JPS6221419B2 - - Google Patents

Description

[Detailed description of the invention] The present invention relates to an automatic equalization method for an optical receiver.

In a PCM optical communication system, a large number of optical repeaters each having a built-in optical receiver are arranged at approximately equal intervals and mainly perform amplification and equalization of transmitted optical signals. This is to compensate for the deterioration of the optical signal that may occur from one terminal station to another. This deterioration of the optical signal directly leads to an increase in the code error rate and significantly deteriorates the transmission quality. Therefore, the optical receiver is provided with an equalization function to improve the waveform distortion of the transmitted optical signal. This waveform distortion is generally expressed quantitatively as the amount of intersymbol interference.

A normal optical receiver uses an optical transmission line (optical fiber)
a photodiode that receives an optical signal from the amplifier and converts it into an electrical signal, an amplifier that AGC amplifies the electrical signal, an equalizer that receives and equalizes the output signal from the amplifier, and the equalizer. and a comparator that identifies and reproduces the codes of "1" and "0" from the equalized output signal from the equalizer, and the equalizer always performs optimal equalization while monitoring the change in the amount of intersymbol interference mentioned above. Let's do it. However, this equalization is only an optimal equalization, and perfect equalization is not possible. This is clear if we consider that the output characteristics of light emitting diodes (laser diodes), frequency band characteristics of optical fibers, etc. in optical communication systems have inherent dispersion factors. Fluctuations in the amount of interference are inevitable, and it must be considered that perfect equalization is impossible.

Therefore, an object of the present invention is to provide an automatic equalization method for an optical receiver, which makes it possible to substantially approach complete equalization, although it is recognized that complete equalization by the equalizer is impossible. It is to make a proposal.

In accordance with the above object, the present invention includes an equalizer that equalizes an optical signal received from an optical transmission line, and a threshold level, and the threshold level and the level of the equalized output signal from the equalizer. Compare “1”
and a comparator for identifying and reproducing a "0" code, the optical receiver detects a magnitude fluctuation in the amount of intersymbol interference contained in the equalized output signal, and the equalizer The present invention is characterized in that the frequency bands are widened and narrowed, and the threshold levels are also varied in height.

The present invention will be explained below with reference to the drawings.

Figure 1 shows one of the optical receivers to which the method of the present invention is applied.
FIG. 2 is a block diagram showing an embodiment. In this figure,
11 is a photodiode (APD: avalanche photodiode) that receives an optical signal P from an optical transmission line (not shown) and converts it into an electrical signal; 12
13 is an equalizer (EQL) that equalizes the output signal of the amplifier 12, and 14 receives the equalized output signal from the equalizer 13 and outputs "1". This is a comparator (COMP) that identifies and reproduces the 0'' code and outputs reproduced data Dout.

As already mentioned, the optical signal P, the electrical signal from the photodiode 11, and the output signal from the amplifier 12 have different signs due to changes in the output characteristics of the transmitting side light emitting diode or changes in the frequency band characteristics in the optical transmission line system. The amount of interference changes. Therefore, first, an intersymbol interference amount extraction circuit 5 is introduced to constantly monitor fluctuations in the amount of intersymbol interference. Specifically, the intersymbol interference amount extraction circuit 15 detects the pulse width of the equalized output signal from the equalizer 13.
If this pulse width is large, the tail portions of adjacent pulses overlap each other, and the amount of intersymbol interference becomes large. Conversely, if the pulse width is small, the tail portions of adjacent pulses will not overlap, and the amount of intersymbol interference will be small. That is, the amount of intersymbol interference and the pulse width are approximately proportional to each other. Therefore, in order to reduce the amount of intersymbol interference, it is required to reduce the pulse width as much as possible. For this purpose, the pass frequency band of the equalizer 13 is widened, and the equalizer 13
It is only necessary to reproduce the input signal as faithfully as possible so that the pulse width does not become large. However, the wider the pass frequency band, the more the so-called S/N deteriorates. That is, if the passing frequency band is simply widened, the S/N will deteriorate even if the amount of intersymbol interference is reduced. Therefore, the pass frequency band is controlled to be wide or narrow so that both the S/N and the amount of intersymbol interference are maintained at appropriate values. This variation in the amount of intersymbol interference is first fed back to the equalizer 13 through the frequency band characteristic control circuit 16. That is, as the amount of intersymbol interference increases or decreases, the pass frequency band of the equalizer is widened or narrowed, respectively. For example, the capacitance component of the filter constituting the equalizer is a variable capacitance element, and the signal from the intersymbol interference extraction circuit 15 is sent to a level suitable for controlling the variable capacitance element via the control circuit 16. This can be achieved by converting it into a voltage signal, applying it to the voltage control terminal of the variable capacitance element, and varying the capacitance value.

