JP2010226627A - Burst signal identifier, burst light receiver, burst signal identifying method and burst light receiving method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a burst light receiver which is easily made into IC and easily improves transmission efficiency. <P>SOLUTION: The burst light receiver as shown in Fig.1 has an APD 1, a transimpedance amplifier 2, an identifier 3, and a ramp voltage output low-pass filter 4. A reference voltage V<SB>R4</SB>generated by the ramp voltage output low-pass filter 4 is a decreasing ramp voltage proportional to a time even in a short time of consecutive identical digit (CID) time and also for a long time of guard time T<SB>G</SB>. V<SB>R4</SB>is then stabilized in the short time of CID time and rapidly decreased for the long time of guard time T<SB>G</SB>. Thus, the burst light receiver as shown in Fig.1, generating the reference voltage V<SB>R4</SB>by the ramp voltage output low-pass filter 4 pulls in burst in a short period of time and improves transmission efficiency. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to identification of burst signals that do not require resetting between bursts, and is particularly suitable for identification of burst signals suitable for use in OLT (Optical Line Terminal) of a PON (Passive Optical Network) system. The present invention relates to a receiver, a burst optical receiver, a burst signal identification method, and a burst optical reception method.

  The OLT is a station-side device and is connected to ONUs (Optical Network Units, optical subscriber line termination devices) provided on a plurality of subscriber sides. Since OLT communicates with each ONU by communicating with each ONU by a time division multiple access (TDMA) method, a signal transmitted to and received from the ONU is a burst signal. . The line length of the optical line connecting the OLT and each ONU is different for each ONU. Therefore, the signal level of the burst signal received from the ONU by the burst optical receiver provided in the OLT is greatly different for each burst. Therefore, a burst optical receiver is required to have a large dynamic range (ratio of minimum power and maximum power of receivable signals), for example, about 15 dB is required.

  As this type of burst optical receiver, Japanese Patent Laid-Open No. 4-342325 discloses a “burst optical transmission optical receiver”. As shown in FIG. 1 of Patent Document 1, this optical receiver for burst light transmission includes a light receiving element, a preamplifier, capacitors 16 and 17, a low-pass filter, and a limiter amplifier. And operates as follows. After the signal from the light receiving element is amplified by the preamplifier, it is branched into two by the branch circuit, and one of the branch circuit outputs is input to the input terminal A of the limiter amplifier via the capacitor 16. The other of the branch circuit outputs is input to the low-pass filter via the capacitor 17. The low-pass filter detects a direct current component of the input signal and supplies the direct current component to the input terminal B of the limiter amplifier as an identification point level signal. The burst light transmission optical receiver of Patent Document 1 uses a capacitor 16 for coupling the preamplifier and the limiter amplifier, and uses a capacitor 17 for coupling the preamplifier and the low-pass filter. Since the capacities of the capacitors 16 and 17 must be large, it is difficult to make an IC. As described above, the optical receiver for burst light transmission in Patent Document 1 has a problem that it is difficult to make an IC.

In order to solve the problem that this IC is difficult, for example, in the academic journal “IEEE JOURNAL OF SOLID-STATE CIRCUITS, VOL.34, NO.1, JANUARY 1999”, “A Wide-Dynamic-Range, High-Transimpedance” Optical reception described in the technical paper of Non-Patent Document 1 disclosed by Kenichi Ohhata, Toru Masuda, Kazuo Imai, Ryoji Takeyari, and Katsuyoshi Washio under the title of “Si Bipolar Preamplifier IC for 10-Gb / s Optical Fiber Links” There is a vessel. If the sum of the capacitances of the capacitors C _int and C _ext of the reference voltage generator disclosed in Fig. 3 and Fig. 4 of Non-Patent Document 1 is reduced, it can function as a burst optical receiver. it can.

  In Fig. 3 of Non-Patent Document 1, a transimpedance amplifier, a differential limiter amplifier, a reference voltage generator, an external capacitor, A burst optical receiver comprising an output buffer is shown in a block circuit diagram. Fig. 4 of Non-Patent Document 1 shows a circuit diagram of a burst optical receiver that embodies each circuit shown in the block in Fig. 3. In the burst optical receiver shown in Fig. 4, the reference voltage generator generates a DC voltage by rectifying the output of the transimpedance amplifier, and supplies this DC voltage as a reference voltage to the limiting amplifier.

The reference voltage generator in the burst optical receiver shown in FIG. 4 of Non-Patent Document 1 is provided with an external capacitor C _ext to constitute a low-pass filter. The time constant in this low-pass filter is a value proportional to R _LP (C _int + C _ext ) obtained by multiplying the sum of the capacities of the capacitors C _int and C _ext (C _int + C _ext ) by the resistor R _LP. Can be said to be a first-order RC filter. As mentioned before, the optical receiver of technical papers described Non-Patent Document 1, by reducing the sum of the capacitance of the capacitor C _int and C _ext in reference voltage generator (Reference voltage generator), as a burst optical receiver In some cases, the capacitor C _ext may be unnecessary when an IC is formed as a burst optical receiver.

  FIG. 9 is a diagram showing a burst optical receiver having substantially the same contents as FIG. 3 of Non-Patent Document 1. However, in FIG. 9, the transimpedance amplifier in the burst optical receiver of FIG. 3 of Non-Patent Document 1 is denoted by reference numeral 2 and the differential limiter amplifier is denoted in order to facilitate correspondence with the embodiments of the present invention described later. Reference numeral 6 denotes a circuit element including a reference voltage generator and an external capacitor by a first-order RC filter indicated by reference numeral 5, an output buffer is omitted, and an APD (means for converting the burst optical signal 100 into a current signal). An avalanche photodiode) 1 is added.

