JP2003218794A - Optical receiver - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To improve the receiving performance of a RZ (Return-to-Zero) optical pulse signal having low quality. <P>SOLUTION: A 10 dB optical wavelength depultiplexer 12 supplies a portion of the RZ optical signal pulse from an input terminal 10 and applies the remaining to an optical-optical gate device 16. An optical clock pulse generator 14 has basic a repetition frequency component of the RZ optical pulse signal inputted from the depultiplexer 12 and generates an optical clock synchronized to the RZ optical pulse signal, to be supplied to the gate device 16. The gate device 16 gates the optical clock from the clock pulse generator 14 by using the RZ optical pulse signal from the demultiplexer 12 as gate control pulse light. A photodiode device 18 converts output light from the gate device 16 into an electric signal. An electric band pass filter 20 extracts the frequency band component of a transmission signal and applies it to an electric determining device 22. The determining device 22 determines whether the frequency band component is '0' or '1' according to a predetermined threshold of the output signal from the filter 20. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an optical receiver for receiving an RZ optical pulse signal that carries binary data.

[0002]

In optical communication systems, in most cases,
Intensity Modulation-Direct Detection (Intensity Module)
ation-Direct Detection; IM
The DD) method is adopted. In this system, the signal light has a binary intensity, that is, "0" and "1" or a space and a mark. The optical receiver directly converts the input optical signal into an electric signal, and the amplitude of the electric signal is "1" or not.
The transmission information is demodulated by determining whether it is "0".

The intensity modulation method is NRZ (Non-Re
Turn-to-Zero modulation method and RZ (Retu)
There is a rn-to-Zero) modulation method. The NRZ modulation method is relatively simple in modulation, and the response frequency band required for the modulator and the light receiver is narrower than that of the RZ modulation method. On the other hand, in the RZ modulation method, the average power is NR.
When the same as the Z modulation method, the pulse energy of "1" bit can be made larger than that of the NRZ modulation method, and in the case of wavelength division multiplexing transmission, the effect of mutual phase modulation between adjacent wavelength channels can be reduced. is there.

[0004]

In the case where an optical signal intensity-modulated by the RZ modulation method is directly received by a photoelectric converter such as a photodiode and a bit determination of "1" or "0" is made with respect to its electrical output. , Determination points are determined on the amplitude axis of the electrical output and the time axis, and the presence or absence of a signal is determined at each determination point. For the judgment on the amplitude axis, the optimum judgment point is determined by the relationship between the average intensity fluctuation of "1" bits and the average intensity fluctuation of "0" bits in the optical signal input to the photoelectric converter. And the optimal decision point is used, the decision error occurrence rate becomes the minimum. However, if the optical signal includes temporal fluctuations, that is, timing jitter, the number of decision errors increases at the optimum decision point on the amplitude axis.
That is, the timing jitter causes a determination error.

The probability of occurrence of a decision error due to timing jitter is higher in the RZ modulated signal than in the NRZ modulated signal. Of course, the optimum decision points on the amplitude axis and the time axis can be tracked up to a certain speed by combining a tracking circuit using an appropriate electronic circuit. However, in a high-speed signal with a bit rate of 10 Gbps or more, the timing jitter has a larger ratio with respect to the time slot per bit and has a higher randomness, and as a result, the optimum timing jitter for each bit on the time axis. Tracking the decision point is difficult.

As described above, according to the conventional technique, even if the receiving apparatus makes a bit determination at the optimum determination point on the amplitude axis, a determination error occurs due to timing jitter, especially for an optical signal with a high bit rate. Cannot be avoided.

Further, if the transmission distance becomes long and the RZ optical pulse waveform is deformed due to polarization mode dispersion (PMD) and nonlinearity of the transmission fiber, it becomes very difficult to restore it to the original RZ optical pulse waveform. In such a case, in the conventional technique, the decision error rate rapidly increases even if the receiving apparatus slightly deviates from the optimum decision points on the amplitude axis and the time axis.

Moreover, in long-distance transmission, the PMD fluctuates with time, so that the fluctuation of the RZ optical pulse waveform at the receiving end becomes large. In that case, it becomes difficult for the receiving apparatus to track the optimum decision points on the amplitude axis and the time axis. In particular, in the case of the RZ optical pulse, the decision error rate with respect to the deviation of the time axis position from the optimum decision point increases almost linearly in the conventional method.

An object of the present invention is to provide an optical receiving device which solves the above problems of the conventional technology.

