Page 308 - Cam Design Handbook

THB10 9/19/03 7:28 PM Page 296

296 CAM DESIGN HANDBOOK

cams has in some instances reduced wear. However, it has been observed that cam surface
finishes better than 15minrms can induce errors in the cam profile that will affect cam
dynamic loading (acceleration curve shape) and as a result nullify the result of the improved
surface finish. The limit on finishing can be determined by experience and by cost.
Finishing of cams is the most critical of the fabrication operations. The basic point for
consideration is how smooth the surface must be for the ultimate fatigue life and follower
performance. This subject was discussed in Chap. 9.
A milled cam surface finish can be held to 125minrms. Grinding is the most accurate
cam-finishing operation and surface finishes as fine as 15minrms can be obtained.
Figure 10.5 shows a DRD ground cam acceleration curve with high-frequency oscillations
at points of maximum acceleration. Barkan (1953) showed that analog ground cams tend
to have high-frequency inaccuracies. However, these high-frequency inaccuracies have
very little effect on the dynamic action of the cam-follower system at moderate speeds.
Roller followers have been finished as smooth as 4minrms. Polishing of automotive
sliding followers as high as 50minrms is acceptable. Lapping is sometimes utilized to
correct minute heat treatment distortion errors in multiple hardened cams. An abrasive
lapping compound is used to fit mating cams and followers. Sometimes if a surface needs
improvement a shear polish, lapping operation, or controlled run-in period is employed,
all limited by cost.
Machine frames should be chosen to be as rigid as possible so as not to detract in any
way from the proper, accurate function of the cam-follower mechanism. If the frame is
“soft,” it may store and release energy in every cycle of operation.
Next, let us discuss the cam curve mathematical analysis as it relates to the accuracy
of manufacturing the actual cam. It should be noted that there is a limit in design to control
the higher order derivatives of the cam curve. In cam analysis, it is conventional to use
the smooth acceleration curve y as a basis for analysis. The curve choice is selected accord-
¨
ingly. This is compatible with the usual manufacturing tolerance of ±0.001in. It has been
shown by Kim and Newcombe (1983), Barkan (1953), and Weiderrich (1973) that extreme
manufacturing limits of ±0.0001in tolerances cannot control the jerk. In general, for indus-
trial production machinery the jerk curve is used basically as a guide and not as a design
tool. Hence, given normal manufacturer tolerance, it is not possible to control higher order
deviation. Thus, it is not significant to elaborate on the jerk curve since it cannot be con-
trolled in manufacturing easily.
Last, let us discuss the total machine tolerance analysis to reduce the accumulated error
in the system. This method consists of a weighting factor for each of the connected parts
of the machine. In this manner the engineer can determine and control the tolerance of
each member that contributes most to the error. The manufacturing tolerances can then be
modified to improve the machine performance. Sometimes a change in manufacturing
methods is helpful.

10.5 CAM SHAPES VERSUS
ACCELERATION CURVE

A cam may appear smooth to the touch or sight and yet be composed of vibratory imper-
fections. It may be dimensionally acceptable yet dynamically poor. In this section, the
blending of the cam profile shapes will be observed to establish their effect on the cam
.
¨
velocity y and acceleration y curves.
In Fig. 10.10a, a cam is cut from a circle that has a 0.0001in flat sliced from its surface.
.
¨
As the cam rotates on the circular portion, the velocity y = 0 and acceleration y = 0. When
the follower reaches point a it receives a velocity shock since the velocity instantaneously