JP2006345284A - Transmission system and station apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a station apparatus durable for long-range transmissions even using a semiconductor laser having multi-mode light emission spectrum characteristics different in frequency at the slave station side, and an economically low-cost transmission system which standardizes required members to facilitate controlling the entire system. <P>SOLUTION: The system comprises a semiconductor laser having multi-mode light emission spectrum characteristics different in frequency, an optical fiber SMF for connecting a station apparatus OLT to a plurality of subscriber terminals ONU through optical couplers OC to propagate signals from the semiconductor laser, and an adaptive control equalizer AE for making distorted signals near to original signal waveforms by wavelength dispersion at passing through an optical fiber SMF. The semiconductor laser is preferably a Fabry-Perot type semiconductor laser FP type LD. The optical fiber is preferably a single mode fiber SMF. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a transmission system using a Fabry-Perot type semiconductor laser and a single mode fiber, capable of supporting transmission over a long distance exceeding 10 km, and capable of compensating a signal distorted by chromatic dispersion after photoelectric conversion, And the station side device.

  Between a master station OLT (Optical Line Terminal: optical subscriber line terminal equipment) as a station side device and a slave station ONU (Optical Network Unit: optical subscriber line termination equipment) as a plurality of subscriber terminals, There are two-way communication systems using an optical data communication network. Then, a single star network configuration in which the master station OLT and each slave station ONU are radially connected by a single optical fiber has been put into practical use. In this network configuration, the system and the equipment configuration are simplified, but if one slave station ONU occupies one optical fiber and the number of slave station ONUs is N, the optical fiber directly connected from the master station OLT N are necessary, and it is difficult to reduce the price of the system.

  Thus, a PON (Passive Optical Network) system (also referred to as PDS (Passive Double Star)) in which one optical fiber drawn from the master station OLT is shared by a plurality of slave stations ONU has been put into practical use. The PON system is one of low-cost access methods for optical subscribers that have been applied to FTTx such as FTTH (Fiber To The Home) and FTTB (Fiber To The Building).

  The PON system consists of a master optical terminal OLT and a passive optical branching device (hereinafter simply referred to as an optical coupler) that branches and multiplexes signals passively from an input signal without requiring an external power supply. The OC is connected by a single mode fiber (hereinafter referred to simply as an optical fiber) SMF having a single propagation mode. There are usually a plurality of slave station ONUs connected by optical fibers SMF corresponding to the number of slave station ONUs. The master station OLT and the slave station ONU of the N station are based on 1-to-N transmission connected via the optical fiber SMF and the optical coupler OC. Thereby, many slave station ONUs can be allocated to one master station OLT, and the overall equipment cost can be suppressed.

A configuration in which another optical coupler OC is further inserted between the optical coupler OC and the plurality of slave stations ONU may be used.
In the PON system, the downstream optical signal transmitted from the master station OLT to the slave station ONU is distributed to all the slave stations ONU, but the identifier of the logical link is added, and the slave station constituting the logical link is transmitted. Only the ONU can capture the optical signal. The upstream optical signal transmitted from the slave station ONU is assigned a transmission time for each slave station ONU so that the slave station ONUs do not collide with each other.

Furthermore, GE-PON (Gigabit Ethernet-Passive) is a technology that incorporates Ethernet (registered trademark) technology in the PON system, improves connection affinity with many devices, and realizes optical fiber access section communication. Optical Network) system has been put into practical use. (For example, refer nonpatent literature 1).
Also, a configuration of an adaptive control equalizer that corrects intersymbol interference, that is, ISI (Inter Symbol Interference), in a data transmission system is shown (for example, see Non-Patent Document 2).
Oki Technical Review January 2004 / No. 197 Vol.71 No.1 p.100-p.105 Digital filter design Tokai University Press ISBN4-486-00894-4 p.237 ~ p.250

  In a so-called GE-PON system as disclosed in Non-Patent Document 1, an inexpensive Fabry-Perot Laser Diode (hereinafter simply referred to as FP LD) is used for communication of less than 10 km. However, an expensive distributed feedback laser diode (hereinafter simply referred to as a DFB type LD) is used for communication exceeding 10 km.

