1: %================================================
2:
3: \documentclass[prb,twocolumn,showpacs,citeautoscript,amsmath]{revtex4}
4:
5: \usepackage{graphicx}
6: \setkeys{Gin}{width=6.6 cm} %figure size
7:
8: \begin{document}
9:
10: \title{Listening to the coefficient of restitution and the
11: gravitational acceleration of a bouncing ball}
12:
13: \author{C. E. Aguiar} \email{carlos@if.ufrj.br}
14: \author{F. Laudares} \email{f_laudares@hotmail.com}
15: \affiliation{Instituto de F{\'\i}sica, Universidade Federal do Rio de Janeiro \\
16: Cx.P.~68528, Rio de Janeiro, 21945-970, RJ, Brasil}
17:
18: \begin{abstract}
19: We show that a well known method for measuring the coefficient of restitution
20: of a bouncing ball can also be used to obtain the gravitational acceleration.
21: \end{abstract}
22:
23: \pacs{01.50.Ht, 01.50.Lc}
24: %\date{\today}
25:
26: \maketitle
27:
28: Three contributions to this journal have described how to measure the coefficient
29: of restitution between a ball and a flat surface using the sound made by
30: the collision of the ball with the surface \cite{Bernstein,Smith,Stensgaard}.
31: The procedure reported in these papers is to drop the ball vertically on a horizontal
32: surface, allow it to bounce several times, while recording the sound produced by the impacts.
33: Analysis of the recording gives the time intervals between successive rebounds, and
34: from these the coefficient of restitution is obtained.
35:
36: The evolution of the techniques described in these papers is a nice example
37: of how the development of microcomputers has changed the science teaching laboratory.
38: In 1977, Bernstein \cite{Bernstein} detected the sound with a microphone,
39: amplified and filtered the signal, and fed it to a pen recorder.
40: Smith, Spencer and Jones \cite{Smith}, in 1981, connected the microphone to
41: a microcomputer via a homemade data collection and interface circuit, and then uploaded the
42: resulting data to a larger computer for analysis and graphical display.
43: In 2001, Stensgaard and L{\ae}gsgaard \cite{Stensgaard} used the microphone input
44: of a PC sound card to make the recording, reducing the experimental equipment to
45: basically a standard microcomputer.
46:
47: To see how the coefficient of restitution $\epsilon$ is related to the time
48: between bounces, note that if $\epsilon$ is constant (independent of velocity),
49: and air resistance is negligible, the velocity of the ball just after the
50: $n$th bounce on the fixed surface is given by
51: \begin{equation}
52: v_n = v_0 \epsilon^n
53: \label{vn}
54: \end{equation}
55: where $v_0$ is the velocity just before the first impact.
56: The time-of-flight $T_n$ between the $n$th and $(n+1)$th collisions is
57: proportional to $v_n$,
58: \begin{equation}
59: T_n = \frac{2 v_n}{g} \; , \quad n=1,2,\dots
60: \label{Tv}
61: \end{equation}
62: where $g$ is the gravitational acceleration. Thus
63: \begin{equation}
64: T_n = T_0 \epsilon^n \;,
65: \label{Tn}
66: \end{equation}
67: where we have defined $T_0 \equiv 2 v_0 / g$ .
68: Taking the logarithm of both sides of Eq.~(\ref{Tn}) we obtain
69: \begin{equation}
70: \log T_n = n \log \epsilon + \log T_0 \; ,
71: \label{logTn}
72: \end{equation}
73: so that the plot of $\log T_n$ vs $n$ is a straight line of slope
74: $\log\epsilon$ and intercept $\log T_0$.
75: Thus, as long as it is independent of velocity, the coefficient of restitution
76: can be obtained by fitting the straight-line of Eq.~(\ref{logTn})
77: to the time-of-flight data.
78:
79: %--------------------------------------------------------------------
80: \begin{figure}[b]
81: \begin{center}
82: \includegraphics{AguiarFig1.eps}
83: \end{center}
84: \caption{
85: The sound of a ball bouncing on a horizontal surface.
86: The zero sound level corresponds to 128 in the vertical axis.}
87: \label{signal}
88: \end{figure}
89: %--------------------------------------------------------------------
90:
91: The purpose of this note is to point out that this straight-line fit can also be
92: used to determine another physical quantity of interest: the gravitational acceleration $g$.
93: The (rather simple) observation is that, if the ball is released from a known
94: height $h$, then $T_0=(8h/g)^{1 / 2}$, and
95: \begin{equation}
96: g = \frac{8h}{{T_0}^2} \; .
