Open Access
Add to BibSonomyAbstract
The backscattering noises introduced by Rayleigh and stimulated Brillouin scattering have been experimentally studied by means of their spectrum broadening, the scattering power variation and their impacts on upstream signals with different transmission fiber lengths and incident powers in a single-fiber bidirectional passive optical network (PON) communication system. The results show that both spontaneous scattering and simulated scattering can take place. The power and spectrum of backscattering noises are determined by the downstream launch power, laser linewidth and transmission fiber length. With the transmission length increasing, the power of backscattering noises gets higher, the spectrum of the backscattering noise broadens and the simulated threshold power decreases. The backscattering noise can beat with uplink light to modulate envelop of upstream signal resulting in degradation of BER greatly. Under the condition of one single channel for the second next generation PON (NG-PON2), the fiber length is 40km and downstream launch power is up to 11dBm. At this time, the backscattering noises are easy to be stimulated and the scattering power rises up from −20dBm to 10dBm, which can overwhelm the US signal. The spectrum of the optical beat interference noise also rises up with fiber length, which causes the uplink’s BER degradation. The experimental results are significant for mitigation of backscattering noises under the condition of bidirectional PONs.
© 2015 Optical Society of America
1. Introduction
It has been widely accepted that the Rayleigh backscattering (RB) noise is a critical issue that can severely degrade the performance of the upstream (US) signal in the full-duplex bidirectional single fiber system in which the upstream and downstream (DS) signals are using the same wavelength [1–3]. With the great increase of the demand of bandwidth and long distance transmission, the next second generation passive optical network (NG-PON2) [4–6] is regarded as a potential technology, in which a high splitting ratio (higher than 1:64) and up to 40km transmission distance are needed. So the launch power up to 11dBm is much higher than standard PON systems such as Gigabit-capable passive optical networks (GPON) [6, 7], therefore the stimulated Brillouin scattering (SBS) should be taken into account [2]. For the purpose of mitigating the backscattering noises, it is necessary to make clear the characteristics of those noise and their impacts on the upstream link.
It has been widely believed that the power spectrum of RB noise concentrates on DC (direct current) and low frequency region and the spectrum does not change with the fiber length [1–3]. Some researchers proposed several models to represent the mechanism of the RB noise [8–13]. P. Gysel and R. K. Staubli analyzed the characteristics of RB based on the statistical properties of the laser and single mode fiber [9, 15]. They got an equation to describe the spectrum of the RB noise, in which the bandwidth of the noise depended on the linewidth of laser only. When the laser linewidth is 60MHz and the bit rate is 140Mb/s, the bandwidth of scattering noise is about 100MHz after 3.3km single mode fiber transmission. In contrast, they provided a 100MHz dithering signal to the laser bias. Then the laser linewidth increased and the bandwidth of scattering noise expanded to more than 200MHz [9].
But under the circumstance of one single Time and Wavelength Division Multiplexed (TWDM) PON channel for NG-PON2 [4–6], we need to research characteristics of backscattering noises in more details. There are some differences between the TWDM-PON channel for NG-PON2 and the system of P. Gysel et al [9]. Firstly, the linewidth of the light source in NG-PON2 is much less than 60MHz. Secondly, the transmission distance is much longer than 3.3km and the launch power of DS signal is very high. Finally, the bit rate of signal can be up to 10Gb/s [6].
Recently, some researchers have investigated the characteristics of the RB with narrow linewidth lasers and interference detection technique [16–18]. T. Zhu et al. [16] found evidences that the stimulated Rayleigh scattering exists in 10km standard single mode fiber (SSMF) and other types of fiber. And in the PON system, Q. Guo et al. realized the influences of SBS on US signal [2]. Their works provided more evidences about the mechanism and characteristics of RB and SBS in fiber. In the experiment of T. Zhu et al [16], the fiber length is much shorter than the transmission distance of NG-PON2. And in the experiment of Q. Guo et al [2], they fixed the DS launch power at 3dBm to avoid SBS. But in a NG-PON2 system, the DS launch power is up to 11dBm [6], thus the influence of SBS cannot be ignored.
