JP7011534B2 - Control circuit - Google Patents

Description

The present invention relates to a control circuit.

The Mach-Zehnder interferometer (MZI) is an optical component used in various fields such as an optical bandpass filter that passes only a specific wavelength and a bit delay interferometer for demodulating a phase-modulated signal. In particular, an optical modulator using MZI (hereinafter referred to as "MZI type optical modulator") is a key component in a large-capacity optical transmission system.

As is well known, MZI branches the input light into two, propagates each arm separately, and then re-waves and outputs the light. The intensity or optical phase of the light output from MZI is determined by the phase difference of the light propagating through each of the two arms, and this phase difference is determined by the difference in the optical path lengths of these two arms.

When MZI is used as an optical filter or a bit delay interferometer, it is necessary to adjust the difference between the optical path lengths of the two arms to a constant value. The difference in optical path length should be determined by how the wavelength characteristics of the optical filter are set or the carrier wavelength and bit rate of the signal demodulated by the bit delay interferometer.

The MZI type optical modulator intentionally modulates this difference in optical path length with a drive signal to modulate the intensity or optical phase of the output light to generate an optical modulation signal. However, even when operating as an MZI type optical modulator, it is necessary to keep the difference between the two arm optical path lengths constant at the moment when the drive signal reaches the zero level. The difference in optical path length should be determined by the wavelength used as the carrier and the signal format of the optical signal.

The control of the difference between the optical path lengths of the two arms of MZI is achieved by changing the refractive index and length of the arms by the delay amount control unit provided on at least one of these arms. The delay amount control unit generally uses an optical non-linear effect such as the Pockels effect or thermal expansion of an optical waveguide by a heater. The delay amount brought to the arm by the delay amount control unit is controlled by an electric signal applied to the delay amount control unit. This electrical signal is commonly referred to as the bias voltage. Although it is possible to control by current instead of voltage, the term "bias voltage" is used in this application according to the convention.

It is generally difficult to predict in advance how much the bias voltage will be required to make the difference between the optical path lengths of the two arms of MZI a desired value. This is because the manufacturing error of the optical path length of each arm is usually an amount comparable to the wavelength, and the wavelength dependence of the above-mentioned optical nonlinear effect and the change in the optical path length due to the environmental temperature cannot be ignored. As a result, the optimum value of the bias voltage differs for each individual MZI, also depends on the wavelength used, and further changes slightly depending on the time course of MZI.

Therefore, in the operation of MZI, an automatic bias control circuit (ABC circuit) including an error detection circuit for detecting a deviation of the bias voltage from the optimum value and a bias control circuit for feeding back the detected error to the bias voltage is provided. It becomes important. Various methods have been proposed in the past for automatic bias control technology, particularly automatic bias control technology for MZI type optical modulators (see, for example, Patent Document 1).

Japanese Patent No. 5261779

In the past, it was common to use the Pockels effect for the delay amount control unit. In this case, the bias voltage and the delay amount of the light propagating through the arm are almost proportional to each other. Since the bias voltage and the optical phase difference are also proportional, the feedback control is relatively easy.

On the other hand, the refractive index of the arm is changed by using the delay amount control unit of the type that adjusts the delay amount by thermally expanding the arm with the heat of the heater arranged near the arm and the nonlinear optical effect of the semiconductor element. There is a type of delay amount control unit that adjusts the delay amount of light propagating through. In these types of delay amount control units, it is important to pay attention to the heat amount of the heater and the non-linearity of the semiconductor when performing feedback control. For example, the amount of heat of a heater is proportional to the product of the heater voltage applied to the heater and the heater current passing through the heater, but the heater current depends on the resistance value of the heater, and the resistance value of the heater depends on the temperature of the heater. .. Therefore, even if the heater voltage and the heater current are controlled by simple PI (Proportional-Integral) control or PID (Proportional-Integral-Differential) control, accurate control is difficult.

Another problem caused by the above-mentioned non-linearity is that the influence of pilot tone or dithering becomes non-linear. As an example, consider controlling the light intensity of the output light from MZI as shown in FIG. 9 to the maximum. Here, for the sake of simplicity, it is assumed that the MZI is not modulated and the output light is CW (Continuous wave) light. The CW light input to the MZI is branched into two by the first optical coupler. A delay amount control unit is provided on one of the arms in which each of the two branched lights propagates. The light propagating through each of the two arms is combined by the second optical coupler and taken out as output light.

Here, the bias voltage is changed to change the delay amount generated in the delay amount control unit so that the difference in optical phase between the two arms is an integral multiple of 2π (unit is rad (radian), the same applies hereinafter). The purpose can be achieved by stopping the change of the delay amount when the light intensity detected by the optical power monitor becomes maximum. However, as mentioned above, the length of the MZI arm is liable to change with time, and the optical power output from the CW light source may also fluctuate, so it is continuously determined whether or not the light intensity is maximum. Difficult to do.

In order to solve this problem, it is widely practiced to add a dither signal of frequency fd to the bias voltage, synchronously detect the output of the optical power monitor at the frequency fd, and feed back the synchronous detection result to the bias voltage and delay amount control unit. There is. Here, it is common to keep the amplitude of the dither signal small so as not to increase the influence on the output light. FIG. 9 also shows a configuration including a functional unit to which a dither signal is applied and a synchronous detection circuit.

A bias voltage, which is the output of the bias power supply, is applied to the delay amount control unit, and a dither signal having a frequency fd output from the oscillator is applied to the bias voltage. The frequency fd or its harmonic intensity modulation component is superimposed on the output light. The output light from the second optical coupler is partially branched by the tap. After the optical power monitor receives the branched light and detects the dither component of the frequency fd or its harmonic wave, the synchronous detection circuit performs synchronous detection using the oscillator output as a reference. As will be described later, the synchronous detection result can be used as an error signal indicating how far the current value of the bias voltage is from the optimum value. That is, the optical power monitor and the synchronous detection circuit function as an error detection circuit. The obtained error signal is sent to the PI control circuit, and the PI control circuit determines the magnitude of the feedback signal to the bias power supply based on the error signal and controls the bias power supply.

FIG. 10 is a schematic diagram showing the influence of the dither signal of the frequency fd on the delay amount control unit, and FIG. 11 is a schematic diagram showing the influence of the dither signal of the frequency fd on the light intensity of the output light. Reference numeral L91 in FIG. 10 and reference numeral L93 in FIG. 11 represent a case where the delay amount generated by the delay amount control unit is proportional to the bias voltage (hereinafter, referred to as “case 1”). In this case, as shown by reference numeral L91 in FIG. 10, the delay amount increases linearly with respect to the bias voltage. However, as shown by reference numeral L93 in FIG. 11, due to the nature of MZI, the output light intensity exhibits a sinusoidal response to the bias voltage.

Reference numeral L92 in FIG. 10 and reference numeral L94 in FIG. 11 are cases where the delay amount generated by the delay amount control unit shows a non-linear response to the bias voltage (hereinafter, referred to as “case 2”). As an example, the case where the delay amount control unit is a heater corresponds to. The amount of heat of the heater is proportional to the product of current and voltage, and the current is the voltage divided by the heater resistance. Therefore, the amount of heat of the heater is approximately proportional to the square of the bias voltage, and the amount of delay of the optical waveguide is also proportional to the square of the bias voltage. Therefore, the sinusoidal response of the output light intensity is distorted, and the higher the bias voltage, the more severe the change in light intensity.

