PERFORMANCE ASSESSMENT OF TWO-CHANNEL DISPERSION SUPPORTED TRANSMISSION SYSTEMS USING SINGLE AND DOUBLE-CAVITY FABRY-PEROT FILTERS AS DEMULTIPLEXERS

Transcription

1 PERFORMANCE ASSESSMENT OF TWO-CHANNEL DISPERSION SUPPORTED TRANSMISSION SYSTEMS USING SINGLE AND DOUBLE-CAVITY FABRY-PEROT FILTERS AS DEMULTIPLEXERS Mário M. Freire Department of Mathematics and Information Technology, University of Beira Interior Rua Marquês d'ávila e Bolama, 600 Covilhã, Portugal Henrique J. A. da Silva Department of Electrical Engineering, University of Coimbra Largo Marquês de Pombal, 3030 Coimbra, Portugal ABSTRACT This letter presents a theoretical performance assessment assuming that single and double-cavity Fabry-Perot filters are used as demultiplexers in two-channel dispersion supported transmission systems operated at 10 GBit/s. A suitable simulation methodology has been used which combines signal simulation with noise analysis in optically amplified multichannel direct detection systems. The results obtained show an improved system performance if a double-cavity Fabry-Perot filter is used as demultiplexer, resulting in a WDM transmission with a crosstalk power penalty within 1-dB in the region of low dispersion penalty of the DST method, for a channel spacing of 1 nm. The robustness against dependencies on data patterns is verified. Page 1

2 INTRODUCTION The method of dispersion supported transmission (DST) has shown to be very powerful for 10 Gbit/s long span transmission via standard singlemode fibers (SMF) [1]. However, accordingly with the principle of dispersion supported transmission [], the dispersion limited link length estimation for 0 and 40 Gbit/s is about 40 and 10 km, respectively. Besides, there is a limitation on the increase of the bit rate due to the insufficient bandwidth of electronic and optoelectronic devices. One solution for long span high capacity links based on SMF is the dispersion supported transmission of wavelength division multiplexed (WDM) channels operated at 10 Gbit/s. If dense wavelength division multiplexed channels are transmitted, a Fabry-Perot filter (FPF) with high finesse is required to select the desired channel with an acceptable crosstalk level [3]. Since it is known that the double-cavity FPF outperforms the single-cavity FPF used as demultiplexer [4], the impact of this two filter types on the performance of WDM systems with closely spaced channels deserves consideration. In this letter we report a theoretical system performance assessment assuming that single and double-cavity Fabry-Perot filters are used as demultiplexers in two-channel dispersion supported transmission systems. The channels, operating at 10 Gbit/s, were wavelength division multiplexed with 1 nm of channel spacing. SYSTEM MODEL AND PERFORMANCE EVALUATION Fig.1 shows a schematic diagram of a two-channel dispersion supported transmission system. A brief description of the system model follows. The pseudo pattern generator (PPG) provides a maximal-length pseudo random binary sequence (PRBS) with 7-1 bits at 10 Gbit/s. Each one of the optical transmitters consists of a laser driver, a MQW-DFB laser and an optical isolator (I). The rate equation model for quantum-well lasers proposed by Nagarajan et al. [5] has been used to describe the carrier dynamics in the quantum wells and in the separate confinement heterostructure (SCH) layers Page

3 and the photon dynamics in the laser cavity. The emission wavelength of the laser 1 was λ 1 =1531 nm (channel 1) and the emission wavelength of the laser was λ =153 nm (channel ). At the WDM optical multiplexer (WDM MUX) output, the total electric field is the sum of the two input electric fields. Synchronous data patterns are assumed to be transmitted in both channels since this is the worst case for crosstalk [4]. For this situation, the complex envelope of the multiplexed electric field in the frequency domain is given by: E ( ω) = E [ ω ( i 1) ω ], (1) mux N i= 1 i where E i (ω) is the complex envelope of the input electric field i, ω is the angular frequency channel spacing and N is the number of channels (N=). The optical amplifiers (EDFAs) have been considered as linear with a noise equivalent bandwidth of 1.5 THz and a spontaneous emission parameter of 0.3 db. The standard singlemode fiber (SMF) was modeled as a phase filter using the low pass transfer function given in [6], just taking into account the first order dispersion term with D=16. ps/(nm.km) at 153 nm. An optical Fabry-Perot filter [4] is assumed to be used as demultiplexer. In this study we have compared the system performance using single-cavity (SC) or double-cavity (DC) FPF with equal cavities. The finesse of the single-cavity FPF is 150 and the finesse of each cavity for double-cavity FPF is also 150. The 3-dB cutoff frequency of the PIN photodiode (PD) was 9.35 GHz [7]. The receiver main amplifier (AMP) and the low-pass filter (LPF) have been jointly modeled as a low-pass RC filter with the 3-dB bandwidth required by the DST method. For performance evaluation, a semi-analytical method [8] has been used, employing the Gaussian approximation. With this method, signal simulation was combined with noise analysis to evaluate the average error probability, which may be estimated by: P e = 1 L L i= 1 Q V ( τ k) V στ ( k ) th, () Page 3

