Page 379 - Fiber Optic Communications Fund
P. 379

360                                                               Fiber Optic Communications


                         [s(t) + n c (t)]e *iω c t    2RA LO Re{[s(t) + n c (t)]e *i(ω IF  +   ∆φ) }
                                           Coherent                               ∑
                                            Rx
                                   LO
                                                                                  n d  (t)
                               A LO e *iω LO  +   φ LO


                                                       Envelope detector
                                         t = T b
                         Decision                   (∙)    LPF       2      H I  (ω)
                          device    r(T b )   r(t)                 (∙)


                    Figure 8.17  A heterodyne receiver with an envelope detector and a matched filter H () for OOK.
                                                                                 I

            In Eq. (8.188), the second term on the right-hand side corresponds to the Fourier transform of s()s ( + T −
                                                                                                b
                                                                                          1
            t) at  =±2 . Since the spectral width of s() is much smaller than  , the second term can be ignored.
                        IF
                                                                      IF
            Therefore,
                                               s (t)
                                                F
                                        I (t)=      cos [ (t − T )+Δ],                   (8.189)
                                                               b
                                                         IF
                                         F
                                                 2
                                              ⏟⏟⏟
                                              envelope
            where
                                                  T b
                                          s (t)=    s()s (T +  − t) d.                 (8.190)
                                           F    ∫       1  b
                                                 0
            The output of the matched filter passes through an envelope detector which can be imagined as a cascade of
            squarer, low-pass filter, and square-rootor, as shown in Fig. 8.17. When we square I (t), we obtain a term
                                                                                 F
            proportional to cos [2 (t − T )+Δ] which is rejected by the low-pass filter. The signal output of the
                               IF
                                     b
            envelope detector is the envelope of I (t) (shown in Eq. (8.189)), which is given by
                                          F
                                                  u(t)= s (t)∕2,                             (8.191)
                                                        F
                                               s (T )  1   T b
                                                F
                                                   b
                                        u(T )=       =       s()s () d,                (8.192)
                                                                 1
                                           b
                                                 2     2 ∫ 0
                                          u(T )= E ∕2 when ‘1’ is sent
                                             b    1
                                               = 0      when ‘0’ is sent.                    (8.193)
            Next, consider the noise component before the matched filter given by Eq. (8.121),
                                                                          n (t)
                                                                           d
                                n(t)= n cos ( t +Δ)+ n  sin ( t +Δ)+   .             (8.194)
                                      cI     IF         cQ     IF
                                                                         2RA
                                                                            LO
            Here, we have dropped the scaling factor 2RA  . Since the ASE is expressed as the modulated noise process,
                                                 LO
            it is convenient to express the detector noise as the modulated noise process as well, i.e.,
                                    n (t)= n cos ( t +Δ)+ n dQ  sin ( t +Δ),         (8.195)
                                           dI
                                                  IF
                                     d
                                                                   IF
   374   375   376   377   378   379   380   381   382   383   384