JP2020010202A - Transimpedance amplifier circuit - Google Patents

Abstract

To reduce an interval period from a termination of one burst optical signal to a starting of the next burst optical signal.SOLUTION: A transimpedance amplifier circuit 11 is an amplification circuit that converts an input current Iapd according to an intermissive burst optical signal into a differential signal Vout and outputs the resultant signal. This transimpedance amplifier circuit 11 comprises: a TIA core part 14 that converts a current signal Iin into a voltage signal Vtia; a feedback control circuit 16 that generates a bypass current Iaoc1; a differential amplifier circuit 17 that generates the differential signal Vout in accordance with a difference between a voltage signal Vtia and a reference voltage signal Vref; and a detection circuit 19 that detects a starting and a termination of a burst optical signal. The detection circuit 19 detects the termination of the burst optical signal on the basis of a peak value of an inverse phase signal Voutn and a peak value of a positive phase signal Voutp, and switches a time constant of the feedback control circuit 16 from a first time constant to a second time constant smaller than the first time constant.SELECTED DRAWING: Figure 3

Description

  The present invention relates to a transimpedance amplifier circuit.

  In EPON (Ethernet (registered trademark) Passive Optical Network) which is an optical access system, a transimpedance amplifier circuit is used for an optical receiver of an OLT (Optical Line Terminal). The optical receiver of the OLT receives burst optical signals from a plurality of optical network units (ONUs) by time division multiplexing (TDM). Since the plurality of ONUs are provided at various distances from the OLT, the transmission path loss from each ONU to the OLT differs depending on the distance. Therefore, for example, the signal strength from the ONU located relatively close to the OLT is large, and the signal strength from the ONU located far from the OLT is small. In this way, burst optical signals having various signal strengths are input to the OLT optical receiver. For this reason, the transimpedance amplifier circuit included in the OLT optical receiver includes a feedback control circuit that extracts a bypass current from a current signal corresponding to the burst optical signal so as to receive burst optical signals having various signal strengths ( For example, see Patent Documents 1 to 4.

JP 2010-213128 A JP 2012-60436 A JP 2012-10107 A WO 2016/035374

  Further, the transimpedance amplifier circuit included in the OLT optical receiver can respond to intermittent burst optical signals at high speed, and the feedback control circuit that controls the amplification operation even if signals of the same code are continuously input is stable. (Having the same code continuous tolerance) is desired. For this reason, for example, in the feedback control circuit described in Patent Document 1, the time constant related to the response speed of the feedback control circuit in the initial stage when the input of the burst optical signal is started is such that the output voltage of the amplifier is quickly stabilized (averaging). ) Is set to a small value. On the other hand, after the initial stage, the time constant is switched to a value larger than the time constant at the initial stage in order to maintain the same code continuity tolerance. Further, after the end of one burst optical signal, the time constant is switched again to the initial small value.

  Switching of these time constants is performed by comparing the voltage value of the control signal obtained by differentially amplifying the average value of the output voltage of the amplifier and the first reference voltage with the second reference voltage. The amount of change per time of the control signal is determined according to the value of the time constant of the feedback control circuit. For example, when the burst optical signal from a certain ONU ends, the time constant of the feedback control circuit is set to a large value, so that the change per unit time of the control signal with respect to the change per unit time in the output voltage of the amplifier. Is small. Therefore, since it takes time until the control signal (voltage value) becomes larger than the value of the second reference voltage, it takes time before the time constant of the feedback control circuit switches to a small value again in the initial stage. As a result, since it takes time for the average value of the output voltage of the amplifier to return to the initial state, the interval period from the end of one burst optical signal input for a predetermined period to the start of the next burst optical signal becomes longer. I will.

  According to the present invention, there is provided a transimpedance amplifier circuit capable of shortening an interval period from the end of one burst optical signal to the start of the next burst optical signal.

  A transimpedance amplifier circuit according to one aspect of the present invention converts an input current generated by a light receiving element in response to an intermittent burst optical signal into a differential signal including a positive-phase signal and a negative-phase signal and outputs the differential signal. It is an amplifier circuit. The transimpedance amplifier includes a single-ended amplifier that converts a current signal into a voltage signal, a first feedback control circuit that has a time constant and generates a bypass current with a response speed adjusted by the time constant, A differential amplifier circuit that generates a differential signal according to the difference between the signal and the reference voltage signal, and a detection circuit that detects the start and end of the burst optical signal based on the differential signal. The first feedback control circuit generates a bypass current according to a difference between the voltage signal and the reference voltage signal, and generates a current signal by subtracting the bypass current from the input current. The detection circuit detects the end of the burst optical signal based on the first peak value that is the peak value of the positive-phase signal and the second peak value that is the peak value of the negative-phase signal, and detects the end of the burst optical signal. At this time, the time constant of the first feedback control circuit is switched from the first time constant to a second time constant smaller than the first time constant for a predetermined period.

  According to the present invention, it is possible to shorten the interval period from the end of one burst optical signal to the start of the next burst optical signal.

FIG. 1A is a block diagram illustrating a PON communication system. FIG. 1B is a schematic diagram illustrating a time change of an optical signal input to the station-side communication device. FIG. 2 is a diagram illustrating a configuration example of the receiving unit illustrated in FIG. FIG. 3 is a circuit diagram illustrating a transimpedance amplifier circuit according to one embodiment. FIG. 4 is a diagram showing an example of the feedback control circuit shown in FIG. FIG. 5 is a circuit diagram showing the detection circuit shown in FIG. FIG. 6A is a circuit diagram showing the differential peak holding circuit shown in FIG. FIG. 6B is a circuit diagram showing the threshold value generation circuit shown in FIG. FIG. 7A is a circuit diagram showing the single-phase peak holding circuit shown in FIG. FIG. 7B is a circuit diagram showing the threshold value generation circuit shown in FIG. FIG. 8A is a circuit diagram showing the edge detection circuit shown in FIG. FIG. 8B is a timing chart showing the operation of the edge detection circuit shown in FIG. FIG. 9A is a circuit diagram showing the switch signal generation circuit shown in FIG. FIG. 9B is a timing chart showing the operation of the switch signal generation circuit shown in FIG. FIG. 10 is a state transition diagram showing the state transition of the feedback control by the feedback control circuit. FIG. 11 is a diagram illustrating a calculation result of an ideal response when a plurality of high-pass filters are connected in series. FIG. 12 is a diagram illustrating a simulation result. FIG. 13 is an enlarged view of a part of the simulation result shown in FIG. FIG. 14 is an enlarged view of a part of the simulation result shown in FIG. FIG. 15 is an enlarged view of a part of the simulation result shown in FIG. FIG. 16 is an enlarged view of a part of the simulation result shown in FIG. FIG. 17 is a diagram illustrating another simulation result. FIG. 18 is an enlarged view of a part of the simulation result shown in FIG. FIG. 19 is a circuit diagram illustrating a feedback control circuit provided in a transimpedance amplifier circuit according to a modification.

[Description of Embodiment of the Present Invention]
First, the contents of the embodiment of the present invention will be listed and described.

  A transimpedance amplifier circuit according to one aspect of the present invention converts an input current generated by a light receiving element in response to an intermittent burst optical signal into a differential signal including a positive-phase signal and a negative-phase signal and outputs the differential signal. It is an amplifier circuit. The transimpedance amplifier includes a single-ended amplifier that converts a current signal into a voltage signal, a first feedback control circuit that has a time constant and generates a bypass current with a response speed adjusted by the time constant, A differential amplifier circuit that generates a differential signal according to the difference between the signal and the reference voltage signal, and a detection circuit that detects the start and end of the burst optical signal based on the differential signal. The first feedback control circuit generates a bypass current according to a difference between the voltage signal and the reference voltage signal, and generates a current signal by subtracting the bypass current from the input current. The detection circuit detects the end of the burst optical signal based on the first peak value that is the peak value of the positive-phase signal and the second peak value that is the peak value of the negative-phase signal, and detects the end of the burst optical signal. At this time, the time constant of the first feedback control circuit is switched from the first time constant to a second time constant smaller than the first time constant for a predetermined period.

  In this transimpedance amplifier circuit, the end of the burst optical signal is detected based on the first peak value of the positive-phase signal and the second peak value of the negative-phase signal, so that the time constant of the first feedback control circuit becomes the first constant. The time constant is switched to a second time constant smaller than the first time constant. While the burst optical signal is being input, a current signal obtained by subtracting the bypass current from the input current corresponding to the burst optical signal is input to the single-ended amplifier, and the voltage converted by the single-ended amplifier is output. A differential signal including a positive-phase signal and a negative-phase signal is generated by the differential amplifier circuit according to the difference between the signal and the reference voltage signal. Since the high-level signal and the low-level signal are mixed in the burst optical signal, the first peak value of the positive-phase signal and the second peak value of the negative-phase signal are substantially constant while the burst optical signal is input. , And the difference between the first peak value and the second peak value is substantially constant. On the other hand, immediately after the end of the burst optical signal, the DC bypass current generated by the first feedback control circuit remains, and only the remaining bypass current becomes an input signal of the single-ended amplifier circuit. Therefore, after the end of the burst optical signal, a state in which the voltage value of one of the positive-phase signal and the negative-phase signal is higher than the voltage value of the other signal continues. For example, when the detection circuit detects the first and second peak values based on the charging voltages of the capacitors according to the positive-phase signal and the negative-phase signal, the voltage value of one signal is higher than the voltage value of the other signal. When the state continues, the difference between the first peak value and the second peak value becomes larger than the difference during the input of the burst optical signal. Therefore, the difference between the first peak value and the second peak value changes with the end of the burst optical signal, so that the end of the burst optical signal is detected based on the first peak value and the second peak value. be able to. Thus, the time constant of the first feedback control circuit is switched to the second time constant smaller than the first time constant immediately after the end of the burst optical signal, and the time until the first feedback control circuit returns to the initial state is shortened. . As a result, it is possible to shorten the interval period from the end of one burst optical signal to the start of the next burst optical signal.

  The detection circuit may include a single-phase peak holding circuit that detects a second peak value, and a first threshold value generation circuit that generates a first threshold value according to the first peak value. When the peak value becomes larger than the first threshold value, the end of the burst optical signal may be detected.

  In this case, the end of the burst optical signal is detected immediately after the end of the burst optical signal by setting the first threshold so that the second peak value becomes larger than the first threshold immediately after the end of the burst optical signal. It is possible to do.

  The detection circuit detects the start of the burst optical signal based on the third peak value, which is the peak value of the differential signal, and the average voltage peak value of the differential signal. During the period, the time constant of the first feedback control circuit may be switched from the first time constant to the second time constant.

  In this case, since the start of the burst optical signal is detected based on the third peak value and the average voltage peak value of the differential signal, the time constant of the first feedback control circuit is smaller than the first time constant. Can be switched to a constant. As a result, it is possible to shorten the time from when the burst optical signal starts to when the value of the bypass current converges.

  The detection circuit may include a differential peak holding circuit that detects a third peak value, and a second threshold value generation circuit that generates a second threshold value in accordance with the average voltage peak value. The start of the burst optical signal may be detected when the peak value becomes larger than the second threshold value.

  In this case, the third peak value of the differential signal increases immediately after the start of the input of the burst optical signal. As a result, the start of the burst optical signal is detected immediately after the start of the burst optical signal by setting the second threshold so that the third peak value becomes larger than the second threshold immediately after the start of the burst optical signal. It becomes possible.

  The burst optical signal may include a preamble signal and a payload signal following the preamble signal, and the predetermined period may be shorter than the period of the preamble signal.

  In this case, by the end of the input of the preamble signal, the time constant of the first feedback control circuit is returned to the first time constant larger than the second time constant, and the same-code continuity tolerance during the input of the payload signal is returned. Can be maintained.

  The transimpedance amplifier circuit may further include a second feedback control circuit that removes a DC offset of the differential signal by performing feedback control on the differential amplifier circuit according to the positive-phase signal and the negative-phase signal.

  In this case, the DC offset generated in the differential amplifier circuit is removed from the differential signal, and the start and end of the burst optical signal can be accurately detected.

[Details of the embodiment of the present invention]
A specific example of the transimpedance amplifier circuit according to the embodiment of the present invention will be described below with reference to the drawings. It should be noted that the present invention is not limited to these examples, but is indicated by the appended claims, and is intended to include all modifications within the meaning and scope equivalent to the appended claims.

