JP2018205550A - Optical trigger pulse generator and label processor using the same - Google Patents

Abstract

To provide an optical trigger pulse generator capable of generating an optical trigger pulse for attenuating optical signals other than an initial optical pulse, and a label processor using the optical trigger pulse generator.SOLUTION: An optical trigger pulse generator is configured to, with an SOA 253 and an SOA driver OEIC 251 for driving the SOA 253, an amplitude initial optical pulse of an incoming optical packet 10 so as to align with a gain peak of the SOA 253. The optical trigger pulse generator is configured to transmit the optical pulse of the optical packet 10 having higher energy than a predetermined threshold, and connect, to an output side of the SOA 253, an optical rectification element 52 for blocking an optical signal having energy less than the predetermined threshold.SELECTED DRAWING: Figure 7

Description

  The present invention relates to an optical trigger pulse generator and a label processor for optical packet switching using the optical trigger pulse generator.

  In optical packet switching technology, which is a technology for transferring optical packets in the optical state at a high speed in units of optical packets, the label processor (LP; Label Processor) is required.

  The inventors have previously developed a burst mode LP for processing optical packets without a preamble. This LP uses a special serial-to-parallel converter (SPC) that requires an optical trigger pulse for operation. Such an optical trigger pulse is generated by a device called an optical trigger pulse generator (TPG).

Salah Ibrahim, Tatsushi Nakahara, Hiroshi Ishikawa, and Ryo Takahashi, "Burst-mode optical label processor with ultralow power consumption", OPTICS EXPRESS, April 4, 2016, vol. 24, no. 7, pp. 6985-6995 Stephen M. Jensen, "The Nonlinear Coherent Coupler", IEEE JOURNAL OF QUANTUM ELECTRONICS, October 1982, vol.QE-18, no. 10, pp. 1580-1583 Lian-Wee Luo, Salah Ibrahim, Arthur Nitkowski, Zhi Ding, Carl B. Poitras, SJ Ben Yoo, Michal Lipson, "High bandwidth on-chip silicon photonic interleaver", OPTICS EXPRESS, October 25, 2010, vol. 18 , no. 22, pp. 23079-23087 C. Koos, P. Vorreau, T. Vallaitis, P. Dumon, W. Bogaerts, R. Baets, B. Esembeson, I. Biaggio, T. Michinobu, F. Diederich, W. Freude and J. Leuthold, "All -optical high-speed signal processing with silicon-organic hybrid slot waveguides ", NATURE PHOTONICS, April 2009, vol. 3, pp. 216-219 Abha Aggarwal and Hardeep Singh, "Analysis of the Effect of Coupling Coefficient in a Nonlinear Directional Coupler based on Cross-Phase Modulation", International Journal for Innovative Research in Science & Technology, May 2015, vol. 1, pp. 406- 412

  The TPG devised by the present inventors (see Non-Patent Document 1) is characterized by extremely low power consumption and low delay, but has the following two problems to be addressed.

1) When a plurality of optical pulses are generated for an incoming optical packet, one optical trigger pulse (initially) is generated by the TPG in order to satisfy the operation requirement of an optical clock transistor array (OCTA). Only the optical pulse) is generated, and the operating conditions are finally adjusted so as to greatly reduce the amplitude of the second and subsequent optical pulses so that the triggered SPC operation is not performed. What you need to do.

2) If the extinction ratio of the “1” and “0” bits of the incoming optical packet is not sufficient, the optical pulse generated by the TPG may be incomplete and not suitable for use in SPC.

  Due to the above problems, the operation of the LP may be deteriorated when the optical signal other than the first optical pulse is not sufficiently attenuated.

  The present invention has been made in view of the above problems, and provides an optical trigger pulse generator capable of generating an optical trigger pulse in which an optical signal other than the first optical pulse is attenuated, and a label processor using the optical trigger pulse generator. With the goal.

  The present invention adds an optical rectifying element (optical threshold element) to the output of the TPG in order to cope with the above problem. In this optical rectifying device, the first optical pulse generated by the TPG passes if it has a sufficient amplitude (energy), and the amplitude (energy) of the optical signal (optical pulse or noise) generated by the TPG is passed. ) Does not reach the threshold is blocked. Therefore, it is possible to take an intensity difference between the first optical pulse and another optical signal.

