Optical signal suppression by a cascaded SOA/RSOA for wavelength reusing reflective PON upstream transmission

Vol. 25, No. 19 18 Sep 2017 OPTICS EXPRESS 22851 Optical signal suppression by a cascaded SOA/RSOA for wavelength reusing reflective PON upstream transmission SANG MIN JUNG, KYOUNG HAK MUN, SOO MIN KANG, AND SANG KOOK HAN* Department of Electrical & Electronic Engineering, Yonsei University, Seoul, South Korea *skhan@yonsei.ac.kr Abstract: An optical signal suppression technique based on a cascaded SOA and RSOA is proposed for the reflective passive optical networks (PONs) with wavelength division multiplexing (WDM). By suppressing the downstream signal of the optical carrier, the proposed reflective PON effectively reuses the downstream optical carrier for upstream signal transmission. As an experimental demonstration, we show that the proposed optical signal suppression technique is effective in terms of the signal bandwidth and bit-error-rate (BER) performance of the remodulated upstream transmission. 2017 Optical Society of America OCIS codes: (060.2330) Fiber optics communications; (060.2360) Fiber optics links and subsystems. References and links 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. G. L. Kazovsky, W.-T. Shaw, D. Gutierrez, N. cheng, and S.-W. Wong, Next-Generation Optical Access Networks, J. Lightwave Technol. 25, 3428 3442 (2007). N. Cvijetic, OFDM for next-generation optical access networks, J. Lightwave Technol. 30(4), 384 398 (2012). F. Saliou, G. Simon, P. Chanclou, A. Pizzinat, H. Lin, E. Zhou, and Z. Xu, WDM PONs based on colorless technology, Opt. Fiber Technol. 26, 126 134 (2015). E. Wong, K. L. Lee, and T. B. Anderson, Directly modulated self-seeding reflective semiconductor optical amplifiers as colorless transmitters in wavelength division multiplexed passive optical networks, J. Lightwave Technol. 25(1), 67 74 (2007). W. Zhan, T. Tanemura, S. Yamauchi, M. Mukaikubo, and Y. Nakano, Uncooled (25 50 C) operation of selfseeded RSOA for low-cost colorless WDM-PON transmitter, ECOC, We1.5.5., Valencia, Spain (2015). R. P. Giddings, E. Hugues-Salas, H. Q. Jin, J. L. Wei, and J. M. Tang, Experimental demonstration of real-time optical OFDM transmission at 7.5 Gb/s over 25-km SSMF using a 1-GHz RSOA, IEEE Photonics Technol. Lett. 22(11), 745 747 (2010). S. Straullu, F. Forghieri, G. Bosco, V. Ferrero, and R. Gaudino, Coherent Reflective PON architecture: can it be made compatible with TWDM-PON? ECOC, We.3.F.1, London, UK (2013). H. Feng, J. Ge, S. Xiao, and M. P. Fok, Suppression of Rayleigh backscattering noise using cascaded-soa and microwave photonic filter for 10 Gb/s loop-back WDM-PON, Opt. Express 22(10), 11770 11777 (2014). H. Shim, H. Kim, and Y. C. Chung, 28-Gb/s Upstream Transmission in RSOA-based WDM PON Using Polar RZ PAM-N Format and Direct Detection, in Proc. Optical Fiber Communication Conference (OFC, 2016), paper Tu2C.3. J. L. Wei, E. Hugues-Salas, R. P. Giddings, X. Q. Jin, X. Zheng, S. Mansoor, and J. M. Tang, Wavelength reused bidirectional transmission of adaptively modulated optical OFDM signals in WDM-PONs incorporating SOA and RSOA intensity modulators, Opt. Express 18(10), 9791 9808 (2010). H. Takesue and T. Sugie, Wavelength channel data rewrite using saturated SOA modulator for WDM networks with centralized light sources, J. Lightwave Technol. 21(11), 2546 2556 (2003). E. Conforti, C. M. Gallep, S. Ho, A. C. Bordonalli, and S. M. Kang, Carrier reuse with gain compression and feed-forward semiconductor optical amplifier, IEEE Trans. Microw. Theory Tech. 50(1), 77 81 (2002). H. Takesue, N. Yoshimoto, Y. Shibata, T. Ito, Y. Tohmori, and T. Sugie, Wavelength channel data rewrite using semiconductor optical saturator/modulator, J. Lightwave Technol. 24(6), 2347 2354 (2006). S. Ó. Dúill, L. Marazzi, P. Parolari, R. Brenot, C. Koos, W. Freude, and J. Leuthold, Efficient modulation cancellation using reflective SOAs, Opt. Express 20(26), B587 B594 (2012). R. Ramaswami and P. A. Humblet, Amplifier induced crosstalk in multichannel optical networks, J. Lightwave Technol. 8(12), 1882 1896 (1990). R. Boula-Picard, M. Alouini, J. Lopez, N. Vodjdani, and J.-C. Simon, Impact of the gain saturation dynamics in semiconductor optical amplifiers on the characteristics of an analog optical link, J. Lightwave Technol. 23(8), 2420 2426 (2005). #300447 Journal 2017 https://doi.org/10.1364/oe.25.022851 Received 20 Jun 2017; revised 6 Sep 2017; accepted 6 Sep 2017; published 11 Sep 2017

