Monday Poster Sessions 87-102

Mday, M arch 24 transmission system was 20 Gb/s, while the symbol rate was 10 Gbaud (=10 GS/s). The second MZ modulator was used as a pulse carver to generate RZ-DQPSK signals. For the 20-Gb/s DQPSK receivers, two 100-ps DIs with two separate balanced receivers are generally required to simultaneously demodulate the two 10-Gb/s data streams contained in the 10-Gbaud DQPSK signal [4]. However, we utilized a single 100-ps DI to demodulate each 10-Gb/s data stream individually by adjusting the differential optical phase between the two DI arms to be π/4 or –π/4. In our experiment, no data encoder circuit was employed. Thus, to allow bit-error ratio (BER) measurements, we programmed the error detector with the expected data sequence. III. Experimental results and simulation Fig. 2(a) shows the measured receiver sensitivity penalty (BER=10-9) for 10-Gb/s DPSK signals as a function of the frequency offset between optical source and DI. The measured results show that NRZ-DPSK and RZ-DPSK exhibit almost identical penalties when changing the frequency offset. For example, both coding schemes experience a 1-dB penalty in receiver sensitivity when the frequency offset amounts to ±300 MHz. Fig. 2(b) shows the measured sensitivity penalties for 20Gb/s (10-Gbaud) DQPSK signals. The results show that, like for DPSK systems, the frequencyoffset tolerance is almost independent of the pulse coding scheme (i.e., NRZ or RZ). However, the DQPSK system is much more sensitive to frequency offsets than the DPSK system: For example, the frequency offset should be kept within ±100 MHz to avoid a power penalty of 1 dB. It should be noted that, in our experiment, the aggregate bit rate for DQPSK is 20 Gb/s while it is 10 Gb/s for DPSK. For a fair comparison, we therefore scale up all measured numbers to the same bit rate. Table I summarizes the allowable frequency offsets, relative to the bit rate and (in brackets: scaled to a 40-Gb/s system,) for a 1-dB and a 2-dB sensitivity penalty. It shows that DPSK is ~6 times more robust to frequency offsets than DQPSK. We also verified the higher robustness of DPSK with respect to frequency offsets by means of simulation. We used a semi-analytical calculation method for the BER, involving numerical solutions to the exact photocurrent probability densities for a 27–1 PRBS, since the Gaussian approximation to the probability densities fails in the case of balanced DPSK reception [5]. Figure 3 shows the simulated receiver sensitivity penalty as a function of optical source frequency offset, normalized to the bit rate, for DPSK and DQPSK, both for NRZ (dashed) and RZ (solid). The simulations assume an optical 4th-order Gaussian filter (3-dB bandwidth of 2.2 times the symbol rate), an electronic receiver bandwidth of 0.6 times the symbol rate, and perfect phase modulation (instantaneous phase shifts), as produced by an MZ modulator biased at its transmission minimum. The simulation results confirm that DQPSK is more sensitive to frequency offsets than DPSK by a factor of 6. The reasons why DQPSK systems are more sensitive to the optical source frequency offsets than DPSK systems by a factor of 6 can be attributed to (a) the reduced DQPSK symbol rate (making up for a factor of 2), (b) the higher eye-closure penalty for DQPSK that is caused by the different number of optical symbol phases as well as by the different operating point on the DI transfer characteristics (making up for a factor of 3.5), and (c) the higher impact of optical noise on DPSK reception (reducing the net difference in frequency offset tolerance between DPSK and DQPSK from a factor of 2×3.5=7 to a factor of 6). IV. Summary We have assessed, both by measurement and simulation, the frequency-offset tolerance between optical source and delay-interferometer for DPSK and DQPSK systems. Our results show that DQPSK is about six times more sensitive to frequency offsets than DPSK operating at the same bit rate. References 1. E. A. Swanson, J. C. Livas, and R. S. Bondurant, “High sensitivity optically preamplified direct detection DPSK receiver with active delayline stabilization.” IEEE Photon. Technol. Lett., 6, 263-265 (1994). 2. W. A. Atia and R. S. Bondurant, “Demonstration of return-to-zero signaling in both OOK and DPSK formats to improve receiver sensitivity in an optically preamplified receiver.” in Proceedings of LEOS, paper TuM3 (1999). 3. A. H. Gnauck et. al., “2.5 Tb/s (64×42.7 Gb/s) transmission over 40 x 100 km NZDSF using RZDPSK format and all-Raman-amplified link.” in Proceedings of OFC, paper FC2 (2002). 4. R. A. Griffin et. al., “10 Gb/s optical differential quadrature phase shift key (DQPSK) transmission using GaAs/AlGaAs integration.” in Proceedings of OFC, paper FD6 (2002). 5. P. J. Winzer and S. Chandrasekhar, “Influence of optical filtering on return-to-zero differential phase shift keying (RZ-DPSK).” in Proceedings of ECOC, paper 9.6.7 (2002).