We propose a novel wavelength-division-multiplexed passive optical network (WDM-PON) architecture with enhanced tolerance toward chromatic dispersion where a DPSK-modulated downstream signal with constant intensity is remodulated at the ONU side with a return to zero (RZ-DPSK). Driving the downstream modulator with a 50% RZ data enabled us to employ the pulse carver at the ONU for both removing downstream data and generating the optical RZ signal for upstream. This offers an attractive alternative to earlier proposed schemes as it allows us to use full modulation depth (FMD) and balanced detection for downstream data restoration. We experimentally demonstrate the system with both balanced and single-ended detection at 2.5Gb/s. Error-free operation has been achieved along a 20Km single mode fiber without dispersion compensation.
©2011 Optical Society of America
WDM PONs offer a great solution to satisfy the increasing demand of bandwidth, in addition to offering a higher level of data security through the virtual point to point connections . A great challenge for producing a cost-effective WDM PON is the need for a transmitter at each Optical Network Unit (ONU) with a dedicated wavelength, which overloads the total cost of the system, in addition to reducing the number of available wavelengths in the system. To reduce these costs, a centralized light source at the Optical Line Terminal (OLT) was introduced; hence the ONU uses the same downstream signal as an upstream data carrier by re-modulating the received light , having each ONU being served by a single wavelength allows to increase network capacity and simplifies network management.
Several remodulation architectures have been proposed, including the use of on-off keying (OOK) for both downstream and upstream ; downstream low extinction-ratio OOK and upstream Differential phase-shift keying (DPSK) ; downstream DPSK and upstream OOK , However, these approaches have several disadvantages, such as high chirp , , limited speed - , and reduced extinction ratio (ER) . Another system has been recently proposed using a reduced modulation depth (RMD) DPSK for downstream signal and a full modulation depth (FMD) DPSK for upstream signal , and this approach has a limitation on the downstream transmission of both the RMD and that balanced detection can't be used in it, the former limitation makes it less tolerant toward phase errors in the DI , while the latter makes it lose the DPSK + 3dB advantage over OOK for downstream link that comes only with balanced detection .
In this paper, we propose a novel wavelength remodulation scheme for WDM-PON using DPSK with full modulation depth for both downstream and upstream signal. We demonstrate the system at 2.5 Gb/s in both directions due to available equipments we currently have, and because the system is robust toward dispersion as we show in this paper, the same principle could be applied to higher speed systems.
Driving the phase modulator with return-to-zero data to produce the downstream signal enabled us to remove all downstream phase information in the ONU using a pulse carver, such that we could write the upstream phase information on the signal coming from the OLT with no phase information, the system has also the advantage of using either single ended or balanced detection unlike prior remodulation schemes , .
2. Operation principle
The operation principle of the proposed system is illustrated in Fig. 1 . In the optical line terminal (OLT) we modulate the light of a laser source using a phase modulator driven by a return-to-zero pre-coded data with a duty cycle of 50%, and this causes writing phase information on the first half of each bit leaving the second half without phase information.
At the ONU side a portion of the downstream signal is detected using a DPSK receiver which restores downstream data in RZ format, and the rest of the signal is passed through a pulse carver which is a device usually implemented as sinusoidally driven Electro Absorption Modulators (EAMs) or Mach-Zehnder Modulators (MZMs) such that it removes the phase-modulated slots of the downstream signal leaving only the un-modulated slots. After that the signal that have no phase information anymore is passed through another phase modulator in the ONU driven by an NRZ signal which is differentially pre-coded using a DPSK pre-coder.The upstream signal exits the ONU as a standard RZ-DPSK that can be transmitted up the link to the OLT to be detected using any DPSK receiver.
3. Experiment and results
A schematic diagram of the experimental setup for the proposed remodulation system is shown in Fig. 1. At the OLT, the downstream 2.5Gb/s DPSK signal was generated by feeding a continuous-wave light source at 1549.1 nm into a LiNbO3 phase modulator driven by an RZ signal, which in turn was generated by converting a 2.5Gb/s NRZ PRBS data stream with a length of 2^23-1 to a 50% RZ signal. The use of 50% RZ produces un-modulated slots that are 200ps wide. We did not actually use a pre-coder in the experiment because the data is random and it would not make any difference. The generated DPSK signal is passed through an optical bandpass filter with a 3dB bandwidth of ~0.8nm which emulates one channel of a 100 GHz arrayed waveguide grating (AWG), and amplified to 7dBm using an EDFA then fed into a 20 km standard single-mode fiber (SMF) through an optical circulator.
At the ONU side, and after passing the received signal through a circulator we split it using a 20:80 optical coupler, taking 20% of the light to a downstream receiver shown in the inset in Fig. 1 to obtain downstream data in a return to zero (RZ) format, because interfering every two un-modulated slots together will produce no power at the output. The rest of light is fed into a 2.5Gb/s Bias-Ready JDSU intensity modulator driven by a 2.5Gb/s clock source, which plays the role of a pulse carver. We use an RF phase-shifter with the clock stream to adjust the process of removing data slots which are contained in the downstream signal. The signal then becomes an RZ of a 50% duty cycle with no phase information and a power level of −8.9dBm. We feed this signal into another LiNbO3 phase modulator driven by a 2.5GB/s NRZ data stream which is also PRBS data with a length of 2^23-1 to obtain the RZ-DPSK upstream signal with −13.5dBm which is in turn launched into the 20 km SMF using the circulator at the ONU side. We receive this signal at the OLT side by amplifying it with an EDFA and then passing it through a DPSK receiver which has the same structure as the one we used to receive the downstream signal.
