Investigation 8 (annotated)

The exponential nature of a bouncing ping-pong ball EX The title tells us all.

I learned an amazing story about exponential grow in my mathematics class. There is a

fascinated geometric phenomena described in the legend of a king and a chessboard. PE It is clear that the student loves the
mathematics of exponential change.
“According to legend, a courtier presented the Persian king with a beautiful, hand made chessboard. The Introducing the report with the legend of a
king and chessboard pulls the reader in.
king asked what he would like in return for his gift and the courtier surprised the king by asking for one
This is a positive aspect of Communications.
grain of rice on the first square, two grains on the second, four grains on the third, etc. The king readily
agreed and asked for the rice to be brought. All went well at first, but the requirement for 2n–1 grains on
the nth square demanded over a million grains on the 21st square, more than a million-million (or
trillion) on the 41st and there simply was not enough rice in the whole world for the final squares.”
http://www.dr-mikes-math-games-for-kids.com/rice-and-chessboard.html

Exponential Growth and Decay

The chessboard story is an example of exponential growth. This occurs when the EX Exponential growth and decay are
explained correctly and in a way relevant to
growth rate of a mathematical function is proportional to the function’s current value. This the study.
is also known as geometric growth. Similarly, there is exponential decay when the rate of
decrease of a mathematical function is proportional to the function’s current value.
http://en.wikipedia.org/wiki/Exponential_growth/
Consider a barrel holding 100 gallons of water. A leak occurs and after 5 minutes only C A clear and direct illustration, and the
technical explanation is good too.
50 gallons remain. After an additional 5 minutes the 50 gallons is reduced to 25 gallons.
And, after another 5 minutes, the 25 gallons is reduced to 12.5 gallons. This continues with
the same mathematical pattern, that is, half the remaining water leaves the barrel in the
same amount of time no matter how much or how little water is in the barrel.
Exponential growth and decay are fundamental properties in many aspects of nature. In EX The concept is nicely explained and
illustrated, and connections are made to
biology this occurs with microorganisms in a culture, with the growth of a virus, sometimes various areas of science.
in human population growth. In the field of economics exponential change occurs in
finance; and it is found in computer technology. In physics the charge and discharge of
capacitors follow exponential functions, and this relationship is well known in nuclear
reactions and radioactive decay, and in heat transfer. Many other areas of natural
phenomena can be described as exponential. That is why I love this topic! PE It sure sounds nice. The student can
explore both hands on and modeling with a
topic they love.
Exponential Decay in This Investigation EX Nicely stated. We know what to expect
In this investigation I will consider the exponential decay of a bouncing ball. The in this investigation.

exponential mathematical model of the bouncing ball will be confirmed, although I did not
know this before the experiment. The bouncing half-life (the number of bounces before
the rebound height reaches one-half a previous height) will also be determined.
To clearly focus this approach to the motion of a bouncing ball I will consider only
vertical bounces. The drop height will also limit the investigation; I use only one initial drop
height and take only one set of data. Too high of an initial drop introduces noticeable spin
and too low an initial drop does not provide enough data.

Physics teacher support material 1

 Investigation 8 (annotated)

Bouncing Balls C The student divides the research project

into manageable sections, and each flows
Under normal bouncing conditions (a reasonable height and a uniform surface) balls nicely one to the next. Not a word is wasted.
rebound to a height that is less than the drop height because kinetic energy is lost on the
rebound. The ball never experiences an elastic collision. Some heat and sound are always
produced. We also know that there is a characteristic unique to a given ball for the amount
of height lost between the drop height and the rebound height. This might be called the
‘bounciness’ or the percentage of rebound. It is related to the coefficient of restitution.
These percentages of rebound express an exponential property for a bouncing ball. A
golf ball, for instance, will bounce back to 36% of the drop height. See the information
found on the web site: http://www.exploratorium.edu/baseball/bouncing_balls.html. Each
consecutive bounce will reach 36% of the previous height, and so on.

Preliminary Work EX The methodology is well thought out.

