WO2019167620A1 - Optical arithmetic unit - Google Patents

Abstract

Provided is an optical arithmetic unit having high speed and a small circuit scale. The optical arithmetic unit is provided with N cascade-connected two-input-one-output Y-junction elements 1-1 through 1-N and N light intensity modulators 5-1 through 5-N. The N light intensity modulators 5-1 through 5-N each individually modulate the intensity of continuous light to a second optical input port different from a first optical input port to which no input light is inputted or to which a signal light from an optical output port of a prior-stage Y-junction element is inputted from among two optical input ports of the cascade-connected N Y-junction elements 1-1 through 1-N, in accordance with a bit corresponding to an N-bit electrical digital signal, and generates a signal light to the second optical input port. Output light obtained from the final-stage Y-junction element 1-N constitutes the result of an N-bit digital/analog operation.

Description

Optical computing unit

The present invention relates to an optical computing unit such as an optical digital-to-analog converter (DAC) using an optical circuit.

A variety of digital-to-analog converters (DACs) that use electrical circuits have been proposed, and performances such as sampling rate, resolution, power consumption, and size are different. In other words, the current situation is that different types of DACs are used depending on the application target. For example, in a current commercial product, the sampling rate is about 1 GS / s, and the resolution is 16 bits or more. With future development of communication or video technology, higher speed, higher resolution, lower power consumption, and smaller DACs are expected. In addition, the latency (delay) required for a high-resolution DAC is expected to become a future bottleneck.

As a DAC that can meet the above requirements, an optical DAC using an optical circuit has been proposed (see Non-Patent Document 1).
However, since some conventional optical DACs are operated by an electric circuit, there is a problem that the speed is limited by the electric circuit, and there is a problem that an enormous element and circuit scale are required. It was.

The present invention has been made to solve the above-described problems, and an object thereof is to provide an optical computing unit that can be mounted at high speed and high density.

The optical computing unit of the present invention includes a 2-input 1-output first Y merging element that receives 1 to 2 signal lights as input, and a 2-input 2-output Y merging / Y that receives 1 to 2 signal lights as inputs. At least one of the branch elements is cascade-connected (N is an integer equal to or greater than 2), and the first of the two optical input ports of each of the N elements connected in cascade has no input light. The continuous light to the second optical input port different from the first optical input port to which the signal light from the optical input port or the optical output port of the previous stage element is input is a corresponding bit of the N-bit electric digital signal, respectively. And N optical modulators that individually modulate in response to the signal and generate the signal light to the second optical input port, and output light obtained from the last-stage element is used as a calculation result. To do.

In addition, one configuration example of the optical computing unit of the present invention includes N first Y merge elements, and the N first Y merge elements exclude the first Y merge element at the most upstream. (N-1) first Y confluence elements use the light output from the light output port of the upstream first Y confluence element as input light to the first light input port, and the most upstream N first Y merging elements including the first Y merging elements are cascaded so that the signal light modulated by the N optical modulators is input to the second optical input port. The N optical modulators individually modulate the N light beams having the same wavelength according to the corresponding bits of the N-bit electric digital signal, respectively. The signal light to the second optical input port is generated and the output light obtained from the first Y junction element at the final stage is N bits. Is characterized in that the result of digital-analog operation.

Also, one configuration example of the optical computing unit of the present invention includes (N−1) Y-merging / Y-branching elements and one first Y-merging element, and these (N−1) elements. The Y merge / Y branch element and one first Y merge element are the (N-2) Y merge / Y branch elements and one first Y merge element excluding the most upstream Y merge / Y branch element. The Y merge element of the upstream is the light output from the first optical output port of the upstream Y merge / Y branch element as the input light to the first optical input port, and the most upstream Y merge / Y branch (N-1) Y combining / Y branching elements including one element and one first Y combining element receive the signal light modulated by the N optical modulators in the second optical input port. The N optical modulators respectively connect N continuous lights of the same wavelength to N-bit electric digital signals. Individually modulated according to the bit to the second optical input port of the (N-1) Y merging / Y branching elements and the second optical input port of the one first Y merging element A second optical output port that generates signal light and is different from the first optical output port that outputs the signal light to the subsequent element among the two optical output ports of the (N−1) Y merging / Y branching elements. The output light obtained from the optical output port and the output light obtained from the first Y-merging element at the final stage are used as the result of 1 to N-bit digital / analog computation.

In addition, one configuration example of the optical computing unit of the present invention includes N Y merging / Y branching elements, and these N Y merging / Y branching elements exclude the most upstream Y merging / Y branching element. (N-1) Y merging / Y branching elements use the light output from the first optical output port of the upstream Y merging / Y branching element as the input light to the first optical input port, and The N Y merge / Y branch elements including the most upstream Y merge / Y branch element use the signal light modulated by the N optical modulators as input light to the second optical input port. The N optical modulators individually modulate N continuous light beams having the same wavelength in accordance with corresponding bits of the N-bit electric digital signal, respectively. Generates signal light to the second optical input port of the Y-branch element, and the cascaded 2-input 1-output (N− ) Second Y-merging elements, and the first optical output port for outputting the signal light to the Y-merging / Y-branching element at the subsequent stage among the two optical output ports of the Y-merging / Y-branching element at the first stage. The signal light obtained from the second optical output port different from the optical output port is input to the first optical input port of the second Y junction element at the first stage, and the (k−1) -th stage (k is 2 to 2). (The integer of (N-1)), the signal light obtained from the optical output port of the second Y-merging element is input to the first optical input port of the k-th second Y-merging element, j + 1) signal light obtained from the second optical output port of the Y-combining / Y-branching element at the j-th stage (j is an integer from 1 to (N-1)) Output light obtained from the first optical output port of the Y-merging / Y-branching element at the final stage is input to the second optical input port of the N-bit digital - as a result of analog operation, an output light obtained from the second Y-combining device in the final stage is characterized in that the result of the N-bit counter operation.

In addition, one configuration example of the optical computing unit of the present invention includes N first Y merge elements, and the N first Y merge elements exclude the first Y merge element at the most upstream. (N-1) first Y confluence elements use the light output from the light output port of the upstream first Y confluence element as input light to the first light input port, and the most upstream N first Y merging elements including the first Y merging elements are cascaded so that the signal light modulated by the N optical modulators is input to the second optical input port. Furthermore, each of the N Y branch elements, except for the most upstream Y branch element that receives a single continuous light, is provided with N Y branch elements having one input and two outputs. Of the two optical output ports of the upstream Y branch element, the light output from the first optical output port is connected in cascade to be input, The output light obtained from the second optical output port of these N Y branch elements is used as the input light to the N optical modulators, and the output obtained from the first Y confluence element in the final stage. The light is the result of N-bit digital / analog operation.

In one configuration example of the optical computing unit of the present invention, at least one of the first Y merging elements or at least one of the Y merging elements constituting the Y merging / Y branching element is: In addition to the first and second light input ports, a bias port to which a fixed intensity bias light is input is provided.
Further, in one configuration example of the optical computing unit of the present invention, the output light obtained from the last stage element is used as the input light to the first optical input port, and the fixed intensity bias light is used as the second light. A Y-merging element serving as input light to the input port is provided, and output light obtained from the Y-merging element is used as a result of N-bit digital / analog computation.
In one configuration example of the optical computing unit of the present invention, the optical modulator is a light intensity modulator.
Further, in one configuration example of the optical computing unit of the present invention, the optical modulator is an optical phase modulator, and further includes a coherent detection unit that extracts an electric signal after digital-analog conversion, and the coherent detection unit includes: A phase shifter that can be adjusted so that a phase difference between the output light resulting from the N-bit digital / analog operation and the reference light having the same wavelength as that of the output light is π / 2, and the N-bit digital A coupler that combines the output light that is the result of the analog operation and the reference light and divides the output light into two equal parts, a first photodetector that converts one output light of the coupler into an electrical signal, A second photodetector for converting the other output light into an electrical signal, and a subtractor for obtaining a difference between the two electrical signals output from the first and second photodetectors. Is.

According to the present invention, a 2-input 1-output first Y-merging element that receives 1 to 2 signal lights as an input and a 2-input 2-output Y-merging / Y-branch element that receives 1 to 2 signal lights as inputs. Are connected in cascade according to the corresponding bits of the N-bit electric digital signal, respectively, and the continuous light to the second optical input port of each of the N elements connected in cascade is connected in cascade. By providing N optical modulators that individually modulate and generate signal light to the second optical input port, an optical arithmetic unit such as an optical DAC that can be mounted at high speed and high density can be realized. .