At the same time, when two lights transmit in opposite directions by single fiber, the backscattering light of one light can beat with the other light. The impacts of the phenomenon are not been investigated sufficiently. Especially, it is unclear how the performance of PON systems change with the backscattering light and beat noises.
Based on all the above, we experimentally investigate the properties of the RB and SBS of DS light and their impacts on the US signal varying from the transmission fiber length and the incident power. We investigate the backscattering noises under the condition of NG-PON2. In our experiments, the US light and DS light are at the same wavelength, which conforms to colorless ONUs required by NG-PON2 [4–6]. The incident power is very high and the fiber length is up to 50km. It is also the case for TWDM-PON standard for NG-PON2, where a maximum power per TWDM-PON channel equals to 11dBm and a maximum fiber length equals to 40km [4–6]. So our experimental results are beneficial for backscattering noises mitigation in NG-PON2, especially when the reach extender is needed [4–6]. We discover that both spontaneous and simulated backscattering can take place in the bidirectional PON system. Especially, when the DS launch power is high enough, the stimulated processes including SBS and simulated RB are motivated.
The rest of this paper is organized as follows. In the Section 2, we introduce the theory for backscattering in PON systems. The Section 3 is the experiment setup and results analysis and the Section 4 is the conclusion and some suggestions.
2. Theoretical introduction
Figure 1 is a schematic diagram of one single TWDM-PON channel for NG-PON2. In PON systems, the incident power of DS signal depends on the transmission distance and the splitting ratio. With a high splitting ratio and a long transmission distance, the incident power must be high enough for recovering the signal accurately after power distribution. According to the standard [6], the typical splitting ratio is 1:64 (up to 1:256) and the transmission reach is 40km in a TWDM-PON system for NG-PON2. So the incident power should be about 11dBm in SSMF [4, 6]. For the spontaneous backscattering, the scattering power is 31~33dB lower than launch power [10–12] and can be about −22dBm to −20dBm. At the same time, the received power of US at the optical line terminal (OLT) is around −10dBm to −20dBm or even lower [2, 6]. So the scattering power is comparable to the received power of US. Besides, the DS incident power is so high that Brillouin scattering can be converted into SBS [19, 20]. The RB always exists in fiber, so the received light intensity at OLT in Fig. 1 can be written as:
In Eq. (1), the denotes the intensity of US signal, and the denotes the intensity of RB and SBS in a backward direction which will lead to SNR degradation of US signal. The RB and SBS lights can be modelled as many small reflections distributed over the fiber [22]. And the summation symbols represent the total intensity of RB and SBS as sum of contribution over the fiber length. The is the interaction term of signal and backscattered light which results in optical beat interference (OBI) noise. The is the interaction term of RB and SBS, which can be regarded as intensity noise. The intensity of RB, SBS and cross terms of them are additive noises, but the interaction terms of signal and backscattering light are multiplicative noises. In order to suppress scattering noises, it is important to investigate its characteristics and influences on signals.
In WDM-PON systems with colorless ONUs, the light sources are centralized in the OLT, so the backscattering light and the US light have the same wavelength and transmit at the same direction. Therefore, the backscattering light can beat with the US light. The cross terms achieve the maximum when all of the light fields have the same polarizations and phases. Especially, when the backscattering is stimulated, the intensity of backscattering light is comparable to that of the US light or even larger. The backscattering noises can interfere with the US signal to make envelop of the US signal be modulated. So the impact of multiplicative noise should become very critical.
From investigation of P. Gysel et al. [9], the power spectrum of the RB noise can be expressed as Eq. (2).
In Eq. (2), is the intensity of the power of RB noise, is the Dirac function and is the linewidth of the laser. In Eq. (2), the spectrum of the RB noise depends only on the linewidth of lasers.So we will investigate the spectra of the backscattering noise under different transmission distances, DS launch powers and bit rates of DS signal. We will also discuss how the scattering power and spectrum change with these factors and provide experimental foundation to mitigate the backscattering noise in the PON system.