  First, a PON communication system will be described as an example of an optical access system according to an embodiment of the present invention. FIG. 1A is a block diagram illustrating a PON communication system. The communication system 1 illustrated in FIG. 1 includes an office-side communication device 2, a plurality of home-side communication devices 3, an optical splitter 4, a communication path L1, and a plurality of communication paths L2.

  The station-side communication device 2 disposed in the office 5 communicates with the home-side communication devices 3 disposed in a plurality of (here, three) homes 6a to 6c via communication paths L1 and L2, which are optical fibers. Connected. In FIG. 1, the illustration of the home-side communication device 3 disposed in the home 6b and the home 6c is omitted. The station-side communication device 2 is, for example, an OLT. The home communication device 3 is, for example, an ONU. The station-side communication device 2 and the optical splitter 4 are connected by one communication path L1. The optical splitter 4 and each home-side communication device 3 are connected via a communication path L2. The optical splitter 4 divides the optical signal output from the station-side communication device 2 to the communication path L1, and outputs an optical signal divided to each communication path L2. The optical signal output from the optical splitter 4 to each communication path L2 is input to the home communication device 3. The optical splitter 4 outputs an optical signal coupled to the communication path L1 by coupling the optical signal output from the home-side communication device 3 to the communication path L2. The optical signal output from the optical splitter 4 to the communication path L1 is input to the station-side communication device 2.

  The station-side communication device 2 includes a control circuit 7, a transmission unit 8, and a reception unit 9. The transmission unit 8 is a transmission circuit that transmits an optical signal to each home-side communication device 3. The receiving unit 9 is a receiving circuit that receives an optical signal from each home communication device 3. The control circuit 7 is a circuit that controls the transmitting unit 8 and the receiving unit 9. The optical signal transmitted from the transmitting unit 8 and the optical signal received by the receiving unit 9 have different wavelengths.

  FIG. 1B is a schematic diagram illustrating a time change of an optical signal input to the station-side communication device. As shown in FIG. 1B, the optical signal input to the receiving unit 9 of the optical line terminal 2 includes intermittent burst optical signals Sa to Sc. In other words, the burst optical signals Sa to Sc are intermittently input to the receiving unit 9. Here, the burst optical signal Sa, the burst optical signal Sb, and the burst optical signal Sc are input to the receiving unit 9 in this order. During the periods Ton1 to Ton3, the burst optical signals Sa to Sc are input to the receiving unit 9. Specifically, during the period Ton1, the burst optical signal Sa output from the home communication device 3 arranged in the home 6a is input to the receiving unit 9. That is, the period from the time when the input of the burst optical signal Sa starts to the time when the input of the burst optical signal Sa ends is equivalent to the period Ton1. During the period Ton2, the burst optical signal Sb output from the home communication device 3 disposed in the home 6b is input to the receiving unit 9. That is, the period from the time when the input of the burst optical signal Sb starts to the time when the input of the burst optical signal Sb ends corresponds to the period Ton2. During the period Ton3, the burst optical signal Sc output from the home communication device 3 disposed in the home 6c is input to the receiving unit 9. That is, the period from the time when the input of the burst optical signal Sc starts to the time when the input of the burst optical signal Sc ends corresponds to the period Ton3.

  Each of the burst optical signals Sa to Sc has a preamble signal and a payload signal following the preamble signal. The preamble signal is a signal for stabilizing an output signal from an amplifier circuit included in the receiving unit 9 while the preamble signal is being input to the receiving unit 9. The period Ts during which the preamble signal included in each of the burst optical signals Sa to Sc is input to the receiving unit 9 is substantially the same. The period Ts is set to a fixed time, for example. The period Ts of the preamble signal is a settling time required until the receiving unit 9 normally receives the burst optical signals Sa to Sc. In other words, the receiving unit 9 needs to complete the preparation so that the payload signal can be normally received during the period Ts (within the settling time).

  In the signal processing unit connected to the subsequent stage of the amplifier circuit in the receiving unit 9, the frequency and phase shifts of the optical signals output from the respective home-side communication devices 3 are corrected, and the payload is reproduced in order to correctly reproduce the optical signals. A signal is used. For this reason, it is desired to realize control (stabilization) of the burst optical signal as fast as possible in the amplifier circuit of the receiving unit 9. That is, a transimpedance amplifier circuit 11 described later, which constitutes the amplifier circuit of the receiving unit 9, appropriately detects the start (rising) and the end (falling) of the burst optical signal, and determines the strength of the burst optical signal during the period Ts. It is desired to eliminate the influence of the above.

  The payload signal is a signal composed of data transmitted from each home communication device 3. The period of the payload signal included in each of the burst optical signals Sa to Sc differs depending on the data amount of each payload signal. Each of the burst optical signals Sa to Sc may include a burst termination signal (BTS) indicating the end of the burst optical signal following the payload signal.

  Each of the burst optical signals Sa to Sc includes high-level and low-level signals. In other words, high-level and low-level signals are mixed in each of the burst optical signals Sa to Sc. For example, a high-level signal is an optical signal having a predetermined amplitude, and a low-level signal is an optical signal whose amplitude is substantially equal to zero. For example, in the preamble signal, a high-level signal and a low-level signal are regularly and alternately repeated. The payload signal is composed of high-level and low-level signals corresponding to data.

  No optical signal is input to the receiving unit 9 during the period Toff1. The period from the time when the input of the burst optical signal Sa ends to the time when the input of the burst optical signal Sb starts corresponds to the period Toff1. During the period Toff2, no optical signal is input to the receiving unit 9. A period from the time when the input of the burst optical signal Sb ends to the time when the input of the burst optical signal Sc starts corresponds to the period Toff2. The periods Toff1 and Toff2 are periods (interval periods) for switching the communication path L2.

  Since the distance between the office-side communication device 2 and each home-side communication device 3 is different from each other, the amplitude of the burst optical signals Sa to Sc output from the home-side communication device 3 and the loss of the optical signal generated in each communication path L2 are , Different from each other. Therefore, the amplitudes of the burst optical signals Sa to Sc input to the receiving unit 9 during the periods Ton1 to Ton3 are different from each other. As described above, optical signals (burst optical signals) output from different homes and having different amplitudes are input to the receiving unit 9 at irregular intervals. In the amplifier circuit of the receiving unit 9, a feedback control circuit is used because the amplitudes of the burst optical signals are different.

  Next, the configuration of the receiving unit 9 of the station side communication device 2 will be described. FIG. 2 is a diagram illustrating a configuration example of the receiving unit illustrated in FIG. The receiving unit 9 is, for example, a 10G-EPON optical receiver. The receiving unit 9 includes a light receiving element 10, a trans-impedance amplifier (TIA) 11, and a limiting amplifier (LIA) 12.

  The light receiving element 10 is an element that converts a burst optical signal input to the receiving unit 9 into an electric signal (current signal). For example, the light receiving element 10 is an avalanche-photodiode (APD). Specifically, the light receiving element 10 generates an input current Iapd including a DC (Direct Current) component as a current signal according to the amplitude of the burst optical signals Sa to Sc. The light receiving element 10 generates an input current Iapd according to the amplitude of the burst optical signals Sa to Sc. For example, as the amplitude of the burst optical signals Sa to Sc increases, the input current Iapd increases. When a low-level signal of the burst optical signals Sa to Sc is input to the receiving unit 9, the light receiving element 10 generates an input current Iapd substantially equal to zero. The light receiving element 10 outputs the generated input current Iapd to the transimpedance amplifier circuit 11. Here, the cathode of the light receiving element 10 is connected to the applied voltage Vapd, and the anode of the light receiving element 10 is connected to the input terminal 11a of the transimpedance amplifier circuit 11 (see FIG. 3).

  The transimpedance amplifying circuit 11 is a circuit that performs impedance conversion and amplifies the input current Iapd and outputs a differential signal Vout that is a voltage signal. The transimpedance amplifier 11 is configured as, for example, an IC (Integrated Circuit). Specifically, the transimpedance amplifier 11 converts the input current Iapd into a differential signal Vout including the positive-phase signal Voutp and the negative-phase signal Voutn, and outputs the differential signal Vout to the limit amplifier 12. The transimpedance amplification circuit 11 performs impedance conversion and amplification at a high amplification factor when the magnitude of the input current Iapd is relatively small, and performs impedance conversion at a low amplification factor when the magnitude of the input current Iapd is relatively large. And amplification. As described above, the transimpedance amplifier circuit 11 controls the gain according to the magnitude of the input current Iapd. The details of the transimpedance amplifier circuit 11 will be described later.

  The differential signal Vout is input to the limit amplifier 12 via the capacitors 13a and 13b. The transimpedance amplifier circuit 11 and the limit amplifier 12 are AC (alternating current) coupled by capacitors 13a and 13b. The capacitance of the capacitors 13a and 13b used for AC coupling is smaller than that of a capacitor mainly used in a receiver for receiving a continuous signal, such as for a trunk system, in order to quickly respond to a burst optical signal. For example, the capacitors 13a and 13b have the same capacitance. The DC component included in the differential signal Vout is removed by the AC coupling. The positive-phase signal Voutp of the differential signal Vout from which the DC component has been removed is input to the limit amplifier 12 as the positive-phase signal Vliap. Is input to the limit amplifier 12.

  The limit amplifier 12 is a circuit that converts differential signals of various intensities into voltage signals of a constant amplitude and outputs the same. The limit amplifier 12 is configured as, for example, an IC. The limit amplifier 12 outputs the voltage values of the positive-phase signal Vliap and the negative-phase signal Vlian, which are input via the capacitors 13a and 13b, in a uniform manner. In other words, the limit amplifier 12 saturates and amplifies the positive-phase signal Vliap and the negative-phase signal Vlian. The constant amplitude voltage signal output from the limit amplifier 12 is input to a CDR (Clock and Data Recovery) (not shown). A predetermined clock signal is extracted from the voltage signal by the CDR, and the waveform is formed by performing the identification and reproduction processing with the clock signal having little jitter.

  Next, details of the transimpedance amplifier circuit 11 will be described. FIG. 3 is a circuit diagram illustrating a transimpedance amplifier circuit according to one embodiment. The transimpedance amplifier circuit 11 shown in FIG. 3 includes a TIA core section 14 (single-ended amplifier circuit), a dummy TIA section 15, a feedback control circuit 16 (first feedback control circuit), and a differential amplifier circuit 17. , A feedback control circuit 18 (second feedback control circuit), and a detection circuit 19.

  The TIA core unit 14 is a single-ended amplifier circuit (transimpedance amplifier) that converts the input current Iapd into a voltage signal Vtia. Specifically, the TIA core unit 14 includes an amplifier 14a and a feedback resistance element 14b, and generates a voltage signal Vtia corresponding to a current signal Iin obtained by subtracting the bypass current Iaoc1 from the input current Iapd. The TIA core unit 14 outputs the voltage signal Vtia to the feedback control circuit 16 and the differential amplifier circuit 17 (differential amplifier 17a). The gain of the TIA core unit 14, which is the ratio of the magnitude of the voltage signal Vtia to the magnitude of the current signal Iin, is determined by the resistance value (transimpedance) of the feedback resistor element 14b.

  The dummy TIA unit 15 is a circuit that generates a reference voltage signal Vref that is a voltage signal. The dummy TIA unit 15 outputs the generated reference voltage signal Vref to the feedback control circuit 16 and the differential amplifier circuit 17 (differential amplifier 17a). The reference voltage signal Vref is used for converting a single voltage signal Vtia into a differential signal Vout in the differential amplifier circuit 17, and has a predetermined voltage value. For example, the voltage value of the reference voltage signal Vref is set to the value of the voltage signal Vtia when the input current Iapd is 0. The dummy TIA unit 15 includes an amplifier 15a and a feedback resistor 15b. For example, the TIA core unit 14 and the dummy TIA unit 15 have the same configuration.