FIG. 1 is a block diagram showing a configuration of an N-bit optical DAC according to the first embodiment of the present invention. FIG. 2 is a diagram showing input / output characteristics of the N-bit optical DAC according to the first embodiment of the present invention. FIG. 3 is a block diagram showing a configuration of an N-bit optical DAC according to the second embodiment of the present invention. FIG. 4 is a diagram showing input / output characteristics of the N-bit optical DAC according to the second embodiment of the present invention. FIG. 5 is a block diagram showing a configuration of an N-bit optical DAC according to the third embodiment of the present invention. FIG. 6 is a diagram showing input / output characteristics of an N-bit optical DAC according to the third embodiment of the present invention. FIG. 7 is a block diagram showing a configuration of an N-bit optical DAC according to the fourth embodiment of the present invention. FIG. 8 is a perspective view showing a configuration of a Y junction element with a bias port of an N-bit optical DAC according to the fourth embodiment of the present invention. FIG. 9 is a plan view showing a configuration of a Y junction device with a bias port of an N-bit optical DAC according to the fourth embodiment of the present invention. FIG. 10 is a perspective view showing another configuration of the Y-merging element with a bias port of the N-bit optical DAC according to the fourth embodiment of the present invention. FIG. 11 is a plan view showing another configuration of the Y-merging element with a bias port of the N-bit optical DAC according to the fourth embodiment of the present invention. FIG. 12 is a diagram showing input / output characteristics of an N-bit optical DAC according to the fourth embodiment of the present invention. FIG. 13 is a block diagram showing a configuration of an N-bit optical DAC according to the fifth embodiment of the present invention. FIG. 14 is a block diagram showing a configuration of an N-bit optical DAC according to the sixth embodiment of the present invention. FIG. 15 is a diagram illustrating the relationship between the number of bits and the operation loss of the optical DAC. FIG. 16 is a diagram showing a specific configuration pattern for actually operating the optical DAC of FIG. FIG. 17 is a block diagram showing a configuration of an optical DAC to be simulated in the sixth embodiment of the present invention. FIG. 18 is a diagram illustrating a result of obtaining a temporal change in the optical signal intensity of each bit in the configuration of FIG. 17 by simulation. FIG. 19 is a diagram illustrating a result of a temporal change in the intensity of the optical signal detected by the photodetector with respect to the configuration of FIG. FIG. 20 is a block diagram showing a configuration of an N-bit optical DAC according to the seventh embodiment of the present invention. FIG. 21 is a block diagram showing a configuration of an optical DAC to be simulated in the seventh embodiment of the present invention. FIG. 22 is a diagram illustrating a result of obtaining a temporal change in the absolute phase of the optical signal of each bit in the configuration of FIG. 21 by simulation. FIG. 23 is a diagram illustrating a result of a temporal change in the intensity of the optical signal detected by the photodetector with respect to the configuration of FIG. FIG. 24 is a diagram illustrating a result of obtaining a temporal change in the absolute phase of the optical signal detected by the photodetector in the configuration of FIG. 21 by simulation. FIG. 25 is a diagram showing a result of obtaining a temporal change in the strength of the electric signal output from the subtractor with the configuration of FIG. 21 by simulation. FIG. 26 is a block diagram showing a configuration when the phase modulation method is applied to the N-bit optical DAC according to the first and fifth embodiments of the present invention. FIG. 27 is a block diagram showing a configuration when the phase modulation method is applied to the N-bit optical DAC according to the second embodiment of the present invention. FIG. 28 is a block diagram showing a configuration when the phase modulation method is applied to the N-bit optical DAC according to the third embodiment of the present invention. FIG. 29 is a block diagram showing a configuration of an N-bit optical DAC according to the eighth embodiment of the present invention. FIG. 30 is a block diagram showing another configuration of the N-bit optical DAC according to the eighth embodiment of the present invention. FIG. 31 is a diagram for explaining the effect of inputting bias light to the bias port of the Y junction element. FIG. 32 is a block diagram showing a configuration when a Y junction element is added to the final stage of the N-bit optical DAC according to the first and fifth embodiments of the present invention. FIG. 33 is a block diagram showing a configuration when a Y junction element is added to the final stage of the N-bit optical DAC according to the second embodiment of the present invention. FIG. 34 is a block diagram showing a configuration when a Y junction element is added to the final stage of the N-bit optical DAC according to the third embodiment of the present invention. FIG. 35 is a block diagram showing a configuration when a Y junction element is added to the final stage of the N-bit optical DAC according to the sixth embodiment of the present invention.

[First embodiment]
Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing the configuration of an N-bit optical DAC which is an optical computing unit according to the first embodiment of the present invention. The N-bit optical DAC is one of the N Y-merging elements 1-1 to 1-N having two inputs and one output connected in cascade (N is an integer of 2 or more) and one of the Y-merging elements 1-1 in the first stage. The optical waveguide 2-1 connected to the input port, the light output port of the (M-1) -th stage Y-merging element 1- (M-1), and one light of the M-th stage Y-merging element 1-M. An optical waveguide 2-M that connects the input port and inputs the signal light output from the Y-merging element 1- (M-1) to the Y-merging element 1-M (M is an integer of 2 to N) , The optical waveguides 3-1 to 3-N connected to the other optical input ports of the Y junction elements 1-1 to 1-N, and the optical waveguide connected to the optical output port of the final Y junction element 1-N. The waveguide 4 is composed of optical intensity modulators 5-1 to 5-N (optical modulators) provided in the optical waveguides 3-1 to 3-N.

Examples of the Y junction elements 1-1 to 1-N and the optical waveguides 2-1 to 2-N, 3-1 to 3-N, 4 include dielectric optical wiring such as PLC (Planar Lightwave Circuit), or Si thin wires Semiconductor optical wiring such as can be used.

Continuous light having the same wavelength and the same intensity is input to the optical waveguides 3-1 to 3 -N. In order to input such continuous light, the continuous laser light from the continuous laser light source is divided into N equal parts by a 1: N splitter, and the continuous light is input to each of the optical waveguides 3-1 to 3 -N. That's fine.

The optical waveguide 2-1 corresponds to zero input. That is, no light is input to the optical waveguide 2-1. The optical input to the optical waveguide 3-i (i is an integer from 1 to N) corresponds to the input of the i-th bit (i = integer from 1 to N). That is, the optical input to the optical waveguide 3-1 corresponds to the least significant bit (LSB: Least Significant へ Bit), and the optical input to the optical waveguide 3-N corresponds to the most significant bit (MSB: Most Significant Bit). To do.

As the light intensity modulators 5-1 to 5-N provided for each bit of the N-bit electric digital signal, a VOA (Variable Optical Attenuator) can be used. The optical intensity modulators 5-1 to 5 -N block the continuous light propagating through the optical waveguides 3-1 to 3 -N when the bit input of the corresponding electric digital signal is “0”, and the bit input is In the case of “1”, continuous light is allowed to pass. Thereby, the continuous light propagating through the optical waveguides 3-1 to 3-N is individually turned on / off according to the corresponding bit of the N-bit electric digital signal. In this way, signal light to be input to the other optical input port of the Y junction elements 1-1 to 1-N is generated.

The Y merge element 1-i (i = 1 to N) merges and propagates the light propagated through the optical waveguide 2-i and the optical waveguide 3-i at an equal ratio. At this time, the transmittance T of the light input from each of the two optical input ports of the Y combining device 1-i to the optical output port of the Y combining device 1-i is both 0.25. A fixed optical attenuator with a loss of 6 dB is provided in place of the first-stage Y junction element 1-1 so that light propagating through the optical waveguide 3-1 is attenuated by 6 dB and introduced into the optical waveguide 2-2. May be. Alternatively, the input / output of the light intensity modulator 5-1 may be adjusted so that the relative intensity of the output light with respect to the input light is 0.25.

Further, it is preferable to adjust so that the phase of the input light from the optical waveguide 2-i to the Y merge element 1-i and the input light from the optical waveguide 3-i to the Y merge element 1-i are in phase. In order to realize such adjustment, a phase shifter may be provided in each of the optical waveguides 3-1 to 3 -N. Examples of such a phase shifter include a heater type phase shifter that controls the phase of guided light by changing the refractive index of the optical waveguide by the thermo-optic effect, and the refractive index of the optical waveguide by the electro-optic effect. And a phase shifter for controlling the phase of guided light.

With the above configuration, the final optical output (DAC output) of the N-bit optical DAC can be obtained from the optical waveguide 4.
FIG. 2 shows a numerical calculation example (N = 8) of the input / output characteristics of the configuration of FIG. The horizontal axis of FIG. 2 is an electric digital signal, and the vertical axis is a value obtained by normalizing the amplitude of the optical output of the N-bit optical DAC, and the N-bit optical for all combinations (“00000000” to “11111111”) of the electric digital signal. The amplitude value of the optical output of the DAC is plotted.