The burst optical receiver of FIG. 9 operates as follows. The APD 1 converts the burst optical signal 100 transmitted from the ONU via the optical fiber into a current signal 110. The transimpedance amplifier 2 converts the current signal 110 into a voltage signal, amplifies it, and outputs a burst signal 210 that is a voltage signal. The burst signal 210 is input to the first input 61 of the differential limiter amplifier 6 and is supplied to the primary RC filter 5. The primary RC filter 5 generates a DC voltage by rectifying the burst signal 210 and supplies the DC voltage to the second input terminal 62 of the differential limiter amplifier 6 as a reference voltage _VR5 . The reference voltage V _R5 represents the average voltage of the burst signal 210.

The differential limiter amplifier 6 amplifies the difference voltage between the burst signal 210 and the reference voltage _VR5, and outputs the amplified voltage as a non-inverted burst signal 611 and an inverted burst signal 612 from the non-inverted output terminal 63 and the inverted output terminal 64, respectively. . When the burst signal 210 is at a signal level within the dynamic range, the differential limiter amplifier 6 detects the burst signal when the magnitude (absolute value) of the burst signal 210 exceeds the magnitude (absolute value) of the reference voltage _VR5. By having a gain sufficient to saturate 210, a so-called limiting operation is performed. Therefore, the non-inverted burst signal 611 and the inverted burst signal 612 have pulse waveforms corresponding to the logical value “1” or “0” and can be processed as digital signals.

FIG. 2A is a timing diagram showing the burst optical signal 100 input to the APD 1. Here, in the burst optical signal 100, the optical burst A is followed by the optical burst B with an interval of the guard time _TG. The peak level V _A0 of the optical burst A and the peak level V _B0 of the optical burst B are large. Different examples are shown. The power ratio between the maximum power receivable optical burst and the minimum power optical burst, that is, the dynamic range of the burst optical receiver is, for example, 15 dB. In the burst optical signal 100, the optical burst B follows the optical burst B one after another at intervals of the guard time _TG . The guard time _TG varies depending on the PON system, but is 100 ns, for example.

FIGS. 2B and 1B are timing diagrams showing the burst optical signal 100 in an enlarged manner at time T1 (a time obtained by combining the guard time _TG and a short time before and after the guard time _TG ) in FIG. It is. FIG. 2B [2] shows a burst signal 210 and a reference voltage _VR5 which are input signals of the differential limiter amplifier 6 at the time T1. In FIG. 2 (B) [2], burst A and burst B in the burst signal 210 are shown. Burst A and burst B correspond to optical burst A and optical burst B in FIG. Here, FIG. 2B [1] is the burst optical signal 100, and FIG. 2B [2] is the burst signal 210. The burst signal 210 is inverted in polarity from the burst optical signal 100 by the transimpedance amplifier 2. The ratio of the peak voltage V _A of burst _A (equivalent to logical value “1”) V _A and the peak voltage of burst B (equivalent to logical value “1”) V _B is the peak level V _A0 of optical burst A and the peak of optical burst B The ratio is the same as the level V _B0 . In the guard time T _G, the level of the burst signal 210 is a zero level V ₀ continues, the capacitor C5 continues to discharge without charging. Therefore, the reference voltage V _R5 gradually increases from −V _A / 2 toward the zero level V ₀ during the guard time _TG as shown in FIG. −V _A / 2 is an average voltage of burst A having a peak voltage (magnitude, absolute value) of _VA . FIG. 2B [4] shows an inverted burst signal 612 that is an output signal of the differential limiter amplifier 6.

Reference voltage _{V R5} is because a DC voltage generated by rectifying the burst signal 210 in the primary RC filter 5, reference voltage _{V R5} are the same polarity as the burst signal 210. In the polarity, when the magnitude (absolute value) of the burst signal 210 exceeds the magnitude (absolute value) of the reference voltage V _R5 , the burst signal 210 of the excess portion is the limiting operation of the differential limiter amplifier 6. Thus, a non-inverted burst signal 611 and an inverted burst signal 612 are obtained. Therefore, the reference voltage V _R5 corresponds to the discrimination point level signal in Patent Document 1 described above. Incidentally, the burst signal 210 and the reference voltage _{V R5} shown in FIG. 2 (B) [2] are all negative signals. At this time, the differential limiter amplifier 6 outputs the non-inverted burst signal 611 and the inverted burst signal 612 when the magnitude of the burst signal 210, that is, the absolute value exceeds the magnitude of the reference voltage _VR5 , that is, the absolute value. To do. Therefore, referring to FIG. 2 (B) [2], the burst signal 210 and the reference voltage _VR5 are compared, and in explaining the operation of the differential limiter amplifier 6, the polarities of both signals are discarded. A description will be given simply by comparing the magnitudes of the two signals, that is, the absolute values.

Now, as shown in FIG. 2B [2], the reference voltage V _R5 at the rising portion of the burst B is set to V _R5P . The magnitude of the reference voltage V _R5 is not sufficiently lowered toward the zero level V ₀ during the guard time _TG , and the magnitude of the reference voltage V _R5P is the peak voltage of the burst B (corresponding to a logical value “1”) V _B 2, the inverted burst signal 612 does not appear at the output of the differential limiter amplifier 6 and the non-inverted burst signal 611 is the same as shown in FIG. 2B [4]. Therefore, the reference voltage peak voltage (logic value "1" or equivalent) of the V _R5P size burst B of V _B size between greater than, the bit of the burst B is lost. As time further elapses, the magnitude of the reference voltage _VR5 gradually decreases. Reference magnitude of the voltage V _R5 is, when reduced to the size below a certain value of the peak voltage (logic value "1" corresponds) V _B of the burst B, it becomes possible to bit detection of the burst B.

