FIG. The perturbation frequency described above is not limited by the resonance frequency of the mirror 230, and can be set to be higher than the resonance frequency. As the perturbation frequency is set higher, the operation amount error can be detected in a shorter time.

  The cone radius, which is the trajectory of the reflected beam when perturbed, is determined to be a value more suitable for the optical characteristics including the input and output fiber collimators. That is, if the radius is too large, the optical power fluctuation becomes too large and it is necessary to consider nonlinearity, whereas if the radius is too small, the fluctuation becomes too small and the S / N in the optical power response The ratio will deteriorate. For this reason, it is desirable that the cone radius at the time of perturbation is constant regardless of the dynamic characteristics of the mirror 230 and the perturbation frequency. Since the cone radius at the time of perturbation has a one-to-one correspondence with the apex angle of the cone, making the cone radius constant is equivalent to making the tilt angle of the perturbation mirror 230 constant. In order to make the tilt angle of the perturbation mirror 230 constant, it is necessary to set a voltage in consideration of the dynamic characteristics of the mirror 230. FIG. 6 shows the gain characteristics of the tilt angle with respect to the operation amount of the mirror around the X axis. When the resonance frequency is exceeded, the gain characteristic is less than 1. When the perturbation frequency is set to exceed the resonance frequency, it is necessary to increase the perturbation voltage in consideration of the attenuation of the gain as compared with the case of perturbation below the mirror resonance frequency. Moreover, since the gain attenuation increases as the perturbation frequency increases, the perturbation voltage must be set higher.

  The dynamic characteristics of the mirror may be different around the X axis and the Y axis of the mirror 230, or may be different between the mirror 230a and the mirror 230b. In this case, since the dynamic characteristics around each axis are different, it is necessary to set a voltage in consideration of the gain characteristics for each axis. For example, FIG. 7 shows the dynamic characteristics of the mirror 230a around the X axis and the Y axis. The resonance frequency around the X axis is lower than the resonance frequency around the Y axis. Therefore, at a frequency higher than the resonance frequency, the attenuation of the gain is greater around the X axis than around the Y axis. The perturbation around the X-axis and Y-axis of the mirror 230a is perturbed at the same frequency. However, when the dynamic characteristics are different as described above, the perturbation voltage around the X-axis needs to be set larger than that around the Y-axis. That is, when the mirror is perturbed at a frequency higher than the resonance frequency, the perturbation voltage needs to be increased as the frequency becomes higher than the resonance frequency, and the perturbation voltage needs to be decreased as the frequency approaches the resonance frequency.

  Here, the perturbation voltage will be described. As described above, the manipulated variable at the time of perturbation is expressed by the above equations (7) to (10). Here, Vr1 and Vr2 are manipulated variables related to the perturbation radius. Focusing on the mirror 230a, the voltage applied to the four electrodes of the mirror 230a is converted into a voltage by the following equations (11) to (14) when the voltage calculation formula of the previous four electrodes is adopted.

V1 = Vo + Vx1 = Vo + Vx10 + Vr1.Cos (2πf1t) (11)
V2 = Vo + Vy1 = Vo + Vy10 + Vr1 · Sin (2πf1t) (12)
V3 = Vo-Vx1 = Vo-Vx10-Vr1.Cos (2πf1t) (13)
V4 = Vo−Vy1 = Vo−Vy10−Vr1 · Sin (2πf1t) (14)

  Since the radial voltage when considering the dynamic characteristics and perturbation frequency of the mirror 230 is different around the X axis and the Y axis, if the parameters Vr1x, Vr1y, Vr2x, Vr2y regarding the dynamic characteristics are introduced, the above formula (11) to (14) are represented by the following expressions (15) to (18). Here, Vr1x, Vr1y, Vr2x, and Vr2y are obtained by multiplying Vr by a reciprocal of gain attenuation determined by the mirror dynamic characteristics and the perturbation frequency. For example, when the gain at the perturbation frequency is 1/10 due to the dynamic characteristics of the mirror, 10 times Vr is set.

