JP3930886B2 - Optical signal quality monitor - Google Patents

Description

  The present invention relates to an optical signal quality monitor for monitoring a signal-to-noise ratio of an optical signal in an optical fiber transmission network in which digital optical signals having different bit rates are transmitted.

  In the SDH (Synchronous Digital Hierarchy), which became a global unified network hierarchy in the 1990s, a parity check called bit interleave parity is performed between repeaters (BIP-8) and between multiplexed terminals (BIP-N × 24). In each case, the failure section identification and the switching activation signal are obtained. Here, N is N of STM-N representing the multiplexing number, and is based on STM-1 of 156 Mbit / s, and a step that is N times the natural number is represented as STM-N. N = 1, 4, 16, and 64 are recognized internationally. BIP-M means a parity check every M bits, and M check bits are obtained. The transmitting side performs a parity check in parallel with the M bits of the signal in the frame, stores the check bit in the next frame, and transmits it along with the main signal. On the receiving side, a similar parity check is performed, and a transmission error is detected by checking with a check bit stored at a predetermined position in the next frame.

FIG. 9 shows a configuration example of a conventional error rate measurement system. In the figure, a part of the optical signal from the transmission line is branched by the optical branching unit 51-1, amplified by the optical amplifier 52, and further branched into two by the optical branching unit 51-2. One optical signal branched by the optical branching device 51-2 is input to the clock extraction circuit 53, and a clock f ₀ corresponding to the bit rate Nf ₀ of the optical signal is extracted. The other optical signal branched by the optical branching unit 51-2 is input to the receiving circuit 54, and the output is further input to an error rate detection circuit 55 including a frame detection circuit, a parity check circuit, and a verification circuit. The reception circuit 54 and the error rate detection circuit 55 operate according to the clock f ₀ extracted by the clock extraction circuit 53, and measure the error rate of the optical signal. Note that the clock extraction circuit 53, the reception circuit 54, and the error rate detection circuit 55 need to be configured according to the bit rate of the optical signal. That is, in order to detect an error rate corresponding to a plurality of bit rates, it is necessary to prepare a circuit corresponding to each bit rate, and it is not possible to cope with a single circuit.

  By the way, in order to evaluate a transmission system, a method of directly measuring a signal error rate is generally used. However, this method has a problem that the measurement time is long when the error rate is very low, and the working efficiency is lowered.

  Therefore, a method has been devised for estimating the error rate at the optimum operating point from the error rate tendency obtained when the threshold value of the discriminator is changed (Non-patent Document 1). FIG. 10 shows an eye pattern of an optical signal and a light intensity histogram. By changing the threshold value of the discriminator at the time point when the eye opening of the eye pattern becomes maximum (discrimination point), the “high” or “1” level in the case of binary transmission and the “low” or “0” level And measure the error rate at that time.

  Actually, a measurement system including the clock extraction circuit 53, the photoelectric converter 56, and the electric signal processing means 57 as shown in FIG. 11 is configured, and the Q value corresponding to the signal-to-noise ratio is obtained from the threshold dependence of the error rate. As an evaluation index. That is, a part of the optical signal from the transmission line is converted into an electric signal by the photoelectric converter 56, and the electric signal and the clock extracted by the clock extraction circuit 53 are input to an electric signal processing means 57 such as a sampling oscilloscope. Then, an eye pattern and a light intensity histogram as shown in FIG. 10 are obtained.

At time t ₀ when the eye opening of the eye pattern is maximum, the signal amplitude (for example, voltage value) is μ (t ₀ ), and the standard deviation of “high” or “1” level noise in binary transmission is σ. _When the standard deviation of noise at ₁ (t ₀ ), “low” or “0” level is σ ₀ (t ₀ ), the Q (t ₀ ) value is Q (t ₀ ) = μ (t ₀ ) / (σ ₁ (t ₀ ) + σ ₀ (t ₀ ))… (1)
Defined by On the other hand, assuming a Gaussian noise amplitude distribution, the error rate P and Q values in the low error rate region are
P = (1 / (Q ( 2π) 1/2)) exp (-Q 2/2) ... (2)
There is a relationship like Therefore, if the Q value can be measured, the transmission channel error rate can be estimated.
NSBergano et al., "Margin Measurement in Optical Amplifier Systems", IEEE Photonics Technology Letters, vol.5, no.3, pp.304-306

  However, since the conventional Q value measurement system measures the Q value at the time when the eye opening becomes maximum, it cannot cope with digital optical signals having different bit rates.

