Realize 7*225Gb/s Nyquist WDM Transmission Technology Using Pulse Carrier Suppression Quadrature Phase Shift Keying Modulation on Traditional Single Mode (SMF-28) Fiber Over 1600 km



 Realization of 7*225Gb/s Nyquist Wavelength Division Multiplexing Transmission Technology Using Pulse Carrier Suppression Quadrature Phase Shift Keying Modulation on Traditional Single Mode (SMF-28) Fiber over 1600 km
Ze Dong, Jianjun Yu, Zhensheng Jia, Hung-Chang Chien, Xinying Li, and Gee-Kung Chang


                                                                        Abstract
    We have transmitted 7 × 224 Gb/s/ch Nyquist over a conventional single-mode fiber (SMF-28) each spanning 80 km and an erbium-doped fiber amplifier for amplification only. wavelength division multiplexed signal. Each channel in the 50GHz network is quadrature phase shift keying using polarization multiplexing, and gives a frequency efficiency of 4 b/s/Hz.bits recording at 56 G baud. Post filter and Viterbi maximum likelihood sequence detection are both introduced into digital signal processing to suppress unwanted noise, improve linear crosstalk and filter effects. After transmission over 1600 km of conventional single-mode fiber, the bit error for all channels is smaller than the previous error correction limit of 3.8×10-3.
Keywords: Coherent Detection, Maximum Likelihood Sequence Detection, Nyquist Wavelength Division Multiplexing, Post Filter, Quadrature Phase Shift Keying, Signal Transmission, Spectral Efficiency. 


                                                                          one. Introduction

       With the commercialization of 100G Ethernet, improving the spectral efficiency of fiber channel transmission has become one of the main challenges for next-generation high-speed and large-capacity transmission systems [1]-[4]. Many solutions beyond 100G rates have been proven feasible in the laboratory, such as electrical/optical orthogonal frequency division multiplexing (OFDM) [5], optical time division multiplexing (OTDM) [6] and multilevel modulation [7] ,[8]. As we know, multilevel modulation not only requires high optical signal-to-noise ratio (OSNR), but also is more sensitive to fiber nonlinear effects and laser phase noise. In addition, in long-distance transmission, multi-level modulation is impractical compared to the established QPSK method, which can use a simpler transmitter structure and efficient Digital Signal Processing Algorithms. Using quadrature phase shift keying at a symbol rate of 56 GSa/s, a 10 × 224-Gbit/s Nyquist WDM signal passes 1890 km of standard single-mode with a spectral efficiency of 2b/s/Hz Optical fiber transmission [9]. In [10], a 224-Gb/s and 112-Gb/s pulse-modulated quadrature-phase-shift keying signal mixed at 50-GHz intervals over 1200 km of dispersion-managed mode large-mode-field fiber with 3b /s/Hz spectrally efficient transmission, and carrier-suppressed nulling pulse shaping is used to soften the filtering effect. Two asymmetric optical crossover wavelength division multiplexers are used to separate the 100-G and 200-G signals. Therefore, even though the frequency grid is only 50 GHz, a 224-Gb/s pulse-modulated quadrature phase-shift keying (PDM-QPSK) signal is endowed with a bandwidth of over 50 GHz [10].
        In order to achieve higher spectral efficiency, in this letter, we applied Nyquist WDM channels spaced at 50 GHz, using a pulse-modulated carrier-suppressed return-to-zero code quadrature phase at 56 G baud Shift keying modulation mode. A digital post filter after maximum likelihood sequence detection is introduced into off-line digital signal processing to suppress noise, crosstalk, and combined filtering interference, which is generated by Nyquist WDM channel model and analog-to-digital conversion due to the bandwidth limitation of the device chip. 7 × 224-Gb/s pulse modulated carrier-suppressed return-to-zero code quadrature phase-shift keying signal is transmitted over 1,600 km of single-mode fiber through an erbium-doped fiber amplifier The ability of this amplifier to amplify spectral efficiency is only 4b/s /Hz, however, this is by far the highest spectral efficiency achievable in pulse-modulated quadrature-phase-shift keying channels at rates in excess of 200 GB/s to our knowledge.


