We demonstrate substantial performance improvements in 256 QAM transmission in terms of both data rate and distance that we realized by using a digital back-propagation (DBP) method. 160 Gbit/s-160 km and 64 Gbit/s-560 km transmissions were successfully achieved with a polarization-multiplexed 256 QAM signal, in which the symbol rate and transmission distance were greatly increased by compensating for the interplay between dispersion and nonlinearity, which is responsible for the transmission impairment especially for a higher symbol rate and longer distance.
© 2012 OSA
Driven by the rapid growth of data traffic in backbone networks, attempts to achieve the ultimate fiber capacity have excited profound research interest. Of the various attempts, higher-order QAM and OFDM with a multiplicity of 64~128 levels have successfully achieved an ultra-large WDM capacity in excess of 100 Tbit/s in a single fiber core by virtue of their spectral efficiency, which is as high as ~10 bit/s/Hz [1, 2]. Achieving a further increase in QAM multiplicity to, for example, 256~1024 levels, has also been the subject of intensive research with a view to extending the spectral efficiency toward the Shannon limit [3–6]. Recently a novel scheme for 256 level modulation has also been demonstrated by means of iterative polar modulation, which allows an OSNR improvement of ~1.2 dB . However, such an extremely high-order QAM inevitably suffers from both reduced tolerance to OSNR and increased susceptibility to fiber nonlinearities, which result in severe limitations as regards transmission distance and data rate. For example, in previous reports on 256 and 512 QAM [3, 4], the symbol rate remained at 3~4 Gsymbol/s, and the maximum transmission reach was limited to 160 km.
Recently, a number of electrical compensation schemes have been proposed in accordance with advances in digital coherent technologies, and applied to the compensation of linear and nonlinear impairments. Among them, the digital back-propagation (DBP) method [8,9], in which the nonlinearity is digitally compensated for by inverse fiber propagation using the nonlinear Schrödinger equation, has the potential to provide significantly improved performance as a result of enhanced OSNR by allowing a higher transmission power. This is because DBP is capable of compensating for the interplay between dispersion and nonlinearity, which is responsible for the transmission impairment especially for the higher symbol rate. DBP has been used to extend the transmission distance in 112 Gbit/s 16 QAM  and 120 Gbit/s 64 QAM  transmissions. Numerical simulations have also shown the benefit of DBP, especially for higher-order QAM . However, there have been no experimental demonstrations of the adoption of DBP for extremely high-order QAM transmission such as 256 QAM.
This paper reports the first application of DBP to 256 QAM, and describes the improved performance in terms of both a higher symbol rate and an extended maximum transmission distance.
2. Experimental setup
Our experimental setup is shown in Fig. 1 . The 256 QAM transmitter comprises a coherent light source, an arbitrary waveform generator (AWG), and an IQ modulator. As a CW laser source, we used an acetylene frequency-stabilized fiber laser emitting at 1538.8 nm with a linewidth of 4 kHz. Part of the laser output was split and its frequency was downshifted by 2.7 GHz, which was co-propagated with a QAM signal and used as a pilot tone for OPLL in the receiver. The remaining CW light was QAM modulated with a 256 QAM baseband signal generated by the AWG, which was operated at 12 Gsample/s with a 10-bit resolution. The bandwidth of the 256 QAM signal was reduced by using a digital Nyquist filter with a roll-off factor of 0.35. In addition, digital pre-equalization was employed to compensate for the non-ideal frequency response of individual components by using a finite impulse response (FIR) filter with 99 taps. The 256 QAM optical signal was then polarization multiplexed andlaunched into a transmission link, which was composed of 80 km of SSMF per span. The loss of each span was compensated for by EDFAs.
After the transmission, the signal was preamplified via a 0.7 nm optical filter, and then coupled into a polarization-diversity coherent receiver together with a local oscillator (LO). Here, MIMO processing for the polarization demultiplexing of a 256 QAM signal is very complex due to its ultra-high multiplicity. Therefore, polarization demultiplexing was carried out with a polarization controller (PC) located before the polarization diversity coherent receiver instead of MIMO processing. In this scheme, a transmitted tone signal was maximized or minimized along the two polarization principal axes of the polarization diversity coherent receiver by controlling the PC so that X- and Y-polarization signals were completely demultiplexed. As the LO, we used a frequency-tunable fiber laser with a 4 kHz linewidth, which was phase-locked to the transmitted pilot tone via OPLL. The received signals were A/D converted at 40 Gsample/s using a digital oscilloscope, and fed into an offline digital signal processor (DSP). In the DSP, we employed the simultaneous compensation of dispersion and nonlinear impairments with DBP. We carried out a split-step Fourier analysis of the Manakov equation, which describes pulse propagation in the presence of dispersion, SPM, and XPM between the two orthogonal polarizations under a randomly varying birefringence :Eq. (1) with the reversed sign of α, β2, and γ. Here we chose an FFT size of 8192, and set the step size at a distance of 10 km. Finally, the compensated QAM signal was demodulated into binary data, and the bit error rate (BER) was evaluated.
