Simultaneous PMD and OSNR Monitoring by Enhanced ... - IEEE Xplore

2790

IEEE PHOTONICS TECHNOLOGY LETTERS, VOL. 17, NO. 12, DECEMBER 2005

Simultaneous PMD and OSNR Monitoring by Enhanced RF Spectral Dip Analysis Assisted With a Local Large-DGD Element Guo-Wei Lu, Man-Hong Cheung, Lian-Kuan Chen, and Chun-Kit Chan

Abstract—In this letter, we propose and experimentally demonstrate an enhanced radio-frequency (RF) spectrum analysis technique to simultaneously monitor optical signal-to-noise ratio (OSNR) and polarization-mode dispersion (PMD) for return-to-zero ON–OFF keying (RZ-OOK) systems. In this method, by cascading a large differential group delay element at the monitoring module, the PMD and OSNR parameters can be derived from the analysis of the position shift and the minimum power of the RF spectral dip, respectively. Experimental results show that this scheme can monitor PMD from 0 to 90 ps with less than 2-ps error and OSNR from 16 to 35 dB with less than 1-dB error in a 10-Gb/s RZ-OOK transmission system with a pulsewidth of 2.5 ps. The scheme possesses the advantages of simple implementation, large monitoring range, and high PMD monitoring sensitivity by using the RF spectrum analysis only at low-frequency range. Index Terms—Optical signal-to-noise ratio (OSNR), performance monitoring, polarization-mode dispersion (PMD), radio-frequency (RF) spectrum.

I. INTRODUCTION

P

OLARIZATION-MODE dispersion (PMD) and optical signal-to-noise ratio (OSNR) are two important parameters to be monitored in high-speed optical networks, especially in the next-generation dynamic reconfigurable wavelength-division-multiplexing networks. These two optical parameters tend to vary with time, environment, or network configurations, thus, proper in-service monitoring is required. Monitoring of PMD has been demonstrated by measuring the degree-of-polarization (DOP), different radio-frequency (RF) notch frequency components, the eye opening penalty, and the phase difference of a given frequency component after the principal state of polarization (PSP) filtering [1]. On the other hand, monitoring of in-band OSNR has been demonstrated by using DOP correlation, polarization-nulling, homodyne signal nulling, and 1/2 clock frequency constellation analysis [2], [3]. However, they were significantly affected by the depolarization effect of PMD. While much research effort has been put on decoupling the two parameters for accurate in-band OSNR monitoring [4], there has been minimal research in simultaneously monitoring of both parameters in a simple manner to provide a more comprehensive performance status [5]. Therefore, it is highly desirable Manuscript received June 17, 2005; revised August 18, 2005. This work was supported in part by a research grant from Hong Kong Research Grants Council under Project CUHK4110/05E. The authors are with the Department of Information Engineering, The Chinese University of Hong Kong, Shatin, N.T., Hong Kong (e-mail: gwlu1@ie.cuhk.edu.hk). Digital Object Identifier 10.1109/LPT.2005.859162

to simultaneously monitor both PMD and OSNR using a single module to reduce the complexity of the monitoring system. Recently, we have demonstrated the simultaneous PMD and OSNR monitoring scheme by RF spectrum dip analysis with locally cascaded large differential group delay (DGD) element [6]. Here, the comprehensive analysis on the proposed scheme with more detailed experimental results is presented. In this letter, we report an effective technique to simultaneously monitor PMD and OSNR for narrow pulses by analyzing the position shift and the minimum power of the RF spectral dip induced by cascading a large DGD element with polarization scrambling at the monitoring module. Using a 2.5-ps 10-Gb/s return-to-zero ON–OFF keying (RZ-OOK) system, we show that PMD can be monitored from 0 to 90 ps with an average error of 1.7% (maximum error 5%) irrespective of the OSNR. Meanwhile, OSNR can be monitored from 15 to 35 dB with errors 1 dB irrespective of the PMD. This simple scheme has large dynamic range and high PMD monitoring sensitivity via only low-frequency RF spectrum analysis. However, as the RF spectrum will be affected by chromatic dispersion and fiber nonlinearity, it is assumed that the chromatic dispersion is compensated, and the system is operated under linear regime. II. OPERATION PRINCIPLE The operation principle of this scheme is mainly based on the RF power fading effect induced by DGD. Due to the DGD’s effect on the signal’s RF spectrum, the signal’s RF spectrum can be theoretically derived as [7] (1) where is the power splitting ratio between the two orthogonal polarizations, is the DGD of the system, is the original signal’s electrical spectrum. This illustrates that the spectrum of the signal is seriously faded by DGD, and is a cosine square function of frequency with the period inversely related to the DGD in the worst case. This means that DGD generates power dips in the resultant electrical spectrum by destructive interference between the two polarization components, and the lowest frequency of the power dip is related to the DGD by . If the DGD of the system is large enough, we can more easily and accurately measure the dip frequency, thus facilitating derivation of the DGD as the dip frequency is pushed into the more readily accessible low-frequency regime of the signal’s RF spectrum. In our proposed simultaneous PMD and OSNR monitoring scheme, a large-DGD element is intentionally introduced at the