  The reason why the FP-type LD is not used for communication exceeding 10 km is that the FP-type LD has multi-mode emission spectrum characteristics with slightly different wavelengths, and therefore, the propagation delay shifts depending on the wavelength due to the dispersion characteristics of the optical fiber. The optical signal waveform at the master station OLT is large between the optical signal from the slave station ONU installed near the master station OLT and the optical signal from the slave station ONU grounded at a place far from the master station OLT. Different. The reason why the optical signal waveform is different is that when the optical signal propagates through the optical fiber, the optical signal waveform is reduced due to the shift in the propagation delay due to the wavelength, but this effect increases as the propagation distance increases. . This is because if the signal modulated by the FP type LD is transmitted over a long distance exceeding 10 km through an optical fiber, the waveform interferes with the adjacent symbol and cannot be accurately decoded.

Therefore, in GE-PON, the slave station ONU using the FP type LD is currently limited to optical communication within approximately 10 km.
Therefore, in order to realize long-distance transmission exceeding 10 km, it is necessary to use a DFB type LD having a narrow emission spectrum width.
However, since the DFB type LD is expensive, the cost is very high.

Non-Patent Document 2 discloses a so-called adaptive control equalizer, but it deals with handling when transmitting light of multi-mode emission spectrum characteristics over a long distance and handling when receiving light from a plurality of slave stations ONU. Not disclosed.
Accordingly, an object of the present invention is to provide a station-side apparatus that can withstand long-distance transmission even when a semiconductor laser having multimode emission spectrum characteristics with different wavelengths is used on the slave station side.

  Another object of the present invention is to provide a transmission system that can manage the entire system easily and economically because the necessary members can be unified.

A transmission system for achieving the above object includes a semiconductor laser having multi-mode emission spectrum characteristics with different wavelengths, an optical fiber that propagates a signal from the semiconductor laser, and chromatic dispersion when passing through the optical fiber. And an adaptive control equalizer that approximates the signal distorted by the output signal waveform before passing through the optical fiber.
According to the present invention, the adaptive control equalizer can bring an optical signal in which a square signal is distorted by long-distance transmission closer to the output signal waveform before passing through the optical fiber. Thereby, it is possible to configure a transmission system that enables long-distance optical communication using a semiconductor laser (for example, a Fabry-Perot semiconductor laser) having multi-mode emission spectrum characteristics with different wavelengths.

The semiconductor laser is a Fabry-Perot type semiconductor laser, and the optical fiber connects a station side device and a plurality of subscriber terminals, and combines a plurality of signals from the plurality of subscriber terminals. A single mode fiber connected through a coupler is preferable.
According to this configuration, the semiconductor laser having multi-mode emission spectrum characteristics with different wavelengths can be an inexpensive Fabry-Perot semiconductor laser. The optical fiber having a single propagation mode may be a single mode fiber connected via an optical coupler that combines a plurality of signals from a plurality of subscriber terminals. Thus, the system can be realized economically and easily by using a widely used member.

The single mode fiber has a distance connecting the subscriber terminal and the station side device exceeding 10 km, and a signal passing through the single mode fiber is a baseband signal of 1.25 Gbps or more. (Claim 3).
According to the present invention, the adaptive control equalizer occurs when a semiconductor laser having multi-mode emission spectral characteristics with different wavelengths, such as a Fabry-Perot semiconductor laser, is used for transmission over a long distance exceeding 10 km. Square signal distortion can be compensated. As a result, an inexpensive Fabry-Perot type semiconductor laser can be used in long-distance optical communication exceeding 10 km, which conventionally requires the use of an expensive distributed feedback type semiconductor laser.

The adaptive control equalizer has a plurality of transmission-system memories, and can switch to a memory corresponding to the subscriber terminal in response to switching of a subscriber terminal that performs uplink transmission. ).
According to the present invention, the adaptive control equalizer can store a tap coefficient which is a coefficient suitable for each subscriber terminal for compensating for the distortion of the square signal. Accordingly, in response to switching of the subscriber terminal for uplink transmission, the adaptive control equalizer instantaneously switches to a coefficient corresponding to the subscriber terminal and performs good communication between the subscriber terminal and the network. be able to.

