Siah Ma Koun 1997
Siah Ma Koun 1997
Siah Ma Koun 1997
magnetic field
Azad Siahmakoun, Valentina A. French, and Jeffrey Patterson
Downloaded 20 May 2013 to 129.174.21.5. Redistribution subject to AAPT license or copyright; see http://ajp.aapt.org/authors/copyright_permission
Nonlinear dynamics of a sinusoidally driven pendulum in a repulsive
magnetic field
Azad Siahmakoun
Department of Physics and Applied Optics, Rose-Hulman Institute of Technology, Terre Haute,
Indiana 47803
Valentina A. French
Department of Physics, Indiana State University, Terre Haute, Indiana 47809
Jeffrey Patterson
Department of Physics and Applied Optics, Rose-Hulman Institute of Technology, Terre Haute,
Indiana 47803
~Received 24 July 1996; accepted 24 October 1996!
The dynamics of a sinusoidally driven pendulum in a repulsive magnetic field is investigated
theoretically and experimentally. The experimental data are acquired using a shaft encoder
interfaced to a PC which measures the angular displacement of the pendulum as a function of time.
Both the theoretical simulations and the experimental measurements exhibit regions of periodic and
chaotic behavior, depending on the system parameters. Amplitude jumps, hysteresis, and bistable
states are also observed. The simplicity of the apparatus makes this experiment suitable for an
advanced undergraduate laboratory. © 1997 American Association of Physics Teachers.
393 Am. J. Phys. 65 ~5!, May 1997 © 1997 American Association of Physics Teachers 393
Downloaded 20 May 2013 to 129.174.21.5. Redistribution subject to AAPT license or copyright; see http://ajp.aapt.org/authors/copyright_permission
Fig. 1. Schematic diagram of the physical pendulum and the forces acting
on it.
II. THEORY
Fig. 2. ~a! Time series and ~b! phase space plots showing periodic motion
A theoretical model is developed considering the four for d570 mm. ~c! Time series and ~d! phase space plots showing more
forces acting on the pendulum: the restoring gravitational complicated periodic orbits when the distance, d, between the magnets is
force, the repulsive magnetic force, the sinusoidal driving decreased to 66.70 mm.
394 Am. J. Phys., Vol. 65, No. 5, May 1997 Siahmakoun, French, and Patterson 394
Downloaded 20 May 2013 to 129.174.21.5. Redistribution subject to AAPT license or copyright; see http://ajp.aapt.org/authors/copyright_permission
Fig. 3. As the distance d is further decreased, transition to chaos occurs near
d566.55 mm. ~a! Time series and ~b! phase space plots showing chaotic
motion for d566.55 mm.
r u 5 A~ L sin u ! 2 1h 2u , ~2!
395 Am. J. Phys., Vol. 65, No. 5, May 1997 Siahmakoun, French, and Patterson 395
Downloaded 20 May 2013 to 129.174.21.5. Redistribution subject to AAPT license or copyright; see http://ajp.aapt.org/authors/copyright_permission
Fig. 5. For d566.53 mm the pendulum is no longer able to overcome the Fig. 6. As the control parameter d is further decreased to 66.40 mm the
repulsive magnetic field and the chaotic motion is confined to the positive u pendulum displays periodic motion again. The motion is confined to the
values. positive u values.
d 2u
I
dt 2
5 ( ti , ~4!
S
3cos u u u 1arctan 2 SU hu
L sin u
U DD , ~6!
u̇ 5 v , ~7!
Ḟ5V. ~8!
Here, M is the mass of the pendulum, g is the gravitational
acceleration, T driver is the maximum value of the periodic
torque produced by the horizontal displacement of the pivot, Fig. 7. Equilibrium positions of the pendulum as a function of the minimum
F and V are, respectively, the phase and the angular fre- separation between the magnets, d.
396 Am. J. Phys., Vol. 65, No. 5, May 1997 Siahmakoun, French, and Patterson 396
Downloaded 20 May 2013 to 129.174.21.5. Redistribution subject to AAPT license or copyright; see http://ajp.aapt.org/authors/copyright_permission
only on the physical constants associated with the system.
