Open Access
Add to BibSonomy
Abstract
We report a polarization-multiplexed 320 Gbaud, 64 QAM coherent Nyquist pulse transmission with a frequency-stabilized mode-locked laser and a modified digital back-propagation method for pulse transmission. Using a combination consisting of a mode-locked laser and a pulse shaper, we obtained a Nyquist pulse with a high OSNR of 51 dB. We achieved error free operation under a back-to-back condition with the OSNR improvement. By developing a new digital back-propagation method for pulse propagation, we achieved a bit error rate below the 7% forward error correction limit of 2x10−3 for all the tributaries of the OTDM signal data after a 150 km transmission. As a result, single-channel 3.84 Tbit/s data were successfully transmitted over 150 km with a spectral efficiency of 10.6 bit/s/Hz.
© 2017 Optical Society of America
1. Introduction
Recently, single-channel transmission beyond 1 Tbit/s with a high spectral efficiency has become an important research subject as regards dealing with the rapid growth in information traffic. Electrical or optical time division multiplexing (ETDM/OTDM) is used to increase single-carrier symbol rates with the aim of increasing the capacity per channel. A transmitter speed of 90 Gbaud for a polarization-division multiplexed (PDM)-64 quadrature amplitude modulation (QAM) signal has already been reported using ETDM, and this approach offers the possibility of achieving a bit rate of 1.08 Tbit/s/ch [1]. Furthermore, a single-carrier 125 Gbaud PDM-16 QAM (1 Tbit/s) transmission over 80 km has also been demonstrated with a combination of 2ETDM and 2OTDM [2]. However, it is difficult to further increase the bit rate by using ETDM because of the bandwidth limitation of electronic devices such as digital-to-analog (DA) and analog-to-digital (AD) converters. On the other hand, OTDM transmission with an optical Nyquist pulse is an attractive method for simultaneously achieving a single-channel transmission beyond Tbit/s and high spectral efficiency [3–6]. In our previous work [6], we demonstrated a 1.92 Tbit/s/ch, PDM-64 QAM coherent Nyquist pulse transmission at 10 Gbaud x 16 OTDM (160 Gbaud) over 150 km with a spectral efficiency of 10.6 bit/s/Hz. In that case, a coherent Nyquist pulse was generated from a frequency-stabilized CW laser combined with an optical comb generator. To further increase the baud rate, it is important to generate shorter coherent Nyquist optical pulses with a high optical signal-to-noise ratio (OSNR) directly from a laser.
In this paper, we demonstrate the first polarization-multiplexed 320 Gbaud, 64 QAM (single-channel 3.84 Tbit/s) 150 km transmission with a 1.55 μm, 10 GHz frequency-stabilized mode-locked fiber laser (MLFL) [7] as a transmitter. Furthermore, by using a new digital back-propagation (DBP) method for optical pulses, we obtained a bit error rate (BER) below the 7% forward error correction (FEC) limit of 2x10−3 for all the tributaries of the OTDM signal data after a 150 km transmission. As a result, we achieved a single-channel 3.84 Tbit/s 150 km transmission with a spectral efficiency of 10.6 bit/s/Hz.
2. Experimental setup for single-channel 3.84 Tbit/s, 64 QAM coherent optical pulse transmission
Figure 1 shows the configuration of a conventional Nyquist pulse source combining a CW laser and an optical comb generator. The comb generator consisted of a dual-drive LiNbO3 (LN) Mach-Zehnder modulator with a half-wave voltage of 1.8 V [8,9]. The input power to the LN modulator was 10 mW, which corresponds to the maximum input power. We set the modulation signal power of the LN modulator at 2W, corresponding to the modulation index of 7.7π. The spectral profile of the output signal from the optical comb generator was designed to be flat over 320 GHz to obtain a Nyquist pulse with a roll-off factor α = 0 for 320 Gbaud OTDM transmission. Figures 2(a), 2(b) and 2(c), respectively, show the optical spectra observed at points A, B and C shown in Fig. 1. Here, the spectral resolution was set at 0.02 nm to allow us to observe the comb structure clearly. The OSNR of the output signal from the comb generator was 48 dB as shown in Fig. 2(a). Since the power level of each comb component was as low as −20 dBm, optical amplification was needed. Figure 2(b) shows the optical spectrum after amplification with an erbium-doped fiber amplifier (EDFA), where the OSNR was degraded from 48 to 40 dB due to the amplified spontaneous emission (ASE) noise from the EDFA. Figures 2(c) and 2(d), respectively, show the spectrum and waveform after shaping with a pulse shaper, where a Nyquist pulse with α = 0 was successfully generated. Although the OSNR after shaping seems to be higher than that before shaping, the actual OSNR is limited by noise contained in the signal band. Thus, the actual OSNR of the obtained Nyquist pulse was 40 dB.
