Eye quality monitor for a 2R regenerator

[0002] The invention is directed to monitoring the quality of a signal received over an optical network, and in particular, to an eye quality monitor for a data regenerator.

[0003] Optical signals suffer between the transmitter and receiver from two principally different groups of degradations that will cause bit errors: noise and distortion. The causes, behavior and remedies for these groups are different. The primary sources of noise are the receiver noise (i.e. shot, thermal), optical bandwidth, interferometric cross-talk, laser noise, reflections, etc.

[0004] Distortion is defined as any inaccurate replication of a signal transmitted over a communication link, and could be referred to any network element (NE) along the link. The pulse distortion in a fiber optic system may, for example, be caused by some parts of the light pulses following longer paths (modes) than other parts. The primary sources for distortion are chromatic dispersion, inter-symbol interference, non-linearity of the elements and transmission medium, receiver frequency response, etc. In addition, in amplified wavelength division multiplexed (WDM) systems, the transmission characteristics vary from one channel to another due to the non-flat gain and noise profile of erbium-doped fiber amplifiers (EDFAs). Distortion can be measured by assessing the difference between the wave shape of the original signal and that of the signal at the network element of interest, after it has traversed the transmission link.

[0005] In the last decade, the transmission rates of data signals have increased very rapidly, along with the demand for receivers with high sensitivity. For high rate transmission, such as rates over 10 Gb/s, the signal corruption introduced by the transmission channel is a critical parameter. Numerous methods have been proposed to overcome the difficult problem of measuring or estimating the signal quality. However, they are often based on the same basic principles, or use the same equipment, and differ mainly in how the primary measurement is analyzed.

[0006] Some signal quality monitoring methods require detection, synchronization, demultiplexing, and then some analysis of the sampled signal. Even in the best cases, the results are uncertain and it is difficult to ascertain if the monitoring method mimics the behavior of the final receiver.

[0007] Other methods are based on power monitoring of each channel. As these are averaging methods, they do not sense the pulse distortion. Moreover, the precision required in the spectral measurement is prohibitive in itself.

[0008] The extent of signal degradations may be directly measured using an eye closure diagram, which is the graphic pattern produced on an oscilloscope when the detected signal is applied to the vertical input of an oscilloscope and is synchronized with the instrument time base. Changes in the eye opening indicate intersymbol interference, amplitude irregularities, or timing problems. For a binary signal, the eye diagram has a single eye, which is open or closed to an extent determined by the signal degradation. An open eye pattern is desired.

[0009] Currently, the time base of the oscilloscope is triggered using a clock signal necessarily extracted from the transmission in order to capture the eye diagram. Consequently, a network provider cannot measure the quality of the optical signal without a clock extract circuit. This prior art method also fails to separate the eye closure due to distortion from that due to noise.

[0010] A receiver regenerates the signal presented to it by interpreting the levels of the received signal according to a decision level, defined also as threshold level, or as a slicing level. Generally, binary data regenerators are provided with a fixed threshold level selected so as to yield the best error rate at a predetermined signal power level.

[0011] When the extent of signal degradations must be assessed, the current way of doing so is to recover the clock at the site of the regenerator, thus destroying bit rate transparency. Such methods not only increase the complexity, and hence the cost of the equipment at the regenerator site, but also are dependent on the rate of the signal traveling along the link.

[0012] For example, U.S. Pat. No. 4,823,360 (Tremblay et al., issued Apr. 18, 1989 and assigned to Northern Telecom Limited) discloses a device for measuring quality of a signal travelling along an optical fiber, using eye closure. The device described in this U.S. patent provides the receiver with three threshold levels for recovering data. Two of the thresholds V1 and V2 are obtained by setting them above and below the center of the eye for a preset error rate, and the third threshold is provided in a selected relationship to the other two. If V1 and V2 are set for equal ‘error’ rates, then the central circuit operates near the middle of the eye and hence with a negligible true error rate.

[0013] The technique described in the '360 patent is based on recovering the signal clock, i.e. implies knowing the rate of the channel. In addition, the measurement does not give an indication as to the separate contribution of the noise and the distortion.

[0014] However, not all regenerators installed in a transmission link are provided with means for recovering the clock. From this point of view, regenerators may be classified as 1R, that only regenerate the signal, 2R that regenerate and reshape the signal, and 3R, that regenerate, reshape and retime the signal. 2R regenerators are used, for example, at sites where the transmission signal needs to be converted from one wavelength (e.g. 1.3 μm short reach) to another (e.g. a wavelength on the ITU grid suitable for dense WDM).