In either case 1 or 2, the output light intensity is intensity-modulated by the dither signal of frequency fd. The frequency component of this intensity modulation is generally fd, but when the output light intensity takes an extreme value, it is 2 fd instead of fd. The reason will be explained below. 10 and 11 show the bias voltages Va1 and Va2 at which the output light intensity becomes the maximum value in the case 1, and the bias voltages Vb1 and Vb2 at which the output light intensity becomes the maximum value in the case 2. At these bias voltages, the output light intensity fluctuates with a maximum value in between due to dithering. That is, the maximum value is once passed in the process in which the bias voltage is slightly increased by dithering, and the maximum value is passed again in the process in which the bias voltage is slightly decreased by dithering. Since these processes are repeated at the frequency fd, the intensity modulation component superimposed on the output light becomes 2 fd, and the frequency component of the fd is suppressed. Therefore, when the control target is to maximize the light intensity, synchronous detection is performed at the frequency fd, the synchronous detection result is used as the error signal, and the absolute value of the error signal is the minimum (that is, the component of the frequency fd is the minimum). It suffices to control so that it becomes.

Here, when correct control is made and the bias voltage is locked to Va1 and Va2 (in the case of case 1) or the bias voltage is locked to Vb1 and Vb2 (in the case of case 2), the amplitude of the intensity modulation of the frequency 2fd superimposed on the output light. Think about what will happen. In the case 1 where the delay amount is proportional to the bias voltage, the amplitude of the intensity modulation of 2fd is uniform in both Va1 and Va2 as shown by the reference numeral L93 in FIG. However, in the case 2 where the delay amount is not proportional to the bias voltage, as shown by the reference numeral L94 in FIG. 11, the amplitude of the intensity modulation of the frequency 2fd in Vb1 is small, and the amplitude of the intensity modulation of the frequency 2fd in Vb2 is large. Further, as is clear from the figure, when locked at a higher bias voltage, the amplitude of the intensity modulation at the frequency 2fd increases at an accelerating rate. As described above, the amplitude of the dither signal is generally kept small so as not to increase the influence on the output light, but even if the amplitude of the dither signal at the time of being applied to the bias voltage is constant. The problem arises that the dither component superimposed on the output light or the amplitude of its double wave is not uniform.

A further problem is that when the delay amount control unit is non-linear with respect to the bias voltage as in the solid line, PI control becomes very difficult. Various studies have been conducted on how to obtain the magnitude of the error signal from the magnitude of the feedback signal, but in the case of reference numeral L94 in FIG. 11, the higher the bias voltage, the higher the bias voltage. The amount of variation in light intensity with respect to the amount of change increases. Therefore, proper convergence cannot be achieved by simple PI control, and in the worst case, divergence occurs.

Moreover, in the explanation so far, the control target is to maximize the light intensity. However, depending on the operation, it may be desirable to set the light intensity to half of the maximum value. This type of control is relatively easy if the amount of delay is proportional to the bias voltage. As shown by reference numeral L93 in FIG. 11, when the light intensity is a sinusoidal response and is half the maximum value of the light intensity, the inclination of the reference numeral L93 is maximum. Therefore, the bias voltage may be controlled so that the intensity modulation component of the frequency fd superimposed on the output light is maximized. That is, synchronous detection may be performed at the frequency fd, the synchronous detection result may be used as an error signal, and the bias voltage may be controlled so that the absolute value of the error signal becomes maximum. However, when the delay amount is not proportional to the bias voltage, as shown by the reference numeral L94 in FIG. 11, the light intensity is not a simple sinusoidal response, and the inclination of the reference numeral L94 when it is half of the maximum value of the light intensity is Not always the maximum. Therefore, the synchronous detection result cannot be used as an error signal, and the control becomes more difficult.

In view of the above circumstances, it is an object of the present invention to provide a control circuit capable of quickly bringing a controlled object having a function of controlling a system by a non-linear response to a target state.

One aspect of the present invention is an error detection unit that detects an error signal that quantifies the difference between the current state of the system and the reference state, and an adjustment unit that adjusts the control target included in the system based on the control signal. By generating the control signal based on the error signal, feedback control or feedforward control to the controlled object is performed via the adjusting unit, and the system is controlled to be close to the reference state. The control unit is a control circuit that performs a non-linear compensation calculation process that compensates for a non-linear response relationship between the control signal and a state change of the controlled object when generating the control signal.

One aspect of the present invention is the control circuit described above, wherein the system has a branch portion that branches the input light into two, and a first arm that transmits one of the branched lights. A Machzenda interferometer having a second arm for transmitting the other light, and a confluence portion for merging the light transmitted through the first arm and the light transmitted through the second arm. The control target is a delay amount control unit that controls the delay amount of the light of at least one of the first arm or the second arm, and the adjustment unit controls the delay amount control unit. Adjusting based on the signal, the control unit compensates for the non-linear response relationship between the delay amount generated in the delay amount control unit and the control signal in the non-linear compensation arithmetic processing.

One aspect of the present invention is the control circuit described above, wherein the adjusting unit adjusts the delay amount control unit to delay the light of the first arm and the light of the second arm. Is push-pulled, and when the delay amount of one of the first arm and the second arm is increased by X, the control unit adjusts the delay amount of the other arm by X. The delay amount control unit is controlled via the unit, and the nonlinear compensation arithmetic processing compensates for the nonlinear response relationship between the control signal and the X.

One aspect of the present invention is the control circuit described above, wherein the delay amount control unit includes a first heater that changes the delay amount of the light transmitted through the first arm by thermal expansion, and the second heater. It consists of a second heater that changes the delay amount of the light transmitted through the arm by thermal expansion, and the adjusting unit supplies voltage or current to the first heater and the second heater based on the control signal. Then, the control unit uses a predetermined offset power, and when the heater power of one of the first heater and the second heater is a value obtained by adding ΔW to the offset power, the other. The delay amount control unit is controlled via the adjustment unit so that the heater power of the above is a value obtained by subtracting the ΔW from the offset power, and the nonlinear compensation calculation process is performed between the control signal and the ΔW. Compensate for non-linear response relationships.

One aspect of the present invention is the control circuit described above, wherein the control unit includes an oscillator oscillating at a predetermined frequency fd, the amplitude of a dither signal generated by the non-linear response relationship, and the adjustment unit. A dithering efficiency compensating unit that compensates for a non-linear correspondence with a change in the state of the system when dithering is applied to the controlled object is provided, and the adjusting unit is provided with the dithering efficiency compensating unit. The dither signal of the frequency fd whose correspondence is compensated is added to the control target, and the error detection unit synchronously detects the state of the system in which the adjustment unit adds dithering to the control target, and the system is detected. The error signal that quantifies the difference between the current state and the reference state of is detected.