Fig. 1 Configuration of a conventional Nyquist pulse source combining a CW laser and an optical comb generator.
Fig. 2 Optical spectrum of the output signal from the comb generator (a) before and (b) after amplification with an EDFA, (c) optical spectrum and (d) waveform after pulse shaping.
Figure 3 shows the configuration of our new Nyquist pulse source with a hydrogen cyanide (HCN) frequency stabilized MLFL [7] instead of a CW laser source. This configuration does not require an EDFA because a short pulse with a wide bandwidth and a high power is directly obtained from the MLFL. Figures 4(a) and 4(b), respectively, show the spectrum and waveform of the output pulse from the MLFL, which was observed at point A in Fig. 3. The output spectral width was 3.2 nm (400 GHz) and the pulse width was 1.1 ps. The time-bandwidth product was 0.44, indicating that the output pulse was a transform-limited Gaussian pulse. The laser output power was 30 mW and the OSNR was 51 dB. Figures 4(c) and 4(d), respectively, show the spectrum and waveform after shaping with a pulse shaper at point B in Fig. 3. The OSNR after shaping is determined by that of the input signal into the pulse shaper as mentioned above. Therefore, the OSNR of the obtained Nyquist pulse can be kept as high as 51 dB, which is 11 dB higher than that obtained with the combination of a CW laser and an optical comb generator shown in Fig. 2(c).
Fig. 4 Optical spectrum and waveform of the MLFL output pulse (b), (c) before and (d), (e) after pulse shaping.
Figure 5 shows the experimental setup for a single-channel 3.84 Tbit/s, 64 QAM coherent optical Nyquist pulse transmission. At the transmitter, the 10 GHz coherent Nyquist pulse signal (α = 0) shown in Fig. 4 was modulated with 10 Gbaud, 64 QAM using an IQ modulator driven by an arbitrary waveform generator (AWG). After modulation, the 10 Gbaud, 64 QAM Nyquist pulses were multiplexed to 320 Gbaud using two OTDM multiplexers based on a planar lightwave circuit (PLC). The OTDM signal was then polarization-multiplexed with a polarization beam combiner. In parallel to these processes, part of the MLFL output was divided in front of the pulse shaper, and the 17th harmonic signal (170 GHz shifted from the center frequency) of the optical comb spectrum was extracted with a narrow bandpass filter. We used it as a pilot tone signal for the optical phase-locked loop (OPLL) at the receiver [10]. The data and pilot tone signals were combined and fed into a transmission link consisting of two 75 km spans with a 50 km super large area (SLA) fiber with a dispersion of 20 ps/nm/km and a 25 km inverse dispersion fiber (IDF) with a normal dispersion of −40 ps/nm/km. The average loss was 17.5 dB/span, which was compensated for by the combination of a Raman amplifier (10 dB gain) and an EDFA (7.5 dB gain). The launch power was set at 4 dBm, which was optimally chosen to maximize the OSNR and minimize the nonlinear impairments.
After a 150 km transmission, the 320 Gbaud OTDM data were synchronously homodyne-detected at the receiver with a 10 GHz Nyquist-local oscillator (LO) pulse (α = 0), in which the demultiplexed 10 Gbaud I and Q data for both polarizations were simultaneously obtained by using the time-domain orthogonality of the Nyquist pulse [11]. Here, the Nyquist-LO pulse was generated as follows. First, a CW-LO (frequency-tunable CW laser) was fed into an optical comb generator driven by a 10 GHz clock extracted from the transmitted OTDM Nyquist pulse data. Then, the CW-LO was phase-locked to the transmitted data signal by using the beat spectrum between the pilot tone and the 16th harmonic of the LO comb signal extracted with a narrow band filter via an OPLL process. After that, a 10 GHz Nyquist pulse train was formed by shaping the optical comb signal using a pulse shaper as shown in Fig. 2(c). The signals received from balanced detectors were A/D-converted at a sampling rate of 80 Gsample/s and demodulated in an offline condition with a digital signal processor (DSP). At the DSP, distortions caused by hardware imperfections were compensated for by using an adaptive 99-tap finite impulse response filter. Furthermore, a digital back-propagation (DBP) method, which was modified for the optical pulse transmission, was newly employed to compensate for the nonlinear phase rotation induced by the self-phase modulation (SPM) and cross phase modulation (XPM) that occurred in the optical fiber during pulse transmission. The modified DBP method will be explained in detail in the next section. Finally, we demodulated the compensated QAM signal into binary data, and measured the BER.
