Unit 2 - Fayetteville Freethinkers

Unit 2 - Fayetteville Freethinkers
© 2008 HarterSoft –LearnIt
Physics by Geometry
Unit 2. Vibration, Resonance, and Wave Motion
C2 r0 r1
02
12
0
0
180
0
C3 r
03
13
23
0
r
0
1
r
0
2
0
-120
120
0
-120
120
0
C24 r 0 r 1 r 2 r 3 r 4 r 5 r 6 r 7 r 8 r 9 r10 r11 r12 r13 r14 r15 r16 r17 r18 r19 r20 r21 r22 r23
0 24
1 24
2 24
3 24
4 24
5 24
6 24
7 24
8 24
9 24
1024
1124
1224
1324
1424
1524
1624
1724
1824
1924
2024
2124
2224
2324
An Honors Course of Study
by
Bill Harter
Department of Physics
University of Arkansas
Fayetteville
Hardware and Software by
HARTER-Soft
Elegant Educational Tools Since 2001
© 2008 W. G. Harter
Introduction
Unit. 2
Road Map for Physics Colloquium (3/25/03)-version
Newton-Hamilton Classical World
(Think Bang-Bang particles ! Waves are Illusory.)
Concepts and Effects
Velocity (1-and-2-particles)
Momentum (1-and-2-particles)
Kinetic Energy
Potential Energy
Force
Force & PE Fields
(n-particles)
Friction
Ways to visualize
Velocity-velocity plots
Space-time and space-space
Energy shell (ellipse)
Matrix rotation & reflection
F versus space
PE versus space
Multi-velocity diagrams
Maxwell-Lorentz View of Classical and Quantum Worlds
(Think resonance! Nature works by persuasion)
Vibration
Action and phase
Hysteresis
1-Particle resonance
2-Particle resonance
n-Particle resonance (Waves)
Phasors
Momentum versus coordinate plot
F versus time & work plots
Lorentzian functions
Smith charts
Wave dispersion
Frequency vs. wavevector plots
Multiple Phasors
Einstein-Planck Relativity-Quantum World
(Think waves! Particles are Illusory.)
Space-Time by Wave Interference
Minkowski Spacetime graph
Doppler shifts tell all
Hyperbolic geometry
Energy Momentum Dispersion
Epstein x-proper time graph
Matter waves vs. No-Matter waves
What is matter?
Dispersion graphs
Why Schrodinger was wrong Wavevector geometry
Waves in accelerated frames
Angular momentum cones
Waves in nano-structures
RE surfaces
Spin and quantum angular momentum
The U(2) spinor slide rule
Correlation (“Entanglement”)
Bose vs. Fermi
2
© 2008 HarterSoft –LearnIt
Unit 2. Vibration, Resonance, and Wave Motion
Chapter 2 - 3
Euclid to Einstein and Beyond: Geometry and Physics
UNIT 2. VIBRATION, RESONANCE AND WAVE MOTION ....................................................... 5
Introduction............................................................................................................................................................................................5
-- The Purest Light and a Resonance Hero – Ken Evenson (1932-2002) --.....................................................................................7
Chapter 1 Keep the phase, baby! Resonant energy transfer ...........................................................................................................9
(a) Swing-hi-swing-lo: A pendulum oscillator..................................................................................................................................9
(b) Energy transfer: Weakly coupled pendulum oscillators............................................................................................................10
The phase-lag-sine-sign-rule: To lead takes energy! .................................................................................................................13
(c) Boxing ellipses ............................................................................................................................................................................17
(d) Relative phase angle versus relative tipping angle . ............................................................................................................21
Chapter 2 Ellipse orbit vector and matrix relations .......................................................................................................................23
(a) Vector mechanics of elliptic oscillator orbits .............................................................................................................................23
(b) Ellipses and matrix operations: Quadratic forms.......................................................................................................................25
Tipped boxes and phasor clockwork...........................................................................................................................................27
Special Tippings: a/b and 1/1 ......................................................................................................................................................31
Chapter 3 Strongly Coupled Oscillators and Beats ........................................................................................................................33
(a) Elliptical potential energy bowls ................................................................................................................................................33
(b) Gradients and modes...................................................................................................................................................................35
(c) Superposition principle and beats...............................................................................................................................................37
Chapter 4 Complex phasor analysis of oscillation, beats, and modes...........................................................................................39
(a) Introducing complex phasors and wavefunctions: Real and imaginary axes ...........................................................................39
(b) Coupled oscillation between identical pendulums ....................................................................................................................40
(c) Summing phasors using Euler’s identity: Beats ........................................................................................................................41
Transverse or Longitudinal?........................................................................................................................................................41
(c) Slow(er) Beats .............................................................................................................................................................................43
(d) Geometry of resonance and phasor beats...................................................................................................................................45
(e) Adding complex phasors ............................................................................................................................................................47
Chapter 5 Quaternion-Spinor Analysis of Oscillators and Beats..................................................................................................51
A classical analog of Schrodinger dynamics .........................................................................................................................51
ABCD Symmetry operator analysis and U(2) spinors ..........................................................................................................52
(a) How spinors and quaternions work ............................................................................................................................................53
The “mysterious” factors of 2 .....................................................................................................................................................55
2D polarization and 3D Stokes vector S.....................................................................................................................................55
Fixed points: A port in the storm of action .................................................................................................................................56
(b) Oscillator states by spinor rotation.............................................................................................................................................57
The A-view in {x1,x2}-basis........................................................................................................................................................59
© 2008 W. G. Harter
Introduction
Unit. 2
4
The C-view in {xR,xL}-basis .......................................................................................................................................................60
(c) How spinors give eigensolutions (Gone in 60 seconds!)...........................................................................................................62
(d) How spinors give time evolution ...............................................................................................................................................63
B-Type Oscillation: Simple examples of balanced beats ...........................................................................................................65
Chapter 6 Multiple Oscillators and Wave Motion ..........................................................................................................................69
(a) The Shower Curtain Model.........................................................................................................................................................69
Nth Roots of unity ........................................................................................................................................................................70
(b) Solving shower curtain models by symmetry............................................................................................................................70
(c) Wave structure and dynamics .....................................................................................................................................................73
Distinguishing and *: Conjugation and time reversal.........................................................................................................73
(d) Wave superposition.....................................................................................................................................................................75
Wave phase velocity ....................................................................................................................................................................75
Group velocity and mean phase velocity....................................................................................................................................77
Wave-zero (WZ) and pulse-peak (PP) space-time coordinate grids...............................................................................................79
© 2008 HarterSoft –LearnIt
Unit 2. Vibration, Resonance, and Wave Motion
Chapter 2 - 5
Unit 2. Vibration, Resonance and Wave Motion
Introduction
Unit 1 began with a car crash. Bang! Wham! Pow! Two superballs careen off each other onto the
floor or ceiling. Huge forces quickly transfer enormous amounts of momentum and energy between the
participants. Seeing this may be a little like watching violent US movies and TV. It looks like the world is
just a lot of punching bags endlessly pounding each other.
Now we consider a view of the world that is more like French and English movies. Here the
participants rarely if ever actually hit each other. Instead they sit for long periods of thought and reflection
and engage in the gentle art of persuasion and being persuaded.
Thought! Reflection!? Mon Dieu! How unpatriotic! But, as we hope to show, this is a far better
analogy for picturing the physics of the world than is an endless series of boring car crashes. The gentle
art of persuasion found deep inside our physics is called resonance and resonance is the single most
important physical process in the entire world, as we presently know it. Let’s think about it.
We hear by resonance, we speak using resonance, and we see only by having delicate resonant
processes in our eyes. Without resonance to amplify tiny vibrating forces, we are blind, deaf, and dumb.
Resonance has been necessary to run our AC power grid. Communications have relied on resonance since
early telegraphs. Without resonance there are no radios, radar, computers, TV (God forbid!), or cell
phones. (God doubly forbid!) Recently, ultra-accurate clocks and the GPS (global positioning system)
rely on resonant process of unimaginable quality and finesse.
Still, all that is not the half of it. What this unit is preparing are ways to see the role of
resonance at the fundamental quantum level where everything, and we do mean everything is resonating.