  The feedback control circuit 16 is a circuit that generates the bypass current Iaoc1 according to the difference between the voltage signal Vtia and the reference voltage signal Vref. The voltage signal Vtia and the reference voltage signal Vref are input to the feedback control circuit 16. The feedback control circuit 16 generates the bypass current Iaoc1 by amplifying only the low-frequency component including the DC component included in the voltage signal Vtia. The output terminal of the feedback control circuit 16 is connected to the input terminal of the TIA core unit 14. Since the bypass current Iaoc1 generated by the feedback control circuit 16 flows toward the feedback control circuit 16, the bypass current Iaoc1 is converted from the input current Iapd. Will be deducted. In other words, the feedback control circuit 16 generates the current signal Iin by subtracting the bypass current Iaoc1 from the input current Iapd. Thereby, the DC component included in the input current Iapd is removed by the bypass current Iaoc1, and only the AC component included in the input current Iapd is input to the TIA core unit 14 as the current signal Iin. By subtracting the bypass current Iaoc1 from the input current Iapd, the average value of the voltage signal Vtia matches the reference voltage signal Vref. Thereby, the DC offset generated in the TIA core unit 14 due to the input current Iapd is removed. As a result, even when the intensity of the burst optical signal is strong and the voltage signal corresponding to the input current Iapd is subjected to the amplitude limitation due to the saturation of the differential amplifier 17b described later, the change in the duty ratio of the voltage signal is suppressed. You.

  The details of the feedback control circuit 16 will be described with reference to FIG. FIG. 4 is a diagram showing an example of the feedback control circuit shown in FIG. The input terminal 16a of the feedback control circuit 16 is connected to the output terminal of the dummy TIA unit 15, and the reference voltage signal Vref is input to the input terminal 16a. The input terminal 16b of the feedback control circuit 16 is connected to the output terminal of the TIA core unit 14, and the voltage signal Vtia is input to the input terminal 16b. A bypass current Iaoc1 corresponding to a difference between the voltage signal Vtia and the reference voltage signal Vref is output to an output terminal 16c of the feedback control circuit 16. The feedback control circuit 16 includes a differential amplifier 21, resistance elements 22a, 22b, 23a, 23b, switches 24a, 24b, a capacitor 25, and an OTA 26. In the present embodiment, the resistance element 22a and the resistance element 22b have the same resistance value R1, and the resistance element 23a and the resistance element 23b have the same resistance value R2.

  The reference voltage signal Vref and the voltage signal Vtia are input to the differential amplifier 21 via the input terminals 16a and 16b. One of the output terminals of the differential amplifier 21 (a negative-phase output terminal) is connected to one end of the resistance element 22a and one end of the switch 24a. The other output terminal (positive-phase output terminal) of the differential amplifier 21 is connected to one end of the resistance element 22b and one end of the switch 24b. The other end of the resistance element 22a and the other end of the switch 24a are connected to one end of the resistance element 23a. The other end of the resistance element 22b and the other end of the switch 24b are connected to one end of the resistance element 23b. That is, the resistance element 22a and the switch 24a are connected in parallel with each other, and the resistance element 22b and the switch 24b are connected in parallel with each other. The other end of resistance element 23a is connected to the positive-phase input terminal of OTA 26, and the other end of resistance element 23b is connected to the negative-phase input terminal of OTA 26. The other end of the resistance element 23a and the other end of the resistance element 23b are connected to each other via a capacitor 25. In other words, the capacitor 25 is inserted between the differential inputs of the OTA 26. The output terminal of the OTA 26 constitutes the output terminal 16c of the feedback control circuit 16.

  The differential amplifier 21 generates a differential signal by amplifying a voltage difference between the reference voltage signal Vref and the voltage signal Vtia. The differential amplifier 21 outputs the generated differential signal. The differential signal output from differential amplifier 21 includes a negative-phase signal output from a negative-phase output terminal and a positive-phase signal output from a positive-phase output terminal. The switches 24a and 24b are, for example, transfer gate switches. Switch signals SW and SWB are input from the detection circuit 19 to the switches 24a and 24b. The switch signals SW and SWB are signals including high-level and low-level states. The logical value of the switch signal SW and the logical value of the switch signal SWB are opposite to each other. For example, the switch signal SWB is generated when the switch signal SW is input to an inversion circuit (NOT circuit). The switches 24a and 24b are controlled by the switch signal SW (switch signal SWB). The switches 24a and 24b are switched between an open state (off state) and a closed state (on state) according to the switch signals SW and SWB. Here, when the switch signal SW is at the high level (the switch signal SWB is at the low level), the switches 24a and 24b are maintained in the open state. When the switch signal SW is at the low level (the switch signal SWB is at the high level), the switches 24a and 24b are maintained in the closed state. Details of the switch signals SW and SWB will be described later.

The OTA 26 is an Operational Transconductance Amplifier, and is a circuit that converts a voltage signal into a current signal. The OTA 26 is a circuit having a known configuration. For example, the OTA 26 has a configuration in which a current mirror circuit is added to a differential amplifier circuit. The OTA 26 has a transconductance Gm, and the input / output impedance of the OTA 26 is, for example, infinite. The bypass current Iaoc1 flowing through a single output terminal (output terminal 16c) of the OTA 26 converts the bypass current Iaoc1 to an input differential voltage, which is a difference between the voltage Vinp input to the OTA 26 and the voltage Vinn, as shown in Expression (1). It is obtained by multiplying the conductance Gm. Since the difference between voltage Vinp and voltage Vinn changes according to the integral value of the difference between voltage signal Vtia and reference voltage signal Vref, bypass current Iaoc1 is equal to the integral value of the difference between voltage signal Vtia and reference voltage signal Vref. Generated accordingly. When the input differential voltage (the current value of the bypass current Iaoc1) is a positive value, the bypass current Iaoc1 flows outward from the OTA 26 (the feedback control circuit 16). When the input differential voltage (the current value of the bypass current Iaoc1) is a negative value, the bypass current Iaoc1 flows from outside the feedback control circuit 16 toward the OTA 26 (the feedback control circuit 16). In this case, the bypass current Iaoc1 is subtracted from the input current Iapd.

The feedback control circuit 16 has a time constant and controls the bypass current Iaoc1 at a response speed adjusted by the value of the time constant. The value of the time constant of the feedback control circuit 16 is determined by the open / close states of the switches 24a and 24b and the constants of the circuit elements. The value of the time constant of the feedback control circuit 16 is a time constant τ1 (first time constant) when the switches 24a and 24b are open, and a time constant τ2 (first time constant) when the switches 24a and 24b are closed. 2 time constant). When the switch signal SW is at the low level, the switches 24a and 24b are in the open state, so that the time constant τ1 is, as shown in Expression (2), the resistance value R1 of the resistance elements 22a and 22b and the resistance element 23a and 23a. It is determined by the resistance value R2 of 23b and the capacitance C1 of the capacitor 25. On the other hand, when the switch signal SW is at the high level, the switches 24a and 24b are in the closed state, so that the resistance value of the parallel circuit formed by the switches 24a and 24b and the resistance elements 22a and 22b is substantially equal to zero. Therefore, the time constant τ2 is determined by the resistance value R2 and the capacitance C1, as shown in Expression (3).

  As shown in Expressions (2) and (3), the time constant τ2 is smaller than the time constant τ1. That is, when the time constant of the feedback control circuit 16 is set to the time constant τ2, the response speed at which the feedback control circuit 16 controls the bypass current Iaoc1 is such that the time constant of the feedback control circuit 16 is set to the time constant τ1. Faster than if you were. In other words, when the high-level switch signal SW (the switch signal SW maintained in the high-level state) is input to the feedback control circuit 16, the bypass current Iaoc1 adjusted by the feedback control circuit 16 per unit time The time change is larger than when the low-level switch signal SW (the switch signal SW maintained in the low-level state) is input to the feedback control circuit 16. Note that the change in the time constant of the feedback control circuit 16 does not cause a change in the DC gain of the loop transfer function in the control loop of the feedback control circuit 16, and the switching of the time constant causes the value of the bypass current Iaoc1 to be discontinuous. No. Although the differential amplifier 21 of the feedback control circuit 16 outputs a differential signal, the feedback control circuit 16 includes a single-ended type differential amplifier that outputs a single output signal instead of the differential amplifier 21. You may. In this case, the feedback control circuit 16 may generate the bypass current Iaoc1 as described above.

  The differential amplifier circuit 17 is a circuit that generates a differential signal Vout including a positive-phase signal Voutp and a negative-phase signal Voutn according to a difference (difference) between the voltage signal Vtia and the reference voltage signal Vref. The differential amplifier circuit 17 includes a differential amplifier 17a and a differential amplifier 17b. The differential amplifier 17a generates a differential signal by amplifying a difference between the voltage signal Vtia and the reference voltage signal Vref. The differential amplifier 17a outputs the generated differential signal to the differential amplifier 17b. The differential amplifier 17b generates a differential signal Vout including the positive-phase signal Voutp and the negative-phase signal Voutn by amplifying the differential signal output from the differential amplifier 17a. The differential amplifier circuit 17 (differential amplifier 17b) outputs the differential signal Vout to the limit amplifier 12 via the output terminals 11b and 11c of the transimpedance amplifier circuit 11. Further, the differential amplifier circuit 17 outputs the differential signal Vout to the feedback control circuit 18 and the detection circuit 19.

  The feedback control circuit 18 is a circuit that removes a DC offset (DC offset) included in the differential signal Vout by performing feedback control on the differential amplifier circuit 17. Even if the average value of the voltage signal Vtia and the voltage value of the reference voltage signal Vref match, the difference between the average value of the voltage signal Vtia and the reference voltage signal Vref is amplified by noise generated in the differential amplifier circuit 17. It is amplified by the circuit 17. For this reason, a DC offset may occur in the differential signal Vout. The DC offset included in the differential signal Vout is a difference between the potential of the DC component included in the positive-phase signal Voutp and the potential of the DC component included in the negative-phase signal Voutn. The positive-phase output terminal and the negative-phase output terminal of the feedback control circuit 18 are respectively connected to output resistance elements (not shown) of the differential amplifier 17a. The feedback control circuit 18 operates so that the potential of the DC component included in the positive-phase signal Voutp matches the potential of the DC component included in the negative-phase signal Voutn. Specifically, the feedback control circuit 18 amplifies only the difference between the low-frequency components including the DC component included in the positive-phase signal Voutp and the negative-phase signal Voutn, so that the bypass currents Iaoc2p and Iaoc2n according to the differences are amplified. Generate Since the feedback control circuit 18 has the same configuration as a known feedback control circuit, a detailed description of the configuration of the feedback control circuit 18 is omitted. When the feedback control circuit 18 performs the feedback control, the potential of the DC component included in the differential signal output from the differential amplifier 17a is adjusted, and the average value of the voltage of the differential signal Vout becomes substantially equal to zero.

  The detection circuit 19 is a circuit that detects the start and end of the burst optical signal. The detection circuit 19 monitors the intensity of the burst optical signal based on the differential signal Vout and adjusts the control threshold. The detection circuit 19 is also called a BTG (Burst Timing Generator). The detection circuit 19 outputs a switch signal SW for switching the time constant to the feedback control circuit 16 by detecting the start and end of the burst optical signal. In the present embodiment, the detection circuit 19 detects the start of the burst optical signal based on the peak value Vp1 (third peak value) of the differential signal Vout and the average voltage peak value Vave of the differential signal Vout. Specifically, the detection circuit 19 detects the start of the burst optical signal when the peak value Vp1 becomes larger than the threshold value Vth1 (second threshold value). The threshold value Vth1 is generated according to the average voltage peak value Vave. The detection circuit 19 detects the end of the burst optical signal based on the peak value Vp2 (second peak value) of the negative-phase signal Voutn and the peak value Vp3 (first peak value) of the positive-phase signal Voutp. Specifically, the detection circuit 19 detects the end of the burst optical signal when the peak value Vp2 becomes larger than the threshold value Vth2 (first threshold value). The threshold value Vth2 is generated according to the peak value Vp3.

  The detection circuit 19 outputs the switch signal SW to the feedback control circuit 16. Note that the switch signal SW is set to a low level in an initial state. That is, in the initial state, the time constant of the feedback control circuit 16 is set to the time constant τ1. When detecting the start or end of the burst optical signal, the detection circuit 19 switches the switch signal SW from a low level to a high level. Here, the detection circuit 19 maintains the switch signal SW at a high level for a predetermined period. When the switch signal SW maintained at the high level is input to the feedback control circuit 16, the time constant of the feedback control circuit 16 is switched from the time constant τ1 to the time constant τ2. After a lapse of a predetermined period, the detection circuit 19 switches the switch signal SW from a high level to a low level. As a result, the time constant of the feedback control circuit 16 is switched from the time constant τ2 to the time constant τ1. In the present embodiment, the predetermined period during which the detection circuit 19 keeps outputting the high-level switch signal SW is shorter than the period Ts of the preamble signal. The detection circuit 19 performs the above-described switching operation of the switch signal SW each time the start or end of the burst optical signal is detected. As described above, when detecting the start or end of the burst optical signal, the detection circuit 19 switches the time constant of the feedback control circuit 16 from the time constant τ1 to the time constant τ2 by the switching operation of the switch signal SW.

