Dispersion post compensated 160 Gbit/s transmission systems are optimized for a wide range of transmission fiber and DCF input powers. The simulation results for SMF, NZDSFHD, NZDSFLD fiber types and for NRZ, RZ, CSRZ and CSRZ-DPSK modulation formats are presented. CSRZ-DPSK modulation, balanced receiver (Bal-Rx) and NZDSFHD transmission fiber combination found to be superior to others, giving 880 km system reach with Q > 6. The alternating dispersion compensation is then applied and optimized for various modulation format and fiber types. This simple dispersion management technique provided significant system reach extension of 60% when CSRZ-DPSK modulation, Bal-Rx and SMF transmission fiber combination is used.
© 2005 Optical Society of America
One alternative of expanding the capacity of transmission systems is to increase the bit rate per DWDM channel. 10 Gbit/s WDM systems are standard in today’s backbone networks and 40 Gbit/s WDM equipment is commercially available and has been tested by network operators in numerous field trials [1,2]. Recently, there has been a considerable interest for the next generation systems with TDM data rate of 160 Gbit/s [3–6]. But with the enhancement of the single-channel data rate to 160 Gbit/s physical impairments, i.e., dispersion and dispersion slope lead to tight system tolerances . The appropriate selection of modulation format and fiber type is critical to manage linear and nonlinear impairments of ultra-high speed optical transmission systems.
Compactness of modulation spectrum, simplicity of transmitter and receiver (Tx/Rx), and tolerance to fiber non-linearity are the three criteria for selecting a modulation format . The performance of different modulation formats are compared and discussed in many studies for 10 and 40 Gbit/s systems [7–9] and Return-to-Zero (RZ) found to be superior compared to conventional Non-RZ (NRZ) systems in terms of nonlinearity tolerance. However, NRZ has a simpler Tx/Rx configuration and a higher dispersion tolerance due to its more compact modulation spectrum. Carrier Suppressed-RZ (CSRZ) and CSRZ Differential-Phase-Shift-Keying (CSRZ-DPSK) are the two other RZ-like modulation formats with improved spectral efficiency and non-linear tolerance [7,10]. The selection of fiber type is the other critical issue for transmission links. It has been shown that G.652 standard single mode fiber (SMF) is advantageous compared to G.655 non-zero dispersion shifted fiber (NZDSF) in 160 Gbit/s systems for NRZ and RZ modulation formats, while NZDSF is advantageous at 40 Gbit/s . The performance comparison of SMF and NZDSF, however, is still unknown for other modulation formats at 160 Gbit/s systems.
In this paper, reach limits of single channel 160 Gbit/s transmission systems are investigated by means of numerical simulations, for NRZ, RZ, CSRZ and CSRZ-DPSK modulation formats and for SMF, NZDSFHD (NZDSF-High Dispersion) and NZDSFLD (NZDSF-Low Dispersion) fiber types. The analysis is extended further for two different receiver configurations of CSRZ-DPSK, single ended receiver (Single-Rx) and balanced receiver (Bal-Rx) and for alternating dispersion compensation scheme . The CSRZ and CSRZ-DPSK modulations are compared at 40 Gbit/s and 160 Gbit/s bit rates. We have demonstrated that at 160 Gbit/s, CSRZ-DPSK-Single-Rx shows the worst performance amongst all the RZ-like modulation formats whereas at 40 Gbit/s it is superior to CSRZ for high input powers as also shown by Miyamoto et. al. The 3 dB receiver sensitivity advantage offered by the DPSK type modulation formats still causes CSRZ-DPSK-Bal-Rx scheme to have the longest system reach at 160 Gbit/s. When we employ alternating compensation scheme with fiber type of SMF and modulation format of CSRZ-DPSK-Bal-Rx, we have found it to be superior at 160 Gbit/s bit rate, thereby giving 1280 km system reach.
The system setup is shown in Fig. 1
At the transmitter side, a pseudorandom bit sequence (PRBS) of length 210-1 is used. For RZ signals, bit rate of B = 160 Gbit/s and a duty-cycle of 0.4 with Gaussian shaped pulses is chosen leading to a temporal full width at half maximum of tFWHM = 2.5 ps. The transmission link is formed of a cascade of post compensated spans. Each span consists of 80 km of transmission fiber (SMF, NZDSFHD, or NZDSFLD) and an appropriate DCF in order to fully compensate for the dispersion slope and accumulated dispersion in the transmission fiber. Our numerical simulations show that (results not presented here) at 160 Gbit/s full dispersion compensation is optimum for RZ modulation. The fiber parameters are given in Table 1. We then varied the input powers of transmission fiber and DCF independently from each other to find the maximum system reach limit . Two EDFAs in front of transmission fiber and DCF, with 4.5dB noise figure each, are used to adjust input power levels. At the receiver side, signal is amplified to 0 dBm power level and filtered by a 1st order optical band-pass Gaussian filter with 2B = 320 GHz bandwidth. The optical signal is transformed into an electrical signal by a PIN photodiode. The electrical signal is filtered by a 5th order low-pass Bessel filter with 1.4B = 224 GHz bandwidth before the analyzer. For CSRZ-DPSK receivers, the structure described in  is used with the same parameters given above. The Q-factor is used to describe transmission performance. The target value of Q-factor is chosen to be 6, which corresponds to BER of 10-9 for a Gaussian distribution of ASE noise.
