Open Access
Add to BibSonomyAbstract
We experimentally demonstrate 8 × 240-Gb/s super-Nyquist wavelength-division-multiplexing (WDM) polarization-division-multiplexing quadrature-phase-shift-keying (PDM-QPSK) signal transmission on a 50-GHz grid with a net spectral efficiency (SE) of 4b/s/Hz adopting hardware-efficient simplified heterodyne detection. 9-ary quadrature-amplitude-modulation-like (9QAM-like) processing based on multi-modulus blind equalization (MMBE) is adopted to reduce analog-to-digital converter (ADC) bandwidth requirement and improve receiver sensitivity. The transmission distance at the soft-decision forward-error-correction (SD-FEC) threshold of 2 × 10−2 is 2 × 420km based on digital post filtering while largely extended to over 5 × 420km based on 9QAM-like processing, which well illustrates 9QAM-like processing is more efficient for heterodyne coherent WDM system. Moreover, only two ADC channels are needed for simplified heterodyne detection of one 60-Gbaud PDM-QPSK WDM channel, and thus only one commercial oscilloscope (OSC) with two input ports can work well for each WDM channel.
© 2014 Optical Society of America
1. Introduction
With the development of large-bandwidth and high-speed electronic analog-to-digital converters (ADCs) and photo detectors (PDs), very recently, coherent detection with digital signal processing (DSP) has been attracting a great deal of interest in research community once again [1–10]. It is well known that coherent detection includes homodyne detection and heterodyne detection. Although heterodyne coherent receiver generally requires a twice larger ADC bandwidth and also has an extra 3-dB signal-to-noise ratio (SNR) penalty compared to heterodyne coherent receiver, heterodyne coherent receiver is much more hardware-efficient and easier to integrate than homodyne coherent receiver, which is because in heterodyne coherent receiver only two ports are needed and also intermediate-frequency (IF) down conversion can be performed in digital frequency domain [7–10]. Moreover, several recently proposed advanced receiver-based DSP algorithms, such as digital post filtering combined with 1-bit maximum likelihood sequence estimation (MLSE) [11, 12], can be used to reduce ADC bandwidth requirement and improve receiver sensitivity for heterodyne coherent receiver. Based on simplified heterodyne coherent detection and digital post filtering, we have experimentally demonstrated 8 × 112-Gb/s Nyquist wavelength-division-multiplexing (WDM) polarization-division-multiplexing quadrature-phase-shift-keying (PDM-QPSK) signal transmission over 1120-km single-mode fiber-28 (SMF-28) on a 25-GHz grid [9] and 4 × 196.8-Gb/s Nyquist WDM PDM carrier-suppressed return-to-zero QPSK (PDM-CSRZ-QPSK) signal transmission over 1040-km SMF-28 on a 50-GHz grid [10]. If 7% hard-decision forward-error-correction (HD-FEC) overhead [13] is considered, 196.8Gb/s is corresponding to a net bit rate of 183.9Gb/s, which, to the best of our knowledge, is the highest bit rate per channel demonstrated for heterodyne coherent WDM transmission system. Meanwhile, it is a potential trend to realize immediate upgrade toward 200G for the existing 100G PDM-QPSK channels on the 50-GHz ITU-T grid, which cannot only increase the spectral efficiency (SE) from 2b/s/Hz to 4b/s/Hz, but also make full use of the advantages of PDM-QPSK modulation in terms of transmission distance and efficient DSP algorithms. However, the upgrade toward 200G will be subject to strong filtering effects and bandwidth-limited ADCs because of the expanded signal spectrum. The conventional coherent DSP algorithms plus digital post filtering have limited tolerance to strong filtering effects and will become inefficient when the large-capacity high-SE WDM PDM-QPSK channels are transmitted over the existing standard fiber links with multiple reconfigurable optical add-drop multiplexers (ROADMs) and transoceanic transmission distance [14, 15]. Recently, we have proposed the 9-ary quadrature amplitude modulation like (9QAM-like) processing for PDM-QPSK signal in presence of very strong filtering effects, which can directly recover the super-Nyquist (signal bandwidth>channel spacing) spectrally-shaped QPSK signal to the 9QAM-like one based on multi-modulus blind equalization (MMBQ) [14]. We have applied the 9QAM-like processing in a homodyne coherent system and experimentally demonstrated 8 × 112-Gb/s super-Nyquist WDM PDM-QPSK signal transmission over 2640-km SMF-28 on a 25-GHz grid [15]. Compared to the classic constant-modulus-algorithm (CMA) equalization plus digital post filtering, the adoption of MMBQ in the homodyne coherent system effectively improves the filtering tolerance and extends the transmission distance.
