Open Access
Add to BibSonomyAbstract
We demonstrate chromatic dispersion (CD) and polarization mode dispersion (PMD) monitoring techniques via simulation and experiment for 2- and 4-level intensity-modulated as well as phase-modulated optical systems. Degree of polarization (DOP) measurement for monitoring PMD up to 100-ps and clock tone measurement for monitoring CD up to 720-ps/nm are demonstrated in 10-Gsymbol/s non-return-to-zero (NRZ) and return-to-zero (RZ) systems. Analysis on dynamic range and monitoring window shows that careful consideration and characterization are necessary when applying these monitoring techniques to multi-level systems.
©2007 Optical Society of America
1. Introduction
Optical performance monitoring has emerged as a desired tool for controlling and managing the physical-state-of-the-network in a ubiquitous and cost-effective manner. There have been several reports of chromatic dispersion (CD) and polarization mode dispersion (PMD) monitoring techniques for a typical binary amplitude-shifted-keyed (ASK) data signal as well as a binary DPSK data channel [1-9]. Many of the current monitoring techniques involves: (i) histogram and eye diagram measurement [1-2], (ii) subcarrier techniques [3], (iii) tracking the power of the RF clock tone [4], (iv) phase or nonlinear detection [5-6], (v) RF spectrum analysis [7], and (vi) polarization modulation or scrambling [8-9]. For system complexity and performance consideration, no extra signal added in the transmitter is a much desired feature. Therefore, monitoring techniques based on directly observing properties of the received data signal will be required to provide real time measurement of key impairments. Consequently, those techniques will not be bit rate and modulation format transparent.
With its advantage of spectral efficiency, the topic of multi-level data modulation formats has gained much recent interest [10-12]. A better dispersion tolerance is also expected, as multi-level signal occupies less optical spectrum and with slower clock rate. Compared with binary system, multi-level system tends to have more complicated setup and smaller operability due to its higher SNR requirement. Monitoring techniques for such multi-level systems are thus highly desired, to either drive an equalizer/compensator or notify the network manager to take appropriate diagnostic or routing action. However, to the best of our knowledge, there has been no report of optical performance monitoring of CD and PMD for multi-level modulation formats.
In this paper, we explore the monitoring techniques of CD and PMD in multi-level data modulation systems. We measure the degree of the polarization (DOP) of the signal for PMD monitoring and the power of the RF clock tone for CD monitoring of multi-level intensity modulatied as well as multi-level phase modulated optical signals in 10-Gsymbol/s NRZ and RZ systems. We find that: (i) the sensitivity of DOP measurement is less in 4-level than in 2-level modulation for intensity modulation systems but almost the same for phase modulation systems, (ii) clock tone measurement is more sensitive in 2-level than in 4-level modulation for NRZ systems, but almost the same in RZ systems, and (iii) DOP value is very sensitive to the phase modulation index in phase modulation systems, and is sensitive to the pulse shape (rising and falling time) in intensity modulation. 1st order PMD (i.e., DGD, differential group delay) monitoring window up to 50-ps for RZ signal and 100-ps for NRZ signal are explored for DOP measurement. The monitoring window of chromatic dispersion up to 720-ps/nm with dynamic range over 15 dB is demonstrated for clock tone measurement. Results show that careful consideration and characterization are necessary when we apply these monitoring techniques to multi-level systems.
2. Concept and experimental setup
Figure 1 shows the conceptual diagram of multi-level modulated optical signals. A multilevel modulation encodes multiple bits (N-bits) on a single symbol. As the spectrum is mainly determined by the symbol rate, the spectral efficiency of a multi-level system is then improved by log2N compared with the conventional 2-level system. When the data signal passes through optical fiber, the deleterious degrading effects, such as CD and PMD, will not only degrade the signal but also change the value of the monitored parameter. We can measure these parameters to monitor the quality of signal after detailed characterization.
Figure 2 illustrates the experimental setup. A multi-level electrical signal is generated by combining two 10-Gbit/s pseudo-random-bit-sequence (PRBS) using an RF power combiner with an appropriate attenuation and decorrelation time delay. Compared with the 2-level signal that is obtained directly from the bit error rate tester (BERT), this multi-level signal suffers extra pulse distortion due to the power-splitter-combiner structure chosen in the experiment and measurement results will be affected if that monitoring method is pulse shape sensitive. A tunable laser (operated at 1551-nm) is then intensity or phase modulated by this multi-level electrical signal. Another intensity modulator (a pulse carver) driven by the clock signal is used for RZ data generation. The multi-level optical signal then propagates through different lengths of single mode fiber to emulate different amount of chromatic dispersion. The signal is detected by a photodiode, and the 10-GHz RF clock tone is then filtered out by an electrical filer for measurement. Meanwhile, the multi-level optical signal passes through different DGD elements to emulate the 1st-order PMD. The depolarized signal is then sent into a polarimeter for DOP measurement.