2. Generation and transmission of WDM Pulse Modulated Carrier Suppression Return-to-Zero Code Quadrature Phase Shift Keying (Nyquist-WDM PDMCSRZ-QPSK) signal at a rate of 7×224-Gbit/s



 
Figure 1. A rate of 7× 224-Gbit/s Nyquist WDM Pulse Modulation Carrier Suppression Return-to-Zero Code Quadrature Phase Shift Keying Signal Generation, Transmission and Inspection
PDMCSRZ-QPSK signal. PM EDFA: polarization-maintaining EDFA.
MZM1-2 : Mach–Zehnder intensity modulator. EA: electrical amplifier.
DL1-2: time delay line. PM-OC: polarization-maintaining optical coupler.
PBC: polarization beam combiner. ATT: optical attenuator. TOF: tunable
optical filter. ADC: analog-digital converter. The eye diagrams of the signalbefore and after CSRZ carver are inserted as (a) and (b), respectively. 

        Figure 1 shows the experimental setup used to generate and transmit a Nyquist WDM pulse-modulated carrier-suppressed return-to-zero code QPSK signal at a rate of 7 × 224-Gbit/s. The 56 Gbaud 0.5 Vp-p binary electrical signal is generated from an electronic multiplexer and a pattern generator. These odd and even channels are implemented with two sets of 1×4 polarization maintaining fiber couplers before independent co- and quadrature modulations. Lightwaves with continuous wavelengths were generated from seven external cavity lasers (E from these lasers with a linear bandwidth of less than 100kHz and an output power of 14.5dBm. Two in-phase and quadrature modulators (I/Q MOD) were used with to modulate two carriers with quadrature phase shift keying signals respectively. The two in-phase and quadrature modulators (I/Q MOD) are separately separated by two sets of random binary sequences with increased power of 56-Gb/s , each binary sequence has a byte length of (213-1) × 4. Each in-phase and quadrature modulator (I/Q MOD) contains two parallel Mach-Zehnder modulators (MZM) , both Mach-Zehnder modulators utilize bias voltage at zero and full swing drive to achieve 0-phase and π-phase modulation. The phase difference between the upper and lower branches of the in-phase and quadrature modulators is controlled at π /2. This 67% Kraft carrier-suppressed return-to-zero code is implemented with a one-arm Mach-Zehnder intensity modulator driven by a 28-GHz sinusoidal radio frequency signal with a DC bias at zero. The eye diagrams before and after the Kraft carrier-suppressed return-to-zero code are inserted at Figure 1(a) and (b), respectively. The polarization multiplexing of each path is achieved by a polarization multiplexer. This polarization multiplexing The device includes a polarization maintaining fiber coupler to split the signal, an optical delay line (DL2 and DL3) to provide a delay of 150 symbols, and a polarization combiner to recombine the signal. These odd and even channels are filtered and combined between them using a programmable wavelength selective switch at 50GHz intervals. The wavelength selective switch has an insertion loss of 7dB. Signals on the 50GHz grid before and after the wavelength selective switch The spectra of , respectively, are given in Figure 2. It can be seen from the figure that the carrier suppression spectrum combined with the wavelength selective switch filter has a Nyquist function, such as a filter shape used to compensate for the effect of narrowband filtering.