As can be seen from FIG. 2, the amplitude value of the optical output increases linearly as the digital input increases, and the N-bit optical DAC operates normally. When all input bits are “1”, the maximum optical output amplitude value is ((2 ^N −1) / 2 ^N ) ^1/2 to 1. Since the total value of the optical input is N, the loss L _DAC due to the operation of the optical DAC is ₁₀ log ₁₀ (N / (2 ^N −1) / 2 ^N )) to ₁₀ log ₁₀ (N) [dB].

[Second Embodiment]
Next, a second embodiment of the present invention will be described. FIG. 3 is a block diagram showing a configuration of an N-bit optical DAC which is an optical arithmetic unit according to the second embodiment of the present invention. The N-bit optical DAC of this embodiment includes (N-1) Y-input / Y-branch elements 10-1 to 10- (N-1) having two inputs and two outputs connected in cascade, and the Y-junction at the final stage. A Y-junction element 13 that uses one optical output of the Y-branch element 10- (N-1) as one input, and an optical fiber connected to one optical input port of the first-stage Y-junction / Y-branch element 10-1. One optical output port of the waveguide 14-1, the (k-1) stage Y junction / Y branch element 10- (k-1) and one of the k stage Y junction / Y branch element 10-k An optical waveguide 14-k that connects the optical input port and inputs the signal light output from the Y combining / Y branching element 10- (k-1) to the Y combining / Y branching element 10-k (where k is 2 to (N−1)), one optical output port of the final stage Y-merging / Y-branching element 10- (N-1) and one optical input of the Y-merging element 13 Optical waveguide 14-N connecting the ports, and Y waveguide / Y branch elements 10-1 to 10- (N-1) and optical waveguides 15-1 to 15- connected to the other optical input port of Y junction element 13 15-N, optical waveguides 16-1 to 16- (N-1) connected to the other optical output ports of the Y-merging / Y-branching elements 10-1 to 10- (N-1), and the final stage An optical waveguide 17 connected to the optical output port of the Y junction element 13 and optical intensity modulators 18-1 to 18-N (optical modulators) provided in the optical waveguides 15-1 to 15-N. The

Each of the Y junction / Y branch elements 10-1 to 10- (N-1) includes a Y input element 11 having two inputs and one output, and a Y branch element 12 having the optical output of the Y junction element 11 as an input. It has a combined configuration.

The Y merge element 11 of each Y merge / Y branch element 10-j (j is an integer of 1 to (N-1)) merges the propagation light of the optical waveguide 14-j and the optical waveguide 15-j at an equal ratio. Output. At this time, the transmittance T ₁₁ of the light input to each of the two optical input ports of the Y junction element 11 to the optical output port of the Y junction element ₁₁ is both 0.5.

The Y branching element 12 of each Y joining / Y branching element 10-j (j = 1 to (N−1)) divides the input light from the Y joining element 11 into two equal parts. At this time, the transmittance T ₁₂ of the light input to the optical input port of the Y branch element 12 to the two optical output ports of the Y branch element ₁₂ is both 0.5.

As in the first embodiment, continuous light having the same wavelength and the same intensity is input to the optical waveguides 15-1 to 15-N. The optical input to the optical waveguide 15-i (i = 1 to N) corresponds to the input of the i-th bit.

As in the first embodiment, the N light intensity modulators 18-1 to 18-N are respectively connected to the optical waveguides 15-1 to 15-N when the bit input of the corresponding electric digital signal is “0”. When the bit input is “1”, the continuous light is allowed to pass. In this way, signal light to be input to the other optical input port of the Y combining / Y branching elements 10-1 to 10- (N-1) and the Y combining element 13 is generated.

As in the first embodiment, the phase of the input light from the optical waveguide 14-j to the Y merge / Y branch element 10-j and the phase of the input light from the optical waveguide 15-j to the Y merge / Y branch element 10-j Is preferably adjusted to be in phase, and the phase of the input light from the optical waveguide 14-N to the Y junction element 13 and the input light from the optical waveguide 15-N to the Y junction element 13 are in phase. It is preferable to adjust. In order to realize such adjustment, a phase shifter may be provided in each of the optical waveguides 15-1 to 15-N.

With the above configuration, the final optical output (DAC output) of the N-bit optical DAC can be obtained from the optical waveguide 17.
Further, in this embodiment, since the Y junction / Y branch elements 10-1 to 10- (N-1) are respectively provided with the Y branch elements 12, (N-1) optical output ports are increased. That is, the optical output of the optical waveguides 16-1 to 16- (N-1) corresponds to a DAC output of 1 to (N-1) bits.

FIG. 4 shows a numerical calculation example (N = 8) of the input / output characteristics of the configuration of FIG. As can be seen from FIG. 4, the N-bit optical DAC of this embodiment is operating normally, as in the first embodiment. In this embodiment, the maximum optical output amplitude value when all input bits are “1” is ((2 ^N −1) / 2 ^N−1 ) ^1/2 to √2.

Advantages of the present embodiment when compared with the configuration of the first embodiment are as follows.
(A) Regardless of the number of bits N, the loss L _DAC due to the operation of the optical _DAC is 3 dB. In this embodiment, since the maximum light output amplitude value is ˜√2 as described above, it is possible to obtain a light output larger than that in the first embodiment.

(B) The Y-junction / Y-branch element of T to 0.5 has a lower reflectivity and is easier to design. The Y junction elements 1-1 to 1-N having T = 0.25 required in the configuration of the first embodiment are not general elements, and in some cases, the light reflectance may increase. In contrast, the T = 0.5 Y merge / Y branch elements 10-1 to 10- (N-1) and the Y merge element 13 used in this example have low loss, low reflectivity, and can be manufactured. Easy design methods are already known.

(C) It is possible to simultaneously generate a DAC with a low bit number. In the present embodiment, 1 to (N−1) -bit DAC output can be taken out from the optical waveguides 16-1 to 16- (N−1) as described above. The low bit number DAC output can be used as a monitor port for operation calibration. Further, a resolution switching function can be obtained by combining the N-bit optical DAC of this embodiment with an optical switch that selects one of the outputs of the optical waveguides 16-1 to 16- (N-1) and 17. Can be realized.

[Third embodiment]
Next, a third embodiment of the present invention will be described. FIG. 5 is a block diagram showing a configuration of an N-bit optical DAC which is an optical arithmetic unit according to the third embodiment of the present invention. The N-bit optical DAC of the present embodiment is one of N cascaded 2-input 2-output Y merge / Y branch elements 10-1 to 10-N and the first stage Y merge / Y branch element 10-1. And the optical waveguide 14-1 connected to the optical input port of (Y), the Y junction of the (K-1) stage, and the optical output port of the Y branch element 10- (K-1) and the Y junction of the K stage. An optical waveguide 14-K connecting one optical input port of the Y branch element 10-K (K is an integer of 2 to N), and the other optical input of the Y junction / Y branch elements 10-1 to 10-N Optical waveguides 15-1 to 15-N connected to the ports, an optical waveguide 17 connected to one optical output port of the Y-junction / Y-branch elements 10-1 to 10-N at the final stage, and an optical waveguide 15 -1 to 15-N provided with light intensity modulators 18-1 to 18-N and cascaded two-input one-output (N-1) Of the Y junction elements 19-1 to 19- (N-1), the other optical output port of the Y junction / Y branch element 10-1, and one optical input port of the Y junction element 19-1. An optical waveguide that connects the waveguide 20-1 with the optical output port of the (k-1) -th stage Y junction element 19- (k-1) and one optical input port of the k-th stage Y junction element 19-k. Waveguides 20-k (where k is an integer from 2 to (N-1)), (j + 1) stage Y junction / Y branch element 10- (j + 1) other optical output port and j stage Y junction element An optical waveguide 21-j for connecting the other optical input port of 19-j (where j is an integer from 1 to (N-1)), and an optical output port of the Y-merging element 19- (N-1) at the final stage And an optical waveguide 22 connected to the.

The N-bit optical DAC of this embodiment has almost the same configuration as that of the second embodiment, and the continuous light input method is the same as that of the second embodiment. The difference from the second embodiment is that the Y junction / Y branch element 10-N is used in place of the Y junction element 13 so that the total number of Y junction / Y branch elements is one, which is N, and further cascade-connected ( N-1) Y junction elements 19-1 to 19- (N-1) are provided, and the output of the final stage Y junction element 19- (N-1) is used as the output of the bit counter (Bit カ ウ ン タ counter output). Is a point. From the output of the bit counter, light having an amplitude value proportional to the number of input digital signals “1” is output.