Another object of the present invention is to provide an optical receiving device capable of receiving an RZ optical pulse signal having timing jitter with a low judgment error.

Another object of the present invention is to provide an optical receiving apparatus capable of receiving an RZ optical pulse signal whose waveform is deformed by long-distance transmission and whose amount of deformation is temporally changed with a low determination error. To do.

[0012]

An optical receiving device according to the present invention is an optical receiving device for receiving an optical pulse signal of a first wavelength that carries a binary signal, and is an optical receiving device that divides the optical pulse signal into two. Depending on the output light of one of the demultiplexer and the optical demultiplexer,
A second synchronization pulse having the same repetition frequency as the basic repetition frequency of the optical pulse signal and synchronized with the optical pulse signal;
An optical clock pulse generator that generates an optical clock pulse of a wavelength, and an optical-optical gate that gates the optical clock pulse output from the optical clock pulse generator according to the intensity information of the other output light of the optical demultiplexer. It is characterized by comprising a gate device, a photoelectric converter for converting an optical clock pulse gated by the optical-optical gate device into an electric signal, and an electric judgment device for binary-valued determination of the output of the photoelectric converter. To do.

[0013]

Embodiments of the present invention will now be described in detail with reference to the drawings.

FIG. 1 is a schematic block diagram of an embodiment of the present invention. An RZ optical pulse signal is input to the input terminal 10 from an optical fiber transmission line (not shown). This RZ optical pulse signal has a wavelength of 1545 nm and an average power of 4.4 m.
Suppose that the rate is W and the rate is 9.95328 Gbit / s.

The 10 dB optical demultiplexer 12 is a part (about 10%, that is, 0.4 m) of the RZ optical pulse signal from the input terminal 10.
W) is supplied to the optical clock pulse generator 14, and the rest is applied to the optical-optical gate device 16. Details will be described later,
The optical clock generator 14 has a basic repetition frequency of 9.95328 of the RZ optical pulse signal input from the optical demultiplexer 12.
A GHz component is extracted, an optical clock having the basic repetition frequency component and synchronized with the RZ optical pulse signal is generated, and supplied to the optical-optical gate device 16.

The optical-optical gate device 16 gates the optical clock from the optical clock pulse generator 14 using the RZ optical pulse signal from the optical demultiplexer 12 as gate control pulse light. The output light of the optical-optical gate device 16 is input to the input terminal 10
The RZ optical pulse signal which is the same as the binary signal carried by the RZ optical pulse signal input to is also a clear waveform with no timing jitter. In other words, the optical clock pulse generator 14 and the optical-optical gate device 16 reproduce the waveform of the RZ optical pulse signal input to the input terminal 10 without passing through the optical / electrical conversion and the electrical / optical conversion to obtain the timing. Eliminate jitter.

The photodiode 18 converts the output light of the optical-optical gate device 16 into an electric signal. The electrical bandpass filter 20 extracts the frequency band component of the transmission signal from the output of the photodiode 18, and the electrical determination device 22
Apply to. The electrical determination device 22 uses the output signal of the electrical bandpass filter 20 as "0" or "depending on a predetermined threshold.
1 ”is determined.

The configuration and operation of the optical clock pulse generator 14 will be described. The photodiode 30 is the optical demultiplexer 12
The RZ optical pulse signal from is converted into an electric signal. The output of the photodiode 30 is amplified by the amplifier 32. The electrical bandpass filter 34 extracts the basic repetition frequency component of the RZ optical pulse signal input to the photodiode 30 from the output of the amplifier 32. PLL circuit 3
6 generates an electric clock whose frequency and phase are synchronized with the output of the electric bandpass filter 34.

The laser diode 38 outputs CW laser light having different wavelengths used for optical signal transmission to the electroabsorption type optical modulator 40. For example, if the signal wavelength is 1545
When the wavelength is nm, the wavelength of the output light of the laser diode 38 is 1555 nm. The electro-absorption optical modulator 40 includes
The output of the PLL circuit 36 is applied as a drive signal. The optical modulator 40 intensity-modulates the output light of the laser diode 38 according to the clock signal from the PLL circuit 36.
As a result, the same frequency as the basic repetition frequency of the RZ optical pulse signal input from the optical transmission line is 9.9328 GHz.
Optical clock pulses are generated. The optical amplifier 42, which is an erbium-doped fiber amplifier, optically amplifies the optical clock pulse output from the optical modulator 40 and applies it to the optical-optical gate device 16. The RZ optical clock pulse output from the optical amplifier 42 has a pulse waveform of almost sech ² type, the half value time width thereof is 12 ps, and the average power is 1 mW.

