JP3343831B2 - Optical signal processor - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an optical signal processor for performing a high-level arbitrary optical filtering process in the fields of optical communication, optical switching, and optical computing.

[0002]

2. Description of the Related Art In recent years, in the fields of optical communication, optical switching, and optical computing, optical signal processing capable of performing a wideband and high-speed filtering process without converting an optical signal into an electric signal without changing the light. The vessel is attracting attention. In particular,
In an optical frequency multiplexing communication in which an optical signal is multiplexed and transmitted, an optical frequency filter that performs a filtering process for each frequency with respect to a propagated frequency multiplexed signal light is an important component.

Conventionally, as an optical signal processor for such a purpose, a transversal type optical signal processor shown in FIG. 8 (JP-A-2-212822) has been reported. In this example, variable directional couplers 103-1 to 103-N are arranged on the input port 101 and 102 sides, and optical signals are split into respective branch waveguides 104-1 to 104-N at an arbitrary branching ratio.
And N-1 3 dB directional couplers 105-1 to 105-1.
A configuration in which each of the branched optical signals is converged again by the coupling unit composed of 105- (N-1) is adopted. A phase controller 1 for individually controlling the phases of the branched optical signals is provided on each of the branch waveguides 104-1 to 104-N.
06-1 to 106-N. The 3 dB directional couplers 105-1 to 105- (N-1) have dummy output ports 107-1 to 107- (N-1), respectively, and the last 3 dB directional coupler 105- (N- The other output port of 1) is the output port 108. In this example, the coupling ratios of the variable directional couplers 103-1 to 103-N are changed, and at the same time, the phase controllers 106-1 to 106-N
By changing the amount of the phase shift, desired transmission characteristics can be realized.

[0004]

However, in an optical circuit that does not include a light return path such as a ring portion or a grating, including the above-described conventional example, it is necessary to obtain a sharp edge at the band edge of the filter characteristic. However, there is a problem that the number of circuit stages must be increased.

The present invention has been made in view of the above-mentioned prior art, and has as its object to firstly realize a filter characteristic having a desired sharpness with a small number of stages.
Second, it is an object of the present invention to provide an optical signal processor that achieves a programmable characteristic that realizes an arbitrary filter characteristic with one circuit.

[0006]

In order to achieve the above object, a first aspect of the present invention is to provide two first and second optical waveguides, and these two first and second optical waveguides. (N +
1) A first directional coupler group having (N + 1) positive or negative amplitude coupling ratios that divides the first and second optical waveguides into N regions by coupling at different positions. When,
A second group of directional couplers having N positive or negative amplitude coupling rates respectively interposed in any one of the first and second optical waveguides of the N regions. A ring optical waveguide group respectively coupled to one of the first and second optical waveguides of the N regions via the N second directional coupler groups; A first phase controller group that is interposed in at least one of the first and second optical waveguides in the region of each of the regions to perform a desired phase shift, and a desired phase that is interposed in each of the ring optical waveguide groups. A second phase controller group for performing a shift, wherein the optical path lengths of the first optical waveguide and the second optical waveguide in the N regions are equal to each other, and the optical path of the ring optical waveguide group is provided. An optical signal processor characterized in that all lengths are equal.

According to a second aspect of the present invention, in the first aspect, all or a part of the first directional coupler group and the second directional coupler group has an amplitude coupling ratio. An optical signal processor characterized by being a variable directional coupler that can be changed to a positive or negative value.

According to a third aspect of the present invention, in the second aspect, the variable directional coupler comprises two fixed coupling rate directional couplers having positive or negative amplitude coupling rates. , Two optical waveguides sandwiched between the two fixed coupling ratio directional couplers, wherein the two optical waveguides have the same optical path length or are shifted by a half wavelength, and An optical signal processor having a Mach-Zehnder interferometer configuration provided with a phase controller for performing a desired phase shift on at least one of the wave paths.

According to a fourth aspect of the present invention, in any one of the first to third aspects, the amplitude coupling ratio θ _1n (−π / 2 to π) of the N + 1 first directional coupler groups is provided. / 2) and the phase shift amount φ _1n (−
π to π), and the N second directional couplers
The amplitude coupling ratio θ _2n (−π / _{2 to} π / 2) and the phase shift amount φ _2n (−π to π) of the N second phase controllers.
Is calculated and set based on the following recurrence formula.

