1

Extrapolating the Damping Coefficient of Air

Resistance for a Pendulum
B. Dolle (Bridgewater State University Physics and Mathematics)

I. A BSTRACT
In this study we report value of the damping coefficient
(γ) as .0021s ± .0001s, based on pendulum mechanics. By
measuring the length of the string, width of pendulum bob
and the transit time taken for the pendulum to pass through
a photogate, we to extrapolate the value of γ from a log-log
plot.

II. I NTRODUCTION
This study is a follow up to the previous simple pendulum
study, titled ”Measuring the Gravity of Earth Based on
Period-Length Relationship of the Simple Pendulum”. One
of the main assumptions of the aforementioned study was
the absence of air resistance. Exploring a more sophisticated
treatment of pendulum mechanics, a drag force caused by air
Fig. 2. Force-body diagram of the pendulum with air resistance.
was added to the model.

In this study, we measure the value of the damping ~

where I is the moment of inertia and θ̈ is the angular
coefficient (γ) by studying a pendulum which is allowed to 2
acceleration. Noting that I = ml , for mass m and length
decay. We derive the mechanics of the pendulum based the l, we find
small angle approximation. The derivation reveals exponential
decay of oscillations directly related to air resistance. τFd = −|l||bv| sin (π/2) = −lbv (2)

and
III. T HEORY
τFg = −|l||mg sin θ| sin (π/2) = −lmg sin θ. (3)
Fig 1 shows the setup of the pendulum we wish to study.
b (Eq 2) represents the drag coefficient of air where g (Eq 3)
represents gravity on earth.

Making the small angle approximation, we find the McLauren

series of sin(θ) to be
θ3 θ5
sin(θ) = θ − + − ... (4)
3! 5!
which to first order approximation becomes

sin(θ) ≈ θ. (5)

Thus, Eq 3 becomes τFg = −lmgθ and we re-write Eq 1 as

X
I θ̈ = ml2 θ̈ = −l2 bθ̇ − lmgθ (6)
Fig. 1. Diagram of a pendulum.
where we substituted v = lθ̇.
From Fig 1 we construct a free-body diagram of the forces
shown in Fig 2. We begin our derivation by using sum of Thus we find the equation of motion of the pendulum
torques: with air resistance to be
X X ~ X
~τ = I θ̈ = ~r × F~ (1) θ̈ + γ θ̇ + ω02 θ = 0 (7)
 2

where γ = b
m and ω02 = gl . Eq 7 has the general solution
The photogate outputs a square wave voltage signal (Fig
θ(t) = θ0 e−γt cos (ω0 t + δ). (8) 2), which we measured over time with a Data Acquisition
Eq 8 models damped harmonic motion where θ0 is the Device. Using a custom program in LabView, we collected
amplitude, cos (ω0 t + δ) is the harmonic motion, and e−γt data in 10.0s intervals at a resolution of .001s. We collected
is the damping term, which acts as the envelope. An energy these intervals at 45.0s intervals over the course of about
1240.0s. We used the midpoint of each interval to be the
representative time value and took an error of ± .05s.

Fig. 4. Diagram of square-wave signal received from Data Acquisition Device.

The axes are voltage (V) and time (t’), where the labeled ∆t0 is the transit
Fig. 3. Alternative diagram of the pendulum used for energy analysis. time of the pendulum.
Assuming a nearly conservative system, we compare energy values at point
a (purely kinetic) and point b (purely potential).

analysis of Fig 3 shows that the energy at point a is E = 21 mv 2 V. DATA A NALYSIS

and the energy at point b is E = mg(l−h). Using conservation
of energy, we find For each 10 second interval, we found the ∆t0 for each transit
event and then took the mean as the reported values. We used
v2 l−h
h= and cos θ = , the standard deviation for the error.
2g l
which rearranged is To find the error in velocity, we used the equation
v2
  s 2  2
θ = arccos 1 − . (9) ∂ 0 ∂
2gl δv = δt + δx . (11)
∂t0 ∂x
Thus we find theta to be a function of velocity and pendulum
length. To find the error in θ, we used
s 2  2
To find the damping coefficient, we discard the oscillatory ∂ ∂
part of Eq 8 and take the natural log δθ = δv + δl (12)
∂v ∂l
ln θ = ln θ0 + (−γ)t (10)
where we took δv from Eq 11.
which is the equation of a straight line with slope −γ.
Taking Eq 10 as our theoretical model and setting both
IV. E XPERIMENT ln θ0 and γ as free parameters, we used reduced χ2 to
measure goodness of fit. We then used Solver in Excel to
In this experiment we measured the damping coefficient of
minimize our reduced χ2 . The standard deviation of the error
a swinging pendulum. To find the damping coefficient, we
in θ was taken for the standard deviation for the χ2 test. The
needed to find velocity and pendulum length. The pendulum
minimized value of γ was taken to be the reported value.
length was directly measured with a meter stick.

To measure the velocity we passed a bob in front of a VI. R ESULTS

photogate and measured its transit time (∆t0 ) (for details
and a diagram of the photogate, see [1]). The pendulum bob The string length of the pendulum (l) in this experiment was
width (∆x) was measured with a vernier caliper. Then we measured to be .800m ± .005m. The width of the bob (∆x)
∆x
calculated velocity, v = ∆t 0. was found to be .0181m ± .0001m. The value of g was taken
as 9.81 m/s2 .
It is worth noting that t and t0 are different variables. t, used in
the derivation, refers to time as the experiment was performed. The log-log plot of measured values for ln θ were plotted for
t0 refers to the times of transit, which were measured each time (Fig 5). We report a value of γ of .0021s−1 ±
independently of t. .0001s−1 .
 3

Fig. 5. Measured values of ln θvsT ime(s)

VII. D ISCUSSION
We found a value of γ of .0021s ± .0001s. This was found
by modifying the free parameters lnθ0 and γ to minimizing
χ2 . The standard deviation for χ2 was found to be .038,
yielding a reduced χ2 value of 1.04, implying a great fit of
the model to the data.

This value of reduced χ2 is interesting when compared

to the value from the study (titled ”Measuring the Gravity
of Earth Based on Period-Length Relationship of the Simple
Pendulum”) of the simple pendulum in absence of air
resistance. One explanation of the over-fitting was the
absence of air resistance. Adding air resistance in this study
seems to have reconciled this discrepancy.

VIII. C ONCLUSION
We report an experimental value of .0021s ± .0001s for γ.
The reduced χ2 value of 1.04 suggests nearly perfect fit of the
model to the data.