In addition, the adaptive control equalizer may include a memory control unit that performs speculative discovery.
According to the present invention, the adaptive control equalizer can have memory control means for performing speculative discovery. This speculative discovery is to prepare a plurality of initial values for adaptive control in advance and cover these initial values in a plurality of discovery procedures, thereby probabilistically achieving the minimum required equalization even in the initial stage. Do and tell the discovery procedure to succeed. Once the discovery procedure is successful, continuous adaptive control can be performed by means for storing the tap coefficients described above, so that better equalization can be performed.

Further, the station side device of the present invention is distorted by chromatic dispersion when passing through the semiconductor laser having a multimode emission spectrum characteristic with different wavelengths, an optical fiber that propagates a signal from the semiconductor laser, and the optical fiber. And an adaptive control equalizer that approximates the output signal waveform to the output signal waveform before passing through the optical fiber.
According to the present invention, the station side device approximates an optical signal output from a semiconductor laser such as an FP type LD and distorted by chromatic dispersion when propagating through an optical fiber to an output signal waveform before passing through the optical fiber. be able to. And this station side apparatus can be provided in the above-mentioned transmission system.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.
FIG. 1 is a schematic diagram illustrating a configuration example of a PON system in which a master station and a plurality of slave stations are connected via an optical coupler via an optical fiber.
In the PON system, a master station OLT provided in a station building and a slave station ONU provided in a plurality of subscriber houses are connected via an optical fiber SMF and an optical coupler OC.

The slave station ONU includes a network interface for connecting a terminal that enjoys optical network services, such as a personal computer installed in the subscriber's home.
The optical coupler OC is composed of a star coupler that passively branches and multiplexes a signal from an input signal without requiring an external power supply.
The optical fiber connected to the master station OLT, the optical coupler OC, the optical coupler OC, and the slave station ONU uses a single mode fiber composed of one optical fiber SMF. That is, one master station OLT is connected to one optical coupler OC by one trunk optical fiber SMF. One optical coupler OC is connected to M second optical couplers OC (M is a number of 4 in this example) by an optical fiber SMF. The second optical coupler OC is connected to N (N is a number of 8 or less in this example) slave stations ONUs via branch optical fibers SMF. Therefore, a signal transmitted / received by one master station OLT is distributed to a maximum of 32 slave station ONUs by 1 + M optical couplers OC.

  The communication system of the present invention incorporates Gigabit Ethernet (Ethernet is a registered trademark) technology into the PON (Passive Optical Network) technology, and an optical fiber at a baseband speed of 1.25 Gbps. The GE-PON (Gigabit Ethernet-Passive Optical Network) system that realizes the access section communication is adopted.

According to the GE-PON system, communication between the master station OLT and the slave station ONU is performed in units of variable length frames. The frame structure is composed of a GE-PON header including a logical link identifier and data of 64 bytes or more. The maximum size of data is generally about 1530 bytes.
First, a predetermined bridge process is performed in the master station OLT for a downstream frame that enters the master station OLT from a higher network, and a logical link to be relayed is specified. Then, it is transmitted to the optical fiber SMF as an optical signal through the master station OLT. At this time, the master station OLT adds a GE-PON header including a logical link identifier to the frame. The optical signal transmitted to the optical fiber SMF is branched by the optical coupler OC and transmitted to the slave station ONU connected to the optical coupler OC, but only the slave station ONU constituting the logical link takes in a predetermined optical signal, Relay the frame to the home network interface.