One aspect of the present invention is the control circuit described above, wherein the system has a branch portion that branches the input light into two, and a first arm that transmits one of the branched lights. A Mach Zenda interferometer having a second arm for transmitting the other light, and a combiner for merging the light transmitted through the first arm and the light transmitted through the second arm. The control target is a delay amount control unit that controls the delay amount of at least one of the first arm and the second arm, and the adjustment unit is the first arm and the first arm. The frequency is added to the delay amount of the light transmitted through at least one of the second arms while controlling the delay amount of the light transmitted through at least one of the first arm and the second arm. The delay amount control unit is controlled so as to add dithering of fd, and the dithering efficiency compensating unit controls the non-linearity between the amplitude of the dither signal and the fluctuation of the delay amount caused by the dithering by the adjusting unit. By compensating for the corresponding relationship, the amplitude of increase / decrease in the amount of delay caused by dithering is kept constant.

One aspect of the present invention is the control circuit described above, wherein the adjusting unit dithers the delay amount Y of the first arm and the second arm in a range of amplitude ΔY, and the first aspect thereof. When the delay amount of one of the arm 1 and the second arm changes from Y + ΔY to Y−ΔY, dithering is performed so that the delay amount of the other arm changes from Y−ΔY to Y + ΔY. Moreover, the dithering efficiency compensating unit compensates for the non-linear correspondence so that ΔY is always constant regardless of the value of Y.

One aspect of the present invention is the control circuit described above, wherein the delay amount control unit has a heater that thermally expands at least one of the first arm and the second arm, and the adjustment unit has a heater. Dithering is applied to the voltage or current applied to the heater, and the dithering efficiency compensator uses the voltage or current applied to the heater so that the amplitude of increase or decrease in the calorific value of the heater due to dithering becomes constant. Compensates for the non-linear correspondence with the calorific value of the heater.

INDUSTRIAL APPLICABILITY According to the present invention, a controlled object having a function of controlling a system by a non-linear response can be quickly brought closer to a target state.

It is a block diagram of MZI and a control circuit by 1st Embodiment. It is a figure which shows the change of the delay amount by the same embodiment. It is a figure which shows the light intensity of the output light by the same embodiment. It is a block diagram of the optical modulator and the control circuit by 2nd Embodiment. It is a figure which shows the change of the light intensity of the output light by the same embodiment. It is a block diagram of the optical modulator and the control circuit by 3rd Embodiment. It is a figure which shows the light intensity of the output light by the same embodiment. It is a block diagram of the control circuit by 4th Embodiment. It is a block diagram of the control circuit of MZI by the prior art. It is a schematic diagram which shows the influence which the dither signal of a frequency fd has on a delay amount control part. It is a schematic diagram which shows the influence which the dither signal of a frequency fd has on the light intensity of an output light.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. This embodiment is suitable for state control of a circuit that responds non-linearly to a control signal. In particular, it is suitable for state control of a Mach-Zehnder interferometer (MZI) of the type in which the control mechanism of the optical path length has a non-linear response or an optical modulator using such MZI.

[First Embodiment]
FIG. 1 is a block diagram of the MZI 1 and the control circuit 2 according to the first embodiment of the present invention. The MZI 1 has a first optical coupler 11, a first arm 12a, a second arm 12b, a heater 13, and a second optical coupler 14. The first optical coupler 11 inputs CW light. The first optical coupler 11 branches the input CW light and outputs one of the branched lights to the first arm 12a and the other branched light to the second arm 12b. The heater 13 is an example of a delay amount control unit that performs a non-linear response. The heater 13 generates heat by the bias voltage applied from the control circuit 2 to thermally expand the first arm 12a, and changes the delay of the optical phase of the light propagating through the first arm 12a to change the MZI1. Controls the light intensity or light phase of the light output of. The second optical coupler 14 combines the modulated light output from the first arm 12a and the light output from the second arm 12b and outputs the light from the MZI1.

The control circuit 2 includes a PI control circuit 21, a first arithmetic circuit 22, a bias power supply 23, an oscillator 24, a second arithmetic circuit 25, an adder 26, a tap 27, and an error detection circuit 28. To prepare for. The heater 13 may be regarded as a component of the control circuit 2.

The PI control circuit 21 instructs the first calculation circuit 22 of the amount of heat of the heater 13. The first arithmetic circuit 22 converts the amount of heat instructed by the PI control circuit 21 into a bias voltage, and controls the bias power supply 23 based on the converted bias voltage. The bias power supply 23 supplies a bias voltage according to the control of the first arithmetic circuit 22. The oscillator 24 oscillates at the frequency fd and generates a reference clock at the frequency fd. The second arithmetic circuit 25 calculates the voltage amplitude for applying dithering so that the amplitude of the amount of heat of the heater 13 becomes constant. The adding unit 26 superimposes the voltage amplitude calculated by the second arithmetic circuit 25 on the output voltage of the bias power supply 23, and applies a dither signal having a frequency fd to the heater 13.

The tap 27 branches the light output from the second optical coupler 14 of the MZI1 and outputs the light to the error detection circuit 28. The error detection circuit 28 includes an optical power monitor 281 and a synchronous detection circuit 282. The optical power monitor 281 measures the optical intensity (optical power) of the light branched by the tap 27, and detects a dither component having a frequency fd or a harmonic thereof. The synchronous detection circuit 282 synchronously detects the measurement result by the optical power monitor 281 by the reference clock of the frequency fd output from the oscillator 24. The PI control circuit 21 determines a target value of the amount of heat of the heater 13 using the synchronous detection result output from the synchronous detection circuit 282 as an error signal, and sends a feedback signal to the first arithmetic circuit 22.

Similar to the example shown in FIG. 9, a case where the optical power of the output light of the MZI1 is controlled to the maximum by the control circuit 2 is taken as an example. As in FIG. 9, for ease of explanation, it is assumed that the MZI1 is unmodulated and the output light is CW light. The delay amount control unit in FIG. 9 is a heater 13 in this embodiment, and the behavior of the heater 13 is non-linear with respect to the bias voltage as shown by reference numeral L92 in FIG. 10 and reference numeral L94 in FIG.

When the bias voltage is defined as Vbias, the amount of heat generated by the heater 13 is proportional to (Vbias) ² / R. Here, R is the resistance value of the heater 13. Strictly speaking, the resistance value R is not a constant but a function of the amount of heat of the heater 13. The increase in the amount of delay caused by the heater 13 is proportional to the amount of heat in the heater 13.

In the present embodiment, the feedback signal generated by the PI control circuit 21 is calculated with reference to the amount of heat of the heater 13. That is, the control circuit 2 does not directly feed back the feedback signal to the bias voltage, but in the first arithmetic circuit 22, the target value of the calorific value of the heater 13 represented by the feedback signal is multiplied by the resistance value R, and then the square root thereof. By taking the above, the target value of the amount of heat of the heater 13 is converted into the bias voltage Vbias and then fed back to the bias power supply 23. For example, whether the PI control circuit 21 applies a proportional coefficient to the error signal output from the error detection circuit 28 (P control), performs integral processing (I control), or performs differential processing (D control). At least one of the above is performed to determine the target value of the calorific value of the heater 13. As described above, the change in the amount of heat of the heater 13 is proportional to the change in the amount of delay in the first arm 12a, but the amount of delay in the first arm 12a and the bias voltage Vbias have a non-linear response relationship. Therefore, when the first arithmetic circuit 22 feeds back the target value of the calorific value determined by the PI control circuit 21 to the calorific value of the heater 13, the non-linear arithmetic processing for compensating the above-mentioned nonlinear response relationship is used as the feedback signal. On the other hand, the bias voltage Vbias is generated.