In this paper, we experimentally demonstrate 8 × 240-Gb/s super-Nyquist WDM PDM-QPSK signal transmission over 5 × 420-km SMF-28 on a 50-GHz grid adopting simplified heterodyne detection and 9QAM-like processing. The net SE is 4b/s/Hz if 20% soft-decision FEC (SD-FEC) overhead [16] is considered. Compared to our previous heterodyne coherent WDM systems based on CMA equalization plus digital post filtering [9, 10], the transmission distance is almost doubled due to the adoption of MMBQ without digital post filtering, which well illustrates the 9QAM-like processing is more efficient for heterodyne coherent WDM system. To our best knowledge, it is the first time to realize long-haul high-SE heterodyne coherent WDM transmission system with the highest baud rate of 60Gbaud per channel.
2. Principle of 9QAM-like processing for PDM-QPSK signal
Figure 1 shows the principle of super-Nyquist spectral shaping for optical single-polarization QPSK signal by an optical Gaussian band-pass filter. The 3-dB pass bandwidth of the optical Gaussian band-pass filter is no more than the symbol rate of the optical QPSK signal. After super-Nyquist spectral shaping, in the time domain, the 4-point QPSK signal becomes a 9QAM-like signal due to the filtering effect, while the channel spectrum becomes much narrower in the frequency domain. At the coherent receiver, the received spectrally-shaped signal is directly recovered to the 9QAM-like signal based on the MMBQ. The MMBQ mainly adopts the cascaded multi-modulus algorithm (CMMA) equalization plus the modified carrier recovery scheme [14]. The CMMA proposed for PDM-8QAM modulation [17] is much more accurate than the classic CMA for the 9QAM-like signal processing. It is because the classic CMA cannot well handle the 9QAM-like signal with several different symbol amplitudes. Moreover, compared to the classic CMA-based equalizer, the frequency response of the CMMA-based equalizer can significantly suppress the high-frequency signal components and protect the subsequent carrier recovery process from the impact of enhanced crosstalk and noise. However, there exists a trade-off for the 9QAM-like processing between filtering tolerance and real-time implementation complexity.
3. Experimental setup and results for heterodyne coherent WDM transmission
Figure 2 shows the experimental setup for 8 × 240-Gb/s super-Nyquist WDM PDM-QPSK on a 50-GHz grid with simplified heterodyne detection and 9QAM-like processing. At the transmitter, there are eight external cavity lasers (ECLs) with linewidth less than 100kHz and maximum output power of 14.5dBm. The operating wavelengths for the odd group of ECLs range from 1548.161nm to 1550.565nm, while those for the even group of ECLs range from 1548.561nm to 1550.965nm. Each group has 100-GHz neighboring frequency spacing. The continuous-wavelength (CW) lightwaves from each group are first combined by a polarization-maintaining optical coupler (PM-OC), then modulated by a 60-Gbaud electrical binary signal via an in-phase/quadrature (I/Q) modulator with 33-GHz 3-dB bandwidth, and finally polarization multiplexed by a polarization multiplexer and power amplified by an erbium-doped fiber amplifier (EDFA). Each 60-Gbaud electrical binary signal has a pseudo-random binary sequence (PRBS) length of 215-1 and is generated from a pulse pattern generator (PPG). For optical PDM-QPSK modulation, the two parallel Mach-Zehnder modulators (MZMs) in each I/Q modulator are both biased at the null point and driven at the full swing. The phase difference between the upper and lower arms of each I/Q modulator is controlled at π/2. Each polarization multiplexer includes a PM-OC to halve the signal into two arms, an optical delay line (DL) to provide a 150-symbol delay, an optical attenuator to balance the power of two arms and a polarization beam combiner (PBC) to recombine the signals. The generated odd-channel and even-channel PDM-QPSK optical signals are combined and spectrally-shaped by a programmable wavelength selective switch (WSS) on a 50-GHz grid, to further generate 8 × 240-Gb/s super-Nyquist WDM PDM-QPSK optical signal.