Simply put, if something doesn’t resonate, it just doesn’t exist! So here we will also be introducing the
most basic process of quantum or Planck-DeBroglie mechanics. We start by using classical coupled
pendulums as one of our many analogies to quantum resonance.
Resonance requires oscillation. The simplest oscillation is harmonic oscillation, that is, oscillation
whose frequency is the same for all amplitudes. So, the simplest resonance requires harmonic oscillation.
Otherwise, it’s usually too hard to stay in tune! Our inside-Earth-orbiting neutron starlet (Sec. 1.6(e)) and
inside-asteroid-orbiting “astroidonaut” did harmonic oscillation as do sub-superballs with linear force
functions F(r)=-kr. (Recall, that linear-force RumpCo. sub-balls refuse to make car-crashes spectacular in
Fig. 1.7.7. They’re looking for some loving resonance, not just a bang!)
© 2008 W. G. Harter
Introduction
Unit. 2
6
As we will see resonance can give a response millions or billions of times more than one lousy
bang. The whole world depends on this beautiful and fascinating process. So get ready to learn about
some things that are very fundamental. Also, prepare to see some very applied physics.
The engineering applications of resonance involve its ability to store and amplify signals, waves,
and energy. If the energy comes in the form of sound waves their properties are described by one of the
oldest physical sciences, that of acoustics. Long before electronic amplifiers were invented, the great
churches and concert halls relied on acoustics to design better ways for speakers and performers to be
heard. Controlling resonance was and still is an important part of good acoustics.
But even before we could assemble crowds in a hall, we had to evolve ways to speak. Again,
resonance is essential. To illustrate this, try to pronounce the word “good” while smiling. You’ll find it
much easier to say “bad” with a smile, but the word “good” comes as “gud” or “gad” and more like a
Southern drawl than proper English!
The reason for this is that “good” requires a larger lower frequency component that in turn needs
the amplification provide by an elongated mouth formed to a tubular resonator. It is unfortunate for all the
smiling car dealers, and perhaps the whole human race, that the physics of resonance prevents “good”
from coming out right with a smile!
The word “good” sounds fine if one’s mouth is shaped as it is when beginning a kiss! Perhaps
reproductive persuasive dynamics trump those of diplomacy in linguistic evolution. Certainly insect,
animal, and bird calls use it more for reproductive advantage than for diplomacy or business. These are
but a few examples of how our world is built on resonant processes that are often not so obvious.
But, at the deepest levels of nature resonance seems to be the main game in town. And unlike
practically any game we’ve ever imagined this one seems, at first, less obvious than anything ever has
been before! But after you see how it works you will wonder why it took us so long to “get it.”
So get ready for the first secrets of the ultimate cosmic quantum game.
© 2008 HarterSoft –LearnIt
Unit 2. Vibration, Resonance, and Wave Motion
Chapter 2 - 7
-- The Purest Light and a Resonance Hero – Ken Evenson (1932-2002) -When travelers punch up their GPS coordinates they owe a debt of gratitude to an under sung hero
who, alongside his colleagues and students, often toiled 18 hour days deep inside a laser laboratory lit
only by the purest light in the universe.
Ken was an “Indiana Jones” of modern physics. While he may never have been called “Montana
Ken,” such a name would describe a real life hero from Bozeman, Montana, whose extraordinary
accomplishments in many ways surpass the fictional characters in cinematic thrillers like Raiders of the
Lost Arc.
Indeed, there were some exciting real life moments shared by his wife Vera, one together with Ken
in a canoe literally inches from the hundred-foot drop-off of Brazil’s largest waterfall. But, such outdoor
exploits, of which Ken had many, pale in the light of an in-the-lab brilliance and courage that profoundly
enriched the world.
Ken is one of few researchers and perhaps the only physicist to be twice listed in the Guinness
Book of Records. The listings are not for jungle exploits but for his lab’s highest frequency measurement
and for a speed of light determination that made c many times more precise due to his lab’s pioneering
work with John Hall in laser resonance and metrology†.
The meter-kilogram-second (mks) system of units underwent a redefinition largely because of
these efforts. Thereafter, the speed of light c was set to 299,792,458ms-1. The meter was defined in terms
of c, instead of the other way around since his time precision had so far trumped that for distance. Without
such resonance precision, the Global Positioning System (GPS), the first large-scale wave space-time
coordinate system, would not be possible.
Ken’s courage and persistence at the Time and Frequency Division of the Boulder Laboratories in
the National Bureau of Standards (now the National Institute of Standards and Technology or NIST) are
legendary as are his railings against boneheaded administrators who seemed bent on thwarting his best
efforts. Undaunted, Ken’s lab painstakingly exploited the resonance properties of metal-insulator diodes,
and succeeded in literally counting the waves of near-infrared radiation and eventually visible light itself.
Those who knew Ken miss him terribly. But, his indelible legacy resonates today as ultra-precise
atomic and molecular wave and pulse quantum optics continue to advance and provide heretofore
unimaginable capability. Our quality of life depends on their metrology through the Quality and Finesse
of the resonant oscillators that are the heartbeats of our technology.
Before being taken by Lou Gehrig’s disease, Ken began ultra-precise laser spectroscopy of
unusual molecules such as HO2, the radical cousin of the more common H2O. Like Ken, such radical
molecules affect us as much or more than better known ones. But also like Ken, they toil in obscurity,
illuminated only by the purest light in the universe.
In 2005 the Nobel Prize in physics was awarded to Glauber, Hall, and Hensch†† for laser optics
and metrology.
† K. M. Evenson, J.S. Wells, F.R. Peterson, B.L. Danielson, G.W. Day, R.L. Barger and J.L. Hall,
Phys. Rev. Letters 29, 1346(1972).
†† The Nobel Prize in Physics, 2005. http://nobelprize.org/
© 2008 W. G. Harter
Introduction
Kenneth M. Evenson – 1932-2002
Unit. 2
8
© 2008 HarterSoft –LearnIt
Unit 2. Vibration, Resonance, and Wave Motion
2-
Chapter 1 Keep the phase, baby! Resonant energy transfer
Instead of trying to transfer energy in a single bang like a superball collision, let’s se how it’s done
steadily by resonance. Imagine two big pendulum bobs, each of mass M = 1000 kg, hanging from ropes,
each of length L=10 meters, and swinging back and forth with an amplitude of A1=10 cm= A2, or so.
First, let’s review the properties of a single pendulum, the common tree swing.
(a) Swing-hi-swing-lo: A pendulum oscillator
If you’ve ever played with a big tall pendulum you may have noticed how easy it is to push it a
little bit off center. (Recall or imagine pushing someone bigger than you on a tall swing.) If you keep
pushing off-and-on in synchrony with the swing, pretty soon it’s zooming. That’s resonance! If the swing
is very very tall (L) it’s like having a frictionless track. (That’s zero-frequency resonance!) Resonance
frequency of an L-long pendulum of mass M depends on its force F(x) or energy U(x) at small distances
x off-center. We now do some geometry using energy in Fig. 1.1a and force vectors in Fig. 1.1b.
In Fig. 1.1a, gravity’s potential energy Mgh varies with horizontal push distance x as seen by a
geometric mean construction. (Recall Fig. 1.6.12(a).) It has a quadratic x-potential for low x. (x<<L)
U(x) = Mgh =
1 Mg 2
x
2 L
(1.1)
A quadratic potential U(x)=1/2kx2 has the linear force F(x)=-kx of a harmonic oscillator with k=Mg/L.
F(x) = dU(x)
Mg
= x
dx
L
(1.2)
Oscillator angular frequency follows using the “starlet” formula =k/m from (1.6.42).
=
k
=
M
g
= 2 ,
L
=
1
2
g
=
2
L
(1.3)
The frequency of an L=10 meter swing is about 1/6.3 Hz, or =6.3 seconds per swing. It’s independent
of mass M, and for low x, independent of amplitude x, too. Fat people swing as fast and high as skinny
ones. Galileo made a big deal of this, but probably not because he was fat. It’s important for resonance to
have frequency not change with amplitude. It meant pendulums could (and did) improve clock precision.