The coupled differential equations ~7!, ~8!, and ~9! are
solved numerically using MATHEMATICA. The following pa-
rameter values and initial conditions are used for the com-
puter simulations: A5110 s22, B50.01 s22, C50.001 s21,
D50.2 m2/s2, u~0!50.1 rad, v~0!50, F~0!50. Data are re-
corded after allowing the transient behavior of the pendulum
to die out ~i.e., after 30 cycles!. While the driver’s frequency
is kept constant at 1 Hz both periodic and chaotic behaviors
are observed by varying the distance, d, between the two
magnets. For example, for values of d between 100 and 70
Fig. 8. Experimental setup for the study of the dynamics of a sinusoidally mm the pendulum exhibits periodic behavior as shown in
driven physical pendulum in a repulsive magnetic field.
Fig. 2. As d is decreased further, more complicated orbits are
observed @Fig. 2~c! and ~d!# and then transition to chaos oc-
quency of the driver, and g is the damping constant. Equa- curs near d566.55 mm, as shown in Fig. 3. As can be seen
tion ~6! can be rewritten in a simpler form from the time series in Fig. 3~a!, after a few oscillations on
both sides of the equilibrium position, the pendulum is no
uuu D longer able to overcome the repulsive magnetic field and its
v̇ 52A sin u 1B sin F2C v 1
u r 2u oscillations are trapped on one side. For d566.54 mm cha-
S SU U DD
otic behavior is observed where the chaotic attractor is
hu bounded in both positive and negative u regions @Fig. 4~b!#.
3cos u u u 1arctan 2 , ~9! In this case the pendulum is at first momentarily trapped on
L sin u
one side, as shown in Fig. 4~a!, and then it suddenly over-
where A, B, C, and D are constant coefficients that depend comes the magnetic field, getting trapped on the other side.
Fig. 9. Experimental time series and phase space plots generated by the Fig. 10. Experimental plots showing transition to chaos as the magnets are
shaft encoder–PC system showing periodic motion for d570 mm. moved closer together at a distance d555.60 mm.
397 Am. J. Phys., Vol. 65, No. 5, May 1997 Siahmakoun, French, and Patterson 397
Downloaded 20 May 2013 to 129.174.21.5. Redistribution subject to AAPT license or copyright; see http://ajp.aapt.org/authors/copyright_permission
Fig. 11. Time series and phase space plots generated from the experimental data showing chaotic behavior for the control parameters: ~a!, ~b! d555.40 mm;
~c!, ~d! d555.30 mm.
Figure 4~c! and ~d! shows the pendulum’s motion for the III. EXPERIMENT
same distance between the magnets, d566.54 mm, but for
slightly different initial conditions: u~0!50.1002 rad, The experimental apparatus, shown in Fig. 8, consists of a
v~0!50, and F~0!50. Notice the dramatic change in the physical pendulum made of an aluminum rod 13 cm long
pendulum’s motion caused by only a very slight change in and of mass 11.4 g. This rod is mounted on a U.S. Digital
the initial conditions ~i.e., a change of 0.0002 rad '0.01 deg softpot ~shaft encoder! model SP-512B such that the pendu-
in the initial angular position of the pendulum!. This is an lum can rotate freely, but the rotation is restricted to a plane.
illustration of one of the characteristic features of chaotic The shaft encoder is used to record angular displacement
motion: its sensitive dependence on initial conditions. Near data via a software package.8 The software package allows
d566.53 mm, the magnetic field becomes sufficiently large immediate viewing of the time series and phase space plots
that the pendulum is no longer able to overcome it and the of the pendulum’s motion. The pendulum is coupled to a
trajectories are confined to the positive u regions ~Fig. 5!. As Pasco Scientific Mechanical Vibrator, model SF-9324, which
d is further decreased, a periodic motion confined to the constitutes the driving force of the pendulum. The amplitude
positive u region is observed about a new equilibrium point of this mechanical driver is adjustable with a maximum of 5
as shown in Fig. 6. The presence of such equilibrium points mm. A sinusoidal signal produced by an SRS synthesized
can be predicted by solving Eq. ~9! for an unforced pendu- function generator model DS 345 is supplied to the driver.