Vx1 = Vx10 + Vr1x · Cos (2πf1t) (15)
Vy1 = Vy10 + Vr1y · Sin (2πf1t) (16)
Vx2 = Vx20 + Vr2x · Cos (2πf2t) (17)
Vy2 = Vy20 + Vr2y · Sin (2πf2t) (18)

  Next, a method for calculating each operation amount error will be described. Here, a case where only the mirror 230a is perturbed so that the reflected light beam locus has a conical shape will be described.

  The relationship between the tilt angles θx and θy of the mirror 230a and the light intensity of the output light has a shape close to a Gaussian distribution in which the optimum mirror tilt angle at which the light intensity of the output light is maximized is a peak. As described above, the mirror 230 applies a voltage so that the tilt angle of the perturbation becomes constant in consideration of the mirror dynamic characteristics and the perturbation frequency. That is, a perturbation that draws a circle is given on the θx and θy planes. When perturbation is given in a state where there is an error in the tilt angle, the light intensity of the output light varies at the same frequency as the perturbation frequency, and the light intensity fluctuation peaks when perturbed in the optimum value direction. Therefore, the peak direction can be known by obtaining the phase difference between the perturbation component and the light output response.

  Further, since the relationship between the light intensity of the output light and the tilt angle can be approximated to a Gaussian distribution, the ratio of the light intensity fluctuation amount at the time of perturbation to the tilt angle amplitude of the perturbation corresponds to a value obtained by differentiating the Gaussian distribution by the tilt angle. When the light intensity of the output light is expressed in dBm, the relationship between the tilt angle and the light intensity is a paraboloid as shown in FIG. As shown in FIG. 9, the ratio of the fluctuation amount of the light intensity at the time of the perturbation to the amplitude of the perturbation at that time is the differential value of the paraboloid in the plane including the initial value of the tilt angle that is the center of the perturbation and the optimum value. The tilt angle error amount can be estimated by multiplying the ratio by a constant. This is because the differential value with respect to the tilt angle of the parabola is a straight line, the differential value is 0 at the optimum position, and the error amount of the tilt angle is proportional to the differential value. Since the perturbation radius is perturbed to be constant, the error amount can be estimated by multiplying the fluctuation range of the light intensity by a constant. From the above, the manipulated variable error amount can be estimated from the phase and amplitude of the light intensity at the perturbation frequency.

  Next, a method for extracting one frequency component from an optical power response in which two mirrors are simultaneously perturbed and two frequencies are mixed will be described. Here, it is assumed that the light intensity p of the output light is expressed by the following equation (19) in which two frequency components are mixed.

p = p1 · Sin (2πf1 t + φ1) + p2 · Sin (2πf2 t + φ2) (19)

  In the above equation (19), f1 is the perturbation frequency of the mirror 230a, and f2 is the perturbation frequency of the mirror 230b. Here, when the product-sum average of p and Cos (2πf1 t) is obtained, the following equation (20) is obtained.

  In the above equation (20), the first term is a term including the phase φ1 and the amplitude p1 necessary for error calculation, and phase information and amplitude information are obtained using this term. The second term, the third term, and the fourth term are unnecessary terms. Therefore, by appropriately selecting the integration time so that these terms become zero, the accuracy of the first term is improved and the error detection accuracy can be improved. If the period of frequency f1 is T1 = 1 / f1, and the period of frequency f2 is T2 = 1 / f2, the second term is 1 / 2f1 = T1 / 2, the third term is 1 / (f2 + f1) = T1 T2 / (T1 + T2), the fourth term is a periodic signal with a period of 1 / (f2−f1) = T1 · T2 / (T1−T2). Therefore, if the least common multiple of T1 and T2 is the integration time, the second, third, and fourth terms can be zero. Similarly, p1 · Cos (φ1) / 2 can be obtained by obtaining the average value of the product-sum average of p and Cos (2πf1 t) at the time interval of the least common multiple of T1 and T2, and the method described above The amplitude p1 and the phase φ1 can be obtained from p1 · Sin (φ1) / 2 obtained in step (1). Further, for the mirror 230b, the amplitude p2 and the phase φ2 can be obtained by obtaining the product-sum average of Sin (2πf2 t) and Cos (2πf2 t) at the time interval of the least common multiple of T1 and T2. This calculation makes it possible to perform error calculation with high accuracy and with a shorter calculation time than FFT. In addition, data acquisition time can be reduced and high accuracy can be achieved by acquiring and calculating data at the time interval of the least common multiple of T1 and T2.