In this way, the equalization function of the equalizer 13 can be improved to a certain optimum level.
However, as described above, this equalization is not perfect equalization and cannot be said to have sufficiently absorbed the variation in the amount of intersymbol interference. Therefore, as a second countermeasure, the comparator 14, especially its threshold level, is controlled by the output of the intersymbol interference extraction circuit 15. This method focuses on the fact that there is a certain relationship between the threshold level and the amount of intersymbol interference. This will be explained with reference to Figures 2A, 2B and 2C. FIG. 2A is a graph showing the relationship between signals and noise, FIG. 2B is a graph showing changes in noise due to the amount of intersymbol interference, and FIG. 2C is a graph showing the relationship between the identification signal and the identification value. Noise in optical reception signals includes thermal noise (TN) and shot noise (SN). The thermal noise (TN) of the is the same as that of a conventional coaxial system, but the shot noise (SN) of the varies depending on the optical input level. This point is different from conventional coaxial systems. Figure 2A shows signal (SIG) and noise (SN, TN)
shows the relationship between t is the time axis. As is clear from FIG. 2A, when there is no amount of intersymbol interference, the magnitude of noise differs greatly between when the signal is "1" and when the signal is "0". Therefore, the optimal discrimination value level is considerably smaller than 1/2, which is around 0.17.

However, if the intersymbol interference of the optical input signal increases, shot noise will appear depending on the magnitude of the signal level even when it is "0". FIG. 2B shows the change in noise at the discrimination point when there is intersymbol interference.

On the other hand, changes in the signal include the so-called eye
is almost the same (however, in absolute value) as the "1" level and the "0" level. Although this change is equivalent to the “1” level and the “0” level in terms of absolute value,
Relatively speaking, the "0" level deteriorates more significantly.
This relationship is shown in FIG. 2C. As shown above, as inter-symbol interference increases, the noise at the "0" level increases, and due to deterioration of the so-called eye, the optimal threshold level becomes
It changes in the direction of becoming larger. This was considered qualitatively and the results of computer simulation are shown in the graph of FIG. This graph shows the normalized threshold level TH on the horizontal axis and the bit error rate ER on the vertical axis.
The curves , , and represent the characteristics of ER versus TH when intersymbol interference amounts are assumed to be 0%, 11.1%, 17.8%, and 28.5%, respectively. Now, assuming that the optical signal is in good condition and the amount of intersymbol interference is 0%, the threshold level value is around 0.17.
By setting the comparator threshold to TH ₀ , the bit error rate can be kept to a minimum. However, the V-shaped curve in this graph shifts, for example, from one curve to another in response to the above-mentioned fluctuations in the frequency band characteristics of the optical fiber. Even if you shift to this curve, you can still change the threshold level.
If TH ₀ is maintained, the increase in the bit error rate ER that occurs will already exceed the compensation range of the equalizer 13. Therefore, TH ₀ will be raised to the point (TH ₁ ) corresponding to the minimum value of the curve in this graph. Specifically, in FIG. 1, the output of the intersymbol interference extraction circuit 15 is obtained and the threshold level of the comparator 14 is raised via the threshold level control circuit 17. In other words, the threshold level is changed in height in response to a change in the amount of intersymbol interference. This results in
Code error rate ER that could not be suppressed by the equalizer 13
can be offset by the comparator 14.

Returning to FIG. 1, 18 is a peak detector (PD) that detects the peak level of the equalized output signal;
19 is a comparator that compares the peak level of the peak detector 18 with a predetermined level, and its output is fed back to the amplifier 12 and is used as a so-called AGC.
(Automatic Gain Control) Achieves amplification.
The output is also fed back to the photodiode 11 to change its multiplication factor. This is mainly due to
This is effective in compensating for optical signal transmission loss occurring in an optical transmission line. In addition, these peak detectors 1
8. The comparator 19 is conventionally used as a part of the main component circuit of an optical receiver.

In this way, in the present invention, four feedback loops are formed as indicated by , , and in FIG. 1, so it is necessary to appropriately determine the transfer function of each loop to stabilize the system. Here, if we capture the loop transfer function as the loop response speed, we get
The loop response speed of the loop is the fastest, and the response speed of the loop is the next fastest,
It is ideal for the stability of the system to keep the response speed of the loop as slow as possible.