[0010]

(Equation 3)

However, the recurrence equation is solved in order from n = N to n = 0, and at the first stage (n = N) of the recurrence equation, h ^[N] (z) = h (z), f ^[N] = (z) f
(Z), g ^[N] (z) = g (z).

Here, h (z) / g (z) represents the desired characteristic by a through port, and f (z) / g (z) represents the following when the desired characteristic is realized by a cross port. In a polynomial, the unitary relation h (z) h _* (z) + f (z) f _*
(Z) = g (z) It is obtained from g _* (z). Also,

[0013]

[Outside 2]

## EQU1 ## j represents √ (-1). Δτ is a delay time corresponding to a constant optical path length difference of the link optical waveguide. Γ _n−1 and γ _n are (n) of g (z).
-1) represents the n-th and n-th zeros. Further, M is any
This is a natural number, which is the number of zeros of g (z).

[0015]

(Equation 4)

[0016]

According to the optical signal processor of the present invention, a filter characteristic having a sharp band edge can be realized by the feedback action due to the presence of the ring optical waveguide. Further, the calculation formulas described in the above-described “Equation 3” and “Equation 4” provide a method of obtaining the coupling ratio of the directional coupler and the phase shift amount of the phase controller for an arbitrary desired filter characteristic. Thus, an optical signal processor having arbitrary filter characteristics can be realized.

Further, by making the coupling ratio of some or all of the directional couplers variable in a positive or negative range,
Programmability of realizing an arbitrary filter characteristic with one circuit can be achieved. Here, "coupling rate"
Means “coupling rate of amplitude”.

[0018]

The present invention will be described below in detail with reference to examples.

(Embodiment 1) The optical signal processor of this embodiment is
It has the circuit configuration shown in FIG. Also, the coupling ratio of the directional coupler and the phase shift amount of the phase controller, which are unknown circuit parameters, were obtained by using the above-mentioned equations (3) and (4).

As shown in FIG. 1, the first and second optical waveguides 1 and 2 are composed of N first directional coupler groups 3- (N + 1) having a positive or negative coupling ratio.
1 to 3- (N + 1) and divided into N first optical waveguides (arms) 1-1 to 1-N and second optical waveguides (arms) 2-1 to 2-N, respectively. Have been. To each of the N first arms 1-1 to 1-N, via N second directional coupler groups 4-1 to 4-N having a positive or negative coupling rate, The ring optical waveguide groups 5-1 to 5-N are respectively coupled one by one. A first phase controller group 6-1 for performing a desired phase shift is provided to each of the N second arms 2-1 to 2-N.
6-N are interposed one by one. further,
Each of the ring waveguide groups 5-1 to 5-N is provided with one second phase controller group 7-1 to 7-N for performing a desired phase shift.

The optical signal processor having such a configuration is composed of N-stage regions 8-1 to 8-N surrounded by a dashed line.
And one end of second optical waveguides 1 and 2 is connected to input port 1
A and 2A, the other ends of which are output ports 1B and 2B.
It is.

Here, the optical path lengths of the first arm 1-1 to 1-N and the second arm 2-1 to 2-N in each of the N-stage regions 8-1 to 8-N are equal. , N ring waveguide groups 5-1 to 5-N have the same optical path length.

In this embodiment, the first phase controller groups 6-1 to 6-N are provided to the second arms 2-1 to 2-N of the N-stage regions 8-1 to 8-N. Although interposed, the first phase controller group may be provided in the first arms 1-1 to 1-N, or may be provided in both. Also, the first arms 1-1 to 1-1
1-N, the second directional couplers 4-1 to 4-1
It may be provided either before or after 4-N or both.
The advantage of providing the first group of phase controllers in both the first and second optical waveguides is that if only one of them is used, it is necessary to change the phase amount in the range of 0 to 2π to control the phase. On the other hand, when both are provided, the amount of phase change required for each phase controller can be suppressed to a half of 0 to π.

