VII. BOUNCE OF A SOLID ELASTIC SPHERE
A brief summary of the model by Maw, Barber, and Fawcett (MBF) for the oblique bounce of a solid elastic sphere is given here because the theoretical predictions are qualitatively consistent with the observed friction force measurements given in Sec. III. The MBF model is numerical, and the approach is to divide the contact circle into small annuli, some of which grip the surface and some of which slip. Because the component of the normal reaction force acting on the outermost annulus is zero, this and several adjacent annuli usually slip when the ball is subject to a horizontal force, while the inner annuli grip the surface if the coefficient of static friction is sufficiently large. Under conditions where Brody’s model predicts that the ball will slide throughout the bounce period, all annuli slip and the two models are equivalent. The primary difference between Brody’s model and the MBF model is that Brody assumed that the ball would commence rolling when vx5Rv, whereas MBF assume that the whole contact area sticks to the surface when vx5Rv because all points within the contact area come to rest at that instant. In Brody’s model, the friction force drops instantaneously to zero when the ball rolls. In the MBF model the friction force does not change instantaneously because MBF assume that the coefficients of static and sliding friction are equal. The ball deforms elastically in the horizontal direction while it is sliding, and it continues to deform and vibrate in the horizontal direction while the contact area is stuck. However, annuli near the outer edge of the contact area quickly become unstuck and begin to slip because the torque on the ball acts to increase the angular velocity of the ball in those annuli. These annuli slide backwards on the surface, reducing the total friction force on the ball. As time progresses, the annuli in slip spread radially inward, reducing the friction force to zero and then reversing it. Near the end of the bounce period the whole contact area slides backwards on the surface. The differences between the Brody, Garwin, and MBF models for a bounce on an infinitely massive surface are summarized in Fig. 8, which is a plot of the dimensionless quantity b2 versus the dimensionless quantity b1 , where
if the ball slides throughout the impact, while b250 if the ball enters a rolling mode. Garwin assumed that ex51 and hence b252b1 . All theoretical results in Fig. 8 are given for a solid sphere with a 50.4, D50, and ey51 so that all three models can be compared using the same parameters. Garwin and MBF considered only a solid sphere with ey51. Also shown in Fig. 8 are experimental data for the superball and for a golf ball filmed when bouncing at low speed ~about 4 m/s! onto a heavy, polished granite slab. The ball was thrown by hand and was incident with negligible spin. Both balls had similar mass and diameter but very different bounce characteristics. A similar set of measurements for a superball is given by Johnson.3 The experimental data in Fig. 8 were plotted using the measured value m50.18 for the golf ball and an assumed value m51.0 for the superball. The value of m for the golf ball represents an average obtained from several low angle bounces where the ball was sliding throughout the bounce. The superball did not slide under any conditions, even at angles of incidence as low as 12° to the horizontal. Consequently, it was not possible to obtain a reliable estimate of m for the superball. A lower limit of 0.9 can be deduced using Eq. ~4! and a maximum value of 2.4 can reasonably be deduced from the data obtained in Ref. 6 for a rubber disk incident on rubber. Experimentally it was found that ey 50.9760.03 for the superball and ey50.9060.02 for the golf ball, for all angles of incidence. The small departures from the ideal value ey51 are not significant. Similarly, experimental values for D were typically less than 0.5 mm for both the superball and the golf ball and this effect is also not very significant. An alternative view of each theoretical model is given in Fig. 9 which shows the horizontal coefficient of restitution, ex , as a function of the angle of incidence, u 1 . Because ex 52b2 /b1 and b151/(mtan u1) when v150, there is no new information in Fig. 9, but the significance of the angle of incidence and the coefficient of friction is more apparent. The golf ball has a much lower coefficient of friction than the superball, and it slides throughout the bounce at angles of incidence up to about 40°. Equation ~7! with D50 and m/M50 indicates that the golf ball should slide at angles of incidence up to 39.9°, and the superball should slide only when u 1,8.1° if m51. It is difficult to obtain accurate data at such low angles of incidence. A more reliable value for m in this case could be obtained if the ball was incident with significant backspin to allow for measurements of sliding at higher angles of incidence, but this was not attempted. The data on ball spin used to generate Figs. 8 and 9 are shown separately in Fig. 10, together with two simple theoretical estimates. If one assumes that the ball enters a rolling mode, then the spin for a solid sphere is given by Rv2 /v1 5(5/7)cos u 1 , where v1 is the incident speed. Garwin’s model with ex51 indicates that Rv2 /v15(10/7)cos u 1 , twice the rolling value. At large angles of incidence, it can be seen from Fig. 10 that the golf ball spins at a rate that is almost the same as that for a rolling ball. At low angles of incidence, the spin is reduced, consistent with sliding. If the ball slides throughout the bounce, then Rv2 /v152.5m(1 1ey)tan u 1 , and hence v2 approaches zero at glancing incidence. The maximum spin of the golf ball, for a given incident speed, occurs at about u 1540°, which is the angle at which the ball switches from a pure sliding to a biting mode. The superball spins faster than the golf ball, but not as fast as predicted by Garwin. We can conclude that the MBF model provides a better qualitative description of the bounce of a ball than the more elementary models. However, Brody’s model provides a better quantitative description in the case of a golf ball, presumably because the storage and recovery of elastic energy due to tangential compliance is less efficient for the golf ball, giving a value for ex of only about 0.1 when the ball grips. The low value of ex is not simply due to the low coefficient of friction. Maw et al.6 obtained good agreement with their model using a steel disk with m50.115. Maw et al.5 considered the situation where a solid sphere is compressed against a half space of the same material. A similar result would be expected for a solid, elastic ball compressed against a rigid surface. However, a tennis ball that is compressed on a rigid surface behaves differently because the normal reaction force is zero at the center of the contact circle as well as at the edge of the circle. If a tennis ball is pushed onto a surface, the ball buckles in such a way that a central section of the contact region lifts off the surface and protrudes inside the ball. The same effect has been observed during a high speed vertical bounce of a tennis ball, in which case the initial contact area bounces up inside the ball while the rest of the ball continues its initial motion downwards.12 As a result, the contact area is an annulus rather than a com plete circular region. The effect is not as dramatic in a low speed bounce, but the distribution of the normal reaction force for a hollow ball is likely to be quite different from the case in Ref. 5 at least when the diameter of the contact area exceeds the wall thickness of the ball. An additional effect that was not considered by MBF is that N does not necessarily act through the center of mass. In this paper it was found that N acts a small distance behind the center of mass. Experiments currently being undertaken by the author and by colleagues at the University of Sheffield using a tennis ball projected at ball speeds greater than 20 m/s indicate that N can act through a point up to about 12 mm ahead of the center of mass during a high speed impact. This work has not yet been published. The effect is analogous to the shift in weight toward the front of a vehicle when the brakes are suddenly applied. The torque due to the friction force causes the vehicle to rotate about its center of mass. The vehicle would roll over if it were not for the fact that the normal reaction force on the front wheels is then larger than the force on the rear wheels.