  Details of the detection circuit 19 will be described with reference to FIGS. FIG. 5 is a circuit diagram showing the detection circuit shown in FIG. FIG. 6A is a circuit diagram showing the differential peak holding circuit shown in FIG. FIG. 6B is a circuit diagram showing the threshold value generation circuit shown in FIG. FIG. 7A is a circuit diagram showing the single-phase peak holding circuit shown in FIG. FIG. 7B is a circuit diagram showing the threshold value generation circuit shown in FIG. FIG. 8A is a circuit diagram showing the edge detection circuit shown in FIG. FIG. 8B is a timing chart showing the operation of the edge detection circuit shown in FIG. FIG. 9A is a circuit diagram showing the switch signal generation circuit shown in FIG. FIG. 9B is a timing chart showing the operation of the switch signal generation circuit shown in FIG.

  The detection circuit 19 shown in FIG. 5 includes a level monitor circuit 31, comparator circuits 32a and 32b, edge detection circuits 33a and 33b, an XOR circuit 34, and a switch signal generation circuit 35. The input terminal 19a of the detection circuit 19 receives the positive-phase signal Voutp, and the input terminal 19b of the detection circuit 19 receives the negative-phase signal Voutn.

  The level monitor circuit 31 is a circuit that monitors the amplitude and the like of the differential signal Vout. Specifically, the level monitor circuit 31 detects the peak value Vp1, the average voltage peak value Vave, the peak value Vp2, and the peak value Vp3. The level monitor circuit 31 generates a threshold Vth1 and a threshold Vth2. The level monitor circuit 31 outputs the peak value Vp1 and the threshold value Vth1 to the comparator circuit 32a, and outputs the peak value Vp2 and the threshold value Vth2 to the comparator circuit 32b. The level monitor circuit 31 includes a differential peak holding circuit 36, a threshold generation circuit 37 (second threshold generation circuit), a single-phase peak holding circuit 38, and a threshold generation circuit 39 (first threshold generation circuit). ing. The positive-phase signal Voutp is input to the differential peak holding circuit 36, the threshold generation circuit 37, and the threshold generation circuit 39 via the input terminal 19a. The negative-phase signal Voutn is input to the differential peak holding circuit 36, the threshold generation circuit 37, and the single-phase peak holding circuit 38 via the input terminal 19b. That is, the differential signal Vout is input to the differential peak holding circuit 36 and the threshold generation circuit 37.

  The differential peak holding circuit 36 is a circuit that detects the peak value Vp1 of the differential signal Vout. The differential peak holding circuit 36 outputs the detected peak value Vp1 to the comparator circuit 32a. As shown in FIG. 6A, the differential peak holding circuit 36 includes transistors 41 and 42, current sources 43 and 44, and a capacitor 45. The base of transistor 41 is connected to input terminal 19a, and the base of transistor 42 is connected to input terminal 19b. The collectors of transistors 41 and 42 are connected to power supply voltage VCC. The emitters of the transistors 41 and 42 are connected to the output terminal 36a. One end of the current source 43 is connected to the output terminal 36a (emitters of the transistors 41 and 42), and the other end of the current source 43 is connected to the ground potential GND. One end of the current source 44 is connected to the power supply voltage VCC, and the other end of the current source 44 is connected to the output terminal 36a (emitters of the transistors 41 and 42). The capacitor 45 is connected to the current source 43 in parallel. That is, one end of the capacitor 45 is connected to the output terminal 36a, and the other end of the capacitor 45 is connected to the ground potential GND. The capacitor 45 has a capacity Ch1. The current source 43 generates a current Ih1, and the current source 44 generates a current Is1. The current sources 43 and 44 allow the currents Ih1 and Is1 to flow in the direction toward the ground potential GND. The charging voltage generated at both ends of the capacitor 45 is output from the differential peak holding circuit 36 as a peak value Vp1.

  In the differential peak holding circuit 36, the capacitor 45 is charged according to an input signal input to the bases of the transistors 41 and 42. When one of the transistors 41 and 42 is on, a charging current flows through the capacitor 45, and charge is stored in the capacitor 45. At this time, the charging voltage generated at both ends of the capacitor 45 corresponds to the peak value Vp1. The current source 43 discharges (discharges) the charge stored in the capacitor 45. The current value of the current Is1 is set smaller than the current value of the current Ih1. Current sources 43 and 44 bias transistors 41 and 42. That is, the operation reference potentials of the transistors 41 and 42 are set by the current sources 43 and 44. Due to the combined current generated by the current sources 43 and 44, a discharge current flows in a direction to discharge the capacitor 45.

  As the amplitude of the differential signal Vout increases, the peak values of the positive-phase signal Voutp and the negative-phase signal Voutn increase. Transistor 41 is turned on when the amplitude of positive-phase signal Voutp is larger than a predetermined value, and transistor 42 is turned on when the amplitude of negative-phase signal Voutn is larger than a predetermined value. The predetermined value is determined, for example, by the voltage value between the base and the emitter when the transistors 41 and 42 change from the off state to the on state. When the amplitude of one of the positive-phase signal Voutp and the negative-phase signal Voutn becomes larger than a predetermined value, the capacitor 45 is charged. When the amplitude of the differential signal Vout stops increasing, the charging current flowing through the capacitor 45 decreases, and the discharging current, which is the difference between the current Ih1 and the current Is1, and the charging current via the transistors 41 and 42 are in an equilibrium state. Thus, the charging voltage of the capacitor 45 is stabilized. The charging voltage at this time corresponds to the peak value Vp1 (peak potential).

  As the amplitude of the differential signal Vout decreases, the charging current flowing through the capacitor 45 further decreases, and discharge from the capacitor 45 starts. At this time, the impedance of the current sources 43 and 44 is very high, and the transistors 41 and 42 are turned off. Therefore, the time constant at the time of discharging the capacitor 45 is larger than the time constant at the time of charging. Therefore, the peak value Vp1 is maintained (held) substantially constant for a predetermined period. The differential signal Vout includes the complementary positive-phase signal Voutp and the negative-phase signal Voutn, and the transistors 41 and 42 are turned on complementarily (alternately). An operation (detection) corresponding to the full-wave rectification of the motion signal Vout is performed. That is, except for the transition section of the data included in the differential signal Vout, the capacitor 45 is almost charged and the peak value Vp1 is maintained.

  The threshold generation circuit 37 is a circuit that generates the threshold Vth1 according to the average voltage of the differential signal Vout. The threshold generation circuit 37 detects the average voltage peak value Vave and outputs a threshold Vth1 for determining the presence / absence (start) of the burst optical signal to the comparator circuit 32a. The average voltage peak value Vave corresponds to the peak value of the average voltage of the differential signal Vout. As shown in FIG. 6B, the threshold generation circuit 37 includes resistance elements 46 and 47, a capacitor 48, transistors 49 and 50, current sources 51 and 52, a resistance element 53, and a capacitor 54. , Is provided. The current source 51 generates a current Ih2, and the current source 52 generates a current Is2. The capacitor 54 has a capacity Ch2.

  In the threshold generation circuit 37, one end of the resistance element 46 is connected to the input terminal 19a, and one end of the resistance element 47 is connected to the input terminal 19b. The other ends of the resistance elements 46 and 47 are connected to each other. The connection point (node) between resistance element 46 and resistance element 47 is connected to ground potential GND via capacitor 48. The connection point between the resistance element 46 and the resistance element 47 is connected to the bases of the transistors 49 and 50. The collectors of transistors 49 and 50 are connected to power supply voltage VCC. The emitters of the transistors 49 and 50 are connected to the ground potential GND via a parallel circuit including a current source 51 and a capacitor 54 connected in parallel with each other. Further, the emitters of the transistors 49 and 50 are connected to the output terminal 37a via the resistance element 53. A current source 52 is connected between the power supply voltage VCC and the output terminal 37a.

  The resistance value of the resistance element 46 and the resistance value of the resistance element 47 are the same as each other. Thereby, the average voltage (common potential) of the differential signal Vout is detected at the connection point between the resistance element 46 and the resistance element 47. The capacitor 48 bypasses common mode high frequency noise. The threshold generation circuit 37 maintains the peak value of the average voltage of the differential signal Vout. As a result, the average voltage peak value Vave of the differential signal Vout is detected as the potential of the emitters of the transistors 49 and 50 (the voltage across the capacitor 54). When the current Is2 flows from the current source 52 to the resistance element 53, the emitter potentials of the transistors 49 and 50 are DC-offset by the voltage Vs2 across the resistance element 53. That is, the threshold value Vth1 is generated by adding the voltage Vs2 generated at both ends of the resistance element 53 to the average voltage peak value Vave (potential of the emitters of the transistors 49 and 50). In other words, the threshold value Vth1 is generated by adding the voltage Vs2 to the peak value of the average voltage of the differential signal Vout.

  In the present embodiment, the transistors 41 and 42 of the differential peak holding circuit 36 and the transistors 49 and 50 of the threshold generation circuit 37 are all configured by transistors of the same size. The current sources 43 and 51 are configured such that the current values of the current Ih1 and the current Ih2 are the same. The current sources 44 and 52 are configured such that the current values of the current Is1 and the current Is2 are the same. By setting these currents Ih1, Ih2, Is1, Is2, the collector current densities of the transistors 41, 42 become the same. Also, in order to match the filtering effects on noise, the capacitors 45 and 54 may be configured such that the capacitances Ch1 and Ch2 are the same.

  The single-phase peak holding circuit 38 is a circuit that detects the peak value Vp2 (corresponding to the bottom value of the burst optical signal) of the negative-phase signal Voutn. The single-phase peak holding circuit 38 outputs the peak value Vp2 to the comparator circuit 32b. As shown in FIG. 7A, the single-phase peak holding circuit 38 includes a transistor 55, current sources 56 and 57, and a capacitor 58. The current source 56 generates a current Ih3, and the current source 57 generates a current Is3. The capacitor 58 has a capacity Ch3. The base of the transistor 55 is connected to the input terminal 19b, and the negative-phase signal Voutn is input to the base of the transistor 55. The collector of transistor 55 is connected to power supply voltage VCC. The emitter of the transistor 55 is connected to the output terminal 38a. The emitter of the transistor 55 is connected to the ground potential GND via a parallel circuit including a current source 56 and a capacitor 58 connected in parallel with each other. A current source 57 is connected between the power supply voltage VCC and the output terminal 38a.

  In the single-phase peak holding circuit 38, the capacitor 58 is charged by the negative-phase signal Voutn. Similarly to the differential peak holding circuit 36, when the amplitude of the negative phase signal Voutn becomes larger than a predetermined value, the transistor 55 is turned on, and the electric charge is stored in the capacitor 58. The current source 56 discharges (discharges) the charge stored in the capacitor 58. The current value of the current Is3 is set smaller than the current value of the current Ih3. Current sources 56 and 57 bias transistor 55. That is, the reference of the operating voltage of the transistor 55 is set by the current sources 56 and 57. The discharge current flows in the direction in which the capacitor 58 is discharged by the combined current generated by the current sources 56 and 57.

  As the amplitude of the negative-phase signal Voutn increases, the peak value of the negative-phase signal Voutn increases. The transistor 55 is turned on when the amplitude of the negative-phase signal Voutn becomes larger than a predetermined value. When the transistor 55 is turned on, a charging current flows through the capacitor 58, and the capacitor 58 is charged. When the amplitude of the negative-phase signal Voutn stops increasing, the charging current flowing through the capacitor 58 decreases, and the discharging current obtained by subtracting the current Is3 from the current Ih3 and the charging current via the transistor 55 are balanced. This stabilizes the voltage across capacitor 58. The voltage across capacitor 58 at this time corresponds to peak value Vp2 of negative-phase signal Voutn. As the amplitude of the negative-phase signal Voutn further decreases, the charging current flowing through the capacitor 58 further decreases, and discharge from the capacitor 58 starts. At this time, the impedance of the current sources 56 and 57 is very high, and the time constant at the time of discharging the capacitor 58 is larger than the time constant at the time of charging. Therefore, the peak value Vp2 is maintained substantially constant for a predetermined period. Unlike the differential peak holding circuit 36, the signal input to the single-phase peak holding circuit 38 is a single-phase opposite-phase signal Voutn, and therefore, the single-phase peak holding circuit 38 supports half-wave rectification of the differential signal Vout. Operation (detection).