The student knows what they are doing.
Ball Selection. A variety of balls were found in the physics lab, and each was dropped
and the bouncing was observed, both in terms of keeping along the vertical and making a
clear sound or well-defined sound upon impact. I decided that a Ping-Pong ball was the
best ball to use.
Drop Height. The Ping-Pong ball dropped from about 60 cm made clear impact sounds
and produced sufficient repetitive bouncing. This value was arbitrary and does not play a
role any of my calculations, only for collecting sufficient data.
Rebound Height. This will be calculated based on the time interval between two
consecutive bounces.
Time Measurement. The sound of the bounce impacts is recorded using a microphone,
data logging interface and a computer. I used the Vernier (see http://www.vernier.com/) Lab
Pro interface and the Logger Pro software with my Mac computer. This is illustrated in
below in Graph 3.
Percentage of Rebound Test. A preliminary test was made by dropping the Ping-Pong
ball from a variety of heights and measuring the rebound height using the Vernier sonic
motion detector. If the percentage of rebound is consistent then the exponential model may
well be appropriate. The results are shown below in Graph 1.
GRAPH 1: Rebound Height against Drop Height

Physics teacher support material 2

 Investigation 8 (annotated)

The data reveals a linear function with a gradient of 0.853. This is the same as saying that
the ball experiences 85.3% efficiency on its bounce for a variety of drops heights. The
systematic shift in all the data is simply an artifact of the measuring technique and does not
affect the gradient value. It is safe, therefore, to assume that this rebound percentage is a
characteristic of the Ping-Pong ball and hence my exponential model is appropriate.

Experimental Design
The purpose of this investigation is to confirm the exponential nature of a EX A perfect research question: well
defined, clear and a most manageable and
bouncing ball and thereby calculate the half-life of the bouncing ball. I will limit my interesting investigation.
study to one type of ball, and I will limit the study to just one initial drop height and the
resulting bouncing heights. Because the ball’s rebound height decreases with time this is
the same as having a number of different drop heights in sequential order; that is to say, for
example, that rebound number 4 is the drop height of bounce number 5, and so on.
Using the microphone, the LabPro data-logging interface and computer system, the
sound of bounces was recorded. Heights were calculated by the spreadsheet. The ball
bounce events were simply counted; the first bounce as n = 1, the second was n = 2, etc.
When a rebound height became one-half the height of some previous height then it will
have bounced a number of times n (not necessarily a whole number, perhaps n = 3.33)
indicating its half-life. With radioactive decay, the half-life is a function of time (seconds,
minutes, years) but with a bouncing ball the event of a bounce is the measure.

Experimental Variables EX Every detail in the method and analysis

is explained and justified. Limitations and
In order to determine the consecutive heights I recorded the sound of the bouncing ball. controlled variables are appreciated too.
Upon impact, the ball made an impact sound. This is the same as saying the air pressure is
constant while there is no sound, but changes or is offset from this constant when there is
sound. Sound is longitudinal wave, a compression and rarefaction of air pressure.
The independent variable is the bounce number. The life of a bouncing ball is
measured as the 1st, 2nd, 3rd, etc. bounce number. This is a counting number, a pure number
with no units and no significant uncertainties. However, identifying the moment of impact
is limited to the precision of the sampling rate, so there is a minimum uncertainty, about 1
ms in this study.
The dependent variable is the rebound height, H, the maximum height reached
between bounces. This was calculated in units of metres (m). To measure H, the time
between consecutive bounces was determined (from a graph of air pressure offset as a
function of time) and H calculated from where . The comes from the

fact that is the time up to the rebound height plus the time down from the rebound
height. It is far more accurate to measure this time interval, and then calculate the height,
than it would be to try to measure the rebound height of a moving bouncing ball. There is
no significant uncertainty in the calculated height as it is based on a very precise timing

Physics teacher support material 3

 Investigation 8 (annotated)

mechanism with the computer and interface. The offset of air pressure is recorded at a rate
of 1000 measurements per second.
The controlled variables included using the same surface and the same ball. This was A The variables are all appreciated and the
quality of data is determined. The student
obvious as I made only one data set. Time limited repeated measures. Nonetheless, I made does a most thorough job at preparing for
a number of trials and selected the cleanest data set, and one where the bouncing was more analysis.
or less always along the vertical. This is good scientific practice thus demonstrating the
nature of science at work. If the ball moved off the vertical while bouncing, then the data
was rejected. A controlled variable was that the bouncing stays more or less along the
vertical.

