Performance Analysis of Direct Detection-Based Modulation Formats for WDM Long-Haul Transmission Systems PRLightCOM Broadband Solutions Pvt. Ltd. Bangalore, Karnataka, INDIA Abstract During the last decade, several new modulation formats were investigated to increase the transmission distance and capacities of WDM long haul transmission systems, operating at bit rates of 10-40 Gb/s using direct detection receivers [1]-[7].The choice of the optimum modulation format is key to building flexible and cost effective high capacity WDM networks. In this white paper, we validate and compare the performance of several key direct detection-based modulation formats such as NRZ-OOK, RZ-DPSK, and RZ-DQPSK using the newly developed software OFC-SystemSoft. The validation and performance comparison results obtained by OFC-SystemsSoft agree well with a reasonable accuracy of the results obtained by the other published results [23]-[28], Hence, this white paper demonstrates the effectiveness of OFC-SystemsSoft as an efficient design tool for systems engineering of WDM long haul terrestrial transmission systems. 1.0 Introduction Since the deployment of WDM technology from mid-1990 s, the capacities of long haul terrestrial transmission systems have been steadily increasing over the years due to increase in data traffic growth in many parts of the world [1]-[7]. Although NRZ-OOK format has been utilized at bit rates up to 10 Gb/s in WDM metro and long haul terrestrial systems, because of its simplicity & low cost. But, NRZ-OOK format require higher OSNR levels and incurs significant transmission penalties due to fiber nonlinearities & PMD effects for 40G transmission systems. Therefore, many long haul experiments and field trials were conducted in the past several years to investigate the performance of other intensity modulation formats such as RZ-OOK, Duobinary, CSRZ, as well as phase modulation formats such as NRZ/RZ-DPSK and NRZ/RZ-DQPSK for 40Gb/s transmission systems. However, RZ format is less susceptible to fiber nonlinearities than NRZ-OOK, but it has a wider signal spectrum yielding lower tolerance to chromatic dispersion. On the contrary, duobinary signal has a compact signal spectrum and hence yields better tolerance to chromatic dispersion, but it is susceptible to fiber nonlinearities. Although the carrier suppressed RZ (CSRZ) format seems to perform well for long haul transmission systems in terms of chromatic dispersion and nonlinearity tolerance levels, but requires about 3 db higher OSNR than the DPSK format. The design goal is to find an optimum modulation format in terms of lower BER performance, better OSNR tolerance, high spectral efficiency, longer transmission distance, higher tolerances to fiber nonlinearities, chromatic, & polarization dispersion effects. During the last decade, several commercial photonic simulators such as VPITransmissionmaker [8], OptiSystem [9], & OptSIM [10] have been utilized to study the new modulation formats by several researchers. These photonic simulators utilize time-domain/wave propagation models and the numerical solution to the wave equation was obtained using split step Fourier method. The system performance in these simulators was measured in terms of signal spectra, eye
diagrams, OSNR, and Q factor or BER. Although the obtained simulation results are accurate, but the simulation time is quite expensive particularly for the performance analysis of WDM transmission systems. Hence, the new software OFC-SystemsSoft, (solely based on analytical models) has been developed to alleviate these issues. Thus, a systems engineer can utilize OFC-SystemsSoft to perform quick analysis and optimization of system performance for WDM long haul links. This paper is organized as follows. A typical WDM long haul transmission system that is used in this paper is described in Section 2.0 along with the description of BER models used in the evaluation NRZ-OOK, RZ-DPSK, and RZ-DQPSK formats. The OFC-SyetmsSoft validation results are presented for each of the above formats in Section 3.0. The channel power optimization results are presented in Section 4 for each of the modulation format. Section 5.0 gives the performance comparison results of these formats Finally, the conclusions are given in Section 6. 2. WDM Long-Haul Transmission Systems We consider dispersion compensated WDM long haul transmission link that comprises of several standard SMF spans having dual stage EDFAs and DCF modules as shown in Figure 1. We assume that the WDM transmitter terminal consist of an array of WDM transmitters, a MUX device and a EDFA booster amplifier. The power per channel or the launch power into the fiber span is determined by the total power of the booster amplifier. The WDM receiver terminal consists of a DMUX, an array of PIN direct detection receivers. A practical optical bandpass filter & a single detection PIN receiver along with an electrical low pass filter are utilized in the case of NRZ-OOK format. But, in the case of RZ-DPSK/DQPSK formats, we assume matched optical bandpass filters and PIN based balanced dual detection receivers. However, it should be noted that optimum bandwidths for practical optical & electrical filters should be chosen to avoid OSNR penalties in DPSK and DQPSK systems [21]. Figure 1. Block diagram of a typical WDM long-haul transmission system Next we discuss the BER models of modulation formats, which we have considered in this paper. In the case of NRZ-OOK format, the Q factor at the end of WDM link is estimated based on the standard Q factor formula, after calculating the variances of ASE noise, receiver noise, and XPM and FWM impairments [5]. We assume optimum filter bandwidths in case of NRZ-OOK format for the variance calculations.