lum. Figure 7 displays a plot of the equilibrium positions of The frequency and amplitude of the driver are varied by
the pendulum as a function of the minimum separation be- changing the frequency and amplitude of the signal produced
tween the magnets d. Notice that two new equilibrium posi- by the function generator. A disk magnet of mass 5 g and of
tions, one on either side of the vertical position, appear as the the same diameter as the aluminum rod is attached to the
distance d is decreased below 70 mm. Similar observations lower end of the rod. An identical disk magnet is attached to
are also reported for a magnetoelastic pendulum.7 the table directly below the pendulum’s vertical position.
398 Am. J. Phys., Vol. 65, No. 5, May 1997 Siahmakoun, French, and Patterson 398
Downloaded 20 May 2013 to 129.174.21.5. Redistribution subject to AAPT license or copyright; see http://ajp.aapt.org/authors/copyright_permission
Fig. 13. Experimental amplitude curve displaying bistable states for driving
frequencies 1.1–1.35 Hz and hysteresis. m points represent the increasing
frequency ~forward! path; s points represent the decreasing frequency ~re-
turn! path.
again and change sides. This happens when the restoring and
driving torques are in step with each other such that the
resultant can overcome the repulsive magnetic torque. The
pendulum displays now two new equilibrium positions, one
on each side of the vertical position, and the orbits change
randomly between oscillations about either one of these po-
sitions and oscillations about the vertical position. Further
decrease in d, however, results in oscillations bounded to one
side of the vertical equilibrium as pointed out in Sec. II.
Notice the drastic change in the pendulum’s trajectories pro-
duced by only a small change in the distance d. All these
features are also found in the computer simulations ~see
Fig. 12. Experimental plots showing chaotic motion for d555.10 mm. Figs. 2–7!.
Using the same experimental setup we are also able to
study amplitude jumps and hysteresis. The angular amplitude
The two magnets are oriented with like poles facing each of the pendulum is plotted as a function of the frequency of
other. The pendulum is suspended from an aluminum rod the driver, while the driver’s amplitude is kept at its maxi-
which is placed on a three-way precision translation stage so mum value and the distance between the two magnets is kept
that the separation distance between the two magnets can be constant at 70 mm. A hysteresis curve is obtained, as shown
varied in steps as small as 10 microns. in Fig. 13. Similar observations are also reported for an in-
Starting with the minimum separation between the two verted pendulum.9
magnets as the control parameter, the driver’s amplitude and Notice that the amplitude of oscillation increases continu-
frequency are kept constant at 5 mm and 1.33 Hz, respec- ously as the frequency increases ~m data points! up to a
tively. All the experimental measurements are taken after the frequency of 1.35 Hz. At this point the amplitude drops dis-
transient behavior of the pendulum is allowed to die out continuously from 0.59 to 0.03 rad. After this point a con-
~after approximately 30 cycles!. Thus the zero on the time tinuous decrease in amplitude is observed as the driving fre-
scale of the experimental plots does not represent the starting quency is increased. The frequency is then decreased through
moment of the pendulum motion. The experimental data the same range starting from 1.5 Hz ~s data points constitute
show similar features of the dynamics that were found in the the return path!. At first the amplitude increases following
computer simulations. Figure 9 shows a periodic motion for the lower branch of the return curve. When the decreasing
a distance d of 70 mm. For distances d in the range 56–55 frequency reaches a value of 1.1 Hz, there is a second dis-
mm, the pendulum displays chaotic motions. Figures 10–12 continuous jump in amplitude from 0.08 to 0.18 rad. After
display the pendulum’s motions at a few distances d in this 1.1 Hz the amplitude decreases continuously with decreasing
range. As can be seen from the time series and phase space frequency as shown in Fig. 13.
plots, the pendulum executes a few oscillations on both sides The above results lead to two observations: ~1! for driving
of the vertical equilibrium position, after which it is trapped frequencies between 1.1 and 1.35 Hz the amplitude curve is
on one side for a few moments. It then suddenly overcomes bistable; ~2! the amplitude of the pendulum is dependent not
the magnetic field and is trapped on the opposite side for a only on the range of the driving frequencies but also on the
few cycles after which is able to overcome the magnetic field history of that frequency range.