  It is an object of the present invention to provide an optical signal quality monitor capable of monitoring a signal-to-noise ratio with a single circuit in an optical fiber transmission network in which digital optical signals of different bit rates are transmitted.

The first invention is a photoelectric converter that converts an optical signal having a bit rate N · f ₀ (bit / s) that is a natural number N times the basic clock frequency f ₀ (Hz) into an electric signal, and the basic clock frequency f _0. a timing clock generating means for generating a timing clock of slightly different repetition 1 a natural number n of ₁ minute (Hz) frequency f ₀ / n ₁ -.DELTA.f (Hz) or _{f 0 / n 1 + Δf (} Hz), the timing clock The level of the electrical signal is sampled for a certain period of time and the histogram thereof is measured. Among all the sampling points constituting the histogram, a point group higher than a predetermined first threshold level is defined as “level 1”. The point group lower than the second threshold level, which is lower than the first threshold level , is defined as “level 0”, and the average value of the level distribution within a certain average time of each of “level 1” and “level 0” And the ratio of the standard deviation values (σ ₁ + σ ₀ ) within the average time of “level 1” and “level 0” Q = μ / (σ ₁ + σ ₀ )
Is obtained as a signal-to-noise ratio coefficient, and electrical signal processing means for inspecting the quality of the optical signal is provided.

The second invention is a photoelectric converter that converts an optical signal having a bit rate N · f ₀ (bit / s), which is a natural number N times the basic clock frequency f ₀ (Hz), into an electrical signal, and the basic clock frequency f _0. a timing clock generating means for generating a timing clock of slightly different repetition 1 a natural number n of ₁ minute (Hz) frequency f ₀ / n ₁ -.DELTA.f (Hz) or _{f 0 / n 1 + Δf (} Hz), the timing clock The level of the electrical signal is sampled for a certain period of time, and the histogram is measured. Among all sampling points constituting the histogram, the points in the high level region of the two regions where the predetermined regions do not overlap The group is “level 1”, the point group in the lower level region is “level 0”, and the difference μ ′ of the average value of the level distribution within a certain average time of each of “level 1” and “level 0” "Level 1" and "level 0" ratio _{Q = μ '/ (σ 1} + σ 0) of the sum (σ ₁ + σ ₀₎ of the standard deviation of each in the mean time
Is obtained as a signal-to-noise ratio coefficient, and electrical signal processing means for inspecting the quality of the optical signal is provided.

  As described above, the optical signal quality monitor according to the present invention obtains a histogram of light intensity from all sampling points obtained within a fixed time, and calculates a signal-to-noise ratio coefficient (Q value) averaged from the histogram. The signal quality of a plurality of bit rates over a wide range from about 1 Mbit / s to several tens of Gbit / s can be monitored by a single monitor circuit.

FIG. 1 shows an embodiment of the optical signal quality monitor of the present invention.
In the figure, an optical signal having a bit rate N · f ₀ (bit / s) (N = 1, 2,...) Which is a natural number N times the basic clock frequency f ₀ (Hz) is input from the transmission line. Is branched by the optical branching device 51. At this time, the branching ratio of the monitor port with respect to the transmission path port is preferably as small as possible in order to suppress transmission characteristic deterioration due to branching loss. The optical signal branched by the optical splitter 51 is converted into an electrical signal by the photoelectric converter 61 and input to the electrical signal processing means 62. If the intensity of the optical signal input to the photoelectric converter 61 is insufficient, it may be amplified using an optical amplifier.

Timing-clock generating unit 63, the basic clock frequency f ₀ frequency f ₀ / n ₁ -.DELTA.f (Hz) after adjusting the offset frequency Delta] f (Hz) from the first natural number n ₁ minute (Hz) or f ₀ / n ₁ + Delta] f Generate (Hz) timing clock.

