We experimentally investigate the transmission performance of 80 × 112-Gb/s polarization-division-multiplexed quadrature phase shift keying (PDM-QPSK) signals over large effective area fiber and standard single mode fiber (SSMF) links with Raman amplification. The large effective area fiber offers higher optimum launch power and longer reach than SSMF. The maximum reach of 5200-km is obtained using large effective area fiber. The Gaussian noise (GN) model is explored to fit with experimental data for optimum power.
© 2014 Optical Society of America
The technology advancements in optical fiber communications have enabled high speed optical transmission system with a data rate of 100-Gb/s and beyond using advanced modulation formats and digital coherent detection. As research continues exploring technologies for transmission systems above 100-Gb/s [1–3] using various advanced modulation formats with high spectral efficiency (SE), commercial transport systems development are focusing on 100-Gb/s systems using polarization-division-multiplexed (PDM) qurdrature phase shift keying (QPSK) modulation with digital coherent detection . Taking advantage of recent developments in digital signal processing (DSP) and optical coherent detection, PDM-QPSK can transmit data with 4 times of its baud rate and be able to compensate chromatic dispersion (CD) and polarization mode dispersion (PMD) electronically at receiver.
Similar to a direct detection system, a key parameter to determine the performance of a 100-G PDM-QPSK dense wavelength division multiplexing (DWDM) system is the optical signal to noise ratio (OSNR) at the receiver, as the amplified spontaneous emission (ASE) is the dominate noise in a system using optical amplifiers. Improve the OSNR at the receiver is a direct method to improve an optical system performance. In general, higher OSNR can be accomplished by increasing signal launch power, reducing fiber loss, and reducing the accumulated optical noises. However, as the optical signal power launched into fiber increasing, the nonlinear impairments of transmission fiber increase and eventually impose the fundamental limits on channel capacity [5,6]. Using large effective area (Aeff) fiber can improve fiber nonlinearities by reducing the optical power density [7–9]. Using large Aeff fiber allows higher power into fiber and significantly enhances the transmission distance in long-haul system. Another method to improve OSNR is to use distributed Raman amplification. Distributed Raman amplification can provide larger power and lower noise compared to traditional erbium doped fiber amplifier (EDFA), especially for large capacity system with high channel count and large spans. Typically, larger Aeff fiber has a relatively low Raman amplification efficiency and requires higher Raman pump power. Therefore, the transmission performance benefit of large Aeff fibers with Raman amplified spans is of interest of research.
In this paper, we experimentally investigate the transmission performance improvement obtained by using large Aeff fiber compared with standard single mode fiber (SMF) for 112-Gb/s fully-loaded 80 channel 50-GHz spaced DWDM systems using distributed all-Raman amplifiers. The transmission distance using large Aeff fiber is extended to 5200-km, about 63% longer than using standard SMF (SSMF). In addition, an analytic Gaussian Noise (GN) [10,11] model for nonlinear impairments estimation is investigated to fit with experimental data and compared with experiments results.
2. Experimental set-up
The experimental setup of the transmission system is shown in Fig. 1. The transmitters consist of 80 distributed feedback (DFB) lasers at wavelengths ranging from 1530.31 to 1563.83-nm in the C-band on the 50-GHz-spaced frequency grid. The odd and even channels are multiplexed separately by two arrayed waveguide grating-router (AWG) and modulated independently by two QPSK modulators. Each modulator is driven by two sets of 28-Gb/s pseudo-random bit sequences (PRBS) with a length of 215-1. The output from each modulator is split into two paths with a relative delay of 84 symbols before being polarization multiplexed by a polarization beam combiner (PBC) to form a PDM-QPSK channel at 112-Gb/s, which simulates a net data rate of 100-Gb/s with a typical 7% overhead for forward error correction (FEC). The odd and even channels are spectrally interleaved through a 50-GHz interleaver (IL) and amplified by an EDFA before it is sent to a re-circulating loop for transmission. In order to measure the bit-error rate (BER) of a channel, a tunable external cavity laser (ECL) with a line-width of ~100-kHz is used. Each channel under measurement is switched from the DFB source to the tunable ECL source.