In this embodiment, an optical signal processor was manufactured using a planar circuit excellent in mass productivity. Each ring waveguide group 5-
The optical path length of 1 to 5-N is 20 GHz at a wavelength of 1.55 μm.
10.37 mm corresponding to the frequency interval of

FIG. 2 shows a cross section taken along the line AA 'of the optical signal processor. In this example, a silicon substrate was used as the substrate 21, the waveguide was formed by a quartz thin film 22 deposited on the substrate by a flame deposition method, and the waveguide structure was manufactured by a photolithography process. Here, the core size of the manufactured waveguide was determined so as to be a single mode. The optical filter of the present invention realizes filter characteristics by utilizing light interference, branching, and coupling, and the presence of higher-order modes inside the waveguide causes performance degradation. For this,
In this embodiment, a single mode waveguide is used in all cases.
The phase controller was manufactured by forming a heater composed of a chromium layer on a quartz thin film. This phase controller controls the phase by utilizing the thermo-optic effect that the refractive index changes when heated.

In this embodiment, the entire optical circuit is mounted on a Peltier element and the temperature of the optical circuit is controlled with an accuracy of about 0.1 degree in order to eliminate the influence of environmental temperature fluctuation.

Since the optical signal processor of the present invention utilizes the coherent interference phenomenon of light, the signal light must be coherent light.

In FIG. 1, the signal light incident from the input port 1A is divided into unit components 8-1 to 8-1 arranged in multiple stages.
The light passes through 8-N and is output to output ports 1B and 2B. At this time, the signal light is split into respective optical waveguides by the respective directional couplers. For example, the signal light traveling in the shortest optical path is the arms 1-1, 1-2,... Of the first optical waveguide.
1-N, or arms 2-1 and 2- of the second optical waveguide
2, ..., 2-N. The signal light traveling the longest optical path goes around the ring waveguides 5-1 to 5-N infinitely. Output light from the ports 1B and 2B is expressed as a sum of signal lights passing through various optical paths. Since the present optical signal processor has a ring waveguide, there is infinite orbital light. Therefore, the complex amplitude of the output light from port 1B is
As represented by equation (1), the denominator and the numerator are each expressed in the form of a z-polynomial.

[0029]

(Equation 5)

Here, a _k and c _k in the equation (1) are complex expansion coefficients. β represents the propagation constant of the optical waveguide, and ω ₀ represents the period of the frequency characteristic. In addition, in Expression (1), as often used in the electric digital filter theory, the following (a) is considered as a complex number z, and the following (b) is extended to a function h (z) on a complex plane. . On this z complex plane,
The frequency function h (ω) is regarded as a function on the circumference of the radius 1 with the absolute value of z = 1. H (z) of numerator, g (z) of denominator
The degree of both polynomials is N. g (z) is a polynomial of degree N and has N zeros, among which NM are z
There is a zero at = 0.

[0031]

[Outside 3]

In the equation (1), the through characteristic h (z) /
g (z) is represented by a polynomial of z. This equation is used in the field of digital filters in IIR (Infini
It can be seen that the transmission characteristic is equal to the transmission characteristic of a digital filter of a type called feedback (teImpulse Response). This fact plays an important role in the synthesis of the optical filter described below.

Similarly, light enters from port 1A and
Output light f (z) / g (z) emitted from B is given by equation (2)
It is represented as

[0034]

(Equation 6)

Here, the order of f (z) is also h (z), g
N as in (z).

Since the present signal processor has two inputs and two outputs, the transmission matrix S of the entire signal processor is represented by a 2 × 2 matrix expressed by equation (3).

[0037]

(Equation 7)

Here, the suffix * represents paraconjugate (h _* (z) = h ^* (1 / z ^* )). Assuming that the circuit is lossless, h (z) / g (z) and f (z) / g
A unitary relationship represented by the following equation (4) is established between (z).

[0039]

(Equation 8)

In this embodiment, the cross characteristic f (z) / g
The filter was designed so that (z) had the desired characteristics.

Specifically, circuit parameters for satisfying the desired filter characteristics with respect to the cross characteristics (the transmission characteristics from the input port 1A to the output port 2B) were determined by the following procedure (FIG. 5).

