Effective Mass of an

Oscillating Spring
Eduardo E. Rodríguez, Universidad de Buenos Aires and Universidad Favaloro, Buenos Aires

Gabriel A. Gesnouin, Universidad Favaloro, Buenos Aires

W
e present an experimental method to or, equivalently, the factor f. The method correlates the
obtain the effective mass of an unloaded behavior of the partial springs formed when only n of
oscillating spring. We measure the period the total number N coils of a given spring are hanging
T(n) of the partial springs that result when hanging and oscillating. First, we review some theoretical treat-
n of the total N coils of a given spring. Data are cor- ments that address the problem and afterward
related with the expectation of a simple model for describe our experimental method derived from a
T(n) that takes into account the effective mass of simple model.
the hanging spring. We analyze results for metal and
plastic springs of different masses and strengths, and The problem
conclude that the method provides the effective mass The problem of finding the value of f is approached
with a good degree of reliability. in several texts by means of energy considerations,1
A typical exercise in mechanics consists of finding starting from the kinetic energy EK of an oscillating
the oscillation period T of a mass m attached to the vertical spring of length L:
free end of a vertical spring of force constant k. If the L
1 2
mass mS of the spring is neglected, the period of free EK = ∫
v dm. (3)
2
oscillations is1 0
(1) The integral is usually solved assuming homogeneous
m
T = 2p . stretching that implies a uniform mass distribution,
k
m
When the mass of the spring is considered, a possibili-
dm = S dz , (4)
ty is to write2 L
and also taking the velocity as a linear function of
m + fmS (2) the position z measured from the fixed point of the
T = 2p ,
k spring,
where the term f mS, a fraction f of the spring’s mass
v
mS, is the “effective mass of the spring,” and with its
v( z ) = 0 z . (5)
inclusion, the period corresponds to that of a mass m L
+ f mS fastened to a massless spring. It is usual to ob- Here, v0 is the instantaneous velocity of the point at
tain the effective mass by analyzing results of T(m) in the free end at z = L. With these assumptions, integra-
the limit m ➝ 0.3 tion of Eq. (3) gives
In this paper we present a new experimental meth-
1 mS 2
od to obtain the effective mass of an oscillating spring
EK = v0 ,
2 3
100 DOI: 10.1119/1.2432087 The Physics Teacher ◆ Vol. 45, February 2007
from which the value f = 1/3 is obtained. This result
can be taken as representative of a stiff spring for
which both uniform stretch and a linear velocity dis-
tribution can be better modeled.
By using the same approach and a noncalculus
method, Ruby also deduces f = 1/3 for a stiff spring.
However, he finds f = 1/2 for a soft spring.4 To obtain
these values, his model takes a spring with a nonuni-
form stretch and an associated mass distribution given
by
az dz
dm = [mS /(1 + a / 2 )] (1 + )
. (6)
L L

This distribution satisfies the condition Fig. 1. Mass distribution of two measured springs. Blue:
Metal spring #1. Red: Plastic spring #3. Lines are fits with
L a quadratic expression derived from Eq. (6).
∫ dm = mS We have measured the mass distribution of several
0
hanging soft springs stretched by their own weight
and introduces an adjustable parameter a: a = 0 for a and found it difficult to apply the nonlinear model of
stiff spring and a >> 1 for a very soft spring. Neverthe- Ref. 4, expressed by Eq. (6), to obtain the parameter
less, this model conserves the assumption of a linear a necessary to estimate f (Fig. 1). For this reason we
velocity distribution [Eq. (5)] to integrate Eq. (3). have thought of an experimental method to obtain
On the other hand, Galloni and Kohen3 treat the the factor f in the case of an unloaded spring. Our
spring oscillations in the framework of wave equations experiments are based on a simple model and consist
and obtain different values of f depending on the ratio of measurements of the oscillation period T(n) and
m/mS. They find the force constant k(n) of the spring that result from
hanging only n of the N coils of the whole spring.
f ➝ 1/3 when m/mS >>1,
The model
and Consider a spring of mass mS, with only n of its N
coils hanging. We propose that the unloaded resulting
f ➝ 4/p2 < 0.41 when m/mS << 1. spring oscillates with a period
fM (n )
The common one-third fraction of the different T (n ) = 2p
, (7)
k (n )
models can be understood as follows. When m >> mS
the stretch is mainly defined by the loading of mass where f is a fraction of the hanging mass M(n) and
m; therefore, a uniform stretch along the whole spring k(n) is the force constant of the spring composed of
can be assumed, like in the energy model for a strong the n active coils. We assume f is independent of n but
spring. However, there is a discrepancy in the results observe that both M and k naturally depend on the
of the models in the limit m << mS when nonlinear number n of hanging coils. On one hand, the hanging
effects can be expected. In this sense, the behavior of a mass is
very soft spring is comparable to the case m/mS
m
<< 1, where the mass of the spring is distributed non-
M (n ) = n S . (8)
uniformly due to the variable stress along the spring N
produced by the different number of coils each coil On the other hand, the spring constant k must vary
supports below. with n—that follows the observation that a spring is

The Physics Teacher ◆ Vol. 45, February 2007 101

 Then, the model predicts a linear increase of T with n
and leads us to an accessible experimental study.