  It should be noted that phase information and amplitude information at each perturbation frequency can also be obtained by frequency analysis of the optical power response using a general FFT calculation tool or the like. Even in this case, the data interval is set to the least common multiple of T1 and T2, so that the accuracy can be improved and the data collection time can be shortened.

  Since the phase obtained by the above equation (20) corresponds to the phase delay from the perturbation drive signal to the optical power response, the delay caused by the dynamic characteristics of the mirror at the perturbation frequency and the initial value of the mirror operation amount deviate from the optimum values. Therefore, it is combined with the phase indicating the optimum value direction. Therefore, in order to accurately obtain the peak direction, it is necessary to subtract the phase delay caused by the dynamic characteristics of the mirror from the phase information obtained by the above-described method. By performing such processing, it is possible to detect errors with higher accuracy.

  The perturbation frequency f1 of the mirror and the perturbation frequency f2 of the mirror 230b can be arbitrarily selected if they are not equal. At this time, if f1 is a frequency that is an integral multiple of f2, or f2 is a frequency that is an integral multiple of f1, an error may occur when the respective operation amount errors are obtained from the light intensity of the output light. This is because the light intensity of the output light may be mixed with the nth power component of the perturbation as well as the primary combination of the same frequency components as the perturbation assumed as described above. Therefore, it is desirable to avoid the above combination of perturbation frequencies.

<Operation of correction unit>
Next, correction value calculation and update operations by the correction unit 112 will be described.

  In the calculation of the operation amount correction value, the fluctuation range of the light intensity of the output light is multiplied by a constant. This constant has a different optimum value depending on the tilt angle when the voltage-tilt angle characteristic of the mirror is nonlinear. Even if the perturbation operation amount width is the same, the smaller the initial value of the mirror tilt angle, the smaller the tilt angle of the perturbation. Therefore, the smaller the initial value is, the larger the constant can be, thereby reducing the influence of nonlinearity. Similarly, when the voltage-tilt angle characteristics differ depending on the mirror, this constant value can be changed by the mirror, that is, by increasing the constant as the tilt angle becomes smaller with the same voltage applied. The influence of the characteristic difference can be reduced.

  When correcting and updating the initial operation amount with the calculated operation amount correction value, if the operation amount is changed stepwise, vibration near the resonance frequency of the mirror occurs. If the optical response data for the next operation amount error calculation is collected in such a vibration state, the accuracy of the error calculation deteriorates. Therefore, in order to correct the operation amount so that the vibration of the mirror does not occur, the initial operation amount is corrected according to a waveform obtained by removing a component near the resonance frequency of the mirror, not in a step shape. By using such a waveform, the mirror is not excited at the resonance frequency, and vibration at the time of correcting the initial operation amount can be suppressed. Further, the mirror can be operated at high speed by increasing the high frequency component in the waveform.

  In order to perform an operation amount correction using such a waveform, the control device 5 stores a coefficient sequence corresponding to the waveform. For example, the time response of the waveform is sampled at regular intervals, and is a value normalized by the difference between the value before the start of the waveform and the end value. Such a value is stored in the waveform storage unit 115. As shown in FIG. 10, the operation amount correction value is multiplied by the coefficient sequence stored in the waveform storage unit 115 and added to the initial operation amount, thereby correcting with an arbitrary response waveform. FIG. 11 shows an example of a drive voltage waveform for each sampling interval. Since this method is easy to calculate and there is no feedback loop like the calculation in the IIR filter, the calculation result does not diverge and has high stability.

