• 沒有找到結果。

TWDM-PON With Signal Remodulation and Rayleigh Noise Circumvention for NG-PON2

N/A
N/A
Protected

Academic year: 2021

Share "TWDM-PON With Signal Remodulation and Rayleigh Noise Circumvention for NG-PON2"

Copied!
7
0
0

加載中.... (立即查看全文)

全文

(1)

Rayleigh Noise Circumvention for NG-PON2

Volume 5, Number 6, December 2013

C. W. Chow, Senior Member, IEEE

C. H. Yeh, Member, IEEE

K. Xu

J. Y. Sung

H. K. Tsang, Senior Member, IEEE

DOI: 10.1109/JPHOT.2013.2285723

1943-0655 Ó 2013 IEEE

(2)

TWDM-PON With Signal Remodulation and

Rayleigh Noise Circumvention for NG-PON2

C. W. Chow,1Senior Member, IEEE, C. H. Yeh,2;3Member, IEEE, K. Xu,4 J. Y. Sung,1and H. K. Tsang,4 Senior Member, IEEE

1Department of Photonics and Institute of Electro-Optical Engineering, National Chiao Tung University,

Hsinchu 30010, Taiwan

2Information and Communications Research Laboratories, Industrial Technology Research Institute

(ITRI), Hsinchu 31040, Taiwan

3Graduate Institute of Applied Science and Engineering, Fu Jen Catholic University,

New Taipei 24205, Taiwan

4Department of Electronic Engineering, The Chinese University of Hong Kong, Hong Kong

DOI: 10.1109/JPHOT.2013.2285723 1943-0655 Ó 2013 IEEE

Manuscript received August 2, 2013; revised September 22, 2013; accepted October 1, 2013. Date of current version October 23, 2013. This work was supported in part by the National Science Council Taiwan under Contracts 101-2622-E-009-009-CC2, 101-2628-E-009-007-MY3, and NSC-100-2221-E-009-088-MY3; by the Ministry of Education, Taiwan; and by the Hong Kong Research Grants Council under Grant GRF CUHK416710. Corresponding author: C. W. Chow (e-mail: cwchow@ faculty.nctu.edu.tw).

Abstract:We propose and demonstrate a 40-Gb/s (4 10 Gb/s) downstream and 10-Gb/s (4  2.5 Gb/s) upstream time–wavelength-division-multiplexed passive optical network (TWDM-PON) using signal remodulation. Here, a downstream differential-phase-shift-keying (DPSK) signal and an upstream remodulated carrier-suppressed single-sideband non-return-to-zero (CS-SSB-NRZ) signal are used, which are wavelength shifted to circum-vent Rayleigh backscattering (RB). A silicon-based optical microring resonator (MRR) filter is preinstalled at the optical networking unit (ONU) to select the desired downstream wavelength and simultaneously demodulate the downstream DPSK signal, which is then detected by a monolithic-integrated germanium-on-silicon (Ge-Si) photodiode (PD). Using silicon-based devices could be cost effective for the cost-sensitive ONU. The future mono-lithic integration of a silicon filter, a silicon detector, and a silicon modulator in the ONU is also discussed. The characteristics of the preinstalled silicon-based MRR and the Ge-Si PD are discussed. Error-free transmission (bit-error rate (BER) G 109) is achieved in both

downstream and remodulated upstream signals after propagating through 20 km of standard single-mode fiber (SMF).

Index Terms: Optical communications, noise mitigation, wavelength-division-multiplex passive optical network (WDM-PON), differential-phase-shift-keying (DPSK).

1. Introduction

A huge increase in bandwidth demand has been observed in recent years due to the introduction of different broadband applications. The traditional time-division-multiplexed passive optical networks (TDM-PON) could not satisfy this bandwidth demand. Hence, different organizations have pro-posed different network architectures to cope with the bandwidth demand. The Full Service Access Network (FSAN) has started to look for potential future-proof access solutions [1]. It has divided the PON evolution into two stages. The first stage is the next-generation PON (NG-PON1), which includes the XG-PON supporting the maximum downstream and upstream data rates of 10 Gb/s and 2.5 Gb/s, respectively. The second stage is the NG-PON2, which is required to provide a

(3)

rators have proposed different solutions [2]. These solutions include using serial 40G TDM-PON [3], [4], orthogonal frequency division multiplexed (OFDM) PON [5], [6], and time-wavelength-division-multiplexed (TWDM) PON [7], [8]. Among these architectures, TWDM-PON is considered as a primary architecture for NG-PON2, and it can highly reuse the existing optical distribution network (ODN), with less destruction and less capital investment. A state-of-the-art TWDM-PON has been reported recently [9]. However, these proposals [7], [8] need tunable transmitters (Txs) and tunable receivers (Rxs) in the cost-sensitive optical networking units (ONUs).