It should be noted that the photodetection noise is non-gaussian, but the Gaussian model compares quite well in practice for the NRZ-OOK format, but does not yield accurate Q factor results for the DPSK and DQPSK formats [11]-[19]. It is well known that the main performance limiting factor in phase modulated systems is the nonlinear phase noise induced along the link which is due to the interaction of ASE noise with fiber nonlinearities [17]-[18]. Hence, the BER performance analysis of WDM transmission systems for DPSK/DQPSK formats is not an easy task due to the non-gaussian nature of the induced nonlinear phase noise and thus require accurate BER models. Several researchers attempted different methods such as Saddle Point approximation technique [15] and semi analytical KLSE approaches to predict the BER for a single channel DPSK [16] or DQPSK transmission systems, including Monte Carlo simulations [17]. These approaches are time consuming for WDM transmission systems and are not feasible. We consider the impact of the accumulated nonlinear phase noise along with ASE noise in the BER evaluation of DPSK and DQPSK systems, by extending the single channel BER theoretical exact model to WDM long haul transmission systems [18]-[19]. The receivers in DPSK/DQPSK formats are based on interferometer based dual detection balanced configurations along with matched optical bandpass filters. The predicted BER performance results for DPSK/DQPSK formats are expected to be 1 to 2 orders of magnitude lower than the practical results due to the use of matched optical bandpass filters. Furthermore, the BER computations in DPSK & DQPSK systems assume accumulated AEE noise and inter-channel XPM [20] as dominate impairments in this study, but the intra-channel XPM & FWM effects, optical and electrical filtering, including receiver imperfections are neglected [21]-[22]. We assume 50% duty cycle for the RZ pulses in DPSK & DQPSK formats in the system performance calculations.. 3.0 OFC-SystemSoft Validation Results In this Section, the performance results of OFC-SystemsSoft for each modulation format are validated & compared with either the simulation or experimental results obtained by other published results [21]-[25] in terms of BER versus channel power and BER versus distance for WDM long haul links. First, the validation results for NRZ-OOK format are presented in Section 3.1 and the results for RZ-DPSK & RZ-DQPSK formats are discussed in Sections 3.2 and 3.3 respectively. 3.1 NRZ-OOK Format The DWDM link parameters given in [23] and [24] are used for validation in terms of Q factor versus channel power & Q factor versus span counts. Figure 3(a) shows the Q factor versus channel power in a nine channel 10 Gb/s NRZ, DWDM transmission system with a spacing of 12.5 GHz over four 80 km NZDSF spans, having 75% inline compensation. It can be seen from Figure 2(a), that the maximum Q factor of close to 6 has been predicted by OFC-SystemsSoft, with an optimum channel power of around -6 dbm, which is in close agreement with the results of [23].
