constant is much longer than the duration of the
impulse.
~2! The length of the brass rod was not sufficient to avoid
reflections off the far end of the rod. The transit time of a
pulse from the upper surface of the piezo to the lower end of
the rod was 30 ms, resulting in a standing wave of period 60
ms or frequency 16.7 kHz. This mode was not excited with
any significant amplitude by any of the balls tested since the
ball contact time was longer than 120 ms in all cases. As a
result, the frequency spectrum of the impulse did not extend
significantly beyond 10 kHz. To avoid reflections off the
table and floor, the rod was isolated from the table with a soft
rubber support, as shown in Fig. 1. Simply holding the rod in
one hand also provided excellent isolation, but the distance
to the laser beams was then not known accurately. In principle,
a much longer rod could have been used to delay the
reflected pulse, but a rod of length at least 10 m would have
been required to avoid the reflected pulse from a tennis ball.
A rod of length about 1.5 m is ideal for studying the impact
of small steel balls, and it also generates textbook examples
of compressional ~nondispersive! and transverse ~strongly
dispersive! wave modes that can be detected with a small
piezo at one or both ends.
~3! A large diameter disk was chosen to avoid saturation
of the force wave form that would occur if the contact area of
the ball exceeded the area of the disk. Even so, measurements
for a tennis ball were restricted to velocities less than
8ms21 since the contact diameter of the ball exceeded 50
mm at ball speeds greater than 8ms21
. In the case of a highspeed
tennis ball, or a large diameter ball such as a basketball,
a piezo larger in diameter than 50 mm would be required. Such piezos are difficult to obtain, but it is relatively
easy to connect any number of small piezos in parallel between
two metal plates, with the same polarity, and bonded
by a very thin layer of epoxy. Piezos extracted from inexpensive
piezo buzzers would be suitable for this purpose.
The lower plate should be quite thick ~40 mm or more! to
avoid transverse oscillations of the structure in the kHz
range, and the upper plate should be relatively thin and light
to minimize the force on the piezos induced by lowfrequency
vibrations transmitted from the soft rubber support
to the upper plate. A suitable plate can be made from doublesided
circuit board, using the upper side as a grounded
shield. Such a system has been constructed by the author to
measure high-speed tennis ball impacts, and the results will
be presented elsewhere.
~4! The combined mass of the piezo and brass rod, 1.8 kg,
was much larger than that of any of the balls tested, so the
energy transferred to the rod was much smaller than the incident
energy of the ball. The momentum transfer was not
entirely negligible, with the result that part of the y displacement
observed at the end of the impact could be attributed to
motion of the brass rod during the impact. The velocity of
the rod after the collision is given by V5m(v11v2)/M,
where m is the ball mass and M51.8 kg is the rod mass.
Since the average speed of the rod during the collision is
approximately V/2, the displacement in time t is approximately
Dy5m(v11v2)t/(2M). This displacement is
shown in the last column of Table I. For the baseball, tennis
ball, and steel ball, motion of the rod accounted for about
half of the final y displacement, and it also accounted for
about 5% of the energy lost by these balls. The results presented
below were not corrected for this effect, in part because
of the unknown effect of the rubber support in restricting
motion of the rod. The displacement of the rod is
significant only toward the end of the impact, and the area of
the hysteresis loop is increased by only a few % as a result.
More precise measurements could be obtained either by using
a heavier rod, or by suspending the rod horizontally to
allow for free motion of the rod during the collision. In the
latter case, an appropriate correction based on the measured
force wave form could then be made for displacement of the
rod.