In this work, we propose and demonstrate a 40 Gb/s (4 10 Gb/s) downstream and 10 Gb/s (4  2.5 Gb/s) upstream TWDM-PON using signal remodulation. Signal remodulation [10]–[12] reuses the downstream signal to generate the upstream signal; hence, only one wavelength is needed for both upstream and downstream signals. This can simplify the wavelength management. Colorless ONU can be implemented using reflective optical modulator. Here, downstream differential-phase-shift-keying (DPSK) signal and upstream remodulated carrier-suppressed single-sideband non-return-to-zero (CS-SSB-NRZ) signal are used. The optical spectrum of the upstream CS-SSB-NRZ signal is wavelength-shifted to circumvent Rayleigh backscattering (RB) [13]. A silicon-based opti-cal micro-ring resonator (MRR) filter is pre-installed at the ONU to select the desired downstream wavelength and simultaneously demodulate the downstream DPSK signal, which is then detected by a monolithic-integrated germanium-on-silicon (Ge-Si) photodiode (PD) [14]. Using silicon-based devices could be cost-effective for the cost-sensitive ONU. Future monolithic integration of silicon filter, silicon detector and silicon modulator in the ONU is also discussed. Part of the downstream DPSK signal is remodulated to produce the upstream CS-SSB-NRZ signal sending to the optical line terminal (OLT). The characteristics of the pre-installed silicon-based MRR and the Ge-Si PD are discussed. Error-free transmission (bit-error-rate (BER)G 109) is achieved in both downstream

and remodulated upstream signals after propagating through 20 km of standard single mode fiber (SMF) without dispersion compensation.

The architecture of the proposed TWDM-PON using signal remodulation is presented in Section 2. The experimental results and analysis for the downstream DPSK demodulation and detection by the monolithic integrated silicon MRR and Ge-Si PD are given in Section 3. The design and fabrication parameters of the silicon devices are also presented in Section 3. The experimental results and analysis of the upstream CS-SSB-NRZ signal with RB circumvention are discussed in Section 4. Finally, a conclusion is given in Section 5.

2. Architecture of the Signal Remodulated TWDM-PON

The system architecture of the proposed signal-remodulated TWDM-PON is shown in Fig. 1. In order to simplify the figure, only one OLT and one ONU for the TWDM-PON are included in the figure. Inside each OLT, four sets of Tx modules at different wavelengths are stacked. A continuous-wave (CW) signal emitted from a distributed feedback laser (DFB) is encoded by a 10 Gb/s, PRBS 223 1

DPSK downstream signal via a phase modulator (PM) in each Tx module. Four different wave-lengths of the DPSK signals are combined by an optical multiplexer to produce the 40 Gb/s down-stream signal. They are then transmitted through 20 km SMF to the ONU.

In each ONU, four sets of Rx modules are used. Inside each Rx module, the downstream DPSK signal is divided by a 3-dB fiber splitter. One half of the downstream signal is launched into a monolithic-integrated silicon-based MRR filter and a Ge-Si PD. The MRR filter can select the desired downstream wavelength and simultaneously demodulate the DPSK signal. Another half of the downstream signal will be selected by a tunable filter (TF) and then launched into a commer-cially available dual-parallel Mach-Zehnder modulator (DP-MZM). A 10 GHz radio-frequency (RF) up-converted 2.5 Gb/s NRZ signal are applied in-phase and quadrature-phase to the upper and lower MZMs of the DP-MZM. By properly adjusting the dc-bias to the DP-MZM, a CS-SSB-NRZ upstream signal can be generated. The optical spectrum of the remodulated upstream NRZ signal is wavelength-shifted by 10 GHz away from the wavelength of the downstream DPSK signal for effective RB circumvention.

(4)

3. Results of Downstream DPSK Demodulation and Detection

In each Rx module in the ONU, there was a monolithic-integrated silicon-based MRR filter and a Ge-Si PD. The scanning electron microscope (SEM) images of the MRR and the Ge-Si PD are shown in Fig. 2(a) and (b), respectively. The MRR filter selects the desired downstream wavelength and simultaneously demodulates the DPSK signal. In the experiment, the downstream DPSK was at the wavelength of 1544.7 nm. The optical signal was coupled to the MRR filter via a tapered waveguide grating coupler with about 5 dB coupling loss, which can be further minimized if apodized grating structure were used [15]. The device was fabricated at the Institute of Micro-electronics (IME). Recently, new foundry service for silicon photonic chips has been established, allowing researchers working on different projects in the world to share the cost of silicon photonic fabrication [16].