Figure 2(a). Q factor versus channel power Similarly, Figure 2(b) shows the comparison of Q factor versus the span count in a 10 Gb/s DWDM link over a standard SMF having 9 channels with 50 GHz spacing for a channel power of 0 dbm, having full in-line compensation [24]. Figure 2(b). Q factor versus span count The comparison of results shown in the above Figure 2(b) depicts that the Q factor is within 1 order of magnitude from the simulation results of [24] over higher span counts. 3.2 RZ-DPSK Format RZ-DPSK format is validated for the WDM link parameters given in [25] in terms of BER versus channel power. Figure 3(a) shows the BER performance versus channel power in a fully compensated WDM long haul link system over a NZDSF having 18 channels with a spacing of 133 GHz at a line rate of 42.8 Gb/s. It can be seen from Figure 3 (a), that the optimum channel power of about -3 dbm has been predicted by OFC-SystemsSoft that is in close agreement with the results of [25].
Figure 3(a). BER versus channel power The BER performance versus transmission distance is shown in Figure 3 (b) for this system with an optimum channel power of -3 dbm, that predicts a maximum transmission distance of over 6000 km for a BER of 1E-6. Figure 3(b). BER versus transmission distance 3.3 RZ-DQPSK Format The DWDM link parameters given in [26] are used to validate RZ-DQPSK format in terms of BER versus transmission distance. The BER performance versus distance is shown in Figure 4 for the transmission of nine 25 Gb/s channels with spacing of 25 GHz over a standard SMF assuming 100% inline compensation at a nominal channel power of 0 dbm. Figure 4. BER versus transmission distance The obtained BER performance versus distance is in close agreement with the results of [26].
4.0 Performance comparison of modulation formats First, Figure 5 shows the BER versus OSNR curves for back to back performances of NRZ-OOK, RZ-DPSK, and RZ-DQPSK formats at a bit rate of 42.8 Gb/s at 1550 nm. Figure 5. BER versus OSNR back to back performances of DD modulation formats. The difference in back to back OSNR requirements to achieve the same BER depends on the pulse shape, format type, optical and electrical filtering, and receiver structures [27]. It can be seen from Figure 5, NRZ-OOK format require much higher OSNR of about 16 db to achieve a BER of 1E-3 compared to the phase modulation formats. On the contrary, RZ-DQPSK and RZ-DPSK formats require about 12 db and 10 db respectively to achieve the same BER. The difference in OSNR levels between RZ-DPSK and RZ-DQSPK formats is due to the crosstalk arising between constructive and destructive components of the QPSK gnal in the MZDI demodulation process, resulting in loss of signal power. However, the use of balanced dual detection receivers in phase modulation formats yield an inherent receiver sensitivity advantage of about 3 db, compared to NRZ-OOK format. Next, the transmission performance comparison of all the above modulation formats at 42.8 Gb/s is discussed in this Section. We assume a fully compensated DWDM system with 16 channels having 50 GHz channel spacings with a fiber span length of 80 km. The obtained results by OFC-SystemsSoft are presented for each format in terms of BER versus channel power & transmission distance over thirty SMF spans. Figure 6. BER versus channel power for NRZ-OOK, RZ-DPSK & RZ-DQPSK formats