  In the flowchart shown in FIG. 3, when the mirror 230 starts or stops perturbation, the mirror 230 moves with vibration near the resonance frequency, which may cause deterioration in accuracy of the operation amount error calculation. Therefore, the perturbation of the mirror 230 is continuously repeated without stopping also in the operation amount error calculation and the initial operation amount correction / update step. As a result, vibration near the mirror resonance frequency is not excited and high accuracy can be achieved. As described above, the detection of the light intensity of the output light is preferably performed at a time interval of the least common multiple of the perturbation period of the mirror 230a and the perturbation period of the mirror 230b. If a phase shift occurs between the perturbation when the first optical power response is detected and the perturbation when the second initial optical power response is detected after the first correction of the initial manipulated variable, the phase shift causes an error in the phase information during the calculation. It becomes. Therefore, as shown in FIG. 12, it is desirable that the total time required for one operation amount error calculation and initial operation amount correction and update so that the phases are aligned is an integral multiple of the detection time of the optical power response.

  Further, when the perturbation is started in the second step after setting the initial operation amount, the mirror vibrates near the resonance frequency. For this reason, it is desirable to start the perturbation from the time when the initial manipulated variable is set in the first step. Further, when setting the initial manipulated variable in the first step, the mirror is excited by vibrations near the resonance frequency. The accuracy can be improved by starting the data acquisition after the vibration is attenuated.

  The radius of the perturbation is set so that the fluctuation of the light intensity of the output light when a point on the perturbation locus is at the optimum value becomes a predetermined value. For example, only the mirror 230a is perturbed, and the perturbation radius is determined so that the fluctuation range of the output optical power response at that time is 0.5 dB. Similarly, the mirror 230b is set so that the power response fluctuation width is 0.5 dB when only the mirror 230b is perturbed.

  Further, when there is an error in the operation amount and the drive voltage deviates from the optimum value, the fluctuation range of the optical response is the predetermined value as long as the perturbation radius is constant, as shown in the light intensity distribution diagram described with reference to FIG. That is, in the above example, it becomes larger than 0.5 dB. Therefore, if the fluctuation range of the light intensity of the output light is smaller than a predetermined value, it can be determined that the optimum value has been reached, and the process proceeds to the end process.

  Even when the perturbation is suddenly stopped as the end process, the vibration of the mirror 230 is generated. To avoid this, the perturbation amplitude is decreased with time, and finally the perturbation amplitude is set to zero. By this process, fluctuations in optical power at the end can be suppressed. As a method of decreasing, as shown in FIG. 13A, it may be decreased step by step for every correction cycle from the second step to the fourth step, or it decreases with time as shown in FIG. 13B. You may make it make it. In this embodiment, the mirror 230a and the mirror 230b are simultaneously perturbed to detect the respective errors at the same time. Therefore, the influence of the errors of the mirrors on the error detection accuracy of the other mirrors can be reduced, and high accuracy can be achieved. become. For example, when the error detection and correction of the mirror 230a and the mirror 230b are sequentially performed, there is an error in the mirror 230b when the error of the mirror 230a is detected. When the optimum value of the mirror 230a is detected in this state, the error of the mirror 230b is detected. The output optical power is maximized at a value deviated from the true optimum value due to the influence. For this reason, the error remains after the correction of the mirror 230a, and the same error detection accuracy deteriorates when the error of the mirror 230b is detected. On the other hand, in this embodiment, since the errors of the mirrors 230a and 2 are detected simultaneously, the accuracy can be improved.

  The present invention can be applied to an optical switch or the like.