[Design Procedure 1] First, in step S1, a polynomial of z which is a denominator of the equation (2) is obtained so as to satisfy the desired filter characteristic. At this time, the fact that the transmission characteristic of the optical filter described in equation (2) is the same as that of a digital filter called IIR type is used. Specifically, various approximation techniques developed in the field of digital filters are used. For example, bilinear z-transform method (bilinear exchange method),
Least squares approximation, Fletche-Powell method and the like are well known (Openpenim & Scaf)
er, "Digital Signal Processing", Prentice-Ha
11 Inc. 1975). At the time of this calculation, the required expansion order is obtained from the required conditions of the desired filter characteristics. This expansion order corresponds to the number of stages N of the unit circuit configuration. When calculating the approximate polynomial, care must be taken that the polynomial f (z) / g (z) representing the cross characteristic satisfies the causality. Specifically, a polynomial f representing the cross characteristic
The pole of (z) / g (z), that is, the zero of g (z) needs to be within the circle of the absolute value of z = 1. Since the optical signal processor of the present invention is a passive circuit, it is essential to satisfy the causality.

[Design Procedure 2] In the design procedure 1, the z-polynomial of the equation (2) was obtained by using the above-described optimal digital filter technique so as to realize desired filter characteristics. If an optical signal processor is realized using the obtained polynomial as it is, the absolute value of the amplitude characteristic may exceed 100%.
However, in this signal processor, which is an analog signal processor using a physical quantity called a light wave, due to the law of conservation of energy,
The absolute value of the complex amplitude in equation (2) cannot essentially exceed 100%. For this reason, in step S2, the polynomial obtained by the digital filter approximation method is normalized by equation (5) so that the absolute value of the amplitude characteristic does not exceed 100%.

[0044]

(Equation 9)

[Design Procedure 3] In step S3, the complex amplitude characteristic having the normalized complex expansion coefficient obtained in the design procedure 2 is cross-outputted from the optical signal processor (from port 1A to port 2A).
(Output to B), an S matrix representing the transmission characteristics of the present optical signal processor is obtained. Specifically, the remaining unknown transfer function h (z) of the S matrix is obtained from the unitary relation of Expression (4). From the unitary relation, h (z) h _* (z) is a polynomial f (z) representing the cross characteristic standardized in the design procedure 2.
By using / g (z) and further expanding z into the form of a product of a function linear expression, it is expressed by Expression (6).

[0046]

(Equation 10)

In this equation (6), a ₀ is obtained as in equation (7).

[0048]

[Equation 11]

Here, γ _N is the N-th zero of g (z).

In equation (6), h (z) and h _* (z)
From the necessity expressed in the form of
The zeros of (z) are obtained as N pairs {α _k , 1 / α _k ^* }. When calculating h (z), the polynomial of h (z) must have one of the root pairs of {α _k , 1 / α _k ^* } as the root. According to the way, there are 2 ^N ways of taking h (z). Also,
g the way of the M of the zero-point selection is not a 0 (z) is _N P _M
(= (N!) / ( N-M)!) Number there, for each, h (z) is because it is required, in the whole, h (z) is (2 _^N · ^N P _M) number of solutions have. These (2 _^N · ^N P
_{The M} ) h (z) solutions have equal amplitude characteristics and different phase characteristics. F (z) is common and h
(Z) are different S-matrix is (2 _^N · ^N P _M) pieces exist, as a result, (2 _^N · ^N P _M) types of circuit parameters is obtained. However, even when making the (2 _^N · ^N P _M) pieces circuit in any circuit parameters, the cross characteristics identical characteristics. In this embodiment, the cross characteristic f (z)
Since interested in the characteristics of the / g (z), (2 N · N P M)
An appropriate one of the h (z) solutions was selected. Needless to say, h (z) can be selected by appropriately setting a selection criterion such as one having low element sensitivity and one having an appropriate phase characteristic.

[Design Procedure 4] In step S4, in order to obtain circuit parameters, the S matrix representing the transmission characteristics of the optical signal processing circuit having the N-stage configuration is replaced by the product of the S matrix representing the transmission characteristics representing the unit configuration. Expand to

The whole S matrix is decomposed as shown in equation (8).

[0053]

(Equation 12)

Here, S _n represents the transmission matrix of the unit configuration circuit of the _n- th stage.