It can be seen from Figure 6, RZ-DPSK format yields lowest BER of about -4.8 at a channel power of 6.3 dbm into SMF at the accumulated nonlinear phase shift of 1 radians. Similarly, RZ-DQPSK format yields a BER of -3 at a channel power of 5.8 dbm into SMF corresponding to a accumulated nonlinear phase shift of 0.9 radians. In the case of NRZ-OOK format, the obtained minimum BER of -1.7 at a channel power of 10 dbm into SMF. The optimum channel powers into DCF modules are 2 dbm & -10 dbm in case of RZ-DPSK/RZ-DQPSK and NRZ-OOK formats respectively (see power contour plots). 5.0 Optimization of Channel Powers The channel powers launched into the transmission and dispersion compensating fibers play a key role towards achieving the minimum BER in a long haul DWDM link. In this Section, the optimized power contour plots generated by OFC-systemsSoft are shown in Figures 7-9 for all the three modulation formats. First, the power contour plot for NRZ-OOK format is shown in Figure 7. As can be seen from Fig 7, the optimized channel powers into SMF and Figure 7. Power contour plot for NRZ-OOK format and DCF are 10 dbm and -10 dbm respectively to achieve a maximum Q factor of 2. Similarly, the optimized channel powers into SMF and DCF modules are found to be 6.3 dbm & 2 dbm for the RZ-DPSK format as depicted in Figure 8. Figure 8. Power contour plot for RZ-DPSK format
The power contour plots generated by OFC-SystemsSoft are shown in Figure 9 for the RZ-DQPSK format. The optimized channel powers into SMF and DCF modules are found to be 5.3 dbm and 2 dbm respectively, as similar to RZ-DQPSK format. Figure 9. Power contour plot for NRZ-DQPSK format Next, Figure 10 shows the BER versus transmission distance curves for all the three modulation formats for the same system operating at 42.8 Gb/s over thirty SMF spans. Figure 10. BER versus transmission distance for NRZ-OOK, RZ-DPSK, & RZ-DQPSK formats As can be seen from Fig. 10, the maximum transmission distance achieved is 700 km, 1500 km, and 2100 km for NRZ-OOK, RZ-DQPSK, and RZ-DPSK formats respectively for a BER of 1E-6. These results indicate that RZ-DPSK format yields the longest transmission distance among all the modulation formats, which agree with the results of [27 ]-[28]. 6.0 Performance comparison summary In this Section, we summarize the performance of all three modulation formats at 42.8 Gb/s over thirty SMF spans in the following Table.1, with respect to optimum launch powers, maximum transmission distance, and PMD penalties. The required OSNR and maximum transmission distance in Table 1 correspond to a BER of 1E-6.
Modulation Format NRZ-OOK RZ-DQPSK RZ-DPSK Optimized Launch powers 10 dbm (SMF) -10 dbm (DCF) 5.8 dbm (SMF) 2 dbm (DCF) 6.3 dbm (SMF) 2 dbm (DCF) Table 1 Required OSNR PMD penalty Maximum Distance 22 db 0.52 db 700 km 16.6 db 0.05 db 1500 km 15.6 db 0.3 db 2100 km It can be seen from Table 1 that NRZ-OOK format require higher OSNR levels and incurs more PMD penalty and yields lower reach than the phase modulation formats. Whereas the RZ-DPSK format requires much lower OSNR than NRZ-OOK format and gives a maximum reach. The difference in OSNR levels in the case of RZ-DPSK and DQPSK formats is about 1 db and the nonlinear tolerance of RZ-DPSK is better than RZ-DQPSK format. But RZ-DQPSK offers less PMD penalty & high spectral efficiency than RZ-DPSK format and yields higher tolerance to narrowband filtering in 50 GHz DWDM systems. 7.0 Conclusion This white paper described the use of OFC-SystemsSoft for the systems engineering of WDM long haul transmission systems using different modulation formats. The performance comparison results obtained by OFC-SystemsSoft reveal that the RZ-DPSK is the optimum modulation format for 40 Gb/s WDM long haul transmission systems, which confirms with the published results of [1]-[3] and [27]-[28]. Hence, OFC-SystemsSoft can be utilized for fast systems engineering calculations to perform analysis, design, and optimization or trade-off studies of WDM long haul transmission systems.