The re-circulating loop consists of four 100-km spans of either SSMF or TeraWave SLA + fiber. The SLA + fiber is a single mode fiber with larger Aeff of 125-μm2. After each fiber span, a bi-directional all-Raman amplifier is used to compensate the span loss. Two semiconductor lasers at 1429 and 1447-nm are used as the co-propagating pumps to provide average forward gain of 3.5-dB. Three fiber lasers at 1429, 1447, and 1465-nm provide counter-propagating pumping with backward gain of 17-dB. The average total pump powers for each SLA + or SSMF span are 1.32-W (235-mW co-pump with 1085-mW counter-pump) and 0.83-W (175-mW co-pump with 655-mW counter-pump), respectively. The Raman amplifier provides a total of 20.5-dB gain to compensate the average total span loss (fiber + components). The estimated effective noise figures for SLA + or SSMF span are −0.85-dB and −1.12-dB, respectively. No optical dispersion compensating fibers or units are used in the loop and the accumulated CD is compensated at the receiver in the electronic domain. After the last span, a loop-synchronous PC is followed by a dynamic GE to flatten the spectrum after each loop. An EDFA is used to compensate the losses of the SW, GE and loop-synchronous PC. The signal launch power into the first fiber span is controlled by the output of the EDFA before the re-circulating loop. The launch powers into the following fiber spans follow the first span since Raman gain in each span is equal to the span loss and will not change with signal power as the distributed Raman amplifications are not operated at depletion regime in this experiment. The signal powers into each loop are controlled by the EDFA within the loop.
At the receiver side, the 80 WDM channels are separated by another IL/AWG combination, before being measured individually. A typical digital coherent receiver, consisting of a polarization-diversity optical hybrid, an OLO using a tunable ECL, and four balanced PDs, is used. The electrical waveforms are digitized by four 50-Gsamples/s analog-to-digital converters (ADCs) in a real-time sampling scope. The digitized waveforms of 1-million samples each are processed offline in a computer to recover the data. The digital signal processing of data uses typical PDM-QPSK algorithms to perform electronic CD and PMD compensation, polarization de-multiplexing, frequency and phase recovery, as well as BER calculation . The final BERs are calculated using direct error-counting and averaged over 1-million samples from which Q factors are calculated.
The key parameters of large Aeff fiber SLA + and SSMF fiber used in the experiments are listed in Table 1. The SLA + is a SMF with germanium-doped core and a depressed-index inner cladding region . The effective area of SLA + at 1550-nm is 125-μm2. Its dispersion at 1550-nm is 19.9-ps/nm-km, about 17% larger than that of SSMF. The peak Raman gain efficiency of SLA + is smaller than that of SSMF, therefore required higher pumps power for Raman amplifiers to provide the same gain.
3. Experimental results and discussions
The noise loaded back to back (B2B) system performance is shown in Fig. 2. The Q-factor is plotted as a function of OSNR (0.1-nm noise bandwidth, both polarizations) for a channel in the middle of transmission band (#38). The square and round markers represent experimental data for single channel and WDM configuration, respectively. The solid line is the linear fitting for WDM configuration. The R-square value is 0.9991, which indicates a good linear relationship of the Q-factor with OSNR within the range of 11.5 ~16.5-dB. There is no difference for two configurations. Consider using 7%-overhead enhanced FEC, the FEC threshold is Q of 8.6-dB (corresponding to a BER of 3.8 × 10−3). The required OSNR at the FEC threshold is about 14.2-dB.