Next, consider the amplitude of the dither signal Vdither superimposed on the bias voltage Vbias. In the present embodiment, dithering is applied so that the amplitude of the heat quantity of the heater 13 is constant, instead of making the amplitude of the dither signal Vdither constant. The amount of change in the amount of heat of the heater 13 caused by dithering is defined as ± ΔW. Further, the bias voltage increased by dithering is defined as VbiasUp, and the bias voltage decreased by dithering is defined as VbiasDn. From these definitions, the following equation (1) is obtained.

(VbiasUp) ² / R- (Vbias) ² / R
= (Vbias) ² / R- (VbiasDn) ² / R
= ΔW ... (1)

From the formula (1), VbiasUp becomes the formula (2), and VbiasDn becomes the formula (3).

VbiasUp = (RΔW + (Vbias) ² ) ^0.5 ... (2)

VbiasDn = (-RΔW + (Vbias) ² ) ^0.5 ... (3)

From the equation (2), the voltage increasing from the bias voltage Vbias due to dithering becomes the equation (4), and from the equation (3), the voltage decreasing from the bias voltage Vbias due to the dithering becomes the equation (5).

VbiasUp-Vbias = (RΔW + (Vbias) ² ) ^0.5 -Vbias ... (4)

VbiasDn-Vbias = (-RΔW + (Vbias) ² ) ^0.5 -Vbias ... (5)

From the equation (4), the upper limit of the amplitude of the dither signal Vdither is the following equation (6), and from the equation (5), the lower limit of the amplitude of the dither signal Vdither is the following equation (7).

Upper limit of Vdither = (RΔW + (Vbias) ² ) ^0.5 -Vbias ... (6)

Lower limit of Vdither = (-RΔW + (Vbias) ² ) ^0.5 -Vbias ... (7)

If the dither signal Vdither is added to the bias voltage Vbias and the Vdither is modulated at the frequency fd within the upper and lower limits shown in the equations (6) and (7), the modulation amplitude of the heat quantity of the heater 13 is determined by Vbias. It can be kept at a constant value ± ΔW.

The second arithmetic circuit 25 has (RΔW + (Vbias) ² ) ^0.5 −Vbias of the equation (6) and (−RΔW + (Vbias) of the equation (7) based on the value of the bias voltage Vbias and the resistance value R. ² ) Calculate ^0.5 -Vbias. The value of the bias voltage Vbias is obtained from the output of the first arithmetic circuit 22. The value of the resistance value R is an amount that depends on the amount of heat of the heater 13 and the characteristics of the heater. Therefore, the second arithmetic circuit 25 reads the value of the resistor R corresponding to the average value (Vbias) ² / R of the amount of heat from the data table stored in advance in the storage unit prepared internally or externally. Perform the calculation of.

When the heat quantity dependence of the resistance R is relatively small, it may be regarded as a constant and the configuration may be such that the reading process from the data table is omitted.

FIG. 2 is a diagram showing a change in the amount of delay obtained by performing the above arithmetic processing, and FIG. 3 is a diagram showing a change in the light intensity (output light power) of the output light obtained by performing the above arithmetic processing. It is a figure which shows. Since these figures are plotted with the horizontal axis as the heater power, the non-linearity in the reference numeral L92 in FIG. 10 and the reference numeral L94 in FIG. 11 is suppressed. The bias voltages Vb1 and Vb2 in FIGS. 10 and 11 are replaced by the heater powers W1 and W2 in FIGS. 2 and 3. The heater power and delay amount are dithered with an amplitude and frequency fd of ± ΔW. In the present embodiment, the purpose is to maximize the output light power, but at this time, the dither component of the frequency 2fd superimposed on the output light intensity becomes the maximum, and the dither component of the frequency fd becomes the minimum.

Therefore, the PI control circuit 21 controls the power of the heater 13 so that the dither component of the frequency fd superimposed on the output light intensity is minimized. The synchronous detection result obtained by the synchronous detection circuit 282 is an error signal indicating how far the current value of the bias voltage is from the optimum value (a state that serves as a reference for calculating the error). The PI control circuit 21 defines a feedback signal so as to adjust the heater power so that the error signal becomes 0. Within the range shown in FIG. 3, the synchronous detection result becomes 0 at three locations of the heater powers W1, W2, and (W1 + W2) / 2. The heater power to be selected is W1 or W2, but when the vicinity of these heater powers is compared with the vicinity of the heater power (W1 + W2) / 2, the gradient of the output light intensity with respect to the heater power is reversed. Since the sign of the synchronous detection result is also reversed by this reversal, the PI control circuit 21 selects the correct heater power (that is, W1 or W2) from the sign of the synchronous detection result in the vicinity of the heater power at which the synchronous detection result becomes 0. It becomes possible to do.

When the heater power is correctly locked by W1 or W2, the dither component superimposed on the output light intensity is 2fd instead of fd as described above, but the amplitude of the intensity modulation is locked to W2 even when locked to W1. Even if you do, it doesn't change. Therefore, the problem that the amplitude of the double wave becomes unpredictable as shown in FIG. 11 does not occur.

[Second Embodiment]
The difference between the present embodiment and the first embodiment is that in the present embodiment, MZI is used as an optical modulator instead of a mere interferometer, and modulated light is output instead of CW light. The delay amount control unit of the light modulator of the present embodiment is a heater as in the first embodiment. The signal format is an intensity modulated signal.

FIG. 4 shows a block diagram of the light modulator 3 and the control circuit 4 according to the second embodiment of the present invention. In the figure, the same parts as the MZI 1 and the control circuit 2 of the first embodiment shown in FIG. 1 are designated by the same reference numerals, and the description thereof will be omitted.

The light modulator 3 has a configuration in which the optical phase controller 31 is installed on the first arm 12a of the MZI1 of the first embodiment. As described above, the light modulator 3 is an MZI in which the optical phase controller 31 and the heater 13 are provided on one of the two arms. The instantaneous light intensity of the output light from the light modulator 3 is modulated by an electrical data signal. The drive amplifier 51 amplifies the data signal, and the optical phase controller 31 modulates the delay amount of the light transmitted through the first arm 12a at high speed by the data signal amplified by the drive amplifier 51.

The difference between the control circuit 4 and the control circuit 2 of the first embodiment is that the control circuit 41 is provided in place of the PI control circuit 21. The output of the drive amplifier 51 fluctuates positively and negatively around the GND level, but the PI control circuit 41 has the maximum power of the optical output from the light modulator 3 at the moment when the output of the drive amplifier 51 reaches the zero level. Vbias is controlled so that it becomes half of the value.