3. Experiment and analysis
3.1 Characteristics of scattering noise
Figure 2 shows the schematic diagram of our experimental setup. The light from a 200kHz linewidth laser is optically modulated by a Mach-Zehnder modulator (MZM) and the bit rates of modulated signal are 1.25Gb/s, 2.5Gb/s and 5Gb/s respectively. Since we don’t have pseudo-random binary sequence generator, the electrical signals are generated by an arbitrary waveform generator (AWG) working at 25GS/s sampling rate with 8-bit resolution. And the pseudo random binary sequence (PRBS) lengths are 217-1, which are used to create the 1.25Gb/s, 2.5Gb/s and 5Gb/s signals. The modulated signal is fed into fiber by an optical circulator. The variable optical attenuator (VOA) and EDFA are used to adjust the power of the DS incident light. The backscattering light of modulated signal is detected by a photodetector (PD) after the circulator. The output signal of PD is then sent to a spectrum analyzer where we measure the spectrum. After the remote terminal of fiber, there should be power splitter and ONUs in actual PON systems. But the backscattering mainly occurs in the trunk fiber line from OLT to the remote terminal and we focus on the power and spectrum of scattering light, so the part in the dashed box of Fig. 2 is omitted in our experimental setup. Different from the experiment of Zhu et al [16], we investigate the intensity of backscattering light and the scattering spectra from different transmission fiber lengths with various DS incident power levels.
When investigating the intensity characteristics of backscattering light, we replace the PD and the spectral analyzer with a power meter to test the power of backscattering light with the power of DS incident light changing. Figure 3 shows the measurement of backscattering light power versus the incident light power.
From Fig. 3, we can get three conclusions. Firstly, with the fixed linewidth of the laser, the backscattering power changes with the DS incident power and the fiber length. Every curve can be divided into two parts: one corresponds to the spontaneous backscattering (with the power of −30~-20dBm) and the other corresponds to the stimulated backscattering (with the power of more than −20dBm). The two processes are separated by a threshold which depends on the transmission fiber length. The power of US signal at OLT is about −20dBm~-10dBm [6].When the DS launch power is lower than the threshold, the backscattering power increases linearly with the launch power and is much lower than that of US signal. The spontaneous backscattering noise can lead to SNR degradation of US signal. But when the DS launch power is higher than the threshold, the backscattering power increases sharply from −20dBm to 10dBm with incident power changing about 4dB for stimulated scattering, which consists of SBS [21] and stimulated Rayleigh backscattering (SRBS) [16]. For the case of the NG-PON2 system, the DS launch power is up to 11dBm and the fiber length is 40km. In this time, the intensity of backscattering light at OLT can be about 5dBm, which is high enough to overwhelm the US signal.
In order to reduce the power of backscattering noises, we can select proper launch power of DS signal. The launch power should below the threshold which is determined by the transmission distance of the PON system. For instance, the threshold of 50km, 40km, 10km and 2km fiber are about 6dBm, 7dBm, 11dBm and 15dBm, respectively. So the launch power of DS signal should be less than 6dBm, 7dBm, 11dBm and 15dBm, respectively.
Then, we measure the spectra of scattering light with different DS incident power and fiber lengths. The central frequency of SBS is about 10.8GHz lower than that of pump light in SSMF. And the spectrum of SBS is stable which has been studied in detail [21]. So we concentrate on the spectra of RB. In order to measure the spectra of RB only, we fix the incident power at 3dBm for suppressing SBS. With the incident power fixed at 3dBm and the bit rate fixed at 5Gb/s, we measure the backscattering spectra with different fiber lengths. As shown in Fig. 4, the bandwidth of backscattering noise increases with fiber length.
As shown in Fig. 4, the bandwidths of the backscattering noises change with the fiber length. The bandwidths of the RB noise for 10km, 20km, 40km and 50km fiber are about 39MHz, 72MHz, 110MHz and 128MHz, respectively. The bandwidth of RB noise in 40km system is about twice wider than that in 20km fiber system. So the bandwidth of the backscattering noise approximatively increases with fiber length linearly.
