169

Carl von Ossietzky Universität Oldenburg – Faculty V - Institute of Physics

Module Introductory laboratory course physics – Part I

Forced mechanical oscillations

Keywords:
HOOKE's law, harmonic oscillation, harmonic oscillator, eigenfrequency, damped harmonic oscillator,
resonance, amplitude resonance, energy resonance, resonance curves

References:
/1/ DEMTRÖDER, W.: „Experimentalphysik 1 – Mechanik und Wärme“, Springer-Verlag, Berlin
among others.
/2/ TIPLER, P.A.: „Physik“, Spektrum Akademischer Verlag, Heidelberg among others.
/3/ ALONSO, M., FINN, E. J.: „Fundamental University Physics, Vol. 1: Mechanics“, Addison-Wesley
Publishing Company, Reading (Mass.) among others.

1 Introduction
It is the object of this experiment to study the properties of a „harmonic oscillator“ in a simple mechanical
model. Such harmonic oscillators will be encountered in different fields of physics again and again, for
example in electrodynamics (see experiment on electromagnetic resonant circuit) and atomic physics.
Therefore it is very important to understand this experiment, especially the importance of the amplitude
resonance and phase curves.

2 Theory
2.1 Undamped harmonic oscillator
Let us observe a set-up according to Fig. 1, where a sphere of mass mK is vertically suspended (x-direc-
tion) on a spring. Let us neglect the effects of friction for the moment. When the sphere is at rest, there is
an equilibrium between the force of gravity, which points downwards, and the dragging resilience which
points upwards; the centre of the sphere is then in the position x = 0. A deflection of the sphere from its
equilibrium position by x causes a proportional dragging force FR opposite to x:

(1) FR ∝ − x

The proportionality constant (elastic or spring constant or directional quantity) is denoted D, and Eq. (1)
becomes the well-known HOOKE’s law:

(2) FR = − D x

Following deflection and release the dragging force causes an acceleration a of the sphere. According to
Newton’s second law and in combination with Eq. (2) we therefore obtain:

d2 x
(3) mK a = mK = mK = −D x
x (t: time)
d t2

the three terms on the left side merely representing different ways to write the relation force = mass ×
acceleration. Eq. (3) is the important differential equation (equation of motion), by means of which all
systems can be described which react with a dragging force on a deflection from their position of rest or
 170

equilibrium that is proportional to the degree of deflection. Such systems will be encountered very often
in different fields of physics.

-x

0
mk
+x

Fig. 1: Mass/spring system.

We are interested in learning which movement the sphere makes when it is deflected from its position at
rest and then released, its initial velocity v at the moment of release being zero. So we look for the
function x(t), which fulfils the differential equation (3) under the condition v(t = 0) = 0. As a resolution
we guess a function x(t), which represents a so-called harmonic (sinusoidal) oscillation:

=
(4) x x0 sin (ω t + ϕ )

x0 is the amplitude, (ωt+ϕ) the phase, ϕ the initial phase and ω the angular eigenfrequency of the oscil-
lation. Inserting Eq. (4) into Eq. (3) and performing differentiation twice with respect to time t we find the
value ω, for which Eq. (4) is a solution of Eq. (3):

D
(5) =ω =: ω0
mK

Thus, the sphere performs oscillations with this angular eigenfrequency ω0 when it is released. Since we
assume that there is no friction, the amplitude x0 of the oscillation remains constant. x0 as well as the ini-
tial phase ϕ are free parameters which have to be chosen such that Eq. (4) is „adjusted“ to the process to
be described, i.e. that Eq. (4) reflects the observed motion with the correct amplitude and initial phase.

Equation (5) is only valid if the mass of the spring, mF , is negligible compared to the mass mK of the
sphere. If this is not true, we have to consider that the spring’s different elements of mass also oscillate
following its deflection and release. The oscillation amplitudes of these elements of mass, however, are
very different: They increase from zero at the point of suspension of the spring to a value x0 at the end of
the spring. An exact calculation (e.g. in /3/) shows that the oscillation of the single elements of mass with
different amplitudes equals the oscillation of one third of the whole spring mass with the amplitude x0.
Therefore, the correct equation for the angular eigenfrequency reads:

D D 1
(6) ω
= 0 =: with =
m : mK + mF
1 m 3
mK + mF
3

In the experiment to be performed the sphere is not directly fixed to the spring but by a bar S with a
indicator Z (Fig. 7). In that case, mK in Eq. (6) has to be replaced by the sum

(7) M = m K + mS + mZ
 171

mS and mZ being the masses of S and Z.