3. Experimental results
Figures 6(a) and 6(b) show the constellations of demultiplexed 10 Gbaud, 64 QAM signals under a back-to-back condition obtained with the conventional and new Nyquist pulse sources shown in Figs. 1 and 3, respectively. By using the new pulse source with an MLFL, the error vector magnitude (EVM) was improved from 3.7% to 3.4%, and an error free for 4096 symbol data was obtained thanks to the higher OSNR.
Fig. 6 Constellations of demultiplexed 10 Gbaud, 64 QAM signals under back-to-back condition obtained with (a) conventional and (b) new pulse sources.
During the optical fiber transmission, the Nyquist pulse signal broadened and overlapped with adjacent pulses due to the fiber dispersion. Therefore, it is important to consider the mutual pulse interference between all the tributaries when compensating for nonlinear effects with a DBP method. Figure 7 shows schematic diagrams of two new DBP methods for optical pulse transmission. Figure 7(a) shows the first scheme (Scheme A) starting from one tributary. At the receiver, one tributary pulse was selectively detected by homodyne detection with an Nyquist LO pulse. The waveform of the detected signal was broadened in the time domain due to the bandwidth limitation of the receiver. Therefore, the signal waveform was reshaped into an ideal Nyquist pulse in the DSP. Then, all the tributaries were prepared with the same pulse data by using a digital MUX circuit with the same construction as that used at the transmitter. After these processes, DBP was applied to the restored OTDM signal. Finally, the target tributary data were detected again by homodyne detection with an Nyquist LO pulse. Figure 7(b) shows the second scheme (Scheme B) with all the tributaries from the beginning. At the receiver, all the tributary pulses were detected by adjusting the timing of the LO pulse to each tributary pulse. The detected signals were reshaped into an ideal Nyquist pulse, and then the reshaped signals were OTDM-multiplexed in the DSP without a digital MUX circuit. The subsequent processes were the same as those in Scheme A.
Fig. 7 Schematic diagrams of modified DBP method with all tributary pulses. (a) Using a digital MUX circuit (Scheme A), (b) receiving all the tributary pulses (Scheme B).
Figure 8 shows the BER of a demultiplexed 10 Gbaud, 64 QAM signal (X-pol) after a 150 km transmission as a function of launch power. The black line shows the BER obtained without DBP. The green and blue lines show the BER obtained with DBP using one tributary pulse and all the tributary pulses (Scheme A), respectively. We set the DBP step size at 2.5 km. In each case, the lowest BER was obtained with a launch power of 4 dBm. The BER was reduced from 2.1x10−3 to 1.3x10−3 by using the present DBP method, which considered the nonlinear effects between all the tributaries starting from one tributary. In Fig. 8, the red circle shows a BER obtained with the DBP method in Scheme B starting from all the tributaries. The BER obtained with Scheme B was almost the same as that obtained with Scheme A. Therefore, we employed the modified DBP method with Scheme A, which is much simpler than Scheme B.
Fig. 8 BER of demultiplexed 10 Gbaud, 64 QAM signal after a 150 km transmission as a function of launch power.
Figures 9(a) and 9(b) show the spectra of the OTDM data signal before and after a 150 km transmission with a launch power of 4 dBm. The spectral fringes in both figures were caused by spectrum interference in our PLC multiplexer. The 20 dB bandwidth of the signal including the pilot tone was 340 GHz. After the transmission, the OSNR had degraded from 36 to 31 dB.
Figure 10(a) shows the radio frequency (RF) spectrum of the intermediate frequency (IF) signal in the OPLL circuit. The OPLL bandwidth was 1.3 MHz, which is limited by the loop length in the OPLL circuit [12]. Figure 10(b) shows the single side-band (SSB) phase noise spectrum. The phase noise was 0.85 degrees, which was well below the 4.7-degree phase difference between the most adjacent symbol points on the 64 QAM constellation.
Figure 11(a) shows the BER characteristics of a demultiplexed 10 Gbaud, 64 QAM signal as a function of the OSNR. The black and blue lines show the BER characteristics before and after a 150 km transmission, respectively. There was an OSNR penalty of 1 dB at a BER of 2x10−3 after a 150 km transmission. Figure 11(b) shows the constellation of a demultiplexed 10 Gbaud, 64 QAM signal with an OSNR of 31 dB after a 150 km transmission. The EVM was increased to 5.0% due to the OSNR degradation during transmission as shown in Fig. 9. The BERs for all the tributaries after a 150 km transmission are shown in Fig. 12. Here, the BER was the average value for two polarizations. By using the new DBP, BERs below the FEC threshold (2x10−3) with a 7% overhead were obtained for the all tributaries.
Fig. 11 (a) BER characteristics of demultiplexed 10 Gbaud, 64 QAM signal as a function of the OSNR, (b) constellation of the signal with an OSNR 31 dB after 150 km transmission.