From Fig. 3 and Fig. 4 when the transmission distance is longer than 20km and DS launch power is higher than 6dBm, both the power and bandwidth of backscattering noise are critical for US signal. We should mitigate the backscattering noises, especially the SBS noise.
3.2 Influence of scattering noise on the uplink signal
We employ experiment setting shown in Fig. 5 to evaluate the influence of backscattering light on US signal. The output signal from a laser is split into two branches by using a 3dB optical coupler. The lower output is fed into fiber after an EDFA, a VOA and an optical circulator, which simulates the DS signal in bidirectional PON. The VOA and EDFA are used to adjust the power of incident light. The upper output is modulated into OOK signal with bit rates of 1.25Gb/s, 2.5Gb/s and 5Gb/s respectively. The electrical signal is generated by an AWG working at 25GS/s sampling rate with 8-bit resolution. And the PRBS lengths of the 1.25Gb/s, 2.5Gb/s and 5Gb/s signals are all 217-1. The upper output simulates the US signal in PON and is fed into fiber by an optical circulator at the other end of fiber. The incident power of US signal is fixed at 0dBm. The US signal and backscattering light of DS signal are detected by a PD after the circulator. The output signal of PD is then sent to a digital oscilloscope. After sampling and decision, we can get the BER. The VOA-2 before PD is used to adjust the received power to measure the BER as a function of received power. We investigate the BER of uplink from different fiber lengths for various DS incident power levels. We focus on the performance of uplink, so the power splitter and receivers in downlink are omitted in the experiment setup.
Figure 6 shows the curves of BER vs. received power with fiber lengths 0km, 10km, 20km, 40km and 50km respectively, when DS incident power is fixed at 3dBm. For the DS launch power much less than the threshold of fiber, the backscattering in the fiber is mainly spontaneous and can be regarded as additive noise. At the same incident power, the longer fiber has higher backscattering noise. Therefore the performance of the system degrades with fiber length, such as at BER of 10−6, the 40km system needs about 7dB higher received power, compared with B2B system.
Figure 7(a) is the oscillograph of received US signal by PD with 50km-length fiber and Fig. 7(b) is the oscillograph of a CW light without being modulated the data in 50km fiber when DS launch power is fixed at 6dBm. For 50km fiber, if the DS launch power higher than 6dBm, the power of scattering noises increases sharply. The US signal can be engulfed by backscattering noise and we cannot recover the US data. As shown in Fig. 7, the envelope of OOK signal is not constant but is modulated by the optical beat interference (OBI) noises. It verifies our assumption that the OBI noise is a kind of multiplicative noise.
When the launch power approximates to the threshold of stimulated scattering, the scattering power is approximate to US signal. Therefore the intensity of cross terms are comparable to that of US light and can induce OBI noises. The backscattering light beats with US light to degrade the SNR of US signal.
Changing the length of fiber, we observed the modulated frequency of signal envelop increases with fiber length. Figure 8 shows how the bandwidths of the modulated frequency verify from the fiber length.
As shown in Fig. 8, for 10km fiber, the bandwidth is about 150 kHz, but for 50km fiber, the bandwidth is about 350 kHz. The bandwidths of the modulated frequency similarly increase with fiber length linearly. It indicates that the stimulated backscattering noises can modulate the US signal’s envelop and the bandwidth of modulating signal broadens with fiber length.
Moreover, the BER of US signal is observably affected by the DS launch power and fiber length, when the DS launch power verges on the threshold of stimulated scattering. The measurement is shown in Fig. 9 and Fig. 10.
As shown in Fig. 9, keeping the fiber length fixed at 20km, we measure the BER of uplink as a function of received power for DS incident power being 6dBm, 7dBm and 8dBm, respectively. From the curves, the performance of uplink degrades with the DS incident power. With the received power fixed at −10dBm, when DS incident power is 6dBm, the BER is less than 10−5; but when the DS incident power is 7dBm, the BER is more than 10−3. Especially, when the DS incident power is larger than the threshold of SBS, the US signal is absolutely engulfed by the scattering light.