That is, the circuit of this embodiment is a circuit that can simultaneously execute an N-bit DAC operation and an N-bit counter operation.
In this embodiment, the input light from the optical waveguide 20- (K-1) to the Y junction element 19- (K-1) and the optical waveguide 21- (K-1) to the Y junction element 19- (K- It is preferable to adjust so that the phase of the input light to 1) is in phase. In order to realize such adjustment, a phase shifter may be provided in at least one of the optical waveguides 20- (K-1) or 21- (K-1).

FIG. 6 shows the result of simulating the input / output characteristics (N = 4) of the configuration of FIG. 5 using OptiSystem, which is an optical communication system simulator of Optiwave. Here, the bit rate of the signal generator is 10 Gbps, the amplitude value of the optical output (DAC （output) of the N-bit optical DAC is processed by a low-pass filter, and the high-frequency noise is cut off, and the optical output (Bit This shows a waveform obtained by processing the amplitude value of (counter output) with a low-pass filter and cutting high-frequency noise. The horizontal axis of FIG. 6 is time (ns).

According to FIG. 6, it was confirmed that the N-bit optical DAC and the N-bit counter are operating normally when all combinations (“0000” to “1111”) of the electrical digital signals are sequentially input. .

[Fourth embodiment]
In the first embodiment, an electrical signal can be obtained by photoelectrically converting the optical output of the optical waveguide 4 by the photodetector 6. In the second embodiment, an electrical signal can be obtained by photoelectrically converting the optical outputs of the optical waveguides 16-1 to 16- (N-1) and 17 by the photodetectors 23-1 to 23-N. In the third embodiment, an electrical signal can be obtained by photoelectrically converting the output of the optical waveguide 17 by the photodetector 24 and photoelectrically converting the output of the optical waveguide 22 by the photodetector 25.

However, in these first to third embodiments, the physical quantity to be detected by the optical output is the optical amplitude value. When the light output is photoelectrically converted by the light detectors 6, 23-1 to 23-N, 24, and 25 to obtain a light intensity value, the amplitude is squared and the output becomes a quadratic function. Therefore, it is necessary to perform the square root processing of the intensity value on the electric circuit side and linearize it. In order to avoid accumulation of calculation delay in the square root process, there are the following two methods.
(I) Control by bias port.
(II) Detection of optical amplitude value by coherent detection.

This example employs the above method (I). FIG. 7 is a block diagram showing the configuration of the N-bit optical DAC according to this embodiment. The same components as those in FIG. The N-bit optical DAC of this embodiment includes N cascaded Y-elements 1a-1 to 1a-N with three inputs and one output and a Y-junction element 1a-1 with a bias port at the first stage. The optical waveguide 2-1 connected to one optical input port, the optical output port of the Y junction element 1a- (M-1) with the (M-1) stage bias port, and the Y with the M stage bias port An optical waveguide 2-M that connects one optical input port of the confluence element 1a-M (where M is an integer of 2 to N), and the other optical input of the Y confluence elements 1a-1 to 1a-N with bias ports Optical waveguides 3-1 to 3-N connected to the ports, an optical waveguide 4 connected to the optical output port of the Y-junction element 1a-N with a bias port at the final stage, and optical waveguides 3-1 to 3-N The light intensity modulators 5-1 to 5-N provided in .

The difference from the first embodiment is that Y confluence elements 1a-1 to 1a-N with bias ports are used instead of Y confluence elements 1-1 to 1-N. FIG. 8 is a perspective view showing the configuration of the Y junction element 1a-1 with a bias port, and FIG. 9 is a plan view showing the configuration of the Y junction element 1a-1 with a bias port.

As shown in FIG. 8, the Y junction element 1a-1 with a bias port is formed on a substrate 101 made of a first dielectric material and on one surface 101a of the substrate 101. From the first dielectric material, An optical waveguide 102, an optical waveguide 103, an optical waveguide 104, and an optical waveguide 105 (bias port) made of a second dielectric material having a high refractive index are provided.

Examples of the first dielectric that forms the substrate 101 include silica (SiO ₂ ) such as quartz.
The second dielectric constituting the optical waveguides 102 to 105 is, for example, silicon (Si). The refractive index of silica is 1.4 in the communication wavelength band (for example, wavelength 1.5 μm), whereas the refractive index of silicon (Si) is 3.5. Therefore, when the optical waveguides 102 to 105 are made of silicon, the substrate and air act as a cladding, and light is confined in the optical waveguides 102 to 105.
Further, by forming the optical waveguides 102 to 105 on one surface 101a of the substrate 101, the Y junction element 1a-1 with a bias port is configured on the planar optical waveguide.

As shown in FIGS. 8 and 9, in the Y junction device with bias port 1a-1 according to the present embodiment, the optical waveguide 102, the optical waveguide 103, and the optical waveguide 104 are connected to each other at their respective ends. It constitutes a confluence element. Each of the optical waveguide 102 and the optical waveguide 103 is an optical waveguide for inputting signal light, and functions as a set of optical input ports. The optical waveguide 104 is an optical waveguide for outputting signal light and functions as an optical output port.
In this embodiment, the optical waveguide 102 and the optical waveguide 103 serving as an optical input port are disposed symmetrically with respect to an extension line of the optical waveguide 104 serving as an optical output port.

On the other hand, the optical waveguide 105 is an optical waveguide for bias light input and acts as a bias port.
The optical waveguide 105 serving as a bias port is disposed between the optical waveguide 102 and the optical waveguide 103. More specifically, the optical waveguide 105 is disposed on an extension line of the optical waveguide 104.

One end of the optical waveguide 105 closer to the Y junction element is formed in a taper shape in plan view. One end of the tapered optical waveguide 105 is referred to as a “tapered portion 105a”. The tapered portion 105a is disposed in close proximity to the optical waveguide 102 and the optical waveguide 103 of the Y junction element with a gap. As a result, the optical waveguide 102 and the optical waveguide 103 and the optical waveguide 105 are optically coupled to each other.

In such a Y-junction element 1a-1 with a bias port, the input signal light propagated through the optical waveguide 102 and the optical waveguide 103, which are optical input ports, interferes with the bias light propagated through the optical waveguide 105, respectively. The output light is output from the optical waveguide 104 serving as an optical output port.

The Y junction element 1a-1 with a bias port can be manufactured by the following process. That is, a silicon-on-insulator (SOI) substrate having a low-loss single crystal silicon layer is prepared, and a photosensitive material applied to the surface of the single crystal silicon layer by a photolithography technique is applied to a predetermined pattern as shown in FIG. If the silicon layer is etched after patterning, an element as shown in FIG. 8 can be obtained.

FIG. 10 is a perspective view showing another configuration of the Y junction element 1a-1 with bias port, and FIG. 11 is a plan view showing another configuration of the Y junction element 1a-1 with bias port.
As shown in FIG. 10, the Y junction element 1a-1 with a bias port is formed on a substrate 201 made of a first dielectric material such as silica (SiO ₂ ) and one surface 201a of the substrate 201. , Optical waveguide 202, optical waveguide 203, optical waveguide 204, and optical waveguide 205 made of a second dielectric material having a higher refractive index than that of the first dielectric material, such as silicon (Si). Here, the optical waveguide 202 and the optical waveguide 203 act as optical input ports, the optical waveguide 204 acts as an optical output port, and the optical waveguide 205 acts as a bias port.

In the Y merge element 1a-1 with bias port of FIGS. 10 and 11, an optical waveguide 205 acting as a bias port is disposed between the optical waveguide 202 acting as an optical input port and the optical waveguide 203. The optical waveguide 202 and the optical waveguide 203 are arranged so as to be symmetric with respect to the optical waveguide 205 in plan view.
The optical waveguide 205 and the optical waveguide 204 acting as an optical output port are connected to each other at one end thereof.

The optical waveguide 202 and the optical waveguide 203 that are optical input ports have coupling portions 202a and 203a that are coupled to the optical waveguide 205 that is a bias port. That is, the coupling portions 202a and 203a are portions corresponding to the length L _{DC in the} vicinity of the tip portions of the optical waveguide 202 and the optical waveguide 203, which are arranged in parallel with the optical waveguide 204 with a gap g _DC therebetween. As described above, the optical waveguide 202 and the optical waveguide 203 are provided with the coupling portions 202a and 203a, respectively, so that the optical waveguide 202 and the optical waveguide 203 and the optical waveguide 205 are separated so as to be coupled to each other and are directionally coupled. A vessel is formed. The length L _DC of the coupling portions 202a and 203a is preferably about 90% of the 3 dB coupling length.