97: \label{g}
98: \end{equation}
99: Thus, just as the slope parameter of Eq.~(\ref{logTn}) fixes the coefficient of
100: restitution, the intercept parameter determines the acceleration of gravity
101: (if the easily measured initial height $h$ is known).
102:
103: In order to check how this works in practice, we have dropped a ``superball'' from
104: a measured height onto a smooth stone surface and recorded the sound produced
105: by the successive impacts.
106: The recording was made with the microphone and sound card of a PC
107: running Windows, using the sound recorder program that comes with the
108: operating system.
109: The sampling frequency was 22,050~Hz, resulting in a time resolution of 45~$\mu$s.
110: The audio file, stored in the binary WAV format, was converted to ASCII text format
111: with the shareware program \textsc{awave~audio} \cite{awave}.
112: The recorded signal is plotted in Fig.~\ref{signal}, where the pulses
113: corresponding to individual impacts are easily recognized (only the first
114: six collisions are shown).
115: We have used 8-bit resolution in the recording, so that data values can only go
116: from 0 to 255.
117: The no-signal value corresponds to 128.
118:
119: The time intervals $T_n$ between collisions $n$ and $n+1$ were obtained directly
120: through inspection of the ASCII sound file.
121: They are plotted in Fig.~\ref{tof} (in logarithmic scale) as a function of $n$.
122: The least-squares fit of the $T_n$ data set to Eq.~(\ref{logTn})
123: gives
124: \begin{subequations}
125: \label{fitpars}
126: \begin {eqnarray}
127: && \epsilon = 0.9544 \pm 0.0002 \; ,
128: \\
129: && T_0 = 0.804 \pm 0.001 \mbox{\ s} \; .
130: \end{eqnarray}
131: \end{subequations}
132: The best-fit line is also shown in Fig.~\ref{tof}.
133:
134: %--------------------------------------------------------------------
135: \begin{figure}[t]
136: \begin{center}
137: \includegraphics{AguiarFig2.eps}
138: \end{center}
139: \caption{Time-of-flight $T_n$ between impacts $n$ and $n+1$.
140: The line is the least-squares fit using Eq.~(\ref{logTn}).}
141: \label{tof}
142: \end{figure}
143: %--------------------------------------------------------------------
144:
145: The ball was released from a height $h = 79.4 \pm 0.1$~cm above the surface.
146: Taking this and the adjusted $T_0$ into Eq.~(\ref{g}), we obtain for the
147: gravitational acceleration
148: \[ g=982 \pm 3 \mbox{\ cm/s}^2 \;. \]
149: For comparison, the value of $g$ in Rio de Janeiro is 978.8~cm/s$^2$.
150:
151: The applicability of the method described above depends on $\epsilon$ being
152: constant over the range of impact velocities involved in the experiment.
153: That this condition is satisfied in the present case is seen in Fig.~\ref{epsT},
154: where the coefficient of restitution for an impact at velocity $v_n$,
155: $\epsilon = v_{n+1}/v_n=T_{n+1}/T_n$, is plotted as a function of $T_n$
156: (recall that $v_n \propto T_n$, see Eq.~(\ref{Tv})).
157: The coefficients of restitution for the different impacts are all very close to
158: the least-squares value given in Eq.~(\ref{fitpars}), indicated by the dashed line
159: in Fig.~\ref{epsT}.
160:
161: %--------------------------------------------------------------------
162: \begin{figure}[t]
163: \begin{center}
164: \includegraphics{AguiarFig3.eps}
165: \end{center}
166: \caption{The coefficient of restitution $\epsilon = T_{n+1}/T_n$ as a function of
167: the time-of-flight $T_n$, for the data of Fig.~\ref{tof}.
168: The dashed line indicates the adjusted value given in Eq.~(\ref{fitpars}).}
169: \label{epsT}
170: \end{figure}
171: %--------------------------------------------------------------------
172:
173: %--------------------------------------------------------------------
174: \begin{figure}[b]
175: \begin{center}
176: \includegraphics{AguiarFig4.eps}
177: \end{center}
178: \caption{Time-of-flight $T_n$ between impacts $n$ and $n+1$.
179: The curve is the least-squares fit using Eq.~(\ref{Tn2}).}
180: \label{tof2}
181: \end{figure}
182: %--------------------------------------------------------------------
183:
184: A case in which the coefficient of restitution depends on the velocity is shown in
185: Fig.~\ref{tof2}, where we display the times of flight of a superball dropped from
186: $h = 27.5 \pm 0.1$~cm onto a wood surface.