The structure and operation of the optical-optical gate device 16 will be described. The basic configuration of the optical-optical gate device 16 is as follows:
It is described in detail in the specification and drawings of 1-132083. Here, the basic operation will be briefly described to the extent necessary to understand the operation of the present embodiment.

The polarization controller 50 adjusts the output light of the optical amplifier 42 of the optical clock pulse generator 14 so that the polarization becomes 45 ° with respect to the slow axis and the fast axis of the birefringent element 52. , To the birefringent element 52. The birefringent element 52 is made of, for example, rutile crystal and has a thickness of 50 p.
It has a birefringence amount corresponding to s. The birefringent element 52 is
The input light, that is, the optical clock pulse generated by the optical clock pulse generator 14 is evenly separated into two polarization components (P wave and S wave), and 50 of the birefringent element 52 on the time axis.
Separate and output only ps.

The 3 dB optical coupler 54 includes a birefringent element 52.
The RZ optical pulse signal of about 4 mW from the optical demultiplexer 12 is multiplexed with the output light of. In the optical-optical gate device 16, the RZ optical pulse signal from the optical demultiplexer 12 is transmitted by the optical-optical gate device 16.
It functions as a gate control optical signal for controlling the gate on / off of. However, at the output stage of the optical coupler 54, the preceding optical clock pulse (for example, P wave optical pulse) and the following optical clock pulse (for example, P wave optical pulse) output from the birefringent element 52 are output.
The optical demultiplexer 1 is applied to the output of the birefringent element 52 so that the control light pulse is positioned during the S wave light pulse).
The RZ optical pulse signal from 2 is multiplexed. The multiplexed light is input to the semiconductor optical amplifier (SOA) 56.

The SOA 56 is forward biased by a DC power supply (not shown). The SOA 56 is, for example,
It is composed of an embedded waveguide type element using an InGaAsP / InP system as an active layer material, and antireflection coating is applied to both ends thereof. Since the optical power of the RZ optical pulse signal is sufficiently larger than the optical powers of the preceding optical clock pulse and the subsequent optical clock pulse, the RZ optical pulse signal is
Many carriers existing in A56 are consumed. Since the trailing optical clock pulse passes through the SOA 56 before its carrier density is completely restored to a steady value by current injection, the trailing optical clock pulse causes the SOA 56 to have a different refractive index value from the leading optical clock pulse. Will pass through. In other words, the subsequent optical clock pulse undergoes cross-phase modulation (XPM) by the RZ optical pulse signal (control optical pulse) in the SOA 56. As a result, the leading optical clock pulse and the trailing optical clock pulse propagate through the optical paths of different execution lengths while passing through the same SOA.
There is a phase difference between the two. The light intensity of the control light pulse and the length and current value of the SOA 56 are set so that the phase change of the subsequent signal light pulse due to the presence or absence of the control light pulse becomes π. As a result, the phase of the subsequent signal light pulse output from the SOA 56 differs by π depending on the presence or absence of the control light pulse.

The P wave component and the S wave component of the optical clock pulse that have passed through the SOA 56 are incident on the birefringent element 58 having the same polarization dispersion amount of 50 ps as the birefringent element 52. However, the polarization plane of the temporally preceding component (for example, S wave component) of the optical clock pulse coincides with the slow axis, and the polarization plane of the temporally subsequent component (for example, P wave component) coincides with the fast axis. The slow axis and the fast axis of the birefringent element 58 are arranged so as to coincide with each other. This axially arranged birefringent element 58
As a result, the two polarization components separated by the birefringence element 52 in the time axis direction are aligned at the same time axis position,
It becomes a single light pulse. A polarization phase adjuster may be arranged on the output side of the birefringent element 58 in order to absorb individual differences of the birefringent element.

The polarization direction of the output light of the birefringent element 58 differs by 90 ° between the presence and absence of the control light pulse, that is, the RZ light pulse input from the optical transmission line. The output light of the birefringent element 58 is generally elliptically polarized because the phase of the trailing light pulse is shifted by the phase difference due to XPM, but the phase change of the trailing light pulse due to the presence or absence of the control light pulse in the SOA 56 is changed. When π, the output light of the birefringent element 52 becomes linearly polarized light.