4 where L is the length of the used PRBS, V(t k ) is the value of the signal waveform at the sampling instant t k, V th is the decision threshold level, s(t k ) is the standard deviation of the noise at the sampling instant, and Q is the well known Q function. After direct detection with a PIN PD, the mean square beat noise current includes signal-ase and ASE-ASE beat noise terms given by [9]: σ s sp= e a sp ηq B hν 4GPL n ( G 1 ) hν, (3) σ sp sp = e a sp ηq B hν 4Ln ( G 1 ) ( hν) B, (4) 0 and, as in [10], the crosstalk-ase beat noise term was also considered. We have found that its variance is σ r sp = N i= 1 ηq 1 B ( ) e 4GPRiL ansp G 1hν hν F πi f + sen 1 π FSR k, (5) where B e is the electrical bandwidth, B o is the optical bandwidth, η is the quantum efficiency of the PIN photo diode, q is electronic charge, h is the Plank constant, ν is the optical frequency, G is the optical preamplifier gain, P is the FPF mean input power for data channel, P Ri is the FPF mean input power for channel i, L a is the loss between the optical preamplifier output and the photodetector input, n sp is the spontaneous emission factor of the EDFA, f is the channel spacing, F is the finesse and FSR is the free spectral range of FPF. The integer k is 1 for SC-FPF and for DC-FPF. Page 4

5 SIMULATION RESULTS AND DISCUSSION The performance assessment was focused on channel. For each fiber length, the system parameters, namely the bias current, the modulation current, the FWHM bandwidth of the FPF and the receiver cutoff frequency, have been adjusted in order to minimize the input mean optical power for an average error probability of The receiver sensitivity for channel, versus fiber length, is shown in figures and 3 assuming that a single-cavity (SC) or a double-cavity (DC) FPF is used as demultiplexer, respectively. For comparison, the receiver sensitivity of the DST system obtained by simulation (theoretical) is also shown. The experimental curve reported by Wedding et al. in [1] is slightly different, which may be explained by the simplifications introduced in the system modeling and performance evaluation reported here. As can be seen in figures and 3, the simulation shows that the power penalty for transmission of two 10 Gbit/s WDM channels is within 1-dB relatively to the single-channel DST transmission for distances ranging: (1) from 0 to 11 km and from 150 to 53 km, assuming that a single-cavity FPF is used as demultiplexer; () from 0 to 11 km and from 6.5 to 300 km, assuming that a double-cavity FPF is used as demultiplexer. As shown in both figures, the same and the complementary data patterns have been considered in the interfering channel (ch. 1), in order to investigate the robustness against pattern dependencies. It was found that the differences in power penalty, using the same and the complementary data patterns, are very small except in the regions of high dispersion penalty of the DST method. In the region of low dispersion penalty (50-50 km), this small differences are less than 0.3 and 0.1 db for single and double-cavity FPF, respectively. Figure 4 shows the receiver sensitivity for channel versus 3-dB FPF bandwidth (FWHM), after single-channel (with FPF) and two-channel dispersion supported transmission via 04 km SMF. A single or a double-cavity FPF is assumed to be used as demultiplexer to select channel. As can be seen in this figure, there is an optimum filter bandwidth (FWHM) Page 5

6 as a compromise between the penalty caused by crosstalk from the adjacent channel, for large filter bandwidth, and the penalty induced by signal distortion, for narrow filter bandwidth. After transmission of two 10 GBit/s WDM channels via 04 km SMF, the power penalty for optimum FPF bandwidth, relatively to the receiver sensitivity obtained for single-channel DST transmission, is less than: (1) 0.9 db, for single-cavity FPF (FWHM=40 GHz); () 0.3 db, for double-cavity FPF(FWHM=50 GHz). This figure shows that power penalty is less than 1-dB, relatively to the receiver sensitivity obtained for single-channel DST transmission, for a 3-dB FPF bandwidth (FWHM) between: (1) 40 and 50 GHz, for single-cavity FPF; () 30 and 90 GHz, for double-cavity FPF. These results indicate that double-cavity FPF are preferred for dense WDM-DST systems, in spite of the additional losses due to the need of isolation between the two filter sections [4]. CONCLUSION A simulation methodology suitable for performance assessment of optically amplified multichannel communication systems has been used to compare the performance of a WDM system with dispersion supported transmission of two 10 Gbit/s channels, using single and double-cavity filters as demultiplexers. If a double-cavity Fabry-Perot Filter is used as demultiplexer, distances ranging from 0 to 11 km and 6.5 to 300 km can be bridged with a power penalty within 1-dB, relatively to single-channel DST transmission. For single-cavity FPF this distance is reduced to 0-11 km and km. The simulation results indicate that double-cavity FPF are preferred for use as demultiplexers and that high density WDM-DST may be realized. Page 6

