Page 108 - Mechatronics with Experiments
P. 108

94   MECHATRONICS
                              impose conditions that CLS poles must have a real part smaller than a negative real value

                                                         Re(p ) < 0  or  − a                    (2.107)
                                                             i
                              or in the frequency domain it can be specified in terms of gain and phase margins,
                                                     GM > GM  min ,  PM > PM min                (2.108)

                                   The response quality is generally divided into two groups: transient response and
                              steady-state response. The transient response is generally specified as the step response.
                              The step response is quantified using percent overshoot, settling time, and rise time spec-
                              ifications: (PO%, t , t ). For a second-order closed loop system, the PO% and t uniquely
                                            s  r                                             s
                              determine the closed loop system poles with damping ratio (  ) and natural frequency (w ):
                                                                                                   n
                                                                √
                                                                      2
                                                   p 1,2  =−  w ± j 1 −    w n                  (2.109)
                                                            n
                                                         4.0
                                                     t =    ;                                   (2.110)
                                                     s
                                                           w n
                                                             √
                                                         −    ∕ 1−   2
                                                   PO = e        ;  for 0 ≤   < 1.0             (2.111)
                              In the frequency domain, the cross-over frequency (w ) of the loop transfer function and
                                                                         cr
                              bandwidth (w ) of the closed loop transfer function, along with the phase margin of the
                                         bw
                              loop transfer function, correlate well with the transient response. Cross-over frequency is
                              defined for the loop transfer function where the magnitude of the loop transfer function
                              is 1 (or 0 dB). Bandwidth frequency is defined for the closed loop transfer function where
                              the magnitude is 0.707 or (−3 dB). Bandwidth closely relates to the speed of response (t ),
                                                                                                    s
                              and phase margin (PM) is closely related to the damping ratio. For a second-order closed
                              loop system as shown in Figure 2.38, it can be shown [6] that
                                                          2  
                                                 −1
                                         PM =tan   √                                            (2.112)
                                                     √
                                                       1 + 4   − 2   2
                                                            2
                                            ≈ 100    for    ≤ 0.6                               (2.113)
                                              √
                                                       √
                                                           2
                                                    2
                                         w =    −2   +   4   + 1 ⋅ w ; w = w at |G(jw)| = 1.0   (2.114)
                                          cr
                                                                 n
                                                                         cr
                                              ;20 log |G(jw)| = 0.0 dB                          (2.115)
                                                    10
                                                  √
                                                              √
                                                                       2
                                                                  2
                                                          2
                                        w bw  = w ⋅  (1 − 2   ) +  (2   − 1) + 1                (2.116)
                                                n
                                              ; w = w bw  at |G(jw)∕(1 + G(jw))| = 0.707        (2.117)
                                              ;20 log |G(jw)∕(1 + G(jw))| =−3 dB                (2.118)
                                                    10
                                                √
                                                         2
                                          w = w   1 − 2 ⋅    ; where |G(jw)∕(1 + G(jw))| is maximum  (2.119)
                                           p    n
                                                         d
                                              ; w = w at   (|G(jw)∕(1 + G(jw))|) = 0.0          (2.120)
                                                    p
                                                        dw
                                              ; for 0 ≤    ≤ 0.707                              (2.121)
                                                                                                (2.122)
                                         M =max |G(jw)∕(1 + G(jw))|
                                           p
                                                  1
                                            =   √       ; for 0 ≤    ≤ 0.707                    (2.123)
                                               2   1 −    2
   103   104   105   106   107   108   109   110   111   112   113