The unit configuration circuit of the n-th stage shown in FIG.
n, 2-n optical waveguides (arms), 5-n ring waveguides, 3- (n + 1), 4-n two directional couplers, 6-n, 7-n And the unit configuration circuit of n = 0 is one directional coupler 3-1.
It is composed of

[0056] degradation of the actual S matrix, the inverse matrix S of the n-th stage of the unit configuration transmission matrix S _n by steps S4 and S5
^{_{n -1 = S n + (S}} n + is transposed conjugate matrix of S _n) a n = n, N
.., 2, 1, 0 in the order of S matrix.

[0057]

(Equation 13)

[0058] In this _{^{case, S [n] = S n}} S n-1 ... S 2 S 1 S
Represents ₀ . Each _Sn is obtained by this repetitive operation.

More specifically, the angle characteristic θ _n2 of the coupling ratio of the directional coupler connecting the two optical waveguides and the ring waveguide and the phase shift φ _n2 of the phase controller on the ring waveguide have a cross characteristic. N-th zero γ _r of the denominator polynomial g (z) of the polynomial
It is obtained from the following equation.

[0060]

[Equation 14]

Assuming that S ^[n] is known in equation (9), the highest order of the numerator polynomial of the element of the product of the matrix on the right side of equation (9) is (n + 1) / 2, but the equation (9) The highest degree of the numerator polynomial of the element of the matrix on the left side of is (n-1) / 2. In order for the equations on the left and right sides to be satisfied, the highest order on the right side must be matched with the highest order on the left side. From these conditions, the angle display θ _n1 of the coupling ratio of the directional coupler connecting the two optical waveguides is obtained. Also, [f ^[n] (γ _n )] / [h
^[n] (γ _n )] is always a real number (here, it is positive), and the phase shift amount φ _n1 of the phase controller provided on the two waveguides is obtained.

[0062]

(Equation 15)

In the equation (11), h ^[n] (z), f ^[n]
(Z) is a numerator polynomial of an element of the n-th unit transmission matrix S ^[n] . γ _n and γ _n-1 are the n-th and n−1-th zeros of g (z), respectively.

As described above, by completing the S matrix decomposition procedure up to the final stage, all circuit parameters are obtained. That is, the following recurrence formula is expressed as n = N, N-1,.
If the solution is performed in the order of 2, 1, 0, all circuit parameters can be obtained. Here, the initial value (n = N) of the recurrence equation is h ^[N] (z) = h (z), f ^[N] obtained in the design procedure 3 ^.
(Z) = f (z), g (z).

[0065]

(Equation 16)

[0066] As described above, in the present invention, the circuit parameters of the (2 _^N · ^N P _M) as to the desired cross filter characteristics are determined (step S6). Regardless of which of these circuit parameters is employed, the same cross filter characteristics can be obtained. From this, it is clear that the present invention is not limited to numerical values relating to circuit parameters, and the present invention extends to all circuit parameters obtained by the above-described circuit synthesis method.

In this embodiment, the circuit parameters are calculated so that the desired filter characteristics can be realized by the cross output.
It is also possible to calculate circuit parameters so as to realize desired filter characteristics with through outputs (outputs from ports 1A to 1B).

The result of the actual design of a Butterworth filter using the above-described design method will be described below.

Optical Filter Having Butterworth Characteristics Butterworth optical filters have the characteristic that the amplitude frequency characteristics are maximally flat in the transmission band. This means
In the Nth-order filter, the zeroth to Nth square amplitude frequency characteristics
This means that the derivatives up to the first order are zero at the center frequency. In this embodiment, a Butterworth type transmission characteristic is realized in the cross characteristic. The required characteristic values of the optical filter are as follows: frequency interval: 20 GHz, transmission band frequency: -2 GH
The amplitude frequency characteristic of the transmission band from z to 2 GHz is constant within 1 dB, and the stop band characteristic is 15 GHz between 3 GHz and 10 GHz and between (-3 GHz) and (-10 GHz).
It was set to dB.

Here, first, a characteristic function of an analog Butterworth filter satisfying the above requirements is obtained, and the characteristic function is obtained by a bilinear conversion method (“Digital Signal Processing” by Oppenheim & Scaffer, “Prentice-Hall”)
Inc. 1975), a sixth-order polynomial approximation formula shown in Expression (13) was obtained.

[0071]

[Equation 17]

FIGS. 6A and 6B show the transmission amplitude of a 6-stage Butterworth type optical filter (N = 6) manufactured based on the circuit parameters obtained through the design procedures 1 to 4 of this embodiment. The result. Transmission amplitude 0.5 at ± 2 GHz
It achieved a rejection of 15 dB at ± 3 GHz.