As already stated, the gist of the present invention is to form the loop shown in FIG. An intersymbol interference amount extraction circuit 15 that supplies a control signal is essential. The circuit 15 can be implemented in various ways, for example, with a configuration as shown in FIG. However, Fig. 4 is suitable when the input signal does not involve mark rate fluctuation, and 13 is an equalizer, 1
4 is a comparator, which is the same as that shown in FIG. 1 (for simplicity, the frequency band characteristic control circuit 16 and threshold level control circuit 17 in FIG.
). 41 is a comparator (COMP), 4
3 is an integrator. Further, FIG. 5 is a diagram showing identification waveforms in the configuration of FIG. 4.

The output of the equalizer 13 is input to comparators 41 and 14. Comparator 41 identifies the input signal with a threshold h ₁ approximately equal to half of its peak value h ₀ and detects the pulse width of this input signal. The reason why the threshold value was set to _h1 is that since the pulse width cannot be detected in the thermal noise TN part of Figure 2A, this part is avoided and the pulse width is detected at a level above this part. . FIG. 5 shows the relationship between the identification waveform and its peak value h ₀ and threshold value h ₁ . Integrator 43
integrates the output of comparator 41 and generates a voltage proportional to the pulse width at threshold h ₁ of the input signal. The comparator 14 uses the signal of the integral value of the integrator 43 as a threshold value to identify and encode the output of the equalizer 13.

As described above, according to the configuration shown in FIG. 4, the amount of intersymbol interference can be determined based on the magnitude of the pulse width at a level of approximately 1/2 of the equalizer output.

However, in the case of the configuration shown in FIG. 4, if the input signal is accompanied by a variation in the mark rate, the integrator output level will change due to the variation in the mark rate, so it cannot be used as is. For this purpose,
The configuration shown in FIG. 6 is preferred. Although this configuration is based on the same operating principle as the case shown in FIG. 4, it can also cope with variations in mark rate. In this figure, the same parts as those in FIG. 4 are designated by the same numbers, and 44 is an integrator and 45 is a differential amplifier.

In FIG. 6, the identification output of comparator 14 is integrated in integrator 44. In FIG. The differential amplifier 45 provides an output corresponding to the difference between the output of the integrator 43 and the output of the integrator 44 to the comparator 14 as a threshold value.

As described above, according to the configuration shown in FIG. 6, even if the output level of the integrator 43 changes due to a change in the mark rate, the output level of the integrator 44 having a similar mark rate change is transferred to the differential amplifier. 45 cancels out changes in the output level of the score divider 43 due to changes in mark rate. As a result, the discrimination level of the comparator 14 and the pass frequency band of the equalizer 13 are controlled to have a voltage proportional only to the magnitude of the pulse width without being affected by mark rate fluctuations.

As described above, according to the present invention, an automatic equalization method for an optical receiver is realized which can significantly reduce the bit error rate compared to the conventional method.

[Brief explanation of the drawing]

Figure 1 shows one of the optical receivers to which the method of the present invention is applied.
The block diagram illustrating the embodiment, FIG. 2A, FIG. 2B, and FIG. 2C are three types of graphs for explaining the relationship between the threshold level and the amount of intersymbol interference, respectively.
Figure 3 is based on computer simulation.
Graph showing the relationship between threshold level and code error rate using the amount of intersymbol interference as a parameter, 4th
1 is a block diagram showing a first configuration example of the circuit 15 shown in FIG. 1, FIG. 5 is a waveform diagram used to explain the operation of FIG. 4, and FIG.
FIG. 2 is a block diagram showing a configuration example. In the figure, 11 is a photodiode, 13 is an equalizer, 14 is a comparator, 15 is an intersymbol interference amount extraction circuit, 16 is a frequency band characteristic control circuit, 17 is a threshold level control circuit, and P is an optical signal. .

Claims (1)

[Scope of Claims] 1. An equalizer including a filter with a variable passing frequency band and equalizing an optical signal received from an optical transmission line; In an optical receiver comprising a comparator that identifies and reproduces "1" and "0" codes by comparing the level of the equalized output signal, the amount of intersymbol interference contained in the equalized output signal is large. In order to detect whether the pulse width is large or small, an intersymbol interference amount extraction circuit is provided which receives the equalized output signal and detects whether the pulse width is large or small. The pulse width is made smaller or larger by changing the pass frequency band of a filter in the equalizer to be wider or narrower depending on whether the pulse width is large or small, and further, the above-mentioned detection 1. An automatic equalization system for an optical receiver, characterized in that the threshold level is changed higher or lower depending on whether the pulse width received is large or small. 2. A difference is created between the response speed for changing the pass frequency band to be wider or narrower and the response speed for changing the threshold level higher or lower, and the former response speed is made higher than the latter response speed. The automatic equalization method according to claim 1, which also sets speedily.