 
Figure 2 Spectra of the signal before and after the wavelength selective switch on the 50GHz grid

        The Nyquist wavelength division multiplexed signal is transmitted in a straight transmission line using 20 x 80 km of single-mode fiber. There is an average loss of 18 dB per span and a dispersion of 17 picoseconds/km/nm at 1550nm, rather than compensating in the dispersion of light. Erbium-doped fiber amplifiers are used to compensate for the energy loss per span. The total transmit power per span (after the EDFA) is 10 dBm, corresponding to 1 dBm per channel at 224 Gb/s. . On the receiver side, a 1 nm tunable optical filter with a 3-dB bandwidth is used to select the desired channel. An external cavity laser with a linewidth of less than 100kHz) is used as a local oscillator. A polarization diversity 90 degree hybrid is used to achieve coherent detection of the different phases of the polarization and local oscillators, and the received signal before balancing the detection. The digital-to-analog converter was implemented on a digital oscilloscope with a sample rate of 80 GSa/s and an electronic bandwidth of 30 GHz. For digital signal processing, electrical polarization recovery is implemented using a three-stage blind equalization scheme: first, the clock is extracted using a "square filtering" method, and then the digital signal is resampled at twice the baud rate based on the recovered clock. Second, a T/2-spaced time-domain finite impulse response filter is used for dispersion compensation, and the filter coefficients in the filter are calculated from the known fiber dispersion transfer function using the frequency-domain truncation method. Third, a two complex-valued, 13-tap, T/2-spaced adaptive finite impulse response filter, based on the classical constant modulus algorithm, is used to recover the modulus of the quadrature phase-shift keying signal. Carrier recovery is performed in a subsequent step where the power of the fourth carrier is used to estimate the frequency offset between the local oscillation and the received optical signal. After frequency offset compensation, phase recovery is performed in phase rotation speed using the power algorithm of the fourth carrier. In addition, a post filter converts the quadrature phase shift keying to quadrature duobinary. Here, the post filter is implemented by adding a finite impulse response filter with a simple transfer function of T delay of H(z) = 1+Z-1 [11]. The post filter not only reduces inter-symbol interference (ISI), but also roughly reverses the frequency response of the Nyquist channel. That's why it can reduce crosstalk from adjacent channels. Finally, four independent two-state maximum likelihood sequence detections based on the Viterbi algorithm are used for both polarized symbol decisions for in-phase and quadrature paths. 


                                                                           three. Experimental results 

         In this experiment, the error exceeds 12 × 106 bits (12 datasets, each containing 106 bits). Differential decoding is used to solve the π/2 phase ambiguity problem. Figure 3 shows the bit error rate (BER) as an optical signal-to-noise ratio (0.1 nm resolution, dual polarization). We measure the performance of a particular channel at 1553.75 nm by adjusting the frequency of the local oscillation. As shown by the star-filled curve in Figure 3, at a bit error rate of 1 × 10-3, for Nyquist WDM Pulse Modulated Carrier Suppression at 56G baud, return-to-zero code QPSK For example, the required optical signal-to-noise ratio is 22.4 dB/0.1 nm. By disabling the adjacent channels at 50 GHz, the bit error rate performance of a single channel using a wavelength selective switch (WSS) is shown by the filled-circle curve in the figure, with a penalty of about 0.5 dB due to crosstalk. As shown by the square-filled curve in the figure, at a bit error rate of 1 × 10-3, for a single channel, the required optical signal-to-noise ratio is 22.2 dB by temporarily turning off adjacent channels and using a wavelength selective switch /0.1 nm. Compared to the single-channel case with a wavelength selective switch at 50 GHz, there is an optical signal-to-noise ratio loss of 0.3 dB. This is because a carrier suppressed return-to-zero (CSRZ) signal has a wide spectrum, and therefore, a portion of the optical power cannot be detected by a coherent receiver due to the limited bandwidth of a real-time oscilloscope or digital-to-analog converter. For a pulse-modulated NRZ code QPSK modulated signal (by turning off the radio frequency of the return-to-zero code), at a bit error rate of 1 × 10-3, the required signal-to-noise ratio of the optical signal (as filled with diamonds The curve shown in ) is 23.2 dB, which is about 0.8 dB higher than the optical signal-to-noise ratio of the pulse-modulated carrier-suppressed return-to-zero code quadrature-phase keying modulation signal. This means that carrier suppressed return-to-zero codes can tolerate a narrower filtering effect than non-return-to-zero codes. In addition, we can measure and confirm that all other channels exhibit similar performance. The results are also compared with those obtained using conventional quadrature-phase key modulation decisions without post-filter and maximum likelihood sequence detection. For single-channel and Nyquist WDM Pulse Modulated Carrier Suppressed Return-to-Zero Code Quadrature Phase Keying modulation channels, the bit error rate is 3. At 8×10 −3 , the optical signal-to-noise ratio requirements (as indicated by the downward-filled triangle and star (*) curves, respectively) are 23 dB and 24 dB, respectively. The improvement after adding the maximum likelihood sequence detection algorithm through the post filter is about 3 to 4 dB. For Nyquist WDM Pulse Modulated Carrier Suppressed Return-to-Zero Code Quadrature Phase Keying modulated signals only, the OSNR gain after detection using the maximum likelihood sequence (as shown by the persimmon-filled curve) is 1 dB .