An example illustrates the described relationships. According to Fig. 1 we observe a sphere of the mass
mK = 0.11 kg suspended by bar and indicator (mS + mZ = 0,07 kg) on a spring with the spring constant
D = 28 kg/s2 and the mass mF = 0.02 kg. The sphere is deflected by x0 = 0.05 m downwards from its
position at rest. Then we release the sphere and it performs oscillations with the amplitude x0 and the
eigenfrequency ν0 = ω0/(2π) ≈ 1.9 Hz (Eq. (6)). If we start to record the motion x(t) of the sphere exactly
when it has achieved its maximum upward deflection, the sinus according to Eq. (4) „starts“ at an initial
phase of ϕ = 3π/2 = 270° (mind the sign of x in Fig. 1!). This situation is represented in Fig. 2.

x(t)

x0

ϕ =3/2π

Fig. 2: Definition of the amplitude x0 and initial phase ϕ.

A system according to the arrangement considered here (also called mass/spring system) that performs
harmonic oscillations is called a harmonic oscillator. The harmonic oscillator is characterized by a
dragging force proportional to the deflection leading to a typical equation of motion in the form of (3)
with a solution in the form of (4). Equally characteristic of the harmonic oscillator is the parabolic
behaviour of its potential energy Ep as a function of the position:

1
(8) Ep = D x2
2

2.2 Damped harmonic oscillator

Now we observe the more realistic case of a mass/spring system under the influence of friction. We start
from the simple case that the system is subject to a force of friction Fb proportional to the velocity v. For
Fb we can write:
dx
(9) Fb =−bv =−b
dt

b being a constant of friction, which represents the magnitude of the friction.

Question 1:
- Which unit does b have? Why is there a minus sign in Eq. (9)?

In this case the equation of motion (3) takes on the form:

d2 x dx
(10) m 2 = −D x − b
dt dt

Usually, this differential equation is written in the form:

 172

d2 x b d x D
(11) + + x=
0
d t2 m d t m

Here again, it is interesting to know what type of motion the sphere performs after being deflected once
from its position at rest and then released with an initial velocity of zero. Thus, we are once again
searching for the function x(t) which resolves the differential equation (11) under the condition
v(t = 0) = 0. As a consequence of damping, we expect a decreasing amplitude of the oscillation and there-
fore try a solution with an exponentially decreasing amplitude (cf. Fig. 3):

=
(12) x e −α t x0 sin (ω t + ϕ ) (α : damping constant)

x(t)
x0
e-α t

T
0 t

Fig. 3: Damped harmonic oscillation.

We now insert Eq. (12) into Eq. (11), perform the differentiations, and find that Eq. (12) represents a
solution of Eq. (11) if the following is true for the parameters α and ω:

b
(13) α= and
2m

2
 b 
ω
(14) = ω0 − 
2

 2m 

We will now interpret this result. First we note that the amplitude of the oscillation decreases more rap-
idly the larger the damping constant (or damping coefficient) α is. In the case of invariable mass this
means according to Eq. (13) that the amplitude of the oscillation decreases more rapidly the larger the
constant of friction b is - which is quite plausible.

From Eq. (14) we can read how the angular eigenfrequency ω of this damped harmonic oscillation
changes with the constant of friction b. We study the following different cases:

(i) b = 0 → ω = ω0

In the case of vanishing friction (b = 0) we have the case of the undamped harmonic oscillator as
discussed in Chapter 2.1; the sphere performs a periodic oscillation at the angular eigenfrequency ω0.
 173

(ii) b/(2m)2 = ω02 → ω=0

This is the case of „critical damping“ in which the sphere does not perform a periodic oscillation any
more. It is therefore called the aperiodic borderline case. The sphere only returns to its starting
position exponentially (cf. remarks).

(iii) b/(m)2 > ω02 → ω imaginary

In the case of „supercritical damping” there is no periodic oscillation either. This case is called aperi-
odic case or over damped case. Here again, the sphere only returns to its starting position, however,
with additional damping, i.e., more slowly (cf. remarks).

(iv) 0 < b < 2mω0 → ω < ω0

This most general case, the oscillation case, leads to a periodic oscillation at a frequency ω, which is
below the angular eigenfrequency ω0 of the undamped harmonic oscillator.