In this way, when the guard time _TG ends and the next burst starts, in order to detect each bit of the burst early, the reference voltage _VR5 has a magnitude between the guard time _TG . It is required to decrease as rapidly as possible. Rapidly reducing the magnitude of the reference voltage _VR5 during the guard time _TG reduces the time constant of the primary RC filter 5 and sets the cutoff frequency of the primary RC filter 5 high. . By setting the cutoff frequency of the primary RC filter 5 high, the magnitude of the reference voltage V _R5 is sufficiently larger than the magnitude of the peak voltage (corresponding to the logical value “1”) V _B of the burst B during the guard time _TG. Therefore, transmission efficiency can be increased. However, when the cutoff frequency of the first-order RC filter 5 is set high as described above, the burst optical receiver shown in FIG. 9 is likely to generate an error in bit detection in the burst as described below.

FIG. 2 (C) [1] is a timing chart showing the time T2 in FIG. 2 (A) in an enlarged manner. Figure 2 (C) [2] shows the burst B and the reference voltage V _R5 is an input signal of the differential limiting amplifier 6 at the time T2. FIG. 2C [4] shows an inverted burst signal 612 that is an output signal of the differential limiter amplifier 6. Here, FIG. 2C [1] is the burst optical signal 100, and FIG. 2C [2] is the burst signal 210. The burst signal 210 is inverted in polarity from the burst optical signal 100 by the transimpedance amplifier 2.

In FIG. 2 (C) [1], B m represents the m-th _bit, is _{B m + 66} represents the first m + 66 th bit. FIG. 2 (C) [1] shows a case where 65 bits from bit B _{m + 1} to bit B _{m + 65} in the optical burst B are continuously at a low level (logical value “0”), that is, in the same burst B. A case where there is a 65-bit identical code sequence (CID, Consecutive Identity Digit) is shown.

In burst B shown in FIG. 2 (C) [2], zero voltage (corresponding to logical value “0”) continues for 65 bits in CID time. Assuming that the speed of the burst optical signal 100 is 10.3125 Gbps, the time for 65 bits is 6.3 ns. If there is such a long CID, the direct current balance (DC balance) of the burst signal 210 fluctuates greatly. Therefore, if the cutoff frequency of the primary RC filter 5 is high, the reference voltage _VR5 is also large within this CID time. May fluctuate. In the example of FIG. 2 (C) [2], the average voltage is zero voltage of burst B (corresponding to a logical value “0”) during the 6.3 ns CID time, and therefore the reference voltage _VR5 is zero within the CID time. The voltage rises toward the voltage (equivalent to logical value “0”). If the reference voltage _VR5 rises to zero voltage (corresponding to logical value “0”) of burst B by the end of CID, the level of the low level bit (corresponding to logical value “0”) within the CID time Cannot exceed (becomes) the magnitude of the reference voltage V _R5 , the logical values “1” and “0” cannot be correctly identified, and a bit error occurs. The magnitude of the reference voltage V _R5 in the CID time, but not lead to the magnitude of the zero voltage of the burst B (logic value "0" or equivalent), considerably more than half of the peak voltage (logic value "1" or equivalent) When the voltage approaches zero (corresponding to a logical value “0”), the logical values “1” and “0” cannot be correctly identified, and a bit error occurs, resulting in a decrease in transmission quality.

On the other hand, in FIG. 2C [1], 65 bits from the bit Bm + 1 to the bit Bm + 65 in the optical burst B may be continuously at a high (equivalent to logical value “1”) level. In this case, the burst B of the burst signal 210 in the CID has a peak voltage (corresponding to a logical value “1”), and the reference voltage _VR5 decreases toward the peak voltage (corresponding to a logical value “1”) within the CID time. . If the reference voltage V _R5 in the CID time has lowered up to the peak voltage of the burst B (logic value "1" or equivalent), bit peak voltage during CID time (logical value "1" or equivalent) is a reference Since the voltage is not lower than _VR5 and the logical values “1” and “0” cannot be correctly identified, a bit error still occurs and the transmission quality is lowered.

The DC balance of the burst signal 210 is greatly broken during the CID time. Whether or not the DC balance is defined depends on the code format employed in the PON system. The DC balance is defined in the code format of 8B10B, but is not defined in the scrambled NRZ or 64B66B code format. Therefore, in order to reduce the bit error in the burst, the reference voltage _VR5 is required not to vary as much as possible with respect to time.

JP-A-4-342325

IEEE JOURNAL OF SOLID-STATE CIRCUITS, VOL.34, NO.1, JANUARY 1999, pp.18-24, “A Wide-Dynamic-Range, High-Transimpedance Si Bipolar Preamplifier IC for 10-Gb / s Optical Fiber Links”, Kenichi Ohhata, et al.

The burst optical receiver described in Patent Document 1 is difficult to be integrated into an IC, whereas the optical receiver shown in FIGS. 3 and 4 of Non-Patent Document 1 has a reference voltage generator (Reference
If the sum of the capacitances of the capacitors C _int and C _ext in the voltage generator) is reduced, it can function as a burst optical receiver, and C _ext is not required when an IC is formed as a burst optical receiver. Can do. However, as described above with reference to FIG. 9 and FIG.
In the burst optical receiver of No. 4, the reference voltage V _R5 is required to be as small as possible with the passage of time even if the DC balance is lost in the burst, while the guard time _TG between bursts is rapidly increased. It is required to decrease.

Therefore, if the cut-off frequency of the first-order RC low-pass filter 5 is lowered and the time constant is increased, the reference voltage _VR5 has a slow change with respect to time, the burst CID has a long time, and the DC balance of the burst signal. even collapses, can maintain a substantially constant value in the CID time, can not be converged to the value of the necessary degree in the guard time T _G, it is necessary to correspondingly long guard time T _G.