  Furthermore, the present invention uses a directional coupler formed of a nonlinear optical material as a suitable example of the optical rectifying element. Although maintaining the amplitude of the first optical pulse without reducing it is beneficial to the operation of the SPC, it is not easy to realize an optical rectifying element with low insertion loss. Various realization methods are conceivable for the optical rectifying element. Among them, a directional coupler formed of a nonlinear optical material is based on the energy of a propagating light pulse, and 1) the first generated by the TPG. The refractive index of the directional coupler is changed so that high energy light pulses are transferred to one output port and 2) all other low energy light outputs from the TPG are transferred to the other output port. Is done. In this directional coupler, high energy light pulses are separated and used to efficiently trigger the SPC. Based on the fact that the insertion loss of a directional coupler is theoretically zero, an optical rectifying device with low loss can be realized by this approach.

  Accordingly, the present invention is characterized by the following configuration.

An optical trigger pulse generator according to a first invention for solving the above-mentioned problems is as follows.
An optical trigger that generates an optical trigger pulse by amplifying an initial optical pulse of an incoming optical packet in accordance with a gain peak of the semiconductor optical amplifier using a semiconductor optical amplifier and a circuit that drives the semiconductor optical amplifier In the pulse generator,
An optical rectifying element that passes the first optical pulse having an energy higher than a predetermined threshold and blocks an optical signal having an energy not reaching the predetermined threshold is connected to the output side of the semiconductor optical amplifier. It is characterized by.

An optical trigger pulse generator according to a second invention for solving the above-mentioned problems is as follows.
In the optical trigger pulse generator according to the first invention,
The optical rectifier element is composed of a directional coupler formed of a nonlinear optical material, and outputs the first optical pulse having energy higher than the predetermined threshold to one output port, and reaches the predetermined threshold. It is characterized in that an optical signal having not-shown energy is output to the other output port.

An optical trigger pulse generator according to a third invention for solving the above-mentioned problems is as follows.
In the optical trigger pulse generator according to the second invention,
One heater in the coupling region of the directional coupler is thermally adjusted relative to the other waveguide, or a heater is provided to thermally adjust two optical waveguides in the coupling region together. Features.

An optical trigger pulse generator according to a fourth invention for solving the above-mentioned problems is as follows.
In the optical trigger pulse generator according to any one of the first to third inventions,
An optical demultiplexer for demultiplexing the optical packet into two;
Receiving one of the demultiplexed optical packets and driving the semiconductor optical amplifier; and
The semiconductor optical amplifier in which the other optical packet that has been demultiplexed is input, a voltage is applied to the anode terminal, and the cathode terminal is connected to the output side of the circuit;
A capacitor having one end connected to the anode terminal of the semiconductor optical amplifier and the other end grounded;
An optical transmission line adjusted so that a first optical pulse of the other optical packet is input in accordance with the gain peak of the semiconductor optical amplifier;
The circuit is
A trigger circuit that receives one of the demultiplexed optical packets and generates one electrical pulse;
The source terminal is grounded, the gate terminal is connected to the output side of the trigger circuit, the drain terminal is connected to the cathode terminal of the semiconductor optical amplifier, and is turned on when the electric pulse is input. A transistor for driving
It is characterized by having.

A label processor according to a fifth invention for solving the above-described problems is as follows.
Using the optical trigger pulse generator according to any one of the first to fourth inventions, a label process for extracting and identifying the label of the optical packet is performed.

  According to the present invention, in the optical trigger pulse generator, an optical trigger pulse in which an optical signal other than the first optical pulse is attenuated can be generated by adding an optical rectifying element to the output side of the semiconductor optical amplifier. . In addition, even when the extinction ratio between the “1” and “0” bits of the optical packet is not sufficient, an optical trigger pulse in which an optical signal other than the first optical pulse is attenuated can be generated. As a result, it is possible to provide an optical trigger pulse generator having such an effect and a label processor using the optical trigger pulse generator.

It is the schematic which shows the structure of a label processor (LP). It is the schematic which shows the structure of an optical clock transistor array (OCTA), (a) is the structure by a S & H scheme, (b) is the structure by a DoH scheme, (c) is the structure of a DB-MSMPD trigger circuit. . It is a figure which shows the comparison of use of DoH scheme, (a) is a case where TL signal with a positive polarity is utilized, (b) is a case where TL signal with a negative polarity is utilized. (A) is the schematic which shows the structure of an optical trigger pulse generator (TPG), (b) is a graph which shows the optical packet isolate | separated with the optical demultiplexer by electric power, (c) is It is a graph which shows the optical packet after the amplification spontaneous emission of SOA and the amplification in SOA by electric power. It is a figure explaining the 2nd problem of an optical trigger pulse generator (TPG), Comprising: The optical trigger pulse in case an extinction ratio is not enough was injected into DB-MSMPD trigger circuit of an optical clock transistor array (OCTA) It is a figure explaining the electric signal in a case. It is the schematic which shows the structure of the label processor (LP) based on this invention. It is the schematic which shows the structure of the optical trigger pulse generator (TPG) based on this invention. It is a figure which shows an example of an optical rectifier, Comprising: It is a figure which shows the directional coupler formed with the nonlinear optical material. It is a figure which shows the example of the output of the optical rectification element shown in FIG. It is a figure which shows before and after the rectification | straightening of an optical trigger pulse.