In Fig. 1.1b, gravitational pendulum restoring force varies linearly with horizontal push distance x
as seen by the vector construction. But, like Fig. 1.1a, it’s linear only if amplitude x is a lot smaller than
lever length L. While the formula F(x)=-(Mg/L)x for force along the circular path is exact, that path is not
linear and distance x=Lsin never exceeds L. The exact angular force function : F =-Mgsin is nonlinear. Only for small angles (sin~) does it approach the desired linear one (F ~-Mg).
© 2008 W. G. Harter
Chapter 1. Resonance phase relations
Unit. 2
Energy geometry
10
Force geometry
L
L
L
L
x=L sinΘ ~ L Θ
L
x=LsinΘ
x
h
Θ
h
x2=h(2L-h) ~ 2hL
1/ (Mg/L)x2 ~ Mgh
2
Fig. 1.1(a) Pendulum energy mean geometry
-Mg sin Θ =F
Θ Mg
F=-Mg sinΘ =-(Mg/L) x
Fig. 1.1(b) Pendulum force vector geometry
(b) Energy transfer: Weakly coupled pendulum oscillators
Imagine we have a pair of identical big and tall swinging pendulums connected by a very weak
rubber band or spring as shown in Fig. 1.2 below. How might the tiny spring start to drain energy out of
one pendulum and into the other? At what rate is energy transferred? Through this simple example we see
a little-appreciated principle of resonance with great and universal significance.
Here we assume the connecting spring constant k12 is small as is its transmitted force F12.
F12 =Fon 1 due to 2 = -k12(x1-x2) = - Fon 2 due to 1
(1.4)
Each pendulum swings at frequency . amplitudes A1 and A2 and phases 1 and 2 are approximately
constant. Only the phase difference or relative phase = 1 - 2 is of interest here so we stick it on x1.
x1= A1 cos( t-) , (1.5a)
x2= A2 cos( t-0 ) ,
(1.5b)
The results are plotted in Fig. 1.3(a-d) for relative phase angles = 0°, 70°, 180°, and –60°. Note the
differences in the ellipse plots of x2 versus x1 like orbit plots in Unit 1 Fig. 9.10. Here the area of the
ellipse determines the work done by x2 on x1. Energy or work is area F12 dx1 under a force function as
given by (7.5a) in Unit 1. Here power loop integral F12 dx1 repeats with each oscillator swing cycle giving
k12 times the area of the x2(x1)-ellipse. For now we say the k12-spring is too weak for this to have a
noticeable effect on amplitudes A1 or A2 or phase lag angle . In Ch. 5 we find what that effect is.
© 2008 Harter
Chapter 2 - 11
Unit 2. Vibration, Resonance, and Wave Motion
Before studying that force and energy dynamics, let us get some plots of the position versus time
curves for the two connected pendulums such as are shown in Fig. 1.2. These are just sine (or cosine)
curves shifted by their respective phase lags. You should notice that a function cos( t-) is shifted /
radians back in time. This is unlike most function translation where f(x-a) shifts f(x) forward by a units in
x. It is due to the fact that we make our clocks go clockwise which is a negative angular rotation. The
phase space plot is made to have positive velocity V/-axis be up.
k12
v1 /ω
v2 /ω
F
ρ
k12
L
x1
v1 /ω
F
ρ
L
x2
ρ
ρ=30
phase lag
ρ
x1
Time
t
v2 /ω
F
L
x1
F
ρ
L
x2
ρ=60
phase lag
ρ
ρ
ρ
ρ
x2
x1
x2
Time
t
Fig. 1.2 Phasor and time plots of phased pendulum oscillation (a) =30° phase lag, (a) =60° phase lag
© 2008 W. G. Harter
Chapter 1. Resonance phase relations
a. In phase (ρ = 0 )
k12
v1 /ω
x1
c. Out of phase (ρ = )
b. Lagging (ρ=70 )
k12
v2 /ω
x2
v2 /ω
v1 /ω
d.Leading (ρ =-60 )
k12
x1
x2
12
Unit. 2
v1 /ω
x1
k12
x2
v2 /ω
v1 /ω
x1
x2
v2 /ω
70
x1
x2
x1
x2
x2
zero
area
x2
x2
positive
area x1
x1 v /ω
2
x2
v2 /ω
180
x1
x2
zero
area
-60
x2
x2
x1
x2
x2
negative
area x
1
x1 v /ω
2
x2
v2 /ω
Fig. 1.3 Four classes (a) “in”, (b) “lag”, (c) “out”, (d) “lead” of relative phase for two oscillators.
As long as the big pendulums swing with the same amplitude and phase, they will transfer the same
amount of energy each cycle. If x1’s phasor clock lags behind that of x2 by any angle from >0 to <,
then pendulum-1 gains while pendulum-2 loses, but if x1’s phasor leads (- >>0), then pendulum-1
becomes a donor and pendulum-2 is a receiver of energy. Only for the cases of in-phase motion (=0) in
Fig. 1.3(a) or -out-of-phase motion (=) in Fig. 1.3(c), is no energy transferred.
Zero energy transfer seems strange (particularly for =) given that the connecting spring is
constantly pushing or pulling. Fig. 1.4(a) shows how energy gained during the upstroke is taken back by
the return stroke. Only if lag angle is between and zero, as in Fig. 1.4(b), does the returning spring leave
behind some of the energy it gained coming in. (x2dx1 is minus if x2 or else dx1 is negative.)
(b) ρ=135 Lagging phase case
What k12 giveth... ... k12 mostly leaveth
x2
x2
[Upstroke]
(a) ρ=180 Out of phase case
What k12 giveth... ... k12 taketh away
x2
x2
[Upstroke]
(-)
(+)
area
(-)
area
x1 area
x1
(+)
area
[Downstroke]
(+)
area
x1
(-)
area
(-)
area
[Downstroke]
x1
(+)
area
Fig. 1.4 Work-cycle area plots for (a) Out-of-phase (=180°) case (b) Lagging phase (=135°) case.
© 2008 Harter
Chapter 2 - 13
Unit 2. Vibration, Resonance, and Wave Motion
The phase-lag-sine-sign-rule: To lead takes energy!
If pendulum-1 leads pendulum-2, as in Fig. 1.3(d), its work-cycle area is swept out in an anticlockwise direction giving an energy deficit for pendulum-1 who now pays pendulum-2 each cycle. The
relation between lag angle in (1.5b) and sign and value of energy flow needs to be derived. First, we do
an algebraic approach. Work integral W12 (t) = F12 dx1 is as follows using (1.4) and (1.5).
t
t
0
0
Work(t ) on1 by 2 = Fon1 by 2dx 1 = F12 v1dt = k12( x 1 x 2 )v1dt where: v 1 =
Work( ) on 1by 2 = K k12 A1 A 2
2 /
2 /
cos( t ) sin( t )dt = k12 0
0
dx1
= A1 sin( t )
dt
(1.6a)
2
is period
(1.6b)
x 2dx 1 where: =
The K term for pendulum-1 integral x1dx 1 is, for a full period , exactly zero for any phase lag angle .
K = k12
2 / dt A1 cos( t ) sin( t ) =
2
0
2 /
1
2
k12 A1 dt sin 2( t ) = 0
2
0
Now sin( t ) = sin t cos cos t sin gives a zero integral plus a sin times an integral
2 /
Work( ) on 1by 2 = k12 A1 A2 0
(1.7)
2/
0
2
cos t = / cos( t)(sin t cos cos t sin )dt = k12 A1 A 2 sin .
(1.8)
Energy flow per cycle goes as a sine of phase lag angle . Compare it to oscillator total energy 12 A 2 .
Energy pendulum1 =
1
1
k A 2 = M 2 A12
2 1 1 2
1
2
1
2
(1.9a) Energy pendulum 2 = k 2 A22 = M 2 A 22
(1.9a)
Power transfer is work-per-cycle (1.8) times number of cycles per second or frequency =2.
Powerto 1 from 2 = 2 2 k12 A1 A 2 sin (1.10)
Such a sweet and powerful relation deserves a geometric construction. An attempt is made in Fig.