3. DOP Measurement for PMD Monitoring
DOP is defined as the ratio of the polarized power versus the total optical power. As shown in Fig. 3, when the signal is passing through a DGD element, the walk-off in time between the two principle states of polarization (PSP) of the signal will gradually decrease their overlap and thus changing the DOP of the signal. In the best case that the signal is aligned with one PSP, the signal is seeing no PMD, no degradation will occur and the DOP of the signal will remain unchanged. In the worst case (i.e., power splits equally between fast and slow axis), the measured DOP as a function of DGD can be calculated by the formula (32) in [13]:
where Rin is the autocorrelation function of the optical signal’s spectrum. We note that the autocorrelation function depends on the pulse width of the signal. The minimum DOP is generally obtained when DGD is equal to the pulse width, which determines the maximum measurable DGD value. DOP is also pulse shape dependent where finite rising time will normally result in larger measured DOP value compared with ideal rectangular pulses.
Fig. 3. DOP of the signal will decrease when the signal is passing through DGD (1st order PMD) element.
Figure 4 shows the experimental and simulated results of DOP measurements versus various DGD values for 10-Gsymbol/s multi-level signals, when the signal is split with equal power between two PSPs. The DOP value of the signal will decrease from 1 to different lower limits with respect to different amount of DGD. For phase-modulated signal, as shown in Fig.-4(a) and (b), both 2- and 4-level phase-modulated signals are almost totally depolarized when DGD equals one pulse width (100-ps for NRZ signal and 50-ps for RZ signal). However, DOP values are sensitive to the insufficient phase modulation. When the driving signal to the phase modulator is half the expected value, the DOP of the received signal becomes much higher. This can be explained by the following calculation derived from equation (1): for phase modulated multi-level NRZ signal, if we assume the input is a series of IID (independently and identically-distributed) with equal probability distribution and has ideal rectangular pulse shape:
where N is the modulation level and φi is phase for each level. If φi is evenly distributed through the whole circle of the constellation, the signal will be totally depolarized (i.e., DOP=0). On the other hand, if φi only lie on part of the circle (i.e. smaller modulation index), the signal will be partially polarized and the minimum DOP will become larger.
Fig. 4. DOP measurement (line: numerical simulation, dot: experiment): (a) NRZ phase-modulated signal, (b) RZ phase-modulated signal, (c) NRZ intensity-modulated signal, and (d) RZ intensity-modulated signal.
Results of DOP vs. DGD for intensity-modulated signals are shown in Fig. 4-(c) and (d). In all cases, the minimum DOP is obtained when DGD equals one pulse width. However, DOP is much less sensitive in 4-level than in 2-level systems. NRZ system will have higher minimum DOP compared with RZ system. In NRZ system, for ideal square wave, the minimum DOP can be calculated from formula (1) assuming that the input is a series of IID:
A higher minimum DOP is expected if N > 2. In RZ systems, the minimum DOP should be 0 with an ideal rectangular pulse shape. However, the finite rising and falling time of the pulse will decrease the DOP variation as more overlapping occurs during auto correlation. Moreover, in this experiment, the 4-level signal is more distorted (with longer rising time) due to the power-splitter-combiner structure and this will further decrease of the DOP variation.
4. Clock Tone Measurement for CD Monitoring
Chromatic dispersion arises from the frequency-dependent nature of the index of refraction of the optical fiber. As illustrated in Fig. 5, the chromatic dispersion will cause a time delay between the upper and lower sidebands of the optical signal and the detected RF component (at symbol rate) will be faded and regenerated periodically. For RZ signal, when CD equals to zero, the upper and lower sideband are in phase and thus the maximum clock components are detected. When the accumulated dispersion is not zero, the regenerated/faded clock components are changed following an approximate sinusoidal curve of
where P0 is the RF power without CD effect, fRF is the RF tone frequency, DL is the total accumulated dispersion and c is the speed of light. For NRZ signal, on the other hand, the upper and lower side band of the clock are out of phase initially in the transmitter. Consequently, compared with equation (4), a sine curve will be obtained. Fiber induced PMD effects will also fade and regenerate the clock components when the accumulated PMD is comparable to the signal time period. For modern fiber, PMD effects usually are shaded by the chromatic dispersion effects. However, there are cases (eg. old fiber with very large PMD value or very long haul with CD totally compensated) in which PMD effects should be compensated or taken into consideration.
Fig. 5. RF components (clock) will fade and regenerate periodically due to the accumulated chromatic dispersion.
For multi-level signals, compared with conventional 2-level signals, we also have the clock components in NRZ and RZ phase modulation systems as well as intensity modulation systems. The largest measurable CD window is again determined by the clock frequency (at symbol rate rather than at the bit rate). The sensitivity/dynamic range may change according to the pulse shape and modulation format.
Figure 6 shows the experimental and simulated results of clock tone measurement. In the experiment, different lengths of single mode fiber are used to emulate different amount of chromatic dispersion and the fiber-induced PMD effects are ignored. For phase modulation, in NRZ system, the maximum regenerated clock power of the 4-level system (i.e., the monitoring dynamic range) is much smaller due to the extra pulse distortion from the RF combiner (Fig. 6-(a)). The monitoring window of chromatic dispersion up to 720-ps/nm is obtained with dynamic range 25 dB and 21 dB for 4-level and 2-level phase-modulated signals, respectively. The dynamic range will be further decreased for the smaller phase modulation. In RZ system, however, since the clock component of the signal mainly comes from the RZ pulse carving, both 2- and 4- level signals have similar maximum clock tone power, even when the phase modulation is not quite sufficient (Fig. 6-(b)).