On the other hand, by increasing the cutoff frequency of the first order RC low-pass filter 5, by reducing the time constant, the reference voltage V _R5 is faster changes with respect to time, converge to the extent needed for the guard time T _G However, when the burst CID time is long, the burst CID time fluctuates rapidly, bit errors are likely to occur, and transmission quality deteriorates.

  As described above, the burst optical receiver of FIG. 9 has a problem to be solved that it is difficult to improve transmission efficiency due to a trade-off between the cutoff frequency of the primary RC filter 5. . Accordingly, the present invention provides a burst optical receiver that can be easily integrated into an IC and that can easily improve transmission efficiency, and a burst signal identifier, a burst signal identification method, and a burst optical reception method that can be applied to the burst optical receiver. Objective.

  In order to solve the aforementioned problems, the present invention provides the following means.

(1) A burst signal and a reference voltage are respectively received at the first and second input terminals, and a logical value “1” and a logical value “0” of the burst signal based on a difference between the voltage of the burst signal and the reference voltage. And a low-pass filter that inputs the burst signal and outputs the reference voltage, and the order of the low-pass filter is second or higher.
(2) comprising a photoelectric conversion element for converting a burst optical signal into a current signal, a transimpedance amplifier for converting the current signal into a voltage signal, and a discriminator having first and second input terminals,
The first input is directly connected to the output of the transimpedance amplifier, the second input is connected to the output of the transimpedance amplifier via a low pass filter, and the order of the low pass filter is a second or higher order burst light Receiver.
(3) The burst signal and the reference voltage are respectively received at the first and second input terminals of the differential amplifier, and the logical value “1” and logical value of the burst signal based on the difference between the voltage of the burst signal and the reference voltage A burst signal method for identifying a value “0”, inputting the burst signal to a low-pass filter having an order of 2 or more, and generating the reference voltage from the low-pass filter.
(4) The burst optical signal is converted into a current signal by a photoelectric conversion element, the current signal is converted into a voltage signal by a transimpedance amplifier, and the first input of the discriminator having first and second input terminals A voltage obtained by directly receiving the output of the transimpedance amplifier at the end and inputting the output of the transimpedance amplifier to a low-pass filter having a second order or higher is supplied to the second input terminal as a reference voltage. Thus, a burst light receiving method for outputting a burst electric signal from the output terminal of the discriminator.

  According to the present invention, it is possible to provide a burst optical receiver that can be easily integrated into an IC and that can easily improve transmission efficiency, a burst signal identifier, a burst signal identification method, and a burst optical reception method that can be applied to the burst optical receiver.

It is a block circuit diagram which shows the structure of the burst optical receiver by one Embodiment of this invention. FIG. 6 is a timing chart showing a burst optical signal 100 input to APD1. FIG. 3 is a timing chart showing signals at various parts at time T1 in FIG. FIG. 3 is a timing chart showing signals at various parts at time T2 in FIG. It is a figure which compares and shows the burst drawing time of the burst optical receiver of embodiment of FIG. 1 and the burst optical receiver of FIG. It is a figure which shows the response characteristic of the low-pass filter with respect to a step input about the low-pass filter whose order n is 1 to 10. It is a figure which shows the step response of a primary RC low pass filter. It is a figure which shows the step response of the secondary RC low pass filter of FIG. FIG. 2 is a circuit diagram showing an example of a lamp voltage output low-pass filter 4 in the burst optical receiver of FIG. It is a circuit diagram which shows an example of an active filter. FIG. 3 is a block circuit diagram showing a burst optical receiver having substantially the same contents as FIG. 3 of Non-Patent Document 1 so as to be easily compatible with the configuration of FIG.

  Hereinafter, the present invention will be described in detail with reference to the drawings.

FIG. 1 is a block circuit diagram showing a configuration of a burst optical receiver according to the first embodiment of the present invention. The burst optical receiver includes an APD (avalanche photodiode) 1, a transimpedance amplifier 2, a discriminator 3, and a lamp voltage output low-pass filter 4. The discriminator 3 and the ramp voltage output low pass filter 4 constitute the burst signal discriminator described above. The burst optical receiver of this embodiment is applied to OLT in the PON system. The ramp voltage output low-pass filter 4 may not strictly output the ramp voltage with respect to the step input. The time T ₁₂ reaches 12% of the convergence voltage V and the time reaches 97% of the convergence voltage V. the ratio of T _{97 is,} as long as it satisfies the _{T 97 /} T 12 <18. The reason for the numerical value will be described later.

The burst optical receiver of FIG. 1 operates as follows. The APD 1 converts the burst optical signal 100 transmitted from the ONU via the optical fiber into a current signal 110. The transimpedance amplifier 2 converts the current signal 110 into a voltage signal, amplifies it, and outputs a burst signal 210 that is a voltage signal. The burst signal 210 is input to the first input terminal 31 of the discriminator 3 and supplied to the ramp voltage output low-pass filter 4. The discriminator 3 is a differential limiter amplifier, and has a first input terminal a, a second input terminal b, a non-inverting output terminal c, and an inverting output terminal d. The non-inverting output terminal c and the inverting output terminal d are connected to the subsequent circuit through the non-inverting output terminal 31 and the inverting output terminal 32, respectively. The ramp voltage output low-pass filter 4 generates a DC voltage by rectification of the burst signal 210 and supplies the DC voltage to the second input terminal b of the discriminator 3 as a reference voltage _VR4 . Reference voltage _VR4 represents the average voltage of burst signal 210.