  The electrical signal processing means 62 samples the electrical signal using this timing clock and measures the light intensity histogram. Then, the difference between the average values of the level distribution within a certain average time of each of “level 1” and “level 0” from the sampling points constituting the histogram, and the average time of each of “level 1” and “level 0” A time-averaged Q value is obtained from the ratio of the sum of the standard deviation values, and the quality of the optical signal is inspected.

That is, the electrical signal processing means 62 detects and analyzes the peak value of the cross-correlated electrical signal, and measures a histogram of light intensity as shown in FIG. Of the sampling points constituting the histogram of light intensity, a point group higher than a predetermined threshold level μ _th1 is set to “level 1”, and a point group lower than a separately determined threshold level μ _th0 is set to “level 0”. The difference μ between the average values μ ₁ and μ ₀ within a certain average time between “level 1” and “level 0”, and the standard deviation within each average time between “level 1” and “level 0” Ratio of sum of values (σ ₁ + σ ₀ ) Q = μ / (σ ₁ + σ ₀ ) (3)
Is obtained as a signal-to-noise ratio coefficient.

Here, an example of a method for determining the threshold levels μ _th1 and μ _th0 will be described with reference to FIG. First, the electric signal processing means 62 obtains a level histogram from sampling points measured in advance within a fixed time, integrates the number of sampling points from the maximum level value of this histogram, and becomes the level when the sampling point number N _middle becomes equal. the an intermediate value mu _m. Note that the sampling point number N _middle is the total number of sampling points N _total , the optical signal duty ratio (pulse width to time slot ratio) is D, and the mark rate (probability of occurrence of mark “1” in digital transmission) is M. sometimes,
N _middle = N _total × D × M (4)
Asking. The average value μ ₁ of “Level 1” is
μ ₁ = 2 (μ _m −μ ₀ ) + μ ₀ (5)
Assume that

Next, the level at which the number of sampling points first reaches the peak value from the level minimum value side in the histogram is defined as the average value μ ₀ of “level 0”. Further, the threshold levels μ _th0 and μ _th1 of “level 0” and “level 1” are set to μ _th0 = μ ₀ + α (μ ₁ −μ ₀ )
μ _th1 = μ ₁ −α (μ ₁ −μ ₀ )… (6)
And set. However, 0 <α <0.5. Substituting (5) into this (6),
μ _th0 = 2αμ _m + (1-2α) μ ₀
μ _th1 = 2 (1-α) μ _m − (1-2α) μ ₀ (7)
Is obtained.

Based on the threshold levels μ _th0 , μ _th1 determined by the above method and the light intensity histogram obtained from all the measured values, the average values μ ₁ , μ ₀ and standard deviations σ ₁ , σ of “level 1” and “level 0” are obtained. ₀ is obtained, and an average signal-to-noise ratio coefficient (Q value) is obtained from these values.

  The signal-to-noise ratio coefficient Q is a physical quantity corresponding to the signal-to-noise ratio of the optical signal on a one-to-one basis. Therefore, the quality of the transmitted optical signal can be inspected by obtaining this Q value.

FIG. 4 shows the relationship between the average Q value obtained by the above method and the Q value at time t ₀ . As can be seen from the figure, when α is set to 0.1, the number of sampling points decreases, and the variation in average Q value increases. On the other hand, when α is 0.4 or more, since the measurement value near the cross point of the eye pattern is included, the average Q value becomes low and the measurement accuracy deteriorates. On the other hand, when 0.1 <α <0.4, the number of sampling points is sufficient and the influence of cross points can be avoided, so that a good correlation is obtained between the average Q value and the Q value at time t ₀ . . In this example, the correlation coefficient has a high value of 0.99.

Therefore, in the equation (7), by setting 0.1 <α <0.4 and determining the respective threshold levels μ _th0 and μ _th1 , the optical signal quality of the transmission line can be accurately monitored.

Also, as shown in FIG. 5, among the sampling points constituting the histogram, the point group in the high level region of the two predetermined regions is “level 1” and the point group in the low level region is “level”. 0 ”, the difference μ ′ of the average value of the level distribution within a certain average time between“ level 1 ”and“ level 0 ”, and the standard within the average time of each of“ level 1 ”and“ level 0 ” Ratio of sum of deviation values (σ ₁ + σ ₀ ) Q = μ ′ / (σ ₁ + σ ₀ ) (8)
May be obtained as a signal-to-noise ratio coefficient (claim 2).