[0015] There is a need to measure the quality of an optical signal, while maintaining bit rate transparency, so that the measurements may be applied to signals of various rates.

[0016] It is an object of this invention to provide a method and apparatus for measuring the quality of a signal of a communication network, that alleviate totally or in part the drawbacks of the current methods and apparatuses.

[0017] It is another object of the invention to provide a method and apparatus for measuring the quality of a signal in a communication network which are bit rate transparent.

[0018] Still another object of the invention is to obtain a measurement of the quality of the optical signal, which can be used to assign the cause of errors to noise or to distortion.

[0019] According to one aspect of the invention, there is provided a method of monitoring the quality of an optical signal at a regenerator site, without performing a clock recovery operation, comprising, generating a threshold voltage V_THto take a plurality of values according to a pattern, applying the threshold voltage V_THon a first input of a slicer, and applying an input voltage V_inon a second input of the slicer to obtain a slicer output voltage V_S, the input voltage V_inbeing an electrical equivalent of the optical signal, for the plurality of threshold voltages, obtaining a corresponding plurality of associated parameters; and processing all the associate parameters as a function of all the threshold voltages to simulate an eye diagram of the optical signal.

[0020] According to a further aspect of the invention, there is provided an eye quality monitor for a data regenerator for providing a simulated eye diagram of an optical signal without performing a clock recovery operation, comprising a pattern generator for generating a threshold voltage V_THaccording to a pattern, a slicer, for receiving a the threshold voltage V_THon a first input and an input voltage V_inon a second input, and generating a slicer output voltage V_S, the input voltage V being an electrical equivalent of the optical signal, an average detector for providing an associated average voltage V_AVof the slicer output voltage V_Scorresponding to the a threshold voltage V_TH, and means for processing the average voltages V_AV, to simulate an eye diagram of the optical signal.

[0021] An important advantage of the present invention is that the quality of the signal can be measured with no need to recover the bit rate of the signal, which is very important in the case of 2R receivers, where the clock is not available. This results in important savings on equipment at such receivers.

[0022] Another advantage of the invention is that it can be applied to various signal rates and is independent of the technology used. Thus, the method can be applied to the existing SONET technology, and also to emerging optical transport networks (OTN).

[0023] The foregoing and other objects, features and advantages of the invention will be apparent from the following more particular description of the preferred embodiments, as illustrated in the appended drawings, where:

[0024] FIG. 1 shows the block diagram of a repeater site equipped with an eye quality monitor according to the invention;

[0025] FIGS. 2A to 2C show the time diagrams for a triangular input signal applied to the eye quality monitor of FIG. 1;

[0026] FIG. 3 shows the ideal transfer function of the eye monitor, i.e. the variation of the output voltage from the eye monitor with the threshold voltage for a triangular input signal;

[0027] FIGS. 4A and 4B show how the eye diagram is reconstructed for a triangular input signal;

[0028] FIGS. 5A to 5C show the simulated eye monitor transfer function for transmitted signals with various degrees of dispersion and noise. FIG. 5A refers to an undistorted signal, FIG. 5B, to a signal with eye closure due to dispersion, and FIG. 5C, to a signal with eye closure due to polarization mode dispersion (PMD);

[0029] FIGS. 6A to 6C show reconstructed eye diagrams for the signals in FIGS. 5A to 5C;

[0030] FIG. 7 is another embodiment of the eye monitor according to the invention; and

[0031] FIG. 8 is still another embodiment of the eye monitor according to the invention.

[0032] FIG. 1 shows an eye monitor 10 provided at a regenerator site A for measuring the quality of the eye of an optical signal S_in. In this Figure, an incoming optical signal S_inarrives at site A over a fiber span 5 and an outgoing optical signal S_outleaves site A over fiber span 5′. Site A is equipped with an optical receiver 12, an amplifier 14 and an optical transmitter 16, which form a 2R regenerator, adapted only to regenerate and reshape the signal. As such, no clock recovery circuitry is provided at site A. Receiver 12 comprises an optical-to-electrical converter 11, for converting S_ininto the electrical equivalent input signal V_in. A comparator 13 compares V_inwith a decision level to separate the signal from noise. The recovered electrical signal V is then amplified in amplifier 14. Optical transmitter 16 receives the amplified variant of V, modulates it over an optical carrier and transmits the optical outgoing signal S_outover fiber 5′.