 
Fig. 3 Bit error rate performance of Nyquist WDM channel at 1553.75nm PF: Post filter

      Fig. 4 shows that only the erbium-doped fiber amplifier (EDFA) is applied for amplification, before 1600km single-mode fiber transmission and after the spectrum. After 1600km single-mode transmission, the signal ensembles obtained before and after passing through the post-filter are (a) and (b) in Figure 4, respectively. After more than 1600km of single-mode fiber transmission, the optical signal-to-noise ratio was 21.3 dB. We can see that the converted duobinary ensemble after applying the post filter is very clear.


 
Fig. 4 Optical spectrum (0.5 nm resolution) before and after transmission over 1600 km of single-mode fiber, and pulse-modulated carrier-suppressed return-to-zero code quadrature phase shift keying before and after transmission through a post-filter The modulated signal groups are shown in (a) and (b) respectively.

       Figure 5(a) shows the bit error rate performance of the Nyquist WDM channel after transmission over 1600 km of single-mode fiber, after transmission as a function of transmit power. A transmit power of 9.5 dBm provides the best bit error rate performance of nearly 1×10-3. Figure 5(b) shows the measured bit error rate results for all channels after each channel is transmitted with an optimal transmit power of 1dBm. After transmission over 1600 km of single-mode fiber, the average received optical signal-to-noise ratio for all channels was 21.3 dB. The bit error rate for all channels is less than the forward error correction (FEC) limit of 3.8 × 10-3 [12].


 
Figure 5(a) Bit error rate performance of Nyquist WDM channel at transmit power after 1553.75-nm and single-mode fiber transmission over 1600 km. (b) Bit error rate of the signal in all channels after transmission over 1600 km of single-mode fiber.

                                                                      Four. in conclusion

        We have demonstrated a 7 × 224-Gb/s Nyquist WDM transmission over a single-mode fiber over 1600 km, recorded at a spectral efficiency of 4 b/s/Hz using only erbium-doped fiber amplifier (EDFA) amplification Generation and transmission of multiplexed pulse modulated carrier-suppressed return-to-zero code quadrature phase shift keying modulated signals. We employ coherent detection combining noise suppression post-filters and 1-bit maximum likelihood sequence detection. After transmission over 1600 km of single-mode fiber transmission, all channels have a bit error rate less than a forward error correction (FEC) with a bit error rate limit of 3.8 x 10-3.

Manuscript received February 21, 2012; revised April 23, 2012; accepted April 24, 2012. Date of current version May 30, 2012.
Z. Dong is with ZTE Corporation, Morristown, NJ 07960 USA, and also with the Georgia Institute of Technology , Atlanta, GA 30332 USA (e-mail:
[email protected]).
J. Yu is with ZTE Corporation, Beijing 100876, China, and also with Fudan University, Shanghai20443, China (e-mail: yu.jianjun @zte.com.cn).
Z. Jia and H.-C. Chien are with ZTE Corporation, Morristown, NJ 07960 USA (e-mail: [email protected]; [email protected] ).
X. Li is with Fudan University, Shanghai 20443, China (e-mail:[email protected]).
G.-K. Chang is with the Georgia Institute of Technology, Atlanta, GA 30332
USA (e-mail: [email protected]).
Color versions of one or more of the figures in this letter are available
online at http://ieeexplore.ieee.org.