  Hereinafter, the configuration, operation, and problems of the conventional LP will be described with reference to the drawings, and then the configuration and operation of the TPG and LP according to the present invention will be described.

  The present inventors have developed the conventional LP20 shown in FIG. 1 in order to validate an optical transmission scheme known as optical packet switching (OPS). The LP 20 can efficiently process the burst mode optical packet 10 without requiring a preamble bit. The optical packet 10 includes a label 11 and a payload 12. In general, the OPS switching node includes a multi-port optical switch 40, and the role of the LP 20 is to extract and identify the label 11 of the optical packet 10 that arrives at each input port of the optical switch 40. . Based on the identified label 11, the optical switch 40 is configured to forward the optical packet 10 to a desired output port in the optical domain without electrical conversion.

  The LP 20 includes a shared controller 21 including a complementary metal oxide semiconductor (CMOS) -based FPGA (Field Programmable Gate Array), a label extractor (LE) 22 connected to the shared controller 21, and a pre-stage of the shared controller 21. The number of LEs 22 is the same as the number of channels of the optical switch 40.

  At each input port of the LE 22, a part of the incoming optical packet 10 is distributed to the LE 22 by the optical demultiplexer OD 1, and most of it is sent to the optical switch 40. First, for correcting the label processing time. Passes through a fiber delay line (FDL) 23. In the conventional serializer / deserializer (SerDes; Serializer / Deserializer), the bits of the entire optical packet 10 are converted, but in the LE 22, only the label 11 of the optical packet 10 is processed. The present invention relates to the LE22 portion of LP20.

  The high-speed bit of the label 11 is divided into a plurality of channels by the optical demultiplexer OD2 after being processed by the TPG 25, and the speed is sufficiently reduced by the serial-parallel conversion executed by the OCTA 24. In addition, the shared controller 21 can be directly latched (signal held) via the off-chip comparator 28. The off-chip comparator 28 adjusts variations in signals output from each channel of the OCTA 24. The label 11 extracted by the LP 20 is searched from the transfer table by the shared controller 21, and when a match is found, the output port of the corresponding optical switch 40 is acquired. At this time, the optical packet 10 is incident on the PD 29 having a low-speed response, and the PD 29 generates an electrical signal (envelope signal) as long as the optical packet 10 does not end. Based on the envelope signal, the shared controller 21 can detect the incidence of the optical packet 10.

  In order to enable the label processing of the label 11 of the optical packet 10 without a preamble, the LE 22 adopts a special operation mechanism. The two basic elements are 1) the optical trigger pulse is supplied. OCTA 24, a burst mode serial / parallel converter that sometimes operates, and 2) a burst mode that selectively uses the first bit of the incoming optical packet 10 to generate a synchronized optical trigger pulse for OCTA 24. TPG25.

  The OCTA 24 is composed of a number of conversion channels. Each channel needs to be supplied with one optical trigger pulse in order to perform serial-parallel conversion with a specific bit of the label 11 of the incoming optical packet 10. The optical trigger pulse generated by the TPG 25 is divided by the optical demultiplexer OD2 into the same number as the SPC channel, and each of the divided time delays is added in accordance with the arrival time of the desired bit of the SPC. And is coupled to the OCTA 24 via the optical head 27. For each conversion channel of the OCTA 24, it is necessary to irradiate the optical trigger pulse at a specific timing according to the corresponding label bit, and the optical head 27 is used to couple these pulse groups to the OCTA 24.

  Depending on the incoming optical packet 10 divided by the optical demultiplexer OD3, the conventional TPG 25 generates several optical pulses with reduced amplitudes instead of one optical pulse. In order to greatly reduce the amplitude of the second light pulse so that the SPC operates correctly, it is necessary to finally adjust the operating conditions. In addition, when the extinction ratio of the “1” and “0” bits of the incoming optical packet 10 is not sufficient, the optical trigger pulse generated by the TPG 25 may be incomplete and may not be suitable for use in SPC. The present invention addresses such problems, and in order to solve these drawbacks, rectification for the optical trigger pulse, which is the output light of the TPG 25, is performed by an optical rectifying element described later.