399 Am. J. Phys., Vol. 65, No. 5, May 1997 Siahmakoun, French, and Patterson 399
Downloaded 20 May 2013 to 129.174.21.5. Redistribution subject to AAPT license or copyright; see http://ajp.aapt.org/authors/copyright_permission
IV. CONCLUSIONS perimental setup is simple and inexpensive to build, making
it a suitable and affordable experiment for an advanced un-
dergraduate physics laboratory.
The physical pendulum in a repulsive magnetic field pre- 1
T. Shinbrot, C. Grebogi, J. Wisdom, and J. A. Yorke, ‘‘Chaos in a double
sented here is a system that exhibits rich nonlinear dynamics. pendulum,’’ Am. J. Phys. 60, 491–498 ~1992!.
Periodic and chaotic behaviors are investigated for different 2
R. B. Levien, and S. M. Tan, ‘‘Double pendulum: An experiment in
values of the control parameters ~driver frequency and mini- chaos,’’ Am. J. Phys. 61, 1038–1044 ~1993!.
3
mum separation between two magnets!. The characteristic B. Duchesne, C. W. Fischer, C. G. Gray, and K. R. Jeffrey, ‘‘Chaos in the
sensitive dependance of the pendulum’s chaotic motion on motion of an inverted pendulum: An undergraduate laboratory experi-
ment,’’ Am. J. Phys. 59, 987–992 ~1991!.
the initial conditions is also demonstrated. Also presented 4
H. J. T. Smith, and J. A. Blackburn, ‘‘Experimental study of an inverted
here are the characteristic features of nonlinear systems such pendulum,’’ Am. J. Phys. 60, 909–911 ~1992!.
5
as bistable states, amplitude jumps, and hysteresis. The ex- James A. Blackburn, H. J. T. Smith, and N. Gronbech-Jensen, ‘‘Stability
perimental data are acquired in real time using a shaft and Hopf bifurcations in an inverted pendulum,’’ Am. J. Phys. 60, 903–
encoder–PC system. The plots generated from these data ex- 908 ~1992!.
6
V. H. Schmidt, and B. R. Childers, ‘‘Magnetic pendulum apparatus for
hibit the same features as those predicted by the theoretical analog demonstration of first-order and second-order phase transitions and
simulations. The experimental setup has up to five control tricritical points,’’ Am. J. Phys. 52, 39–43 ~1984!.
7
parameters that are experimentally accessible in real time: F. C. Moon and P. J. Holmes, ‘‘A magnetoelastic strange attractor,’’ J.
the frequency and the amplitude of the driver, the minimum Sound Vib. 65, 275–296 ~1979!.
8
Trevis J. Litherland, and Azad Siahmakoun, ‘‘Chaotic behavior of the
separation between the magnets, and the relative position of
Zeeman Catastrophe Machine,’’ Am. J. Phys. 63, 426–431 ~1995!.
the magnets in a horizontal plane. The system holds much 9
N. Alessi, C. W. Fischer, and C. G. Gray, ‘‘Measurement of amplitude
promise for further studies since not all of the above control jumps and hysteresis in a driven inverted pendulum,’’ Am. J. Phys. 60,
parameters are examined in this paper. Furthermore, the ex- 755–756 ~1992!.
400 Am. J. Phys., Vol. 65, No. 5, May 1997 Siahmakoun, French, and Patterson 400
Downloaded 20 May 2013 to 129.174.21.5. Redistribution subject to AAPT license or copyright; see http://ajp.aapt.org/authors/copyright_permission