As described above, the configuration of the Y junction element 1a-1 with the bias port has been described. However, the configurations of the other Y junction elements 1a-2 to 1a-N with the bias port are the same as those of the Y junction element 1a-1 with the bias port. It is.

In the present embodiment, bias light having a fixed intensity P _bias is input from a continuous laser light source (not shown) to the bias ports of all Y junction elements 1a-1 to 1a-N with bias ports. The phases of the signal light input to the two optical input ports of the Y junction elements 1a-1 to 1a-N with the bias ports and the bias light input to the bias port are in phase.

Similar to the first embodiment, the light input from each of the two optical input ports of the Y merge element 1a-i (i = 1 to N) with the bias port is output to the optical output port of the Y merge element 1a-i. FIG. 12A to FIG. 12C show input / output characteristics under different bias light intensities when both transmittances T are 0.25. 12A to 12C, the horizontal axis represents an electric digital signal, and the vertical axis represents a value obtained by normalizing the optical output intensity of the N-bit optical DAC.

12A shows a case where the product T _bias P _bias of the transmittance T _bias of bias light and the intensity P _{bias of the} bias light is 0, and FIG. 12B shows a case where T _bias P _bias is 0.0625. FIG. 12C shows a case where T _bias P _bias is 0.25. A characteristic indicated by a broken line 301 in FIGS. 12A to 12C represents a characteristic obtained by linearly approximating an actual input / output characteristic indicated by a solid line 300. R ² is the mean square error between the actual input / output characteristics and the linear approximation. In the examples of FIGS. 12A to 12C, the mean square error R ² is 0.9373, 0.9926, and 0.9973. Thus, as T _bias P _bias increases, the input / output relationship of the N-bit optical DAC gradually approaches a linear function from a quadratic function.

Therefore, if the intensity P _{bias of the} bias light is appropriately adjusted, the relationship between the input digital signal and the output of the photodetector 6 can be made linear, and the output of the photodetector 6 can be square root on the electric circuit side. No need to process.

However, as is apparent from FIGS. 12A to 12C, the intercept (light output intensity when all input bits are “0”) gradually increases from zero as T _bias P _bias increases. Increase. Therefore, subtraction processing for removing the intercept on the electric circuit side is required instead of square root processing after photoelectric conversion, but this processing can be realized relatively easily and at high speed using a threshold processor.

Note that this embodiment can also be applied to the second and third embodiments. When applied to the second and third embodiments, the Y junction element 11 and the Y junction element 13 of each of the Y junction / Y branch elements 10-1 to 10-N are respectively used as Y junction elements with bias ports. Replace it.

[Fifth embodiment]
Next, a fifth embodiment of the present invention will be described. In this example, the above method (II) is adopted. FIG. 13 is a block diagram showing the configuration of an N-bit optical DAC that is an optical computing unit according to the present embodiment. The same components as those in FIG. 1 are denoted by the same reference numerals. In the example of FIG. 13, N = 8, and the optical DAC calculation unit 30 and the coherent detection unit 31 corresponding to FIG. 1 are integrated.

The continuous laser light from the continuous laser light source 32 is divided into two equal parts by the Y branch element 33. One continuous light branched by the Y branch element 33 is input to the Y branch element 35 via the optical waveguide 34. The Y branch element 35 divides the continuous light input from the optical waveguide 34 into two equal parts. One continuous light branched by the Y branch element 35 is input to the Y branch element 38 via the optical waveguide 36, and the other continuous light branched is input to the Y branch element 39 via the optical waveguide 37. The

The Y branch element 38 divides the continuous light input from the optical waveguide 36 into two equal parts, and the Y branch element 39 divides the continuous light input from the optical waveguide 37 into two equal parts. One continuous light branched by the Y branch element 38 is input to the Y branch element 44 via the optical waveguide 40, and the other continuous light branched is input to the Y branch element 45 via the optical waveguide 41. The One continuous light branched by the Y branch element 39 is input to the Y branch element 46 via the optical waveguide 42, and the other continuous light branched is input to the Y branch element 47 via the optical waveguide 43. The

The Y branch element 44 divides the continuous light input from the optical waveguide 40 into two equal parts, and the Y branch element 45 divides the continuous light input from the optical waveguide 41 into two equal parts. The Y branch element 46 divides the continuous light input from the optical waveguide 42 into two equal parts, and the Y branch element 47 divides the continuous light input from the optical waveguide 43 into two equal parts.

One continuous light branched by the Y branch element 44 is input to the optical waveguide 3-8, and the other continuous light branched is input to the optical waveguide 3-5. One continuous light branched by the Y branching element 45 is input to the optical waveguide 3-4, and the other continuous light branched is input to the optical waveguide 3-1. One continuous light branched by the Y branch element 46 is input to the optical waveguide 3-2, and the other continuous light branched is input to the optical waveguide 3-3. One continuous light branched by the Y branching element 47 is input to the optical waveguide 3-6, and the other continuous light branched is input to the optical waveguide 3-7. In this way, continuous light input to the optical waveguides 3-1 to 3-8 described in the first embodiment can be realized.
Reference numerals 48-1 to 48-8 in FIG. 13 denote the phase shifters described in the first embodiment.

The coherent detection unit 31 equalizes the optical waveguide 50 connected to the other optical output port of the Y-branch element 33, the phase shifter 51 provided in the optical waveguide 50, and the propagation light of the optical waveguide 4 and the optical waveguide 50. A 3 dB coupler 52 (MMI coupler) that merges at a ratio and outputs the result by dividing into two, a photodetector 53 that converts one output light of the 3 dB coupler 52 into an electrical signal, and the other output light of the 3 dB coupler 52 are electrically It is comprised from the photodetector 54 converted into a signal.

For the phase shifter 51, before operating the configuration of FIG. 13 as an 8-bit optical DAC, the output light propagating through the optical waveguide 4 from the optical DAC arithmetic unit 30 and input to the 3 dB coupler 52, and the optical waveguide 50 It may be adjusted in advance so that the phase difference from the reference light that is propagated and input to the 3 dB coupler 52 becomes π / 2.

The two outputs from the 3 dB coupler 52 are received by two different photodetectors 53 and 54, and the difference between the two electrical signals output from the photodetectors 53 and 54 is obtained by a subtracter (not shown). In this way, the optical amplitude value can be detected by using a so-called balanced detector configuration.

In a normal coherent optical receiver, the phase of the local light different from the light source of the signal light is synchronized and used as the reference light. This phase synchronization requires a large cost. On the other hand, in the case of the optical computing unit as in this embodiment, since the transmitter and the receiver are integrated at a close distance, the optical DAC computation is performed only by preparing the coherent detection unit 31 for the reference light. The homodyne detection using the same continuous laser light source 32 as that of the unit 30 is possible.

In this embodiment, the example in which homodyne detection is applied to the first embodiment has been described, but it may be applied to the second to fourth embodiments.
When applied to the second embodiment, light from the same continuous laser light source as the optical waveguides 15-1 to 15-N is input to one optical input port of the 3 dB coupler, and the optical waveguides 16-1 to 16- A coherent detector may be provided instead of the photodetectors 23-1 to 23-N so that the light from (N-1) and 17 is input to the other optical input port of the 3 dB coupler. .

When applied to the third embodiment, light from the same continuous laser light source as the optical waveguides 15-1 to 15-N is input to one optical input port of the 3 dB coupler, and light from the optical waveguides 17 and 22 is input. May be provided in place of the photodetectors 24 and 25, respectively, so that is input to the other optical input port of the 3 dB coupler.

When applied to the fourth embodiment, light from the same continuous laser light source as the optical waveguides 3-1 to 3 -N is input to one optical input port of the 3 dB coupler, and light from the optical waveguide 4 is 3 dB. What is necessary is just to provide a coherent detection part instead of the photodetector 6 so that it may input into the other optical input port of a coupler.

[Sixth embodiment]
Next, a sixth embodiment of the present invention will be described. In the first, fourth, and fifth embodiments, the two optical input ports (except the bias port of the fourth embodiment) of the Y junction elements 1-i and 1a-i (i = 1 to N) are respectively used. The transmittance T of the input light to the optical output ports of the Y converging elements 1-i and 1a-i is 0.25. In the second and third embodiments, the signals are input to the two optical input ports of the Y junction element 11 constituting the Y junction / Y branch element 10-j (j is an integer from 1 to (N-1)). The transmitted light T ₁₁ to the optical output port of the Y-merging element 11 is set to 0.5, and the light input to the optical input port of the Y-branching element 12 constituting the Y-merging / Y-branching element 10-j The transmittance T ₁₂ to the two optical output ports of the Y branch element ₁₂ is 0.5. As a result, in the first to fifth embodiments, (N−1) light beams corresponding to the respective bits excluding the LSB of the N-bit electric digital signal are four times as large as the light beams corresponding to the adjacent lower bits. A light intensity difference is given to the N lights so as to have a light intensity of (6 dB).

