Page 362 - Analog and Digital Filter Design
P. 362

359
                                                         Introduction to Digital Filters




                                  w,
                            h(0) = -
                                  n
                            where n = 0.


                      To  produce the desired frequency response we  must have an impulse response
                      that produces an output before the impulse arrives! This is impossible. The solu-
                      tion is to delay the signal, so that some signal processing takes place before the
                      peak of  the impulse response arrives at the output. The longer the delay: the
                      closer we get to the ideal frequency response. Limiting the period during which
                      signal processing takes place is known as truncation. This can lead to rounding
                      of  the passband edge and ripples in the stopband.


                      A  1 radts lowpass filter can now be designed using discrete logic. First build a
                      chain of delay elements. usually these are D-type flip-flops clocked by the master
                      clock. Each delay element is  10 bits wide. From the output of  each stage take
                      the digitized signal and multiply  it by  the value of  the impulse response that
                      corresponds to that moment in time. An example will help explain this further.
                      Suppose we  have 21 delay elements. Delay elements 1 to  10 produce the nega-
                      tive time outputs, delay element 11 corresponds with the zero time output, and
                      delay elements 12 to 21 are the positive time outputs.


                      The  output  from  delay  element  1  will  be  multiplied  by  the  value  of  the
                      impulse response at -20  seconds, to give product one. In Figure 15.2 the impulse
                      response  value  is  approximately  0.01453  at  -20  seconds,  where  t  = -2000
                      hundredths  of  a  second.  The  next  delay  element  output  will  be  multiplied
                      by  the  impulse response  at  -18  seconds,  to  give product  two.  Further  delay
                      element outputs are multiplied by  the impulse response at -16,  -14,  -12.  -10.
                      -8.  -6,  -4,  -2,  0, 2, 4,  6, 8, 10,  12, 14, 16, 18, and 20 seconds. to give products
                      3  to  21.  All  these products  must  now be  added  together, to form  the  filter’s
                      output signal.


                      The hardware we  need  is  as follows: 21 x  10 D-type flip-flops; and  21 rnulri-
                      pliers and a summing circuit with  21 inputs. This list  omitted  the analog-to-
                      digital (MD) and digital-to-analog (D/A) conkerters at the input and output. A
                      circuit diagram is given in Figure  15.3. This type of  filter is known as a finite
                      impulse response (FIR) filter because, if  there are 21 taps. after 21 clock pulses
                      an impulse signal will have passed through  and no longer affect the output. In
                      this diagram, “D” represents a delay, “x” represents a multiplication, and ”Sum”
                      represents an addition.
   357   358   359   360   361   362   363   364   365   366   367