Procedure and Results

For the experiments, each spring is mounted using
a thin metal plate and located close to a wall (Fig. 2).
The setup easily allows hanging only n turns of the
total N of the spring. The other N – n coils remain
demobilized, piled on the plate. On the wall we glue
a long strip of millimetric paper on which we mark
the position of each spire to study the mass distribu-
tion of the spring. To do that, the spring is suspended
from the uppermost turn, remaining elongated due to
its own weight. The same experimental arrangement
allows us to measure the force constant of the spring
when an extra load is attached to the lowermost turn
of the n hanging coils.
Our experiments test the model through two steps.
First, for a given spring we measure k(n) to validate
Fig. 2. Experimental setup to determine the effective
mass of a spring.
Eq. (9) and obtain C as the slope of the linear graph of
k as a function of 1/n. Second, we measure T(n), verify
stiffer when shorter. Thus, we propose a simple the linear relationship between T and n, and obtain
dependence of k with n: the slope A. With the values A and C we finally calcu-
late the fraction f.
C

k (n ) = , (9) In order to prevent “Slinky” oscillations, the mea-
n surements of the period T for each chosen n are made
where C is a constant. Combining (7) with (8) and maintaining small-oscillation amplitudes. To reduce
(9), the period results in uncertainties, we measure with a digital stopwatch the
time t10 of 10 oscillations and calculate T = t10/10.
fmS

T (n ) = 2p n = An. (10) The spring constant for each n is obtained under static
C N conditions. By means of a simple measurement, we ob-

Fig. 3. k(n–1) of the spring #1. The uncertainty of k is of Fig. 4. T(n) of the spring #1. Solid line is a fit with Eq.
the order of 2%, and it is covered by the size of the data- (10).
point markers. Solid line is a fit with Eq. (9).

102 The Physics Teacher ◆ Vol. 45, February 2007

 Table I. Data of different springs.
tain k as the ratio of the weight W of a small load and
the resulting spring’s length change DL, k = W/DL. Full
Num-
We show results for a metal helical spring (spring length, Frac-
ber of Mass, A C
Spring L, and tion,
#1) with a mass mS = 111.4 g and N = 99 coils. Figure coil dia-
coils, mS (g) (s) (N/m)
f
N
3 shows k as a function of 1/n, and Figure 4 contains meter
T(n). In the figures, the solid lines represent in each #1 1190 mm
99 111.4 0.021 41.7 0.42
case the best fit of the experiment points with Eq. metal 55 mm

(9) and (10), respectively. We obtain the best values: #2 1200 mm

98 117.0 0.023 42.2 0.43
C = 41.7 N/m and A = 0.021 s. For this spring, the metal 55 mm
fraction results in f = 0.42. Results for two pairs of #3 700 mm
springs, metal and plastic, are summarized in Table plastic 45 mm 39 32.4 0.035 11.5 0.43

I. While data comparison of the fractions f obtained

#4 680 mm
for springs #1 and #2 (or #3 and #4) show the repro- 38 33.0 0.039 10.0 0.43
plastic 45 mm
ducibility of the measurements performed on similar
specimens, the close agreement of the results obtained
with the metal and plastic springs with different char-
acteristics [note that C(#1)/C(#3) = 3.8] seems to vali- References
date the proposed method to obtain the effective mass 1. F. Sears, M. Zemansky, and H. Young, University Phys-
of a hanging spring. ics (Addison-Wesley, Reading, MA, 1976).
2. See, for example, J. Christensen, “An improved calcula-
Analysis tion of the mass for the resonant spring pendulum,”
Due to the regularity observed in the values of f for Am. J. Phys. 72, 818–878 ( June 2004).
the springs analyzed, we conclude that our method 3. E.E. Galloni and M. Kohen, “Influence of the mass
provides consistent results for the unloaded springs. of the spring on its static and dynamic effects,” Am. J.
Phys. 47, 1076–1078 (Dec. 1979).
We observe that the method produces a fraction f
closer to the value f = 4/p2 of Ref. 3 when m/mS 4. L. Ruby, “Equivalent mass of a coil spring,” Phys. Teach.
38, 140–141 (March 2000).
<< 1 (a discrepancy of +5%, which is within the ac-
curacy of our determination of f ) than to the value ½ PACS codes: 01.55.+b, 01.50.Pa, 46.00.00
(discrepancy of –12%) predicted for soft springs by
the particular nonlinear model of Ref. 4. For compari- Eduardo E. Rodríguez is a physicist and science teacher
in Argentina. He currently teaches physics to students
son, Christensen2 also determined in the limit m/mS majoring in biology, geology, medical physics, and engi-
<< 1 a fraction f slightly higher than the theoretical neering. He also coordinates science workshops for high
school teachers. The results of these activities are posted
value 4/p2 of Ref. 3.
on the following websites: http://www.fisicarecreativa.com/
Our data treatment is based on a simple model and http://www.cienciaredcreativa.org/. He is also the
tested through graphical analysis. Because this exer- author of books on experimental physics.
cise involves a number of important concepts and Departamento de Física, Facultad de Ciencias Exactas
y Naturales, Universidad de Buenos Aires, Ciudad
activities—modeling, measuring, graphical analysis, Universitaria, Pabellón I, (1428) Ciudad Autónoma de
and comparison with other models and results—the Buenos Aires; eduardo@df.uba.ar
experiments are useful in a basic experimental course
Gabriel Gesnoiun is an advanced student major-
in the first university year, and constitute a suitable ing in medical physics and a teaching assistant at the
complement to the study of springs in the context of Universidad Favaloro, Buenos Aires;
oscillatory motion. Departamento de Física y Química, Facultad
de Ingeniería y Ciencias Naturales y Exactas,
Universidad Favaloro, Av. Belgrano 1723, (1093)
Acknowledgments Ciudad Autónoma de Buenos Aires; gabriel
We thank Fundación Antorchas of Argentina and gesnouin@hotmail.com
Suillar Argentina for financial support.

The Physics Teacher ◆ Vol. 45, February 2007 103