4. Conclusion
We successfully achieved a single-channel 3.84 Tbit/s, 64 QAM coherent optical Nyquist pulse transmission over 150 km within an optical bandwidth of 340 GHz with a frequency-stabilized MLFL and a new DBP method for pulse transmission. The present result is scalable to a net spectral efficiency as high as 10.6 bit/s/Hz even when the 7% FEC overhead is taken into account. To the best of our knowledge, the transmission capacity is the highest in a single-carrier transmission with a spectral efficiency exceeding 10 bit/s/Hz.
Funding
JSPS Grant-in-Aid for Specially Promoted Research (26000009).
References and links
1. G. Raybon, J. Cho, A. Adamiecki, P. Winzer, A. Konczykowska, F. Jorge, J. Dupuy, M. Riet, B. Duval, K. Kim, S. Randel, D. Pilori, B. Guan, N. K. Fontaine, and E. Burrows, “Single carrier high symbol rate transmitter for data rates up to 1.0 Tb/s,” in Proceedings of Optical Fiber Conference (OFC, 2016), Th3A.2. [CrossRef]
2. H. Yamazaki, A. Sano, M. Nagatani, and Y. Miyamoto, “Single-carrier 1-Tb/s PDM-16QAM transmission using high-speed InP MUX-DACs and an integrated OTDM modulator,” Opt. Express 23(10), 12866–12873 (2015). [CrossRef] [PubMed]
3. M. Nakazawa, T. Hirooka, P. Ruan, and P. Guan, “Ultrahigh-speed “orthogonal” TDM transmission with an optical Nyquist pulse train,” Opt. Express 20(2), 1129–1140 (2012). [CrossRef] [PubMed]
4. H. Hu, D. Kong, E. Palushani, J. D. Andersen, A. Rasmussen, B. M. Sørensen, M. Galili, H. C. H. Mulvad, K. J. Larsen, S. Forchhammer, P. Jeppesen, and L. K. Oxenløwe, “1.28 Tbaud Nyquist signal transmission using time-domain optical Fourier transformation based receiver,” in Proceedings of the Conference on Laser and Electro-Optics (CLEO, 2013), CTh5D.5. [CrossRef]
5. D. Suzuki, K. Harako, T. Hirooka, and M. Nakazawa, “Single-channel 5.12 Tbit/s (1.28 Tbaud) DQPSK transmission over 300 km using non-coherent Nyquist pulses” in Proceedings of the European Conference on Optical Communication (ECOC, 2016), paper W.4.P1.SC5.49.
6. D. O. Otuya, K. Kasai, T. Hirooka, and M. Nakazawa, “Single-channel 1.92 Tbit/s, 64 QAM coherent Nyquist orthogonal TDM transmission with a spectral efficiency of 10.6 bit/s/Hz,” J. Lightwave Technol. 34(2), 768–775 (2016). [CrossRef]
7. M. Yoshida, K. Yoshida, K. Kasai, and M. Nakazawa, “1.55 μm hydrogen cyanide optical frequency-stabilized and 10 GHz repetition-rate-stabilized mode-locked fiber laser,” Opt. Express 24(21), 24287–24296 (2016). [CrossRef] [PubMed]
8. T. Sakamoto, T. Kawanishi, and M. Izutsu, “Asymptotic formalism for ultraflat optical frequency comb generation using a Mach-Zehnder modulator,” Opt. Lett. 32(11), 1515–1517 (2007). [CrossRef] [PubMed]
9. T. Sakamoto, T. Kawanishi, and M. Izutsu, “Optimization of electro-optic comb generation conventional Mach-Zehnder modulator,” in Proceedings of IEEE Topical Meeting Microwave Photonics (IEEE, 2007), pp. 50–53. [CrossRef]
10. K. Kasai, T. Omiya, P. Guan, M. Yoshida, T. Hirooka, and M. Nakazawa, “Single-channel 400-Gb/s OTDM-32 RZ/QAM coherent transmission over 225 km using an optical phase-locked loop technique,” IEEE Photonics Technol. Lett. 22(8), 562–564 (2010). [CrossRef]
11. K. Harako, D. O. Otuya, K. Kasai, T. Hirooka, and M. Nakazawa, “High-performance TDM demultiplexing of coherent Nyquist pulses using time-domain orthogonality,” Opt. Express 22(24), 29456–29464 (2014). [CrossRef] [PubMed]
12. S. Beppu, K. Kasai, M. Yoshida, and M. Nakazawa, “2048 QAM (66 Gbit/s) single-carrier coherent optical transmission over 150 km with a potential SE of 15.3 bit/s/Hz,” Opt. Express 23(4), 4960–4969 (2015). [CrossRef] [PubMed]