  The threshold generation circuit 39 is a circuit that generates a threshold Vth2 according to the peak value Vp3. The threshold generation circuit 39 detects the peak value Vp3 and outputs the threshold Vth2 to the comparator circuit 32b. As shown in FIG. 7B, the threshold generation circuit 39 includes a transistor 59, current sources 60 and 61, a resistance element 62, and a capacitor 63. The current source 60 generates a current Ih4, and the current source 61 generates a current Is4. The capacitor 63 has a capacity Ch4. The threshold generation circuit 39 has the same configuration as the single-phase peak holding circuit 38, except for the resistance element 62. The base of the transistor 59 is connected to the input terminal 19a, and the base of the transistor 59 receives the positive-phase signal Voutp. The collector of transistor 59 is connected to power supply voltage VCC. The emitter of the transistor 59 is connected to the ground potential GND via a parallel circuit including a current source 60 and a capacitor 63 connected in parallel with each other. The emitter of the transistor 59 is connected to the output terminal 39a via the resistance element 62. A current source 61 is connected between the power supply voltage VCC and the output terminal 39a.

  The threshold generation circuit 39 performs an operation of maintaining a peak value for a single-phase input signal, similarly to the single-phase peak holding circuit 38, and the peak value Vp 3 is detected as the emitter potential of the transistor 49. When the current Is4 flows through the resistance element 62, a potential that is DC-offset by a voltage Vs4 generated at both ends of the resistance element 62 with respect to the emitter potential is generated as the threshold value Vth2. The threshold generation circuit 39 is a circuit for detecting a peak value in the same manner as the single-phase peak holding circuit 38, but since the input signal is the positive-phase signal Voutp, the positive- The voltage value of the signal Voutp decreases, and the threshold value Vth2 decreases at a higher speed from a high voltage value as compared with the peak value Vp2.

  In the present embodiment, the transistor 55 of the single-phase peak holding circuit 38 and the transistor 59 of the threshold generation circuit 39 are configured by transistors of the same size. Current sources 56 and 60 are configured such that current values of current Ih3 and current Ih4 are the same. The current sources 57 and 61 are configured such that the current values of the current Is3 and the current Is4 are the same. By setting the current values of the currents Ih3, Ih4, Is3, and Is4, the collector current densities of the transistors 55 and 59 become equal to each other. Further, in order to match the filtering effect on noise, the capacitors 58 and 63 may be configured such that the capacitance Ch3 of the capacitor 58 and the capacitance Ch4 of the capacitor 63 are equal to each other.

  Furthermore, the capacitances Ch3 and Ch4 may be set so that a malfunction due to discharge does not occur when a burst optical signal having a continuous signal of the same sign is input. For example, in a communication system in which a signal transmission speed is 10 Gb / s (bps; bits per second), the capacities Ch3 and Ch4 are set so as to satisfy the same code continuity tolerance with 72 bits as the set number of bits. In a communication system having a transmission rate of 10 Gb / s, 100 psec is assigned to each symbol signal. In this case, in the capacitors 58 and 63, the single-phase peak holding circuit 38 and the threshold value are set so that the discharge is performed with a discharge time constant having a value at least about twice as long as the signal period corresponding to the set number of bits. The generation circuit 39 may be configured. For example, the discharge time constant when the capacitors 58 and 63 discharge may be set to 14.4 nsec or more.

  On the other hand, in detecting the end of the burst optical signal, if the discharge time constant is too large, the detection may be delayed. The discharge time constant in the single-phase peak holding circuit 38 is obtained by dividing the value obtained by subtracting the current value of the current Is3 from the current value of the current Ih3 by the capacitance Ch3, assuming that the impedances of the current sources 56 and 57 are infinite. Depends on the value obtained. The discharge time constants of the differential peak holding circuit 36, the threshold generation circuit 37, and the threshold generation circuit 39 are determined in the same manner as the single-phase peak holding circuit 38. For example, assuming that the value obtained by subtracting the current value of the current Is3 from the current value of the current Ih3 is 10 μA and the capacitance Ch3 is 10 pF, assuming that the signal amplitude is 100 mV, the voltage change at 14.4 nsec is 14 .4 mV, which corresponds to a 14.4% decrease in amplitude. The voltage change at 100 nsec is 100 mV, and the amplitude decreases to 0%. Therefore, the delay time affecting the detection of the end of the burst optical signal is at most about 100 nsec.

  In detecting the start of the burst optical signal, the edge at the start of the burst optical signal is detected by comparing the peak value Vp1 with the threshold value Vth1. The peak value Vp1 is a voltage obtained by full-wave rectifying the differential signal Vout, and can respond to the start of the burst optical signal at high speed. On the other hand, if the peak value Vp1 is used to detect the fall at the end of the burst optical signal, the response of the peak value Vp1 is changed to a change in the residual DC offset where the feedback control circuit 16 performs the low-speed feedback control. Since the tracking is performed with a delay of the time constant, the response time of the peak value Vp1 is delayed. On the other hand, the fall at the end of the burst optical signal is detected by comparing the peak value Vp2 with the threshold value Vth2. Since the peak value Vp2 and the threshold value Vth2 are voltages obtained by half-wave rectification of the differential signal Vout, the discharge capacity for holding the peak value (peak hold discharge capacity) is incorrect when the same sign continues. If the threshold value Vth2 is reduced within a range where detection (malfunction) does not occur, the threshold value Vth2 drops faster than the peak value Vp2, and the falling edge at the end of the burst optical signal is detected.

  In the transimpedance amplifier circuit 11 of the present embodiment, when the burst optical signal is input to the receiving unit 9, the input current Iapd including both the high level and the low level is input, and the reference voltage signal Vref and the voltage signal Vtia are input. And a difference is generated between Therefore, a differential signal Vout having a predetermined amplitude is generated, and the peak value Vp1 of the differential signal Vout increases. Before the input of the burst optical signal (interval period), the peak value Vp1 and the average voltage peak value Vave are substantially the same. Since the threshold value Vth2 is obtained by adding the voltage Vs2 to the average voltage peak value Vave, the peak value Vp1 is smaller than the threshold value Vth2. On the other hand, when the burst optical signal is started (input), the threshold value generation circuit 37 detects the peak value of the average voltage of the differential signal Vout as the average voltage peak value Vave, so that the peak value Vp1 becomes the average voltage peak value Vave. (Substantially equal to twice the average voltage peak value Vave). Therefore, immediately after the start of the burst optical signal, the start of the burst optical signal is detected by setting the threshold value Vth1 so that the peak value Vp1 exceeds the threshold value Vth1 corresponding to the average voltage peak value Vave. For example, the threshold value Vth1 is set by measuring the peak value Vp1 and the average voltage peak value Vave at the start of the burst optical signal in advance.

  When the burst optical signal ends, the input current Iapd becomes low level, and only the remaining bypass current Iaoc1 is input to the TIA core unit 14 as the current signal Iin. Therefore, after the end of the burst optical signal, the state where the amplitude of the negative-phase signal Voutn is higher than the amplitude of the positive-phase signal Voutp continues. At this time, the positive-phase signal Voutp is at a low level. While the burst optical signal is being input, the amplitude of the negative-phase signal Voutn and the amplitude of the positive-phase signal Voutp are substantially the same, so that the peak value Vp2 and the peak value Vp3 are substantially the same. Since the threshold value Vth2 is obtained by adding the voltage Vs4 to the peak value Vp3, the peak value Vp2 is smaller than the threshold value Vth2. On the other hand, when the burst optical signal ends, the positive-phase signal Voutp goes to a low level state, so that the charge of the capacitor 63 is discharged in the threshold generation circuit 39, and the peak value Vp3 decreases. In the single-phase peak holding circuit 38, the peak value Vp2 is maintained substantially constant. As a result, the peak value Vp2 becomes larger than the peak value Vp3. Therefore, immediately after the end of the burst optical signal, the end of the burst optical signal is detected by setting the threshold value Vth2 so that the peak value Vp2 exceeds the threshold value Vth2 corresponding to the peak value Vp3. For example, the threshold value Vth2 is set by measuring peak values Vp2 and Vp3 at the end of the burst optical signal in advance.

  The comparator circuits 32a and 32b are, for example, comparator circuits having hysteresis characteristics. Specifically, the comparator circuit 32a compares the peak value Vp1 with the threshold value Vth1, and outputs a signal SD in either a high level or a low level to the edge detection circuit 33a according to the comparison result. When the peak value Vp1 is equal to or smaller than the threshold value Vth1, the comparator circuit 32a outputs a low-level signal SD. On the other hand, when the peak value Vp1 is larger than the threshold value Vth1, the comparator circuit 32a outputs a high-level signal SD.

  The comparator circuit 32b compares the peak value Vp2 with the threshold value Vth2, and outputs a signal LOS in either a high level or a low level to the edge detection circuit 33b according to the comparison result. When the peak value Vp2 is equal to or smaller than the threshold value Vth2, the comparator circuit 32b outputs a low-level signal LOS. On the other hand, when the peak value Vp2 is larger than the threshold value Vth2, a high-level signal LOS is output. For example, the hysteresis amounts of the comparator circuits 32a and 32b are set to values that can prevent erroneous detection due to noise generated by the modulation state of the input signal in the level monitor circuit 31.

  The edge detection circuits 33a and 33b are circuits that detect a change in a logical value in each of the signal SD and the signal LOS. Specifically, the edge detection circuit 33a detects a switch (rising edge) from a low level to a high level in the signal SD. When detecting the switching, the edge detection circuit 33a outputs a high-level signal RE to the XOR circuit 34 for a predetermined period. The edge detection circuit 33b detects switching of the signal LOS from a low level to a high level. The edge detection circuit 33b outputs a high-level signal FE to the XOR circuit 34 for a predetermined period when switching is detected. The edge detection circuits 33a and 33b output low-level signals RE and FE to the XOR circuit 34 when switching is not detected. The edge detection circuits 33a and 33b have the same configuration as each other.

  As shown in FIG. 8A, the edge detection circuits 33a and 33b include a resistance element 65, a capacitor 66, an inversion circuit 67, and an AND circuit 68. Signals SD and LOS are input to input terminals 64a of the edge detection circuits 33a and 33b. The input terminal 64a is connected to one end of the resistance element 65 and one input terminal of the AND circuit 68. The other end of the resistance element 65 is connected to the ground potential via the capacitor 66 and to the input terminal of the inversion circuit 67. The output terminal of the inverting circuit 67 is connected to the other input terminal of the AND circuit 68. In the edge detection circuits 33a and 33b, a delay circuit is configured by the resistance element 65 and the capacitor 66.

  FIG. 8B shows a time change (timing chart) of the signals SD and LOS, the delay signal Va, the inverted signal Vb, and the signals RE and FE. As shown in FIG. 8B, a delay signal Va obtained by delaying the signals SD and LOS by the time constant τd1 is input to the input terminal of the inversion circuit 67. The time constant τd1 is obtained by multiplying the resistance value of the resistance element 65 by the capacitance of the capacitor 66. An inverted signal Vb obtained by inverting the logic value of the delay signal Va by the inverting circuit 67 is output from the output terminal of the inverting circuit 67. The AND circuit 68 calculates the logical product of the signals SD and LOS and the inverted signal Vb, and outputs the calculation result as the signals RE and FE from the output terminal 64b. The signals RE and FE output from the output terminal 64b are at a high level only when the signals SD and LOS as input signals are switched from a low level to a high level. As described above, the rising edges of the signals SD and LOS are detected by the edge detection circuits 33a and 33b, and the edge detection circuits 33a and 33b output the high-level signals RE and FE having a pulse width substantially equal to the time constant τd1. I do. The pulse widths of the high-level signals RE and FE are shorter than the pulse widths of the high-level signals SD and LOS. The pulse width corresponds to a period during which the signals RE, FE, SD, and LOS are maintained at a high level.

  The XOR circuit 34 calculates the exclusive OR of the signal RE and the signal FE, and outputs the calculation result to the switch signal generation circuit 35 as the signal TRIG. Specifically, when one of the states of the signal RE and the signal FE is at a high level, a high-level signal TRIG is output. Note that the pulse width of the high-level signal TRIG is substantially equal to the pulse width of the high-level signals RE and FE. When both the signal RE and the signal FE are in one of a high level and a low level, the XOR circuit 34 outputs a low level signal TRIG. Note that the detection circuit 19 may include, instead of the XOR circuit 34, an OR circuit that calculates the logical sum of the signal RE and the signal FE.