187: A plot of $\epsilon$ at each collision, shown in Fig.~\ref{epsT2}, reveals
188: a clear dependence of the coefficient of restitution on the time-of-flight
189: (or impact velocity).
190: Assuming a linear relation between $\epsilon$ and $T$, as suggested by
191: Fig.~\ref{epsT2},
192: \begin{equation}
193: \epsilon = \epsilon_0 (1 + \alpha T) \;,
194: \label{linear}
195: \end{equation}
196: we obtain an extension of Eq.~(\ref{Tn})
197: \begin{equation}
198: T_n = T_0 {\epsilon_0}^n \prod_{i=0}^{n-1}(1 + \alpha T_i) \; .
199: \label{Tn2}
200: \end{equation}
201:
202: The least-squares fit of Eq.~(\ref{Tn2}) to the data shown in Fig.~\ref{tof2}
203: gives
204: \begin{subequations}
205: \label{fitpars2}
206: \begin{eqnarray}
207: && \epsilon_0 = 0.921 \pm 0.001 \; ,
208: \\
209: && \alpha = 0.078 \pm 0.003 \mbox{\ s}^{-1} \; ,
210: \\
211: && T_0 = 0.4752 \pm 0.0005 \mbox{\ s} \; .
212: \end{eqnarray}
213: \end{subequations}
214: The curves corresponding to these parameters are also shown in
215: Figs.~\ref{tof2} and \ref{epsT2}.
216: The above value for $T_0$ yields
217: \[
218: g = 974 \pm 5 \mbox{\ cm/s}^2 \;,
219: \]
220: again a very reasonable value.
221: Consideration of the velocity dependence of the coefficient
222: of restitution was important in order to get an accurate result;
223: had we assumed a constant $\epsilon$, we would obtain
224: $g = 935 \pm 10$~cm/s$^2$.
225:
226: %--------------------------------------------------------------------
227: \begin{figure}[t]
228: \begin{center}
229: \includegraphics{AguiarFig5.eps}
230: \end{center}
231: \caption{The coefficient of restitution $\epsilon = T_{n+1}/T_n$ as a function
232: of $T_n$, for the data of Fig.~\ref{tof2}.
233: The dashed line is the linear relation of Eq.~(\ref{linear})
234: with the adjusted parameters given in Eq.~(\ref{fitpars2}).}
235: \label{epsT2}
236: \end{figure}
237: %--------------------------------------------------------------------
238:
239: To summarize, we have seen that the value of the gravitational acceleration
240: is a useful by-product of experiments devised to ``hear'' the coefficient of
241: restitution of a bouncing ball.
242: The measurement of $g$ is particularly simple if the coefficient of restitution
243: is independent of the impact velocity, but more complicated cases can also be
244: handled.
245:
246: \
247:
248: After this work was completed we learned of a recent paper by Cavalcante \emph{et al.}
249: \cite{Cavalcante}, in which $g$ was measured using the sound of a bouncing ball.
250: The analysis presented in the paper is, however, somewhat different from ours.
251: Another related reference is the article by Guercio and Zanetti \cite{Guercio}
252: in this journal.
253:
254: %====================================================================
255:
256: \begin{thebibliography}{99}
257:
258: \bibitem{Bernstein} A.D. Bernstein,
259: ``Listening to the coefficient of restitution,''
260: Am. J. Phys. \textbf{45}, 41-44 (1977).
261:
262: \bibitem{Smith} P.A. Smith, C.D. Spencer, and D.E. Jones,
263: ``Microcomputer listens to the coefficient of restitution,''
264: Am. J. Phys. \textbf{49}, 136-140 (1981).
265:
266: \bibitem{Stensgaard} I. Stensgaard and E. L{\ae}gsgaard,
267: ``Listening to the coefficient of restitution - revisited,''
268: Am. J. Phys. \textbf{69}, 301-305 (2001).
269:
270: \bibitem{awave} AwaveAudio (audio file format converter),
271: FMJ-Software, \url{http://www.fmjsoft.com}
272:
273: \bibitem{Cavalcante}M.A. Cavalcante, E. Silva, R. Prado, and R. Haag,
274: ``O estudo de colis\~oes atrav\'es do som,''
275: Revista Brasileira de Ensino de F{\'\i}sica \textbf{24}, 150-157 (2002).
276:
277: \bibitem{Guercio} G. Guercio and V. Zanetti,
278: ``Determination of gravitational acceleration using a rubber ball,''
279: Am. J. Phys. \textbf{55}, 59-63 (1987).
280:
281: \end{thebibliography}
282:
283: \end{document}