1041-1135/$20.00 © 2005 IEEE

LU et al.: SIMULTANEOUS PMD AND OSNR MONITORING BY ENHANCED RF SPECTRAL DIP ANALYSIS

Fig. 1. Experimental setup of the simultaneous PMD and OSNR monitoring by enhanced RF spectrum analysis. FMLL: fiber mode-locked laser.

monitoring module to increase the total DGD value, thus simultaneously implementing PMD monitoring with improved monitoring sensitivity, and OSNR monitoring based on orthogonal delayed-homodyne principle [8]. Fig. 1 shows our proposed module for simultaneously monitoring PMD and OSNR. The monitoring signal is sent to an optical filter, a polarization controller (PC), and then a piece of polarization-maintaining fiber (PMF) with a large DGD value. To monitor PMD, we scramble the PC in front of the PMF and notice the position shift of the RF spectral dip (i.e., the minimum and the maximum dip frequency ) by dip frequency an RF spectrum analyzer. By considering the transmission fiber as one span of fiber and the PMF in the monitoring module as another, we can express the overall DGD of the cascaded two spans of fibers as [9] (2) ) are the DGD of the two spans, where , (assuming respectively, and is the coupling angle between the two spans. From (2), the maximum and the minimum overall DGD can be expressed as (3) (4) Hence, (5) (6) As is proportional to while is proportional to , thus we can calculate and . The cascade of a large-DGD element and the use of position shift as a monitoring parameter in our scheme actually present some advantages over the previously proposed RF power-based PMD monitoring scheme. In previous scheme, when the transmission link has a small DGD, the RF spectral dip would be located at a very high frequency and the power of the low frequency components changed very little, causing low monitoring sensitivity. The cascade of a local large-DGD element can help move the dip position to the low-frequency regime of the spectrum even when the transmission link DGD is small. Monitoring at low-frequency components with an improved sensitivity is highly desirable because it not only eliminates the use of high-speed electronics, but also desensitizes the monitoring

2791

to higher order PMD [1]. In addition, since the position shift, instead of the absolute RF power, of the dip is measured, our PMD monitoring scheme is strictly dependent on the transmission link DGD and is independent of the signal bit rate. Therefore, no prior knowledge between the RF power level and the DGD values is needed. For the PMD monitoring, one limitation of the proposed scheme is that the PMD monitoring accuracy may be affected by the input state of polarization (SOP) [10]. To eliminate the input SOP effect for accurate measurement, in practice, the modulated signal can be fast scrambled to assure the even power split between PSPs [11]. In order to monitor OSNR by the same proposed module, we note the minimum dip power during PC scrambling. The minimum power corresponds to the receiver noise power within a narrow resolution bandwidth after signal elimination. This noise power is predominantly the beat noise power, which is inversely proportional to the OSNR. The relationship of minimum dip power versus OSNR can be precalibrated at different optical powers. We can then estimate the OSNR by measuring the optical power and the electrical noise power. This method has been demonstrated to monitor OSNR in the absence of PMD, and is also referred to as orthogonal delayed homodyne method [8]. It is found that the presence of the link PMD may change the dip frequency but the minimum power will remain fairly constant within a considerable PMD range. III. EXPERIMENTS AND RESULTS The proposed monitoring scheme was experimentally demonstrated in a 2.5-ps 10-Gb/s RZ-OOK system. A pulse source, generated from a 10-Gb/s 1550.3-nm fiber mode-locked laser, was externally modulated by a LiNbO intensity modulator with pseudorandom binary sequence. An amplified spontaneous emission (ASE) noise source was generated by cascading two erbium-doped fiber amplifiers. Two attenuators were inserted in both signal and noise branches. Thus, the OSNR can be changed by combining the signal with the ASE noise source at different power levels. A DGD emulator (ProDelay, General Photonics), which can digitally vary the amount of first-order PMD from 0 to 90 ps with small second-order PMD of less than 85 ps , was inserted in the link to simulate different system DGD values. At the receiver side, a small portion (10%) of the signal power was sent to the monitoring module. The PMF used had a DGD of 409 ps and the nulling frequency was at 1.22 GHz. A tradeoff has to be made when considering the DGD value of the PMF. A large DGD will move the spectral dip further to the low frequency part for the ease of measurement; however, the position shift will be smaller for the same transmission link DGD value, resulting in a smaller monitoring sensitivity. It also reduces the measurement dynamic range hindered by the lower dip frequency. A suggested DGD value used was 200 ps which produced a dip at 2.5 GHz. In this case, the dip position changed for about 50 MHz when the transmission link DGD was only 2 ps. Fig. 2(a) shows the RF spectrum when the PMD emulator was set at 40 ps with the local-DGD element of 409 ps. The minimum and maximum dip frequencies measured were 1.113 and 1.355 GHz, respectively, giving an estimated PMD of 40.1 ps. Whereas in Fig. 2(b), a DGD of 68.1 ps was derived, with the PMD emulator set at 70 ps, from the measured minimum and