The generated signal is then launched into a re-circulating fiber loop, which consists of five spans of 84-km SMF-28. Each span has 18-dB average fiber loss and 17-ps/km/nm chromatic dispersion (CD) at 1550nm without optical dispersion compensation. An EDFA is used before each span to compensate for the fiber loss. The total launched power (after EDFA) into each span is ~1dBm per channel. After 5 × 84-km SMF-28 transmission, the optical signal passes through a programmable WSS on a 50-GHz grid to remove the amplified spontaneous emission (ASE) noise. An EDFA is used after the WSS to compensate for the switch loss in the loop. Thus, for each circulation, the optical signal is transmitted over 5 × 84-km SMF-28 and passes through one WSS. After five circulations, the optical signal is transmitted over 5 × 420-km SMF-28 and passes through five WSSs.
At the receiver, another programmable WSS on a 50-GHz grid is used to select the desired channel. An ECL with linewidth less than 100kHz is used as local oscillator (LO) and has 32.5-GHz frequency offset relative to the selected channel. Two polarization beam splitters (PBSs) and two OCs are used to implement polarization diversity of the received signal and LO in optical domain before heterodyne beating. Only two balanced PDs (BPDs) with 50-GHz 3-dB bandwidth are needed here. A broadband electrical amplifier (EA), with 25-dB gain and DC~60GHz frequency range, is used after each BPD. The analog-to-digital conversion is realized in the real-time oscilloscope (OSC) with 160-GSa/s sampling rate and 65-GHz electrical bandwidth. Two ADC channels are enough for each WDM channel. Offline DSP is implemented after analog-to-digital conversion. Firstly, the received signals are down-converted to the baseband by multiplying synchronous cosine and sine functions, which are generated from a digital LO. Secondly, a T/2-spaced time-domain finite-impulse-response (FIR) filter is used for CD compensation, where the filter coefficients are calculated from the known fiber CD transfer function using the frequency-domain truncation method. Thirdly, two complex-valued, 17-tap, T/2-spaced adaptive FIR filters, based on CMMA, are used to retrieve the modulus of the 9QAM-like signal and realize polarization de-multiplexing. The carrier recovery, including frequency offset estimation and two-stage carrier phase estimation, is performed after QPSK-based joint-polarization partition and rotation [14]. The frequency offset estimation is based on the fast Fourier transform (FFT) method, which is identical to the operation for QPSK signal. The two-stage carrier phase estimation is based on the first-stage Viterbi-Viterbi algorithm and the second-stage maximum-likelihood algorithm [14]. Finally, 1-bit MLSE before bit-error-ratio (BER) calculation is used for symbol decoding and inter-symbol-interference (ISI) suppression. Figures 3(a) and 3(b) show optical spectra (0.1-nm resolution) before and after 5 × 420-km SMF-28 transmission, respectively. Each amplifier span of 84-km SMF-28 has a total loss of 18dB. Figure 3(c) shows optical spectrum (0.1-nm resolution) of the fifth WDM channel selected by the 50-GHz WSS at the receiver. Figure 3(d) shows electrical spectrum centered on 32.5-GHz IF after analog-to-digital conversion.
Fig. 3 Optical spectra: (a) before and (b) after 5 × 420-km SMF-28 transmission. (c) Optical spectrum selected by WSS. (d) Electrical spectrum after analog-to-digital conversion.
The selected fifth channel at 1549.761nm is denoted by ch5 if all other channels are turned off (single-channel case) and by CH5 if all other channels are also turned on (WDM-channel case). In comparison with 9QAM-like processing, we also carry out the process of digital post filtering for the selected fifth channel, where a linear digital delay-and-add FIR filter, with a transfer function of H(z) = 1 + Z−1, and 1-bit MLSE are adopted after conventional coherent QPSK algorithm [11, 12]. Figure 4(a) shows back-to-back (BTB) optical spectra for ch5 based on 9QAM-like processing in three different cases. Two cascaded 50-GHz WSSs are adopted in our BTB experiment. The optical spectrum becomes much narrower after one 50-GHz WSS, while the optical spectrum after two cascaded 50-GHz WSSs is quite similar to that after one 50-GHz WSS. Figure 4(b) shows BTB BER versus optical SNR (OSNR) for ch5 based on 9QAM-like processing with and without the first 50-GHz WSS, respectively. The adoption of the first 50-GHz WSS introduces 1-dB OSNR penalty at the BER of 3.8 × 10−3 and 1.5-dB OSNR penalty at the BER of 2 × 10−2. It is because the multilevel detection for the 9QAM-like signal requires a larger OSNR than that for the QPSK signal. We also verify that the second WSS does not cause any OSNR penalty with 9QAM-like processing.
