Open Access
Add to BibSonomyAbstract
A novel coherent multi-level OTDM transmission scheme is presented by using an RZ-CW conversion technique in a coherent receiver, in which the transmitted coherent RZ pulses are OTDM demultiplexed and converted into a CW signals for demodulation with a CW-LO. This scheme enables us to achieve a high SNR demodulation and demultiplexing performance even for ultrahigh-speed OTDM signals with ultra multi-level modulation, which would be difficult to realize with a conventional scheme of simultaneous demodulation and demultiplexing using a pulsed-LO. The advantage of this scheme is successfully demonstrated with 10 Gsymbol/s, 4- and 8-OTDM, 32-RZ/QAM demodulation experiments.
©2011 Optical Society of America
1. Introduction
Increases in the bit rate per wavelength channel are significant with respect to achieving high-speed and high-capacity optical networks. Coherent optical time division multiplexing (OTDM) transmission combined with a multi-level modulation format has been recently proposed for realizing high-speed and spectrally efficient transmission [1–3]. The speed and bandwidth limitations of electrical components such as A/D and D/A converters in a conventional coherent system can be overcome by using a multi-level amplitude and phase-modulated coherent return-to-zero (RZ) OTDM pulse train. Furthermore, OTDM demultiplexing and homodyne-detection can be simultaneously realized using a pulsed local oscillator (LO) [4] without an optical demultiplexer. Employing this scheme, 640 Gbit/s, QPSK [1] and 1.28 Tbit/s, 16 QAM [2] transmissions have been demonstrated at 160 Gsymbol/s (10 Gsymbol/s × 16 OTDM). In addition, by increasing the OTDM multiplicity to 128, a 10.2 Tbit/s, 16 QAM data was transmitted with a self-homodyne detection method [5]. We have also achieved a QAM multiplicity of up to 32 levels in RZ pulses and demonstrated a 400 Gbit/s-225 km transmission by using an optical phase-locked loop [3].
However, in all above experiments, due to the input power limitation of coherent detection circuits such as 90-degree optical hybrid and balanced photo detectors (B-PD), it is difficult to achieve a demodulation of an ultra-short coherent RZ data signal with a high signal to noise ratio (SNR). It inevitably causes SNR degradation in one symbol with an increase in the multiplicity in OTDM and QAM.
In this paper, we propose a new demodulation and demultiplexing scheme based on RZ-CW conversion that can realize ultra multi-level, high-speed coherent transmission and thus overcome the limitation in OTDM and QAM multiplicity. The effectiveness of the scheme has been successfully demonstrated in both 4- and 8-OTDM 10 Gsymbol/s, 32 RZ/QAM demodulation experiments.
2. Principle of coherent OTDM transmission with RZ-CW conversion scheme
A conventional coherent OTDM transmission has some drawbacks as illustrated in Figs. 1 and 2 . Figure 1 is a block diagram of a conventional coherent homodyne-detection circuit. When coherent N-OTDM RZ data signals are OTDM-demultiplexed and homodyne-detected with a pulsed-LO in the circuit, the detected signal power of one tributary is reduced to 1/N of the total received power. This is attributed to the input power limitation of the optical and electrical circuits. In addition, compared to a CW carrier case shown in Fig. 2(a), the RZ-carrier uses only a small fraction of the broad-bandwidth RZ data signal power which contributes to demodulation as shown in Fig. 2(b). These difficulties cause a serious SNR degradation in one symbol when increasing the multiplicity in OTDM and modulation format.