[OCTA]
Here, the operation mechanism of the OCTA 24 will be described.

  2 (a) and 2 (b) are composed of several serial-parallel conversion channels 241 and 242 installed on a common transmission line TL through individual high electron mobility transistors (HEMTs) Tm. 1 shows a configuration of an OCTA 24 that is a monolithic integrated circuit.

  As shown in FIG. 1, an avalanche photo detector and transimpedance amplifier (APD-TIA) 26 in burst mode is an electric signal for the optical packet 10 divided by the optical demultiplexer OD 4 of the LE 22. This signal is later input to OCTA 24 and propagated by its common transmission line TL. Thereafter, conversion of each label bit of the label 11 is performed using these serial-parallel conversion channels 241 and 242. The timing of each optical trigger pulse is adjusted according to the presence of a bit assigned to the common transmission line TL, and the optical trigger pulse is used for each serial-parallel conversion channel 241 and 242 to perform bit conversion.

  In order for the OCTA 24 to operate correctly, only one bit of the label 11 is converted, so the high electron mobility transistor Tm needs to be switched on for a short time. In order to generate one narrow electric pulse for controlling the high electron mobility transistor Tm, a discharge-based metal-semiconductor-metal photodetector (DB-MSMPD; Discharge based-Metal Semiconductor) shown in FIG. Metal Photo Detector) trigger circuit 243 is used. The DB-MSMPD trigger circuit 243 will be described later.

  For example, if the label 11 is composed of 25 Gbps bits, the electrical pulse generated by the DB-MSMPD trigger circuit 243 must be sufficiently smaller than 40 picoseconds to control the high electron mobility transistor Tm. Don't be. If the generated electrical pulse is not sufficiently narrow, the output of the OCTA 24 conversion channel does not correspond to the desired label bit required and a false output is generated.

The serial / parallel conversion in the OCTA 24 was initially performed by the sample and hold (S & H) scheme shown in FIG. The serial-parallel conversion channel 241 according to the S & H scheme has a high electron mobility transistor Tm having a drain connected to the common transmission line TL and a gate connected to the DB-MSMPD trigger circuit 243, and a source of the high electron mobility transistor Tm. The capacitor C _hold has one end connected and the other end grounded, and an amplifier Amp having one end connected to the source of the high electron mobility transistor Tm and the other end serving as an output.

Here, the charge corresponding to the corresponding bit level is sampled in the capacitor C _hold through the high electron mobility transistor Tm. The change in voltage generated by capacitor C _hold is amplified to produce the output of the channel. In this SPC scheme, the difference between the charge samples corresponding to the “1” and “0” bits, respectively, is not sufficient.

The reason is that when C _hold is charged and becomes a bit of “1”, the voltage of the source terminal S of the high electron mobility transistor Tm increases and switches off. Further, since it is inevitable that a part of the bit voltage disappears due to the characteristic impedance of the common transmission line TL that exists in parallel with the turn-on conversion channel, C _hold cannot be charged efficiently.

Therefore, instead of the S & H scheme, the discharge or hold (DoH) scheme shown in FIG. 2B has been used. The series-parallel conversion channel 242 according to the DoH scheme has a high electron mobility transistor Tm whose source is connected to the common transmission line TL and whose gate is connected to the DB-MSMPD trigger circuit 243, and a drain of the high electron mobility transistor Tm. One end is connected to the capacitor C _hold having one end connected to the other end and the drain of the high electron mobility transistor Tm, one end connected to the drain of the high electron mobility transistor Tm, and one end connected to the drain of the high electron mobility transistor Tm. And a resistor R _high to which a voltage is applied to the other end.

Here, the charge initially in C _hold is discharged to the common transmission line TL or held without change. The discharge of C _hold is performed more efficiently than the charge, and a larger charge difference is generated at C _hold .

  The advantage of the DoH scheme is further enhanced by using a negative voltage span, that is, a label signal having a negative voltage for a “1” bit and a voltage of 0 for a “0” bit.

FIGS. 3A and 3B show a comparison of the use of the DoH scheme using a TL signal having a positive or negative polarity. The gate voltage signal V _GS generated by the light trigger pulse is indicated by a solid line pulse. When the gate voltage signal V _GS exceeds the threshold voltage V _th and the high electron mobility transistor Tm is turned on, the bias voltage ΔV _DS between the drain terminal D and the source terminal S has a positive polarity shown in FIG. Compared with the case of FIG. 3, the case of negative polarity shown in FIG.