Figure 10 is the curves of BER vs. received power for fiber lengths that are 0km, 10km, 20km, 40km and 50km respectively, when DS incident power is fixed at 6dBm. With the fiber length increasing, the performance of system degrades. When DS incident power is fixed at 3dBm and the BER is 10−3, the 40km system needs about 5dB higher received sensitivity compared with 10km system. When DS incident power is 6dBm, the 40km system needs over 10dB higher received sensitivity, because the DS incident power is close to the threshold of stimulated backscattering. For DS incident power approximating to stimulated threshold, the performances of 20km 40km, and 50km system degrade obviously.
With the experiments, we investigate the characteristics of backscattering noises of DS light and their influences on US signals. The intensity of RB and SBS (), the cross terms of RB and SBS () can be regard as intensity noise. For the spontaneous process, the intensity increases approximatively linearly, but for the stimulated process, the intensity increases sharply with DS launch power. The fiber length also has influence on the backscattering power. Especially, when the fiber is long enough and DS launch power is higher enough, the cross terms of US signal and DS backscattering noises can beat with each other to modulate the US signal and degrade the performance of uplink.
4. Conclusion
In this paper, we investigate the intensity and spectrum characteristics of backscattered noises and their influence in a bidirectional PON system. In low incident power, the intensity of backscattering light increases linearly with incident power, but when incident power is higher than the threshold, the scattering power increases sharply from −20dBm to 10dBm with incident power changing about only 4dB. And the threshold of the stimulated backscattering is relative to fiber length. With the increasing of the transmission length, the threshold is decreased. So in long haul and high splitter ratio PON systems, the backscattering noise’s influences are much bigger than that in short distant and low splitter ratio systems. Under the NG-PON2 circumstance, since the DS launch power is up to 11dBm and the fiber length is 40km, the scattering power is about 5dBm. It is necessary to mitigate the scattering noise for US signal recovering. Besides, we observe the spectrum bandwidth of backscattering noise increases with fiber length and incident power.
From our experiments, we can give some suggestions for mitigating the influence of backscattering noises. Firstly, for the threshold of stimulated scatterings being determined by fiber length, we should select appropriate the downstream incident power. Avoiding US signal being overwhelmed, the DS incident power should lower than the threshold of SBS. Secondly, we can improve the threshold of the stimulated backscattering, such as modulating the phase of DS carrier with a sine wave of 200MHz~400MHz. Finally, it is beneficial for suppressing beat noise to break the coherent condition and reduce the spectral overlap between US signal and backscattering light. Most of the existing solutions for RB mitigation are by means of spectrum shifting to reduce the spectrum overlap. Our results are helpful for selecting optimized frequency range of spectrum shifting according to fiber lengths and incident power.
Acknowledgments
This work is supported by the China National Science Foundation Project (under granted: 6177063) and the Open Foundation of State Key Laboratory of Optical Communication Technologies and Networks (Wuhan Research Institute of Posts & Telecommunications).
References and links
1. 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). [CrossRef] [PubMed]
2. Q. Guo and A. V. Tran, “Mitigation of Rayleigh noise and dispersion in REAM-based WDM-PON using spectrum-shaping codes,” Opt. Express 20(26), B452–B461 (2012). [CrossRef] [PubMed]
3. R. Matsumoto, T. Kodama, S. Shimizu, R. Nomura, K. Omichi, N. Wada, and K. Kitayama, “40G-OCDMA-PON System With an Asymmetric Structure Using a Single Multi-Port and Sampled SSFBG Encoder/Decoders,” J. Lightwave Technol. 32(6), 1132–1143 (2014). [CrossRef]
4. D. Nesset, “NG-PON2 technology and standards [Invited],” J. Lightwave Technol. 33(5), 1136–1143 (2015). [CrossRef]
5. Y. Luo, X. Zhou, F. Effenberger, X. Yan, G. Peng, Y. Qian, and Y. Ma, “Time and wavelength division multiplexed passive optical network (TWDM-PON) for next generation PON stage 2 (NG-PON2),” J. Lightwave Technol. 31(4), 587–593 (2013). [CrossRef]