Figure 3 shows a new coherent OTDM transmission system employing an RZ-CW conversion technique in front of the 90-degree optical hybrid. The RZ-CW conversion greatly improves the SNR at the receiver. A multi-level coherent OTDM data signal is generated using a pulse carving circuit, an IQ modulator and an OTDM multiplexer at the transmission point. After transmission, the coherent OTDM RZ data are OTDM demultiplexed followed by the conversion of the demultiplexed RZ data signal to a CW data signal using an RZ-CW conversion circuit. The RZ-CW conversion circuit is composed of a dispersive medium and an optical phase modulator. Here, the RZ signal is passed through the dispersive medium, after which it is given an opposite chirp using the phase modulator. An RZ signal can be converted into a CW data signal by driving the phase modulator with an appropriate modulation depth. Then, the converted CW data signal is homodyne-detected with a CW-LO at a 90-degree optical hybrid and B-PDs. This technique narrows the broadened RZ data signal spectrum hence an ultra multi-level, high-speed coherent OTDM signal can be demodulated with a high SNR, which is comparable to that of CW-carrier coherent transmission.
Figure 4 shows our experimental setup for demonstrating RZ-CW conversion. A CW, 1538.8 nm C2H2 frequency-stabilized erbium fiber laser (EFL) [6] is used as a coherent light source. Two kinds of RZ pulse carving schemes are used to generate a 10 GHz coherent pulse train. The first scheme consists of an LN phase modulator and an SMF as shown in Fig. 4(b). The second scheme has an optical comb generator [7], followed by an optical filter and an SMF as shown in Fig. 4(c). The generated RZ pulses are converted into a CW signal using an RZ-CW conversion circuit.
Fig. 4 Experimental setup for demonstration of RZ-CW conversion scheme (a), and detailed configuration of pulse carving circuit using an LN phase modulator (b) and an optical comb generator (c).
For the pulse carving scheme in Fig. 4(b), an ideal RZ-CW conversion can be realized by using dispersion and chirp that are exactly opposite to those used for pulse carving. For general RZ pulses, including those generated using the setup in Fig. 4(c), an RZ to quasi CW conversion can be achieved, in which the spectrum is compressed at a ratio equal to the inverse of the broadening ratio at the dispersion medium. This can be realized by exactly canceling the dispersion-induced chirp using the phase modulator. For example, for a Gaussian pulse Aexp(−t2/2T02), it can be shown that dispersion D = β2L and linear chirp exp(iKt2/2) compress the spectrum by a factor F = 1/[1 + (D/T02)2], where the required chirp is K = D/(T04 + D2). This indicates that a large D is preferable for improving the RZ-CW conversion, but at the same time a large K is required. Spectrum compression is also possible with the time-domain optical Fourier transformation (OFT) technique [8]. Indeed, the configuration of the RZ-CW conversion circuit is similar to that used for time-to-frequency conversion based on OFT [9, 10]. However, it should be noted that the present RZ-CW conversion technique does not require a special relationship between D and K, while OFT requires the condition D = 1/K.
We undertook an RZ-CW conversion experiment with the two RZ pulse carving schemes shown in Figs. 4(b) and (c). Figure 5 shows the optical spectra and the pulse waveforms before and after RZ-CW conversion when the RZ pulse was generated with an LN phase modulator. Here, the SMF length was 2.1 km (33.6 ps/nm). The LN phase modulators for carving and conversion were both driven with a modulation depth of 2.5 π. By using a DCF with a dispersion of −30.7 ps/nm in the conversion circuit, the RZ signal was successfully converted to a CW signal with a side-mode suppression ratio of 30 dB. The magnitude of the fiber dispersion was determined experimentally to achieve the maximum suppression of the side modes considering the differences in phase modulator performance. This is a consequence of the symmetry of the pulse carving and RZ-CW conversion configuration.
Fig. 5 Optical spectra and pulse waveforms (a) before RZ-CW conversion, (b) after RZ-CW conversion when an RZ pulse is generated with an LN phase modulator.
The experimental results obtained with an optical comb generator for pulse carving are shown in Fig. 6(a) and (b) . Here, the filter bandwidth was 1 nm and the SMF length was 1.4 km. In the RZ-CW conversion circuit, the modulation depth and the dispersion of the DCF were set at 2.5 π (K = 0.016 ps−2) and −65 ps/nm (D = 82.9 ps2), respectively, satisfying the K and D requirements described above. The estimated spectral compression ratio is F = 9.3. In this case, the small hump that accompanies the RZ-quasi CW conversion is attributed to the intensity modulation components of the RZ pulse, which is introduced at the optical comb generator and cannot be removed at the converter.