Four Experimental Assumptions EX This is very good style and the detail
again demonstrate that the student knows
(1) The evidence for exponential decay has already been discussed and it is a reasonable exactly what they are doing.
model for the bouncing ball. This was confirmed in Graph 1. I then select one drop
height and let the ball bounce again and again and then see if this motion is indeed
exponential. This is the first mathematical model assumption I made.
(2) The value of free-fall gravity will be assumed in the calculations. If this is in error it will
be a systematic error, a constant, and hence will make no difference to the results.
(3) It is true by definition that the distance up equals the distance down. However, if the
ball spins and moves off a purely vertical path then the data trail was rejected. Again,
in this experiment a number of trials were done but only one set of data was used to
determine the results.
(4) If the distance up equals the distance down then assuming uniform acceleration we
can say that the time up equals the time down. This assumes that air friction plays no
role in the motion of the ball. From a practical point of this, air friction can be ignored,
but technically the time up is not equal to the time down if we account for air friction.
On the upward journey the weight is directed downward and the air resistance retards
the motion; air friction is directed downwards. The net force causing the ball to
decelerate is then Fweight + Fair. When the ball falls from the maximum height to the
ground the weight is again directed downwards but the air resistance, which retards
the ball’s motion, is directed upwards. Hence the force causing the ball to accelerate is
Fweight – Fair. This means that the time going up is less than the time coming down when
air friction is accounted for. In my experiment the balls maximum speed is rather small,
and the height distance is small too, and I assume air resistance is negligible. Moreover,
a test with the sonic motion detector (Vernier) revealed that any asymmetry between
the up and down times is equal the period of the sampling frequency. That is, no
difference is detected other than the precision of 1 ms and any noise or random effects,
artifacts of the measuring process. My second mathematical model assumption then EV Assumptions are an important part of
an investigation. The student appreciates
is the basic equation of uniform accelerated motion, relating height h, gravity g and the nature of science here.
time t, as .

Physics teacher support material 4

 Investigation 8 (annotated)

Gathering Data
The time is determined by recording the impact sound of a bouncing ball. Time
intervals are read off a graph of sound pressure against time.
FIGURE 2: Sample of Raw Data
This shows a sample of the raw data of the air
pressure (in arbitrary units), the offset pressure when
there is sound, and time measurements (in seconds).
The microphone was connected to the Vernier Lab
Pro interface and then this was connected to the
computer. The Vernier’s LoggerPro 3.4.1 software
automatically sensed the microphone and displayed
graph axes of sound level against time. The sampling
rate was set at 1000 Hz. http://www.vernier.com/.
The sampling rate was set at 1000 Hz and I took 6 seconds of data. Only a spreadsheet
could handle all this data. See the last datum at number 6000 in the figure below. A Only a spreadsheet can handle this much
number crunching.
FIGURE 3: Last Few Data Values

Close-Up View of Graphed Data

Graph 2: Close Up of Air Pressure Offset against Time

This is a close-up view of the next graph. By clicking on a peak (maximum or minimum A Nice detail here and in other places
demonstrates the student’s competence in
intensity offset) the time datum is given. Although there are several data points before the this project.
peak (where the ball first hits the table) I used the first maximum or minimum peak
(greatest offset intensity) in all the time measurements.

Graphing Data
The following graph (Graph 3) includes all 6000 data points.

Physics teacher support material 5

 Investigation 8 (annotated)

GRAPH 3:
Graph of Offset Sound Pressure (arbitrary units) against Time (s)

The sharp peaks, both positive and negative, indicate a pulse of sound; they represent
the sound made at impact. I decided to use the largest datum value as the impact time. Even
if there is a systematic error here, using the same reference point for all measurements of
time makes this assumption reliable. The computer calculates the consecutive times for the
first peaks of each bounce and uses this to calculate the rebound height H and the natural
logarithm of the height H. A value of is used but as this is a constant through
out the experiment, it could have been normalized, i.e. .

Processing Data
The processing of data was done in the spreadsheet automatically once I programmed
the calculations. The Logger Pro graphing program nicely determined consecutive values as
well as doing the various calculations.
FIGURE 4: Spreadsheet Processed Data

Physics teacher support material 6

 Investigation 8 (annotated)

The following shows of how the spreadsheet made the calculations. I use just one set of A The student is focused and clear on what
is needed here. It is a rather clever
values to illustrate this. technique for determining the rebound
Delta Time Calculation: heights.

For example, the interval from N = 3 to N = 4 is as follows.

Rebound Height Calculation:

For example, H for the interval is as follows.

Natural Logarithm of Height Calculation:

Analysis of Data
The calculated heights are now graphed against bounce number. You can see the
gradient decreases as the bounce number increases. This could be exponential.
GRAPH 4: Height (m) against Bounce Number

From Graph 3 we can see, for example, that H reduces from 0.10 m to 0.05 m in about 5
bounces (4.9) and then from 0.05 to 0.025 in a further 5 bounces, thus indicating an
exponential decay. That is, changes from to have equal bounce intervals, the half-
life of the bouncing ball.

Physics teacher support material 7

 Investigation 8 (annotated)

Assume the mathematical model where and is the decay constant and n
is the bounce number so that a plot of the natural logarithm of H against n should give a
straight line graph the gradient of which is . This graph (Graph 4) is plotted below.

