Page 330 - Satellite Communications, Fourth Edition
P. 330

310  Chapter Ten

                                With QPSK, the signal can be represented by the formulas given in
                              Table 10.1, which may be written generally as

                                                                      n
                                                  e(t)   22 cosa  t      b              (10.27)
                                                                 0
                                                                      4
                              Quadrupling this, followed by some trigonometric simplification,
                              results in


                                       4
                                       e (t)    3    2 cos 2a  0 t    n  b    1  cos 4a  0 t    n  b  (10.28)
                                             2                 4     2            4
                              The last term on the right-hand side is selected by the bandpass filter
                              and is

                                             1           n      1
                                               cos 4a  0 t    b     cos(4  0 t   n )    (10.29)
                                             2            4     2

                                This is seen to consist of the fourth harmonic of the carrier, including
                              a constant-phase term that can be ignored. The fourth harmonic is
                              selected by the bandpass filter, and the operation of the circuit proceeds
                              in a similar manner to that for the BPSK signal.
                                Frequency multiplication can be avoided by use of a method known
                              as the Costas loop. Details of this, along with an analysis of the effects
                              of noise on the squaring loop and the Costas loop methods, will be
                              found in Gagliardi (1991). Other methods are also described in detail
                              in Franks (1980).



                              10.8 Bit Timing Recovery
                              Accurate bit timing is needed at the receiver in order to be able to sample
                              the received waveform at the optimal points. In the most common
                              arrangements, the clocking signal is recovered from the demodulated
                              waveform, these being known as self-clocking or self-synchronizing systems.
                              Where the waveform has a high density of zero crossings, a zero-crossing
                              detector can be used to recover the clocking signal. In practice, the received
                              waveform is often badly distorted by the frequency response of the trans-
                              mission link and by noise, and the design of the bit timing recovery circuit
                              is quite complicated. In most instances, the spectrum of the received wave-
                              form will not contain a discrete component at the clock frequency. However,
                              it can be shown that a periodic component at the clocking frequency is
                              present in the squared waveform for digital signals (unless the received
                              pulses are exactly rectangular, in which case squaring simply produces a
                              dc level for a binary waveform). A commonly used baseband scheme is
   325   326   327   328   329   330   331   332   333   334   335