Fig. 6 Optical spectra and pulse waveforms (a) before RZ-CW conversion, (b) after RZ-CW conversion when an RZ pulse is generated with an optical comb generator.
3. Demodulation performance of 10 Gsymbol/, 4- and 8-OTDM 32 RZ/QAM signals with RZ-CW conversion scheme
We used this technique to demodulate and demultiplex 10 Gsymbol/s, 4- and 8-OTDM 32 RZ/QAM signals. Figure 7 shows the experimental setup. At the transmitter, a combination of an optical comb generator, an optical filter and a 1.4 km-SMF were used to generate 10 GHz, 7 and 5 ps RZ coherent pulse signals. These pulse signals were used for the 4- and 8-OTDM experiments, respectively. The RZ pulse was 32 QAM modulated at an IQ modulator driven by an arbitrary waveform generator (AWG). The 10 Gsymbol/s, 32 RZ/QAM data were then optically time-division multiplexed to 40 or 80 Gsymbol/s with a planar lightwave circuit (PLC)-type OTDM multiplexer. Figures 8(a) and (b) show the optical spectrum and pulse waveform of 40 and 80 Gsymbol/s, 32 RZ/QAM data, respectively. The OSNRs of these spectra obtained with a resolution bandwidth of 0.01 nm were both 34 dB. At thereceiver, an OTDM 32 RZ/QAM signal was first OTDM demultiplexed by using a nonlinear optical loop mirror (NOLM). Here, with the use of a control pulse generation circuit as shown in Fig. 9 , a 10 ps control pulse was generated by using a CW-DFB LD, and an optical comb generator [7] which was driven by a clock signal extracted from the QAM data [11]. The 10 Gsymbol/s, 32 QAM signal was then fed into the RZ-CW conversion circuit. The converted quasi CW 10 Gsymbol/s, 32 QAM signal was self homodyne-detected with part of the transmitter signal as a CW LO for simplicity. Thereafter, the data signal was demodulated with a digital signal processor (DSP) in an offline condition.
Fig. 7 Experimental setup for 10 Gsymbol/s, 4- and 8-OTDM, 32 RZ/QAM modulation/demodulation with an RZ-CW conversion scheme.
Figures 10(a) and (b) show the optical spectra of the demultiplexed 10 Gsymbol/s 32 RZ/QAM signal before and after RZ-CW conversion obtained with 4- or 8-OTDM signals, respectively. The spectral widths were both narrowed with the SNR increasing by 5 and 4.5 dB, respectively. Figure 11 shows the BER characteristics after demodulation with an RZ-CW conversion technique. The demodulation results without RZ-CW conversion are also shownfor comparison. Here, part of the RZ signal before QAM modulation was used as a pulsed-LO for self-homodyne detection. The BER for a 10 Gsymbol/s, CW-32 QAM signal demodulated with a CW-LO is also included. When using a 4-OTDM data signal, the power penalty with the RZ-CW conversion improved by as much as 5.5 dB at a BER of 1x10−4 compared to without the RZ-CW conversion. With an 8-OTDM signal, the demodulation result is 9 dB better without the RZ-CW conversion. The difference between the BER improvements in the 4- and 8-OTDM experiments are due to the difference in the power Nth increase per 10 Gsymbol/s data signal for demodulation resulting from RZ-CW conversion. These results indicate that the RZ-CW conversion scheme has a great advantage for higher multiplicity coherent OTDM transmission. However, compared with the result obtained with CW-32QAM, there is still a residual power penalty. This is mainly due to incomplete CW conversion as described in section 2.
Fig. 10 Optical spectra (0.01 nm resolution) of 10 Gsymbol/s, 32 RZ/QAM signal before and after RZ-CW conversion obtained with (a)4 and (b)8-OTDM signal.
