Page 419 - Cam Design Handbook
P. 419

THB13  9/19/03  7:56 PM  Page 407

                                CAM SYSTEM DYNAMICS—RESPONSE               407

               For many years much research has been directed toward identifying and tabulating suit-
            able functions to define cam motion segments which produce good dynamic response.
            Neklutin (1952) contributed much in this area.
               Freudenstein (1960) observed that cam dynamics could be improved by minimizing
            the harmonic content of the motion and developed a family of low harmonic curves that
            minimized the peak acceleration.
               The next four sections of this chapter each explore a method for determining optimal
            or at least near-optimal cam curves for any particular set of requirements (mainly lift, oper-
            ating speed, system natural frequency, and duration). These methods allow a designer to
            systematically  obtain  a  near-optimal  cam  curve  directly  for  whatever  system  is  being
            developed. The designer does not have to choose a curve which is known to have good
            vibration characteristics and hope for the best nor must a variety of curves be selected to
            find the best one through analysis.
               Wiederrich (1981) showed, by applying modal analysis methods, that once the cam
            curve has been designed using a single-DOF model to minimize vibration over the entire
            speed range, from zero to maximum, then the actual system, with its many degrees of
            freedom, has also been so optimized (i.e., the vibration of each of the modes has been
            minimized). However, if the one-DOF model is used to optimize only at a single speed or
            over a limited range of speeds, the vibration from the higher modes cannot be assumed to
            be optimized and their combined vibration response could be significant compared to that
            predicted by the single DOF analyzed. Therefore, it is generally recommended that opti-
            mization using a single-DOF model be applied over the entire speed range. Of course, the
            single-DOF model must also satisfy the conditions stated in Sec. 13.1 before any conclu-
            sions can be made as to the adequacy of the design.
               The methods presented are all relatively complex mathematically and thus require solu-
            tion by computer using a program written for that purpose. Once the computer program
            is available, application is relatively easy.




            13.4 DYNAMIC SYNTHESIS OF CAMS USING
            FINITE TRIGONOMETRIC SERIES

            13.4.1 Introduction

            In this section, we present cam design methods originated by Wiederrich and Roth (1978).
            In these methods, follower or cam displacement is defined as a finite trigonometric series.
            One  method  generates  a  “tuned”  cam  design,  in  which  the  cam  profile  is  designed
            to provide exactly the follower lift as a given finite trigonometric series function at one
            particular  cam  speed.  With  this  method,  the  performance  at  other  speeds  can  be  very
            poor. Another method uses a finite trigonometric series function for cam lift and a mean
            square error minimization technique to optimize the cam performance over a range of cam
            speeds.
               The methods employ the usual linear single degree-of-freedom mathematical system
            model. The original 1975 paper provides a mathematical method to estimate the limita-
            tions  of  such  a  model  for  any  particular  case.  Basically,  the  method  shows  the  single
            degree-of-freedom model to be valid when the highest significant frequency of a Fourier
            representation of the lift event is much lower than the second mode natural frequency of
            the system. This is almost always true of a good design, where the highest significant fre-
            quencies are significantly below the first mode natural frequency and therefore far below
            that of the second mode.
   414   415   416   417   418   419   420   421   422   423   424