Vol. 25, No. 19 18 Sep 2017 OPTICS EXPRESS 22852 1. Introduction Flexible and dynamic multiple access in optical access networks has become one of the most important requirements for the evolution of future access networks to support mobile broadband, Internet-of-Things (IoT), and enterprise connectivity [1,2]. Especially, using wavelength division multiplexing (WDM) technologies, a passive optical network (PON) can provide a high transmission capacity supporting large number of ONUs [1]. For these reasons, WDM PONs based on the reflective semiconductor optical amplifier (RSOA), the reflective WDM PONs, have been a spotlight in the area of the next-generation optical access network by the virtue of the laserless optical network unit (ONU) with colorless operation [3 10]. Reflective WDM PONs can provide high performance optical sources to all ONUs by distributing the centralized light source (CLS) from the optical line terminal (OLT) to ONU. Each ONU then modulates the distributed optical source using reflective modulators like RSOA, reflective electro-absorption modulator (REAM), and semiconductor optical amplifier (SOA) with fiber-loop, leading to laserless and colorless operation [3 6]. Nowadays, in spite of the limited bandwidth of the RSOA, the upstream transmission capacity is increased more than 10 Gb/s in many demonstrations [8 10]. However, most of the proposed reflective WDM PON networks used two wavelengths for bidirectional transmission: one for the downstream transmission and another for the upstream transmission. Consequently, as the number of ONUs are increasing in the access network, a required optical bandwidth is also increased twice because of the increasing number of available wavelength channels, and this lead to the inefficient wavelength usages in the limited optical bandwidth in the near future. In order to solve this issue, wavelength reuse techniques that use a gain-saturated SOA have been proposed [11 14]. These systems achieved wavelength reuse of downstream signal for upstream transmission, and enhanced the capacity in the limited optical bandwidth by employing a single wavelength for the both upstream and downstream transmission. This solution has become a one of the most promising solutions for the WDM-PON by reducing the system complexity and enhancing the optical spectral efficiency. However, these techniques have limitations in terms of the usable signal bandwidth for both upstream and downstream transmission and the usable modulation depth of the downstream signal because of the limited relaxation time of the SOA [11]. Other works have been proposed for wavelength reuse like SOA with external modulator [12], cascade SOAs [13], and RSOA [14]. These approaches enhance the upstream capacity by using additional modulators, however, downstream capacity is still limited by the response of the SOA. Because the upstream transmission performance is very sensitive to the signal suppression ratio of the downstream signal, the modulation depth of the downstream signal was required to be minimized for enhancing the signal suppression efficiency. These two critical limitations lead to limited transmission capacity for access network traffic. In this paper, we proposed a novel approach for wavelength reuse technique based on a cascaded SOA/RSOA configuration. Using the proposed technique, the recovered optical carrier effectively suppresses the modulated downstream signal and reduces the dependency of the bandwidth and modulation depth of downstream signal. In the experimental demonstrations, the achieved signal suppression bandwidth was 4 GHz with the modulation depth of the 0.9 for bidirectional transmission. 2. Signal suppression in cascaded SOA/RSOA configuration The basic principle of the proposed signal suppression technique is retaining the optical signal-to-noise ratio (OSNR) of the optical carrier by lowering the optical power of the modulated signal. The proposed signal suppression technique can be divided into three steps. In the first step, the optical power of the incoming signal is reduced to a sufficiently low level with an optical power attenuator. Then, the incoming optical signal (comprising two optical components: the optical carrier and the modulated signal) is attenuated to the same

Vol. 25, No. 19 18 Sep 2017 OPTICS EXPRESS 22853 scale. In this step, the modulated signal and the optical carrier maintain the same OSNR in spite of the absolute power reduction of the each components. Fig. 1. Proposed signal suppression technique based on the cascaded SOA/RSOA. In the second step, the first SOA amplifies the incoming optical signal to increase the total optical power. In this step, the small input optical power of the SOA causes significant amplified spontaneous emission (ASE) noise, leading to significant degradation of the OSNR for each of the components of the incoming signal. Each of the amplified incoming optical signal components therefore has the optical gain with an enhanced noise which related with the bias current. The amount of ASE noise depends on the bias current of the SOA in the small incoming signal power condition. Therefore, the OSNR degradation of the modulated signal components is introduced with the total optical power increasing. In the third step, a deeply saturated RSOA is used for noise and signal suppression. In the deep saturated SOA, the gain distribution caused by the differential gains for each optical frequency components, and it depends on the distributions of photon densities [15-16]. Because the power of the optical carrier