The polarizer 60 is an RZ input from the optical transmission line.
The birefringent element 58 is arranged with a polarization direction that extracts the output light in the presence of a light pulse. By the polarizer 60, pulse signal light in which the optical clock pulse is gated according to "0" and "1" of the RZ optical pulse input from the optical transmission line is obtained. For example, the optical clock pulse output from the birefringent element 58 when the RZ optical signal pulse is present, that is, when the bit is “1” is transmitted through the polarizer 60, and when the RZ optical signal pulse is not present, that is, The optical clock pulse output from the birefringent element 58 when the bit is “0” is cut by the polarizer 60. The polarization direction of the polarizer 60 is orthogonal to the above description, and the input RZ optical pulse signal is "
It may be arranged so that it is blocked when it is “1” and is transparent when it is “0.” In that case, it is obvious that the logical value determined by the electrical determination device 22 described below must be inverted.

The optical bandpass filter 62 includes a polarizer 60.
Only the wavelength component of the laser diode 38 is extracted from the output light of. As a result, unnecessary RZ optical pulse signals and SO
The spontaneous emission light generated in A56 can be removed. The output light of the optical bandpass filter 62 is applied to the photodiode 18. The photodiode 18 converts the input signal light into an electric signal and applies it to the electric bandpass filter 20. The electrical bandpass filter is a photodiode 18
Placed for the purpose of removing beat noise from the output of. The electrical determination device 22 identifies the binary signal value from the output of the electrical bandpass filter 20 by a method similar to the conventional IMDD method.

In this embodiment, the optical clock pulse regenerator 14 regenerates an optical RZ clock pulse synchronized with the input RZ optical pulse signal, and the optical-optical gate device 16 regenerates the binary value of the input RZ pulse signal. After being transferred to the optical RZ clock pulse, it is converted into an electric signal. In particular,
The optical-optical gate device 16 performs optical cross-phase modulation (XPM) on the phases of the following optical pulses of the preceding optical clock pulse and the following optical clock pulse obtained by separating the RZ optical clock pulse in the time axis direction. Since the RZ optical pulse signal (gate control optical pulse) is used for modulation, the phase of the subsequent optical clock pulse can be relatively stably modulated with respect to the timing jitter of the RZ optical pulse signal (gate control optical pulse).
That is, the optical-optical gate device 16 has a high tolerance for the timing jitter of the RZ optical pulse signal (gate control optical pulse), and as a result, the input R is generated with a very low transfer error.
The binary value of the Z pulse signal can be transferred to the optical RZ clock pulse. As a result, in this embodiment, the code error rate becomes lower than that in the case where the input RZ optical pulse signal is directly converted into an electric signal. The reason will be briefly described below.

The light-light gating device 16 has an R
The XPM for the trailing light pulse that the Z light signal pulse causes in the SOA 56 is utilized. Assuming that the optical pulse of the mark level (binary value “1”) of the input RZ optical signal pulse is a Gaussian waveform and its full width at half maximum (FWHM) is 25 ps, the time waveform I _ctrl (t) of this optical pulse is It is given by the following formula. That is,

[0030]

[Equation 1] However, σ = FWHM / 2.355, Tc indicates the time position of the peak of the optical pulse, and I ₀ indicates the peak crest value of the optical pulse.

The preceding light pulse causes the SOA 56 at time t = 0.
Suppose you pass through. Further, the time delay Tc from the preceding optical pulse at the peak of the input RZ optical pulse is 30 ps, and the PMD amount ΔT of the birefringent element 52 is 60 ps. At this time, the preceding light pulse receives the XPM amount Δθ in the SOA 56 due to the input RZ light pulse (gate control light pulse).
_lead is

[0032]

[Equation 2] Is. However, Δφ ₀ is a constant, τ is the gain recovery time of the SOA 56 (here, assumed to be 60 ps), and T ₀ is
The time is 0 or less (for example, -50 ps) when the input RZ light pulse (gate control light pulse) has a sufficiently small peak value.

On the other hand, the amount of XPM Δφ _trail received by the trailing light pulse in the SOA 56 is

[0034]

[Equation 3] Is.

Input from the optical clock pulse generator 14
The amount of gain modulation received by the optical clock pulse is negligible.
If it is small, the output power of the optical-optical gate device 16 is small.
-P _outIs   P_out= P₀{(Cos (Δφ_lead) −cos (Δφ_trail))^Two         + (Sin (Δφ_lead) −sin (Δφ_trail))^Two} (4) Is. However, P₀Is the power of the input optical clock pulse
Is.