It is a block diagram which shows the structure of the optical switch which concerns on this invention. It is a block diagram which shows the structure of a control apparatus. It is a flowchart which shows adjustment operation | movement of the tilt angle of a mirror. (A) is a figure which shows the reflected light beam locus | trajectory at the time of perturbation of the mirror 230a, (b) is a figure which shows the reflected light beam locus | trajectory at the time of the perturbation of the mirror 230b. It is a figure which shows the gain characteristic of the tilt angle with respect to the controlled variable of a mirror. It is a figure which shows the gain characteristic of the tilt angle with respect to the control amount of the X-axis periphery. It is the figure which piled up the dynamic characteristic of the X-axis periphery and the Y-axis rotation. It is a figure which shows the relationship between a tilt angle and light intensity. It is a figure explaining the ratio of the variation | change_quantity of the light intensity with respect to the amplitude of a perturbation. It is a figure for demonstrating correction | amendment operation | movement of control amount. It is a figure which shows an example of the drive voltage waveform for every sampling interval. It is a figure explaining the relationship between the time required for control amount error calculation and correction / update of the initial control amount and the detection time of the light intensity of the output light. (A) is a figure which shows an example of the relationship between operation amount and time, (b) is a figure which shows an example of the relationship between operation amount and time. It is a perspective view which shows the structure of an optical switch typically. It is a perspective view which shows the structure of a mirror apparatus typically. It is sectional drawing which shows the structure of a mirror apparatus typically.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1a ... Input port, 1b ... Output port, 3a, 3b ... Micromirror device, 4 ... Output light measuring device, 5 ... Control device, 6 ... Drive part, 7 ... Detection part, 8 ... Control part, 9 ... Storage part, DESCRIPTION OF SYMBOLS 11 ... Control part, 111 ... Error calculating part, 112 ... Correction | amendment part, 113 ... Initial value production | generation part, 114 ... Perturbation production | generation part, 115 ... Waveform memory | storage part.

Claims (11)

