Page 486 - Fiber Optic Communications Fund
P. 486

Nonlinear Effects in Fibers                                                        467


           and  post , respectively. Dispersion of the TF is partially compensated by the in-line dispersion compensation.
           Let  res  be the residual accumulated dispersion of a single span, i.e.,

                                           res  = D L  + D inline inline ,               (10.314)
                                                             L
                                                 TF TF
           where D and L denote the dispersion parameter (see Chapter 2) and length, respectively, and the subscripts
           TF and inline correspond to the transmission fiber and inline DCF, respectively. We have assumed that the
           total accumulated dispersion from transmitter to receiver is zero, i.e.,
                                            pre  + N res  +  post  = 0,                (10.315)

           where N is the number of TF spans. The following parameters are used in the numerical simulation of the
           direct detection OOK system. Bit rate = 40 Gb/s, pulsewidth (FWHM) = 5ps, N = 10, peak powers launched
           to TF and DCF are 10 dBm and 0 dBm, respectively. The lengths of pre-, inline, and post-compensating
           fibers are chosen so that   = 100 ps/nm and   =   =−500 ps/nm. The amplifier noise is turned off.
                                res              post   pre
           Two pulses centered at 25 ps and 50 ps are launched to the fiber-optic link. Owing to IFWM, echo pulses are
           generated around 0 ps and 75 ps. The solid and broken lines in Fig. 10.29 show the echo pulses after 10 spans
           obtained by the analytical expression (Eq. (10.281)) and numerical simulations, respectively. In this example,
           a small pulse width is chosen so that the echo pulse is not affected by the ISI from the pulse centered at 25 ps.
           In practice, short-duty-cycle pulses are rarely used because of the large bandwidth which leads to cross-talk
           in WDM systems.
            The percentage pre-compensation ratio is defined as
                                                             pre  × 100
                                     %pre-compensation ratio =        .                   (10.316)
                                                               + 
                                                             pre   post
            Fig. 10.30 shows the variance as a function of pre-compensation ratio. The following parameters are used for
           Fig. 10.30: pulsewidth (FWHM) = 12.5 ps, D res  = 100ps/nm, peak powers launched to TF and DCF are 0 dBm


                                   0
                                                                     numeric
                                                                     analytic
                                 –20

                                Power (dBm)  –40



                                 –60


                                 –80


                                 –100
                                       –10      –5      0       5       10
                                                     Time (ps)
           Figure 10.29  Comparison of the echo pulse power at the output obtained by the analytical expression (Eq. (10.281))
           and numerical simulations. Two signal pulses centered around 25 ps and 50 ps are launched to the fiber (not shown in the
           figure).
   481   482   483   484   485   486   487   488   489   490   491