INSTITUTE OF PHYSICS PUBLISHING EUROPEAN JOURNAL OF PHYSICS

Eur. J. Phys. 26 (2005) 651–655 doi:10.1088/0143-0807/26/4/011

Real-world damping of a physical

E-mail: [email protected]

Received 11 February 2005, in final form 4 April 2005

Abstract
Damped periodic motion is ubiquitous in the physical world and is a subject
of study at all levels of undergraduate education. In this paper we investigate
the damping of a metre stick acting as a physical pendulum subject to air drag.
We do not limit our investigation to small angles and find that the air drag is
well described by a retarding torque equal to a term proportional to the angular
velocity together with a term proportional to the square on the angular velocity.
The study is made possible by the use of a video camera, video capture and
analysis software and an easy-to-use intuitive, icon-based, simulation program
to numerically solve the equation of motion. Suggestions are made for further
study.

Introduction

Generally the treatment of damped periodic motion given in undergraduate textbooks is limited
to an analytically solvable model based on the simple harmonic oscillator with damping
proportional to velocity [1, 2]. Another model that has been treated fairly extensively, with
varying degrees of sophistication in the undergraduate literature, is the simple harmonic
oscillator subjected to so-called dry damping (a constant damping force) [3].
Although it is possible to construct laboratory systems of varying complexity, which will
behave in accordance with these analytically solvable models (see for example [4]), real-
world oscillating systems, in and of themselves, seldom behave in accordance with these
simple models [5, 6].
With the increasing availability of computers and computer-related technology it has
become feasible for undergraduates to investigate, both experimentally and theoretically, the
behaviour of real-world damped periodic motion in far more detail than was possible even a
few years ago.
As an example, we show in the present paper, the results of a study of the behaviour of
a simple physical pendulum constructed from a metre stick, subjected to damping from the
0143-0807/05/040651+05$30.00

c 2005 IOP Publishing Ltd Printed in the UK 651
652 M E Bacon and D D Nguyen

(a)
FV
A
FH

τD

mg
(b)

Figure 1. (a) Schematic of experimental set-up. (b) Free body diagram.

surrounding air. These experiments can be easily repeated and expanded on by undergraduates
and are invaluable in honing the skills necessary for extracting detail from experimental data.
In the next section we briefly outline the basic theory. This is followed by a discussion of
the experiment and a presentation of the results and conclusions in the final section, together
with some suggestions for further study.

Theory

Consider a metre stick (length L) pivoted at one end (point A) and free to move in a vertical plane
as shown schematically in figure 1(a). The free body diagram including a phenomenological
damping torque (τD ) is shown in figure 1(b).
Since the metre stick is rotating about a fixed axis, we are free to use the equation


τ = I θ̈ , (1)
A
where the τ s are the torques of the forces evaluated about the point A, I is the moment of
inertia of the metre stick about A and θ̈ is the angular acceleration of the metre stick.
Applying this equation to the situation depicted in the free body diagram yields (with the
sign convention that counterclockwise torques are positive)
τD − (mgL/2) sin θ = (1/3)mL2 θ̈ . (2)
Simplifying yields
θ̈ = −(1.5g/L) sin θ + D (3)
where D = 3τD /(mL ). In elementary treatments of the physical pendulum θ is assumed
2

small and therefore sin θ ≈ θ , and D is set equal to zero. With these assumptions
equation (3) reduces to the equation for the simple harmonic oscillator and is analytically
solvable.
Real-world damping of a physical pendulum 653

1.50
Damping proportional to angular velocity
1.00
c1 = 0.55, c 2 = 0

Angle (radians)
c1 = 0.32, c 2 = 0
0.50

0.00

-0.50
data

-1.00
0.0 2.0 4.0 6.0 8.0 10.0 12.0
Time (s)

Figure 2. Data for an initial angle of 1.01 radians (solid dots). Theoretical curves for damping
proportional to θ̇. Coefficient c2 = 0 and c1 as indicated.

In more advanced treatments [2], air drag proportional to the velocity is included and
again the problem can be solved analytically provided θ is limited to small values.
In the present study we assume that in general [7]
D = −c1 θ̇ − c2 θ̇ |θ̇|, (4)
and keep sin θ , so that we are not limited to small angles. The negative signs are included
because θ̇ is negative for the situation shown in figure 1(b). A little thought will indicate that
the negative sign remains valid in general for all θ as the pendulum oscillates back and forth.
Equation (3) can then be written as
θ̈ = −(1.5g/L) sin θ − c1 θ̇ − c2 θ̇|θ̇ |. (5)
Equation (5) does not have an analytic solution and we must use numerical methods to
solve for the motion. We have used the icon-based software package STELLA1 to perform the
numerical calculations, since it is easy to use, requires no knowledge of differential equations
or programming languages and the learning curve is relatively short.