1.5 for the case of equal amplitudes A1=A=A2. Each step draws a different time snapshot in a cycle in
which “Follower” pendulum-1 lags behind the “Leader” pendulum-2 by a constant angle =60°. At first,
the fearless Leader L is stationary (v2=0) at its maximum righthand point x2=A and tugging on Follower F
who is back at x1=Acos with velocity v1=A sin . A little later at time t=/2 the Leader is starting back
with velocity v2=-A sin/2 at x2=Acos/2. Then x2 equals the off-center distance x1=Acos/2 of the follower
who is coming forward with velocity v2=+A sin/2. This is the work-ellipse +45° apogee point at major
axis a=2Acos/2. There the k12 spring stops pulling and starts pushing. Later at t=/ comes a –45° perigee
point or minor axis b=2Asin/2. The area of the resulting ellipse is ab. Work is k12 times this.
2
2
Work( ) on 1by 2 = k12 ab= k12 2A sin/2 cos/2 = k12 A sin (1.11)
This geometric result equals the algebraic one in (1.8) for A1=A=A2. But, how do we do a A1A2 case?
© 2008 W. G. Harter
Chapter 1. Resonance phase relations
Step 1a. Establish lag-angle ρ
of Follower F behind Leader L k
12
with Leader L on x-axis
F
x1
ρ
v2 /ω
v1 /ω
F
x2
ρ
F
L
Step 3b. Locate right
contact point Crt
for ellipse
x2
F
x2
x2
L
ρ
v2 /ω
Step 5.Repat as necessary
to complete work-cycle
ellipse
Cb
Minor axis b
b=A 2 sin(ρ/2)
a
Major axis
a=A 2 cos(ρ/2)
Fig. 1.5 Work-cycle geometry for two weakly-coupled oscillators
L
Asin(ρ/2)
L
x2
ρ
Step 4b.
Locate diagonal
contact point
Cb for ellipse
x1
Crt
x1
x1
ρ
L
L
ρ
Step 4a. Rotate Follower F and Leader L
until their lag-angle ρ is bisected by v-axes
v2 /ω
F
x1
F
L
x2
x2
F
ρ
Ca Acos(ρ/2)
v2 /ω
Time
t=ρ/ω
L
Step 2b.
Locate diagonal
contact point Ca
for ellipse
x2
Step 3a. Rotate Follower F and Leader L
until the Follower F is on the x-axes
v1 /ω
ρ
x1
L
Step 1b. Locate top
contact point Ctop
for ellipse
x1
v2 /ω
L
ρ
L
x2
Time
t=ρ/2ω
F
x2
ρ
L
x1
F
x2 Ctop
v1 /ω
F
v2 /ω
ρ
v2 /ω
Time
t=0
v2 /ω
F
Step 2a. Rotate Follower F and Leader L
until their lag-angle ρ is bisected by x-axes
L
x2
v1 /ω
14
Unit. 2
F
ρ
L
© 2008 Harter
Chapter 2 - 15
Unit 2. Vibration, Resonance, and Wave Motion
Also, shouldn’t work stop increasing and start decreasing after the apogee when k12(x1-x2) goes
negative? These two questions are related and help us understand an oscillator energy shell game that
might seem, at first, to be as convoluted as an Enron accounting scandal.
First, the Work( ) on 1by 2 expression averages a whole period =2/ ignoring W fluctuations.
t
t
t
0
0
0
Work(t ) on1 by 2 = k12 x 2dx 1 k12 x 1dx1 = k12 x 2dx 1 k12
x 12
2
(1.12)
The K-term K = k12 x 1dx1 = k12 x 12 / 2 in (1.6b) adds an oscillation to the elliptic loop area A = k12 x 2dx 1 as
seen in Fig. 1.6 below. Work(t ) on1 by 2 in (1.12) has both K and A terms. The K-term subtracts a 45° right
triangle just so that when k12(x1-x2) changes sign, as it’s doing in Fig. 1.6(a), so also does the growth of
instantaneous work W=K+A. Total work W goes down briefly as it is in Fig. 1.6(b) even as ellipse area A
is still growing, but then the thieving K-triangle spits back its area in Fig. 1.6(c) and A starts grabbing area
below the x1 axis in Fig. 1.6(d). Together this makes W surge ahead in Fig. 1.6(e). Then K again eats some
more area but again spits it back. But, finally the effect of K for a full period is zero.
Elliptical area-sweeping represented by A or W does not accumulate area at a constant rate like the
Kepler area sweeps in (1.16) below that are angular sweeps by a radius line. Here the A-sweep in Fig. 1.5
is by a Cartesian vertical x2=y-line. The A-sweep first gobbles area voraciously and then gives some back
and then becomes voracious again but has to give some back, again, during each period. This bi-cyclic
binge-and-purge behavior of A is tempered somewhat by the thieving K-triangle which partially “starves”
A and spits energy back with almost, but not quite, the same bi-cyclic schedule.
As we see in Fig. 1.6, K is too late each time to make W eat its area at a constant rate. To see this
with algebra, we work out the W-integral (1.6) or (1.12) completely as a function of time.
t
t
0
0
Work(t ) on1 by 2 = k12 A12 cos( t ) sin( t )dt k12 A1 A2 cos( t) sin( t )dt
t
A A
A2
= k12 A1 A2
sin + k12 1 2 cos(2 t ) k12 1 cos(2 t 2 ) + const.
2
4
4
(1.13)
(The constant term is k12(A12-A1A2)/4.). The first term is the constant area growth that a Kepler ellipse might
predict. With A1=A2, the oscillating terms might cancel each other were not that the K-part (last term) is
late by 2 or twice the phase lag of follower x1 and the other term. So feast-to-famine is unavoidable.
Finally, let us see how geometry treats the case of unequal (A1A2) amplitudes. We have been
plotting force F12 versus x1 to represent work by area under the curve. It happened that the force was
proportional to the coordinate x2 so we plotted that instead. But, when you plot things of different
dimension you may scale them however you wish. This means you can turn what might have been an
ellipse into a circle as we do for our phasor plots. We may do the same with F12 versus x1 plots.
© 2008 W. G. Harter
(a)
Chapter 1. Resonance phase relations
W
=
A
+
Unit. 2
16
K
-1/ k x 2
2 12 1
(b)
W
=
A
+
K
-1/ k x 2
2 12 1
(c)
(d)
W
=
A
+
K
W
=
A
+
K
Work W
or
Area A
(e)
W
=
A
+
K
A
(f)
W
=
A
+
K
W
K
Time
t
Fig. 1.6 Work-area F12 versus x1 plot-W made from sum of x2 versus x1 plot A and x1 versus x1 plot K.
An ellipse-in-a-square has its axes on the ±45° diagonals of the square. But, rescaling a dimension
of an ellipse-in-a-square results in an ellipse-in-a-rectangle whose axes are not on a rectangle diagonal.
© 2008 Harter
Chapter 2 - 17
Unit 2. Vibration, Resonance, and Wave Motion
This is shown by Fig. 1.7(b) in which the A1 amplitude is twice that of the A2 amplitude, and the
ellipse axis is clearly off the rectangle’s diagonal. Calculating the new ellipse inclination is an interesting
geometry problem that we will consider below. But, this isn’t needed to see that the lag-sine-rule (1.8) is
derived by geometry for any amplitude A1 or A2. Scaling one amplitude scales area by the same amount.
(c) Boxing ellipses
You can put a rectangular box at any angle relative to an ellipse so that all four sides of the box are
ellipse tangents as shown in Fig. 1.8. If the ellipse has major-by-minor (a-by-b)-radii then the box parallel
to ellipse axes is a (2a-by-2b)-rectangle as indicated in Fig. 1.8(a). A box tipped at =30° to the ellipse
axes is fatter as shown in Fig. 1.8(b). A box at =45° to the ellipse axes is a perfect square as shown in
Fig. 1.8(c). Note, all boxes in Fig. 1.8 share a diagonal diameter R=(a2+ b2) of a circle that inscribes
them including the skinniest =0° (2a-by-2b)-box (a) or fattest =45° (2a-by-2a) box (c).