The Rx was fabricated on a silicon-on-insulator (SOI) wafer, having the thickness of the top silicon and buried oxide (BOX) of 220 nm and 2 m, respectively. The silicon single-mode wave-guide connecting the grating coupler and the MRR has a dimension of 500 nm 160 nm. For the MRR filter, the coupling length, the gap distance between the waveguide and the MRR, and the MRR radius are 4 m, 200 nm, and 7.5 m, respectively. A thermal heater on the MRR was used for tuning the resonance wavelength. The thermal heater was produced by deposition of a 120 nm thick titanium nitride (TiN) layer on top of the silicon MRR. Fig. 3(a) and (b) shows the measured transmission spectrum of the MRR, with the applied dc-bias voltage from 0 V to 4 V, and the corresponding power consumption, respectively. When the applied dc bias voltage was increased

Fig. 2. SEM images of (a) silicon MRR filter for wavelength selection and DPSK demodulation and (b) Ge-Si PD for downstream detection.

Fig. 1. Proposed signal-remodulated TWDM-PON with RB noise circumvention. DFB: distributed feedback laser, PM: phase modulator, Rx: receiver, SMF: single mode fiber, OLT: optical line terminal, ODN: optical distribution network, ONU: optical networking unit, TF: tunable filter.

(5)

from 0 V to 4V, it produces a wavelength tuning of 8.9 nm, covering about a free-spectral range (FSR) of the MRR. The Q and the extinction ratio of the MRR are3500 and 22 dB, respectively. The MRR demodulated downstream DPSK signal was then directly launched to the Ge-Si PD via a tapered single-mode single waveguide. The Ge region has the thickness, length, and width of 800 nm, 6 m, and 90 m, respectively. After the optical-to-electrical conversion in the PD, the downstream signal was measured by a BER tester (BERT) connected via a signal-ground-signal (SGS) electrical probe. The responsivity and bandwidth of the Ge-Si PD are 0.7 A/W (at bias of 3 V) and 12.5 GHz, respectively. Shrinking the Ge window size to increase the PD bandwidth and using balanced detection will be considered in the next version of the Rx.

Then, BER measurements were performed for the 10 Gb/s DPSK demodulated and Ge-Si PD detected downstream signal at back-to-back (B2B) and after 20 km SMF transmission without any dispersion compensation. A1 dB power penalty was measured at BER of 109, with wide-open

eye-diagram shown in the inset of Fig. 4.

4. Results of Upstream CS-SSB-NRZ Generation With RB Circumvention

The signal-remodulated upstream CS-SSB-NRZ signal was generated by a commercially available DP-MZM (also known as different quadrature-phase-shift-keying (DQPSK) modulator), which was used as the colorless upstream modulator in the ONU. The DP-MZM was electrically driven by a

Fig. 4. Measured BER of the 10 Gb/s DPSK demodulated and Ge-Si PD detected downstream signal. Inset: detected eye-diagram after 20 km SMF transmission.

Fig. 3. (a) Measured transmission spectrum of the MRR with the applied DC-bias voltage from 0 V to 4 V, and (b) the corresponding power consumption.

(6)

10 GHz RF up-converted 2.5 Gb/s NRZ signal at in-phase and quadrature-phase respectively applied to the upper and lower MZMs of the DP-MZM. At the output of the DP-MZM, a 10 GHz CS-SSB-NRZ signal-remodulated optical upstream signal was produced. The original center wavelength at 1544.7 nm was suppressed, producing an upstream NRZ signal at wavelength of 1544.78 nm. The signal-remodulated upstream signal was then sent back to the OLT through the same fiber path and then was measured using a BERT.

Fig. 5 shows the measured BER of the signal-remodulated conventional NRZ signal and the proposed CS-SSB-NRZ signal. These two upstream signals were generated from the remodulation of the downstream DPSK signal, which have propagated through a total of 40 km SMF. We can observe that the conventional NRZ signal (without wavelength-shifting) was highly degraded by the RB at the OLT Rx, with an error-floor at BER of 102. The proposed signal-remodulated CS-SSB-NRZ signal can significantly circumvent the RB noise, and error-free transmission (BER G 109)

was successfully achieved. In a typical signal remodulation network, the RB generated by the downstream signal and the remodulated upstream signal are propagating in the same direction toward the Rx at OLT. The RB and the upstream signal will produce the interferometric beat noise, which will be at the baseband and fall within the Rx bandwidth, degrading the upstream signal. In the proposed scheme, the upstream signal is wavelength shifted by 10 GHz; hence, the beat noise will be shifted and will fall outside the Rx bandwidth (the Rx 3-dB bandwidth is 2.5 GHz). As a result, the RB can be mitigated.