When the output of the drive amplifier 51 is positive, the optical phase controller 31 changes the optical phase so that the output light intensity of the light modulator 3 becomes large (or small), and the output of the drive amplifier 51 is negative. In this case, the optical phase controller 31 changes the optical phase so that the output light intensity of the light modulator 3 becomes small (or large). As a result, the light intensity of the output from the light modulator 3 is intensity modulated corresponding to the data signal. When the band of the optical power monitor 281 is smaller than the bit rate of the data signal, the optical power monitor 281 cannot detect the change in the instantaneous intensity and monitors the average value. Therefore, the output optical power measured by the optical power monitor 281 is substantially the same as that shown in FIG.

Also in the present embodiment, as in the first embodiment, the PI control circuit 41 determines the feedback signal based on the heater power, and the first arithmetic circuit 22 converts the heater power into Vbias. Further, by using the second arithmetic circuit 25, the fluctuation of the heater power due to dithering is kept constant regardless of the value of Vbias. However, in the present embodiment, the PI control circuit 41 adjusts the heater power so that the absolute value of the synchronous detection result is maximized.

FIG. 5 is a diagram showing the relationship between the heater power and the output light intensity of the light modulator 3. The synchronous detection result obtained by the synchronous detection circuit 282 becomes an error signal, and the PI control circuit 41 defines a feedback signal for adjusting the heater power so that the absolute value of the error signal becomes maximum. This means that the heater power is adjusted so that the value obtained by differentiating the output light intensity of the light modulator 3 with the heater power is maximized. That is, the optimum value of the heater power is W3 or W4 shown in the figure. As intended, the output light intensity is half of the maximum value at this heater power.

It should be noted here that if dithering is applied so that the amplitude of the bias voltage, not the amount of heat of the heater 13, is constant, the work of maximizing the absolute value of the synchronous detection result is indicated by the reference numeral 11 in FIG. The work is to maximize the inclination of L94. However, since this gradient increases at an accelerating rate in the region where the bias voltage is large, it cannot be locked to the optimum point. In the present embodiment, since dithering is applied so that the amplitude of the amount of heat of the heater 13 becomes constant, it can be locked to an appropriate heater power (W3 or W4 in FIG. 5).

Since the slope of the output light power with respect to the heater power is opposite between the case where the heater power is W3 and the case where the heater power is W4, the sign of the synchronous detection result is reversed. When this sign is inverted, the correspondence between the magnitude of the data signal and the magnitude of the instantaneous intensity of the modulated light is inverted, and the logic is inverted. It is possible to generate light intensity modulation regardless of which logic is selected, but if you want to control the presence or absence of logic inversion, limit the sign of the synchronous detection result in advance and then synchronize the detection result. Control may be performed so that the absolute value of is maximized.

[Third Embodiment]
FIG. 6 shows a block diagram of the light modulator 6 and the control circuit 7 according to the third embodiment of the present invention. In the figure, the same parts as those of the second embodiment shown in FIG. 4 are designated by the same reference numerals, and the description thereof will be omitted.

The light modulator 6 differs from the light modulator 3 of the second embodiment shown in FIG. 4 in that the first arm 12a is replaced with the optical phase controller 31 and the heater 13, and the first optical phase controller 61a and The point is that the first heater 62a is provided, and the second arm 12b is provided with the second optical phase controller 61b and the second heater 62b. As described above, the light modulator 6 is an MZI having an optical phase controller and a heater in each of the two arms.

The first optical phase controller 61a applies a data signal amplified by the differential output type drive amplifier 52 to modulate the optical phase of the light transmitted through the first arm 12a. The first heater 62a generates heat by the bias voltage applied from the control circuit 7 to thermally expand the first arm 12a, and modulates the intensity or optical phase of the light transmitted through the first arm 12a. ..

The second optical phase controller 61b applies a data signal amplified and inverted by the differential output type drive amplifier 52 to change the delay of the optical phase of the light propagating through the second arm 12b. The intensity or optical phase of the light of the light output of MZI is controlled. The second heater 62b generates heat by the bias voltage applied from the control circuit 7 to thermally expand the second arm 12b, and changes the delay of the optical phase of the light propagating through the second arm 12b. Controls the light intensity or the light phase of the light output of the MZI.

The difference between the control circuit 7 and the control circuit 5 of the second embodiment shown in FIG. 4 is that the PI control circuit 41, the first arithmetic circuit 22, the bias power supply 23, the second arithmetic circuit 25, and the adder 26 are replaced. A point including a PI control circuit 71, a first arithmetic circuit 72a, a first bias power supply 73a, a second arithmetic circuit 75a and a first adder 76a, and a third arithmetic circuit 72b and a second bias. A power source 73b, a fourth arithmetic circuit 75b, and a second addition unit 76b are further provided. The first heater 62a and the second heater 62b may be regarded as components of the control circuit 7.

The PI control circuit 71 instructs the first calculation circuit 72a of the heater power of the first heater 62a by the feedback signal based on the error detection signal, and the third calculation circuit 72b is instructed by the heater power of the second heater 62b. Is indicated by a feedback signal.

The first arithmetic circuit 72a calculates a bias voltage based on the heater power instructed by the feedback signal from the PI control circuit 71, and controls the first bias power supply 73a based on the calculated bias voltage. The first bias power supply 73a, the second calculation circuit 75a, and the first addition unit 76a operate in the same manner as the bias power supply 23, the second calculation circuit 25, and the addition unit 26 of the second embodiment, respectively. That is, the first bias power supply 73a supplies the bias voltage Vbias1 according to the control of the first arithmetic circuit 72a. The second arithmetic circuit 75a calculates a dither signal Vdither1 having a frequency fd for applying dithering so that the amplitude of the amount of heat of the first heater 62a becomes constant. The first addition unit 76a superimposes the dither signal Vdither1 calculated by the second arithmetic circuit 75a on the bias voltage Vbias1 output from the first bias power supply 73a, and applies the first bias voltage to the first heater 62a. Apply to.

The third arithmetic circuit 72b, the second bias power supply 73b, the fourth arithmetic circuit 75b, and the second adder 76b are each the first arithmetic circuit with respect to the second bias voltage applied to the second heater 62b. The same processing as the 72a, the first bias power supply 73a, the second arithmetic circuit 75a, and the first addition unit 76a is performed. That is, the third arithmetic circuit 72b calculates the bias voltage based on the heater power instructed by the feedback signal from the PI control circuit 71, and controls the second bias power supply 73b based on the calculated bias voltage. The second bias power supply 73b supplies the bias voltage Vbias2 according to the control of the third arithmetic circuit 72b. The fourth arithmetic circuit 75b calculates the dither signal Vdither2 having a frequency fd for applying dithering so that the amplitude of the amount of heat of the second heater 62b becomes constant. However, the dither signal Vdither2 lags the phase of the dither signal Vdither1 by 180 degrees. The second addition unit 76b applies a second bias voltage to the second heater 62b by superimposing the dither signal Vdither2 output by the fourth arithmetic circuit 75b on the bias voltage Vbias2 of the second bias power supply 73b. ..