FIG. 2 shows the change in P _out when the delay time Tc of the trailing component of the optical clock pulse with respect to the input RZ optical pulse signal changes based on the equation (4). The horizontal axis represents the delay time Tc, and the vertical axis represents P _out . From FIG. 2, even when the incident timing of the input RZ optical pulse signal to the optical-optical gate device 16 is largely deviated by the timing jitter, the input optical power P of the photodiode 18 is increased.
It can be seen that _out does not show a sharp drop. Therefore, the output signal light of the optical-optical gate device 16 is transferred to the photodiode 1
Even in the configuration in which photoelectric conversion is performed in 8 and binary determination is performed in the electrical determination device 22, the code error rate is significantly reduced.

In the conventional example in which the input RZ optical pulse signal is directly input to the photodiode, if there is timing jitter in the input RZ optical pulse signal, the electrical output linearly decreases according to the timing jitter, so the binary judgment is made. The bit error rate at the point abruptly deteriorates.

FIG. 3 shows the intensity I of the input RZ optical pulse signal.
The relationship between _ctrl and the output light power P _out of the optical-optical gate device 16 is shown. The horizontal axis represents I _ctrl and the vertical axis represents P _out .
From Figure 3, in the vicinity of the intensity I _ctrl is 1 input RZ optical pulse signal, even if I _ctrl is changed, does not change much P _{ou t.} Due to such a trigonometric function response of the light-light gate device 16, the output signal light of the light-light gate device 16 is photoelectrically converted by the photodiode 18, and the output signal light of the output light is determined by the electrical determination device 22.
In the value determination configuration, even if the determination threshold of the electrical determination device 22 fluctuates, it is possible to suppress a sharp increase in the bit error rate.

FIG. 4 shows an example of measuring the Q value of this embodiment and the conventional example with respect to the optical SN ratio (OSNR) of the input RZ optical pulse signal. The horizontal axis represents the optical SN ratio (OSNR) of the input RZ optical pulse signal, and the vertical axis represents the Q ² value (dB).
The bit error rate value (BER) measured in this example and the conventional example was converted into a Q value by the following formula, and displayed as a dB value. ^{That, BER = (1/2) erfc (} Q / 2 1/2) (5) Q 2 (dB) = 20log 10 (Q) (6) where, erfc () is a complementary error function.

As can be seen from FIG. 4, when the OSNR of the input RZ optical pulse is 10 dB, Q ² is 1 in the conventional example.
In contrast to ^2.5 dB, Q ² is 13 in this embodiment.
It was dB. As described above, the present embodiment exerts a greater effect on the input optical signal containing a large amount of timing jitter. Therefore, the poorer the signal quality of the input optical signal, the larger the difference from the conventional example.

FIG. 5 shows the results of comparison of the change in BER with respect to the judgment threshold value when the electric signal is binary-valued, in the present embodiment and the conventional example. The horizontal axis shows the deviation from the optimum threshold voltage (m
V), and the vertical axis represents BER in logarithm. As can be easily understood from FIG. 5, in comparison with the conventional example, B when the threshold voltage is moved particularly to the plus side (mark level side)
Deterioration of ER is suppressed.

As described above in detail, in the present embodiment, the judgment error can be reduced even in the case of a signal whose signal quality is remarkably deteriorated due to the timing jitter generated by transmission or the like. In other words, the timing jitter becomes strong. Also,
Accumulates spontaneous emission noise generated from optical amplifiers
Since it can be removed by the optical gate device 16, the electrical determination device 22
Then, even if the determination threshold changes, the change in BER can be suppressed. That is, there are remarkable effects such as a significant reduction in the code error rate and a reception performance that is strong against variations in transmission characteristics.

FIG. 6 shows another configuration example of the optical clock pulse generator 14. The RZ optical pulse signal demultiplexed by the optical demultiplexer 12 is input to the port A of the optical circulator 70 and is incident on the semiconductor mode-locked laser 72 from the port B.