[0033] Site A is also equipped with a control unit 18 which provides the decision level for receiver 12 and controls operation of receiver 12, amplifier 14, and transmitter 16. A digital-to-analog converter (DAC) 15 converts the slicing signal output by control unit 18 into the decision level for receiver 12.

[0034] Receiver 12, amplifier 14 and transmitter 16 are illustrated for showing the connection of the eye monitor 10.

[0035] Eye monitor 10 comprises a slicer 17 and an average detector 19. The slicer 17 receives a threshold voltage V_THon the negative input and input signal V_inand on the positive input. The slicer output V_Sof the slicer 17 is above logic ‘zero’ whenever V_inis higher than the threshold voltage V_TH.

[0036] Average detector 19 receives V_Sand provides the mean value V_AVof slicer output V_Sover a given interval. This mean value decreases when the number and duration of pulses in V_Sdecreases. In general, the closer V_THis to the maximum value of V_in, the shorter the time when V_inis above V_TH, which results in shorter pulses for V_S. Shorter V_Spulses means lover V_AVvalues.

[0037] FIGS. 2A to 2C show time diagrams for a triangular input signal V_inapplied to the eye quality monitor of FIG. 1, and FIG. 3 illustrates the ideal transfer function of the eye monitor, i.e. the variation of V_AVwith the threshold voltage V_THfor a triangular input signal.

[0038] When slicing level V_THis at a level of 50% of the level of a logic “1”, the mean value of V_Sis 0.5. This is shown in FIG. 2A. If the slicing level is moved up to 75% of the “1” level, the mean of V_Sbecomes 0.375, as shown in FIG. 2B. If the slicing level is moved up to 99.9% of the “1” level, the mean of V_Sbecomes 0.25, shown in FIG. 2C.

[0039] FIG. 3 illustrates a V_AV−V_THdiagram for a triangular input signal. Both V_AVand V_THare normalized with respect to the “1” level of V_in. A value of V_AV=0.875 is approximated for a threshold V_THcorresponding to V_TH=0 (“long 0s threshold) and a V_AV=0.125 is approximated to a threshold corresponding to V_TH=1 (“long 1s threshold”). The points marked with A, B and C correspond to the mean values of V_Sin FIG. 2A, 2B and respectively 2C.

[0040] To obtain the V_AV−V_THdiagram, the threshold voltage V_THis linearly swept across the entire range of the output of average detector 19, V_AV. A pattern generator 21 also shown in FIG. 1, gives various values to the threshold voltage V_THaccording to a pattern. In this example, pattern generator 21 is illustrated as being part of control unit 18, where it can be implemented in software. Pattern generator 21 may also be provided as an independent unit, so that a threshold voltage V_THas necessary for the invention may be obtained by any other known means.

[0041] The signal at the output of block 19 is shown in FIG. 4A, illustrating diagrams a, b, c and d obtained with the same procedure of linearly increasing the threshold voltage form 0% to 100% of V_AV, and repeating this pattern. These diagrams are similar to the V_AV−V_THdiagram shown in FIG. 3, with the abscissa and the ordinate axes interchanged, and the values of V_AVscaled by a factor of 2. The diagrams are preferably one bit apart.

[0042] The average voltage thus obtained is input to a reflected version unit 23 that provides a variant of the average voltage, “reflected” with respect to the V_THaxis of coordinates. This reflected variant is denoted herein with RV_AV. A succession of reflected variants is also illustrated on FIG. 4B, where diagrams denoted with a′, b′ and respectively c′, are reflected versions of diagrams a, b and c.

[0043] To construct the eye of the signal, the output of the average detector 19 and the output of the reflected version unit 23 are added in add unit 25, together with lines “1” and “0” at 100% and at 0% respectively. The output of the add unit 25 is the reconstructed eye for V_in, which is denoted on FIG. 1 with V_EYE.

[0044] The reflected variants may be obtained in any suitable way, in software or in hardware. It is to be understood that the reflected variants may be obtained by linearly decreasing the threshold voltage V_THacross the entire range of the average voltage V_AV, rather than calculating the reflected variants.