3. 64 Gbit/s (4 Gsymbol/s), 256 QAM transmission over 560 km
We first set the symbol rate at 4 Gsymbol/s and evaluated the transmission performance of a 64 Gbit/s, 256 QAM transmission over 160 km that we described in our previous report . Here we compare the performance of the simultaneous linear and nonlinear compensation with DBP for a conventional scheme of average individual compensation. In the latter case, we employed pre-compensation for the dispersion, exp(iβ2ω2L/2), and SPM, exp(iγ|Ax,y|2L), individually for the baseband data in the AWG, and XPM was not compensated.
Figure 2 shows a comparison of BER as a function of the launched power. With DBP, the optimum launched power was increased by 2 dB, which resulted in an improved BER as a consequence of the increased OSNR. The BER characteristics against the received power with the optimum launched power is shown in Fig. 3 , and the constellation diagrams after 160 km transmission with and without DBP are plotted in Fig. 4 . The blue curve in Fig. 3 is the result reported in . It can be seen that the power penalty at the FEC threshold (BER = 2x10−3) was greatly reduced from 5.3 to 0.8 dB by employing DBP. The error vector magnitude (EVM) was also improved from 2.10% to 1.82%. We confirmed that the interplay between dispersion and nonlinearity is still not significant at a symbol rate of 4 Gsymbol/s, and therefore this improvement is mainly attributed to the compensation of XPM between two orthogonal polarizations (the third term on the right hand side of Eq. (1)) with DBP.
Next we extended the transmission distance from 160 km, and examined the maximum transmission reach with DBP. The relationship between the transmission distance and BER is shown in Fig. 5 . As shown by the red closed symbols, DBP enables us to realize a BER below the FEC threshold after 320 km. In addition, the transmission distance was further extended by newly introducing a frequency-domain equalizer (FDE)  in the receiver instead of the FIR pre-equalizer, which yields better equalization performance due to its potentially high resolution. We also adopted Raman amplifiers together with EDFAs, which also contribute greatly to the OSNR increase. Here, we set the Raman gain at 11 dB with a backward pump power of 500 mW, and chose an optimal signal launch power of –4 dBm. As a result, the maximum transmission distance was extended to 560 km as shown by the red open symbols, and the performance improvement with DBP can also be clearly seen in the hybrid EDFA and Raman amplified system.
4. 160 Gbit/s (10 Gsymbol/s), 256 QAM transmission over 160 km
To further clarify the advantage of DBP, we next explored the possibility of realizing a higher data rate by employing DBP. Figure 6 compares the BER with DBP and a conventional compensation scheme for various symbol rates in a 256 QAM-160 km transmission. Here, the Nyquist filter is not employed, because oversampling of more than twice cannot be realized for a symbol rate of > 6 Gsymbol/s due to the sampling rate limitation of 12 Gsample/s in the AWG. The results for a symbol rate of ≤ 6 Gsymbol/s were also obtained without employing the Nyquist filter for consistency. It can be seen that the symbol rate can be increased from 4 to 8 Gsymbol/s with DBP. Furthermore, by introducing Raman amplifiers combined with EDFAs, even a symbol rate of 10 Gsymbol/s is possible in a 160 km transmission, whichcorresponds to a bit rate of 160 Gbit/s. These enhancements are a consequence of the DBP’s ability to compensate for the interplay between dispersion and nonlinearity, which leads to a serious penalty at such a high symbol rate.