4. Conclusion
A new coherent OTDM transmission system employing an RZ-CW conversion scheme was presented. With this scheme, an ultra multi-level RZ data signal can be demodulated with a high SNR which overcome the limitations in OTDM and QAM multiplicity. The scheme is shown to be highly advantageous in 10 Gsymbol/s, 4- and 8-OTDM, 32 RZ/QAM signal demodulation experiments. The present technique is more advantageous for higher multiplicity coherent OTDM transmission.
References and links
1. C. Zhang, Y. Mori, M. Usui, K. Igarashi, K. Katoh, and K. Kikuchi, “Straight-line 1,073-km transmission of 640-Gbit/s dual-polarization QPSK signals on a single carrier,” in 35th European Conference on Optical Communication, 2009. ECOC '09(2009), postdeadline paper PD2. 8.
2. C. Schmidt-Langhorst, R. Ludwig, L. Molle, D. Gros, R. Freund, and C. Schubert, “Terabit/s single-carrier transmission system based on coherent time-division demultiplexing,” in Optical Fiber Communication Conference, OSA Technical Digest (CD) (Optical Society of America, 2010), paper OThV3.
3. K. Kasai, T. Omiya, P. Guan, M. Yoshida, T. Hirooka, and M. Nakazawa, “Single-channel 400 Gbit/s, OTDM-32 RZ/QAM coherent transmission over 225 km using an optical phase-locked loop technique,” IEEE Photon. Technol. Lett. 22(8), 562–564 (2010). [CrossRef]
4. F. Ito, “Interferometric demultiplexing experiment using linear coherent correlation with modulated local oscillator,” Electron. Lett. 32(1), 14–15 (1996). [CrossRef]
5. T. Richter, E. Palushani, C. Schmidt-Langhorst, M. Nölle, R. Ludwig, and C. Schubert, “Single wavelength channel 10.2 Tb/s TDM-data capacity using 16-QAM and coherent detection,” in Optical Fiber Communication Conference, OSA Technical Digest (CD) (Optical Society of America, 2011), paper PDPA9.
6. K. Kasai, A. Suzuki, M. Yoshida, and M. Nakazawa, “Performance improvement of an acetylene (C2H2) frequency-stabilized fiber laser,” IEICE Electron. Express 3(22), 487–492 (2006). [CrossRef]
7. T. Sakamoto, T. Kawanishi, and M. Izutsu, “Asymptotic formalism for ultraflat optical frequency comb generation using a Mach-Zehnder modulator,” Opt. Lett. 32(11), 1515–1517 (2007). [CrossRef] [PubMed]
8. M. Nakazawa, T. Hirooka, F. Futami, and S. Watanabe, “Ideal distortion-free transmission using optical Fourier transformation and Fourier transform-limited optical pulses,” IEEE Photon. Technol. Lett. 16(4), 1059–1061 (2004). [CrossRef]
9. P. Guan, H. C. H. Mulvad, K. Kasai, T. Hirooka, and M. Nakazawa, “High time-resolution 640-Gb/s clock recovery using time-domain optical Fourier transformation and narrowband optical filter,” IEEE Photon. Technol. Lett. 22(23), 1735–1737 (2010). [CrossRef]
10. H. C. H. Mulvad, E. Palushani, H. Hu, H. Ji, M. Galili, A. T. Clausen, M. Pu, K. Yvind, J. M. Hvam, P. Jeppesen, and L. K. Oxenlowe, “Ultra-high-speed optical serial-to-parallel data conversion in a silicon nanowire,” in 37th European Conference and Exposition on Optical Communications, OSA Technical Digest (CD) (Optical Society of America, 2011), paper Th.13.A.2.
11. C. Boerner, V. Marembert, S. Ferber, C. Schubert, C. Schmidt-Langhorst, R. Ludwig, and H. G. Weber, “320 Gbit/s clock recovery with electro-optical PLL using a bidirectionally operated electroabsorption modulator as phase comparator,” in Optical Fiber Communication Conference and Exposition and The National Fiber Optic Engineers Conference, Technical Digest (CD) (Optical Society of America, 2005), paper OTuO3.