6. ITU-T G.989.1, 40-Gigabit-capable passive optical networks (NG-PON2): General requirements [S], (2013)
7. ITU-T G.984.1, Gigabit-capable passive optical networks (GPON): General characteristics [S], (2008).
8. S. Wu, A. Yariv, H. Blauvelt, and N. Kwong, “Theoretical and experimental investigation of conversion of phase noise to intensity noise by Rayleigh scattering in optical fibers,” Appl. Phys. Lett. 59(10), 1156–1158 (1991). [CrossRef]
9. P. Gysel and R. K. Staubli, “Spectral Properties of Rayleigh Backscattered Light from Single-Mode Fibers Caused by a Modulated Probe Signal,” J. Lightwave Technol. 8(12), 1792–1798 (1990). [CrossRef]
10. J. M. Mackintosh and B. Culshaw, “Analysis and observation of coupling ratio dependence of Rayleigh backscattering noise in a fiber optic gyroscope,” J. Lightwave Technol. 7(9), 1323–1328 (1989). [CrossRef]
11. Y. Takushima and K. Kikuchi, “Observation of spectral gain hole burning and modulation instability in a Brillouin fiber amplifier,” in Proceedings of Lasers and Electro-Optics Society 1999 12th Annual Meeting(LEOS 99.)1,58 (1999). [CrossRef]
12. T. H. Wood, R. A. Linke, B. L. Kasper, and E. C. Carr, “Observation of Coherent Rayleigh Noise in Single-Source Bidirectional Optical Fiber Systems,” J. Lightwave Technol. 6(2), 346–352 (1988). [CrossRef]
13. W. K. Burns and R. P. Moeller, “Rayleigh Backscattering in a Fiber Gyroscope with Limited Coherence Sources,” J. Lightwave Technol. 1(2), 381–386 (1983). [CrossRef]
14. L. Dong, “Stimulated thermal Rayleigh scattering in optical fibers,” Opt. Express 21(3), 2642–2656 (2013). [CrossRef] [PubMed]
15. R. K. Staubli and P. Gysel, “Crosstalk Penalties Due to Coherent Rayleigh Noise in Bidirectional Optical Communication Systems,” J. Lightwave Technol. 9(3), 375–380 (1991). [CrossRef]
16. T. Zhu, X. Bao, L. Chen, H. Liang, and Y. Dong, “Experimental study on stimulated Rayleigh scattering in optical fibers,” Opt. Express 18(22), 22958–22963 (2010). [CrossRef] [PubMed]
17. O. Okusaga, J. Cahill, A. Docherty, W. Zhou, and C. R. Menyuk, “Guided entropy mode Rayleigh scattering in optical fibers,” Opt. Lett. 37(4), 683–685 (2012). [CrossRef] [PubMed]
18. O. Okusaga, J. P. Cahill, A. Docherty, C. R. Menyuk, and W. Zhou, “Spontaneous inelastic Rayleigh scattering in optical fibers,” Opt. Lett. 38(4), 549–551 (2013). [CrossRef] [PubMed]
19. H. W. Hu and H. Anis, “Degradation of Bi-Directional Single Fiber Transmission in WDM-PON Due to Beat Noise,” J. Lightwave Technol. 26(8), 870–881 (2008). [CrossRef]
20. M. Ferrario, L. Marazzi, P. Boffi, A. Righetti, and M. Martinelli, “Impact of Rayleigh backscattering on Stimulated Brillouin Scattering threshold evaluation for 10 Gb/s NRZ-OOK signals,” Opt. Express 17(20), 18110–18115 (2009). [CrossRef] [PubMed]
21. A. Kobyakov, M. Sauer, and D. Chowdhury, “Stimulated Brillouin scattering in optical fibers,” Adv. Opt. Photon. 2(1), 1–59 (2010). [CrossRef]
22. E. Brinkmeyer, “Forward-backward transmission in birefringent single-mode fibers: interpretation of polarization-sensitive measurements,” Opt. Lett. 6(11), 575–577 (1981). [CrossRef] [PubMed]