As described above, in the first to fifth embodiments, the light intensity difference is given to the N lights corresponding to the N-bit electric digital signal on the output side of the modulator. A light intensity difference is given to N pieces of light not only on the output side but also on the input side.
FIG. 14 is a block diagram illustrating a configuration of an N-bit optical DAC that is an optical arithmetic unit according to the present embodiment.

The N-bit optical DAC 60 according to the present embodiment includes 1-input 2-output N Y branch elements 61-1 to 61-N (N = 4 in the present embodiment) cascaded and the first stage Y branch element 61- The optical waveguide 62-1 connected to one optical input port, one optical output port of the (M-1) -th stage Y branch element 61- (M-1), and the M-th stage Y branch element 61-. An optical waveguide 62-M for connecting the M optical input port and inputting the light output from the Y branch element 61- (M-1) to the Y branch element 61-M (an integer from M = 2 to N) ), Optical waveguides 63-N to 63-1 connected to the other optical output ports of the Y branch elements 61-1 to 61-N, and light intensity modulators provided in the optical waveguides 63-1 to 63-N 64-1 to 64-N (optical modulator) and one optical input port are connected to the optical waveguides 63-1 to 63-N, and the other optical input Two Y-input N-output elements 65-1 to 65-N cascaded so that the ports are connected to the optical output port of the Y-junction element at the previous stage, and the Y-junction element 65-1 at the first stage The optical waveguide 66-1 connected to the other optical input port, the optical output port of the (M-1) stage Y junction element 65- (M-1), and the M stage Y junction element 65-M. An optical waveguide 66-M that connects the other optical input port to input the light output from the Y combining element 65- (M-1) to the Y combining element 65-M, and the Y combining element 65 in the final stage. -Phase shifters 68-1 to 68-N that can be adjusted so that the phases of the light combined by the optical waveguide 67 connected to the -N optical output port and the Y combining elements 65-1 to 65-N are in phase. It consists of. As described above, N = 4 in the example of FIG.

Each Y branch element 61-i (i = 1 to N) divides the propagation light of the optical waveguide 62-i into two equal parts (branch ratio 1: 1). At this time, the transmittance T of the light input to the optical input port of the Y branch element 61-i to the two optical output ports of the Y branch element 61-i is both 0.5. Thus, each Y branch element 61-i is one of the two optical output ports of the upstream Y branch element except for the most upstream Y branch element that receives a single continuous light. Are connected in cascade so as to receive light output from the optical output port.

Thereby, the continuous laser light from a single continuous laser light source (not shown) is branched into N continuous lights corresponding to the respective bits of the N-bit electric digital signal, and each of the N-bit electric digital signals excluding the LSB. A difference in light intensity is applied to the N continuous lights so that (N−1) continuous lights corresponding to the bits have twice (3 dB) light intensity as compared to the continuous lights corresponding to the adjacent lower bits. Can be granted.

The light corresponding to the i-th bit counted from the MSB of the N-bit electric digital signal is not connected to the subsequent Y-branch element of the two optical output ports of the i-th Y-branch element 61-i from the most upstream. Output from the optical output port. The optical input to the optical waveguide 63-i corresponds to the input of the i-th bit counted from the LSB.

Similar to the light intensity modulators 5-1 to 5 -N, the light intensity modulators 64-1 to 64 -N provided for each bit of the N-bit electric digital signal each have a bit input of the corresponding electric digital signal “ When 0, the continuous light propagating through the optical waveguides 63-1 to 63-N is blocked, and when the bit input is 1, the continuous light is allowed to pass.

Similar to the optical waveguide 2-1, the optical waveguide 66-1 corresponds to zero input. That is, no light is input to the optical waveguide 66-1. The Y merge element 65-i merges the light propagated through the optical waveguide 66-i and the optical waveguide 63-i at an equal ratio (merging ratio 1: 1) and outputs the merged light. At this time, the transmittance T of the light input from each of the two optical input ports of the Y merge element 65-i to the optical output port of the Y merge element 65-i is both 0.5. In this way, each Y merge element 65-i receives the signal light intensity-modulated by the light intensity modulator 64-i as one optical input, and each Y merge element excluding the most upstream Y merge element is located upstream. Cascade connection is performed so that the light output from the light output port of the Y junction element is used as the other light input.

As a result, the N signal lights intensity-modulated by the light intensity modulators 64-1 to 64-N are merged into one and correspond to each bit except the LSB of the N-bit electric digital signal (N−1). It is possible to give a light intensity difference to the N signal lights so that each of the signal lights has a light intensity twice (3 dB) with respect to the continuous light corresponding to the adjacent lower bits.

The phase shifters 68-1 to 68-N are configured so that the light output intensities of the Y junction elements 65-1 to 65-N are maximized when the light intensity modulators 64-1 to 64-N are in the passing state. The phase is adjusted in advance (so that the phases of the light combined by the Y combining elements 65-1 to 65-N are in phase).

With the above configuration, the final optical output (output) of the N-bit optical DAC 60 can be obtained from the optical waveguide 67.
The calculation loss Loss in the present embodiment can be defined as follows:

Fixed in the formula (1), the maximum light output intensity when P _out _ _max were all input N-bit electric digital signal "1", the P _in = 1 in P _in the optical input intensity (in this embodiment ), a _out _ _max is the light output amplitude is the square root of the P _out _ _max.

When the intensity of the input light (input The) to above as N bit optical DAC 60 and the P _in = 1, the amplitude A _out _ _max of the light output of N bits light DAC 60 (output) is a recurrence formula as follows Given in.

Here, T is the light intensity transmittance (ideally 0.5) of the Y junction elements 65-1 to 65-N. Dividing both sides of equation (2) by √ (2 ⁿ ) gives the following equation.

Therefore, since the _{a n = pa n-1 +} q, the following equation is obtained by solving Equation (3).

When substituting n = N into equation (4), it can be arranged as the following equation.

The relationship between the number of bits N and the operation loss Loss of the N-bit optical DAC, which is obtained by substituting Equation (5) into Equation (1), is as shown in FIG. In FIG. 15, 400 indicates the operation loss Loss of the N-bit optical DAC 60 of this embodiment, and 401 indicates the operation loss Loss of the N-bit optical DAC of the second and third embodiments. In the second and third embodiments, the operation loss Loss monotonously increases with respect to the number of bits N, whereas in this embodiment, the operation loss Loss monotonously decreases with respect to the number of bits N and the loss is zero. Asymptotically,
In this embodiment, since the operation loss Loss decreases with respect to the number of bits N, a configuration suitable for high resolution can be realized.

FIG. 16A to FIG. 16D show specific configuration patterns when the N-bit optical DAC 60 of FIG. 14 is actually operated. FIG. 16A shows the case of taking out light output as it is. In this case, continuous laser light from the continuous laser light source 70 is input to the N-bit light DAC 60.

FIG. 16B shows a case where the optical output of the N-bit optical DAC 60 is directly detected by a single photodetector 71. In this case, an electrical signal can be obtained by photoelectrically converting the optical output of the N-bit optical DAC 60 by the photodetector 71.

FIG. 16C shows a case where continuous light having a specific amplitude and phase is added by the Y confluence element 74 and then directly detected by a single photodetector 75. In the example of FIG. 16C, the continuous laser light from the continuous laser light source 70 is divided into two equal parts by the Y-branch element 72, one continuous light is input to the N-bit light DAC 60, and the other continuous light is combined into Y. The element 74 combines with the output light of the N-bit optical DAC 60. At this time, the phase shifter 73 is phase-adjusted in advance so that the light output intensity of the Y junction element 74 is maximized.

FIG. 16D shows the case of so-called coherent detection. In the example of FIG. 16D, the continuous laser light from the continuous laser light source 70 is divided into two equal parts by the Y branching element 76, one continuous light is input to the N-bit light DAC 60, and the other continuous light is input to the 3 dB coupler. (MMI coupler) 78 joins the output light of the N-bit optical DAC 60. For the phase shifter 77, the phase difference between the output light input from the N-bit light DAC 60 to the 3 dB coupler 78 and the other continuous light (reference light) branched by the Y branch element 76 is π / 2. It may be adjusted in advance.