Another way to view this is to tip the ellipse while keeping it against a corner wall and floor frame
as shown in Fig. 1.9. The center of the ellipse stays on a circle of the same radius R=(a2+ b2) because
total energy must be the same no matter what box frame or we use to frame an elliptic orbit. (Energy of
an elliptic starlet orbit in Unit 1 Fig. 9.10 is the same viewed from any latitude .)
E=
1 2
mv
2 1
1
m 2 x 12
2
+
+
1 2
mv
2 2
1
m 2 x 22
2
+
1
2 1
2 1
2 1
2
2
2
m( A1 sin ( t )) + m ( A1 cos( t )) + m(A2 sin( t)) + m ( A 2 cos( t))
2
2
2
2
1
1
=
mA 2 2 sin 2 ( t ) + cos2 ( t )
+
mA 22 2 sin 2 ( t) + cos 2 ( t )
2 1
2
=
(
)
(
(1.14)
)
Each -ellipse has different box components x1=A1 cos( t-) and x2=A2 cos( t) but the same total energy.
E=
(
1
m 2 A12 + A22
2
)
=
(
1
1
m 2 R 2 = m 2 a 2 + b2
2
2
)
(1.15)
Each -ellipse orbit also has constant angular momentum L=|rxp|=m|rxv|.
L=
(
mx1
)(
v2
) (
mx2
v1
)(
)
= m A1 cos( t ) A2 sin ( t ) m A2 cos( t ) A1 sin ( t )
=
(
)
(1.16a)
m A1 A2 cos( t ) sin ( t ) sin ( t ) cos( t ) = - m A1 A2 sin Angular momentum L is proportional to ellipse area ab as is work-per-cycle in the lag-sine-rule (1.8).
L = -m A1 A2 sin = -m ab sin
= - m ab
2
(1.16b)
Area |rxv| is constant so equal area |rxdr|= |rxv|dt is swept each time interval dt. (Kepler’s law).
Phase lag angle varies with angle and is /2 for =0. Clockwise or left-handed. rotation
(viewed as negative motion in the Northern hemisphere) gives negative L in (1.16) but positive power in
(1.10). Right-handed rotation yields negative work and power. (Is there political analogy here?)
© 2008 W. G. Harter
Chapter 1. Resonance phase relations
(b) Unequal amplitudes (A1=2A2 )
(a) Equal amplitudes (A1=A2 )
v1 /ω
v2 /ω
v1 /ω
Unit. 2
ρ
ρ
x1
v2 /ω
ρ
x2
x1
x2
ρ
v2 /ω
v2 /ω
ρ
x2
x1
x1
x2
x2
x2
ρ
Fig. 1.7 Comparing equal-amplitude resonance (a) to a more generic case (b).
________________________________________________________________________________________________________________
(a) Alligned (ϕ=0)
(b) Tipped (ϕ=30 )
b R
(c) Diagonal (ϕ=45 )
ϕ=45
ϕ=30
a
R
R
(d) More tipped (ϕ=75 )
R
ϕ=75
R
Fig. 1.8 Different tipped boxes for the same ellipsehave the same diagonal radius R.
________________________________________________________________________________________________________________
(c) Diagonal (ϕ=-45 )
a
(b) Tipped (ϕ=-30 )
a
ϕ
a
b
b
b
b
(a) Alligned (ϕ=0)
a
R
R
R
R
Fig. 1.9 “Cornered” ellipse rotated by various angles maintains central radius R.
________________________________________________________________________________________________________________
18
© 2008 Harter
Chapter 2 - 19
Unit 2. Vibration, Resonance, and Wave Motion
The difference D= E1-E2 between the energy E1 in one oscillator and the energy E2 in the other is
an important quantity like total energy E= E1+E2 (1.15) and angular momentum L (1.16).
1
1
1
1
mv 12 +
m 2 x 12 mv 22 m 2 x 22
2
2
2
2
1
2
2
2
m A1 A2
=
2
D=
(
(1.17)
)
To express this in terms of the tipping angle we will make use of a geometric construction in Fig. 1.10
of an ellipse’s box tangents. The oscillation energy in the u1-axis tipped by from x1 is a sum of energy
1
/2m2(a cos)2 from x1-component a cos and energy 1/2m2(b sin)2 from x2-component b sin while
oscillation associated with the u2-axis tipped by from x2 has the sum of energy 1/2m2(a sin)2 from x1component asin and energy 1/2m2(b cos)2 from x2-component bcos .
A12 = ( a cos ) 2 + (bsin ) 2
(1.18)
A2 2 = ( asin ) 2 + ( bcos ) 2
Taking the difference (1.17) gives the third oscillator quantity D in angular form.
D=
=
(
)
(
1
1
m 2 A12 A22 = m 2 a 2 b 2
2
2
(
1
m 2 a 2 b 2
2
=0
)
) ( cos 2 sin 2 )
=
(
)
1
m 2 a 2 b 2 cos2
2
for: = 0
for: = ±
(1.19a)
(1.19b)
4
(1.19c)
The horizontal x1-oscillation along the a-axis of the ellipse has a phase angle exactly =90°
behind the vertical x2-oscillation along the b-axis. This makes the amplitudes in (1.18) add in quadrature,
that is, like the two sides of a right triangle. If x1-oscillation and x2-oscillation were in phase then we
would have A12 = ( a cos + bsin ) 2 and A2 2 = ( asin + bcos ) 2 , but if a-oscillation and b-oscillation were
180° out of phase then we would have A12 = ( a cos bsin ) 2 and A2 2 = ( asin b cos ) 2 .
For a tipped ellipse the horizontal and vertical oscillations have some phase angle other than one
of the four E, W, N, or S compass points of heading 0°, 180°, 90°, or 270°=-90°, respectively. This was
shown already in Fig. 1.3. Now we will precisely quantify just how the all-important phase lag angle varies with the tipping angle of an ellipse relative to its defining oscillators.
Chapter 1. Resonance phase relations
through center O
u
2
ac
osϕ
ϕ
bs
inϕ ϕ
ϕ
a
-a
x1
as b
inϕ co
sϕ
b
a
20
Step 2 From the center arc-off bsinϕ and bcosϕ
from one tipped axis to the other
Draw tipped
Draw tipped segment
segment
+b
for new bsinϕ
for new bcosϕ
Step 0
Draw ϕ-tipped orthogonal axes
x2
Unit. 2
+a
bc
osϕ
as
inϕ
© 2008 W. G. Harter
u
1
Step 1 Draw ϕ-tipped lines
through axis points a and b
+b
os
bc
ac
osϕ
bs
os
+a
a
inϕ ϕ
ϕ
-a
+a
bc
osϕ
in
ϕ
ac
A
1=
as
a
inϕ ϕ
in
-a
bs
inϕ
osϕ
as
ac
as bc
inϕ osϕ
b
+b
Note segments
a sinϕ and bcosϕ
as
Note segments
a cosϕ and bsinϕ
Step 3 From the ends of Step-2 arcs arc-off
the tangent radii and draw tipped box tangents
[(a
cos
ϕ) 2
+(
x2
bs
inϕ
) 2]
) 2]
u
A
2=
[(a
x1
sin
ϕ) 2
+(
b
cos
ϕ
2
u
1
Fig. 1.10 Construction of -tipped amplitudes and box-tangents for ellipse.
© 2008 Harter
Chapter 2 - 21
Unit 2. Vibration, Resonance, and Wave Motion
(d) Relative phase angle versus relative tipping angle .
Fig. 1.11 shows a horizontal x1-oscillation and vertical x2-oscillation that are 90° out of phase and
making an (a-by-b)-ellipse. This is compared directly with a =-30°-tipped u1-oscillation and
+90°=60°-tipped u2-oscillation that are 57° out of phase and making the very same (a-by-b)-ellipse. To
understand this diagram and construction is to better understand the clockwork relation between any pair
of quantum states in the universe! It’s certainly an important piece of geometry.