The burst optical receiver shown in FIG. 1 employs a lamp voltage output low-pass filter 4 as a low-pass filter that generates a reference voltage _VR4 . The lamp voltage output low pass filter 4 is a filter that outputs a lamp voltage in response to a step input. When the burst signal 210 is input to the ramp voltage output low-pass filter 4, a pulse for each bit in the burst signal 210 becomes a step input. The reference voltage V _R4, which is the output voltage of the ramp voltage output low-pass filter 4, changes toward the peak voltage of the pulse (corresponding to a logical value “1”) in proportion to the pulse duration, and the period in which the pulse is missing Changes to zero voltage (corresponding to logical value “0”) in proportion to the time.

The discriminator 3 amplifies the difference voltage between the burst signal 210 and the reference voltage _VR4, and outputs the amplified voltage as a non-inverted burst signal 311 and an inverted burst signal 312 from the non-inverted output terminal c and the inverted output terminal d, respectively. The discriminator 3 determines the burst signal 210 when the burst signal 210 has a signal level within the dynamic range, and when the magnitude (absolute value) of the burst signal 210 exceeds the magnitude (absolute value) of the reference voltage _VR4. By having a gain sufficient to saturate, a so-called limiting operation is performed. Therefore, the non-inverted burst signal 311 and the inverted burst signal 312 have pulse waveforms corresponding to the logical value “1” or “0” and can be processed as digital signals. However, it is not an essential requirement of the burst optical receiver that the discriminator 3 performs the limiting operation. If the gain of the discriminator 3 is not so high that the discriminator 3 performs the limiting operation, or the level of the burst signal 210 is small and the limiting voltage is not amplified, a limiter amplifier is provided at the subsequent stage of the discriminator 3, A digital signal can be obtained by performing a limiting operation using an amplifying circuit including the discriminator 3 and its limiter amplifier. The discriminator 3 and the subsequent limiter amplifier can be connected by DC coupling or AC coupling. In the case of AC coupling, the AC coupling cutoff frequency is preferably lower than the cutoff frequency of the lamp voltage output low-pass filter 4.

FIG. 2A is a timing diagram showing the burst optical signal 100 input to the APD 1. Here, in the burst optical signal 100, the optical burst A is followed by the optical burst B with an interval of the guard time _TG. The peak level V _A0 of the optical burst A and the peak level V _B0 of the optical burst B are large. Different examples are shown. The power ratio between the maximum power receivable optical burst and the minimum power optical burst, that is, the dynamic range of the burst optical receiver is, for example, 15 dB. In the burst optical signal 100, the optical burst B follows the optical burst B one after another at intervals of the guard time _TG . The guard time _TG varies depending on the PON system, but is 100 ns, for example.

2B and [1] are timing charts showing the burst signal 100 in an enlarged manner at time T1 in FIG. 2A (a time obtained by combining the guard time _TG and the short time before and after the guard time _TG ). is there. Figure 2 (B) [2] shows the burst signal 210 and the reference voltage _{V R4} is the input signal of the discriminator 3 at that time T1. In FIG. 2 (B) [2], burst A and burst B in the burst signal 210 are shown. Burst A and burst B correspond to optical burst A and optical burst B in FIG. Here, FIG. 2B [1] is the burst optical signal 100, and FIG. 2B [2] is the burst signal 210. The burst signal 210 is inverted in polarity from the burst optical signal 100 by the two transimpedance amplifier. The ratio of the peak voltage V _A of burst _A (equivalent to logical value “1”) V _A and the peak voltage of burst B (equivalent to logical value “1”) V _B is the peak level V _A0 of optical burst A and the peak of optical burst B The ratio is the same as the level V _B0 . At the guard time _TG , the level of the burst signal 210 continues to be the zero level V ₀ (corresponding to a logical value “0”), so the reference voltage V _R4 of the output of the ramp voltage output low-pass filter 4 is as shown in FIG. B) as shown in [2], during the guard time _{T G,} it rises toward the _-V a / 2 in proportion to time zero level _{V 0} (logic value "0" or equivalent). −V _A / 2 is an average voltage of burst A having a peak voltage (magnitude, absolute value) of _VA . FIG. 2B [3] shows an inverted burst signal 312 which is an output signal of the discriminator 3.

Since the reference voltage V _R4 is a DC voltage that varies toward the average voltage of the burst signal 210 in the ramp voltage output low-pass filter 4, the reference voltage V _R4 has the same polarity as the burst signal 210. The average voltage of the burst signal 210 is the aforementioned convergence voltage V. In its polarity, when the magnitude of the voltage of the burst signal 210 (absolute value) exceeds the magnitude of the reference voltage V _R4 (absolute value), constant by limiting the operation of the discriminator 3 is a burst signal 210 of exceeded part The non-inverted burst signal 311 and the inverted burst signal 312 are obtained. Therefore, the reference voltage V _R4 corresponds to the discrimination point level signal in Patent Document 1 described above. Incidentally, the burst signal 210 and the reference voltage _{V R4} shown in FIG. 2 (B) [2] are all negative signals. At this time, the discriminator 3 constituted by a differential limiting amplifier, the magnitude of the burst signal 210, that is, the absolute value, the magnitude of the reference voltage V _R4, i.e. when exceeding the absolute value, the non-inverted burst signal 311 and the inverted burst The signal 312 is output. Therefore, in the following, referring to FIG. 2 (B) [2], the burst signal 210 and the reference voltage _VR4 are compared, and in describing the operation of the discriminator 3, the polarities of both signals are discarded, and both The magnitude of the signal, that is, the absolute value, will be mainly described for comparison (same as described when the operation of the differential limiter amplifier 6 in the burst optical receiver of FIG. 9 is described earlier).

