IV. BALL SPEED AND SPIN MEASUREMENTS
Each of the bounces described in Sec. III was filmed with a video camera to obtain simultaneous measurements of the ball speed and spin and the speed of the block. In all cases the change in the horizontal momentum of the ball was equal to the momentum of the block after the collision (within experimental error) as expected. The change in horizontal momentum of the ball is equal to the time integral of F. This information was used to calibrate the response of the piezo disk used to record the F waveform. Similarly, the piezo blocks used to record the N waveform were calibrated by using the fact that the time integral of N is equal to the change in vertical momentum of the ball. A summary of the information obtained for each bounce is given in Table II. The angle u 1 is the angle between the incident ball and the horizontal, v1 is the incident ball speed, v1 is the angular velocity of the incident ball, and v2 is the angular velocity of the ball after bouncing. The horizontal velocity vxR5vx22V2 is defined with respect to the horizontal velocity V2 of the block after the bounce. The ratio Rv2 /vxR equals 1.0 if the ball commences to roll during the bounce and is less than 1.0 if it slides throughout the bounce. A value Rv2 /vxR.1 implies that the ball slides backwards on the surface at the end of the bounce period. In Table II, the quantity D(mv)5m(vx12vx2) is the change in horizontal momentum of the ball and D(Iv) 5I(v22v1) is the change in angular momentum of the ball. In theory, the change in angular momentum of the ball is given by the time integral of the torque FR. For the low speed impacts studied in this paper, R is essentially constant. A video of the ball in contact with the block indicated that R decreased by no more than 3 or 4 mm during the bounce, which is consistent with the relatively small values of the normal reaction force and the known stiffness of each ball type. However, the quantity R*F dt was consistently less than the change in angular momentum of the ball, typically by 30 or 40% ~except for the superball where it was only 3% smaller!. Allowing R to be smaller than the actual ball radius makes the difference even greater. This result indicates that the torque acting on the ball is significantly larger than RF. Because the only other force acting on the ball is N, the normal reaction force, there must be a time interval during the impact when N acts vertically through a line passing a distance D behind the center of mass. The additional torque ND'0.3FR, where F is typically about N/4 so D is typically about R/10. Such a result indicates that the ball tends to lean forward during the bounce. A tennis ball served with heavy topspin is called a ‘‘kick’’ serve because the ball bounces at a steep angle off the court and bounces typically around head height. Players and commentators often remark that the ball ‘‘really bites’’ in this situation, but the results in Fig. 4 and in Table II are not consistent with this interpretation. The ball did not bounce steeply and it did not grip strongly. In a kick serve the ball is incident at much higher speed than in Fig. 4 and spins much faster. The ball will kick up at a steep angle if there is large decrease in the horizontal velocity or if the vertical coeffi- cient of restitution is enhanced. Our evidence shows that the horizontal velocity is not decreased substantially when the ball is incident with significant topspin. The fact that the ball bounces to around head height indicates that the vertical coefficient of restitution is enhanced under conditions where the ball rotates by a significant fraction of a revolution during the bounce.4 In Fig. 4 the ball rotated by only 25° during the bounce.