The chosen (a-by-b)-ellipse has a=2 and b=1. With =-30°-tipping, amplitudes (1.18) are
2
(
) (
(
) + ( bcos )
2
A12 = a cos A2 2 = a sin + bsin 2
)
2
2
2
1
3
13
= 2 + 1 = ,
2
2
4
A1 =
13
2
(1.20a)
2
2
1
3
7
= 2 + 1 = ,
4
2
2
A2 ==
7
2
With these amplitudes we calculate total energy E, energy difference D, and angular momentum L. Total
energy is proportional to the total area A12+ A22 of the two tipped phasors in Fig. 1.11 or area R2.
(
)
(
(
)
)
1
1
1
m 2 A12 + A22 = m 2 a 2 + b 2 = m 2 R 2
2
2
2
1
13
7
1
1
= m 2 + = m 2 22 + 12 = m 2( 5)
2
4 4
2
2
E=
(1.20b)
The angular momentum L and also work-per-cycle W depends on the sine of the phase-lag-angle and
while L and W, like E above, must each be the same for the two oscillator pairs, the lag angle differs.
L = -m A1 A2 sin = -m ab = W
= -m
13 7
sin = -m 2 1 = 2 ,
4 4
or: sin =
24
(1.20c)
13 7
This give a tipped lag angle of sin-1 (8/91)=56.996°. That is consistent with the angle 57° angle obtained
by geometric construction in Fig. 1.11. It’s quite a bit less than the 90° lag between a-and-b horizontalvertical oscillators that give the same ellipse.
Finally we compare the energy difference D-function (1.19) for the two pairs of oscillators.
(
)
(
)
1
1
m 2 A12 A22 = m 2 a 2 b2 cos2
2
2
1
7
13
1
=
m 2 = m 2 22 12 cos( 2 30°)
4 4
2
2
1
1
1
1
3
6
=
m 2
= m 2
= m 2 ( 3) 2
2
4
2
2
2
D=
(
)
(1.20da)
Both the D and L are functions of a relativity angle. For D it is the tipping angle in oscillator coordinate
space. For L it is the relative phase lag-angle or phase-space tipping. These are important insights.
© 2008 W. G. Harter
Chapter 1. Resonance phase relations
Unit. 2
22
a
b
a cos ϕ
b a
a sin ϕ
b sin ϕ
ϕ
Vx1 /ω
b cos ϕ
x1
ρ=90
Area
a2=4
u
2
x2
x2
ρ=90
x1
u
u
1
Vu
Vu
Ar
A 2 ea
2 =
7
4
1 /ω
Ar
A 2 ea
1 =
13
4
ρ=
5
7
b
2
a
ρ=
5
Area
b2=
1 /ω
Vx2 /ω
7
u
1
Fig. 1.11 Untipped (a-by-b) ellipse of (x1,x2)-phasors, (A1-by-A2) ellipse, =30°-tipped (u1,u2)-phasors.
© 2008 HarterSoft –LearnIt
Unit 2. Vibration, Resonance, and Wave Motion
2-
Chapter 2 Ellipse orbit vector and matrix relations
The construction in the preceding Fig. 1.11 neglects the geometry of oscillator velocity,
acceleration, and other kinematic quantities that help greatly in understanding ellipse geometry. Vector
algebra as well as matrix algebra helps to make calculation and visualization easier as shown in Unit 1 for
superball velocity ellipse mechanics (Ch.4 Fig. 4.10) and for oscillator orbits (Ch. 9 fig. 9.10).
(a) Vector mechanics of elliptic oscillator orbits
Oscillator orbits have a peculiar property. According to (1.9.5) their position r(t), velocity v(t)/,
acceleration a(t)/ 2, and jerk j(t)/ 3, etc., all follow the same elliptical
a cos( t) v(t) a sin( t)
r(t) = , = b cos t ,
( )
bsin ( t) a(t) a cos ( t)
=
,
2 bsin ( t) a cos t + 2 ,
=
bsin t + 2 2 a cos t + 2 ,
=
bsin t + 2 2 x2
a
2
+
y2
b
2
=1
path if scaled by .
j(t) asin ( t) =
,
3 bcos( t)
3 a cos t + 2 ,
=
bsin t + 3 2
(2.1)
Each is 90° ahead in angle or 90° behind in time phase. The next derivative, inauguration i(t)/ 4, in the
sequence above will be identical to position r(t). So the sequence of vectors repeats after four quadrants.
Fig. 2.1 uses an (a,b)-circle construction from Fig. 3.4 of Unit 1 to construct each quadrant vector
of position r(t) (Step-0), velocity v(t)/ (Step-1), acceleration a(t)/ 2 (Step-2), and jerk j(t)/ 3 (Step-3).
The (t+n/2)-phase radius (Step-n) is from (2.1).
In Step-1 the velocity vector v(t)/ is copied to sit head-to-tail on the position vector . It is seen
that v(t)/ is tangent to the ellipse at r(t), as it should be in order to represent velocity correctly and be the
rate of change of position. Note: acceleration a(t) equals 2 r(t) and restates F(t) = M a(t) =-k r(t).
Tangency relations for elliptical oscillator orbits apply vice-versa, that is, position vector r(t) is
tangent to the ellipse at velocity point v(t)/. Thus we can work this construction in reverse as well as
forward. The reverse construction shown at the bottom of Fig. 2.1 starts with a tangent line direction v(t),
such as the construction in Fig. 1.10, and locates the position point of contact r(t). So geometric
integration (finding r(t) given v(t)) for elliptic oscillator orbits is as easy as geometric differentiation
(finding v(t) given r(t)). (The construction in Fig. 1.10 also gives tangents for a particular tipping angle or t but only approximate location of tangent contact points.) Area of the r-v parallelogram equals |rxv|
and is proportional to the angular momentum L in (1.16) that is a Kepler-law constant.
L=|rxp|=m|rxv| = m | r || v | sin rv
© 2008 W. G. Harter
Chapter 2. Ellipse orbit relations
Unit. 2
Arc-copy
vectors r(t)
and v(t)/ω
Step 1 Advance t by 90
to find velocity vector v(t)/
Step 0 Construct
position vector r(t)
for a particular time t b
(Recall Fig. 1.9.1)
r(t)
r(t)
a
b-circle
90
v(t)/ω
ωt
Ω
24
ωt
Note that velocity vector v(t)/ω
is tangent to ellipse at position r(t)
... and vice-versa...
Steps 2-3 Advance ω t by 90 again
a-circle
to find acceleration vector a(t)/ω2
and jerk vector j(t)/ω3
Note: velocity vectors v(t) make
ellipse box tangents at r(t)
velocity
v(t)/ω
90
r(t)
ωt
position
ωt
90
acceleration
a(t)/ω2
jerk
j(t)/ω3
90
90
Reverse Construction Steps
Use velocity vectors v(t)
and (a,b) circles
to find ω t radus
velocity
v(t)/ω
90
ωt
ωt
Use ω t radus
and (a,b) circles
to find point r(t)
of tangency
r(t)
position
Fig. 2.1 Construction of ellipse box tangents using position, velocity, acceleration and jerk vectors.
© 2008 Harter
Chapter 2 - 25
Unit 2. Vibration, Resonance, and Wave Motion
(b) Ellipses and matrix operations: Quadratic forms
x2
Ellipse equation
a
2
1
2
y)• a
0
(x
+
y2
b
2
=1
can be expressed by a matrix
0 x
•
= 1 = (x
1 y b2 1/ a2
0 Q = 2 1/ b 0
x 2 x2 y2
y ) • ay = 2 + 2
b
a
b2 and vectors
x
r = = (x
y
r •Q• r =1
or:
y) .
(2.2a)
We use dot product introduced (10.30) and matrix product introduced in (5.2) of Unit 1. One advantage of
matrix notation r•Q•r=1 is it describes tipped ellipses
A B x
y)• • = 1 = (x
B C y
(x
Ax 2 + Bxy + Byx + Cy 2 = 1 = Ax 2 + 2Bxy + Cy 2 ,
Ax + By 2
2
y) • = Ax + 2Bxy + Cy
Bx + Cy or:
too.
r •Q• r =1
(2.2b)
Mathematicians call r•Q•r a quadratic form QF. Physics uses QF’s (metric ds2=gmndxmdxn, energy, etc.) a lot.