  On the other hand, the upstream optical signal includes an upstream frame from each slave station ONU. The upstream optical signal needs to be transmitted so that the optical signals from the respective slave stations ONU do not compete with each other in time. For this purpose, the master station OLT assigns a window (hereinafter simply referred to as a window) during which an upstream optical signal may be transmitted to each slave station ONU and notifies it as a control frame. The slave station ONU to which the window is assigned transmits an upstream optical signal to the assigned window. Therefore, the competition of the upstream optical signal between each slave station ONU is avoided. Each slave station ONU, when given a certain window, may transmit frames continuously as long as it fits in that window. At this time, it is necessary to share a clock between the master station OLT and the slave station ONU. The time adjustment of this clock includes time information in the control frame when the control frame is communicated. Can be done. The master station OLT can receive a burst optical signal including a series of frame signals from each slave station ONU. A burst optical signal is an optical signal sequence of a finite time in which the light emission state is changed by a 1.25 Gbps basepand signal. Before the significant frame sequence, the light emission state and the light reception state are stabilized, and the receiving side Have a synchronization time for recovering the baseband signal (referred to collectively as a synchronization adjustment period (Sync Time) ST), and a unique signal sequence is transmitted during this period. In addition, after a significant frame sequence, there is an adjustment time for quenching.

The optical signal waveform received by the master station OLT differs depending on the characteristics of the optical transmission line such as the characteristics of the light emitting element of the slave station ONU and the length of the optical fiber SMF. Therefore, the master station OLT performs a process described later on the upstream burst optical signal, and then transmits the restored significant frame sequence to the upper network.
FIG. 2 is a schematic diagram showing a basic configuration in the GE-PON system.

The slave station ONU includes an optical transceiver OT that uses an FP-type LD as a light emitting element, and a GE-PON LSI.
The FP-type LD is one of semiconductor lasers, which uses a crystal cleavage plane on two light emission surfaces (light extraction side and emission surface on the monitor photodiode PD side), and transmits light between them. Laser oscillation is realized by reflection. Since this FP-type LD does not have a wavelength selection mechanism and laser light in a plurality of oscillation modes is output as it is, the optical signal transmitted from the slave station ONU has a spectrum that is relatively wide in the 1300 nm band.

  The master station OLT connects an optical transceiver OT that transmits and receives signals to and from the slave station ONU, a GE-PON LSI that transmits and receives signals to and from a higher-level network, and an optical transceiver OT and a GE-PON LSI, and A serializer / deserializer SerDes (hereinafter simply referred to as SerDes) that performs mutual conversion between the signal and the parallel signal.

The optical transceiver OT has an adaptive control equalizer (AE) for equalizing a signal transmitted from the slave station ONU.
Then, the optical transceivers OT of the slave station ONU and the master station OLT emit light strongly for a period corresponding to 1 of a baseband signal, which is an electric signal representing information to be transmitted, and weakly emit light for a period corresponding to 0. Thereby, a rectangular signal of NRZ (Non Return to Zero) can be transmitted at both ends of the optical transceiver OT of each of the slave station ONU and the master station OLT.

Hereinafter, a series of flows of optical signals exchanged between the slave station ONU and the master station OLT will be described.
A WDM (Wavelength Division Multiplexing) filter (hereinafter simply referred to as WDM) of the master station OLT separates an optical signal in the 1300 nm band that is emitted from the slave station ONU and arrives via the optical fiber SMF. Lead to diode PD. The weak current signal detected by the photodiode PD is converted into a weak voltage signal by the transimpedance amplifier TIA, and the weak voltage signal is further amplified by the post amplifier PA.

A signal from the slave station ONU amplified by the post-amplifier PA is input to the adaptive control equalizer AE, equalized to the original signal transmitted from the slave station ONU by a mechanism described later, and sent to SerDes.
Incidentally, the conventional optical transceiver binarizes the output of the post-amplifier PA as it is and sends it to SerDes, which is different from the present invention.

SerDes converts the equalized input signal from clock data recovery and conversion from a serial signal to a parallel signal in code units, and sends a code stream to the GE-PON LSI. The GE-PON LSI sends a frame represented by the sent code stream to a higher-level network (for example, the Internet).
Processing beyond the PCS layer (Physical Coding Sublayer) of GE-PON is performed by GE-PON LSI. The code stream between SerDes and GE-PON LSI is a 10-bit signal of 1000BASE-X, one of the Gigabit Ethernet standards, and the serialized 1.25 Gbps signal is a baseband in which 0 and 1 frequently alternate. Signal.