The delay amount control unit of the light modulator 6 of the present embodiment is a heater as in the first and second embodiments. In this embodiment, as in the second embodiment, MZI is used not as a mere interferometer but as an optical modulator, but the signal format is Carrier Suppressed Return to Zero (CS-RZ). There are two differences between the present embodiment and the second embodiment due to the difference in the signal format. First, it is necessary to drive the two arms of the MZI in the light modulator 6, that is, the first arm 12a and the second arm 12b by push-pull. Therefore, instead of the drive amplifier 51 of the second embodiment shown in FIG. 4, a differential output type drive amplifier 52 is used, and one of the two contradictory outputs of the differential output type drive amplifier 52 is the first light. In addition to the phase controller 61a, the other is added to the second optical phase controller 61b. As a result, the optical phases of the two arms of the first arm 12a and the second arm 12b of the light modulator 6 are modulated to push-pull. That is, when the first optical phase controller 61a advances (or delays) the optical phase of the CW light transmitting the first arm 12a, the second optical phase controller 61b transmits the second arm 12b. Delays (or advances) the optical phase of CW light.

Another difference from the second embodiment is that the PI control circuit 71 minimizes the output optical power of the light modulator 6 at the moment when the output of the differential output type drive amplifier 52 is at 0 level. The point is to control the bias voltage. In this embodiment, the bias voltage is also controlled by push-pull. That is, the PI control circuit 71 uses the first heater 62a of the first arm 12a and the second heater 62b of the second arm 12b, and when the heater power of one is increased, the heater power of the other is decreased. Take control.

In generating the CS-RZ signal, push-pull drive is indispensable for the drive signal as described above, but push-pull is not necessarily required for the control of the bias voltage. However, by using push-pull control, it is possible to secure the same optical phase difference control range with half the calorific value compared to the configuration using a single heater, so the current of the first and second bias power supplies There is an advantage that the capacity can be kept small.

However, unlike the control of the optical phase by the Pockels effect, the heat generation of the heater and the change in the delay amount do not depend on the positive or negative of the applied voltage. Therefore, in order to control the heater power to push-pull, an _offset heater power Woffset is prepared, the heater power of the first heater 62a is ( _Woffset + W _bias ), and the heater power of the second heater 62b is set. (W _offset -W _bias ) is controlled. Here, W _offset is a positive value, but W _bias can take both positive and negative values. Since the amount of heat does not take a negative value, (W _offset + W _bias ) and (W _offset -W _bias ) are always positive values, so that the limitation of -W _offset <W _bias <W _offset arises.

Also in this embodiment, dithering is added to the heater power (W office + W _bias ) of the first heater _62a and the heater power (W office − W _bias ) of the second heater _62b . In order to improve the detection efficiency of the synchronous detection circuit 282, the dithering applied to the heater power of the first heater 62a and the dithering applied to the heater power of the second heater 62b have the same frequency fd and are in opposite phase. It is desirable to do.

FIG. 7 shows the change in the light intensity of the output light from the light modulator 6 with respect to W _bias . As mentioned above-W _offset and _Woffset respectively are the lower and upper limits of the change from _Woffset . When paying attention to the intensity modulation component derived from dithering superimposed on the output light intensity, the fact that the frequency 2fd component is the maximum and the frequency fd component is the minimum is that the output light intensity is the same as in the first embodiment. When it becomes the maximum or the minimum. In the range shown in FIG. 7, when the output light intensity becomes the maximum or the minimum, it is when W _bias = W5, W _bias = w6, or W _bias = (W5 + W6) / 2. At this time, the synchronous detection result becomes zero as in the first embodiment. The choice should be W _bias = W5 or W _bias = W6, and W _bias = (W5 + W6) / 2 must be excluded. Similar to the first embodiment, the PI control circuit 71 can correctly select W _bias = W5 or W _bias = W6 from the sign of the synchronous detection result in the vicinity of W _bias where the synchronous detection result becomes 0. ..

As described above, the synchronous detection result obtained by the synchronous detection circuit 282 becomes an error signal, but the PI control circuit 71 determines a feedback signal for changing the W _bias so that the error signal approaches 0.

By the way, in the present embodiment, since the first bias voltage applied to the first heater 62a and the second bias voltage applied to the second heater 62b are different, the control circuit 7 is connected to the first bias power supply 73a. It is necessary to prepare a second bias power supply 73b. The bias voltage Vbias1 output from the first bias power supply 73a is obtained by converting the power Woffset + _Wbias into a voltage, and the _bias voltage Vbias2 output from the second _bias power supply 73b is the power Woffset _−Wbias . Obtained by converting to voltage. Moreover, since the W _office is a constant, the variable to be controlled is only W _bias , and the feedback signal defined by the PI control circuit 71 is calculated with only W _bias as the control target. This feedback signal is fed back to the first arithmetic circuit 72a and the third arithmetic circuit 72b. In the same manner as in the first embodiment, the first arithmetic circuit 72a converts the power W _offset + W _bias into _Vbias1 , and the third arithmetic circuit 72b converts Woffset −W _bias into Vbias 2.

In this embodiment, since dithering is applied to the two arms of the first arm 12a and the second arm 12b, two dither signals of Vdizer1 and Vdiser2 are used. The amplitude of Vdither1 is set so that the amplitude of the heater power of the first heater 52a is constant, and the amplitude of Vdither2 is set so that the amplitude of the heater power of the second heater 52b is constant. This setting is calculated by the second calculation circuit 75a and the fourth calculation circuit 75b by the same method as in the first embodiment. That is, the second arithmetic circuit 75a sets Vdizer1 based on the values of Vbias1 and the heater resistance value R, and the fourth arithmetic circuit 75b sets Vdizer2 based on the values of Vbias2 and the heater resistance value R. do. However, the fourth arithmetic circuit 75b delays the phase of Vdiser2 by 180 degrees with respect to Vdiser1 in order to improve the efficiency of dithering. This can be easily realized by inverting the voltage of the dither signal.

[Fourth Embodiment]
In the first to third embodiments described above, the feedback system that generates the feedback signal by the PI control circuit has been described. This embodiment is applied to a feed forward system. Hereinafter, the difference from the second embodiment will be mainly described, but this difference may be applied to the first embodiment and the third embodiment.

FIG. 8 shows a configuration diagram of the control circuit 8 of the present embodiment. In the figure, the same parts as those of the second embodiment shown in FIG. 4 are designated by the same reference numerals, and the description thereof will be omitted. The control circuit 8 shown in FIG. 4 differs from the control circuit 4 of the second embodiment shown in FIG. 4 in that it further includes a feedforward control circuit 81 and an optical attenuator 82. With such a configuration, the control circuit 8 of the present embodiment controls the optical attenuator 82 provided in the subsequent stage of the tap 27 by the feedforward signal from the feedforward control circuit 81.

As described in the second embodiment, the PI control circuit 41 defines a feedback signal for adjusting the heater power so that the absolute value of the error signal (that is, the synchronous detection result) is maximized. Therefore, when the error signal is close to zero, the heater power is far from the target value, and the light intensity input to the transmission line may change suddenly until the feedback loop stabilizes. is expected. In particular, when the coefficient of P in PI control is set to be large, the vibration is repeated in the vicinity of the optimum value of the light intensity until the light intensity converges. This corresponds to an irregular and unpredictable disturbance from the viewpoint of the transmission system, and in some cases causes unstable operation of the optical device included in the transmission system.