Experiment

A 1/8 hole was drilled through the narrow side of a metre stick and a well-lubricated rod
inserted through the hole to form an axis of rotation in order to ensure that any drag due to
pivot resistance would be small relative to air drag. We have tried to establish this by using
different diameter rods and varying the lubrication. The metre stick was displaced through
varying angles (θ ) from the vertical and released. Both small and large angle oscillations were
investigated. For each initial angle of displacement 8 to 10 periods were recorded using a
JVC GR-DV2000U digital video camera. The video was then transferred to a Pentium-based
computer via a fire wire connection using dvvcap1.01 and analysed using videopoint2 . Typical
data for a large initial angular displacement (1.01 radians) are shown in figures 2–4.
1 Available from isee Systems (formerly HPS): 46 Centerra Parkway, Suite 2000 Lebanon, NH 03766 (www.hps-
inc.com).
2 Available from Pasco Scientific: 10101 Foothills Blvd., Roseville, CA 95747 (www.pasco.com).
654 M E Bacon and D D Nguyen

1.50
Damping proportional to square on the angular velocity

1.00 c1 = 0, c 2 = 0.25

Angle (radians)
c1 = 0, c2 = 0.2
0.50

0.00

-0.50
data
-1.00
0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0
Time (s)

Figure 3. Data for an initial angle of 1.01 radians (solid dots). Theoretical curves for damping
proportional to θ̇ 2 . Coefficient c1 = 0 and c2 as indicated.

1.50
Damping proportional to angular velocity and angular velocity squared
c1 = 0.10, c 2 = 0.17
1.00
Angle (radians)

0.50

0.00

-0.50

-1.00
0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0
Time (s)

Figure 4. Data for an initial angle of 1.01 radians (solid dots). The solid line is the theoretical
curve for damping equal to −0.10θ̇ − 0.17θ̇ 2 .

The behaviour of the metre stick was simulated using STELLA and equation (5). Initially
c2 was set equal to zero (damping proportional to θ̇ only) and c1 varied to see if the data were
compatible with damping proportional to θ̇ . Figure 2 compares the experimental data with the
theoretical curves for c1 = 0.32 and 0.55. From the figure we can see that with c1 = 0.32 the
theoretical curve fits the late time data but at early times the theoretical curve is significantly
under damped. The reverse is true for c1 = 0.55. Clearly there is no way to fit the data with
damping proportional to θ̇ .
Figure 3 compares the data with two theoretical curves generated with damping
proportional to θ̇ 2 . The fit is a lot better but careful inspection shows that the data cannot
be fitted with pure θ̇ 2 damping either. Note that for θ̇ 2 damping, it is the larger coefficient
(c2 = 0.25) that fits the late time data and shows slightly too much damping at early times.
The reverse is true for the smaller c2 = 0.2.
Real-world damping of a physical pendulum 655

The results shown in figures 2 and 3 suggest that a combination of θ̇ and θ̇ 2 damping
could be made to fit the data. This is indeed the case as figure 4 demonstrates. The solid curve
in figure 4 is the theoretical curve using c1 = 0.10 and c2 = 0.17. Through trial and error
fitting of the theoretical calculations to the data we estimate the uncertainty in these constants
to be about ±0.01. As can be seen the fit is quite remarkable. The same combination was
applied to the fitting of the data for five or six different initial angles and in all cases the fit was
excellent.
Extensions to this easy to assemble and perform experiment could include an investigation
of the effect of cross-section area on the damping by taping varying size cards to the metre
stick. In addition, students could take video of other real life oscillating systems such as a
playground swing, long dangly earrings or a variety of other objects that swing in the air and
compare the data with theoretical models using STELLA.

References
[1] Halliday D, Resnick R and Walker J 1993 Fundamentals of Physics 4th edn (New York: Wiley) pp 397–8
Young H D and Freedman R A 2000 University Physics 10th edn (New York: Addison-Wesley) pp 411–3
Huggins E R 1999 Physics 2000 (New Hampshire: Moose Mountain Digital Press) pp 14–23
[2] Fowles G R and Cassidy G 1990 Analytical Mechanics 6th edn (Fort Worth: Saunders) p 56
Bairlein R 1973 Newtonian Dynamics (New York: McGraw-Hill) pp 45–54
Barger V and Olsson M 1973 Classical Mechanics: A Modern Perspective (New York: McGraw-Hill) pp 17–24
Arya A P 1990 Introduction to Classical Mechanics (Boston: Allyn and Bacon) pp 63–76
[3] Molina M I 2004 Exponential versus linear amplitude decay in damped oscillators Phys. Teach. 42 485–7 and
references therein
[4] Zonetti L F C, Camargo A S S, Sartori J, de Sousa D F and Nunes L A 1999 A demonstration of dry and viscous
damping of an oscillating pendulum Eur. J. Phys. 20 85–8
[5] LoPresto M C and Holody P R 2003 Measuring the damping constant for underdamped harmonic motion Phys.
Teach. 41 22–4
Mallinckrodt J 2003 Drag forces Phys. Teach. 41 261–2
LoPresto M C and Holody P R 2003 Drag forces authors responses Phys. Teach. 41 262
[6] Alexander P and Indelicato E 2005 A semi empirical approach to a viscously damped oscillating sphere Eur. J.
Phys. 26 1–10
[7] French A P 1971 Newtonian Mechanics (New York: W W Norton) problem 7–22, p 240
Lock J A 1982 The physics of air resistance Phys. Teach. 20 158–60
Fowles G and Cassiday G 1990 Analytical Mechanics 6th edn (Fort Worth: Saunders) p 56