Remarks:
Under the conditions discussed above (v(t = 0) = 0) there is no considerable difference between the
case of critical damping and supercritical damping: In both cases the sphere returns to its starting
position along an exponential path; in the case of supercritical damping there is only a stronger
damping. We find a different situation in the case v(t = 0) ≠ 0. If we do not only release the sphere, but
push it thus giving it a certain starting velocity, it is possible in the case of critical damping that the
sphere oscillates beyond its position at rest once, and only then returns to its starting position along an
exponential path. In the case of supercritical damping such an oscillation beyond that position does
not occur. The sphere always returns to its position at rest along an exponential path. Detailed
calculations (solution of the differential equation (11) under the conditions (2) and (3)) confirm these
relationships.

2.3 Forced harmonic oscillations

In Chapters 2.1 and 2.2 we have observed how the sphere oscillates if we deflect it once from its position
at rest and then release it. Now we will investigate which oscillations the sphere performs if the system is
subject to a periodically changing external force Fe (Fig. 4), for which the following is true:

(15) Fe = F1 sin (ω1 t )

F1 is the amplitude of the external force and ω1 its angular frequency. The sign is chosen such that the
forces directed downwards are counted as positive and upward forces are counted as negative.

-x

0 Fe
m
+x

Fig. 4: Oscillation generation of a mass/spring system with an external force Fe, m being the mass
according to Eqs. (6) and (7).
 174

The external force Fe additionally acts on the spring; the equation of motion thus gets the form (cf. Eqs.
(10) and (11)):

d2 x dx
(16) m 2
=
− D x−b + Fe
dt dt
and hence

d2 x b d x D 1
(17) 2
+ + x= F1 sin (ω1 t )
dt m dt m m

It is expected that the motion of the sphere following a certain transient time, i.e., after the transient os-
cillation, occurs at the same frequency as does the change of the external force. (There would be no plau-
sible explanation for another frequency.) However, a phase shift φ between the stimulating force and the
deflection of the sphere could be assumed. We may expect the oscillation amplitude to remain constant
upon completion of the transient oscillation since the system is provided with new external energy again
and again. Based on these considerations the following ansatz is suggested for the differential equation
(17):

=
(18) x x0 sin (ω 1 t + φ )

In this case φ is the phase shift between the deflection x and the external force Fe. For φ <0 the deflection
lags behind the stimulating force. By inserting Eq. (18) into Eq. (17) we find that Eq. (18) represents a
solution of Eq. (17) if the following is true for the amplitude x0 and the phase shift φ (cf. Appendix):

F1
(19) x0 = m
ω b
2

(ω 0
2
−ω 1 )
2 2
+ 1 
 m 

 
 ω 02 − ω 12  π
=(20) φ arctan   −
 ω1b  2
 m 

Contrary to the cases discussed in Chapters 2.1 and 2.2, the amplitude x0 and the phase φ are no longer
freely selectable parameters, rather they are definitely determined by the quantities F1, ω1, m, b and ω02 =
D/m.

Eq. (19) shows that the amplitude of the sphere's motion depends on the stimulating frequency. Plotting x0
over ω1, we obtain the amplitude resonance curve. Fig. 5 (top) shows some typical amplitude resonance
curves for different values of the friction constant b. In the stationary case, i.e. for ω1 = 0, we obtain the
amplitude x0 = F1/D known from HOOKE's law from Eq. (19). This is the value by which the sphere is
deflected if it is affected by a constant force F1. The position of the maximum of x0 as a function of ω1 is
found by means of the condition dx0/dω1 = 0. From Eq. (19) follows:
 175

0.3
←ω b=0
0

x /m
0.15
b = 0.1
b = 0.2
0
0
0 5 ω / Hz 10 15
1
0
←ω b=0
0
φ / r ad

-1.5 b = 0.1
b = 0.2
-3
0 5 ω / Hz 10 15
1

Fig. 5: Amplitude resonance curves (top) and phase curves (bottom) for a damped harmonic oscillator.
(F1 = 0.1 N, m = 0.1 kg, D = 2 kg/s2, b in kg/s).

b2
(21) ω 1= ω 02 − → x0 = Max.
2m 2

Except for the case b = 0, the maximum of the amplitude resonance curve is thus not found at the angular
eigenfrequencies ω0, but at lower angular frequencies ω1 < ω0.

The lower part of Fig. 5 shows the so called phase curves which determine the development of the phase
shift φ as a function of the angular frequency ω1. From Eq. (20) it follows that the phase angle φ is always
negative, i.e., the deflection of the sphere always lags behind the stimulating force except for the case
ω1 = 0.