The bit-error-rate (BER) measurement results for both downstream and upstream signals are shown in Fig. 2 . For single ended detection we used a PP-10G Nortel receiver while for balanced detection we used an R410 with limiting Transimpedance Amplifier (TIA) Discovery Semiconductors balanced detector, and to investigate the issue of Rayleigh noise on the US signal, we measured the BER in both the cases of single-fiber and dual-fiber configuration.
In the upstream we measured a receiver sensitivity at BER = 10^-9 in the case of single-fiber −20.4dBm compared to −24.2dBm in the dual-fiber case and −23.9dBm in the B2B one, which means a great tolerance toward dispersion and ~4dB penalty coming from RBS noise . As for the downstream we measured a receiver sensitivity at BER = 10^-9 in the case of single ended detection around −7.4dBm for both B2B and single-fiber, while there was around ~5dB advantage in receiver sensitivity for balanced detection over single ended detection.
The limitation in the receiver sensitivity for downstream compared to upstream is coming from the photodetector itself rather than the system.
While the pulse carver in our setup adds cost and complexity to the ONU compared to the scheme discussed in , it should be noted that the pulse carver can be integrated with the phase modulator. In addition the setup in  has more tolerance to RBS noise but less tolerance toward dispersion. Moreover our setup has ~6dB advantage in receiver sensitivity of upstream. Our scheme also offers the option of using a balanced receiver for downstream, for a higher cost.
The use of electrical RZ for downstream was also studied in , and while it has higher tolerance to the RBS noise than our setup, it has an overall ~3dB less receiver sensitivity for upstream.
Figure 3 shows the eye diagrams for upstream and downstream in both single ended and balanced detection. We notice ripples at the zero level inside the eye in (d) we show them more clearly in (e) which we obtain by stopping the data of upstream while keeping all the system working. These ripples are caused by the imperfect removal of downstream data by the intensity modulator, as the latter is driven by a sinusoid signal, rather than an ideal squared signal.
To compare our system with the conventional scheme that uses DPSK for DS and OOK for US, we show in Fig. 3 the eye of DS DPSK (c) and the eye of US OOK in (f), we notice a significance advantage in the upstream case, and this comes from the tolerance of our system toward Rayleigh noise when compared to the case of DPSK-OOK, as we are using single-fiber configuration.
We have proposed and experimentally demonstrated a novel wavelength remodulation scheme with enhanced tolerance toward chromatic dispersion using DPSK with electrical RZ for downstream and DPSK with optical RZ for upstream. The system operated error-free after down and upstream transmission along a single 20-Km fiber without dispersion compensation.
References and links
1. M. Attygalle, T. Anderson, D. Hewitt, and T. Nirmalathas, “Wavelength reuse scheme for source free optical network units inWDM passive optical networks,” presented at the ECOC 2006 Cannes, France, Paper Th 3.5.3.
2. N. Calabretta, M. Presi, R. Proietti, G. Contestabile, and E. Ciaramella, “A Bidirectional WDM/TDM-PON Using DPSK Downstream Signals and a Narrowband AWG,” IEEE Photon. Technol. Lett. 19(16), 1227–1229 (2007). [CrossRef]
3. L. Y. Chan, C. K. Chan, D. T. K. Tong, F. Tong, and L. K. Chen, “Upstream traffic transmitter using injection-locked Fabry–Perot laser diode as modulator for WDM access networks,” Electron. Lett. 38(1), 43–45 (2002). [CrossRef]
4. J. Zhao, L. K. Chen, and C. K. Chan, “A novel re-modulation scheme to achieve colorless high-speed WDM-PON with enhanced tolerance to chromatic dispersion and re-modulation misalignment,” in Proc. OFC, Anaheim, CA, Mar. 25–29, 2007, Paper OWD2.
5. W. Hung, C. K. Chan, L. K. Chen, and F. Tong, “An optical network unit for WDM access networks with downstream DPSK and upstream remodulated OOK data using injection-locked FP laser,” IEEE Photon. Technol. Lett. 15(10), 1476–1478 (2003). [CrossRef]
6. J. Xu and L. K. Chen, “Optical Phase Remodulation for 10-Gb/s WDM-PON with Enhanced Tolerance to Rayleigh Noise”, OFC/NFOEC 2010, paper OThG3, San Diego.
7. J. Xu and L. K. Chen, “A New Remodulation Scheme for WDM-PONs With Enhanced Tolerance to Chromatic Dispersion and Remodulation Misalignment,” IEEE Photon. Technol. Lett. 22(7), 456–458 (2010). [CrossRef]
8. A. H. Gnauck and P. J. Winzer, “Optical phase-shift-keyed transmission,” J. Lightwave Technol. 23(1), 115–130 (2005). [CrossRef]
9. J. Xu, L. K. Chen, and C. K. Chan, “High Extinction Ratio Phase Re-modulation for 10-Gb/s WDM-PON with Enhanced Tolerance to Rayleigh Noise, ” International Conference on Optical Internet, COIN 2010, Paper P1061, Korea, Jeju, Jul. 2010.