In this case, for example, assuming that the maximum value and the minimum value are μ _max and μ _min , respectively, among the sampling points measured in advance within a certain period of time, the range of “level 1” as shown in FIG. _{May be} “μ _th1 or more and μ _max or less”, and the range of “level 0” may be “μ _min or more and μ _th0 or less” (Claim 4).

FIG. 7 shows a configuration example of the timing clock generator 63.
In FIG. 7 (a), bit rate timing generator 13 generates a basic clock frequency f clock frequency f ₀ / n ₁ is a natural number n 1 of ₁ minute ₀ from the optical signal of the bit rate N · f _0. The oscillator 14 generates an offset frequency Δf. The mixer 15 mixes the clock frequency f ₀ / n ₁ and the offset frequency Δf to generate a timing clock of frequency f ₀ / n ₁ ± Δf (Hz), and the bandpass filter 18 outputs f ₀ / n ₁ from the mixer output. Only the frequency component of −Δf (Hz) or f ₀ / n ₁ + Δf (Hz) is output.

Here, a method for setting the offset frequency Δf will be described. FIG. 8 is a time chart showing the relationship on the time axis between the electrical signal, the timing clock, and the generated sum frequency signal. Since the repetition frequency of the timing clock is smaller by Δf than the 1 / n ₁ frequency division of the repetition frequency f ₀ of the electrical signal, the timing clock pulse has a relative position of ΔT each time as shown in FIG. Sweep while shifting. This relative position shift ΔT is the difference between the timing clock repetition period Ts (= 1 / fs) and the 1 / n ₁ frequency division period n ₁ T ₀ (T ₀ = 1 / f ₀ ) of the input signal. And

It is expressed. This relative position shift ΔT is the sum of n ₂ times the period T ₀ of the input signal and the sampling step time δT. That is,
_{ΔT = n 2 T 0 + δT} ... (10)
And This means that the input signal waveform is sampled at the sampling step time δT. At this time, from the equations (9) and (10), Δf is

It becomes. In addition, n _1, n ₂ is a natural number. Therefore, the signal waveform can be sampled with a desired sampling step amount δT by setting the offset frequency Δf by the equation (11). Further, by appropriately selecting a combination of n ₁ and n ₂ , the offset frequency and timing clock can be set in accordance with the band of the signal processing system.

The timing clock generating unit 63, as shown in FIG. 7 (b), the bit rate f ₀ / from network synchronization clock signal m, the basic clock frequency f natural numbers n ₁ minute 1 clock frequency f _{0 0} / Basic bit rate timing generation means 17 for generating n ₁ may be used.

Further, FIG. 7 (a), the place of the configuration of FIG. 7 (b), as shown in FIG. 7 (c), the clock frequency f ₀ / n ₁ of the first natural number n ₁ minute of the basic clock frequency, ( An oscillator that oscillates the timing clock f ₀ / n ₁ −Δf or f ₀ / n ₁ + Δf obtained by adding or subtracting the offset frequency Δf in equation (11) may be used as the timing clock generating means.

  Here, when the timing clock generating means 63 has the configuration of FIG. 7A, an optical branching device is disposed between the optical branching device 51 and the photoelectric converter 61, and the branched optical signal is sent to the timing clock. Input to the generating means 63. When the configuration shown in FIG. 7B is adopted, a synchronous network clock signal is input to the timing clock generating means 63.

  The feature of this embodiment is that the quality of an optical signal having an arbitrary bit rate is monitored by sampling at the electric stage. However, the bit rate of the optical signal that can be handled is limited to about several tens of Gbit / s depending on the bandwidth and processing speed of the photoelectric converter and the electric signal processing means. However, it is applicable to fiber optic transmission networks where the maximum bit rate is known not to exceed this limit.