The semiconductor mode-locked laser 72 has a surface 72a facing the port B of the optical circulator 70 as a cleavage surface and an opposite surface 72b as a high reflection surface. Face 72
A laser cavity is formed by a and 72b. Laser 72
The resonator length of is set so that the oscillation repetition frequency determined by this resonator length substantially matches the fundamental repetition frequency of the input RZ optical pulse signal. Surface 72 in the resonator
A saturable absorption region 72c is provided near a, and the rest is a gain region 72d. That is, in the gain region 72d, the current source 74
Supplies a constant current necessary for laser oscillation, and applies a constant voltage from the voltage source 76 to the saturable absorption region 72c. The output voltage of the voltage source 76 is such that when the RZ optical pulse is input from the port B of the optical circulator 70 to the saturable absorption region 72c, the saturable absorption region 72c is saturated, and loss occurs when the RZ optical pulse is not input. Is set. By setting in this way, the mode-locked laser 72 can receive the R light from the port B of the optical circulator 70.
Pulse oscillation is performed in synchronization with the Z optical pulse signal. That is, the semiconductor mode-locked laser 72 generates an RZ optical clock pulse having an oscillation repetition frequency locked to the basic repetition frequency of the input RZ optical pulse signal.

The optical clock pulse output from the semiconductor mode-locked laser 72 is input to the port B of the optical circulator 70 and applied to the optical-optical gate device 16 from the port C.

FIG. 7 shows another structural example of the optical-optical gate device 16. The RZ optical pulse signal from the optical demultiplexer 12 is 3d
It is divided into two by the B coupler 80, one of which is the variable optical delay line 8
2 and the other is an input port 86 of an optical circuit 84 formed of an optical waveguide having a symmetrical interference structure formed on a quartz substrate.
Enter in a. The optical circuit 84 also includes an input port 86.
b, 86c, 86d are provided, but the input port 86b is not used. The optical circuit 84 also includes a 3 dB optical coupler 8
8, 90, 92, SOA 94, 96, 3 dB optical coupler 98, and output ports 100a, 100b, 100c,
It comprises 100d.

The output light of the variable optical delay line 82 is the input port 8
Enter in 6d. The optical clock pulse from the optical clock pulse generator 14 is input to the input port 86c. The delay time of the variable optical delay line 82 is adjusted so that each pulse of the RZ optical pulse signal is located in the middle of each pulse of the optical clock pulse. For example, for a transmission rate of 10 Gbps, the delay time of the variable optical delay line 82 is set to 50 ps.

The optical clock pulse input to the input port 86c is divided into two by the 3 dB optical coupler 88, and each divided light is applied to the 3 dB optical couplers 90 and 92. The optical coupler 90 multiplexes the RZ optical pulse signal from the input port 86 a with the optical clock pulse from the optical coupler 88, and applies the combined light to the SOA 94. Also, the optical coupler 92
Combines the delayed RZ optical pulse signal from the input port 86d with the optical clock pulse from the optical coupler 88,
The combined light is applied to the SOA 96.

Within the SOA 94, the RZ optical pulse signal and the optical clock pulse are present at substantially the same time to the extent that the RZ optical pulse signal causes the optical clock pulse to undergo cross-phase modulation. In other words, in the range in which the pulsed light of the RZ optical pulse signal can bring the optical clock pulse into mutual phase modulation,
The RZ optical pulse signal leads the optical clock pulse in time.

On the other hand, in the SOA 96, the RZ optical pulse signals are preferably separated in time such that the corresponding optical clock pulses do not cause mutual phase modulation. Of course, it should not be delayed enough to affect the subsequent pulses of the optical clock pulse.

The optical coupler 98 combines the output lights of the SOAs 94 and 96 and supplies the combined light to the external optical bandpass filter 102 via the output port 100c. That is, S
In the OA94, an optical clock pulse which is / is not phase-modulated by the presence / absence of a pulse of the RZ optical pulse signal, and SO
The optical clock pulse that has passed through A96 is multiplexed and applied to the optical bandpass filter 102. If the mutual phase modulation amount in the SOA 94 is set to π, the optical coupler 98
Outputs an optical clock pulse to the output port 100c when the input RZ optical pulse signal is "1", and outputs an optical clock pulse to the output port 100b when the input RZ optical pulse signal is "0". By this switching, output port 1
00c to the optical bandpass filter 102, the optical clock pulse from the optical clock pulse generation circuit 14
A gated optical pulse signal is output depending on the presence or absence of a pulse of the optical pulse signal. The optical bandpass filter 102 extracts only the wavelength component of the optical clock pulse from the input light.

The output light from the output port 100b may be adopted. However, in that case, it should be noted that the logical value determined by the electrical determination circuit 22 is inverted from the logical value of the input RZ optical pulse signal.