  FIG. 9A shows a switch signal generation circuit 35 that generates a switch signal SW. The switch signal generation circuit 35 is a circuit that generates a switch signal SW based on the signal TRIG. The switch signal generation circuit 35 outputs the switch signal SW to the feedback control circuit 16. The switch signal generation circuit 35 is also called a One Shot Timer circuit. As shown in FIG. 9A, the switch signal generation circuit 35 includes an RS flip-flop circuit 69, a MOS transistor 70, a resistor 71, and a capacitor 72. The RS flip-flop circuit 69 has two NOR circuits 69a and 69b.

  The input terminal 35a of the switch signal generation circuit 35 is connected to the S (set) terminal of the RS flip-flop circuit 69. That is, the signal TRIG is input to the S terminal of the RS flip-flop circuit 69. The S terminal is connected to one input terminal of the NOR circuit 69a. One input terminal of the NOR circuit 69b is connected to the Q terminal of the RS flip-flop circuit 69, which is the output terminal of the NOR circuit 69a. The QB terminal of the RS flip-flop circuit 69, which is the output terminal of the NOR circuit 69b, is connected to the other input terminal of the NOR circuit 69a. The QB terminal is connected to the gate terminal of the MOS transistor 70. The MOS transistor 70 is an N-type (n-channel) MOS transistor. The source terminal of MOS transistor 70 is connected to the ground potential. The drain terminal of MOS transistor 70 is connected to power supply voltage VDD via resistance element 71 (pull-up resistance element) and to ground potential via capacitor 72.

  The ethical value of the signal output from the QB terminal is inverted at the drain terminal of the MOS transistor 70. The signal whose logic value is inverted at the drain terminal of the MOS transistor 70 is delayed by a time constant τd2 determined by the product of the resistance value of the resistance element 71 and the capacitance of the capacitor 72, thereby generating a delay signal Vd2. The drain terminal of the MOS transistor 70 is connected to the R (reset) terminal of the RS flip-flop circuit 69, and the delay signal Vd2 is input to the other input terminal of the NOR circuit 69a. The Q terminal of the RS flip-flop circuit 69 is connected to the output terminal 19c.

  FIG. 9B shows a time change (timing chart) of the TRIG signal, the signal at the QB terminal, the delay signal Vd2, and the switch signal SW. As shown in FIG. 9B, when a high-level signal TRIG is input to the input terminal 35a, the signal (switch signal SW) at the Q terminal (output terminal 19c) of the RS flip-flop circuit 69 becomes low. Change from level to high level. At this time, the MOS transistor 70 changes from the on state to the off state, and the delay signal Vd2 slowly rises from a low level to a high level at a speed defined by a time constant τd2. When the voltage value of the increased delay signal Vd2 exceeds a predetermined value, a reset instruction is input to one input terminal (R terminal) of the NOR circuit 69a. As a result, the signal at the Q terminal changes from the high level to the low level. When the reset instruction is input to the R terminal, the signal at the QB terminal returns to the initial high level. That is, the switch signal SW output from the output terminal 19c changes from the high level to the low level.

  Thus, the reset operation is performed by the time corresponding to the rising speed of the voltage value of the delay signal Vd2 (reset signal). Since the reset operation is reliably performed after a lapse of a predetermined time from the rising edge of the signal TRIG, the signal TRIG returns to a low level before the logical value of the reset signal becomes valid (the reset signal is asserted). Is desired. For this purpose, the edge detection circuits 33a and 33b and the switch signal generation circuit 35 may be configured such that the time constant τd2 is larger than the time constant τd1. Further, the time constant τd2 is set according to the period Ts (settling time) of the preamble signal of the burst optical signal. In the present embodiment, the time constant τd2 is set such that the period during which the switch signal SW output from the input terminal 19a is at a high level is shorter than the period Ts.

  In the detection circuit 19, when the peak value Vp1 exceeds the threshold value Vth1, the comparator circuit 32a outputs a high-level signal SD to the edge detection circuit 33a. Similarly, when the peak value Vp2 exceeds the threshold value Vth2, the comparator circuit 32b outputs a high-level signal LOS to the edge detection circuit 33b. The edge detection circuits 33a and 33b output the high-level signals RE and FE to the XOR circuit 34 when detecting that the signals SD and LOS have switched from low level to high level. The signal TRIG output from the XOR circuit 34 changes from a low level to a high level when one of the signal RE and the signal FE is at a high level. Then, when the signal TRIG is at the high level, the switch signal generation circuit 35 changes the switch signal SW from the low level to the high level. The pulse width during the period when the switch signal SW is at the high level is larger than the pulse width during the period when the signal TRIG is at the high level. As described above, when either the peak value Vp1 exceeds the threshold value Vth1 or the peak value Vp2 exceeds the threshold value Vth2, the switch signal SW maintained at the high level for a predetermined period is output from the detection circuit 19. Output to the feedback control circuit 16.

  Next, the state transition of the feedback control by the feedback control circuit 16 will be described with reference to FIG. FIG. 10 is a state transition diagram showing the state transition of the feedback control by the feedback control circuit. The state of the feedback control by the feedback control circuit 16 includes a state I in which the feedback control is performed at a low speed and a state II in which the feedback control is performed at a high speed. In the following, the feedback control by the feedback control circuit 16 is simply referred to as “feedback control” for convenience of description. In state I, the time constant of the feedback control circuit 16 is set to the time constant τ1, and in state II, the time constant of the feedback control circuit 16 is set to the time constant τ2. In the initial state where the transimpedance amplifier 11 starts operating, the state of feedback control is state I. When the burst optical signal is not input, the switch signal SW is at the low level, so that the state of the feedback control is maintained at the state I. When the burst optical signal is input to the receiver 9, a differential signal Vout in which the current signal Iin is converted by the transimpedance amplifier circuit 11 is generated. When the detection circuit 19 detects that the peak value Vp1 has become larger than the threshold value Vth1, the signal TRIG changes from a low level to a high level, and the switch signal generation circuit 35 starts operating. As a result, the switch signal SW output from the detection circuit 19 (switch signal generation circuit 35) changes from the low level to the high level, and the state of the feedback control changes to the state II. That is, the time constant of the feedback control circuit 16 switches from the time constant τ1 to the time constant τ2.

  While the switch signal SW is at the high level, the state of the feedback control is maintained in the state II. The time during which the feedback control is maintained in the state II is determined by the time constant τd2 of the switch signal generation circuit 35. Immediately after the start of the burst optical signal, since the current signal Iin contains a DC component, the feedback control is performed at high speed so as to remove the DC offset (DC component) included in the voltage signal Vtia. When the switch signal SW changes from the high level to the low level (when the switch signal generation circuit 35 times out), the state of the feedback control changes to the state I. That is, the time constant of the feedback control circuit 16 switches from the time constant τ2 to the time constant τ1. Then, while the payload signal of the burst optical signal is input, the state of the feedback control is maintained at the state I. That is, during the period of the payload signal, the time constant of the feedback control circuit 16 is maintained at the time constant τ1.

  When the burst optical signal ends, the current value of the input current Iapd output from the light receiving element 10 suddenly becomes zero. In the state at the moment when the burst optical signal ends, the state of the feedback control is maintained at the state I, and the residual DC offset, which is the potential difference of the DC component between the voltage signal Vtia and the reference voltage signal Vref, increases. When the detection circuit 19 detects that the peak value Vp2 has become larger than the threshold value Vth2, the signal TRIG changes from a low level to a high level, and the switch signal generation circuit 35 starts operating. As a result, the switch signal SW output from the detection circuit 19 (switch signal generation circuit 35) changes from the low level to the high level, and the state of the feedback control changes to the state II. That is, the time constant of the feedback control circuit 16 switches from the time constant τ1 to the time constant τ2. While the switch signal SW is at the high level, the state of the feedback control is maintained in the state II, and the feedback control circuit 16 performs the feedback control at high speed so as to remove the residual DC offset. When the switch signal SW changes from the high level to the low level (when the switch signal generation circuit 35 times out), the state of the feedback control transits to the state I. That is, the feedback control state returns to the initial state, and the feedback control circuit 16 enters a standby state for receiving the next burst optical signal. In the present embodiment, the period during which the state of feedback control is maintained in state II after the end of the burst optical signal is the same as the period during which the state of feedback control is maintained in state II after the start of the burst optical signal.

  Next, the feedback control circuit 16, the feedback control circuit 18, and a method of setting each time constant in the AC coupling between the transimpedance amplifier circuit 11 and the limit amplifier 12 will be described. FIG. 11 is a diagram illustrating a calculation result of an ideal response when a plurality of high-pass filters are connected in series. In FIG. 11, a feedback control circuit 16, a feedback control circuit 18, and three high-pass filters (HPF: High-Pass filter) assuming AC coupling between the transimpedance amplifier circuit 11 and the limit amplifier 12 are continuously arranged. The result of calculating the ideal response when the connection is made as shown in FIG. In the following description, the AC coupling between the transimpedance amplifier circuit 11 and the limit amplifier 12 is simply referred to as “AC coupling”. Each high-pass filter is a first-order high-pass filter, and the transfer function of each high-pass filter is shown as a function of the angular velocity. In FIG. 11, the angular velocity ωaoc1 is the cutoff angular velocity of the feedback control circuit 16, the angular velocity ωaoc2 is the cutoff angular velocity of the feedback control circuit 18, and the angular velocity ωac is the cutoff angular velocity of the AC coupling. For example, the time constant of the feedback control circuit 16 is the reciprocal of the cutoff angular velocity ωaoc1, and the cutoff frequency faoc1 is a value obtained by dividing the cutoff angular velocity ωaoc1 by twice the value of the circular constant π.

  The calculation result shown in FIG. 11 is a calculation result of the step response at a time of 10 μsec. Here, the calculation was performed by setting the cutoff frequency faoc2 to 16 kHz and the cutoff frequency fac to 1.6 MHz. Note that the cutoff frequency fac is set to 1.6 MHz as a lower limit frequency at which a signal having the same code can be transmitted without error, and the cutoff frequency faoc2 is set to a value one hundredth of the set value of the cutoff frequency fac. did. FIG. 11 shows a calculation result of the step response when the cutoff frequency faoc1 is changed N times with respect to the cutoff frequency fac. Here, the calculation results when the cutoff frequency faoc1 is set to 0.005 times, 0.05 times, 0.5 times, and 5 times the cutoff frequency fac are shown. For example, when it is desired to set the settling time to 400 nsec or less, in the step response indicated by the calculation result when N is 0.5, the error does not converge within 10% at 10.4 μsec. In the step response indicated by the calculation result when N is 5, the error is substantially equal to 0 at 10.4 μsec, and the step response converges. Similarly, the step response indicated by the calculation result when N is 0.05 converges at 10.4 μsec. From these facts, when the cut-off frequency faoc2 is set to 16 kHz and the cut-off frequency fac is set to 1.6 MHz, the feedback control circuit 16 sets the target value of the cut-off frequency to 8 MHz in the high-speed feedback control. It is understood that the target value of the cutoff frequency may be set to 80 kHz in the feedback control at a low speed by the control circuit 16.

  Note that the voltages Vs2 and Vs4 for generating the thresholds Vth1 and Vth2 for detecting the start and end of the burst optical signal are the maximum values that the circuit inside the transimpedance amplifier 11 can linearly amplify without saturating the signal. It may be set to a value lower than the value. In some cases, a burst optical signal having a weak signal intensity that does not exceed the threshold is input, and the feedback control state of the feedback control circuit 16 is maintained at a low speed. Even in this case, the DC offset at the input to the limit amplifier 12 is removed within the settling time due to the AC coupling effect between the transimpedance amplifier circuit 11 and the limit amplifier 12.