In practical case, the upstream signal is operated in burst mode (transmitting the TDM optical packets); hence, the Rx at OLT should be gated to only measure the payload of the optical packets. As the wavelength of the upstream signal is shifted to mitigate the RB, we believe that the RB mitigation efficiency should be similar to that shown in Fig. 5 (in continuous mode). It is also worth pointing out that although the lithium niobateðLiNbO3Þ based DP-MZM was used in the

proof-of-concept demonstration, it is also believed that this modulator can be fabricated in SOI platform [17]. Hence, the MRR, PD, and the modulator can be monolithic integrated in the ONU to reduce cost.

5. Conclusion

TWDM-PON is a cost-effective and primary solution for the NG-PON2. It is also particularly desirable for the brown-field deployment, in which the conventional TDM-PON has already been built. In this work, we proposed and demonstrated a TWDM-PON using signal remodulation. The downstream DPSK signal was divided into two parts in the ONU; one part of the downstream DPSK signal was remodulated to produce the upstream CS-SSB-NRZ signal sending to the OLT. The optical spectrum of the upstream CS-SSB-NRZ signal was wavelength-shifted by 10 GHz to cir-cumvent RB. The experimental results showed that the proposed signal-remodulated CS-SSB-NRZ

Fig. 5. Measured BER of the signal-remodulated conventional NRZ signal and the proposed CS-SSB-NRZ signal, showing RB noise can be successfully circumvented. Inset: corresponding eye-diagrams.

(7)

G 10

the conventional signal-remodulated NRZ had an error floor at BER of 102. Another part of the downstream DPSK signal was demodulated by a silicon-based optical MRR filter, which had the Q and extinction ratio of3500 and 22 dB, respectively. A TiN thermal heater was deposited on top of the silicon MRR for wavelength tuning. By applying a dc-bias voltage from 0 V to 4 V, the wave-length tuning can cover the FSR of the MRR (8.9 nm), with power consumptionG 60 mW. Then, the demodulated DPSK signal was detected by a monolithic-integrated Ge-Si PD, which had the responsivity and bandwidth of 0.7 A/W (at bias of3 V) and 12.5 GHz, respectively.

References

[1] R. W. Heron and E. Harstead,BFSAN NG-PON2 updates,[ presented at the OFC Conf., Anaheim, CA, USA, 2013, Paper OW4D.5.

[2] P. Vetter,BNext generation optical access technologies,[ presented at the ECOC Conf., Amsterdam, The Netherlands, 2012, Paper Tu.3.G.1.

[3] D. Van Veen, V. E. Houtsma, P. Winzer, and P. Vetter,B26-Gbps PON transmission over 40-km using duobinary detection with a low cost 7-GHz APD-based receiver,[ presented at the ECOC Conf., Amsterdam, The Netherlands, Paper Tu.3.B.1.

[4] C. W. Chow and C. H. Yeh,B40-Gb/s downstream DPSK and 40-Gb/s upstream OOK signal remodulation PON using reduced modulation index,[ Opt. Exp., vol. 18, no. 25, pp. 26 046–26 051, Dec. 2010.

[5] N. Cvijetic,BOFDM in optical access networks,[ presented at the OFC Conf., Los Angeles, CA, USA, 2011, Paper OMG3.

[6] C. W. Chow, C. H. Yeh, C. H. Wang, C. L. Wu, S. Chi, and C. Lin,BStudies of OFDM signal for broadband optical access networks,[ IEEE J. Sel. Areas Commun., vol. 28, no. 6, pp. 800–807, Aug. 2010.

[7] Y. Ma, Y. Qian, G. Peng, X. Zhou, X. Wang, J. Yu, Y. Luo, X. Yan, and F. Effenberger,BDemonstration of a 40Gb/s time and wavelength division multiplexed passive optical network prototype system,[ presented at the OFC Conf., Los Angeles, CA, USA, 2012, Paper PDP5D.7.