2792

IEEE PHOTONICS TECHNOLOGY LETTERS, VOL. 17, NO. 12, DECEMBER 2005

Fig. 2. RF spectra with the power dip shift with the large-DGD element of 409 ps, PMD emulator of (a) 40, (b) 70, and (c) 20 ps under 35-dB OSNR.



maximum dip frequencies of 1.047 and 1.465 GHz, respectively. Fig. 2(c) shows the RF spectrum when the OSNR was set at 35 dB and the PMD emulator at 20 ps with the local-DGD element of 409 ps. The minimum power measurement gave an OSNR estimation of 34.8 dB. In our experiment, the monitoring time is around 2 min, which needs further reduction to desirable millisecond range in PMD compensation systems. The monitoring time in our experiment is mainly limited by our RF spectrum analyzer. It can be further reduced by using specifically designed RF spectrum analysis circuit. Fig. 3 confirms the PMD and OSNR monitoring functionality of our proposed scheme. By analyzing the RF power dip shift during scrambling the PC in front of the local large-DGD element, the link PMD can be estimated. The PMD monitoring has an average error of 1.7% and maximum error 5%, with OSNR varying from 15 to 35 dB. On the other hand, by using the minimum RF power dip to derive the noise power and measuring the total optical power by power-meter, the OSNR can be estimated. The results show that the monitoring errors for OSNR were 1 dB, with PMD varying from 0 to 90 ps. Due to the equipment availability in our laboratory, the DGD and OSNR monitoring was only experimentally demonstrated with DGD varying from 0 to 90 ps. The main factors that limit the maximum amount of DGD are the DGD value of the locally cascaded DGD element, the duty cycle and bit rate of the signal. IV. CONCLUSION A new and simple monitoring technique is demonstrated to simultaneously monitor both PMD and OSNR of narrow pulses. This technique analyzes the position shift of the RF spectral dip for PMD estimation and the minimum dip power for OSNR estimation. In a 10-Gb/s RZ-OOK system with a pulsewidth of 2.5 ps, the proposed scheme can monitor PMD from 0 to 90 ps with an average error of 1.7% and OSNR from 15 to 35 dB with errors 1 dB. Even though the technique is demonstrated in a 2.5-ps 10-Gb/s RZ-OOK system, it is applicable to NRZ and

Fig. 3. (a) PMD monitoring results for 10-Gb/s 2.5% RZ-OOK by proposed method versus that by PMD emulator, OSNR varying from 15 to 35 dB; (b) OSNR monitoring results for 10-Gb/s 2.5% RZ-OOK by proposed method against that by optical spectrum analyzer (OSA), PMD varying from 0 to 90 ps.

RZ-OOK systems at high speed ( 40 Gb/s), where monitoring PMD effect is desirable. REFERENCES [1] A. E. Willner, S. M. R. M. Nezam, L.-S. Yan, Z. Pan, and M. C. Hauer, “Monitoring and control of polarization-related impairments in optical fiber systems,” J. Lightw. Technol., vol. 22, no. 1, pp. 106–125, Jan. 2004. [2] M. Petersson, H. Sunnerud, M. Karlsson, and B.-E. Olsson, “Performance monitoring in optical networks using stokes parameters,” IEEE Photon. Technol. Lett., vol. 16, no. 2, pp. 686–688, Feb. 2004. [3] D. C. Kilper, “Optical performance monitoring,” J. Lightw. Technol., vol. 22, no. 1, pp. 294–304, Jan. 2004. [4] M.-H. Cheung, L.-K. Chen, and C.-K. Chan, “A PMD-insensitive OSNR monitoring scheme based on polarization-nulling with off-center narrow-band filtering,” in Proc. OFC 2004, Los Angeles, CA, 2004, Paper FF2. [5] L.-S. Yan, Y. Q. Shi, S. Yao, and A. E. Willner, “Simultaneous monitoring of both OSNR and PMD using polarization techniques,” in Proc. ECOC 2003, Rimini, Italy, 2003, Paper We4.P.133. [6] 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,” in Proc. ECOC 2004, Stockholm, Sweden, 2004, Paper We4.P.092. [7] G. Ishikawa and H. Ooi, “Polarization-mode dispersion sensitivity and monitoring in 40-Gbit/s OTDM and 10-Gbit/s NRZ transmission experiments,” in Proc. OFC’98, San Jose, CA, 1998, Paper WC5. [8] C. Y. Joun, K. J. Park, J. H. Lee, and Y. C. Chung, “OSNR monitoring technique based on orthogonal delayed-homodyne method,” IEEE Photon. Technol. Lett., vol. 14, no. 10, pp. 1469–1471, Oct. 2002. [9] M. Wegmuller, S. Demma, C. Vinegoni, and N. Gisin, “Emulator of firstand second-order polarization-mode dispersion,” IEEE Photon. Technol. Lett., vol. 14, no. 5, pp. 630–632, May 2002. [10] I. P. Kaminow and T. L. Koch, Optical Fiber Telecommunications, IIIA. New York: Academic, 1997. [11] T. Luo et al., “PMD monitoring by tracking the chromatic-dispersioninsensitive RF power of the vestigial sideband,” IEEE Photon. Technol. Lett., vol. 16, no. 9, pp. 2177–2179, Sep. 2004.