[0045] It is to be understood that units 23 and 25 can also be implemented as separate units, or can be implemented in control unit 18, as shown in FIG. 1. Preferably, control unit 18 is a microcontroller. A DAC 27 is also needed to convert the digital value of V_THinto the analog value applied at the input of slicer 17.

[0046] FIGS. 5A-5C shows the transfer function V_TH−V_AVof the slicer measured with and without noise, and for input signals V_inwith various grades of distortion. The transfer functions in FIG. 5A are prepared for an undistorted of V_in. The transfer functions in FIG. 5B are for a V_inwith dispersion, and in FIG. 5C, for a V_inwith polarization mode dispersion (PMD).

[0047] FIGS. 6A to 6C show simulated eye diagrams for the signals V_inused for preparing the transfer characteristics of the slicer shown in FIGS. 5A to 5C, i.e. for an undistorted signal, a signal with dispersion and a signal with PMD. Superimposed on the eye diagrams for V_inare the reconstructed eyes V_EYEfrom the invention both with and without noise added to the signal V_in.

[0048] The results could be communicated by control unit 18 to the network management system, or they could be used to raise an alarm when the signal quality is under a provisioned value. This provisioned value may be any parameter defining the size of the eye.

[0049] Other implementations of the eye quality monitor are also possible based on similar principles, as shown in FIGS. 7 and 8.

[0050] FIG. 7 shows another embodiment of the eye monitor according to the invention. The output V_Sof slicer 17 is applied to a frequency counter 31, which counts the number of times V_inrises above V_THin a given period. The threshold V_THis varied to cover the entire range of V_in, using a microcontroller shown at 18. The output of the microcontroller 18 is converted into an analog value by a DAC 25, so that V_THtakes as many values as necessary for a given resolution. The values measured by the frequency counter 31 are provided to microcontroller 18.

[0051] As in the previous case, a transfer characteristic is plotted, this time a ‘frequency count (F)’−‘threshold voltage (V_TH)’ diagram, by varying V_THfor a given V_inand measuring the corresponding values of the frequency counter 31. F is plotted in relative units with respect to the maximum value of the counter for the respective V_in. V_THis plotted relative to the peak-to-peak voltage.

[0052] The transfer characteristics are interpreted to determine a measure of the distortion of the eye. This method is still under investigation for determining what other parameters of V_inmay be determined on F−V_THdiagrams.

[0053] FIG. 8 shows the block diagram of still another embodiment of the eye quality monitor according to the invention. In this embodiment, the optical incoming signal V_inis fed to one input of slicers 17 and 17′, the second input of the slicers receiving a corresponding threshold V₁, V₂. V₁and V₂are received from microcontroller 18 and a respective first and second DAC 25 and 25′. The output V_S1of slicer 17 and output V_S2of slicer 17′ are fed to an exclusive OR gate 35, and the output V_Sof the XOR gate is averaged to obtain V_AVin average detector 19. As is well known, the output of XOR gate 35 becomes ‘logic 1’ whenever one input is ‘logic 1’ and the other is ‘logic 0’, and the output is ‘logic 0’ whenever the inputs are both ‘logic 1’ or ‘logic 0’.

[0054] Appropriate manipulation of thresholds V₁and V₂results in an average output V_AVthat estimates the probable eye shape of the optical signal with sufficient accuracy for determining the quality of transmission.

[0055] One method of exploring various aspects of the eye diagram of V_inaccording to the embodiment of FIG. 8 is performed by first setting V₁to a voltage above signal V_in, therefore setting one input to XOR gate 35 permanently low. Threshold V₂is then set so that the output of average detector 19 V_AVis 0.5. In other words, V_Sis ‘logic 1’ for half the time and “0” for half the time. The value measured will correspond to the ‘50% level’ on the eye diagram.

[0056] Next, threshold V₂is set to a voltage above signal V_inand the settings for V₁required for values of V_AVof 0.125 and 0.875 are determined. The values measured correspond to the 87.5% level and the 12.5% level on the eye diagram.

[0057] V₂is then returned to the value corresponding to a V_AVof 0.5, and V₁is swept between the 0.125 and 0.875 values while measuring the averaged output at each point.

[0058] An estimation of the average eye shape may now be obtained by plotting various values of the V₁vs (0.25+V_S)*2 when V₁>V₂and V₁vs (0.25−V_out)*2 when V₁<V₂.

[0059] Still another way of obtaining information about the eye of V_inis to scan V₁and V₂through the eye with a small offset of for example 2%.