At least one input port for inputting input light;
At least one output port for outputting output light;
It comprises a first mirror which is rotatably supported with respect to the Y-axis orthogonal to the X-axis and the X-axis, and a first electrode which is the mirror disposed opposite to the tilting angle of the first mirror a first mirror device for deflecting said input light by a driving voltage is applied to the first electrode to tilt the first mirror in accordance with the first operation amount corresponding to,
It includes a second mirror that is rotatably supported with respect to the Y-axis orthogonal to the X-axis and the X-axis, and a second electrode which is the mirror disposed opposite to the tilting angle of the second mirror output to the output port to deflect the light reflected by said first mirror device by tilting the second mirror is applied to the second electrode of the driving voltage corresponding to the second operation amount corresponding to the A second mirror device,
An initial value generator that generates a first initial value that is an initial value of the first manipulated variable and a second initial value that is an initial value of the second manipulated variable ;
An operation amount that periodically changes around the first initial value is set as the first operation amount, and the first mirror is perturbed at a first perturbation frequency by the set operation amount, and A perturbation unit that sets an operation amount that periodically changes around a second initial value as the second operation amount, and perturbs the second mirror at a second perturbation frequency by the set operation amount ;
When perturb the first and second mirror, a detector for detecting the light intensity of the output light the input light is output from the output port from said input port,
The light intensity of the output light is a error in the optimal become the first and the second respectively of the respective optimal operation amounts respectively against the first and the second operation amount corresponding to the tilt angle of the mirror An error calculation unit for calculating the first and second errors;
The first initial value based on the first error, the second initial value based on the second error is corrected respectively with the predetermined time response waveform, and a correction unit that updates each With
The perturbation part is
The first perturbation frequency and the second perturbation frequency are made different so that the first mirror draws a substantially conical trajectory in which the reflected light beam has an apex located on the first mirror. The first and second manipulated variables so that the second mirror perturbs the second mirror so that the reflected light beam draws a substantially conical locus whose apex is located on the second mirror. Set, and further, start the perturbation of the first mirror and the perturbation of the second mirror synchronously ,
The error calculator is
Detecting the light intensity of the output light in synchronization with the perturbation of the first and second mirrors;
The detected light intensity of the output light is subjected to frequency analysis, and the amplitude and phase of the first light intensity component related to the first perturbation frequency mixed with the light intensity of the output light and the second perturbation frequency. And calculating the first error based on the obtained amplitude and phase of the first light intensity component, and calculating the obtained second light intensity component. The optical switch, wherein the second error is calculated based on an amplitude and a phase of the optical switch. In the perturbation unit, a ratio of the first perturbation frequency to the second perturbation frequency is m: k (m and k are integers), and the first perturbation frequency is n (2) of the second perturbation frequency. n so is not a 1 or n fraction not an integer) times, the optical switch according to claim 1, wherein setting the first and second operation amount. The error calculating unit, the first and the first perturbation period specified by the perturbation frequency, the output light in a time zone which is a least common multiple of the second perturbation period specified by the second perturbation frequency The optical switch according to claim 2 , wherein the first and second errors are calculated based on the light intensity. The error calculation unit is configured to calculate the first based on an average value of a product-sum operation between the light intensity of the output light in the time zone and each of the first and second manipulated variables generated by the perturbation unit . The optical switch according to claim 3 , wherein the first and second errors are calculated. The error calculation unit calculates the first error by subtracting a phase lag component at the first perturbation frequency specified by a dynamic characteristic of the first mirror from a phase of the first light intensity component. The second error is calculated by subtracting the phase lag component at the second perturbation frequency specified by the dynamic characteristic of the second mirror from the phase of the second light intensity component. the optical switch of claim 4 Symbol mounting. Wherein the correction unit, wherein the first calculated by the error calculating unit, the first using a correction value obtained by multiplying a predetermined constant on each of the second error, characterized in that the second initial value is updated each The optical switch according to claim 1 . A waveform storage unit that stores a coefficient sequence corresponding to the time response waveform;
The optical switch according to claim 6 , wherein the correction unit calculates a correction value sequence obtained by multiplying the correction value by the coefficient sequence. The optical switch according to claim 7 , wherein the time response waveform is a step response waveform obtained by attenuating resonance frequency components in the tilt direction of the first and second mirrors. At least one input port for inputting input light, at least one output port for outputting output light, and a first mirror supported so as to be rotatable with respect to the X axis and a Y axis orthogonal to the X axis; A first electrode disposed opposite to the mirror, and applying a driving voltage corresponding to a first operation amount corresponding to a tilt angle of the first mirror to the first electrode to apply the first electrode . a first mirror device for deflecting said input optical mirror by tilting the of the second mirror which is rotatably supported with respect to the Y-axis orthogonal to the X-axis and the X axis, and the mirror and a second electrode facing each other, tilts the second second said driving voltage corresponding to the operation amount is applied to the second electrode a second mirror corresponding to the tilt angle of the mirror Thereby deflecting the reflected light from the first mirror device. A second and a mirror device, method of controlling an optical switch that includes a detector for detecting the light intensity of the output light output from said output port for outputting to the output port is,
A first step of generating a first initial value that is an initial value of the first manipulated variable and a second initial value that is an initial value of the second manipulated variable ;
An operation amount that periodically changes around the first initial value and causes the first mirror to perturb so that a reflected light beam draws a substantially conical locus whose apex is located on the first mirror. The first operation amount is set, and the first mirror is perturbed at the first perturbation frequency according to the set operation amount, and periodically changes around the second initial value, An operation amount for causing the mirror 2 to perturb the reflected light beam so as to draw a substantially conical locus whose apex is located on the second mirror is set as the second operation amount, and the set operation is performed. The second mirror is perturbed by a second perturbation frequency different from the first perturbation frequency according to the amount, and the first mirror perturbation and the second mirror perturbation are started synchronously . Two steps,
The errors of the first and second manipulated variables respectively corresponding to the respective optimum manipulated variables corresponding to the tilt angles of the first and second mirrors at which the light intensity of the output light is optimum are respectively set to first. And the second error, the light intensity of the output light is detected in synchronization with the perturbation of the first and second mirrors, the light intensity of the detected output light is frequency-analyzed, and the output light The amplitude and phase of the first light intensity component related to the first perturbation frequency mixed with the light intensity and the amplitude and phase of the second light intensity component related to the second perturbation frequency were obtained and obtained. The third error is calculated based on the amplitude and phase of the first light intensity component, and the second error is calculated based on the obtained amplitude and phase of the second light intensity component . And the steps
A fourth step of updating each of the first initial value based on the first error and the second initial value based on the second error by correcting each using a predetermined time response waveform; Have
The method for controlling an optical switch, wherein the second to fourth steps are repeated. Before Symbol third and fourth step, according to claim 9, wherein the continuously give it without stopping the periodic perturbations in the first Mirror Contact and the second Mirror Control method of optical switch Time of the second step, and the third, the sum of the time of the fourth step is identified and the first perturbation period specified by the first perturbation frequency, by said second perturbation frequency The method of controlling an optical switch according to claim 10, wherein the control time is an integral multiple of a time zone that is a least common multiple of the second perturbation period.

Images