As described above, the present invention is different from a conventional error rate measurement method in which data is taken in and discriminated at a frequency equal to the bit rate of the optical signal at the best time of eye pattern opening. That is, the optical signal quality monitor of the present invention obtains a histogram of light intensity from all sampling points obtained within a certain time, and monitors the signal-to-noise ratio coefficient (Q value) averaged from this histogram. Yes, it can cope with an optical signal having a bit rate Nf ₀ (bit / s) which is an arbitrary integral multiple of the basic clock frequency f ₀ (Hz).

The figure which shows embodiment of the optical signal quality monitor of this invention. The figure explaining the setting method of the level of the light intensity histogram measured by an electrical signal processing means. The figure explaining the determination method of a threshold level. Diagram showing the relationship between the Q value in the average Q value and time t _0. The figure explaining the setting method of the level of the light intensity histogram measured by an electrical signal processing means. The figure explaining the determination method of a threshold level. The block diagram which shows the structural example of a timing clock generation means. The time chart which shows the relationship on the time-axis of an electrical signal, a timing clock, and the generated sum frequency signal. The figure which shows the structural example of the conventional error rate measurement system. The figure which shows the eye pattern and light intensity histogram of an optical signal. The block diagram which shows the measurement system of the eye pattern of an optical signal.

Explanation of symbols

13, 17 Basic bit rate timing generation means 14, 19 Oscillator 15 Mixer 18 Band pass filter 51 Optical splitter 52 Optical amplifier 61 Photoelectric converter 62 Electric signal processing means 63 Timing clock generation means

Claims (7)