In the case of negative polarity, this makes it possible to discharge the C _hold with a high current enabled with a high bias voltage ΔV _DS more efficiently. Further, when the energy of the light trigger pulse is reduced, the amplitude of the electric pulse of the gate voltage signal V _GS is also correspondingly reduced as shown by the dotted line pulse. Even if the gate voltage signal V _GS is reduced in this way, if the value of the bias voltage ΔV _DS is initially high, efficient conversion can be performed as in the case of positive polarity in which the light trigger energy is not reduced. In FIGS. 3A and 3B, V _D is a voltage at the drain terminal, V _S is a voltage at the source terminal, and ΔV _S is a voltage span at the source terminal.

[OCTA DB-MSMPD trigger circuit]
Next, the DB-MSMPD trigger circuit 243 will be described.

When the optical trigger pulse is incident, the DB-MSMPD trigger circuit 243 shown in FIG. 2C outputs a positive electric pulse on the positive output side, and outputs a negative electric pulse on the negative output side. A resistor R _bias having one end attached to the PD and the positive output side of the MSM-PD and the other end grounded, and a resistor R _bias having one end attached to the positive output side of the MSM-PD and a voltage applied to the other end and _bias, one end attached to the negative output side of the MSM-PD, a resistor R _in the voltage V _in applied to the other end, one end attached to the negative output side of the MSM-PD, the other end of which is grounded and a capacitor C _in. The positive output side of the MSM-PD is connected to the gate of the high electron mobility transistor Tm.

  The discharge-based DB-MSMPD trigger circuit 243 is a basic element of the burst mode device and is used to convert the optical trigger pulse into a narrow electric pulse capable of controlling the gate terminal G of the high electron mobility transistor Tm. . The DB-MSMPD trigger circuit 243 uses MSM-PD having a surface structure suitable for manufacturing. In the case of PIN (Positive-Intrinsic-Negative) photodetectors, the reaction is limited by the RC time constant, but in the case of MSM-PD, the rise time is very fast and this limitation is due to the very low capacitance. Only electronic travel time.

However, the MSM-PD reaction has a long tail due to the low mobility of holes. Therefore, a discharge-based one is used to generate a sufficiently narrow electric pulse in MSM-PD. The input capacitor C _in of the DB-MSMPD trigger circuit 243 is charged from the beginning, and maintains a high bias voltage. Next, when a light pulse enters the MSM-PD, a photocurrent is generated. Resistor R _in can not be put corresponding current by the input voltage V _in to high, the input capacitor C _in is discharged, resulting in the bias voltage is lowered. One important feature of MSM-PD is that no current can flow even when carriers are present because there is no bias voltage.

If multiple optical trigger pulse to the DB-MSM-PD trigger circuit 243 is continuously supplied, the first light pulse is depleted turned into a carrier of the input capacitor C _in, sufficient energy to eliminate the influence of subsequent light pulses It is necessary to have. Only in this completely depleted state, the DB-MSMPD trigger circuit 243 can generate one electric pulse having a sufficiently narrow amplitude required by SPC.

To the input capacitor C _in the full depletion state, high power consumption in order to obtain a sufficient amplitude for optical trigger pulse is determined indirectly. For example, when it is necessary to use an EDFA to generate an optical trigger pulse, energy is continuously consumed by the EDFA even when it is used only when the optical packet 10 arrives. In order to enable power-efficient operation, as will be described later, in the conventional TPG 25 of the inventors, a semiconductor optical amplifier (SOA) that consumes energy only when an optical packet arrives and does not consume energy at other times. ; Semiconductor Optical Amplifier) is used.

  The TPG 25 generates a second pulse with a smaller amplitude than the first optical pulse in the optical trigger pulse, but it is beneficial to maximize the difference between the first and second optical pulses. It is necessary to work on.

[Conventional TPG]
Next, the conventional TPG 25 will be described.

  The TPG 25 generates an optical trigger pulse by selectively amplifying the first bit of the incoming optical packet 10 using the SOA. Since the optical trigger pulse was originally a part of the incoming optical packet 10, the generated optical trigger pulse is also a synchronized trigger and can be subjected to serial-parallel conversion without jitter when applied to the OCTA 24.