We will now discuss some special cases:

(i) In the case ω1 << ω0 the amplitude x0 ≈ F1/D is independent of b for „not too large“ b. The phase
shift φ goes to 0 in this case: φ ≈ 0°. Thus the sphere directly follows the stimulating force.
(ii) In the resonance case ω1 ≈ ω0, the amplitude equals x0 ≈ F1/(ω0b), i.e., it is dependent on b. The
smaller b is, the larger x0 becomes; for b → 0, x0 → ∞. In this case the sphere's deflection lags behind
the generating force by 90° (φ ≈ -π/2).
(iii) In the case ω1 >> ω0 we find x0 ≈ F1/(mω12), i.e., the amplitude drops by 1/ω12. The phase shift is
φ = -π in this case, i.e., the sphere's deflection lags behind the generating force by 180°.

From the amplitude resonance curves the damping behaviour of a mass-spring-system can be read, i.e. of
a vibration isolating table, which is frequently used in optical precision metrology. The eigenfrequencies
of such tables are in the range of about 1 Hz. If an external disturbance has a very low frequency
(ω1 → 0), its amplitude is transferred onto the table without damping. In the range of the eigenfrequency
(ω1 ≈ ω0) it is (unintentional) amplified and in the range of high frequencies (ω1 >>ω0) it is damped
strongly.
 176

The damping behaviour of such a system can be influenced by changing the mass m. Fig. 7 shows that a
larger m reduces the eigenfrequency with the other parameters remaining unchanged and that the damping
for frequencies above the eigenfrequency can be increased significantly.

= ω 1 x 0 cos (ω 1 t + φ ) := v 0 cos (ω 1 t + φ )
dx
(22) v=
dt

0.4
m = 0.1
m = 0.2
m = 0.3
x /m

0.2
0

0
0 5 ω / Hz 10 15
1

Fig. 6: Amplitude resonance curves for different masses m (in kg) with other parameters remaining
unchanged (F = 0.1 N, D = 2 kg/s2, b = 0.1 kg/s)

Finally we will examine at which frequency the maximal energy transfer occurs from the generating sys-
tem to the oscillating system. As we know that the maximal kinetic energy is equivalent to the maximum
velocity, we first calculate the temporal course of the velocity v of the sphere using Eq. (18):

With Eq. (19) we obtain for the maximal velocity v0:

ω 1 F1
v0 ω=
(23) = m
1x 0
ω b
2

(ω 2
1 −ω 0)
2 2
+ 1 
 m 

and hence:

F1
(24) v0 =
2
 D
 m ω 1 −  + b
2

 ω1 

v0 becomes maximal when the denominator of Eq. (24) becomes minimal, i.e., if the following is true (for
b ≠ 0):

D
(25) − m ω 1 =0 → v 0= Max.
ω1

Hence it follows:
 177

D
(26) ω1 = =ω 0 → v 0 = Max.
m

Thus the velocity and also the kinetic energy become maximal if the system is stimulated with its angular
eigenfrequency ω0. Therefore, this case is called energy resonance, a case in which the generating system
can transfer the maximal energy to the oscillating system.

Question 2:
- What is the typical course of energy resonance curves (~ v02 (ω1 ) )? Sketch a diagram with the
principal course of v02 (ω1 ) for the cases b ≈ 0, b = b1 and b = b2 (according to Fig. 5).

Summing up we note the important result that a maximum transfer of energy and a maximum oscillation
amplitude are achieved at different generating frequencies in the case b ≠ 0.

3 Experimental procedure
Equipment:
Spring (D = (22.7 ± 0.5) kg/s2, mF = (0.0575 ± 10-4) kg), sphere with suspension bar and indicator
(M = (0.1727 ± 3 × 10-4) kg), generation system on stand with motor and light barrier, power supply
for motor and light barrier of the generation system, electronic speed controller for motor, halogen
bulb with condenser and power supply, lens (focal length f ≈ 200 mm, ∅ ≈ 80 mm), photo diode with
pinhole diaphragm (∅ = 0.5 mm) on stand, power supply for photo diode, observation screen on stand,
CCD camera with TV monitor and line selector, stand material for mounting lens, bulb, and CCD
camera, 2 glasses with different glycerine/water mixtures (b = (0.72 ± 0.03) kg/s for the more viscous
mixture at T = 20°C), desk for lens, lamp and CCD camera, digital oscilloscope, electronic counter,
metal measuring tape,.