FIG. 8 shows still another configuration example of the optical-optical gate device 16. The 3 dB optical coupler 110 is the optical demultiplexer 1
RZ optical pulse signal from 2 and optical clock generator 14
The optical clock pulse from the optical waveguide is combined and the combined light is incident on the electro-absorption type optical waveguide 112. Before entering the electro-absorption optical waveguide 112, the input RZ optical signal pulse precedes the optical clock pulse by about 10 ps in time.
It is assumed that the timing is adjusted. The electro-absorption optical waveguide 112 is reverse biased. That is, the negative DC voltage 114 is connected to one electrode of the electroabsorption type optical waveguide 112, and the other electrode is connected to the ground via the terminating resistor 116.

In the electro-absorption type optical waveguide 112, the input RZ
When a pulse is present in the optical pulse signal, that is, when it is "1", mutual absorption modulation occurs, whereby the optical clock pulse is low loss or lossless and the electro-absorption type optical waveguide 112.
Through. When there is no pulse in the input RZ optical pulse signal, that is, when it is "0", the electro-absorption type optical waveguide 11
At 2, the mutual absorption modulation does not occur, so that the optical clock pulse is absorbed. In this way, the reverse-biased electro-absorption optical waveguide 112 transmits the optical clock pulse when the input RZ optical pulse signal has a pulse, and when the input RZ optical pulse signal has no pulse,
It functions as an optical gate device that blocks an optical clock pulse.

The optical bandpass filter 118 extracts only the wavelength component of the optical clock pulse from the output light of the electroabsorption type optical waveguide 112 and supplies it to the photodiode 18.

The configuration shown in FIG. 8 requires relatively few basic components. However, the output fluctuation amount due to the fluctuation of the incident timing of the input RZ optical pulse signal (= timing jitter) becomes larger than that in the case of the configuration shown in FIG.

Conventionally, in so-called optical reproduction in which another optical clock pulse is generated based on an input optical pulse signal and the optical clock pulse is gated by the input optical pulse signal, the signal quality (Q value) is the same or It has been thought to deteriorate. However, this is because NRZ transmission is assumed.

In the NRZ optical signal, signal quality deterioration is mainly caused by intensity noise. As shown in FIG. 3, the optical reproduction causes the intensity distribution to change between the input signal light and the reproduced signal light. However, this alone is only useful for expanding the strength judgment threshold voltage window, and the reception Q
The value does not improve. This is because there is no means for discriminating whether the original signal was the mark level or the space level once the same level was generated by the intensity noise. By this, theoretically, the received Q is generated by the optical regeneration.
It has been considered that the value cannot be improved.

On the other hand, due to the higher bit rate and wavelength multiplexing, the RZ signal light has recently been used. In RZ signal light, signal quality deterioration due to timing jitter becomes significant. As shown in FIG. 2, in the present embodiment, it is possible to suppress the change in the reception intensity that occurs because the mark level temporally fluctuates due to the timing jitter. Conventionally, the improvement of the reception Q value due to the suppression of the timing jitter has not been considered at all.

The optical reproduction in this embodiment is greatly different from the optical reproduction on the optical fiber transmission line in the following points. That is, it is desirable that the optical clock generator 14 generate an RZ optical clock pulse having a narrow optical pulse width regardless of the optical pulse width of the input RZ optical pulse signal. This is because the narrower the optical pulse width of the optical clock pulse, the more the fluctuation of the reception level due to the timing jitter of the input RZ optical pulse signal can be suppressed. On the other hand, in the optical regenerator on the optical transmission line,
Since it is premised on the re-transmission after the optical regeneration, it is necessary to generate the optical clock pulse which is almost equal to the optical pulse width before the transmission of the input RZ optical pulse signal. For example, in the case of the present embodiment, the optical pulse width of the input RZ optical pulse signal pulse before transmission is 15 ps, but the pulse width of the optical clock pulse generated by the optical clock generator 14 is 8 ps.

Further, the wavelength of the optical clock pulse generated by the optical clock generator 14 may be any wavelength different from the wavelength of the input RZ optical pulse signal. This means that the range of selection of optical devices after optical reproduction is expanded. On the other hand, the optical regenerator on the optical transmission line needs to regenerate an optical signal having the same wavelength as the input RZ optical signal.

Although the embodiment of RZ transmission has been described, the present invention can be applied to NRZ transmission. For NRZ transmission, a converter for converting an NRZ optical pulse into an RZ optical pulse may be arranged immediately before the optical demultiplexer 12 just before the optical demultiplexer 12 or on the output side of the optical demultiplexer 12. good.

