Normal Modes of Beaded String
Consider a mechanical system consisting of a taut string that is stretched
between two immovable walls. Suppose that
identical beads of
mass
are attached to the string in such a manner that they cannot slide along it. Let the beads be equally
spaced a distance
apart, and let the distance between the first and the last beads and
the neighboring walls also be
. See Figure 4.1. Consider transverse
oscillations of the string; that is, oscillations in which the string
moves in a direction perpendicular to its length.
It is assumed that the inertia of the string
is negligible with respect to that of the beads. It follows that the sections of the string
between neighboring beads, and between the outermost beads and the walls,
are straight. (Otherwise, there would be a net tension force acting on the sections, and
they would consequently suffer an infinite acceleration.) In fact, we expect the instantaneous configuration of the string to be a
set of continuous straight-line segments of varying inclinations, as shown in the figure. Finally, assuming that
the transverse displacement of the string is relatively small, it is
reasonable to suppose that each section of the string possesses the
same tension,
. [See Equation (4.11).]
Figure 4.1:
A beaded string.
|
It is convenient to introduce a Cartesian coordinate system such that
measures distance along the string from the left wall, and
measures the transverse displacement of the string from its
equilibrium position. See Figure 4.1. Thus, when the string is in its equilibrium position
it runs along the
-axis. We can define
 |
(4.1) |
where
. Here,
is the
-coordinate of the
closest bead to the left wall,
the
-coordinate of the second-closest bead, et cetera. The
-coordinates of the beads
are assumed to remain constant during their transverse oscillations.
We can also define
and
as the
-coordinates of the left
and right ends of the string, respectively. Let the transverse displacement
of the
th bead be
, for
. Because each displacement
can vary independently, we are dealing with an
degree of freedom system.
We would, therefore, expect such a system to possess
unique normal modes of
oscillation.
Consider the section of the
string lying between the
th and
th beads, as shown in Figure 4.2.
Here,
,
, and
are the distances of the
th,
th, and
th beads, respectively,
from the left wall, whereas
,
, and
are the corresponding transverse displacements of these beads.
The two sections of the string that are attached to the
th
bead subtend angles
and
with the
-axis, as illustrated
in the figure. Standard trigonometry reveals that
 |
(4.2) |
and
 |
(4.3) |
However, if the transverse displacement of the string is
relatively small—that is, if
for all
—which we shall assume to be
the case, then
and
are both small angles. Thus,
we can use the small-angle approximation
. (See Appendix B.) It follows that
Figure 4.2:
A short section of a beaded string.
|
Let us find the transverse equation of motion of the
th bead. This
bead is subject to two forces: namely, the tensions in the sections
of the string to the left and to the right of it. (Incidentally, we are neglecting any gravitational forces
acting on the beads, compared to the tension forces.) These tensions are of
magnitude
, and are directed parallel to the associated string sections, as shown
in Figure 4.2. Thus, the transverse (i.e.,
-directed) components of
these two tensions are
and
, respectively.
Hence, the transverse equation of motion of the
th bead becomes
 |
(4.6) |
However, because
and
are both small angles, we
can employ the small-angle approximation
.
It follows that
 |
(4.7) |
Finally, making use of Equations (4.4) and (4.5),
we obtain
 |
(4.8) |
where
. Because there is nothing special
about the
th bead, we deduce that the preceding equation of motion
applies to all
beads; that is, it is valid for
. Of course,
the first (
) and last (
) beads are special cases, because there
is no bead corresponding to
or
. In fact,
and
correspond to the left and right ends of the string, respectively. However,
Equation (4.8) still applies to the first and last beads, as long as
we set
and
. What we are effectively demanding
is that the two ends of the string, which are attached to the left and right walls, must both have zero transverse
displacement.
Incidentally, we can prove that the tensions in the two sections of the string shown in Figure 4.2 must be
equal by considering the longitudinal equation of motion of the
th bead. This equation takes the
form
 |
(4.9) |
where
and
are the, supposedly different, tensions in the sections of the string to the
immediate left and right of the
th bead, respectively. We are assuming that the motion
of the beads is purely transverse; that is, all of the
are constant in time. Thus, it follows from the
preceding equation that
 |
(4.10) |
However, if the transverse displacement of the string is such that all of the
are small then
we can make use of the small-angle approximation
. (See Appendix B.) Hence, we
obtain
 |
(4.11) |
A straightforward extension of this argument reveals that the tension is the same in all sections of the string.
Let us search for a normal mode solution to Equation (4.8) that
takes the form
 |
(4.12) |
where
,
,
, and
are constants.
This particular type of solution is such that all of the beads execute transverse simple harmonic oscillations
that are in phase
with one another. See Figure 4.4. Moreover, the oscillations have an amplitude
that varies sinusoidally along the length of the string (i.e., in the
-direction). The pattern of oscillations is thus periodic in space. The spatial repetition period, which is generally termed the wavelength, is
. [This follows from Equation (4.12) because
is a periodic function with period
: i.e.,
.] The constant
, which determines the wavelength, is usually referred to as the wavenumber. Thus, a
small wavenumber corresponds to a long wavelength, and vice versa.
The type of solution specified in Equation (4.12) is generally
known as a standing wave. It is a wave because it is periodic
in both space and time. (An oscillation is periodic in time only.) It is
a standing wave, rather than a traveling wave, because the points
of maximum and minimum amplitude oscillation are stationary (in
). See Figure 4.4.