component is larger than the modulated optical signal component, the most of gains of RSOA dissipate by the optical carrier components, and the rest of optical gains are distributed to the modulated optical signal components and the ASE from the second step. As a result, the optical carrier components of the output signal regains its OSNR which is sufficient for using as the upstream optical carrier. In addition, because of the nonlinear characteristics of the gain saturated SOA, the RSOA suppressed the enhanced noise of SOA (see Fig. 1). By suppressing the noise and modulated signals with cascaded SOA/RSOA, the recovered optical carrier effectively reduces the dependency of the modulation depth and the interference between the remodulated upstream signal and the modulated downstream signal regardless of the signal bandwidth. Signal suppression techniques with the gain saturated SOAs were actively investigated in many previous works [11 14]. Most of the works have a limitation for the usable signal bandwidth, and it degrades the attractive advantages of wavelength reuse in the WDM architecture. In order to enhance the usable bandwidth, a cascade gain-saturated SOAs also was proposed. The use of multiple gain-saturated cascaded SOAs can enhance the usable

Vol. 25, No. 19 18 Sep 2017 OPTICS EXPRESS 22854 bandwidth by accumulating the effect of the gain-saturated SOAs, however, it shows unflatted signal suppression on the signal bandwidth (a strong signal suppression on the lower frequency components and a relatively weak signal suppression on the higher frequency components). However, the proposed signal suppression uses different mechanism for the signal suppression. It also uses cascaded two SOAs; the SOA in small signal amplification condition and the RSOA in the deep saturation condition. By using the nonlinear gain distribution between the optical carrier and the modulated optical signals, the signal suppression technique provides an effective wavelength reuse regardless of the usable signal bandwidth. The signal suppression techniques based on the gain-saturation of SOA requires high input optical power to operate SOAs in the gain saturation condition. Therefore, in order to compensate the power loss in the transmission link, additional preamplifier is required before the signal suppression. Another biggest problem of the multiple gain-saturated SOA is the linewidth broadening caused by the nonlinear response of the gain-saturated SOA. However, the proposed cascade SOA/RSOA only introduces small nonlinearities on the carrier by using only one gain-saturated SOA. Instead, the proposed cascade SOA/RSOA may introduce large noise enhancements compared with the gain-saturated SOA. This noise enhancement reduces the quality of the upstream optical carrier. However, gain-saturated RSOA can suppress the addictive noise of SOA with the weak downstream signal because of the nonlinear characteristics of the gain saturated SOA and the gain distribution difference. As a result, the proposed signal suppression techniques overcome the noise penalty by introducing effective signal suppression against the multiple gain-saturated SOAs in case of the larger usable signal bandwidth for bidirectional transmission system. 3. Experiments and Results Figure 2 shows an experimental proof-of-concept setup of the proposed wavelength reuse reflective PON. An external cavity laser (ECL) with a center wavelength of 1550.223 nm and linewidth of 50 khz was used as the optical source for the bidirectional transmission. Here, a Mazh-Zhender modulator (MZM) was used for the downstream signal modulation, and the used modulation format was a discrete multi-tone (DMT), the baseband version of the OFDM, in order to analyze the effect of the signal suppression in the frequency domain in detail. Because the DMT used multiple orthogonal subcarriers in the signal bandwidth, the signal performance can be analyzed dividing the signal bandwidth in the order of the subcarrier numbers. We configured cascaded SOA/RSOA configurations for wavelength reuse by suppressing downstream signals. The polarization dependent gains were 1.5 db for the used SOA and 20 Fig. 2. Experimental setup

Vol. 25, No. 19 18 Sep 2017 OPTICS EXPRESS 22855 Fig. 3. Modulation depth of input and output signal for various modulated signal bandwidths with respect to SOA input power of (a) + 5 dbm, (b) 5 dbm, (c) 10 dbm, (d) 20 dbm, and (e) 30dBm. db for the used RSOA. Therefore, polarization controller was used before the RSOA. The upstream signals can be modulated with the RSOAs or the external modulator. However, for the pair performance comparisons between the downstream and upstream signals, additional MZM was used for the upstream modulation. For the best modulation efficiency, we used polarization controllers (PCs) before both MZMs. The bandwidth of the photo detector (PD) was 6 GHz with peak responsivity of 0.95 A/W at 1550 nm. In this experiment, we have only demonstrated signal suppression in the case of optical back-to-back