  On the other hand, the GE-PON LSI adds a GE-PON header to a frame received from an upper network for sending to a slave station ONU, converts the frame into a code stream, and sends the code stream to SerDes. SerDes converts the code stream from a parallel signal to a serial signal and sends it to the LD driver in the optical transceiver. Then, it is converted into a drive current of the DFB type LD by the LD driver, and the light emission of the DFB type LD is controlled.

The monitor photodiode PD has a mechanism for detecting a part of the laser beam and feeding it back to the LD driver in order to stabilize the light emission of the DFB type LD. The WDM guides the optical signal in the 1480 nm band emitted from the DFB LD to the optical fiber SMF, and the optical signal reaches the slave station ONU via the optical fiber SMF and the optical coupler OC.
Thus, since the optical wavelength band used for transmission in the master station OLT and the optical wavelength band used for transmission in the slave station ONU are different, the slave station ONU and the master station OLT can be connected by a single optical fiber SMF. Is possible.

  By the way, in the IEEE802.3ah (IEEE: Institute of Electrical and Electronics Engineers) standard, which is a specification of 1 Gbps using an optical fiber, including the specification of the spectrum width used in GE-PON is 10 km. The 1000BASE-PX10 standard (hereinafter referred to as PX10) for transmission up to 20 km and the 1000BASE-PX20 standard (hereinafter referred to as PX20) for transmission up to 20 km are defined. Considering the waveform distortion caused by the chromatic dispersion of the optical fiber SMF, the spectral width of PX20 is defined narrower than that of PX10.

  The distorted signal may be binarized to a state different from the state of the transmission source. Generally, when the communication between the slave station ONU of PX10 and the master station OLT is connected with an optical fiber SMF exceeding 10 km, the bit error rate BER (Bit Error Rate) required by IEEE802.3ah is 10 ^ (-12) The following quality (^ represents a power multiplier. The same shall apply hereinafter) cannot be achieved. In addition, it is usually difficult for the FP type LD to satisfy the spectrum width specification of PX20, and an optical transceiver in optical communication exceeding 10 km using the FP type LD of the prior art may satisfy the spectrum distribution of PX10. In many cases, the spectral distribution of PX20 is not satisfied.

Hereinafter, the relationship between the spectrum width and the transmission wavelength at the time of long distance transmission in the wavelength of 1300 nm band emitted from the FP type LD will be described.
FIG. 3 is a diagram showing the relationship between the transmission wavelength from 1260 nm to 1360 nm and the allowable range of the spectrum width. (A) of this figure is a figure which shows the relationship between the transmission wavelength in the uplink transmission (1000BASE-PX10-Upsteam) of PX10 for 10 km transmission, and the allowable range of a spectrum width. Moreover, (b) of this figure is a figure which shows the relationship between the tolerance | permissible_range of the spectrum width in the uplink transmission (1000BASE-PX20-Upsteam) of PX20 for 20 km transmission, and a transmission wavelength. In these figures, the vertical axis indicates the allowable spectrum width (rms value), and the horizontal axis indicates the transmission wavelength (unit: nm).

However, if the master station OLT according to the present invention is used, the adaptive control equalizer AE, which will be described later, reshapes the signal into a clean square signal. Can be easily connected in excess of 10 km.
Hereinafter, the adaptive control equalizer AE of the present invention will be described in detail.
In FIG. 2, the adaptive control equalizer AE in the optical transceiver OT of the master station OLT includes an adaptive transversal filter (Transversal Filter) TF integrated with a slicer between the post-amplifier PA and SerDes.

FIG. 4 is a block diagram of the adaptive control equalizer shown in FIG. FIG. 5 is a block diagram of the tap coefficient adaptive control unit shown in FIG.
The adaptive transversal filter TF of the adaptive control equalizer AE includes a known delay element Z ⁻¹ , a multiplier X, an adder Σ, and a tap coefficient control unit T as adaptive control means.
The adaptive transversal filter TF is operated based on a known LMS (Least Mean Square) algorithm. The constant μ here is a system constant multiplied by the multiplier X, and affects the convergence speed and stability. The adaptive control equalizer AE shapes the signal transmitted through the long-distance optical fiber SMF and distorted by chromatic dispersion, and approximates the square signal transmitted by the slave station ONU.