3. Results and Discussion
The Q-factor was calculated for varying input powers of transmission fiber and DCF at the end of each span. The number of spans with Q > 6 is found for each transmission fiber and DCF input power pair. Figure 2 shows system reach contour plots of four modulation formats with respect to average input powers of transmission fiber and DCF when SMF is used.
It is obvious that system reach has a maximum at certain sets of input power levels balancing the ASE noise and nonlinear signal distortions. For RZ modulation this maximum number of spans is 10 for only one power set, 2 dBm and -3 dBm (2/-3 dBm) for SMF and DCF inputs, respectively. Since, there is no margin for power variations of the system, 10 span reach has no practical importance. So, it is more realistic to accept larger area of 9 spans as the maximum system reach for RZ. NRZ modulation shows a much worse performance compared to RZ as it only reaches 7 spans with the desired signal quality. CSRZ modulation, with its enhanced spectral efficiency and nonlinearity tolerance, reaches to a maximum of 10 spans in a wide range of input power sets. The center of the 10 span region is found to be around 3.5/-2 dBm. CSRZ-DPSK-Bal-Rx has a maximum system reach of 11 spans for 6 dBm SMF input power and 1 dBm DCF input power. Since power variation tolerance at that input power set is very tight, maximum system reach of 10 spans is a more realistic value for CSRZ-DPSK-Bal-Rx. Bal-Rx configuration inherently offers 3dB higher sensitivity . Therefore, the system reach contour plot of Single-Rx configuration is also presented to allow an accurate performance comparison with other modulation formats. As shown in Fig.2, when Single-Rx is used CSRZ-DPSK only reaches to a maximum of 7 spans. This suggests that at 160 Gbit/s the superior performance of CSRZ-DPSK is only due to the 3 dB higher sensitivity of the Bal-Rx configuration.
Before making any further comments on this result, we first compared CSRZ with CSRZ-DPSK-Single-Rx at 40 Gbit/s for SMF type fiber. Optical and electrical filter bandwidths chosen to be 2B = 80 GHz and 1.4B = 56 GHz, respectively, for fair comparison with 160 Gbit/s. The results are shown in Fig. 3. Unlike the results at the bit rate of 160 Gbit/s, there is a broad region where CSRZ-DPSK-Single-Rx has a longer reach than CSRZ at 40 Gbit/s. Although CSRZ reaches 21 spans, it is clear that for higher powers (PSMF > 5 dBm) CSRZ-DPSK-Single-Rx has a longer system reach than CSRZ, which is because of the enhanced self phase modulation (SPM) tolerance . At lower powers CSRZ still shows better performance. When we compare CSRZ-DPSK-Single Rx and CSRZ in Fig. 2 and Fig. 3, we come to a conclusion that the system impairments due to increased bit rate have a larger effect on CSRZ-DPSK modulation than CSRZ.
In Fig. 4 modulation formats are compared for NZDSFHD fiber type. RZ modulation format has a system reach limit of 9 spans and NRZ reaches to 7 spans, equal to the results of SMF type fiber. However, maximum system reach is achievable in a smaller region than SMF for RZ and NRZ modulations. CSRZ reaches to 10 spans as in SMF but this time the center of the maximum system reach region is found to be around 2.5/-3.5 dBm. CSRZ-DPSK-Bal-Rx reaches to 11 spans in a much wider region than for SMF, with a center at 5/-2 dBm. The CSRZ-DPSK-Single-Rx reaches to 8 spans for a single input power set 4/-3 dBm. This is the longest system reach for CSRZ-DPSK-Single Rx amongst all the fiber types.