As shown in FIG. 2 (B) [2], reference voltage _{V R4} is the time when the burst A peak voltage (logic value "1" corresponds) _{V A} is completed, the peak voltage (logic value "1" corresponds ) _A value in the vicinity of the peak voltage V _A / 2, which is the average voltage of V _A. It is _assumed that the reference voltage V _R4 at the rising _edge of the burst B (at the end of the guard time _TG ) is V _R4P . During the guard time _TG , the reference voltage V _R4 increases from a value in the vicinity of −V _A / 2 toward the zero level V ₀ in proportion to the time (the magnitude of the reference voltage V _R4 is V _A / 2). Therefore, the reference voltage V _R4 is sufficiently increased during the guard time _TG , and the reference voltage V _R4P is increased to the average voltage of the burst B, that is, half of the peak voltage V _B. It is possible to make the value about V _B / 2. Then, since the magnitude of the burst B exceeds the magnitude of the reference voltage V _R4P , the discriminator 3 can detect the burst B having a level exceeding the reference voltage V _R4P . At this time, the non-inverted burst signal 311 and the inverted burst signal 312 are high-level or low-level digital signals corresponding to the logical value “1” or “0”. FIG. 2B [3] shows the inverted burst signal 312. As described above, the guard time _TG necessary for identifying the burst B in the inverted burst signal 312 can be shortened, and the transmission efficiency of the PON system can be improved.

FIG. 2 (C) [1] is a timing chart showing the time T2 in FIG. 2 (A) in an enlarged manner. Figure 2 (C) [2] shows the burst B and the reference voltage _{V R4} is the input signal of the discriminator 3 at that time T2. FIG. 2C [3] shows an inverted burst signal 312 that is an output signal of the discriminator 3. Here, FIG. 2C [1] is the burst optical signal 100, and FIG. 2C [2] is the burst signal 210. The burst signal 210 is inverted in polarity from the burst optical signal 100 by the two transimpedance amplifier.

In FIG. 2 (C) [1], B m represents the m-th _bit, is _{B m + 66} represents the first m + 66 th bit. FIG. 2 (C) [1] shows a case where 65 bits from the bit B _{m + 1} to the bit B _{m + 65} in the optical burst B are continuously at a low level, that is, 65 bits of the same code sequence ( A case with CID (Consecutive Identical Digit) is shown.

In burst B shown in FIG. 2 (C) [2], zero voltage (corresponding to logical value “0”) continues for 65 bits in CID time. Assuming that the speed of the burst optical signal 100 is 10.3125 Gbps, the time for 65 bits is 6.3 ns. When there is such a long CID, the direct current balance (DC balance) of the burst signal 210 varies greatly. At this time, in the embodiment of FIG. 1, the reference voltage _VR4 varies in proportion to the passage of time in the CID time within the CID time. In the example of FIG. 2 (C) [2], the average voltage is zero voltage of burst B (corresponding to a logical value “0”) during the 6.3 ns CID time, so that the reference voltage V _R4 is within the CID time. It rises toward zero voltage (equivalent to logical value “0”). If the reference voltage _VR4 rises to the zero voltage (corresponding to logical value “0”) of burst B by the end of CID, the voltage of the low-level bit (corresponding to logical value “0”) within the CID time becomes the reference. The voltage _VR4 cannot be exceeded (exceeded), and the logical values “1” and “0” cannot be correctly identified. However, in the embodiment of FIG. 1, the longest in the CID scheduled time the proportional coefficient of the reference voltage V _R4 that varies in proportion to time, can be designed to the extent that almost does not vary. Moreover the ramp voltage output low pass filter 4 to a reference voltage V _R4 has designed terribly not to vary in this way CID time, also the period of the guard time T _G described above, the reference voltage V _R4 between guard time T _G It is possible to change the reference voltage V _R4P sufficiently so that the average voltage of the burst B, that is, the peak voltage (corresponding to the logical value “1”) V _B , is half the value of V _B / 2. It is.

FIG. 3 is a diagram in the case where the burst optical receiver of FIG. 9 that generates the reference voltage V _R5 by the primary RC low-pass filter 5 and the lamp voltage output low-pass filter 4 generates an ideal lamp voltage as the reference voltage V _R4 . It is a figure which compares and shows burst pull-in time about 1 burst optical receiver. Here, the pull-in time indicates a time during which the reference voltage changes from V _A0 / 2 to V _B0 / 2 between bursts. The burst optical signal 100 (input signal) has an operation speed of 10.3125 Gbps, CID: 65 bits, tCID (CID time): 6.30 ns, and dynamic range: 15 dB. The allowable burst penalty corresponds to the difference in input power required to obtain the same bit error rate with respect to an ideal input signal, and represents the quality of the input signal. In Figure 3, the actually obtain an acceptable burst penalty 0.5dB required when applied to the product, it is preferable to 54.7ns case, the time constant for generating a reference voltage V _R by the primary RC filter of FIG. 9 The pull-in time at this time is 348.5 ns. Further, in order to obtain the same allowable burst penalty 0.5 dB, the ramp voltage output low pass filter 4 of FIG. 1, when generating the ideal ramp voltage as the reference voltage V _R, the rate of change dv / dt of the lamp voltage 0.0173 ( 1 / ns) is preferable, and the pull-in time at this time is 57.9 ns.

As shown in the table of FIG. 3, when the allowable burst penalty was 1.5 dB, pull time than the burst optical receiver of FIG. 9 for generating the reference voltage V _R by the primary RC filter, the lamp voltage output If the low-pass filter 4 to generate an ideal ramp voltage as the reference voltage V _R, the burst optical receiver of the embodiment of the invention of FIG. 1, pull time is shortened to about one-fifth. The same applies to other permissible burst penalties. The pull-in time of the burst optical receiver according to the embodiment of the invention of FIG. 1 is 1/5 to 1/7 of the pull-in time of the burst optical receiver of FIG. As an improvement effect, the effect is about 4.5 to 7 times. Here, the improvement effect is (the pull-in time in the burst optical receiver of FIG. 9) / (the pull-in time in the first embodiment (FIG. 1) of the present invention). As described above, in the burst optical receiver of the embodiment of FIG. 1, the pull-in time is shortened, and the transmission efficiency can be improved.