Figure 5(a) gives BTB BER versus OSNR for CH5 with 9QAM-like processing and digital post filtering, respectively. The BER performance based on 9QAM-like processing is much better than that based on digital post filtering, which is because the WDM channel based on 9QAM-like processing is more robustness to inter-channel crosstalk. The BER of 3.8 × 10−3 can be attained at the OSNR of 24.5dB for the WDM channel based on 9QAM-like processing. We have experimentally demonstrated the required OSNR at the BER of 3.8 × 10−3 is 27dB for the homodyne detection of 28-Gbaud Nyquist WDM PDM-QPSK on a 25-GHz grid when only conventional coherent QPSK algorithm is adopted [4], and therefore in theory the required OSNR at the BER of 3.8 × 10−3 will be over 30dB for the homodyne detection of 60-Gbaud super-Nyquist WDM PDM-QPSK on a 50-GHz grid. Thus, we can conclude our demonstrated heterodyne coherent system has relatively good receiver sensitivity due to the adoption of 9-QAM processing. Figure 5(b) gives BTB BER versus OSNR for ch5 and CH5, respectively. Both ch5 and CH5 adopt 9QAM-like processing. The BER performance for CH5 is similar to that for ch5, which well demonstrates 9QAM-like processing can almost remove inter-channel crosstalk completely. The BER of 3.8 × 10−3 can be attained at the OSNR of 24.5dB. Insets (I) and (II) in Fig. 5(b) give the received X- and Y-polarization 9QAM-like BTB constellations for CH5 at the OSNR of 25dB and the BER of 2.3 × 10−3. The unequal spacing between the constellation points is due to the non-optimized bias of the I/Q modulator with a limited bandwidth of about 30GHz. Figure 6(a) gives BER versus transmission distance for CH5 with 9QAM-like processing and digital post filtering, respectively. The launched power into fiber is 1dBm/channel. The BER for CH5 based on 9QAM-like processing increases from 3.8 × 10−3 to 2 × 10−2 when the transmission distance increases from 2 × 420km to over 5 × 420km, which is because the OSNR is reduced with the increase of the transmission distance. The transmission distance for CH5 at the BER of 2 × 10−2 is 2 × 420km based on digital post filtering and over 5 × 420km based on 9QAM-like processing, which well demonstrates that compared to digital post filtering, the adoption of 9QAM-like processing can effectively improve system performance and largely extend transmission distance. Figure 6(b) gives BERs of all channels after 5 × 420-km SMF-28 transmission with 9QAM-like processing. The launched power into fiber is 1dBm/channel. The BER of all channels is under the SD-FEC threshold of 2 × 10−2. Inset (I) in Fig. 6(b) gives the received X-polarization 9QAM-like constellation for CH5 after 5 × 420-km SMF-28 transmission.
Fig. 5 (a) BTB BER versus OSNR. (b) BTB BER versus OSNR for single and WDM channels. Insets (I) and (II) give X- and Y-polarization 9QAM-like constellations.
Fig. 6 (a) BER versus transmission distance. (b) BER of all channels. Inset (I) gives X-polarization 9QAM-like constellation.
4. Conclusion
We experimentally demonstrate 8 × 240-Gb/s super-Nyquist WDM PDM-QPSK signal transmission on a 50-GHz grid with a net SE of 4b/s/Hz adopting simplified heterodyne detection. The introduction of MMBE-based 9-QAM processing into the demonstrated heterodyne coherent transmission system can reduce ADC bandwidth requirement and improve receiver sensitivity while maintaining a hardware-efficient system architecture. To the best of our knowledge, it is the first time to realize the long-haul high-SE heterodyne coherent WDM transmission system with the highest baud rate of 60Gbaud per channel. This work was partially supported by NNSF of China (61325002, 61250018, 61177071), NHTRDP (863 Program) of China (2012AA011303, 2013AA010501), NKTR&DP of China (2012BAH18B00), KPSSTA of Shanghai (12dz114300, 12510705600 and 13JC1400700).
References and links
1. X. Zhou and J. Yu, “Multi-level, multi-dimensional coding for high-speed and high spectral-efficiency optical transmission,” J. Lightwave Technol. 27(16), 3641–3653 (2009). [CrossRef]
2. E. Ip, A. P. T. Lau, D. J. F. Barros, and J. M. Kahn, “Coherent detection in optical fiber systems,” Opt. Express 16(2), 753–791 (2008). [CrossRef] [PubMed]
3. S. Zhang, M. F. Huang, F. Yaman, E. Mateo, D. Qian, Y. Zhang, L. Xu, Y. Shao, I. Djordjevic, T. Wang, Y. Inada, T. Inoue, T. Ogata, and Y. Aoki, “40×117.6 Gb/s PDM-16QAM OFDM transmission over 10,181 km with soft-decision LDPC coding and nonlinearity compensation,” in Proc. OFC/NFOEC2012, Los Angeles, CA, paper PDP5C.4.