To obtain the best performance, the signal launch power needs be optimized. Figure 3 shows the measured Q-factor (derived from BER) of channel #38, as a function of signal launch power, for SLA + and SSMF links, respectively. The round and triangle markers represent experiment results at transmission distance of 3200-km for SLA + and SSMF, respectively. At low power range, the Q-factor increases linearly as launch power increasing. The Q-factor for SLA + link is about 1-dB larger than that of SSMF link for the same launch power, which can be attributed to the larger Aeff of SLA + fiber. The large Aeff fiber has much less double Rayleigh scattering associated multiple path interference (MPI) noises  compared with SSMF, and furthermore, the large Aeff fiber has much high Raman gain saturation threshold, allowing the distributed Raman amplification in linear regime even for all Raman systems. At high power range, due to high non-linear effects, the Q-factor decreases as launch power increasing. The optimal signal launch power is a trade-off between the accumulated ASE noise and the fiber nonlinear impairments. The optimum power per channel for SLA + system and SSMF system are −3.2-dBm and −5.8-dBm, respectively. The optimum power for SLA + span is about 2.6-dBm higher than that for SSMF span. The optimum Q-factor in SLA + system at 3200-km is about 1.5-dB higher than that in SSMF system. With this higher Q-factor, the SLA + link can transmit longer distance compared with SSMF. The maximum reach of SLA + observed in the experiment is 5200-km. The optimal signal launch power is −4.1-dBm, about 0.9-dB lower than the optimal value required at 3200-km reach. It is worth to note that the SLA + requires higher Raman pump power to achieve better performance. The Raman pump powers for each SLA + span is about 1.32-W, which is 0.5-W higher than that for each SSMF span. Therefore, SLA + requires an extra total 16-Watts pump power to achieve 1.5-dB Q-factor improvement at 3200-km and an extra 42-Watts to extend reach to 5200-km.
The solid lines in Fig. 3 are theory fittings based on the GN model . GN model assumes that, for an uncompensated coherent optical system, non-linear interference (NLI) on dense WDM signals can be modeled as additive Gaussian noise due to the unmitigated dispersion effects. A modified effective OSNR which includes both ASE and NLI noise contribution is introduced to describe the system performance. The effective OSNR can be calculated as follows [10,11]:Fig. 2. The PASE can be calculated through measured OSNR at the receiver. To get the best fitting as shown in Fig. 3, α is 0.59. The η parameters are chosen as 0.012 and 0.0058 for SLA + at 5200-km and 3200-km, respectively, and 0.024 for SSMF at 3200-km. The analytic GN model is found to fit well with these experimental data.
The system performances at optimum launch power for SLA + system and SSMF system are plotted in Fig. 4. The square and round markers represent Q-factor and OSNR for SLA + system at 3200-km and 5200-km, respectively. The triangle markers are experiment results for SSMF system at 3200-km. The average Q-factors are 9.3-dB and 9.1-dB for SLA + at 5200-km and SSMF at 3200-km, respectively. All 80 channels for SLA + after 5200-km transmission are above the FEC limit of 8.6-dB and would yield a BER below 10−13 after correction by enhanced FEC.
The OSNR and Q-factor for SLA + and SMF are summarized in Table 2. It can be seen that the mean OSNR value after 3,200-km transmission is about 2.3-dB higher for SLA + compared with SSMF. A 2.3-dB OSNR increase will have about 2-dB Q improvement based on B2B performance (Fig. 2). After 3200-km transmission, an average OSNR difference of 2.3-dB only results in a mean Q difference about 1.5-dB for SLA + and SSMF from the Table 2, which means that fiber transmission causes about 0.5-dB transmission penalty, which is reasonable.
Figure 5 shows Q-factor as functions of distance for SLA + system of different channel #2, #38, #78, respectively. Solid markers represent experiment results for SLA + system and unfilled markers represent the results for SSMF system at 3200-km.
As expected, Q-factor decreases as the transmission distance increasing. Q-factor of channel #38 is about 0.6 ~0.8-dB higher than that of channel #2 and #78. This can be attributed to the OSNR difference of three channels, as shown in Fig. 4(b). Due to the limitation of the system, the OSNRs at long and short edge wavelength are approximately 0.5 ~1-dB lower than those at center. The system performance is determined by the worst performing channel. In this experiment, the SLA + achieves the reach of 5200-km which is about 63% improvement of reach compared with SSMF system.