The 3 dB coupler 78 combines the output light of the N-bit light DAC 60 and the reference light phase-adjusted by the phase shifter 77 at an equal ratio, and divides the output light into two equal parts. Each of the photodetectors 79.80 converts the two output lights of the 3 dB coupler 78 into electrical signals. The subtractor 81 obtains the difference between the two electrical signals output from the photodetectors 79 and 80.
In the case of direct detection in FIGS. 16B and 16C, a nonlinear output of a quadratic function is obtained, but in the case of coherent detection, a linear output is obtained.

Hereinafter, simulation results using Optiwave manufactured by Optiwave are shown for the configuration of this example. Here, simulation results for the configuration of FIG. 17, that is, the configuration combining FIG. 14 and FIG. 16B are shown. The simulation conditions are as follows.

(I) For the laser light source 70, the wavelength is 1550 nm, the light intensity is 1 mW, the line width is 10 MHz, and the initial phase is −90 °.
(II) The optical intensity modulators 64-1 to 64-4 are assumed to have no loss, the LSB bit rate is 10 Gbps, the extinction ratio is infinite, the rise time and the fall time are 0.05 bits (8 ps). Electric digital signals “0000” to “1111” are sequentially input to the light intensity modulators 64-1 to 64-4 of each bit.
(III) The optical waveguide and coupler used in the configuration of FIG. 17 are assumed to have no loss, and further, no propagation delay difference and phase shift of the optical signal of each bit due to the optical path length difference. Therefore, since the light always merges in the same phase, the phase shifters 68-1 to 68-N for adjustment are omitted in the configuration of FIG.
(IV) For the photodetector 71, the conversion efficiency is 1 A / W, no noise, and unlimited bandwidth.

FIG. 18 shows the results obtained by simulating the temporal change in the optical signal intensity of each bit in the configuration of FIG. In FIG. 18, the vertical axis represents the optical signal intensity of each bit, and the horizontal axis represents time. According to FIG. 18, it can be seen that the intensity difference is already doubled (3 dB) between the bits before joining by the Y joining elements 65-1 to 65-4.

FIG. 19 shows a result of _obtaining a temporal change in the intensity _Pout of the optical signal detected by the photodetector 71 by simulation. According to FIG. 19, by sequentially input from the electric digital signal "0000" to "1111" to the light intensity modulators 64-1 to 64-4, it can be seen that P _out is increased gradually. When the electric digital signal is “1111”, P _{out is} 879 μW, so the operation loss is Loss to 0.56 dB. The value of the optical signal intensity _Pout matches the value obtained by the equation (5).

Further, according to FIG. 19, there is a bit transition, particularly a signal spike seen between “0111” and “1000”. This spike is caused by the fact that signals that are not intended instantaneously overlap each other because the rise time and fall time of the light modulation signal are finite, but they can be removed by using an appropriate low-pass filter.

[Seventh embodiment]
Next, a seventh embodiment of the present invention will be described. This embodiment is an example in which an optical phase modulator is used instead of the optical intensity modulator in the sixth embodiment. FIG. 20 is a block diagram showing a configuration of an N-bit optical DAC that is an optical computing unit according to the present embodiment. The same components as those in FIG. 14 are denoted by the same reference numerals.

The N-bit optical DAC 60a of this embodiment includes Y branch elements 61-1 to 61-N (N = 4 in this embodiment), optical waveguides 62-1 to 62-N, 63-1 to 63-N, 66. -1 to 66-N, 67, Y junction elements 65-1 to 65-N, and optical phase modulators 69-1 to 69-N (optical modulators) provided in the optical waveguides 63-1 to 63-N. ). As described above, N = 4 in the example of FIG.

The optical phase modulators 69-1 to 69-N provided for each bit of the N-bit electric digital signal respectively pass the optical waveguides 63-1 to 63-N when the bit input of the corresponding electric digital signal is “0”. When the phase of the continuous light propagating is output without being changed (in phase), and the bit input of the corresponding electric digital signal is “1”, the phase of the continuous light propagating through the optical waveguides 63-1 to 63-N is π. Output by shifting only (reverse phase). In this way, in-phase (0) or anti-phase (π) phases are individually assigned to the continuous light propagating through the optical waveguides 63-1 to 63-N according to the corresponding bits of the N-bit electric digital signal.

Other configurations are as described in the sixth embodiment. In this embodiment, the phase shifter in the N-bit optical DAC 60a is not necessary, and the number of components can be reduced.
However, in this embodiment, coherent detection is indispensable in order to distinguish between positive and negative of the signal amplitude. That is, the specific configuration pattern when actually operating the N-bit optical DAC 60a of FIG. 20 is only FIG.

Next, the simulation result using the OptiSystem manufactured by Optiwave is shown in the same manner as the sixth embodiment for the configuration of the present embodiment. Here, simulation results for the configuration of FIG. 21, that is, the configuration combining FIG. 20 and FIG. The simulation conditions are the same as the above (I) to (IV) except that optical phase modulators 69-1 to 69-4 are used instead of the optical intensity modulators 64-1 to 64-4.

FIG. 22 shows a result obtained by simulating the temporal change of the absolute phase of the optical signal of each bit in the configuration of FIG. Since the present embodiment is a phase modulation system, the phase difference (relative phase) between each bit is modulated to 0 / π. Although there is a light intensity difference of 4 times (6 dB) between each bit, the light intensity of each bit is constant with respect to time. Since the line width of the laser beam is 10 MHz, the absolute phase fluctuates without maintaining the initial phase, but since all the bits share the same laser beam, the relative phase between the bits is kept at 0 or π. ing.

FIG. 23 shows a result obtained by simulating the temporal change of the optical signal intensity _Pout obtained when 100% light is coupled to the photodetector 82 provided in the optical waveguide 67 of FIG. FIG. 24 shows a result obtained by simulating the time change of the absolute phase of the detected optical signal. It was confirmed that the optical phase intensity P _out is symmetric at the boundary between “0111” and “1000” of the input electric digital signal, whereas the absolute phase is inverted.

FIG. 25 shows the results obtained by simulating the temporal change in the electric signal intensity output from the subtracter 81 in FIG. The positive and negative electric field amplitudes of the optical signal were detected by homodyne detection, and it was confirmed that the output increased linearly with increasing digital input.

In this embodiment, the case where the phase modulation method is applied to the sixth embodiment has been described. However, the phase modulation method can be applied to the first to fifth embodiments.
FIG. 26 shows the configuration when the phase modulation method is applied to the first and fifth embodiments, FIG. 27 shows the configuration when the phase modulation method is applied to the second embodiment, and FIG. FIG. 28 shows a configuration when the phase modulation method is applied to the. As described above, coherent detection is indispensable in the phase modulation method, but in FIGS. 26 to 28, configurations related to homodyne detection (the coherent detection unit 31 and the Y branch element 33 in FIG. 13 and the Y branch element in FIG. 16D). 76, the phase shifter 77, the 3 dB coupler 78, the photodetectors 79 and 80, and the subtractor 81) are omitted.

[Eighth embodiment]
In the fourth embodiment, all Y junction elements 1-1 to 1-N of the first and fifth embodiments are replaced with Y junction elements with bias ports, and all of the second and third embodiments are replaced. The Y merge element 11 and the Y merge element 13 of the Y merge / Y branch elements 10-1 to 10-N are replaced with Y merge elements with bias ports. Similarly, all of the sixth and seventh embodiments are used. The Y junction elements 65-1 to 65-N may be replaced with Y junction elements with bias ports.

FIG. 29 shows a configuration in which the Y junction elements 65-1 to 65-N in the sixth embodiment are replaced with Y junction elements 65a-1 to 65a-N with bias ports. The Y junction element 65 in the seventh embodiment is shown in FIG. FIG. 30 shows a configuration in which −1 to 65-N are replaced with Y junction elements 65a-1 to 65a-N with bias ports. Similarly, in the configuration shown in FIGS. 26 to 28, the Y junction elements 1-1 to 1-N and the Y junction element 11 and the Y junction element 13 of the Y junction / Y branch elements 10-1 to 10-N are biased. You may make it replace with the Y junction element with a port.

The essential meaning of inputting the bias light to the bias port of the Y junction element is to improve the linearity between the input digital signal and the output of the optical DAC by giving an offset to the light intensity. FIG. 31 is a diagram showing the relationship between the input digital signal and the intensity of the optical signal detected by the optical DAC detector 82, 310 indicates the operation region of the optical DAC when no bias light is input, and 311 The operation region of the optical DAC when bias light is input is shown.

Therefore, any one of the Y merge elements in front of the photodetector may be replaced with a Y merge element with a bias port. Specifically, any one of the Y merge elements 1-1 to 1-N in FIGS. 1, 13, and 26 may be replaced with a Y merge element with a bias port, or FIGS. 27 and 28, any one of the Y merge element 11 and the Y merge element 13 of the Y merge / Y branch elements 10-1 to 10-N may be replaced with a Y merge element with a bias port. 14. One of the Y junction elements 65-1 to 65-N in FIG. 20 may be replaced with a Y junction element with a bias port.