  FIG. 4 (a) shows a SOA driver opto-electronic integrated circuit (OEIC) SOA that has been developed to generate electrical pulses with high peak current (> 600 mA) and narrow amplitude (up to 1 ns). The driver OEIC 251 is shown. The SOA driver OEIC 251 includes a DB-MSMPD trigger circuit 252 having the same configuration as the DB-MSMPD trigger circuit 243 described above, a pulse shape changer Mod connected to the output side of the DB-MSMPD trigger circuit 252, and a pulse shape changer A gate terminal is connected to the output side of Mod, a source terminal is grounded, and a high electron mobility transistor Tr is turned on when an electric pulse is input.

In the SOA 253, the drain terminal of the high electron mobility transistor Tr is connected to the cathode terminal via the inductor L, and the voltage V _high is applied to the anode terminal of the capacitor C _s . one end is connected, the other end of the capacitor C _s is grounded. The SOA 253 amplifies the incoming optical packet 10 when the high electron mobility transistor Tr is turned on.

  The optical demultiplexer OD5 demultiplexes the optical packet 10 into two, one of which enters the DB-MSMPD trigger circuit 252 of the SOA driver OEIC 251 and the other of which enters the SOA 253. In the SOA driver OEIC 251, a part (see FIG. 4 (b)) separated from the incoming optical packet 10 by the optical demultiplexer OD 5 is generated in order to generate one electric pulse. The light enters the MSMPD trigger circuit 252. The electric pulse generated here is used to turn on the high electron mobility transistor Tr so that a high current temporarily flows through the SOA 253.

  The dotted curve in FIG. 4 (c) shows the normalized amplified spontaneous emission (ASE) of the SOA 253, which means that the electric pulse of the SOA driver OEIC 251 is applied to the SOA 253 without supplying optical input. It corresponds to the gain profile when used. For example, the optical packet 10 arriving at the SOA 253 is roughly adjusted by the optical transmission line 254 so as to be close to the gain peak. At this time, it is desirable to adjust so that the first optical pulse of the optical packet 10 matches the gain peak. Thereby, the first optical pulse of the optical packet 10 is greatly amplified, the carriers accumulated in the SOA 253 are lost, and the subsequent optical pulse is gradually extinguished due to the lack of gain (solid line in FIG. 4C). . The power consumption of the SOA 253 is limited to a short time after the arrival of the optical packet 10, and the power consumption of the TPG 25 can be greatly reduced.

  To avoid confusion, the DB-MSMPD trigger circuit 243 used in the OCTA 24 and the DB-MSMPD trigger circuit 252 used in the input stage of the TPG 25 use different parameters because they have different purposes. Designed.

  Since the DB-MSMPD trigger circuit 243 of each channel of the OCTA 24 needs to generate one electric pulse during picoseconds, and the optical trigger pulses used for all the channels of the OCTA 24 are generated by the same TPG 25, the OCTA 24 There is a strong demand to minimize the trigger energy required to operate the DB-MSMPD trigger circuit 243 in each channel.

  On the other hand, the DB-MSMPD trigger circuit 252 of the TPG 25 is required to generate electrical pulses at a more generous interval (several hundred picoseconds to nanoseconds) and is separated from the input optical packet 10. Triggered by the department.

[Problems with conventional optical trigger pulses]
Next, a problem with the conventional optical trigger pulse will be described.

  As described in the description of the DB-MSMPD trigger circuit 243 above, when the optical trigger pulse used to trigger the channel of the OCTA 24 is composed of a plurality of optical pulses, the first optical pulse and the second optical pulse are used. The pulse difference needs to be maximized.

  A method of directly increasing the difference between the first optical pulse and the second optical pulse generated by the TPG 25 is to shorten the time during which the gain of the SOA 253 exists, but this is difficult to achieve as will be described later. . The SOA driver OEIC 251 needs to pass high current instantaneously in the SOA 253, and maintains a very fast rise time for the gain of the SOA 253 even when the SOA drivers OEIC 251 and the SOA 253 are located very close to each other. Therefore, it is necessary to keep the inductance of the wiring between them very low.

  Therefore, whether the rise time is improved depends on whether the physical limitation is solved, but the fall time is controlled by the rate of carrier recombination in the SOA 253. Increasing the length of the SOA 253 increases the gain disappearance effect due to the input optical signal instead of increasing the capacitance of the SOA 253. Therefore, the temporary gain profile of the SOA 253 and the amplification of the first optical pulse propagated by the SOA 253 have a trade-off relationship.