We have experimentally compared the system performance of 112-Gb/s fully-loaded 50-GHz spaced DWDM transmission over large Aeff fiber TeraWave SLA + and SSMF spans using distributed all-Raman amplifiers. The SLA + fiber offers ~2.6-dBm higher optimum signal launch power with an improvement of Q-factor about 1.5-dB for a reach of 32 × 100-km systems. The average OSNR improvement for SLA + is about of 2.3-dB. A maximum transmission distance of 52 × 100-km is obtained with SLA + fiber spans, which is about 63% longer than SSMF link. The GN model is explored to fit with experimental data for optimum power. Good match is obtained for optimum power curve.
This work is partially support by NSF MRI grant 1040223.
References and links
1. X. Zhou, L. Nelson, P. Magill, R. Issac, B. Zhu, D. Peckham, P. Borel, and K. Carlson, “4000km transmission of 50GHz spaced, 10x494.85-Gb/s hybrid 32-64QAM using cascaded equalization and training-assisted phase recovery,” in Proc. OFC/NFOEC 2012, Los Angeles, California, Mar. (2012), PDP5C.6. [CrossRef]
2. 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,” in Proc. OFC/NFOEC 2012, Los Angeles, California, Mar. (2012), PDP5B.3. [CrossRef]
3. P. J. Winzer, A. H. Gnauck, S. Chandrasekhar, S. Draving, J. Evangelista, and B. Zhu, “Generation and 1,200-km transmission of 448-Gb/s ETDM 56-Gbaud PDM 16-QAM using a single I/Q modulator,” in Proc. of ECOC 2012, Turin, Italy, Sept. (2012), PDP2.2.
4. Optical Internetworking Forum, 100G Ultra Long Haul DWDM Framework Document.
5. O. Bertran Pardo, J. Renaudier, G. Charlet, P. Tran, H. Mardoyan, M. Salsi, and S. Bigo, “Impact of nonlinear impairments on the tolerance to PMD of 100Gb/s PDM-QPSK data processed in a coherent receiver,” in Proc. of ECOC 2008, Brussels, Belgium, Sept. (2008), We.3.E.1. [CrossRef]
7. G. Charlet, M. Salsi, P. Tran, M. Bertolini, H. Mardoyan, J. Renaudier, O. Bertran-Pardo, and S. Bigo, “72x100Gb/s transmission over transoceanic distance, using large effective area fiber, hybrid Raman-Erbium amplification and coherent detection,” in Proc. of OFC/NFOEC 2009, San Diego, California, Mar. (2009), PDPB6. [CrossRef]
8. Y. Yamamoto, M. Hirano, K. Kuwahara, and T. Sasakiet, “OSNR-enhancing pure-silica-core fiber with large effective area and low attenuation,” in Proc. of OFC/NFOEC 2010, San Diego, California, Mar. (2010), OTuI2. [CrossRef]
9. B. Zhu, S. Chandrasekhar, X. Liu, and D. W. Peckham, “Transmission performance of a 485-Gb/s CO-OFDM superchannel with PDM-16QAM subcarriers over ULAF and SSMF-based links,” IEEE Photon. Technol. Lett. 23(19), 1400–1402 (2011). [CrossRef]
10. P. Poggiolini, “The GN model of non-linear propagation in uncompensated coherent optical systems,” J. Lightwave Technol. 30(24), 3857–3879 (2012). [CrossRef]
11. J.-X. Cai, H. Zhang, H. G. Batshon, M. Mazurczyk, O. V. Sinkin, D. G. Foursa, A. N. Pilipetskii, G. Mohs, and N. S. Bergano, “200 Gb/s and dual wavelength 400 Gb/s transmission over transpacific distance at 6.0 b/s/Hz spectral efficiency,” J. Lightwave Technol. 32(4), 832–839 (2014).
13. P. Parolari, L. Marazzi, L. Bernardini, and M. Martinelli, “Double Rayleigh scattering noise in lumped and distributed Raman amplifiers,” J. Lightwave Technol. 21(10), 2224–2228 (2003). [CrossRef]