(Embodiment 2) This embodiment is also implemented to realize the first object.

FIG. 7 shows a circuit configuration of this embodiment. In addition,
1 are denoted by the same reference numerals, and duplicate description will be omitted.

In the first embodiment, all the ring waveguides are on the optical waveguide 1 side. In practice, 2 ^N waveguide arrangements can be considered depending on whether the ring waveguide is on one side or two sides of the two optical waveguides. In the present embodiment, an embodiment will be described in which a Butterworth filter similar to that of the first embodiment is manufactured using the following two types of arrangements out of 2 ^N waveguide arrangements.

[Waveguide arrangement 1] is an arrangement in which the ring waveguides alternately appear on the waveguide 1 side and the waveguide 2 side of the two optical waveguides.

[Waveguide arrangement 2] is based on k =
From 0 to n, there is a ring waveguide on the waveguide 1 side, and from k = n + 1 to N in the unit configuration circuit, there is a ring waveguide on the waveguide 2 side.

The circuit parameters for these waveguide arrangements can be easily obtained from the calculation results of the circuit parameters obtained in the first embodiment. This can be easily performed by the following procedure.

In [Procedure 2], similarly to the first embodiment, circuit parameters for realizing a Butterworth filter are obtained when all the ring waveguides are on the waveguide 1 side.

[Procedure 2] is that the waveguides 1 and 2 where the ring waveguide is on the waveguide 2 side in the desired waveguide arrangement
The coupling ratios of the directional couplers at both ends of the symmetric Mach-Zehnder interferometer composed of are reduced by π. This operation is repeated for all the symmetric Mach-Zehnder interferometers whose ring waveguides are on the waveguide 2 side.

A filter circuit was manufactured for each waveguide arrangement, and the filter characteristics were measured. As a result, the same filter characteristics as those in Example 1 were obtained.

As can be seen from this embodiment, in the present invention, for all 2 ^N types of waveguide arrangements which differ depending on whether the ring waveguide is taken on the waveguide 1 side or the waveguide 2 side,
Equivalent circuit parameters can be designed. From this, it is clear that the present invention includes all circuit configurations optically equivalent to FIG.

(Embodiment 3) This embodiment is an example in which a configuration for realizing the second object is implemented.

The variable directional coupler used in this embodiment is shown in FIG.
Shown in The advantage of using a variable directional coupler is that
One object is to obtain programmability in which arbitrary circuit characteristics can be realized by one circuit. The variable directional coupler of FIG.
A symmetric Mach-Zehnder interferometer composed of two optical waveguides 11 and 12 having the same optical path length and 3 dB directional couplers 13-1 and 13-2 disposed at both ends thereof is configured. On each optical waveguide, phase controllers 14-1 and 14-2 made of a chrome heater are arranged. Even if this phase controller is installed on only one waveguide, a variable directional coupler is formed. The variable directional coupler can change the coupling ratio from 100% to 0% by changing the phase difference between the two optical waveguides from 0 to π by the phase controller. Specifically, the phase controllers 14A,
When the phase shift amount of 14B is 0, it becomes a directional coupler with 100% coupling, and when the phase shift amount is π, it becomes a directional coupler with 0% coupling. Although not adopted here, when the phase shift amount of the phase controller is 0, the optical path lengths of the two waveguides may be shifted in advance by a half wavelength so that the directional coupler has 0% coupling. It is possible. Although not adopted here, the 3 dB directional coupler 1 forming the variable directional coupler 1
It is also possible to replace 3-1 and 13-2 with a variable directional coupler having a symmetric Mach-Zehnder interferometer configuration shown in FIG. Variable directional coupler is 0-100 as designed
To be completely variable up to% coupling, the 3d
The coupling rate of the B coupler needs to be 3 dB with considerable accuracy. In order to accurately set the coupling ratio of the 3 dB coupler to 3 dB, it is effective to replace the 3 dB coupler with a variable directional coupler. Further, the 3 dB directional couplers 13-1 and 13-2 constituting the variable directional couplers
Is converted to a wavelength-independent directional coupler (Electon) having an asymmetric Mach-Zehnder interferometer having an optical path length difference on the order of wavelength.
ics Letters; “Mach-Zehnde
r Interfermenter type op
tical waveguide couler wi
th wavelength-flattened c
coupling ratio ", Vol. 26, No.
17, pp. 1326-1327, 1990). In this way, the manufactured optical signal processor can operate irrespective of the wavelength.