[8] Z. Li, L. Yi, M. Bi, J. Li, H. He, X. Yang, and W. Hu,BExperimental demonstration of a symmetric 40-Gb/s TWDM-PON,[ presented at the OFC Conf., Anaheim, CA, USA, 2013, Paper NTh4F.3.

[9] Y. Luo, X. Zhou, F. Effenberger, X. Yan, G. Peng, Y. Qian, and Y. Ma,BTime- and wavelength-division multiplexed passive optical network (TWDM-PON) for next-generation PON stage 2 (NG-PON2),[ J. Lightwave Technol., vol. 31, no. 4, pp. 587–593, Feb. 2013.

[10] M. Bi, S. Xiao, H. He, L. Yi, Z. Li, J. Li, X. Yang, and W. Hu,BSimultaneous DPSK demodulation and chirp management using delay interferometer in symmetric 40-Gb/s capability TWDM-PON system,[ Opt. Exp., vol. 21, no. 14, pp. 16 528– 16 535, Jul. 2013.

[11] R. Wang, S. Fu, P. P. Shum, and C. Lin, B10 Gb/s WDM-PON using downstream PolSK coded by polarization modulator and upstream intensity re-modulation,[ Electron. Lett., vol. 46, no. 6, pp. 428–430, Mar. 2010.

[12] C. W. Chow, C. H. Wang, C. H. Yeh, and S. Chi,BAnalysis of the carrier-suppressed single-sideband modulators used to mitigate Rayleigh backscattering in carrier-distributed PON,[ Opt. Exp., vol. 19, no. 11, pp. 10 973–10 978, May 2011.

[13] C. W. Chow and C. H. Yeh, BUsing downstream DPSK and upstream wavelength-shifted ASK for Rayleigh backscattering mitigation in TDM-PON to WDM-PON migration scheme,[ IEEE Photon. J., vol. 5, no. 2, p. 7900407, Apr. 2013.

[14] K. Xu, L. Wang, G. K. P. Lei, Z. Cheng, Y. Chen, C. Y. Wong, C. Shu, and H. K. Tsang,BDemodulation of 20 Gbaud/s differential quadrature phase-shift keying signals using wavelength-tunable silicon microring resonators,[ Opt. Lett., vol. 37, no. 16, pp. 3462–3464, Aug. 2012.

[15] X. Chen, C. Li, C. K. Y. Fung, S. M. G. Lo, and H. K. Tsang,BApodized waveguide grating couplers for efficient coupling to optical fibers,[ IEEE Photon. Technol. Lett., vol. 22, no. 15, pp. 1156–1158, Aug. 2010.

[16]BFoundry service aids silicon photonic chips,[ in SPIE Professional Magazine, Jul. 2011.

[17] K. Xu, L. G. Yang, J. Y. Sung, Y. M. Chen, Z. Z. Cheng, C. W. Chow, C. H. Yeh, and H. K. Tsang,BCompatibility of silicon Mach-Zehnder modulators for advanced modulation formats,[ J. Lightwave Technol., vol. 31, no. 15, pp. 2550–2554, Aug. 2013.

數據

Fig. 2. SEM images of (a) silicon MRR filter for wavelength selection and DPSK demodulation and (b) Ge-Si PD for downstream detection.
Fig. 4. Measured BER of the 10 Gb/s DPSK demodulated and Ge-Si PD detected downstream signal
Fig. 5 shows the measured BER of the signal-remodulated conventional NRZ signal and the proposed CS-SSB-NRZ signal

參考文獻

相關文件

Biases in Pricing Continuously Monitored Options with Monte Carlo (continued).. • If all of the sampled prices are below the barrier, this sample path pays max(S(t n ) −

The AR part is designated for the desired adjustment to the target; the exogenous part is corresponding the offset between the center of target and the aiming point wherein

For MIMO-OFDM systems, the objective of the existing power control strategies is maximization of the signal to interference and noise ratio (SINR) or minimization of the bit

This objective of this research is to develop water based sol-gel process to apply a protective coating on both optical molds and glass performs, which can effectively prevent glass

The peak detector and loop filter form a feedback circuit that monitors the peak amplitude, A out, of the output signal V out and adjusts the VGA gain until the measured

The advantages of using photonic crystal fibers (PCFs) as gas sensors include large overlap and long optical path interaction between the gas and light mode field and only a

C., “Robust and Efficient Algorithm for Optical Flow Computation,” Proceeding of IEEE International Conference on Computer Vision, pp. “Determining Optical Flow.” Artificial

Its principle is to use Gyrator structure such that active MOSFET can equivalent to inductor ,along with parasitic capacitor to form a resonator tank for an oscillator.. The second