Page 209 - Mechanical Engineers' Handbook (Volume 2)
P. 209

198   Signal Processing
















                                      Figure 8 Prototype circuit element for construction of analog filter.


           3.1 z-Transforms

                          The z-transform is used to analyze the frequency response and stability of a system of
                          difference equations in much the same way that the Laplace transform is used to analyze
                          the frequency response and stability of a system of differential equations. The z-transform
                          of Eq. (24) is
                                                                           M   j
                                               a   az  1          az  M    0  az
                                                                             j
                                         H(z)    0   1          M                             (25)
                                                 1   bz  1        bz  N  1    j 1  bz  j
                                                                            N
                                                               N
                                                      1
                                                                               j
                                                               j
                          We determine the frequency response by 
H(e )
, where 0       2  and corresponds to
                          the scaled frequencies of our sampled system from 0 Hz up to the sampled frequency. The
                          function e  j   is a periodic signal. For         2 , e  j     e  j(   ) , and for 2        4 , e
                          j     e  j(  2 ) . Thus frequencies greater than  , corresponding to the Nyquist frequency, or
                          one-half of the sampling frequency assume an identical characteristic to an analogous fre-
                          quency less than  . This phenomenon is called aliasing and is illustrated in Fig. 9.
           3.2  Design of FIR Filters
                          It is possible to use the Fourier transform to determine the coefficients to an FIR filter.
                          However, the Fourier coefficients are generally complex numbers, and when working with
                          real signals, it is desirable to have a real coefficients in our filter. To do this, we apply Euler’s
                          identity and observe from Eq. (4) that the coefficients will be real if the inner products with


                                        Table 1 Circuit Elements for the Construction of Basic
                                        Transfer Function Prototypes
                                        Single pole                  Z 1 ← resistor
                                                                     Z 2 ← RC in parallel
                                        Single zero                  Z 1 ← RC in parallel
                                                                     Z 2 ← resistor
                                        Complex-conjugate pole pair  Z 1 ← LRC in series
                                                                     Z 2 ← capacitor
                                        Complex-conjugate zero pair  Z 1 ← capacitor
                                                                     Z 2 ← LRC series
   204   205   206   207   208   209   210   211   212   213   214