Page 418 - Cam Design Handbook
P. 418

THB13  9/19/03  7:56 PM  Page 406

          406                      CAM DESIGN HANDBOOK


                2.0         11th-degree polynomial
                             Cycloidal
                                Berzak-Freudenstein 3-4-5-6-7
             Envelope of normalized maximum residual acceleration response  1.0  Gutman F-3  Modified sine  Berzak-Freudenstein
                                Polynomial-E curve
                1.5
                                Freudenstein 1-3-5 harmonic





                                                     Polynomial-D curve
                0.5
                               3-4-5 Polynomial
                      4-5-6-7 Polynomial                     Modified trapezoidal
                 0
                  2        4       6       8       10       12
                                       T / T N
                                        1
           FIGURE  13.5.  Comparison  of  normalized  peak  acceleration  magnitude  envelopes  for  residual
           vibration.



             No single cam form is best for all applications. A form that is optimized for its response
          spectrum in certain frequency regions will not be optimized for response in others. Other
          complicating factors that are difficult to account for are variations in external loads, torque
          fluctuations, fabrication errors, and alignment and other inherent factors. Good perform-
          ance in the field is the final test that the system design is a good one.



          13.3 VIBRATION MINIMIZATION

          13.3.1 Introduction to Vibration Minimization
          Many  investigators  have  studied  optimization  criteria  for  cam  profile  design.  Their
          methods minimized many important response parameters, particularly vibration of the cam
          follower.  One  of  the  first  presentations  of  a  form  of  vibration  optimization  was  by
          Hussmann (1938). He established specific harmonics near the follower natural frequency
          equal to zero to minimize vibrations. Chew and Chuang (1990) applied Lagrange multi-
          pliers and polynomial lift curves to minimize the integral of the end of the rise residual
          vibrations over the desired speed range. They developed a direct procedure for minimiz-
          ing residual vibrations when designing cam motions. They concluded (as did Wiederrich
          and Roth, 1978) that for high-speed applications, specification of vanishing cam bound-
          ary conditions for derivatives higher than the velocity is inappropriate when using an opti-
          mization procedure that accounts for the dynamic response. Perhaps the most important
          feature of these optimization methods in cam design is that they can readily be applied to
          design the entire cam motion rather than just a segment of the motion. Cams designed in
          this way have been found to work well in practice. Optimization methods can also be
          applied to optimize the geometry of the cam and follower mechanism.
   413   414   415   416   417   418   419   420   421   422   423