  Next, the effects of the transimpedance amplifier circuit 11 will be described with reference to FIGS. FIG. 12 is a diagram illustrating a simulation result. 13 to 16 are diagrams in which a part of the simulation result shown in FIG. 12 is enlarged. FIG. 17 is a diagram illustrating another simulation result. FIG. 18 is an enlarged view of a part of the simulation result shown in FIG. FIG. 12 shows a time change of each current value and each voltage value and a state of feedback control of the feedback control circuit 16. The simulation results show that the cutoff frequency faoc2 is 16 kHz, the cutoff frequency fac is 1.6 MHz, the value obtained by subtracting the current value of the current Is1 from the current value of the current Ih1 is 20 μA, and the current value of the current Is2 is subtracted from the current value of the current Ih2. It is a calculation result when the calculated value is set to 20 μA. The capacitances Ch1 and Ch2 are 10 pF, the value obtained by subtracting the current value of the current Is3 from the current value of the current Ih3 is 10 μA, the value obtained by subtracting the current value of the current Is4 from the current value of the current Ih4 is 10 μA, and the capacitances Ch3 and Ch4 are The calculation was performed by setting 5 pF, the time constant τd1 to 50 nsec, the time constant τd2 to 500 nsec, and the voltages Vs2 and Vs4 to 50 mV.

  12 to 18, the change in the feedback control state of the feedback control circuit 16, the input current Iapd, the voltage signal Vtia and the reference voltage signal Vref, the positive-phase signal Voutp and the negative-phase signal Voutn, and the positive-phase signal Time changes of Vliap and the negative-phase signal Vlian, peak value Vp1 and threshold value Vth1, peak value Vp2 and threshold value Vth2, signal TRIG and switch signal SW, bypass current Iaoc1, and bypass currents Iaoc2p and Iaoc2n are shown. ing. In this simulation, when the time t is 100 nsec, the first burst optical signal having a strong signal intensity is input first, and when the time t is 2 μsec, the first burst optical signal ends. The average value of the input current Iapd by the first burst optical signal is 1 mA. Then, when the time t is 2.3 μsec, a second burst optical signal having a weak signal intensity is input. The average value of the input current Iapd by the second burst optical signal is 20 μA. The interval (interval period) between the first burst optical signal and the second burst optical signal is set to 300 nsec. The input current Iapd according to the burst optical signal which is a modulation signal imitates a signal having the same sign. Here, the input current Iapd is a signal (repetition signal) in which a signal having a continuous positive sign (high level) and a signal having a continuous negative sign (low level) are alternately repeated in a cycle of 10 nsec.

  With the input of the first burst optical signal, the peak value Vp1 increases and exceeds the threshold value Vth1. Thus, the start of the burst optical signal is detected, and the signal TRIG changes from the low level to the high level. The high-level signal TRIG is input to the switch signal generation circuit 35, so that a high-level switch signal SW having a pulse width of about 500 nsec is generated. At this time, the state of the feedback control of the feedback control circuit 16 transits to the state II, and the feedback control circuit 16 performs the feedback control at high speed. The bypass current Iaoc1 has sufficiently converged at about 200 nsec from the time when the start of the burst optical signal is detected, and when the time t is around 600 nsec (before the first burst optical signal shifts from the preamble signal to the payload signal). 2), the feedback control state of the feedback control circuit 16 continuously transitions to the state I.

  When the time t is 2 μsec, the first burst optical signal ends. Thereafter, the amplitude (potential) of the positive-phase signal Voutp of the differential signal Vout gradually decreases, and the amplitude (potential) of the negative-phase signal Voutn gradually increases. These amplitude decreases and increases are caused by the difference between the high-frequency gain and the low-frequency gain of the differential amplifier 17b due to AC coupling, because the modulation signal component disappears with the end of the burst optical signal. Since the peak value Vp1 is a result of detection corresponding to full-wave rectification of the differential signal Vout, the peak value Vp1 follows the fluctuation of the potential in the negative-phase signal Voutn. The peak value Vp2 also changes with time similarly to the peak value Vp1, and the potential of the threshold value Vth2 changes with time in a decreasing direction. When the peak value Vp2 becomes larger than the threshold value Vth2, the end of the burst optical signal is detected. As a result, the signal TRIG changes from a low level to a high level. The high-level signal TRIG is input to the switch signal generation circuit 35, so that a high-level switch signal SW having a pulse width of about 500 nsec is generated. At this time, the feedback control of the feedback control circuit 16 transits to the state II, and the feedback control circuit 16 performs the feedback control at high speed. The bypass current Iaoc1 has sufficiently converged in about 200 nsec from the time when the end of the burst optical signal is detected. .

  When the time t is 2.3 μsec, the second burst optical signal has started. At this time, the state of the feedback control of the feedback control circuit 16 is the state II due to the influence of the state transition of the feedback control accompanying the end of the first burst optical signal. The feedback control of the feedback control circuit 16 is started in a high-speed control state. Before the feedback control is started, the residual DC offset of the voltage signal Vtia (differential signal Vout) has been sufficiently removed.

  FIG. 13 is an enlarged view (timing chart) of part A of FIG. FIG. 13 shows a time change of the current value and the like when the time t is around 0.1 μsec. With the input of the first burst optical signal, the peak value Vp1 increases, and the peak value Vp1 becomes larger than the threshold value Vth1. As a result, the switch signal SW (signal TRIG) changes from low level to high level. Immediately after the start of the first burst optical signal, the average value of the voltage signal Vtia has a lower value than the reference voltage signal Vref. Then, as the DC component of the input current Iapd starts to be extracted by the feedback control of the feedback control circuit 16, the average value of the voltage signal Vtia starts to increase and is gradually controlled to match the value of the reference voltage signal Vref. You. During this time, the bypass currents Iaoc2p and Iaoc2n hardly change because the time constant is long.

  FIG. 14 is an enlarged view (timing chart) of the portion B in FIG. FIG. 14 shows a temporal change in the current value and the like when the time t is around 0.57 μsec. In addition, the influence of the continuation of the same sign on the discharge of the capacitor in the differential peak holding circuit 36 and the like is shown. When the time t is about 572 nsec, the switch signal SW changes from the high level to the low level, and the feedback control of the feedback control circuit 16 changes from the state II to the state I. At this time, no particular disturbance occurs in the bypass current Iaoc1 generated by the feedback control circuit 16. Since the peak value Vp1 and the threshold value Vth1 are full-wave rectified voltages of the differential signal Vout, no noise due to the repetitive signal is seen at the peak value Vp1 and the threshold value Vth1. On the other hand, the peak value Vp2 and the threshold value Vth2 are voltages obtained by half-wave rectification of the differential signal Vout, and the discharge time constants in the single-phase peak holding circuit 38 and the threshold value generating circuit 39 are set short. Values fluctuate slightly. The erroneous detection (malfunction) caused by the slight fluctuation is suppressed by the hysteresis characteristics of the comparator circuits 32a and 32b of the detection circuit 19.

  FIG. 15 is an enlarged view (timing chart) of a portion C in FIG. FIG. 15 shows a time change of the current value and the like when the time t is around 2.0 μsec. When the time t is 2.0 μsec, the first burst optical signal has ended. When the burst optical signal ends, the current value of the input current Iapd becomes substantially equal to 0, so that the voltage signal Vtia changes so as to decrease from a high amplitude. Since the voltage signal Vtia is subjected to the amplitude limitation in the differential amplifier circuit 17, the positive-phase signal Voutp changes from a low level and the negative-phase signal Voutn changes from a high level. The amplitude (potential) of the positive-phase signal Voutp gradually decreases, and the amplitude (potential) of the negative-phase signal Voutn gradually increases. This change is caused by the disappearance of the modulation signal component in the input current Iapd with the end of the burst optical signal, and is caused by the difference between the high frequency gain and the low frequency gain in the differential amplifier 17b. By reducing the cutoff frequency fac due to the AC coupling, the change in the amplitudes of the positive-phase signal Voutp and the negative-phase signal Voutn becomes slow. Even if this change becomes slow, a burst end condition occurs because the amplitude of the positive-phase signal Voutp becomes smaller with the end of the burst optical signal, so that the threshold Vth2 is not greatly affected by the cutoff frequency fac. When the peak value Vp2 becomes larger than the threshold value Vth2 around the time when 14 nsec has elapsed after the end of the burst optical signal, the end of the burst optical signal is detected, and the switch signal SW (signal TRIG) is changed from low level to high level. Change to a level. As a result, the feedback control of the feedback control circuit 16 switches from the low speed control (state I) to the high speed control (state II). With this switching, the temporal changes of the bypass current Iaoc1 and the voltage signal Vtia become large, and the positive-phase signal Voutp and the negative-phase signal Voutn subsequently converge.

  FIG. 16 is an enlarged view (timing chart) of a portion D in FIG. FIG. 16 shows a temporal change of the current value and the like when the time t is around 2.3 μsec. When the time t is 2.3 μsec, the second burst optical signal is input, but before that, the residual DC offset between the positive-phase signal Voutp and the negative-phase signal Voutn has sufficiently converged (removed). Have been). Since the intensity of the second burst optical signal is weak, the peak value Vp1 does not become larger than the threshold value Vth1, and the state transition of the feedback control of the feedback control circuit 16 does not occur. Even if the signal strength is weak, a DC offset occurs in the transimpedance amplifier circuit 11 due to the second burst optical signal. However, since the feedback control of the feedback control circuit 16 is maintained in the state II due to the state transition accompanying the end of the first burst optical signal, the feedback control circuit 16 starts the feedback control in the high-speed control state. Therefore, the DC offset generated between the positive-phase signal Voutp and the negative-phase signal Voutn is immediately eliminated by the feedback control in the high-speed state by the feedback control circuit 16. The positive-phase signal Vliap and the negative-phase signal Vlian almost follow the responses of the positive-phase signal Voutp and the negative-phase signal Voutn.

  The setting of another simulation result shown in FIG. 17 mainly differs from the setting of the simulation result shown in FIG. 12 at the input timing of the second burst optical signal. In this simulation, when the time t is 2.8 μsec, the second burst optical signal is input. Changes in each voltage value and the like due to the input of the first burst optical signal are similar to the simulation results shown in FIG. FIG. 18 is an enlarged view (timing chart) of a portion E in FIG. FIG. 18 shows a temporal change of the current value and the like when the time t is around 2.8 μsec. The second burst optical signal is input at a timing that is not affected by the switching of the feedback control of the feedback control circuit 16 due to the end of the first burst optical signal. Since the intensity of the second burst optical signal is weak, the peak value Vp1 does not become larger than the threshold value Vth1, and the state transition of the feedback control of the feedback control circuit 16 does not occur. Even if the signal strength is weak, a DC offset occurs in the transimpedance amplifier circuit 11 due to the second burst optical signal. Unlike the simulation result shown in FIG. 12, the second burst optical signal is input when the feedback control state of the feedback control circuit 16 is the state I (the state of the low-speed control). Therefore, the DC offset generated between the positive-phase signal Voutp and the negative-phase signal Voutn does not immediately converge. On the other hand, since the cutoff frequency fac is set to 1.6 MHz, the DC offset is removed from the positive-phase signal Vliap and the negative-phase signal Vlian within the settling time. Further, since the intensity of the second burst optical signal is weak and the signal is not saturated in the transimpedance amplifier circuit 11, the occurrence of duty distortion in the differential signal Vout is suppressed.

  As described above, in the transimpedance amplifier circuit 11, the end of the burst optical signal is detected based on the peak value Vp3 of the positive-phase signal Voutp and the peak value Vp2 of the negative-phase signal Voutn. The time constant is switched from the time constant τ1 to the time constant τ2 smaller than the time constant τ1. While the burst optical signal is being input, the current signal Iin obtained by subtracting the bypass current Iaoc1 from the input current Iapd corresponding to the burst optical signal is input to the TIA core unit 14 and converted by the TIA core unit 14. The differential amplifier circuit 17 generates a differential signal Vout including the positive-phase signal Voutp and the negative-phase signal Voutn according to the difference between the voltage signal Vtia and the reference voltage signal Vref. Since the high-level signal and the low-level signal are mixed in the burst optical signal, while the burst optical signal is being input, each of the peak value Vp2 and the peak value Vp3 is kept substantially constant, and the peak value Vp2 The difference from the peak value Vp3 is substantially constant.