Logarithmic Graph

GRAPH 4: Natural Logarithm of Height against Bounce Number

The best-fit line is clearly linear indicating an exponential relationship. Note that both A The student could have used the
statistical software in the graphing
logarithms and counting numbers have no units; they are pure numbers. software to determine an uncertainty in the
gradient of the graph. So many other areas
were given attention by the student that the
Half-Life Analysis moderator here does not think this
omission is significant. The results stand on
The computer calculated the gradient m of the graph where . The half- their own.
life is calculated from .

There is no reason to expect the half-life of height to be a whole number of bounces. In

fact, as the ball continues to bounce the time between bounces decreases. This decay in
time is not exponential. A graph of the natural logarithm of height against time reveals this
but is not included here.

Uncertainty in The Exponential Model

Instead of propagating the errors based on precision and random estimates, I decided EV The student is justified in this approach.
The research question still ends up with a
to graph the characteristic that I assumed in my exponential model, namely the ratio of degree of uncertainty.
rebound against bounce number, and then find a value of uncertainty from the scatter of
these values (one unique value per bounce event, see Graph 5 below). The average of this
can easily be determined and an uncertainty for the average can be found from the range of
ratio values.

Physics teacher support material 8

 Investigation 8 (annotated)

GRAPH 5: Rebound Ratio against Bounce Number

The LoggerPro software statistical analysis function tells us the minimum and the
maximum values of this rebound ratio. It also tells us the average or mean value, which is
0.861375 or 86.1%.
The uncertainty in this can be found using the range of date as follows, where y is the A Uncertainties are appreciated where
relevant. See, for example, in the conclusion.
mean value of the ratio.

As a percentage, we find:

The percentage of rebound and its uncertainty are thus (87±4)%. However, if we take a EV A degree of confidence is appropriately
determined here.
statistical approach and use standard deviation, we get a ratio and its uncertainty as
0.86 ± 0.02 or only 2.3%. When rounded off this is about 2%, which is half the uncertainty
when calculated by the range method. I can assume that this limit is due to random errors
as well as less than ideal exponential motion (air resistance, slight spin, etc.).
In both sets of uncertainty calculations I assume that under perfect conditions the
exponential model applies. My assumptions has been confirmed to a high degree but not C The student is reflective and appreciates
what is being done. The language is clear
proved. Recall the 38 arc-seconds difference between Newton’s and Einstein’s theories of and direct.
gravity when it came to the perihelion precession of the planet Mercury. Small details can
make big differences in theory. And now, for my Conclusions and Evaluative comments.

Physics teacher support material 9

 Investigation 8 (annotated)

The Exponential Model. The experimental results make a beautiful fit of the exponential
model. The consecutive decreases in rebound heights follow the model, and this was estimated as
being within about 2% uncertainty, a most acceptable conclusion.
The Half-Life of a Bouncing Ball. Following the exponential trend, the half-life of a
bouncing ball (the number of bounces for the rebound to become half an initial value) was
determined to be about 4.84 bounces. No uncertainties were determined for this.
Obtaining Results. The measurements would not have been possible without the data-
logging and a computer as well as the data-processing using a spreadsheet. Using a meter
stick and eyeing the bouncing ball would have been impossible.
Limitations. One limitation in this experiment is that there was insufficient time to take EV The student mentions relevant and
realistic details for improvements.
more data in the form of repeated trials from the same initial drop height. No doubt the data
would vary slightly due to possible ball spin produced by the release method, slightly different
impact conditions (perhaps in a repeated trail the ball is every so slightly warmer), and so on. I
doubt if the results would be different.
One advantage of repeated trials would be to establish a degree of confidence, an uncertainty
in the results by error propagation method. More data would have helped to test the validity and
or limitations of the exponential model. I could have also tested this model’s validity for balls of
different material dropped on different surfaces.
An important improvement would be to record more precise intervals of sound,
perhaps changing the sample rate from 1000 Hz to 2000 Hz. Also, assuming slight spin
caused the slight variation in the data, some sort of mechanical release mechanism might
help improve the already high degree of precision.
My preliminary investigations showed me that much higher drop heights (as well as
very low drop heights) do not produce quality data, so my conclusions in this investigation
are limited to the scope and range that I covered.
Further studies, or an extension to this, would be to determine the half-life of different
types of balls. Now that I have thoroughly analyzed the Ping-Pong ball, I a happy to leave
my study there.
Overall, I was amazed that I could clearly demonstrate the exponential nature of a PE Sure enough, the student was involved
here.
bouncing ball, just one more fascinating example of this mathematical function found in the
natural world.

10

Physics teacher support material 10