In order to avoid such a situation, in the present embodiment, the feed forward control circuit 81 is set to the optical attenuator 82 during the period when it is determined that the error signal is close to zero and the heater power is far from the target value. A feed-forward signal is sent to increase the optical loss of the optical attenuator 82. The feed forward control circuit 81 sends a feed forward signal to the optical attenuator 82 in a state where the absolute value of the error signal becomes large, the deviation between the heater power and the target value is within a predetermined value, and it is determined that the heater power is sufficiently close to the target value. To reduce the optical loss of the optical attenuator 82.

In the present embodiment, by adopting such a configuration, it is possible to suppress a sudden change in the optical input to the transmission line.

[Variations of each embodiment]
In each of the embodiments described so far, the delay amount control unit is assumed to be a heater. However, the present invention is not limited to this, and the delay amount control unit may use the non-linear optical effect of the semiconductor. In this case, the control target of the PI control circuits 21, 41, 71 is not the heater power but the delay amount of the arm, and the first calculation circuit 22, 72a and the third calculation circuit 72b shift the delay amount to the bias voltage. Control the conversion.

Further, in each of the embodiments described so far, the initial value of the feedback system is not mentioned. At the time of starting up the feedback system, the heater power of the first or second embodiment may be started from zero, or the W _bias of the third embodiment may be started from zero. More preferably, the control circuits 2, 4, 7, and 8 record the bias voltage when the feedback system converges, or the heater power and W _bias values, which are parameters of the control amount, in the non-volatile memory. The record may be maintained even after the system is lowered, and the feedback may be started using the numerical value recorded in the memory as the initial value when the feedback system is started again. Since there is a time difference between the time when it was last written to the non-volatile memory and the time when the startup starts, it is necessary to correct the bias voltage (or heater power or _Wbias ) recorded in the non-volatile memory. However, it has the effect of shortening the time required for convergence.

According to the above embodiment, the control circuit controls the system that makes a non-linear response to the control signal, particularly MZI of the type in which the control mechanism of the optical path length gives a non-linear response, or optical modulation using such MZI. In controlling the device, quick control can be achieved, and the influence of the dither signal used for synchronous detection on the system can be suppressed to a certain value or less.

According to the embodiment described above, the control circuit includes an error detection unit, an adjustment unit, and a control unit. The error detection unit detects an error signal that quantifies the difference between the current state of the system and the reference state. The adjusting unit adjusts the control target included in the system based on the control signal. For example, the adjusting unit is a bias power supply 23 and an addition unit 26, a first bias power supply 73a and a first addition unit 76a, a second bias power supply 73b and a second addition unit 76b. Further, for example, the control target is the heater 13, the first heater 62a, and the second heater 62b, and the control signal is the bias voltage applied to these. By generating a control signal based on the detected error signal, the control unit performs feedback control or feedforward control to the controlled object via the adjusting unit, and controls the system so as to be close to the reference state. For example, the control unit is a PI control circuit 21, 41, a first arithmetic circuit 22, a PI control circuit 71, a first arithmetic circuit 72a, and a third arithmetic circuit 72b. When performing feedforward control, for example, the adjusting unit is an optical attenuator, and the control signal is a feedforward signal. The control signal and the state change of the controlled object controlled based on the control signal have a non-linear response relationship. Therefore, the control unit generates a control signal that has been subjected to nonlinear compensation arithmetic processing to compensate for this nonlinear response relationship, and outputs the control signal to the adjustment unit. For example, the control unit performs at least one of multiplying the error signal by a proportional coefficient (P control), performing integral processing (I control), and performing differential processing (D control). Further, the nonlinear compensation calculation process is performed to generate a control signal as a feedback signal, which is output to the adjustment unit.

The system may be an MZI or MZI type optical modulator. The MZI transmits a branch portion (for example, the first optical coupler 11) that branches the input light into two, a first arm that transmits one of the branched lights, and the other branched light. It has a second arm, and a merging unit (for example, a second optical coupler 14) that combines the light transmitted through the first arm and the light transmitted through the second arm. The control target is a delay amount control unit that controls the delay amount of light of at least one of the first arm and the second arm of MZI. The adjusting unit adjusts the delay amount control unit based on the control signal. In this case, the control target by the delay amount control unit is the delay amount of the light transmitted through the arm, but may be the phase or light intensity of the MZI output light controlled by the delay amount. The control signal is an electric signal, and the control unit compensates for the non-linear response relationship between the delay amount generated in the delay amount control unit and the control signal in the non-linear compensation arithmetic processing.

The adjusting unit may change the first arm and the second arm of the Mach-Zehnder interferometer to push-pull by adjusting the delay amount control unit. When the control unit increases the delay amount by X with respect to the light transmitted through one of the first arm and the second arm, the control unit decreases the delay amount by X with respect to the light transmitted through the other arm. .. There is a non-linear response relationship between the value of the delay amount X and the control signal to the adjustment unit, and this non-linear correspondence is compensated by the non-linear compensation arithmetic processing.

The adjusting unit is for the first heater that changes the delay amount of the light transmitted through the first arm by thermal expansion and the second heater that changes the delay amount of the light transmitted through the second arm by thermal expansion. It may be configured to supply voltage or current. When the heater power of one of the first heater and the second heater is offset power W _offset + bias power W _bias , the control unit sets the other heater power to offset power W _offset -bias power W _bias . The delay amount control unit is controlled via the adjustment unit. The non-linear compensation arithmetic process compensates for the non-linear response relationship between the control signal and the bias power.

Further, the control unit may include an oscillator and a dithering efficiency compensation unit. Since the control signal and the state change of the controlled object based on the control signal have a non-linear response relationship, the amplitude of the dither signal and the state of the system when the adjusting unit applies dithering to the controlled object. A non-linear correspondence with fluctuations occurs. The dithering efficiency compensator compensates for this non-linear correspondence. The adjusting unit applies a dither signal having a frequency based on the output of the oscillator and having the dithering efficiency compensating unit compensating for the above-mentioned non-linear correspondence to the controlled object. For example, the dithering efficiency compensation unit is the second arithmetic circuit 25, 75a and the fourth arithmetic circuit 75b, and the adjustment unit is the addition unit 26, the first addition unit 76a, and the second addition unit 76b. .. The error detection unit synchronously detects the state of the system in which the adjustment unit adds dithering to the controlled object based on the control signal, and detects an error signal that quantifies the difference between the result of the synchronous detection and the reference state.

If the system is an MZI or MZI type light modulator, the control signal is an electrical signal. This may be an analog electric signal, or may be a digital signal indicating a control amount numerically. The oscillator outputs a reference signal with a frequency of fd. The control unit adjusts the dithering of the frequency fd to the delay amount of the light transmitted through at least one of the first arm and the second arm based on the dither signal of the electric signal fluctuating by the frequency fd. To control. The dithering efficiency compensator compensates for the non-linear correspondence between the amplitude of the dither signal and the fluctuation of the delay amount caused by the dithering by the adjusting unit, so that the amplitude of the increase / decrease in the delay amount caused by dithering is constant. Keep in.