Furthermore, one Y junction element may be added to the final stage, and bias light may be input to this Y junction element. FIG. 32 shows the configuration when the Y junction element 83 is added to the final stage of the optical DAC of the first and fifth embodiments, and the Y junction element 83 is added to the final stage of the optical DAC of the second embodiment. FIG. 33 shows the configuration in this case, FIG. 34 shows the configuration when the Y junction element 83 is added to the final stage of the optical DAC of the third embodiment, and Y in the final stage of the optical DAC of the sixth embodiment. FIG. 35 shows a configuration when the junction element 83 is added. In order to make the configurations of FIGS. 32 to 35 into the phase modulation system, the light intensity modulators 5-1 to 5-N, 18-1 to 18-N, 64-1 to 64-N are used as the optical phase modulators. Replace it.

The present invention can be applied to, for example, a technique for converting an electrical digital signal into an analog signal using an optical circuit.

1, 11, 13, 19, 65, 74, 83... Y junction element, 1a, 65a... Y junction element with bias port, 2 to 4, 14 to 17, 20 to 22, 34, 36, 37, 40 to 43 , 50, 62, 63, 66, 67, 102-105, 202-205 ... optical waveguide, 5, 18, 64 ... light intensity modulator, 6, 23-25, 53, 54, 71, 75, 79.80. , 82 ... photodetector, 10 ... Y converging / Y branching element, 12, 33, 35, 38, 39, 44 to 47, 61, 72, 76 ... Y branching element, 30 ... optical DAC arithmetic unit, 31 ... coherent Detection unit, 32, 70 ... Continuous laser light source, 48-1 to 48-8, 51, 68, 73, 77 ... Phase shifter, 52, 78 ... 3dB coupler, 60, 60a ... N-bit light DAC, 69 ... light Phase modulator, 81 ... subtractor, 101, 201 ... base .

Claims (9)

1. At least one of a 2-input 1-output first Y combining element that receives 1 to 2 signal lights and a 2-input 2-output Y combining / Y-branching element that receives 1 to 2 signal lights N elements (N is an integer of 2 or more) cascaded,
   Of the two optical input ports of each of the N elements connected in cascade, the first optical input to which signal light is input from the first optical input port without input light or the optical output port of the preceding element. The continuous light to the second optical input port different from the port is individually modulated according to the corresponding bit of the N-bit electric digital signal to generate the signal light to the second optical input port. With a light modulator,
   An optical computing unit characterized in that an output light obtained from an element at the final stage is used as a computation result.
2. The optical computing unit according to claim 1,
   N first Y confluence elements,
   These N first Y merge elements are the optical output ports of the upstream first Y merge elements except for the most upstream first Y merge element (N−1) first Y merge elements. The light output from the first optical input port is used as input light to the first optical input port, and N first Y combining elements including the most upstream first Y combining element are formed by the N optical modulators. The modulated signal light is cascaded so as to be input light to the second optical input port,
   The N optical modulators individually modulate the N continuous lights of the same wavelength according to the corresponding bits of the N-bit electric digital signal, and the N optical modulators of the first Y confluence elements. 2 to generate signal light to the optical input port
   An optical computing unit characterized in that the output light obtained from the first Y-merging element in the final stage is the result of N-bit digital / analog computation.
3. The optical computing unit according to claim 1,
   (N-1) comprising the Y merge / Y branch elements and the first Y merge element.
   These (N-1) Y merge / Y branch elements and one first Y merge element exclude (N-2) Y merge / Y branches except the most upstream Y merge / Y branch elements. The light output from the first optical output port of the upstream Y-merging / Y-branching element is input to the first optical input port, and the first Y-merging element and the first Y-merging element (N−1) Y merge / Y branch elements including one upstream Y merge / Y branch element and one first Y merge element receive the signal light modulated by the N optical modulators. Cascaded to provide input light to the second optical input port;
   The N optical modulators individually modulate the N continuous lights of the same wavelength according to the corresponding bits of the N-bit electric digital signal, respectively, and (N−1) Y combined / Y Generating signal light to the second optical input port of the branch element and the second optical input port of the one first Y-merging element;
   Of the two optical output ports of each of the (N-1) Y merging / Y branching elements, it is obtained from a second optical output port different from the first optical output port that outputs the signal light to the subsequent element. An optical computing unit characterized in that the output light and the output light obtained from the first Y-merging element at the final stage are obtained as a result of 1 to N-bit digital / analog computation.
4. The optical computing unit according to claim 1,
   N pieces of the Y merge / Y branch element are provided,
   These N Y merging / Y branching elements exclude the most upstream Y merging / Y branching element (N−1) Y merging / Y branching elements are the first of the upstream Y merging / Y branching elements. The light output from the optical output port is used as the input light to the first optical input port, and the N Y merging / Y branching elements including the most upstream Y merging / Y branching element are the N light beams. The signal light modulated by the modulator is cascaded to be input light to the second optical input port;
   The N optical modulators individually modulate the N continuous lights of the same wavelength according to the corresponding bits of the N-bit electric digital signal, respectively. 2 to generate signal light to the optical input port
   (N-1) second Y merging elements having two inputs and one output connected in cascade are further provided,
   Obtained from a second optical output port different from the first optical output port that outputs the signal light to the Y merging / Y branching element in the subsequent stage, out of the two optical output ports of the Y merging / Y branching element in the first stage The received signal light is input to the first optical input port of the second Y junction element at the first stage, and the second light at the (k−1) th stage (k is an integer of 2 to (N−1)). The signal light obtained from the optical output port of the Y-merging element is input to the first optical input port of the second Y-merging element of the k-th stage, and the (j + 1) -th stage (j is 1 to (N−1) )), The signal light obtained from the second optical output port of the Y-combining / Y-branching element is input to the second optical input port of the j-th second Y-merging element,
   The output light obtained from the first optical output port of the Y-junction / Y-branch element at the final stage is the result of N-bit digital / analog operation,
   An optical computing unit characterized in that the output light obtained from the second Y-merging element at the final stage is the result of N-bit counter computation.
5. The optical computing unit according to claim 1,
   N first Y confluence elements,
   These N first Y merge elements are the optical output ports of the upstream first Y merge elements except for the most upstream first Y merge element (N−1) first Y merge elements. The light output from the first optical input port is used as input light to the first optical input port, and N first Y combining elements including the most upstream first Y combining element are formed by the N optical modulators. The modulated signal light is cascaded so as to be input light to the second optical input port,
   Furthermore, it has N Y branch elements with 1 input and 2 outputs,
   Among these N Y branch elements, each Y branch element other than the most upstream Y branch element that receives a single continuous light is the first optical output of the two optical output ports of the upstream Y branch element. Cascade connection so that the light output from the port is the input,
   The output light obtained from the second optical output port of these N Y branch elements is used as input light to the N optical modulators,
   An optical computing unit characterized in that the output light obtained from the first Y-merging element in the final stage is the result of N-bit digital / analog computation.
6. In the optical arithmetic unit according to any one of claims 1 to 5,
   At least one of the first Y-merging elements or at least one of the Y-merging elements constituting the Y-merging / Y-branching element is in addition to the first and second optical input ports. An optical computing unit comprising a bias port to which a fixed intensity bias light is input.
7. In the optical arithmetic unit according to any one of claims 1 to 5,
   Furthermore, the output light obtained from the element at the last stage is used as input light to the first optical input port, and a Y-merging element is used that uses bias light having a fixed intensity as input light to the second optical input port,
   An optical computing unit characterized in that the output light obtained from the Y junction element is the result of N-bit digital / analog computation.
8. The optical computing unit according to any one of claims 1 to 7,
   The optical calculator is an optical intensity modulator.
9. The optical computing unit according to any one of claims 1 to 7,
   The optical modulator is an optical phase modulator;
   In addition, it has a coherent detector that extracts the electrical signal after digital-analog conversion.
   The coherent detection unit is
   A phase shifter that can be adjusted so that the phase difference between the output light resulting from the N-bit digital / analog operation and the reference light having the same wavelength as that of the output light is π / 2;
   A coupler that combines the output light resulting from the N-bit digital / analog operation and the reference light, and bisects the output light;
   A first photodetector for converting one output light of the coupler into an electrical signal;
   A second photodetector for converting the other output light of the coupler into an electrical signal;
   An optical computing unit comprising: a subtractor for obtaining a difference between two electrical signals output from the first and second photodetectors.

Images