Also in this embodiment, the same 6 as in Embodiment 1 is used.
A Butterworth filter having two stages was manufactured. It was also confirmed that a Chebyshev filter could be realized with the same circuit. This embodiment is characterized in that it is more resistant to manufacturing errors than the first embodiment. This is because error correction is possible because a variable directional coupler is used.

The configuration and the design method of the present invention have been described using the embodiments, but the present invention is not limited to the embodiments. For example, although a quartz planar optical circuit is used in the present invention, a semiconductor such as an InGaAsP-based semiconductor, LiNbO ₃
It is also possible to form a planar circuit with another material such as an electro-optical material such as described above and an organic optical material. It is also possible to use an optical fiber instead of a planar circuit. In the present embodiment, the thermo-optical effect and the electro-optical effect were used to perform the phase control. However, other physical phenomena, such as the nonlinear optical effect represented by the Kerr effect,
It is also possible to use a magneto-optical effect or the like. As described above, the present invention relates to a circuit composition technique for calculating a circuit parameter by giving a circuit configuration representing a combination of elements and a desired filter characteristic, and is not restricted by a physical realizing means of the circuit.

In the present invention, as shown in the second embodiment, there are a number of optically equivalent circuit configurations. However, all of them can be regarded as the same, so that a circuit which is optically equivalent to the circuit configuration of the present invention. All configurations are the subject of the present invention. For example, 2
When the coupling ratio of the directional coupler connecting the optical waveguide and the ring optical waveguide is 0, the existence of the ring optical waveguide becomes unnecessary, and therefore, it is conceivable to use a circuit configuration in which unnecessary ring optical waveguides are removed in advance. . Also in this case, setting the coupling ratio of the directional coupler to 0 and removing the corresponding ring optical waveguide in advance are optically equivalent, and can be considered to be the same circuit configuration.

A feature of the claims of the present invention resides in that a design procedure for obtaining circuit parameters is required in addition to the circuit configuration. In order to actually achieve the desired characteristics, a formula for calculating the circuit parameters for realizing the desired characteristics is required. It is actually impossible to determine circuit parameters by trial and error,
Even if an attempt is made to obtain circuit parameters by a successive approximation technique using a computer, it is almost impossible to obtain optimal circuit parameters. The design method sought in the present invention is unique to the present invention, and is characterized in that, unlike the successive approximation method, an optimum circuit parameter can be obtained with a finite number of calculations. Even if a patent having a circuit configuration similar to the present invention exists, the present invention claims differentiation from other patents due to the existence of a unique design method of the present invention. Specifically, the frequency filter described in the embodiment is a characteristic that cannot be obtained without using this design method, and even when the circuit configuration is similar, by comparing the embodiment, it is possible to discriminate from other patents Can be performed.

[0089]

As described above, the optical signal processor of the present invention employs the above-described circuit configuration and design method to perform an arbitrary filtering process at a maximum transmittance of 100%, which could not be realized conventionally. Feasible. In particular, the present invention is characterized in that by providing a ring waveguide, a steep filter characteristic can be realized with a small number of stages using the same feedback effect as a thin film filter. In addition, by using a variable directional coupler that freely describes a coupling ratio and a phase controller that can give an arbitrary phase shift as circuit components, it is possible to realize an arbitrary filter characteristic with the same circuit. It also has the feature.

The optical signal processor of the present invention having such excellent characteristics provides a high-performance optical signal processor in the field of advanced information processing such as frequency multiplexing communication requiring high speed and wide bandwidth. Is possible.

[Brief description of the drawings]

FIG. 1 is a diagram showing a circuit configuration of an optical signal processor according to a first embodiment of the present invention.

FIG. 2 is a sectional view taken along line AA ′ of the optical signal processor according to the first embodiment.

FIG. 3 is a diagram illustrating a circuit configuration of a variable directional coupler used in the first embodiment.