  On the other hand, if the temporary gain profile of the SOA 253 is reduced, it becomes difficult to adjust the time of the first bit of the optical packet 10 and the electrical pulse of the SOA driver OEIC 251. In order to improve the reaction of TPG25, these difficult conditions are presupposed, so that it is desirable to rectify the output of TPG25 instead.

  The second problem regarding the reaction of TPG25 is shown in FIG. Here, the case where the extinction ratio between the “1” and “0” bits of the incoming optical packet 10 is not sufficient (in the case of a low ER packet) and the case where the extinction ratio is sufficient (in the case of a high ER packet). Show.

In the case of a high ER packet, in the DB-MSMPD trigger circuit 243, when the first optical pulse arrives at the MSM-PD, the charge Q _high of the capacitor C _in is almost completely depleted by the first optical pulse. The electrical signal at has a desired waveform.

On the other hand, in the case of a low ER packet, since the output of the SOA 253 generated before the first optical pulse is significantly increased, the DB-MSMPD trigger circuit 243 passes the capacitor through the MSM-PD before the first optical pulse arrives. A part of the charge Q _high of C _in leaks. In this state, since the bias of the entire MSM-PD terminal has already decreased due to the occurrence of leakage of charge Q _high , an electrical output having a sufficiently large amplitude is generated when the first optical pulse arrives at the MSM-PD. The electrical signal at node G has an incomplete waveform. In order to prevent this state, it is desirable to rectify the output of the TPG 25.

[Adjustment of TPG output signal]
Next, adjustment of the TPG output signal will be described.

In order to solve the above drawbacks, in the present invention, an optical rectifying element (optical threshold element) 52 is added to the output side of the SOA 253 in the TPG 25. That is, as shown in FIGS. 6 and 7, the TPG 51 of the present invention includes a conventional TPG 25 and an optical rectifying element 52 connected to the output side of the SOA 253 of the TPG 25. 6 and 7, the same reference numerals are given to the same components as those shown in FIGS. 1 and 4, and a duplicate description is omitted here. With such a configuration, the optical trigger pulse generated by the TPG 25 is rectified. The optical rectifying element 52 passes the first optical pulse generated by the TPG 25 having an amplitude (energy) higher than a predetermined threshold P _th set in advance, and the amplitude (energy) does not reach the threshold P _th. The optical signal (light pulse or noise) is blocked.

  Realizing the optical rectifying element 52 with low insertion loss is beneficial to the operation of the SPC, but it is not easy to realize such a device. Here, among several possible realization methods, as shown in FIG. 8, a directional coupler 521 formed of a nonlinear optical material 522 is used. The nonlinear optical material 522 may be any material as long as it has a large self-phase modulation (for example, Kerr effect) due to a nonlinear refractive index. For example, a polymer waveguide described in Non-Patent Document 4 is preferable.

  Here, the use of a directional coupler 521 in which two optical waveguides W1 and W2 are monolithically integrated on the same substrate will be considered. For the desired length Le, coupled mode excitation between the two optical waveguides W1, W2 is enabled, and the gap Ga between the two optical waveguides W1, W2 is reduced. In the current manufacturing process, a small gap Ga of about several tens of nanometers can be easily manufactured.

  The interaction length between the two optical waveguides W1 and W2 needs to include a nonlinear optical material 522 whose refractive index changes depending on the intensity of propagating light. The inventors' conventional TPG 25 can generate optical trigger pulses exceeding 6.4 pJ and can trigger 16 channels of OCTA 24 that require approximately 0.4 pJ optical pulses for each channel. Therefore, the amount of energy of the main light pulse generated by the TPG 25 is sufficient to cause a large change in the refractive index of the nonlinear optical material 522.

In order to obtain an answer for the propagation of the optical pulse of the directional coupler 521 of the nonlinear optical material 522, it is necessary to solve the following equations shown in Non-Patent Document 2 and Non-Patent Document 5. In order to explain the propagation loss in the optical waveguide, the loss parameter α is added to the ideal condition without loss studied in Non-Patent Document 2 and Non-Patent Document 5. Since the operation of the device is mainly based on the self-phase modulation period Q (z), the cross-phase modulation period is ignored. In the following equation, z is the mode propagation direction, A ₁ is the mode profile as a function of the propagation direction of one optical waveguide, A ₂ is the mode profile as a function of the propagation direction of the other optical waveguide, and κ is the two light beams. This is the coupling coefficient between the waveguides.