First, the signal input to the adaptive control equalizer AE is shifted in time by the respective delay elements Z- ¹ . The input signals shifted in time are respectively multiplied by the tap coefficient C _i (i = −2, −1,0, + 1, + 2) and combined by the adder Σ ₊ . The combined signal is replaced with a square signal by the slicer, output from the adaptive control equalizer AE, and sent to SerDes.
Further, the error e is obtained by the adder Σ ₋ from the signal output from the adder Σ ₊ and the signal output from the slicer, and μe is multiplied by the multiplier X to output μe. Each tap coefficient control unit T _i can calculate an optimum tap coefficient C _i from μe and x _i by the least square method.

For example, as shown in FIG. 5, in the tap coefficient control unit T _i , μe multiplied by the multiplier X and x _i calculated by the multiplier X _i are multiplied by the multiplier X _T. Then, by adding by the adder sigma ₊ and those calculated by the [mu] e · x _i and delay elements Z ^-1 calculated by the multiplier X _T, new tap coefficients C _i are obtained.
In GE-PON, the uplink signal is time division multiplexed (TDM) based on an instruction from the master station OLT. That is, the slave station ONU sends a burst optical signal to the window designated by the master station OLT. The burst optical signal has a duration of the order of several μs to ms, and several hundred ns to 1 μs of the first bit group is used for the synchronization adjustment period ST. The slave station ONU transmits a signal of a fixed pattern (IDLE sequence) of 20 bits during the synchronization adjustment period ST. The value of μ is determined in advance so that the adaptive control converges within the synchronization adjustment period ST. The state of adaptive control is initialized for each burst and does not carry over to the next burst optical signal.

Here, the initial values of the tap coefficients are C ₀ = 1, C ₋₂ = C ₋₁ = C ₊₁ = C ₊₂ = 0.
The tap coefficients C ₀ , C ₋₂ , C ₋₁ , C ₊₁ and C ₊₂ may be other initial values. In these cases, initial values may be set according to the installation situation and environment.
By the way, if μ is increased in order to speed up convergence, stability of convergence may be impaired. On the other hand, when μ is reduced, the convergence time is generally increased, but the convergence stability is improved. However, in that case, the tap coefficients may not converge within the synchronization adjustment period ST. Therefore, it is preferable that the value of μ can be adjusted.

FIG. 6 is a configuration diagram of another form of the tap coefficient control unit.
The tap coefficient control unit T _i has J t _i , C _i1 , C _i2 ,..., C _ij ,... And C _iJ (j = 1, 2,..., J). The data is stored in or taken out from a specific storage means by the switch SW. Thereby, adaptive control requiring a long convergence time for a plurality of slave station ONUs can be performed by one adaptive control equalizer.

The switch SW is controlled by the switch controller SC. The tap coefficient C _i0 corresponds to the initial value in FIG. The switch control unit SC switches the switch SW so that continuous adaptive control is performed by spinning a plurality of discrete burst optical signals emitted from each slave station ONU. That is, j of tap coefficients C _i1 , C _i2 ,..., C _ij ,... And C _iJ correspond to the slave stations ONU _j (j = 1, 2,..., J). The switch controller SC is informed in advance of window information (from which slave station ONU to when the burst optical signal arrives) from the GE-PON LSI. Here, it is assumed that the burst optical signal from the slave station ONU _j arrives from time Ts to Te.

When the time reaches Ts, the switch control unit SC takes out the tap coefficient C _ij and reflects it in the tap coefficient C _i . Further, when the time reaches Te, the tap coefficient C _i is stored in the tap coefficient C _ij . Such processing is performed for each window. Here, the storage means is provided for each slave station ONU, but the slave stations ONU having similar light emitting element characteristics and transmission path characteristics may be grouped, and the storage means may be assigned to each group.

It should be noted that the reaction time of the switch control unit SC does not need to start at the times Ts and Te, but may start at a time slightly shifted from the times Ts and Te.
By the way, the substation ONU joining procedure in GE-PON is also determined as a plug-and-play method as automatic discovery. Automatic discovery means that the master station OLT opens a logical link with the slave station ONU in response to a request from the slave station ONU, and the master station OLT recognizes the slave station ONU that is newly requesting subscription. Polling.