3.1 Description of experimental set-up

The experiments are performed in a set-up as sketched in Fig. 7. By optically imaging the sphere’s
oscillation on a) a TV monitor and b) an observation screen it is possible to measure amplitude resonance
curves and phase curves without contact. This set-up is described in the following:

A sphere of mass mK is suspended on a spring by means of a bar S. The sphere is plunged into a glass GL
filled with a glycerine/water mixture to damp its oscillation. We assume that the frictional force is
proportional to the velocity of the sphere. There is an indicator Z fixed on the bar S, which is illuminated
with a halogen bulb from the right side. The indicator (with this illumination, its shadow) is imaged on a
screen B by means of the lens L (focal length f ≈ 200 mm) such that a magnified shadow Z' of the
indicator is produced on the screen. The image distance b' is chosen to be approx. 2 m, which yields an
object distance g of approx. 0.2 m according to the laws of geometric optics (imaging equation):

In order to obtain a good image, we place the lamp at a distance of 30 cm from the indicator and place the
condenser K such that the opening of the lens L is illuminated uniformly and in a straight line in the
direction of the optical axis. The indicator and the centre of the lamp and lens should be at the same level
when the plunged sphere (x = 0, central position of the bar S1) is at its zero position.

Now we look at the suspension of the spring. It is connected to a bar P of the length l via a joint G1 with a
bar S which runs in a guide F. The bar P is fixed on a rotary disk D via a ball bearing joint G2. The disk
can be turned at an angular frequency ω1 via a motor. Thus, the suspension point of the spring (i.e. the
position of the joint G1) is set in a periodic motion and a time-dependent external force Fe is produced on
the spring.
 183

4 Appendix
We want to demonstrate that the resonance amplitude x0 and the phase shift φ can be calculated with a
few simple calculation steps, if we change over to complex representation. In complex representation Eq.
(17) reads:

d2 x b d x D 1 iω t
(34) 2
+ + x= F1 e 1
dt m dt m m

In analogy to Eq. (18) we choose as a complex approach:

(
i ω t +φ ) iω1 t
=
(35) x x=
0 e
1
x0 e ei φ

iω1 t
Following differentiation and division by e insertion of Eq. (35) into Eq. (34) yields:

b D F1
(36) −ω 2
1 x0 ei φ + i ω 1 x0 ei φ + x0 ei φ =
m m m

Hence it follows with the definition of the angular eigenfrequency ω 0 according to Eq. (6):

F1
=
(37) x0 ei φ = m : z
b
ω 0 − ω 1 + iω 1
2 2

As already demonstrated in the experiment on the measurement of capacities, Eq. (37) is one representa-
tion form of a complex number z, whose absolute value (modulus) |z| = x 0 is given by zz * , with z*
being the conjugate complex quantity of z. Hence it follows:

 F1  F1 
 m   m 
(38) = zz ∗
=    b
x0
b
 ω 02 − ω 12 + iω 1   ω 02 − ω 12 − iω 1 
 m  m

from which we obtain Eq. (19) by simple multiplication.

For calculating the phase angle we again use (cf. experiment on the measurement of capacities) the sec-
ond representation of complex numbers, namely z = α + iβ, α being the real part and β, the imaginary part
of z. As is generally known, the phase angle φ can be calculated from these quantities as

 β  + π for α < 0 and β ≥ 0 

(39) φ = arctan   
 α  − π for α < 0 and β < 0 

In order to convert Eq. (37) into the form α + iβ, we extend the fraction in Eq. (37) with the conjugated
complex denominator:
 184

F1  2 b
m
 ω 0 − ω 1 − iω 1 
2

m
F1
m
( ω 02 − ω 12 ) − i 1 ω 1
F
m m
b
(40) x0 ei φ =
 2 b  2 b  ω1 b 
2

 ω 0 − ω 1 + iω 1  ω 0 − ω 1 − iω 1  ( )
2 2
2 2
ω 2
− ω +  
 m  m 0 1
 m 

from which we can read off the quantities α and β :

F1
( ω 02 − ω 12 )
F1
ω1
b
(41) α= m und β= − m m
 ω1 b   ω1 b 
2 2

(ω 0 − ω 1 ) +  m 
2 2 2
(ω 0 − ω 1 ) +  m 
2 2 2

   

which yields by insertion into Eq. (39):

 ω1b 
 
(42) φ =arctan  − 2 m 2  {− π für ω 1 > ω }
 ω 0− ω 1 
0

 

With

1 π
(43) =
arctan ( − y ) arctan   −
 y 2

it finally yields Eq. (20).