FIG. 4 is a diagram showing response characteristics of the Bessel type low-pass filter with respect to the step input for the Bessel type low-pass filter having the order n of 1 to 10. In this figure, the horizontal axis represents time t, and the vertical axis represents the relative value of the output voltage. The characteristic lines in this figure correspond to the reference voltages V _R4 and V _R5 during the guard time _TG shown in FIG. As shown in the response characteristic diagram of FIG. 4, when the order n of the Bessel type low-pass filter is 1, the response of the Bessel type low-pass filter to the step input is more rapid than the Bessel type low-pass filter of other orders. However, the slope of the response gradually decreases, and finally reaches a level of about 0.01 before reaching zero, and hardly fluctuates. On the other hand, when the order n is 2 or more, the response of the Bessel type low-pass filter to the step input rises with a much gentler slope than that of the first-order Bessel type low-pass filter, and after that, almost in the time period. It increases proportionally towards zero and gradually reaches zero after a certain period of time while gradually easing the slope.

The response characteristics of FIG. 4 indicate the following. The output of the first-order Bessel type low-pass filter fluctuates exponentially, and when it approaches to zero, it maintains an almost constant value for a long time, and it takes a very long time to reach substantially zero. . On the other hand, the output of the Bessel low-pass filter whose order n is 2 or more is a ramp voltage that varies substantially in proportion to time, and becomes substantially zero in a short time. In this embodiment of FIG. 1, since a means for generating a reference voltage V _R4 Bessel low-pass filter are constituted by using a ramp voltage output low-pass filter 4, a reference voltage V _R4 response characteristics close to the lamp voltage It becomes.

The reference voltages V _R4 and V _R5 in the burst optical receiver of FIGS. 1 and 9 have been described with reference to FIGS. 2B and 2C and FIG. 2C and FIG. The response characteristics of the Bessel low-pass filter correspond well to the difference between the reference voltages V _R4 and V _R5 . The guard time _TG in FIG. 2 (B) [2] is, for example, 100 ns, which is considerably longer than the 6.3 ns CID time in FIG. 2 (C) [2]. Reference voltage V _R is briefly the stable called CID time, it is desirable to change rapidly to a long time of the guard time T _G is described above.

_VR5 generated by the first-order RC low-pass filter 5 shows the response characteristic of n = 1 in FIG. 4, and therefore decreases rapidly in a short time of CID time and slowly decreases in a long time of guard time _TG. However, it does not reach zero for a long time. In contrast, V _R4 for generating a ramp voltage output low-pass filter 4, since shows the response characteristics of the n = 2 or n> 2 in FIG. 4, in a short time of CID time, the long time of the guard time T _G Also decreases in proportion to time. _{Therefore, V R4} is compared to the _{V R5,} briefly the stable called CID time, decreases rapidly for a long time that the guard time _{T G.} Thus, burst optical receiver of the embodiment of Figure 1 using a ramp voltage output low pass filter 4 as a low-pass filter is a means for generating a reference voltage V _R generates a reference voltage V _R the primary RC low-pass filter 5 Figure 9 compared with the burst optical receiver, it is possible to draw a burst with a short guard time T _G, thereby improving transmission efficiency.

Ramp voltage output low-pass filter 1 4 is strictly may not and outputs the ramp voltage to the step input, 97% of the time T ₁₂ to reach 12% of the convergence voltage V converges voltage V the percentage of time _{T 97} to reach _may as long as it satisfies the _{T 97 /} T 12 <18. Here, 12% of the convergence voltage V is a value corresponding to the allowable burst penalty of 0.5 dB in FIG.
10 ^ (0.5 / 10) = 112%
112%-100% = 12%
It becomes. On the other hand, 97% of the convergence voltage V is a dynamic range, that is, a value corresponding to the ratio of V _A0 and V _B0 in FIG. If the required dynamic range is 15 dB,
10 ^ (15/10) = 31.6 times 100%-(1 / 31.6) = 97%
It becomes.

FIG. 7 is a circuit diagram showing a specific circuit of the lamp voltage output low-pass filter 4 in the burst optical receiver of FIG. The lamp voltage output low-pass filter of FIG. 7 is formed by connecting a first primary RC low-pass filter composed of a resistor R1 and a capacitor C1 and a second primary RC low-pass filter composed of a resistor R2 and a capacitor C2 in cascade. The next RC low-pass filter. FIG. 6 shows the step response of the second order RC low pass filter of FIG. For comparison, FIG. 5 shows the step response of the first-order RC low-pass filter. 5 and 6 are set so that the response (vertical axis) is 0.1 at t = 1. As shown in FIG. 6, in the second-order RC low-pass filter, T ₉₇ / T ₁₂ = 16, but as shown in FIG. 5, in the first-order RC low-pass filter, T ₉₇ / T ₁₂ = 27, so that the second-order RC low-pass filter. The filter is 27/16 = 1.69 times better than the first-order RC low-pass filter. In general, with a second-order RC low-pass filter, it is difficult to set T ₉₇ / T ₁₂ to 16 or less. Therefore, as a T ₉₇ / T ₁₂ value for obtaining an improvement effect of 1.5 times, T ₉₇ / T ₁₂ = 18 is preferable.

  The burst optical receiver according to the embodiment of FIG. 1 does not require a coupling capacitor that is necessary for the burst optical receiver of Patent Document 1, and thus can be easily integrated into an IC.

  The above has been described in detail with reference to an example in which the present invention is implemented as a burst optical receiver. However, it is apparent from the above description of the embodiments of the burst optical receiver that the present invention can be realized not only as a burst optical receiver but also as a burst signal identifier, a burst signal identification method, and a burst optical reception method.