  First, the mechanism of automatic discovery will be described. The master station OLT notifies a discovery window (a kind of transmission permission) to all the slave station ONUs. A slave station ONU that wishes to join newly transmits a registration request in a random period within the discovery window. The master station OLT receives this registration request, and grants registration permission if the subscription conditions are satisfied. Upon receiving registration permission, the slave station ONU returns registration confirmation and the registration procedure is completed. However, when a plurality of slave stations ONU try to join at the same time, the registration request may collide and the master station OLT may not receive the registration request. Due to the discovery sequence retry and the random delay of the transmission period, it is statistically guaranteed that the registration request can be received without collision.

The master station OLT allocates a normal window so that the slave station ONU can send a burst optical signal including registration confirmation at the stage of granting registration permission to the slave station ONU to be newly joined, Send window information in advance. The normal window is for a specific child station ONU, while the discovery window is generally for an unspecified child station ONU. Such window information is also given to the switch controller SC. The switch control unit SC performs the above-described processing in the corresponding window, but since the slave station ONU is not specified in the discovery window, the switch control unit SC extracts the tap coefficient C _i0 instead of the tap coefficient C _ij . _{Alternatively} , the tap coefficient C _i0 may be preset to the tap coefficient C _ij . This means that continuous adaptive control is not performed in the discovery window, but it is natural that the first transmission is performed in the discovery window as seen from the slave station ONU. In this automatic discovery procedure, information that must be communicated from the slave station ONU to the master station OLT is 128 bytes (about 1 kbit). If the bit error rate BER is 10 ^ (-3) or less, the success of the discovery procedure can be expected. Therefore, when distortion due to chromatic dispersion is not so severe, automatic discovery is possible even with the configuration of FIG. 6 by taking measures to disable equalization in the discovery window.

  However, if the distortion is significant, automatic discovery is difficult. Therefore, it is desirable to perform speculative discovery. This speculative discovery is to prepare a plurality of initial values for adaptive control in advance and cover these initial values in a plurality of discovery procedures, thereby probabilistically achieving the minimum required equalization even in the initial stage. Do and tell the discovery procedure to succeed. Once the discovery procedure is successful, continuous adaptive control can be performed by means for storing the tap coefficients described above, so that better equalization can be performed.

The speculative discovery procedure will be described below.
FIG. 7 is a configuration diagram of a switch corresponding to speculative discovery.
Here, FP type LD models and transmission line models are assumed to have several patterns, and tap coefficients to be equalized are calculated in advance as C _sk (k = 1, 2,..., K). . Then, as shown in the figure, it is provided side by side with C _i0 so that the switch SW can be selected.

The GE-PON LSI includes type information indicating whether it is a discovery window or a normal window in the window information.
When the burst optical signal to be controlled is the discovery window, the switch controller SC selects one of the tap coefficients C _sk (k = 1, 2,..., K) and starts the discovery window. It is reflected in the tap coefficient C _i at the time. The tap coefficient C _sk may be selected statistically to cover possible values of k, and is most simply selected in order. Then, even if the automatic discovery fails, the retry is performed. Therefore, even with this method, it is possible to equalize within a limited number of retries, and the automatic discovery succeeds. Therefore, the optimum tap coefficient C _i can be obtained.

Here, the form on the assumption that μ is set to 0 has a specific effect of simplifying the circuit configuration and reducing the cost. The equalizer of this form no longer performs application control, but is treated as a kind of adaptive control equalizer here.
FIG. 8 is a block diagram of an adaptive control equalizer AE that is simplified on the assumption that μ is zero. Figure 9 is a block diagram of a tap coefficient control unit T _i shown in FIG.

Here, for example, the transmission path distance is divided into J ways, and tap coefficients for compensating for waveform distortion due to transmission at each distance are C _ij (i = −2, −1,0, + 1, + 2, j = 1, 2, ..., J). The GE-PON LSI adds distance information between the master station OLT and the slave station ONU to the window information. The switch controller SC determines j corresponding to the distance information, gives a selection signal S to the switch SW for a period corresponding to the window, and selects a tap coefficient C _i corresponding to the switch SW.