Matrix
1/ a2
0 Q = 2 1/ b 0
operates on vector
x
r = y
to give a vector p perpendicular to ellipse tangent ˙r =
x cos x = rx = a cos = a cos t
x a 2 a •
=
=
where:
y
sin 1
y = ry = bsin = bsin t
y 2
2 b
b
b
cos 1
px = cos r˙x = a sin
px a a
r˙y • = ( a sin b cos ) sin where: ˙
and:
1
ry = bcos py py = sin b b
1
2
p = Q • r == a
0
(
.
0
)
˙r • p = 0 = r˙x
dr
d
(2.3a)
(2.3b)
Also, the resulting p-vector lies on the ellipse of the inverse quadratic formQ-1F r•Q-1•r=1 whose axes are
the inverses (1/a,1/b) of the original Q-ellipse. This is shown by the equations below and in Fig. 2.2.
( px
a2
py • 0
)
0 px • = 1 = px
b2 py (
a2 p p y • 2 x = a2 px + b2 py
b py )
or:
p • Q1 • p = 1
(2.4)
The original vector r, obtained by inverse operation Q-1•p on p is perpendicular to the Q-1F ellipse tangent.
r=Q
(
p˙ • r = 0 = p˙ x
1
a2 0 • p == 0 b 2
rx 1
p˙ y • = sin
ry a
)
1
cos a
where:
1
py = sin b
1
˙px = sin
rx = a cos a
where:
and:
1
ry = b sin p˙ y = cos b
px a 2 p a cos • = 2 x = py b py bsin 1
a cos cos b sin b
px =
(2.5a)
(2.5b)
Note that vectors p and r maintain a unit mutual projection, that is, their dot-product p•r always is 1.
(
p • r = 1 = px
rx 1
py • = cos ry a
)
1
a cos sin = p • Q 1 • p = 1 = r • Q • r
b sin b
(2.5c)
What gorgeous geometric-algebraic symmetry! It works for tipped ellipses, too. We scale our ellipse plots
by the geometric mean S=(ab) so that minor radius b/S=(b/a) is the inverse of major radius a/S=(a/b),
that is bS=1/aS. So to get the inverse ellipse, just switch axes (or rotate by 90°)! This shows the symmetry.
© 2008 W. G. Harter
Chapter 2. Ellipse orbit relations
Unit. 2
•
Step 1 Construct velocity or tangent vector r(t)
(Recall Fig. 2.2.1)
Step 0 Construct ellipse position vector r(φ)
for a particular angle-time φ (Recall Fig. 1.9.1)
r(φ)
φ=ω t
b
26
φ=ω t
•
r(
φ)
π/2
=r(φ+π/2)
r(φ)
•
v(φ)=r(φ)
r(φ)
a
a
b-circle
b-circle
b
a-circle
a-circle
Step 3 Use (a,b)-circle rectangles to locate
•
φ)
perpendicular vector p(φ) and its tangent p(
Step 2 Flip (a,b)-circle triangles into rectangles.
Note that rectangle corners are on inverse ellipse
Inverse ellipse
p•Q-1•p = 1
(a,b)-circle
rectangle
(a,b)-circle
triangle
•
r(φ) r(φ)
r(φ)
•
=r(φ+π/2)
•
Note vector p(φ) is
p(φ)
•
Note vector p(φ) is
perpendicular to r(φ)
p(φ)
•
r(φ)
•
r(φ)
p(φ)
r(φ)
•
r(φ)
b
=r(φ+π/2)
a
perpendicular to r(φ)
•
φ=ω t
b
p(φ)
r(φ)
b-circle
a
Original ellipse
r•Q•r = 1
p(φ)
•
Original ellipse
r•Q•r = r•p = 1
p(φ)
•
r(φ)
p(φ)
•
•
r(φ)
b-circle
Inverse ellipse
p•Q-1•p =p•r = 1
a-circle
a-circle
r(φ)
φ=ω t
r(φ)
p(φ)
r(φ)
p(φ)
•
•
r(φ)
p(φ)
Fig. 2.2 Quadratic form and orbit geometry. From ellipse orbit r() and velocity tangent ˙r( ) =dr()/d ,
constructs the inverse or “perpendicular” ellipse p() and its velocity tangent p˙ ( ) =dp()/d. (Here =1)
© 2008 Harter
Unit 2. Vibration, Resonance, and Wave Motion
Chapter 2 - 27
Tipped boxes and phasor clockwork
Before studying tipped-ellipse equations for coupled-oscillator analysis, let us review geometry of
tipped-tangent ellipse boxes. A tangent vector ˙r( ) in Step 2 of Fig. 2.2 (upper left) lies on one corner of
an (a,b)-circle rectangle and p˙ ( ) lies on the opposite corner in Step 3. Between ˙r( ) and p˙ ( ) passes the
radial diagonal at angle +90° that connects the other two rectangle corners.
A 90° clockwise rotation of the ˙r( ) - p˙ ( ) rectangle gives the r( ) - p( ) rectangle at phase angle in Step 3 of Fig. 2.2. The vectors ˙r( ) and p( ) define tangent and normal to point r( ) on the Q-ellipse.
They also are parallel to sides of an ellipse box with a tipping angle . We now relate phase angle- to
tipping angle-, slope angle r of position r( ) , and slope angle ˙r of tangent (velocity) ˙r( ) .
Below is a summary of the slope angles of the position r( ) , velocity ˙r( ) , and perpendicular or
normal vector p( ) using (2.3) through (2.5). The geometric relations are constructed in Fig. 2.3.
b
r = ATN tan( ) r
y
b
sin
b
y
a
r
Slope of tangent contact position r( ) : tan = = =
= tan( ) or :
a
x rx a cos a
= ATN tan r b
b
r˙ = ATN cot ( ) r˙y
b cos b
a
˙r
Slope of tangent velocity vector ˙r( ) : tan = =
=
cot( ) or :
a
r˙x a sin
a
r˙ = ACT tan b
a
= ATN tan( ) p
asin
a
y
b
p
Slope of tangent perpendicular p( ): tan = =
=
= tan( ) or :
b
px bcos b
= ATN tan() a
( )
( )
( )
(
( )
)
(2.6a)
(2.6b)
(2.6c)
The object of the construction in Fig. 2.3 is to box an ellipse by a tangent rectangle at a given
tipping angle of the tangent perpendicular vector p( ) . Fig. 2.3 is a more precise tipped box than Fig.
1.10-11. The key relation is (2.6c) between phase angle = t and the polar angle of the tangent
perpendicular vector p()=Q•r(). If position r-vector is on an ellipse axis (=0 or =) then so is the pvector (p(0)=r(0)). Otherwise, they are separated by one of those (a,b)-circle rectangles developed in Step
2 of Fig. 2.2 and Step 1 of Fig. 2.3 on the next page.
Each rectangle has a phase angle as its radial diagonal, and is found using a tan = b tan (2.6c) in Step 0 of Fig. 2.3. The phase angles c = c+/2 and = +/2 in Fig. 2.3-Step 0 are for two
velocity vectors ˙r( c ) and ˙r( ) that will lie along the -tipped ellipse box. In Step 1 their companion
position vectors for the tangent contact points r( c ) and r( ) are easily found, too.
© 2008 W. G. Harter
Chapter 2. Ellipse orbit relations
Unit. 2
28
Step 0 Given tipping ϕ-angles (or tangent r(φ) directions) locate velocity phase φ′−angles
ϕ
• φ ) direction
tangent r(
c
Transfer a-elevation at ϕ
to b-elevation to find angle φ′
φc′=φc+90
ϕc
a tan ϕc
ϕc
a
b
ϕ
ϕ
ϕc
a tan ϕc
ϕ
=a tan φ′c
ϕc
ϕ
a
b
ϕ
ϕ
ϕc
b-circle
a tan ϕ
ϕ
ϕc
a tan ϕ
φ′=φ+90
= b tan φ
a-circle
• φ) direction
tangent r(
•
•
•
•
Step 1 Use phase φ′−angles to construct vectors r(
φ), p(
φ), r(φ), p(φ), r(
φc), p(
φc), r(φc), and p(φc)
and the box tangents
ϕ
•
ϕc
φc′
p(φc)
p(φc)
φc
•
p(
• φc)
p(φc)
ϕc
•
r(
φc)
a
r(φc)
b
p(φ)
•
r(
φ)
r(φ)
φ
-r(φ)
a
-r(φc)
φ′
r(φc)
•
r(
φc)
ϕ
b
r(φ)
p(φ)
•
r(
φ)
p(φ)
•
•
•p(φ)
Fig. 2.3 Precise construction of ellipse tangent box with a given vertical inclination =90°c.