[0063]

As can be easily understood from the above description, according to the present invention, the receiving performance of the optical pulse signal can be greatly improved.

[Brief description of drawings]

FIG. 1 is a schematic block diagram of an embodiment of the present invention.

FIG. 2 is a diagram showing a change in P _out with respect to a delay time Tc of a trailing component of an optical clock pulse with respect to an input RZ optical pulse signal.

FIG. 3 is a diagram showing the relationship between the intensity I _ctrl of the input RZ optical pulse signal and the output optical power P _out of the optical-optical gate device 16.

FIG. 4 shows an optical SN ratio (OSN) of an input RZ optical pulse signal.
It is an example of measurement of the Q value for R).

FIG. 5 is a comparative example of a change in BER with respect to a determination threshold value when an electrical signal is binary-determined.

FIG. 6 is another configuration example of the optical clock pulse generator 14.

FIG. 7 is another configuration example of the optical-optical gate device 16.

FIG. 8 is still another configuration example of the optical-optical gate device 16.

[Explanation of symbols]

10: Input terminal 12:10 dB optical demultiplexer 14: Optical clock pulse generator 16: Optical-optical gate device 18: Photodiode 20: Electric bandpass filter 22: Electricity determination device 30: Photodiode 32: Amplifier 34: Electric bandpass filter 36: PLL circuit 38: Laser diode 40: Electroabsorption type optical modulator 42: Optical amplifier 50: Polarization control device 52: Birefringent element 54: 3 dB optical coupler 56: Semiconductor optical amplifier (SOA) 58: Birefringent element 60: Polarizer 62: Optical bandpass filter 70: Optical circulator 72: Semiconductor mode-locked laser 72a, 72b: reflective surface 72c: saturable absorption region 72d: gain area 74: Current source 76: Voltage source 80: 3 dB optical coupler 82: Variable optical delay line 84: Optical circuit 86a, 86b, 86c, 86d: input port 88, 90, 92: 3 dB optical coupler 94, 96: Semiconductor optical amplifier (SOA) 98: 3dB optical coupler 100a, 100b, 100c, 100d: output ports 102: Optical bandpass filter 110: 3 dB optical coupler 112: Electroabsorption type optical waveguide 114: DC voltage 116: Termination resistance

   ─────────────────────────────────────────────────── ─── Continued front page    (72) Inventor Komura Nishimura             2-15 Ohara Stock Exchange, Kamifukuoka City, Saitama Prefecture             Company KDI Research Institute F term (reference) 5F088 AA01 BA03 BB01                 5K002 AA03 BA02 BA04 BA05 BA13                       BA21 CA02 CA13 CA14 DA05                       FA01

Claims (4)

[Claims] 1. An optical receiving device for receiving an optical pulse signal of a first wavelength that carries a binary signal, the optical demultiplexer dividing the optical pulse signal into two, and one of the optical demultiplexers. An optical clock pulse generator having an optical clock pulse of a second wavelength, which has the same repetition frequency as the basic repetition frequency of the optical pulse signal according to the output light, and which is synchronized with the optical pulse signal, and the optical clock An optical-optical gate device that gates the optical clock pulse output from the pulse generator according to the intensity information of the other output light of the optical demultiplexer, and an optical clock pulse gated by the optical-optical gate device. An optical receiving device comprising: a photoelectric converter for converting into an electric signal; and an electric judging device for judging the output of the photoelectric converter in binary. 2. The optical receiver according to claim 1, wherein the optical pulse signal is an RZ optical pulse signal. 3. The optical-optical gate element divides the optical clock pulse output from the optical clock pulse generator into two, and the phase of one output light of the demultiplexer is set to the optical phase. A phase modulator that modulates according to the intensity information of the other output light of the demultiplexer, a multiplexer that combines the other output light of the demultiplexer and the output light of the phase modulator, and a The optical receiver according to claim 1, further comprising an optical filter that extracts the component of the second wavelength from the output light. 4. The optical clock pulse generator is a mode-locked laser having a gain region and a saturation absorption region in a laser resonator, the laser resonator being substantially the fundamental repetition frequency of the optical pulse signal. The optical receiving device according to claim 1, further comprising a mode-locked laser having a cavity length corresponding to, wherein the one output light of the optical demultiplexer is incident on the saturation absorption region of the mode-locked laser. .