transmission with separated fiber branches in order to exclude other channel impairments. DMT signals were generated in an offline process and loaded into the AWG with an oversampling ratio of two. The size of the fast Fourier transform (FFT) for the DMT was 512 with 256 effective subcarriers. Subcarriers were mapped with 4 quadrature amplitude modulation (4-QAM) for the channel performance analysis. The signal bandwidth varied with the sampling rate of the AWG. The received signals were captured using a digital phosphor oscilloscope (DPO, MSO7160C by Tektronix) with a sampling rate of 50 GS/s and evaluated through offline processing. The average bit-error-rate (BER) was obtained over multiple transmissions of 5,000 serial DMT symbols. In the experiment, we used training symbols which contains uniform bits (4 QAM) and power among the entire subcarriers, and evaluate the channel response of each subcarriers. The used equalization process was simple one-tap equalizer of each subcarriers based on the evaluated channel response. Figure 3 shows the signal suppression efficiency of our setup by measuring the modulation depths of the input and output signals of the proposed cascaded SOA/RSOA configuration with respect to SOA input power and modulated signal bandwidth. The RSOA was set to the deeply saturated condition by adjusting the input optical power by 4 dbm with a variable optical attenuator (VOA). The ratio between the modulation depth of input and output signals shows the signal suppression efficiency of the proposed signal suppression technique. Lower slope of the modulation depth ratio means that the output signal suppressed modulation depth of input signal effectively, therefore, the signal suppression efficiency increases. In Fig. 3(a), the first SOA is operating in the gain-saturated region with an input

Vol. 25, No. 19 18 Sep 2017 OPTICS EXPRESS 22856 optical power of 5 dbm. In this condition, the signal suppression efficiency rapidly degrades, showing a steeper slope of modulation depth ratio as the signal bandwidth increases. Because of the limited relaxation time of the SOA, the gain response of the SOA cannot follow the signal amplitude variation. Consequently, the possible signal suppression bandwidth (suppression intensity noise) is very limited. Figures 3(b) 3(e) show the signal suppression in the case of an unsaturated SOA with a gain saturated RSOA. Because the SOA is gain saturated for input optical powers higher than - 5 dbm, we adjusted the input optical power in steps from - 5 dbm to - 30 dbm. Even with the decreasing input optical power of the SOA, the modulation depth ratio of each signal s bandwidth was unchanged. This is because the slope of the modulation depth ratio is determined by the signal suppression effect of the gain-saturated SOA. Therefore, in Figs. 3(b) 3(e), the slope of the modulation depth ratio is only determined by the gain-saturated RSOA. On the other hand, in Fig. 3(a), all the input signal bandwidths show a lower slope than in Figs. 3(b) 3(e), and show better signal suppression efficiency with smaller output modulation depth compared with Fig. 3(b). Proposed technique adjusts the incoming optical power of the SOA to be lower, then the modulated signal become weaker with the OSNR degradations. At the RSOA, the gain-saturated SOA suppressed the addictive noise of SOA with the weak downstream signal because of the nonlinear characteristics of the gain saturated SOA and the gain distribution difference. As a result, the enhanced noise from the SOA and the modulated signal are both suppressed providing a large signal bandwidth and a small dependency on the modulation index of the downstream signal as shown in Figs. 3(c) 3(e). Figure 4 shows the BER performance of the remodulated upstream signal with the proposed downstream signal suppression. The signal bandwidth was set to 4 GHz with DMT modulation leading to a transmission capacity of 8 Gb/s for both upstream and downstream signal transmission. Because the unsuppressed downstream signal causes interference with the upstream signal, leading to reduced transmission performance, the proposed cascaded SOA/RSOA should suppress the downstream signal as strongly as possible. Without downstream signals [the without Remod case in the Figs. 4(a) and 4(b)), the upstream signal performance shows better BER performance by excluding the influence of the modulated downstream signals, and the ASE noise from the SOA decreases as the input optical power to the SOA increases. However, providing high input optical power to the SOA gives poor BER performance when the downstream signal is modulated because of the inefficient suppression of the high-frequency signal components. This penalty can be mitigated by reducing the Fig. 4. BER performance of the remodulated upstream signal versus received optical power with respect to the modulation depth of the downstream signal; (a) 0.9 and (b) 0.5. The used signal bandwidth for both the upstream and downstream signals was 4 GHz with the DMT modulation.