For intensity modulation, in NRZ system, the 4-level signal has smaller clock component and therefore the maximum power of the clock tone is much smaller (Fig. 6-(c)). The experimental data is further smaller due to the extra pulse distortion. The clock component in a RZ system, however, again is mainly from the RZ pulse carving. Therefore, there is little difference between 2- and 4-level systems (Fig. 6-(d)).
Fig. 6. Clock tone measurement (line: numerical simulation, dot: experiment): (a) NRZ-phase modulated signal, (b) RZ phase-modulated signal, (c) NRZ intensity modulated signal, (d) RZ intensity modulated signal.
5. Conclusion
In this paper, we analyze the modulation format dependency of chromatic dispersion and polarization mode dispersion monitoring techniques by simulation and experiment. The comparisons are carried out for 10-Gsymbol/s NRZ and RZ, multi-level intensity and phase-modulated signals. We demonstrate the monitoring of 1st-order PMD up to 100-ps using DOP measurement and chromatic dispersion up to 720-ps/nm with dynamic range over 15 dB using clock tone power measurement. The measurable window, dynamic range, and system parameter sensitivity are investigated. Results show that multi-level systems may have smaller dynamic range but generally the measureable window is not changed if the symbol rate keeps the same. Careful consideration and characterization are necessary when applying these monitoring techniques to multi-level systems.
References
1. S. D. Dods and T. B. Anderson, “Optical performance monitoring technique using delay tap asynchronous waveform sampling,” in Conf. Proc. of Optical Fiber Communication (OFC), 2006, OThP5.
2. Z. Li and G. Li, “Chromatic dispersion and polarization-mode dispersion for RZ-DPSK signals based on asynchronous amplitude-histogram evaluation,” J. Lightwave Technol. 24, 2859–2866 (2006). [CrossRef]
3. S. B. Jun, H. Kim, P. K. J. Park, J. H. Lee, and Y. C. Chung, “Pilot-tone-based WDM monitoring technique for DPSK systems,” IEEE Photon. Technol. Lett. 20, 2171–2173 (2006). [CrossRef]
4. Y. K. Lizé, L. Christen, J.-Y. Yang, P. Saghari, S. Nuccio, A. E. Willner, and R. Kashyap, “Independent and simultaneous monitoring of chromatic dispersion and polarization-mode dispersion in OOK and DPSK transmission,” IEEE Photon. Technol. Lett. 19, 3–5 (2007). [CrossRef]
5. A. L. Campillo, “Chromatic dispersion-monitoring technique based on phase-sensitive detection,” IEEE Photon. Technol. Lett. 17, 124122013;1243 (2005). [CrossRef]
6. S. Wielandy, M Fishteyn, and B. Zhu, “Optical performance monitoring using nonlinear detection,” J. Lightwave Technol. 22, 784–793 (2004). [CrossRef]
7. G.-W. Lu, M.-H. Cheung, L-K. Chen, and C.-K. Chan, “Simultaneous PMD and OSNR monitoring by enhanced RF spectral dip analysis assisted with a local large-DGD element,” IEEE Photon. Technol. Lett. 17, 2790–2792 (2005). [CrossRef]
8. Y. Shi, M. Chen, and S. Xie, “Simultaneous polarization mode dispersion and chromatic dispersion monitoring method in 40 Gbit/s system by polarization modulation,” in Conf. Proc. of Optical Fiber Communication (OFC), 2006, JThB39.
9. K. J. Park, H. Kim, J. H. Lee, C. J. Youn, S. K. Shin, and Y. C. Chung, “Polarization-mode dispersion monitoring techniques based on polarization scrambling,” Electron. Lett. 38 ,83–85 (2002). [CrossRef]
10. Y. Han and G. Li, “Theoretical sensitivity of direct-detection multilevel modulation formats for high spectral efficiency optical communications,” IEEE J. Sel. Top. Quantum Electron. 12, 57122013;580 (2006). [CrossRef]
11. S. G. Park, K. D. Hong, Y. S. Jang, and K. Kim, “On the WDM transmission using multilevel (M>4) DPSK modulation format,” IEEE Photon. Technol. Lett. 17, 154622013;1548 (2005). [CrossRef]
12. N. Kikuchi, K. Sekine, S. Sasaki, and T. Sugawara, “Study on cross-phase modulation (XPM) effect on amplitude and differentially phase-modulated multilevel signals in DWDM transmission,” IEEE Photon. Technol. Lett. 17, 1549–1551 (2005). [CrossRef]
13. S. M. R. M. Nezam, J. E. McGeehan, and A. E. Willner, “Theoretical and experimental analysis of the dependence of a signal’s degree of polarization on the optical data spectrum,” J. Lightwave Technol. 22, 763–772 (2004). [CrossRef]