FIG. 4 is a diagram showing an n-th unit configuration circuit of the circuit configuration of the first embodiment.

FIG. 5 is a flowchart illustrating a design procedure according to the first embodiment.

FIG. 6 is a diagram illustrating a Butterworth type transmission characteristic of the optical frequency filter realized by the circuit configuration of the first embodiment.

FIG. 7 is a diagram illustrating two types of waveguide configurations according to a second embodiment.

FIG. 8 is a diagram showing a circuit configuration of a transversal optical signal processor according to a conventional example.

[Explanation of symbols]

1, 2 Optical waveguide 1-1, 1-N, 2-1 to 2-N Arm 1A, 2A Input port 1B, 2B Output port 3-1, 3- (N + 1), 4-1 to 4-N Positive or Fixed (variable) directional coupler having a negative coupling ratio 5-1 to 5-N Ring waveguide 6-1 to 6-N, 7-1 to 7-N Phase controller

──────────────────────────────────────────────────続 き Continuation of the front page (56) References Kaname Jinguji, Masao Kawauchi, Synthesis of complex-coefficient optical FIR filter with multi-stage Mach-Zehnder configuration, IEICE Autumn Meeting 1993, SC-2-4, Chapter 4- Pages 381 to 382 (58) Fields investigated (Int.Cl. ⁷ , DB name) G02F 1/00-1/125 G02F 1/29-7/00

Claims (4)

(57) [Claims] 1. Two first and second optical waveguides, and the two first and second optical waveguides are coupled at (N + 1) different positions to form the first and second optical waveguides. Divide the wave path into N regions (N + 1) positive or negative amplitudes
A first directional coupler group having a coupling ratio of a width, and N positive or negative amplitudes respectively interposed in one of the first and second optical waveguides of the N regions . A second directional coupler group having a coupling ratio, and one of the first and second optical waveguides of the N regions via the N second directional coupler groups, respectively. A ring optical waveguide group to be coupled, and the first and second ring optical waveguides in the N regions.
A first phase controller group interposed in at least one of the optical waveguides and providing a desired phase shift, and a second phase controller interposed in each of the ring optical waveguide groups and performing a desired phase shift Wherein the first optical waveguides and the second optical waveguides in the N regions have the same optical path length, and the ring optical waveguide groups all have the same optical path length. Optical signal processor. 2. The method according to claim 1, wherein all or a part of the first directional coupler group and the second directional coupler group change an amplitude coupling ratio to a positive or negative value. An optical signal processor characterized in that it is a variable directional coupler capable of performing the following. 3. The directional coupler according to claim 2, wherein the variable directional couplers have two fixed coupling ratio directional couplers having a coupling ratio of positive or negative amplitude, and the two fixed coupling ratio directional couplers. And two optical waveguides having the same optical path length or being shifted by a half wavelength, and providing a desired phase shift to at least one of the two optical waveguides. An optical signal processor comprising a Mach-Zehnder interferometer provided with a controller. 4. The coupling ratio θ of the amplitudes of the (N + 1) first directional coupler groups according to claim 1,
_{1n (-π} / _2~π / 2) and said N first phase shift amount phi _1n the phase control unit group (-π~π), and the N second directional coupler group of the binding rate of amplitude θ _2n (-
π / 2 to π / 2) and the phase shift amounts φ _2n (−π to π) of the N second phase controllers are calculated and set based on the following recurrence formula. Optical signal processor. (Equation 1) Here, the recurrence equation is solved in the order of n = N to n = 0, and in the first stage (n = N) of the recurrence equation, h ^[N]
(Z) = h (z), f ^[N] (z) = f (z), g
^[N] (z) = g (z). Here, h (z) /
g (z) is a through port for desired characteristics, and f (z) / g
(Z) is a polynomial expressed by the following when a desired characteristic is to be realized by a cross port, and a unitary relation h (z) h
_* (Z) + f (z) f _* (z) = g (z) g _*
(Z). Z is expressed by the complex angular frequency ω. Is a complex variable represented by j represents √ (-1). Δ
τ is a delay time corresponding to a constant optical path length difference of the link optical waveguide. Γ _n−1 and γ _n are g (z)
(N-1) -th and n-th zeros of Furthermore, M is
An arbitrary natural number, which is the number of zeros of g (z). (Equation 2)