A photoelectric converter that converts an optical signal having a bit rate N · f ₀ (bit / s) that is a natural number (N) times the basic clock frequency f ₀ (Hz) into an electrical signal;
A timing clock having a repetitive frequency f ₀ / n ₁ −Δf (Hz) or f ₀ / n ₁ + Δf (Hz) slightly different from a natural number (n ₁ ) of the basic clock frequency f ₀ (Hz) is generated. Timing clock generation means;
The level of the electrical signal is sampled for a certain period of time with the timing clock, and the histogram is measured. Among all sampling points constituting the histogram, a point group higher than a predetermined first threshold level is set to “level”. 1 ”, a point group lower than the first threshold level and lower than the second threshold level is set to“ level 0 ”, and the average value of the level distribution within a certain average time of each of“ level 1 ”and“ level 0 ” And the ratio of the standard deviation values (σ ₁ + σ ₀ ) within the average time of “level 1” and “level 0” Q = μ / (σ ₁ + σ ₀ )
An optical signal quality monitor comprising: an electric signal processing means for obtaining a signal-to-noise ratio coefficient and inspecting the quality of the optical signal. A photoelectric converter that converts an optical signal having a bit rate N · f ₀ (bit / s) that is a natural number (N) times the basic clock frequency f ₀ (Hz) into an electrical signal;
A timing clock having a repetitive frequency f ₀ / n ₁ −Δf (Hz) or f ₀ / n ₁ + Δf (Hz) slightly different from a natural number (n ₁ ) of the basic clock frequency f ₀ (Hz) is generated. Timing clock generation means;
The level of the electrical signal is sampled for a certain period of time with the timing clock, the histogram is measured, and among all sampling points constituting the histogram, a high level region of two regions where the predetermined regions do not overlap The point group in the level is “level 1”, the point group in the low level region is “level 0”, and the difference between the average values of the level distribution within a certain average time of “level 1” and “level 0” μ ′ And the ratio of the standard deviation values (σ ₁ + σ ₀ ) within the average time of “level 1” and “level 0” respectively Q = μ ′ / (σ ₁ + σ ₀ )
An optical signal quality monitor comprising: an electric signal processing means for obtaining a signal-to-noise ratio coefficient and inspecting the quality of the optical signal. The optical signal quality monitor according to claim 1,
The electrical signal processing means obtains a level histogram from sampling points measured in advance within a certain period of time, integrates the number of sampling points from the maximum level value of this histogram, N sampling is the _total number of sampling points, and the duty ratio of the optical signal ( When the ratio of the pulse width to the time slot is D, and the mark rate (probability of occurrence of “level 1” in digital transmission) is M,
N _middle = N _total × D × M
Level when it becomes equal to the number of sampling points N _middle which is obtained by an intermediate value mu _m,
In the histogram, the level at which the number of sampling points first reaches the peak value from the level minimum value side is the average value μ ₀ of “level 0”,
The threshold levels μ _th0 and μ _th1 of “level 0” and “level 1” are set to μ _th0 = 2αμ _m + (1-2α) μ ₀
μ _th1 = 2 (1-α) μ _m − (1-2α) μ ₀
0.1 <α <0.4
An optical signal quality monitor. The optical signal quality monitor according to claim 2,
The electrical signal processing means obtains a level histogram from sampling points measured in advance within a certain period of time, integrates the number of sampling points from the maximum level value of this histogram, N sampling is the _total number of sampling points, and the duty ratio of the optical signal ( When the ratio of the pulse width to the time slot is D, and the mark rate (probability of occurrence of “level 1” in digital transmission) is M,
N _middle = N _total × D × M
Level when it becomes equal to the number of sampling points N _middle which is obtained by an intermediate value mu _m,
In the histogram, the level at which the number of sampling points first reaches the peak value from the level minimum value side is the average value μ ₀ of “level 0”,
The threshold levels μ _th0 and μ _th1 of “level 0” and “level 1” are set to μ _th0 = 2αμ _m + (1-2α) μ ₀
μ _th1 = 2 (1-α) μ _m − (1-2α) μ ₀
0.1 <α <0.4
age,
Among the sampling points measured within the predetermined time, the maximum value and minimum value of the level are μ _max and μ _min , respectively, the range of “level 0” is set to μ _{min to} μ _th0, and the range of “level 1” _Is an optical signal quality monitor characterized by having a value of μ _{th1 to} μ _max . In the optical signal quality monitor according to claim 1 or 2,
Timing clock generation means
A frequency divider for dividing the basic clock frequency f ₀ (Hz) of the clock signal extracted from the optical signal branched from the transmission path into f ₀ / n ₁ (Hz);
When n ₂ is a natural number and δT is a sampling step time,
Δf = f ₀ (n ₁ + f ₀ δT) / n ₁ (n ₁ + n ₂ + f ₀ δT)
An oscillator that oscillates at an osset frequency Δf represented by:
A mixer that mixes the divided clock signal and the output of the oscillator to generate a timing clock having a frequency of f ₀ / n ₁ ± Δf (Hz);
An optical signal quality monitor comprising: a band-pass filter that outputs only a frequency component of f ₀ / n ₁ −Δf (Hz) or f ₀ / n ₁ + Δf (Hz) from the mixer output. In the optical signal quality monitor according to claim 1 or 2,
Timing clock generation means
A multiplier that multiplies or divides a synchronous network clock signal having a clock frequency f ₀ / m (Hz) to a frequency f ₀ / n ₁ (Hz) that is a natural number (n ₁ ) of the basic clock frequency f ₀ (Hz). Or with a divider,
When n ₂ is a natural number and δT is a sampling step time,
Δf = f ₀ (n ₁ + f ₀ δT) / n ₁ (n ₁ + n ₂ + f ₀ δT)
An oscillator that oscillates at an osset frequency Δf represented by:
A mixer that mixes the multiplied or divided clock signal and the output of the oscillator to generate a timing clock having a frequency of f ₀ / n ₁ ± Δf (Hz);
An optical signal quality monitor comprising: a band-pass filter that outputs only a frequency component of f ₀ / n ₁ −Δf (Hz) or f ₀ / n ₁ + Δf (Hz) from the mixer output. In the optical signal quality monitor according to claim 1 or 2,
Timing clock generation means
From a frequency f ₀ / n ₁ (Hz) that is a fraction of the natural number (n ₁ ) of the basic clock frequency f ₀ (Hz), when n ₂ is a natural number and δT is a sampling step time,
Δf = f ₀ (n ₁ + f ₀ δT) / n ₁ (n ₁ + n ₂ + f ₀ δT)
An optical signal quality monitor characterized by comprising an oscillator that oscillates at a frequency f ₀ / n ₁ −Δf (Hz) or f ₀ / n ₁ + Δf (Hz) obtained by adding or subtracting the Osset frequency Δf represented by

Images