In Fig. 5 modulation formats are compared for NZDSFLD fiber type. RZ modulation format reaches 8 spans and NRZ reaches to 6 spans, both 80 km shorter than SMF system reach limits. These results suggest that SMF is favorable for RZ and NRZ modulation formats at 160 Gbit/s, in agreement with the conclusion of . CSRZ again reaches to a maximum of 10 spans for NZDSFLD where the center of the 10 span region is at around 1.5/-5.5 dBm. CSRZ-DPSK-Bal-Rx reaches to 11 spans in a slightly narrower range than NZDSFHD with a center at 3.5/-1 dBm. CSRZ-DPSK-Single-Rx reaches to 7 spans.
As the dispersion decreases from 17 ps/nm/km (SMF) to 4 ps/nm/km (NZDSFLD) the length of DCF required to compensate accumulated dispersion in transmission fiber gets shorter, and this reduces the influence of nonlinear effects due to the DCF. In agreement with this statement, the simulation results of three different fiber types in Figs. 2, 4, and 5 show that dependence of the system performance on the DCF input power level decreases with decreasing dispersion coefficient of transmission fiber.
If we compare the results of all the fiber types for RZ and NRZ modulation formats there is a reduction in the system reach with the decreasing dispersion coefficient. This shows that higher local dispersion causes RZ and NRZ pulses to be more robust against SPM, and reduces the effect of intersymbol interference. Higher dispersion causes rapid pulse broadening which reduces the nonlinear distortion of the pulses. So, decreased dispersion degrades performances of RZ and NRZ modulations due to nonlinear pulse distortion effects.
CSRZ and CSRZ-DPSK modulation formats showed different behavior than RZ and NRZ modulations when fiber type changed. This result can be attributed to their enhanced nonlinearity tolerance and spectral efficiency. As opposed to RZ and NRZ, CSRZ performance did not change significantly as the dispersion changed. System reach of 10 spans is identical for all fiber types. The medium dispersion coefficient value caused CSRZ-DPSK modulation to show its best performance in NZDSFHD fiber type which can be best explained by the optimized interplay of nonlinearity and dispersion at this dispersion value. However, if we pay attention to the center of maximum system reach region, it has a clear trend moving to the bottom left corner of the plots as the dispersion coefficient decrease from 17 ps/nm/km (SMF) to 4 ps/nm/km (NZDSFLD), for all of the modulation formats. This can be an indication for CSRZ and CSRZ-DPSK modulation formats to have a larger SPM tolerance when fibers with higher dispersion used.
There is a 20% increase in maximum achievable system reach from 9 spans (720km), when RZ modulation is used, to 11 spans (880km), when CSRZ-DPSK modulation and Bal-Rx configuration is used. Although, the transmission system configuration is different, the 20% increase is in agreement with . A further significant extension of system reach can be achieved by applying the concept of alternating compensation, which means to have an alternating order of post- and pre-compensated sections as described in . The alternating effects of pulse broadening in a post-compensated span and pulse compression in a pre-compensated span cancel each other and thus the initial pulse shape is restored efficiently. Figure 6 shows the corresponding setup.
The components and parameters are identical to those used for dispersion post-compensated system simulations. The optimization results of CSRZ-DPSK modulation format in combination with SMF and NZDSFHD type fibers are presented in order to find the maximum system reach that is possible by using alternating dispersion compensation. Figure 7 shows the results. This technique provides 6 span longer transmission with Q > 6, reaching to 16 spans when SMF is used. NZDSFHD fiber type gives a maximum of 15 spans, 4 spans longer than the pure post-compensated scheme. This simple dispersion management technique results a significant 60% (37%) extension in the transmission length for SMF (NZDSFHD) fiber type. Although all the results are not presented our simulations showed that alternating dispersion compensation is most efficient when fiber type with the largest dispersion (SMF) is used. For RZ and CSRZ modulation formats the improvement gained by this technique is found less than 25% for all the fiber types.
We have optimized the transmission fiber and DCF input power levels of dispersion-post-compensated 160 Gbit/s transmission systems for a wide range of parameters. We have used three different fiber types (SMF, NZDSFHD, NZDSFLD) and NRZ, RZ, CSRZ and CSRZ-DPSK modulation formats to find the most efficient combination, giving longest system reach. We have demonstrated that performance degradation due to physical impairments at 160 Gbit/s is more severe for CSRZ-DPSK modulation format compared to CSRZ. However, 3 dB receiver sensitivity advantage of Bal-Rx still makes CSRZ-DPSK modulation format favorable.
We have shown that for RZ and NRZ modulation, SMF transmission fiber is advantageous compared to others. However for CSRZ modulation, system reach does not change significantly when different fiber types are used. Finally for CSRZ-DPSK modulation, NZDSFHD fiber type is optimum.