In speculative discovery, the switch control unit SC determines to cover statistically j in the range of 1, 2,.
As described above, the transmission system according to the present invention can approximate the optical signal output from the FP type LD and distorted by the transmission of the long-distance optical fiber to the output signal waveform before passing through the optical fiber. Thereby, in long-distance optical transmission, an FP LD having multi-mode emission spectrum characteristics with different wavelengths can be used, and an economical transmission system can be constructed. In addition, FP type LD and single mode fiber can be used over a wide range, so this transmission system can use PX10 slave station ONU regardless of transmission distance, so it is unified with PX10 slave station ONU You can also Thereby, since this transmission system can unify and utilize the member currently spread, it can implement | achieve a system economically and easily.

Since the baseband signal of GE-PON is as high as 1.25 Gbps, analog processing is suitable for current device capability and cost reduction. 6 and 7, the tap coefficient is assumed to be a voltage, and the storage means is assumed to be a capacitor. However, some or all of the components can be digitally processed, and the present invention is effective for such a configuration.
Note that the communication method has been described above using GE-PON, but other methods may be used.

Further, as described above, the system using the FP type LD has been described. However, if the cost of the system can be reduced by using another LD, an optical transmitter including another LD may be used. .
In addition, various design changes can be made within the scope of matters described in the claims.

1 is a schematic diagram illustrating a configuration example of a PON system in which a master station and a plurality of slave stations are connected via an optical fiber via an optical coupler. It is the schematic which shows the basic composition in a GE-PON system. It is a figure which shows the relationship between the transmission wavelength from 1260 nm to 1360 nm, and the tolerance | permissible_range of a spectrum width. It is a block diagram of the adaptive control equalizer shown in FIG. It is a block diagram of the tap coefficient control part shown in FIG. It is a block diagram of another form of a tap coefficient control part. It is a block diagram of a switch corresponding to speculative discovery. It is a block diagram of a simplified equalizer on the assumption that μ is zero. It is a block diagram of the tap coefficient control part shown in FIG.

Explanation of symbols

ONU Slave station (subscriber terminal)
OLT Master station (station side device)
OC optical coupler SMF optical fiber (single mode fiber)
PD photodiode FP type LD Fabry-Perot type laser diode DFB type LD distributed feedback semiconductor laser TIA transimpedance amplifier PA postamplifier PD photodiode TF adaptive transversal filter T tap coefficient control unit AE adaptive control equalizer OT optical transceiver

Claims (6)

A semiconductor laser having multi-mode emission spectral characteristics of different wavelengths;
An optical fiber that propagates a signal from the semiconductor laser;
A transmission system comprising: an adaptive control equalizer that approximates a signal distorted by chromatic dispersion when passing through the optical fiber to an output signal waveform before passing through the optical fiber. The semiconductor laser is a Fabry-Perot semiconductor laser,
The optical fiber is a single mode fiber that connects a station-side device and a plurality of subscriber terminals, and is connected via an optical coupler that combines a plurality of signals from the plurality of subscriber terminals. The transmission system according to claim 1. The single mode fiber is
The distance connecting the subscriber terminal and the station side device exceeds 10 km, and the signal passing through the single mode fiber is a baseband signal of 1.25 Gbps or more. The transmission system described in 1. The adaptive control equalizer is:
4. The transmission system according to claim 1, wherein the transmission system has a plurality of transmission-system memories, and switches to a memory corresponding to the subscriber terminal in response to switching of a subscriber terminal that performs uplink transmission. The adaptive control equalizer is:
The transmission system according to claim 4, further comprising memory control means for performing speculative discovery. A semiconductor laser having multi-mode emission spectral characteristics of different wavelengths;
A station-side apparatus, comprising: an adaptive control equalizer that approximates a signal distorted by chromatic dispersion when passing through an optical fiber that propagates a signal from the semiconductor laser to an output signal waveform before passing through the optical fiber.

Images