Vol. 25, No. 19 18 Sep 2017 OPTICS EXPRESS 22857 Fig. 5. BER performance of the bidirectional transmission versus ODN loss with respect to the modulation depth of the downstream signal; (a) 0.5 and (b) 0.9. The used signal bandwidth for both the upstream and downstream signals was 4 GHz with the DMT modulation. modulation depth of the downstream signal as shown in Fig. 4(b). In addition, the small optical input power of the SOA suppresses the downstream signal efficiently, with small BER penalties for the remodulated upstream signals which come from the slightly degraded OSNR of the reused optical carrier. Figure 5 shows the BER performance of the bidirectional transmission with regard to the ODN loss by varying the VOA in the experiment. The performance of the downstream transmission was decreasing with the ODN loss increasing because of the receiver sensitivity. The small ODN loss increased the input power of the SOA leading to the inefficient signal Fig. 6. EVM performance of the remodulated upstream signal for each individual subcarrier in the DMT frame. The modulation depth was 0.5 for the downstream signal, and the received optical power was - 6 dbm.

Vol. 25, No. 19 18 Sep 2017 OPTICS EXPRESS 22858 suppression effect in spite of the increasing the BER performance of the downstream signals (Interference dominant). Instead, as the ODN loss increases, the transmission performance shows better performance, and it decreased the transmission performance at above the optimized point (Loss dominant) because of the optical power loss. Therefore, the sufficient ODN loss was required to achieve the best efficiency of the signal suppression and provide the sufficient input optical power for the up/downstream signal detection. In the all transmission lengths of which we demonstrated, the modulation index of 0.5 shown the better BER performance than the modulation index of 0.9 by easy suppression of the incoming modulated signal power. Figure 6 shows the error vector magnitude (EVM) performance of the subcarrier in the DMT frame. The remodulated upstream signal was received with optical power of - 6 dbm, with a modulation depth of 0.5 for the downstream signal. Without a downstream signal, higher input optical power to the SOA shows better EVM performance for upstream transmission by reducing the ASE enhancement. However, the higher optical power into the SOA left the high-frequency components unsuppressed after the cascaded SOR/RSOA, leading to a worse EVM for the upstream signal, especially in high-frequency subcarriers. Instead, by lowering the input optical power of the SOA, the performance penalty due to the remaining downstream signals was minimized in spite of the ASE-induced OSNR degradation. 4. Conclusions We have demonstrated wavelength reused remodulation for reflective PON upstream transmission. The proposed wavelength reuse technique effectively suppresses downstream signals with an unsaturated SOA and a gain-saturated RSOA in cascade configuration. The proposed carrier reuse scheme shows the outstanding performance in terms of signal suppression efficiency regardless of the suppressed signal bandwidth, and has great potential to satisfy the transmission throughput requirements of future optical access networks. In the experimental demonstrations, we achieved a signal suppression bandwidth of 4 GHz with a modulation depth of 0.9 for bidirectional transmission. However, the maximum usable signal bandwidth is potentially much larger than the demonstrated bandwidth of 4 GHz, considering the fact that the proposed signal suppression technique is independent of the signal bandwidth. By reusing the downstream wavelength for the upstream transmission, the proposed signal suppression technique can enhance optical spectral efficiency and solve the problem of lack of bandwidth in future WDM access networks. Funding ICT R&D program of MSIP/IITP, Republic of Korea [2014-0-00538(B0101-17-0131), Nextgeneration coherent optical access physical network]