References [1] P. J. Winzer and Rene-Jean Essiambre, Advanced Optical Modulation Formats, Proc. of the IEEE, vol. 94, No. 5, May 2006, pp. 952-985. [2] Rajdi Agalliu and M. Lucki, Benefits and limits of modulation formats for optical communications, Optics and Optoelectronics, Vol. 12, No. 2, 2014, PP. 160-167. [3] T. Hoshida et al., Optimal 40Gb/s modulation formats for spectrally efficient long-haul DWDM systems, IEEE J. Lightwave Technology, vol. 20, No. 12, 2002, pp. 1989-1996. [4] N.S. Bergano, & C.R. Davidson, Wavelength division multiplexing in long-haul transmission systems, IEEE Journal of Lightwave Technology, Vol.14, No. 6, 1996, pp. 1299-1308. [5] R. Papannareddy, Lightwave Communication Systems: A Practical Perspective, Penram International Publishing, Bombay, India, 2005. [6] G. Keiser, Optical Fiber Communications, McGraw Hill, New Delhi, 5 th Edition, 2013 [7] R. Ramaswami and K. Sivarajan, Optical Networks: A Practical Perspective, San Francisco: Morgan Kaufman, 1998. [8] VPITtransmissionMaker, VPI Photonics Inc., Berlin, Germany. [9] OptiSystem, OptiWave Corporation, Ottawa, Canada. [10] OptSim, Synopsys Inc, Ossining, New York, U.S.A. [11] G. Bosco and P. Poggiolini, On the Q factor inaccuracy in the performance analysis of optical direct detection DPSK systems., IEEE Photonic Tech. Letters, Vol. 16, No. 2, 2004, pp. 655-667. [12] C.C. View et al., BER estimation of optical WDM RZ-DPSK systems through the differential phase Q, IEEE Photonic Tech. Letters, Vol. 16, No.12, 2004, pp. 2619-2621.. [13 ] K. Ho, Impact of phase noise to DPSK signals: A comparison of different models, IEEE Photonic Tech. Letters, Vol. 10, No. 3, 2004, pp. 1403-1405. [14] C. Hu et al., Differential phase-shit keying for high spectral efficiency optical transmissions, IEEE J. Selected Areas in Quantum Electronics, Vol. 10, No.2, 2004., pp. 281-293. [15] N.S Avlonitis and E. M. Yeatman, Performance evaluation of optical DQPSK using saddle point approximation, IEEE J. Lightwave Tech., Vol. 24, no. 3, 2006, pp 1176-1185. [16] L.D. Coelho, et al., Modeling nonlinear phase noise in differentially phase-modulated optical communication systems, Optics Express, Vol. 17, no. 5, pp. 3326-3239. [17] Y Yadin, et al., Nonlienar phase noise in phase modulated WDM fiber optic communications,, IEEE Photonic Tech. Letters, Vol. 16, No. 5, 2004, pp. 1307-1309. [18] K. Ho, Performance degradation of phase modulated systems due to nonlinear phase noise,, IEEE Photonic Tech. Letters, Vol. 15, No. 3, 2003, pp. 1213-1215.. [19] K. Ho, Exact error probability of DQPSK signal with nonlinear phase noise, IEEE Lasers and Electro-Optics, CLEO/Pacific Rim 2003. P. 91. [20] T-K Chiang et al., Cross phase modulation in fiber links with multiple optical amplifiers and dispersion compensators, IEEE Journal of Lightwave Technology, Vol. 14, No. 3, 1996, pp. 249-260. [21] P. J. Winzer et al., Impact of filtering on RZ-DPSK reception, IEEE Photonics Technology Letters,, Vol. 15, 2003, pp. 840-842. [22] P. J. Winzer et al., Degradations in Balanced DPSK receivers, IEEE Photonics
Technology Letters, Vol. 15, 2003, pp. 1282-1284. [23] H. Louchet et al., Analytical model for the design of multispan of DWDM transmission systems, IEEE Photonics Technology Letters, Vol. 17, No. 1, 2005, pp. 247-249. [24] S. Pachnicke et al., Physically constrained routing in 10 Gb/s DWDM networks including fiber nonlinearities and polarization effects, IEEE Journal of Lightwave Technology, Vol. 24, No. 9, 2006, pp. 3418-3426. [25] J.X, cai et al., Transmission of 40 Gb/s WDM signals over 6250 km of conventional NZ-DSF with > 4 db FEC margin, Proc. of OFC Post deadline papers, PDP 26 [26] P. S. Cho, Transmission of 25 Gb/s RZ-DQPSK signals with 25G HZ spacing over 1000 km of SMF 28 fiber, IEEE Photonics Technology Letters,, Vol. 15, No.3, 2003, pp. 473-475. [27] M. Daikoku et al., Performance comparison of modulation formats for 40 Gb/s DWDM transmission systems, Optical Communication, paper OFN2. [28] B. Linlin et al., Comprehensive assessment of new modulation techniques in 40 Gb/s optical communication systems, 3 rd International Photonics and OptoElectronics Meetings, Journal of Physics: Conference Series, 276, 2011, pp. 1-6.