In addition, the embodiments have been described above and the present invention has been specifically described in detail. However, the present invention is not limited to these embodiments. For example, the ramp voltage output low-pass filter 4 in FIG. 1 does not necessarily strictly indicate a ramp voltage response to the step input. The time T ₁₂ for reaching 12% of the convergence voltage V and the convergence voltage V of 97 are obtained. It is sufficient that the ratio of the time T ₉₇ to reach% satisfies T ₉₇ / T ₁₂ <18. Further, for example, the lamp voltage output low-pass filter 4 may be a secondary RC low-pass filter as shown in FIG. 7 or an active filter. An example of the active filter is shown in FIG. Further, the differential limiter amplifier constituting the discriminator in the present invention does not necessarily perform the limiting operation, and the differential limiter amplifier has a small gain or the level of the burst signal 210 is small. As described above, when the limiter amplifier cannot amplify the saturation voltage, it is possible to obtain a burst signal that can be handled as a digital signal by providing another limiter amplifier in the subsequent stage.

1 APD (avalanche photodiode)
2 Transimpedance amplifier 3 Discriminator 4 Lamp voltage output low-pass filter 5 Primary RC low-pass filter a First input terminal b Second input terminal c Non-inverted output terminal d Inverted output terminal 31 Non-inverted output terminal 32 Inverted output terminal 6 Differential limiter amplifier 61 First input terminal 62 Second input terminal 63 Non-inverted output terminal 64 Inverted output terminal 100 Burst optical signal 110 Current signal 210 Burst signal 311 Non-inverted burst signal 312 Inverted burst signal _TG guard time V ₀ Zero level V _R4 , V _R5 reference voltage

Claims (18)

  The burst signal and the reference voltage are respectively received at the first and second input terminals, and the logical value “1” and the logical value “0” of the burst signal are identified based on the difference between the voltage of the burst signal and the reference voltage. And a low-pass filter that inputs the burst signal and outputs the reference voltage, wherein the order of the low-pass filter is second or higher.   The burst signal discriminator according to claim 1, wherein the step response of the low-pass filter is a ramp voltage. When the convergence voltage of the lamp voltage is V, the ratio of the time T ₁₂ reaching 12% of the convergence voltage V and the time T ₉₇ reaching 97% of the convergence voltage V satisfies T ₉₇ / T ₁₂ <18. The burst signal discriminator according to claim 1 or 2, characterized by the above.   4. The burst signal discriminator according to claim 1, wherein the low-pass filter is realized by a resistor element and a capacitor element.   4. The burst signal identifier according to claim 1, wherein the low-pass filter is an active filter.   6. The burst signal discriminator according to claim 1, wherein the discriminator is a differential limiter amplifier.   7. The burst signal identification according to claim 6, wherein a limiter amplifier is connected after the differential limiter amplifier, and an output of the differential limiter amplifier and an input of the limiter amplifier are connected by DC coupling. vessel.   A limiter amplifier is connected to the subsequent stage of the differential limiter amplifier, and an output of the differential limiter amplifier and an input of the limiter amplifier are connected by an AC coupling having a cutoff frequency lower than the cutoff frequency of the low-pass filter. The burst signal discriminator according to claim 6, wherein: A photoelectric conversion element that converts a burst optical signal into a current signal, a transimpedance amplifier that converts the current signal into a voltage signal, and a discriminator having first and second input terminals,
The first input is directly connected to the output of the transimpedance amplifier, the second input is connected to the output of the transimpedance amplifier via a low-pass filter, and the order of the low-pass filter is second or higher. A burst optical receiver.   The burst optical receiver according to claim 9, wherein the step response of the low-pass filter is a lamp voltage. When the convergence voltage of the lamp voltage is V, the ratio of the time T ₁₂ reaching 12% of the convergence voltage V and the time T ₉₇ reaching 97% of the convergence voltage V satisfies T ₉₇ / T ₁₂ <18. The burst signal discriminator according to claim 9 or 10, wherein:   The burst optical receiver according to claim 9, wherein the low-pass filter is realized by a resistance element and a capacitance element.   The burst optical receiver according to claim 9, wherein the low-pass filter is an active filter.   The burst optical receiver according to claim 9, wherein the discriminator is a differential limiter amplifier.   15. The burst according to claim 9, wherein a limiter amplifier is connected to a subsequent stage of the differential limiter amplifier, and an output of the differential limiter amplifier and an input of the limiter amplifier are connected by DC coupling. Optical receiver.   A limiter amplifier is connected to the subsequent stage of the differential limiter amplifier, and an output of the differential limiter amplifier and an input of the limiter amplifier are connected by an AC coupling having a cutoff frequency lower than the cutoff frequency of the low-pass filter. The burst optical receiver according to claim 9, wherein the optical receiver is a burst optical receiver.   A burst signal and a reference voltage are respectively received at the first and second input terminals of the differential amplifier, and a logical value “1” and a logical value “0” of the burst signal based on a difference between the voltage of the burst signal and the reference voltage. ”, The burst signal is input to a low-pass filter having an order of 2 or more, and the reference voltage is generated from the low-pass filter.   A burst optical signal is converted into a current signal by a photoelectric conversion element, the current signal is converted into a voltage signal by a transimpedance amplifier, and the first input terminal of the discriminator having first and second input terminals is connected to the first input terminal. By directly receiving the output of the transimpedance amplifier and supplying the voltage obtained by inputting the output of the transimpedance amplifier to a low-pass filter having a second order or higher order as a reference voltage to the second input terminal, A burst optical receiving method, wherein a burst electric signal is output from an output terminal of the discriminator.

Images