                                                                          注  解
[1] C. R. S. Fludger, et al., “Coherent equalization and POLMUX-RZDQPSK for robust 100-GE transmission,” J. Lightw.Technol., vol. 26, no. 1, pp. 64–72, Jan. 1, 2008.

[2] J.-X. Cai, et al., “112×112 Gb/s transmission over 9360 km with channel spacing set to the baud rate (360% spectral efficiency),” in Proc. ECOC 2010, Torino, Italy, Sep., pp. 1–3, paper PD2.1.
[3] J. Yu, et al., “112.8-Gb/s PM-RZ-64QAM optical signal generation and transmission on a 12.5 GHzWDM grid,” in Proc. OFC 2010, San Diego, CA, pp. 1–3, paper OThM1.

[4] Z. Dong, J. Yu, C.-H. Chien, S. Shi, Y. Xia, and C. Ge, “24 Tb/s (24×1.3 Tb/s) WDM transmission of terabit PDM-CO-OFDM superchannels over 2400 km SMF-28,” in Proc. OECC, Jul. 2011, pp. 756–757, paper PDP6.

[5] B. Zhu, X. Liu, S. Chandrasekhar, D. W. Peckham, and R. Lingle, “Ultralong-haul transmission of 1.2-Tb/s multicarrier no-guard-interval CO-OFDM superchannel using ultralarge-area fiber,” IEEE Photon. Technol. Lett., vol. 22, no. 11, pp. 826–828, Jun. 1, 2010.

[6] Y.-K. Huang, et al., “10×456-Gb/s DP-16 QAM transmission over 8×100 km of ULAF using coherent detection with a 30-GHz analog-todigitalconverter,” in Proc. OECC, Jul. 2010, pp. 1–2, paper PDP2.

[7] A. H. Gnauck, et al., “Generation and transmission of 21.4-Gbaud PDM 64-QAM using a high-power DAC driving a single I/Q modulator,” in Proc. OFC, Mar. 2011, pp. 1–3, paper PDPB2.

[8] X. Zhou, et al., “8×450-Gb/s, 50-GHz-spaced, PDM-32 QAM transmission over 400 km and one 50 GHz-grid road,” in Proc. OFC, 2011, pp. 1–3, paper PDPB3.

[9] A. H. Gnauck, P. J. Winzer, G. Raybon, M. Schnecker, and P. J. Pupalaikis, “10×224-Gb/s WDM transmission of 56-Gbaud PDMQPSK signals over 1890 km of fiber,” IEEE Photon. Technol. Lett., vol. 22, no. 13, pp. 954–956, Jul. 1, 2010.

[10] C. Xie, G. Raybon, and P. J. Winzer, “Hybrid 224-Gb/s and 112-Gb/s PDM-QPSK transmission at 50-GHz channel spacing over 1200-km dispersion managed LEAF spans and 3 ROADMs,” in Proc. OFC, Mar. 2011, pp. 1–3, paper PDPD2.

[11] J. Li, Z. Tao, H. Zhang, W. Yan, T. Hoshida, and J. C. Rasmussen, “Enhanced digital coherent receiver for high spectral-efficiency dualpolarization quadrature duobinary systems,” in Proc. ECOC 2010, Torino, Italy, Sep., pp. 1–3, paper Th.10.A.3.

[12] Forward Error Correction for High Bit-Rate DWDM Submarine Systems, ITU-T Standard G.975.1, 2004.

 

Guess you like

Origin http://43.154.161.224:23101/article/api/json?id=326381388&siteId=291194637