Substituting Equation (4.12) into Equation (4.8), we obtain
which yields
![$\displaystyle -\omega^{\,2}\,\sin(k\,x_i)=\omega_0^{\,2}\left(\sin[k\,(x_i-a)] -2\,\sin(k\,x_i)
+\sin[k\,(x_i+a)]\right).$](img917.png) |
(4.14) |
However, because
(see Appendix B), we get
![$\displaystyle -\omega^{\,2}\,\sin(k\,x_i)=\omega_0^{\,2}\left[\cos(k\,a)-2+\cos(k\,a)\right]\sin(k\,x_i),$](img919.png) |
(4.15) |
which gives
 |
(4.16) |
where use has been made of the trigonometric identity
. (See Appendix B.) An expression, such as Equation (4.16), that determines the angular frequency,
, of
a wave in terms of its wavenumber,
, is generally termed a dispersion relation.
The solution (4.12) is only physical provided
. In other words,
provided the two ends of the string remain stationary. The first constraint
is automatically satisfied, because
. See Equation (4.1). The second constraint
implies that
![$\displaystyle \sin(k\,x_{N+1}) = \sin[(N+1)\,k\,a] = 0.$](img923.png) |
(4.17) |
This condition can only be satisfied if
 |
(4.18) |
where the integer
is known as a mode number. A small mode number translates to a small wavenumber, and, hence, to a
long wavelength, and vice versa. We conclude that the possible wavenumbers,
, of the normal modes of the system are
quantized such that they are integer multiples of
.
Thus, the
th normal mode is
associated with the characteristic pattern of bead displacements
 |
(4.19) |
where
 |
(4.20) |
Here, the integer
indexes the beads, whereas the mode number
indexes the normal modes. Furthermore,
and
are arbitrary constants determined by the initial conditions.
Of course,
is the peak amplitude of the
th normal mode, whereas
is the associated phase angle.
How many unique normal modes does the system possess? At first sight, it
might seem that there are an infinite number of normal modes, corresponding to the
infinite number of possible values that the integer
can take. However, this is not
the case. For instance, if
or
then all of the
are zero.
Clearly, these cases are not real normal modes. Moreover, it can be
demonstrated that
provided
and
, as well as
provided
and
. We,
thus, conclude that only those normal modes that have
in the
range
to
correspond to unique modes. Modes with
values lying outside
this range are either null modes, or modes that are identical to other modes
with
values lying within the prescribed range. It follows that
there are
unique normal modes of the form (4.19). Hence, given that we are dealing with an
degree of
freedom system, which we would expect to only possess
unique normal modes, we can be sure that we have found all the normal modes.
Figure: 4.3
Normal modes of a beaded string with eight equally-spaced
beads. The modes are all
shown at the instances in time when they attain their maximum amplitudes: in other words,
at
.
|
Figure 4.3 illustrates the spatial variation of the normal modes of a beaded string possessing eight
beads. That is, an
system. It can be seen that the low-
—that is, long-wavelength—modes cause the string to oscillate in a
fairly smoothly-varying (in
) sine wave pattern. On the other hand, the high-
—that is, short-wavelength—modes cause the string to oscillate in a
rapidly-varying zig-zag pattern that bears little resemblance to a sine wave.
The crucial distinction between the two different types of mode is that the
wavelength of the oscillation (in the
-direction) is much larger
than the bead spacing in the former case, while it is similar to
the bead spacing in the latter. For instance,
for the
mode,
for the
mode, but
for the
mode.
Figure 4.4:
Time evolution of the
normal mode of a beaded string with eight equally-spaced
beads. The mode is shown at
,
,
,
,
,
,
,
and
.
|
Figure 4.5 displays the temporal variation of the
normal mode of
an
beaded string. It can be seen
that the beads oscillate in phase with one another. In other words, they all attain their maximal
transverse displacements, and pass through zero displacement, simultaneously.
Moreover, the mid-way point of the string always remains stationary. Such a point is
known as a node. The
normal mode has two nodes (counting the
stationary points at each end of the string as nodes), the
mode has three nodes, the
mode four nodes, et cetera. In fact, the existence
of nodes is one of the distinguishing features of a standing wave.
Figure 4.5:
Normal frequencies of a beaded
string with eight equally-spaced beads.
|
Figure 4.5 shows the normal frequencies of an
beaded string plotted as a function of the normalized wavenumber. Recall that, for an
system,
the relationship between the normalized wavenumber,
, and the
mode number,
, is
.
It can be seen that the angular frequency increases as the wavenumber
increases, which implies that shorter wavelength modes have
higher oscillation frequencies. The dependence of the angular frequency
on the normalized wavenumber,
, is approximately linear when
. Indeed, it can be seen from Equation (4.20) that if
then the small-angle approximation
yields a linear dispersion relation
of the form
 |
(4.25) |
We, thus, conclude that those normal modes of a uniformly-beaded string whose wavelengths greatly
exceed the bead spacing (i.e., modes with
) have approximately linear dispersion
relations in which their angular frequencies are directly proportional
to their mode numbers. However, it is evident from the figure that this linear relationship
breaks down as
, and the mode wavelength consequently becomes
similar to the bead spacing.