  On the other hand, immediately after the end of the burst optical signal, the DC bypass current Iaoc1 generated by the feedback control circuit 16 remains, and only the remaining bypass current Iaoc1 becomes an input signal of the TIA core unit 14. Therefore, after the end of the burst optical signal, the state where the voltage value of the negative-phase signal Voutn is higher than the voltage value of the positive-phase signal Voutp continues. In the detection circuit 19, the peak values Vp2 and Vp3 are detected based on the charging voltages of the capacitors 58 and 63 corresponding to the negative-phase signal Voutn and the positive-phase signal Voutp, respectively. When the voltage value of the negative-phase signal Voutn continues to be higher than the voltage value of the positive-phase signal Voutp, the difference between the peak value Vp2 and the peak value Vp3 is larger than the difference during the input of the burst optical signal. It is getting bigger. Accordingly, the difference between the peak value Vp2 and the peak value Vp3 changes with the end of the burst optical signal, so that the end of the burst optical signal can be detected based on the peak value Vp2 and the peak value Vp3. As a result, the time constant of the feedback control circuit 16 is switched to the time constant τ2 smaller than the time constant τ1 immediately after the end of the burst optical signal, and the time until the feedback control circuit 16 returns to the initial state is shortened. As a result, it is possible to shorten the interval period from the end of one burst optical signal to the start of the next burst optical signal.

  Immediately after the end of the burst optical signal, the threshold value Vth2 is set so that the peak value Vp2 becomes larger than the threshold value Vth2 corresponding to the peak value Vp3. Can be detected.

  The start of the burst optical signal is detected based on the peak value Vp1 and the average voltage peak value Vave of the differential signal Vout, so that the time constant of the feedback control circuit 16 is smaller than the time constant τ1 from the time constant τ1. Can be switched to As a result, it is possible to shorten the time from the start of the burst optical signal until the value of the bypass current Iaoc1 converges.

  Immediately after the burst optical signal is input, the peak value Vp1 increases. As a result, the threshold value Vth1 is set such that the peak value Vp1 becomes larger than the threshold value Vth1 corresponding to the average voltage peak value Vave immediately after the start of the burst optical signal. The start of the optical signal can be detected.

  The period in which the time constant of the feedback control circuit 16 is switched from the time constant τ1 to the time constant τ2 is shorter than the period Ts of the preamble signal. As a result, by the end of the input of the preamble signal, the time constant of the feedback control circuit 16 is returned from the time constant τ2 to the time constant τ1, so that the same code continuity immunity can be maintained while the payload signal is input. Becomes

  The DC offset generated in the differential amplifier circuit 17 is removed from the differential signal Vout by the feedback control circuit 18. As a result, the peak values Vp1, Vp2, Vp3 and the average voltage peak value Vave for detecting the start and end of the burst optical signal are accurately detected, so that the end and start of the burst optical signal can be accurately detected. It becomes possible.

  The receiving unit 9 including the transimpedance amplification circuit 11 receives the payload signal of the burst optical signal without being strongly restricted by the strength of the burst optical signal or the order (length of the interval period) between the burst optical signals. be able to. Further, since the DC component included in the input current Iapd is suppressed in the TIA core unit 14, the gain of the TIA core unit 14 is increased, and the reception sensitivity of the transimpedance amplifier circuit 11 (the receiving unit 9) can be increased. Becomes Further, since the reception sensitivity is increased, even if the signal is saturated inside the transimpedance amplifier circuit 11, the change in the duty ratio of the signal is suppressed. Further, the communication efficiency can be improved by shortening the interval period.

  Note that the transimpedance amplifier circuit according to the present invention is not limited to the above embodiment.

  FIG. 19 is a circuit diagram illustrating a feedback control circuit provided in a transimpedance amplifier circuit according to a modification. As shown in FIG. 19, the feedback control circuit 16A provided in the transimpedance amplifier circuit according to the modification includes OTAs 81, 82, 83 and a capacitor 84. Like the feedback control circuit 16, the feedback control circuit 16A generates the bypass current Iaoc1 according to the difference between the voltage signal Vtia and the reference voltage signal Vref, and subtracts the bypass current Iaoc1 from the input current Iapd to generate the current signal Iin. Generate The reference voltage signal Vref is input to the input terminal 16d of the feedback control circuit 16A, and the voltage signal Vtia is input to the input terminal 16e of the feedback control circuit 16A. The positive phase input terminal of OTA81 is connected to input terminal 16d, and the negative phase input terminal of OTA81 is connected to input terminal 16e. The negative phase output terminal of the OTA 81, one end of the capacitor 84, the positive phase input terminal of the OTA 82, and the positive phase input terminal of the OTA 83 (the negative phase output terminal of the OTA 82) are connected to each other. The positive phase output terminal of the OTA 81, the other end of the capacitor 84, the negative phase input terminal of the OTA 82, and the negative phase input terminal of the OTA 83 (the positive phase output terminal of the OTA 82) are connected to each other. Capacitor 84 has capacitance C2.

The OTA 81 has a transconductance Gm, and the ideal input / output impedance of the OTA 81 is infinite. As shown in Expressions (4) and (5), an output differential current obtained by multiplying the input differential voltage by a half of the transconductance Gm flows through the differential output terminal of the OTA 81. Note that the voltage Vinp in the equations (4) and (5) corresponds to the reference voltage signal Vref, and the voltage Vinn corresponds to the voltage signal Vtia. The input differential voltage is a voltage obtained by subtracting the voltage signal Vtia from the reference voltage signal Vref. When the input differential voltage is positive, current Ioutp flows externally from the positive-phase output terminal of OTA81, and current Ioutn flows externally from the negative-phase output terminal of OTA81. On the other hand, when the input differential voltage is negative, the current Ioutp flows from the outside to the positive-phase output terminal of the OTA 81, and the current Ioutn flows from the reverse-phase output terminal of the OTA 81 to the outside.

The OTA 82 has a transconductance Gm like the OTA 81. The positive phase input terminal of the OTA 82 is connected to the negative phase output terminal of the OTA 82, and the negative phase input terminal of the OTA 82 is connected to the positive phase output terminal of the OTA 82. That is, the input and output terminals of the OTA 82 are connected in negative feedback. Thereby, the OTA 82 functions as a resistor having a resistance value (1 / Gm) equivalently. The OTA 83 is a single-ended operational transconductance amplifier that receives a differential voltage (the voltage across the capacitor 84) and outputs an output current to a single output terminal 16f, similarly to the OTA 26 of the feedback control circuit 16. The OTA 83 has a transconductance Gmo. The transfer function of the feedback control circuit 16A is determined by the transconductances Gm and Gmo and the capacitance C2 of the capacitor 84 as shown in Expression (6). Here, the current Iout corresponds to the bypass current Iaoc1, and the voltage Vin corresponds to the difference between the voltage signal Vtia and the reference voltage signal Vref.

  The switch signals SW are input to the OTAs 81 and 82. Note that the switch signal SW is output from the detection circuit 19. The value of the transconductance Gm of the OTAs 81 and 82 changes according to the logical value of the switch signal SW. Here, the amounts of change in the transconductances Gm of the OTAs 81 and 82 according to the logical value of the switch signal SW are set to be the same as each other. For example, when the switch signal SW is at the low level, the value of the transconductance Gm is set to the transconductance gm1, and when the switch signal SW is at the high level, the value of the transconductance Gm is set to the transconductance (A × gm1). Is set to Note that the constant A is larger than 1. As described above, if the amount of change according to the logical value of the switch signal SW is the same in the OTAs 81 and 82, the DC gain of the feedback control circuit 16A shown in Expression (6) does not change.

The time constants τ1 and τ2, which are the values of the time constant of the feedback control circuit 16A, are determined as shown in Expressions (7) and (8). Here, when the switch signal SW is at a low level, the time constant of the feedback control circuit 16A is a time constant τ1, and when the switch signal SW is at a high level, the time constant of the feedback control circuit 16A is a time constant τ2. It is. As described above, when the switch signal SW is at the high level, the time constant of the feedback control circuit 16A is (1 / A) times that when the switch signal SW is at the low level, and the feedback control circuit 16A operates at high speed. Perform feedback control. Note that the feedback control circuit 16A may include a single-ended OTA instead of the OTA 81. In this case, the time constant of the feedback control circuit 16A may be switched in the same manner as described above.

  In the transimpedance amplifier including the feedback control circuit 16A, the same effect as in the transimpedance amplifier 11 can be obtained.

  The transimpedance amplifier circuit may include a feedback control circuit having a configuration different from the feedback control circuits 16 and 16A, instead of the feedback control circuits 16 and 16A. The feedback control circuit included in the transimpedance amplifier circuit only needs to have a configuration capable of generating the bypass current Iaoc1 according to the difference between the reference voltage signal Vref and the voltage signal Vtia and switching the time constant of the feedback control circuit.

  The transimpedance amplifier circuit may have a configuration in which the voltage value of the positive-phase signal Voutp continues to be higher than the voltage value of the negative-phase signal Voutn due to the residual bypass current Iaoc1 after the end of the burst optical signal. In this case, the detection circuit 19 may be configured to detect the end of the burst optical signal by making the peak value Vp3 of the positive-phase signal Voutp larger than a threshold value corresponding to the peak value Vp2 of the negative-phase signal Voutn. Good.

  The detection circuit 19 may be configured to switch the switch signal SW from a low level to a high level only when detecting the end of the burst optical signal. That is, at the start of the burst optical signal, the time constants of the feedback control circuits 16 and 16A are not switched, and the detection circuit 19 changes the time constant of the feedback control circuits 16 and 16A only when the end of the burst optical signal is detected. The time constant τ1 may be switched to the time constant τ2.

  The circuit for detecting the peak values Vp1, Vp2, Vp3 and the average voltage peak value Vave is not limited to the differential peak holding circuit 36 or the like. The detection circuit 19 only needs to include a circuit capable of detecting the peak values Vp1, Vp2, Vp3 and the average voltage peak value Vave and generating the threshold values Vth1, Vth2.

  The differential amplifier circuit 17 may be configured by one differential amplifier, or may be configured by three or more differential amplifiers. An output terminal of the feedback control circuit 18 may be connected to an input terminal of the differential amplifier 17a. The transimpedance amplifier does not need to include the feedback control circuit 18.

  11: Transimpedance amplifier circuit, 14: TIA core unit, 15: Dummy TIA unit, 16: Feedback control circuit, 17: Differential amplifier circuit, 18: Feedback control circuit, 19: Detection circuit, 36: Differential peak holding circuit , 37: threshold generation circuit, 38: single-phase peak holding circuit, 39: threshold generation circuit.

Claims (6)

A transimpedance amplifier circuit that converts an input current generated by a light receiving element according to an intermittent burst optical signal into a differential signal including a positive-phase signal and a negative-phase signal and outputs the differential signal,
A single-ended amplifier circuit for converting a current signal to a voltage signal,
A first feedback control circuit having a time constant and generating a bypass current at a response speed adjusted by the time constant;
A differential amplifier circuit that generates the differential signal according to a difference between the voltage signal and a reference voltage signal,
A detection circuit for detecting the start and end of the burst optical signal based on the differential signal,
The first feedback control circuit generates the bypass current according to a difference between the voltage signal and the reference voltage signal, and generates the current signal by subtracting the bypass current from the input current.
The detection circuit detects the end of the burst optical signal based on a first peak value that is a peak value of the positive-phase signal and a second peak value that is a peak value of the negative-phase signal, and detects the end of the burst optical signal. When the end of the time is detected, the time constant of the first feedback control circuit is switched from a first time constant to a second time constant smaller than the first time constant for a predetermined period,
Transimpedance amplifier circuit. The detection circuit includes a single-phase peak holding circuit that detects the second peak value, and a first threshold value generation circuit that generates a first threshold value according to the first peak value,
The detection circuit detects the end of the burst optical signal when the second peak value is larger than the first threshold value.
The transimpedance amplifier circuit according to claim 1. The detection circuit detects the start of the burst optical signal based on a third peak value that is a peak value of the differential signal and an average voltage peak value of the differential signal, and detects the start of the burst optical signal. And switching the time constant of the first feedback control circuit from the first time constant to the second time constant during the predetermined period.
The transimpedance amplifier circuit according to claim 1. The detection circuit includes a differential peak holding circuit that detects the third peak value, and a second threshold value generation circuit that generates a second threshold value according to the average voltage peak value,
The detection circuit detects the start of the burst optical signal when the third peak value is larger than the second threshold value.
The transimpedance amplifier circuit according to claim 3. The burst optical signal has a preamble signal and a payload signal following the preamble signal,
The predetermined period is shorter than a period of the preamble signal;
The transimpedance amplifier circuit according to claim 1. A second feedback control circuit that removes a DC offset of the differential signal by performing feedback control on the differential amplifier circuit according to the positive-phase signal and the negative-phase signal,
The transimpedance amplifier circuit according to claim 1.