When alternating dispersion compensation scheme is applied to the transmission system, we have achieved a significant 60% reach extension compared to the pure post-compensation for CSRZ-DPSK modulation with balanced receiver and SMF fiber type, giving 1280 km system reach. It has been found that the reach extension is highest when CSRZ-DPSK is used as the modulation format. The advantage of alternating dispersion compensation technique reduces when the dispersion of transmission fiber is decreased.
References and links
1. Lucent Technol. Press Release, http://www.lucent.com/press/0501/010523.nsa.html, 2001
2. Lightreading News, “MCI Claims ULH Speed Record”, http://www.lightreading.com/document.asp?doc_id=54478&site=lightreading, 2004
3. U. Feiste, R. Ludwig, C. Schubert, J. Berger, C. Schmidt, H.G. Weber, B. Schmauss, A. Munk, B. Buchold, D. Briggmann, F. Kueppers, and F. Rumpf, “160 Gbit/s transmission over 116 km field-installed fibre using 160 Gbit/s OTDM and 40 Gbit/s ETDM,” Electron. Lett. 37, 443–445 (2001) [CrossRef]
4. J.P. Turkiewicz, E. Tangdiongga, G. Lehmann, H. Rohde, W. Schairer, Y.R. Zhou, E.S.R. Sikora, A. Lord, D.B. Payne, G.-D Khoe, and H. de Waardt, “160 Gb/s OTDM networking using deployed fiber,” IEEE J. Lightwave Technol. 23, 225–235 (2005) [CrossRef]
5. T. Tokle, B. Zsigri, C. Peucheret, and P. Jeppesen, “Generation and transmission of 160 Gbit/s polarisation multiplexed RZ-DBPSK-ASK signal,” Electron. Lett. 41, 433–435 (2005) [CrossRef]
6. S. Vorbeck and R. Leppla, “Dispersion and Dispersion Slope Tolerance of 160-Gb/s Systems, Considering the Temperature Dependence of Chromatic Dispersion,” IEEE Photonics Technol. Lett. 15, 1470–1472 (2002) [CrossRef]
7. Y. Miyamoto, A. Hirano, S. Kuwahara, M. Tomizawa, and Y. Tada, “Novel modulation and detection for bandwidth-reduced RZ formats using duobinary-mode splitting in wideband PSK/ASK conversion,” IEEE J. Lightwave Technol. 20, 2067–2078 (2002) [CrossRef]
8. M.I. Hayee and A.E. Willner, “NRZ versus RZ in 10-40-Gb/s dispersion-managed WDM transmission systems,” IEEE Photonics Technol. Lett. 11, 991–993 (1999) [CrossRef]
9. D. Breuer and K. Peterman, “Comparison of NRZ- and RZ- modulation format for 40 Gbit/s TDM standard-fiber systems,” IEEE Photonics Technol. Lett. 9, 398–400 (1997) [CrossRef]
10. A. Hirano, Y. Miyamoto, and S. Kuwahara, “Performances of CSRZ-DPSK and RZ-DPSK in 43-Gbit/s/ch DWDM G.652 single-mode-fiber transmission,” in OFC 2003, 2, pp. 454–456
11. B. Konrad and K. Petermann, “Optimum fiber dispersion in high-speed TDM systems,” IEEE Photonics Technol. Lett. 13, 299–301 (2001) [CrossRef]
12. D. Breuer, K. Jürgensen, F. Küppers, A. Mattheus, I. Gabitov, and S. K. Turitsyn, “Optimal schemes for dispersion compensation of standard monomode fiber based links,” Opt. Commun. 140, 15–18 (1997) [CrossRef]
13. C. Peucheret, N. Hanik, R. Freund, L. Molle, and P. Jeppesen, “Optimization of pre- and post-dispersion compensation schemes for 10-Gbits/s NRZ links using standard and dispersion compensating fibers,” IEEE Photonics Technol. Lett. 12, 992–994 (2000) [CrossRef]
14. P.J. Winzer, S. Chandrasekhar, and Kim Hoon, “Impact of filtering on RZ-DPSK reception,” IEEE Photonics Technol. Lett. 15, 840–842 (2003) [CrossRef]
15. K. Yonenaga, S. Aizawa, N. Takachio, and K. Iwashita, “Reduction of four-wave mixing induced penalty in unequally spaced WDM transmission system by using optical DPSK,” Electron. Lett. 32, 2118–2119 (1996) [CrossRef]
16. Xu Zhenbo, K. Rottwitt, and P. Jeppesen, “Evaluation of Modulation Formats at 160 Gb/s Transmission Systems Using Raman Amplification,” in LEOS 2004. The 17th Annual Meeting of the IEEE, pp 613–614