We have successfully demonstrated substantial improvement in 256 QAM performance as regards both data rate and transmission distance by using DBP. By using EDFAs and Raman amplifiers, the transmission distance was extended from 160 to 560 km at a symbol rate of 4 Gsymbols (64 Gbit/s), and the symbol rate was increased from 4 to 10 Gsymbol/s in a 160 km transmission. These substantial improvements were successfully achieved as a result of the increased OSNR and the compensation of the interplay between dispersion and nonlinearity. Improved WDM transmission performance is also expected by reducing XPM between different WDM channels using DBP.
References and links
1. D. Qian, M. Huang, E. Ip, Y. Huang, Y. Shao, J. Hu, and T. Wang, “101.7-Tb/s (370×294-Gb/s) PDM-128QAM-OFDM transmission over 3×55-km SSMF using pilot-based phase noise mitigation,” in National Fiber Optic Engineers Conference (Los Angeles, Calif., 2011), PDPB5.
2. A. Sano, T. Kobayashi, S. Yamanaka, A. Matsuura, H. Kawakami, Y. Miyamoto, K. Ishihara, and H. Masuda, “102.3-Tb/s (224 x 548-Gb/s) C- and extended L-band all-Raman transmission over 240 km using PDM-64QAM single carrier FDM with digital pilot tone,” in Optical Fiber Communication Conference, (Los Angeles, Calif., 2012), PDP5C.3.
3. M. Nakazawa, S. Okamoto, T. Omiya, K. Kasai, and M. Yoshida, “256-QAM (64 Gb/s) coherent optical transmission over 160 km with an optical bandwidth of 5.4 GHz,” IEEE Photon. Technol. Lett. 22(3), 185–187 (2010). [CrossRef]
4. S. Okamoto, K. Toyoda, T. Omiya, K. Kasai, M. Yoshida, and M. Nakazawa, “512 QAM (54 Gbit/s) coherent optical transmission over 150 km with an optical bandwidth of 4.1 GHz,” in the 36th European Conference and Exhibition onOptical Communication (2010), PD2.3.
6. M.-F. Huang, D. Qian, and E. Ip, “50.53-Gb/s PDM-1024QAM-OFDM transmission using pilot-based phase noise mitigation,” in Optical Fiber Communication Conference (2011), PDP1.
7. X. Liu, S. Chandrasekhar, T. Lotz, P. J. Winzer, H. Haunstein, S. Randel, S. Corteselli, B. Zhu, and D. W. Peckham, “Generation and FEC-decoding of a 231.5-Gb/s PDM-OFDM signal with 256-iterative-polar-modulation achieving 11.15-b/s/Hz intrachannel spectral efficiency and 800-km reach,” National Fiber Optic Engineers Conference (Los Angeles, Calif., 2012), PDP5B.3.
8. C. Paré, A. Villeneuve, P.-A. Bélanger, and N. J. Doran, “Compensating for dispersion and the nonlinear Kerr effect without phase conjugation,” Opt. Lett. 21(7), 459–461 (1996). [CrossRef] [PubMed]
9. X. Li, X. Chen, G. Goldfarb, E. F. Mateo, I. Kim, F. Yaman, and G. Li, “Electronic post-compensation of WDM transmission impairments using coherent detection and digital signal processing,” Opt. Express 16(2), 880–888 (2008). [CrossRef] [PubMed]
10. S. Makovejs, D.S. Millar, V. Mikhailov, G. Gavioli, R.I. Killey, S.J. Savory, and P. Bayvel, “Experimental investigation of PDM-QAM16 transmission at 112 Gbit/s over 2400 km,” in Optical Fiber Communication Conference (Los Angeles, Calif., 2010), OMJ6.
11. T. Kobayashi, A. Sano, A. Matsuura, E. Yamazaki, E. Yoshida, Y. Miyamoto, T. Nakagawa, Y. Sakamaki, and T. Mizuno, “120-Gb/s PDM 64-QAM transmission over 1,280 km using multi-staged nonlinear compensation in digital coherent receiver,” Optical Fiber Communication Conference, (Los Angeles, Calif., 2011), OThF6.
12. D. Rafique, J. Zhao, and A. D. Ellis,”Performance improvement by fibre nonlinearity compensation in 112 Gb/s PM M-ary QAM,” Optical Fiber Communication Conference, (Los Angeles, Calif., 2011), OWO6.
14. K. Ishihara, T. Kobayashi, R. Kudo, Y. Takatori, A. Sano, and Y. Miyamoto, “Frequency-domain equalization for coherent optical single-carrier transmission systems,” IEICE Trans. Comm. E 92-B(12), 3736–3743 (2010).