4. J. Yu, Z. Dong, H. C. Chien, Z. Jia, X. Li, D. Huo, M. Gunkel, P. Wagner, H. Mayer, and A. Schippel, “Transmission of 200 G PDM-CSRZ-QPSK and PDM-16 QAM with a SE of 4 b/s/Hz,” J. Lightwave Technol. 31(4), 515–522 (2013). [CrossRef]
5. P. J. Winzer, A. H. Gnauck, C. R. Doerr, M. Magarini, and L. L. Buhl, “Spectrally efficient long-haul optical networking using 112-Gb/s polarization-multiplexed 16-QAM,” J. Lightwave Technol. 28(4), 547–556 (2010). [CrossRef]
6. R. Zhu, K. Xu, Y. Zhang, Y. Li, J. Wu, X. Hong, and J. Lin, “QAM coherent subcarrier multiplexing system based on heterodyne detection using intermediate frequency carrier modulation,” in Proc. Of APMP, 165–168 (2008).
7. X. Li, J. Zhang, F. Li, and J. Xiao, “Over 2000-km transmission of 60-Gbaud PDM-QPSK signal with heterodyne detection and SE of 4b/s/Hz,” in Proc. OFC/NFOEC 2014, San Francisco, CA, paper Th4F.4. [CrossRef]
8. J. Zhang, Z. Dong, J. Yu, N. Chi, L. Tao, X. Li, and Y. Shao, “Simplified coherent receiver with heterodyne detection of eight-channel 50 Gb/s PDM-QPSK WDM signal after 1040 km SMF-28 transmission,” Opt. Lett. 37(19), 4050–4052 (2012). [CrossRef] [PubMed]
9. X. Li, Z. Dong, J. Yu, J. Yu, and N. Chi, “Heterodyne coherent detection of WDM PDM-QPSK signals with spectral efficiency of 4b/s/Hz,” Opt. Express 21(7), 8808–8814 (2013). [CrossRef] [PubMed]
10. X. Li, J. Yu, N. Chi, Z. Dong, J. Zhang, and J. Yu, “The reduction of the LO number for heterodyne coherent detection,” Opt. Express 20(28), 29613–29619 (2012). [CrossRef] [PubMed]
11. J. Li, E. Tipsuwannakul, M. Karlsson, and P. A. Andrekson, “Low complexity duobinary signaling and detection for sensitivity improvement in Nyquist-WDM coherent system,” in Proc. OFC/NFOEC2012, Los Angeles, CA, paper OM3H.2. [CrossRef]
12. K. Igarashi, T. Tsuritani, I. Morita, Y. Tsuchida, K. Maeda, M. Tadakuma, T. Saito, K. Watanabe, K. Imamura, R. Sugizaki, and M. Suzuki, “Super-Nyquist-WDM transmission over 7,326-km seven-core fiber with capacity-distance product of 1.03 Exabit/s · km,” Opt. Express 22(2), 1220–1228 (2014). [CrossRef] [PubMed]
13. ITU-T Recommendation G.975.1, “Forward error correction for high bit-rate DWDM submarine system,” 2004.
14. J. Zhang, J. Yu, N. Chi, Z. Dong, J. Yu, X. Li, L. Tao, and Y. Shao, “Multi-modulus blind equalizations for coherent quadrature duobinary spectrum shaped PM-QPSK digital signal processing,” J. Lightwave Technol. 31(7), 1073–1078 (2013). [CrossRef]
15. J. Zhang, B. Huang, and X. Li, “Improved quadrature duobinary system performance using multi-modulus equalization,” IEEE Photon. Technol. Lett. 25(16), 1630–1633 (2013). [CrossRef]
16. T. Mizuochi, Y. Miyata, K. Kubo, T. Sugihara, K. Onohara, and H. Yoshida, “Progress in soft-decision FEC,” in Proc. OFC/NFOEC2011, Los Angeles, CA, paper NWC2. [CrossRef]
17. X. Zhou, J. Yu, and P. D. Magill, “Cascaded two-modulus algorithm for blind polarization de-multiplexing of 114-Gb/s PDM-8-QAM optical signals,” in Proc. OFC/NFOEC2009, San Diego, CA, Paper OWG3.