7 ACKNOWLEDGMENT Part of this work was supported by Junta Nacional de Investigação Científica e Tecnológica (JNICT), Portugal, within Programme CIÊNCIA. Page 7

8 REFERENCES [1] B. Wedding, B. Franz, and B. Junginger, "Dispersion supported transmission at 10 Gbit/s via up to 53 km of standard singlemode fibre", Proc. ECOC'93, paper TuC4.3, Montreux, [] B. Wedding, "New method for optical transmission beyond dispersion limit", Electron. Lett., Vol. 8, No. 14, pp , 199. [3] C. M. Miller and J. W. Miller, "Wavelength-Locked, Two-Stage Fibre Fabry-Perot Filter for Dense Wavelength Division Demultiplexing in Erbium Doped Fibre Amplifier Spectrum", Electron. Lett., Vol. 8, No. 3, pp , 199. [4] P. A. Humblet, and W. M. Hamdy, "Crosstalk analysis and filter optimization of singleand double-cavity fabry-perot filters", IEEE J. Select. Areas Commun., Vol. 8, No. 6, pp , [5] R. Nagarajan, M. Ishikawa, T. Fukushima, R. S. Geels, and J. E. Bowers, "High speed quantum-well lasers and carrier transport effects", IEEE J. Quantum Electron., Vol. 8, No. 10, pp , 199. [6] H. J. A. da Silva, R. S. Fyath, and J. J. O'Reilly, "Sensitivity degradation with laser wavelength chirp for direct-detection optical receivers", IEE Proceedings, Vol. 136, Pt. J, No. 4, pp , [7] J. E. Bowers, and C. A. Burrus, "Ultrawide-band long-wavelength p-i-n photodetectors", IEEE J. Lightwave Tech., Vol. LT-5, No. 10, pp , Page 8

9 [8] M. C. Jeruchim, P. Balaban, and K. S. Shanmugan, "Simulation of communication systems", Plenum Press, N. York, 199. [9] K. Inoue, H. Toba, and K. Nosu, "Multichannel amplification utilizing an Er 3+ -doped fiber amplifier", IEEE J. Lightwave Tech., Vol. 9, No. 3, pp , [10] A. E. Willner, "SNR analysis of crosstalk and filtering effects in an amplified multichannel direct-detection dense-wdm system", IEEE Photon. Technol. Lett., Vol. 4, No., pp , 199. Page 9

10 FIGURE CAPTIONS Figure 1. Block diagram of the simulated WDM-DST system. Figure. Receiver sensitivity for channel versus fiber length assuming that a single-cavity (SC) FPF is used as a demultiplexer. For comparison, the receiver sensitivity for a single-channel DST system it is also displayed. Figure 3. Receiver sensitivity for channel versus fiber length assuming that a double-cavity (DC) FPF is used as a demultiplexer. For comparison, the receiver sensitivity for a single-channel DST system it is also displayed. Figure 4. Receiver sensitivity for channel versus FWHM using a single-cavity (SC) or a double-cavity (DC) FPF as demultiplexer after single-channel (ch.) or twochannel (ch.1+ch.) transmission over 04 km SMF. Page 10

11 FIGURES MQW-DFB I PPG TG DRIVER λ 1 WDM MUX EDFA MQW-DFB I PPG TG DRIVER λ SMF EDFA SMF EDFA SMF FPF EDFA PIN PD AMP LPF Figure 1. Block diagram of the simulated WDM-DST system. Page 11

12 -10 Sensitivity for cha [dbm] Fibre length [km ] SC (same pattern) SC (complementary pattern) DST (theoretical) Figure. Receiver sensitivity for channel versus fiber length assuming that a single-cavity (SC) FPF is used as a demultiplexer. For comparison, the receiver sensitivity for a single-channel DST system it is also displayed. Page 1

13 -10 Sensitivity for cha [dbm] Fibre length [km ] DC (same pattern) DC (complementary pattern) DST (theoretical) Figure 3. Receiver sensitivity for channel versus fiber length assuming that a double-cavity (DC) FPF is used as a demultiplexer. For comparison, the receiver sensitivity for a single-channel DST system it is also displayed. Page 13

14 Sensitivity for cha [dbm] FW H M [GHz] DC (Channel ) SC (Channel ) DC (Channels 1 and ) SC (Channels 1 and ) Figure 4. Receiver sensitivity for channel versus FWHM using a single-cavity (SC) or a double-cavity (DC) FPF as demultiplexer after single-channel (ch.) or twochannel (ch.1+ch.) transmission over 04 km SMF. Page 14