Further, the control unit may control the adjustment unit so that the delay amount Y of the first arm and the second arm of the MZI or MZI type optical modulator is reciprocally dithered within the range of the amplitude ΔY. When the delay amount of one of the first arm and the second arm changes from Y + ΔY to Y−ΔY, the adjusting unit causes the delay amount of the other arm to change from Y−ΔY to Y + ΔY. Do dithering. The dithering efficiency compensating unit compensates for the non-linear correspondence so that the amplitude ΔY is constant regardless of the value of the delay amount Y.

Further, the delay amount control unit has a function of supplying a voltage or current to a heater that thermally expands at least one of the first arm and the second arm of the MZI or MZI type optical modulator, and the adjustment unit has a function. , Add dithering to the voltage or current applied to the heater. The dithering efficiency compensator compensates for the non-linear correspondence between the voltage or current applied to the heater and the calorific value of the heater so that the amplitude of increase / decrease in the calorific value of the heater due to dithering is constant. ..

Although the embodiments of the present invention have been described in detail with reference to the drawings, the specific configuration is not limited to this embodiment, and includes designs and the like within a range that does not deviate from the gist of the present invention.

1 ... MZI, 2 ... control circuit, 3 ... optical modulator, 4 ... control circuit, 5 ... control circuit, 6 ... optical modulator, 7 ... control circuit, 8 ... control circuit, 11 ... first optical coupler, 12a ... 1st arm, 12b ... 2nd arm, 13 ... heater, 14 ... second optical coupler, 21 ... PI control circuit, 22 ... first arithmetic circuit, 23 ... bias power supply, 24 ... oscillator, 25 ... 2nd arithmetic circuit, 26 ... adder, 27 ... tap, 28 ... error detection circuit, 31 ... optical phase controller, 41 ... PI control circuit, 51 ... drive amplifier, 52 ... differential output type drive amplifier, 52a ... 1st heater, 52b ... 2nd heater, 61a ... 1st optical phase controller, 61b ... 2nd optical phase controller, 62a ... 1st heater, 62b ... 2nd heater, 71 ... PI control Circuit, 72a ... 1st arithmetic circuit, 72b ... 3rd arithmetic circuit, 73a ... 1st bias power supply, 73b ... 2nd bias power supply, 75a ... 2nd arithmetic circuit, 75b ... 4th arithmetic circuit, 76a ... 1st adder, 76b ... 2nd adder, 81 ... Feed forward control circuit, 82 ... Optical attenuator, 281 ... Optical power monitor, 282 ... Synchronous detector circuit

Claims (8)

An error detector that detects an error signal that quantifies the difference between the current state of the system and the reference state, and
An adjustment unit that adjusts the control target included in the system based on the control signal,
By generating the control signal based on the error signal, feedback control or feedforward control to the controlled object is performed via the adjusting unit, and the control unit controls the system so as to be close to the reference state. Equipped with
The control unit includes a PI control circuit and a non-linear compensation arithmetic circuit.
The PI control circuit determines a target value of a state change of the controlled object by performing P control for multiplying the error signal by a proportional coefficient and I control for performing integration processing.
The nonlinear compensation arithmetic circuit performs nonlinear compensation arithmetic processing for compensating for the nonlinear response relationship between the control signal and the state change of the controlled object based on the target value instructed by the PI control circuit . Above, the control signal is generated.
Control circuit. The system includes a branch portion that branches the input light into two, a first arm that transmits the branched light, and a second arm that transmits the branched light. It is a Mach Zenda interferometer having the light transmitted through the first arm and the combined wave portion for merging the light transmitted through the second arm.
The control target is a delay amount control unit that controls the delay amount of the light of at least one of the first arm and the second arm.
The adjusting unit adjusts the delay amount control unit based on the control signal.
The nonlinear compensation arithmetic circuit compensates for the nonlinear response relationship between the delay amount generated in the delay amount control unit and the control signal in the nonlinear compensation arithmetic processing.
The control circuit according to claim 1. By adjusting the delay amount control unit, the adjusting unit controls the delay amount of the light of the first arm and the light of the second arm in a push-pull manner.
When the delay amount of one of the first arm and the second arm is increased by X, the control unit via the adjustment unit so as to decrease the delay amount of the other arm by X. Control the control unit,
The non-linear compensation arithmetic processing compensates for the non-linear response relationship between the control signal and the X.
The control circuit according to claim 2. The delay amount control unit changes the delay amount of the light transmitted through the first arm by thermal expansion of the first heater and the light delayed amount transmitted by the second arm by thermal expansion. It consists of a second heater that
The adjusting unit supplies a voltage or a current to the first heater and the second heater based on the control signal.
The control unit uses a predetermined offset power, and when the heater power of one of the first heater and the second heater is a value obtained by adding ΔW to the offset power, the other heater. The delay amount control unit is controlled via the adjustment unit so that the electric power is a value obtained by subtracting the ΔW from the offset power.
The non-linear compensation arithmetic processing compensates for the non-linear response relationship between the control signal and the ΔW.
The control circuit according to claim 2. The control unit
An oscillator that oscillates at a predetermined frequency fd,
A dithering efficiency compensating unit that compensates for a non-linear correspondence between the amplitude of the dither signal generated by the non-linear response relationship and the fluctuation of the state of the system when the dithering unit applies dithering to the controlled object. , Equipped with
The adjusting unit adds the dither signal of the frequency fd whose non-linear correspondence is compensated by the dithering efficiency compensating unit to the controlled object.
The error detection unit detects the error signal in which the adjustment unit synchronously detects the state of the system in which dithering is applied to the control target and quantifies the difference between the current state of the system and the reference state. do,
The control circuit according to claim 1. The system includes a branch portion that branches the input light into two, a first arm that transmits the branched light, and a second arm that transmits the branched light. It is a Mach Zenda interferometer having the light transmitted through the first arm and the combined wave portion for merging the light transmitted through the second arm.
The control target is a delay amount control unit that controls the delay amount of the light of at least one of the first arm and the second arm.
The adjusting unit controls the delay amount of light transmitted through at least one of the first arm and the second arm, and at least one of the first arm and the second arm. The delay amount control unit is controlled so as to add dithering of the frequency fd to the delay amount of the light transmitted through the arm.
The dithering efficiency compensating unit compensates for the non-linear correspondence between the amplitude of the dither signal and the fluctuation of the delay amount caused by the dithering by the adjusting unit, thereby causing the delay amount due to dithering. Keep the amplitude of increase / decrease constant,
The control circuit according to claim 5. The adjusting unit dithers the delay amount Y of the first arm and the second arm in a range of amplitude ΔY, and causes one of the first arm and the second arm. When the delay amount changes from Y + ΔY to Y−ΔY, dithering is performed so that the delay amount of the other arm changes from Y−ΔY to Y + ΔY, and
The dithering efficiency compensating unit compensates for the non-linear correspondence so that ΔY is always constant regardless of the value of Y.
The control circuit according to claim 6. The delay amount control unit has a heater that thermally expands at least one of the first arm and the second arm.
The adjusting unit applies dithering to the voltage or current applied to the heater.
The dithering efficiency compensating unit compensates for a non-linear correspondence between the voltage or current applied to the heater and the calorific value of the heater so that the amplitude of increase / decrease in the calorific value of the heater due to dithering becomes constant. do,
The control circuit according to claim 6.

Images