FIG. 9 shows an example of an output that can be generated by the optical rectifying element 52 when appropriately designed. Here, as shown in FIG. 8, it is considered that the output from the SOA 253 of the TPG 25 is coupled with the input port of the optical waveguide W1, and the energy of the optical pulse input to the input port of the optical waveguide W1 ( If the power is low, the transmittance to the output port OP1 is low and the transmittance to the output port OP2 is high. On the other hand, if the energy (power) of the optical pulse is high, the transmittance to the output port OP1 is high. And the transmittance to the output port OP2 is low. That is, according to the energy (power) of the input optical pulse, the input optical pulse is transferred to the output port OP1 or transferred to the output port OP2 (transfer function). According to such a transfer function, an optical pulse having energy (power) sufficiently exceeding the threshold P _th is transferred (transmitted) to the output port OP1 of the optical waveguide W1. The other part of the optical trigger pulse generated by the TPG 25 is subject to loss when viewed from the output port OP1 because a part of the optical signal is also transferred to the output port OP2. As described above, the optical rectifying element 52 has a transfer function that selectively enables transmission of a high-energy light pulse.

  FIG. 10 shows before and after rectification of the optical trigger pulse generated by the TPG 25 using the optical packet 10 having an insufficient extinction ratio. The dotted line is before rectification, and the solid line is after rectification. Here, the optical signal generated before the first optical pulse is greatly reduced (up to 10 db or more) due to an insufficient extinction ratio of the incoming optical packet 10. The second optical pulse has a small decrease in amplitude, but is reduced by several db from before the rectification, and the difference from the first optical pulse is large.

In order to increase the design flexibility of the optical rectifying element 52, a fine adjustment function of the refractive index may be added to the coupling region of the directional coupler 521. This enables precise control of the threshold value P _th and strict management of the transfer function. In order to change the temperature of the optical waveguides W1 and W2 together or relatively, for example, as described in Non-Patent Document 3, in the vicinity of the optical waveguides W1 and W2 or the optical waveguides W1 and W2 A micro heater 523 may be added in the vicinity of either one of the above.

  The present invention is suitable for an optical trigger pulse generator and a label processor for optical packet switching using the optical trigger pulse generator.

10 Optical packet 11 Label 21 Shared controller 22 Label extractor (LE)
23 Fiber Delay Line (FDL)
24 Optical clock transistor array (OCTA)
25 Optical trigger pulse generator (TPG)
26 Burst Mode APD-TIA
40 Optical switch 50 Label processor 51 Optical trigger pulse generator (TPG)
52 Optical rectifier

Claims (5)

An optical trigger that generates an optical trigger pulse by amplifying an initial optical pulse of an incoming optical packet in accordance with a gain peak of the semiconductor optical amplifier using a semiconductor optical amplifier and a circuit that drives the semiconductor optical amplifier In the pulse generator,
An optical rectifying element that passes the first optical pulse having an energy higher than a predetermined threshold and blocks an optical signal having an energy not reaching the predetermined threshold is connected to the output side of the semiconductor optical amplifier. An optical trigger pulse generator characterized by The optical trigger pulse generator according to claim 1.
The optical rectifier element is composed of a directional coupler formed of a nonlinear optical material, and outputs the first optical pulse having energy higher than the predetermined threshold to one output port, and reaches the predetermined threshold. An optical trigger pulse generator that outputs an optical signal having a non-energy to the other output port. The optical trigger pulse generator according to claim 2,
One heater in the coupling region of the directional coupler is thermally adjusted relative to the other waveguide, or a heater is provided to thermally adjust two optical waveguides in the coupling region together. Features an optical trigger pulse generator. In the optical trigger pulse generator according to any one of claims 1 to 3,
An optical demultiplexer for demultiplexing the optical packet into two;
Receiving one of the demultiplexed optical packets and driving the semiconductor optical amplifier; and
The semiconductor optical amplifier in which the other optical packet that has been demultiplexed is input, a voltage is applied to the anode terminal, and the cathode terminal is connected to the output side of the circuit;
A capacitor having one end connected to the anode terminal of the semiconductor optical amplifier and the other end grounded;
An optical transmission line adjusted so that a first optical pulse of the other optical packet is input in accordance with the gain peak of the semiconductor optical amplifier;
The circuit is
A trigger circuit that receives one of the demultiplexed optical packets and generates one electrical pulse;
The source terminal is grounded, the gate terminal is connected to the output side of the trigger circuit, the drain terminal is connected to the cathode terminal of the semiconductor optical amplifier, and is turned on when the electric pulse is input. A transistor for driving
An optical trigger pulse generator comprising: A label processor for extracting and identifying a label of the optical packet by using the optical trigger pulse generator according to any one of claims 1 to 4.

Images