The next Fig. 2.4(a-b) uses the box tangents to compare -tipped (u1,u2)-axis phasor components
with those of ellipse’s own (x1,x2)-axes. Phasors rotate clockwise by a -angle all with the same rate .
© 2008 Harter
Chapter 2 - 29
Unit 2. Vibration, Resonance, and Wave Motion
Vx1
All phasors
go clockwise
x2
Relative
tipping ϕ
angle
ϕ
u2
ϕ
x1
x1
u1
x2
This particular orbit
is going clockwise
All phasors x
go clockwise 2
•••
r(
φ)
φ
r(φ)
Vx2
All phasors
go clockwise
x1
u2
r(φ)
•
••
r(
φ)
φ
Vu2
All phasors
go clockwise
Vu1
u1
Relative
phase lead
angle
ρ
2 leads 1
by
ρ
u2
u1
Fig. 2.4 (a) Two pairs of phasor components that make a clockwise orbit.
In Fig. 2.4(a) above the ellipse orbit also goes clockwise, but in the next Fig. 2.4(b) the ellipse
orbit is counter clockwise even though the phasor rotation is clockwise there as it is above. Setting the x2
phasor ahead of x1 by /2 (or u2 ahead of u1 by ) causes the orbit to be clockwise ((+) work W and (-) L)
© 2008 W. G. Harter
Chapter 2. Ellipse orbit relations
Unit. 2
30
Vx1
x2
Relative
tipping
angle ϕ
ϕ
u2
x1
ϕ
All phasors
go clockwise
x1
φ
u1
x2
x2
All phasors
go clockwise
φ
•
r(
0)
r(0)
••
Vx2
This particular orbit
is going counter-clockwise
r(φ)
•
r(
φ)
r(0)
••
r(
φ)
x1
All phasors
go clockwise
u2
r(φ)
•••
r(0)
•••
Vu2
φ
Vu1
All phasors
go clockwise
φ
u1
Relative
phase lag
angle
ρ
2 lags 1
by
ρ
u2
u1
Fig. 2.4 (b) Two pairs of phasor components that make a counter-clockwise orbit.
It is important to see how the phasors are adjusted to make the orbits come out as they do. In the
case above the x2 phasor is behind x1 by /2 giving a counter clockwise orbital rotation. These are
elementary examples of a wave going around a circular or elliptical enclosure. Very useful!
© 2008 Harter
Unit 2. Vibration, Resonance, and Wave Motion
Chapter 2 - 31
Special Tippings: a/b and 1/1
The perpendicular p( ) starts out at zero-phase =0 in the same direction as the position radius
vector r( ) , and would remain so if a=b. But, if a>b, as in Fig. 2.5, the slope of perpendicular
p( ) = Q • r ( )
is greater by a/b than phase slope tan while slope of position radius r( ) is less than by
the same factor.
Fig. 2.5 Step 0-1shows a special case. The phase angle slope tan is precisely that of the ellipse
diagonal, that is, tan =b/a. By (2.6a) position radius has slope b2/a2. The tangent slope, given cot =a/b,
equals –1 by (2.6b) so its perpendicular has slope is +1, that is, angle 45° in agreement with (2.6c). For
Step 2-3 the phase slope is instead tan c =a/b so the position radius has slope +1 , its tangent has a slope b2/a2, and its tangent perpendicular has a slope a2/b2. This is shown by geometric construction in Fig. 2.5
Step 4 of a box tipped to -b2/a2 that touches the ellipse at the 45° line.
In Fig. 2.6, the phase angle slope is set to one (tan =1) so that the position radius rises up to the
ellipse diagonal, that is, tan r = b / a . A box tangent to the ellipse diagonal is tipped to -b/a (2.6b) while the
perpendicular slope (2.6c) is a/b. Applying matrix Q moves slope closer to the minor (b=1/a)-axis of the
Q-ellipse by a factor a2/b2, while inverse matrix Q-1 moves slope closer to the major (a=1/b)-axis of the
Q-ellipse by a factor b2/a2, that is, by the inverse factor.
This Q-matrix geometry is important. One of the matrices seems to want to sweep all vectors onto
the y-axis while the other is just as determined to send every one to the x-axis. The x-axis and y-axis are
the ellipse’s own axes or, in partial-German, eigen-axes. The concept of eigenvectors which lie along
eigen-axes is the single most important mathematical concept in quantum theory.
© 2008 W. G. Harter
Chapter 2. Ellipse orbit relations
Step 0
Set phase-angle φ to diagonal of (a,b) of ellipse box.
Using circle-(a,b) box, find ellipse position vector r(φ)
and its tangent-perpendicular p(φ) (Recall Fig. 2.2.2)
Unit. 2
Step 2
Do Steps 0-2 for complimentary angle φc=π/2−φ
•
X-reflection of r(φ) and p(φ) gives r(φc) and p(φc)
•
Atan a/b=φc
p(φ) p(φc)
•
p(φ)
•
Atan b/a=φ
φ=ω t
p(φ)
•
r(φ)
b-circle
•
•
X-reflection of r(φ) and p(φ) gives r(φc) and p(φc)
r(φ)
p(φc)
•
r(φc)
r(φ)
r(φc)=p(φ)
r(φc)
•
•
φc)=r(
φ)
p(
•
a
φc=π/2−φ
b
ϕ =45
ϕ45 =Atan b2/a2
a-circle
Step 1
Rotate Step-0 construction 90 to get velocity vector
•
•
r(φ) and its tangent-perpendicular p(φ). 45 contact
•
φ) and p(φ)
box of both Q and Q-1 ellipses has axes r(
Step 3
•
φc) and its perpendicular p(φc)
Use velocity r(
to construct box tangent at 45 position r(φc).
45 contact slope angle is ϕ45 =Atan b2/a2.
Fig. 2.3 Square with =45° tipping contacts both Q and Q-1 ellipses. Slope b2/a2 box contacts at 45°.
Step 1 Set phase-angle φ to 45 .Using circle-(a,b) box,
•
Step 2 Use velocity r(
φ) and its perpendicular p(φ)
find ellipse position vector r(φ) and p(φ)
to construct box tangent at position r(φ).
and ellipsevelocity vector r(φ) and p(φ)
Contact slope angle is ϕ=Atan b/a.
•
•
p(φ)
p(φ)
•
•
p(φ)
φ=45
r(φ)
r(φ)
•
b-circle
a-circle
p(φ)
r(φ)
r(φ)
•
a
a
b
φ=45
b
ϕ =Atan b/a
Fig. 2.4 Box with =45° phase angle contacts Q ellipse at its (a,b)-diagonal with slope-b/a.
32
© 2008 HarterSoft –LearnIt
Unit 2. Vibration, Resonance, and Wave Motion
2-
33
Chapter 3 Strongly Coupled Oscillators and Beats
So far, two identical swinging pendulums were coupled by a k12-spring so weak that it would take
a long time to noticeably affect the motion of either one. We derived the sine-lag work-energy per-cycle
transfer formula (1.8) and the power transfer formula (1.10) using geometry of phasors and ellipses.
Work()on 1 by 2=k12A1 A2 sin (1.8)repeat
Power()on 1 by 2=22 k12A1 A2 sin (1.10)repeat
Now finally, we will work out how a k12-spring of arbitrary strength will affect each of the
pendulum’s motion. As before, we use geometry and algebra to help understand the resonance behavior.
(a) Elliptical potential energy bowls
The energy expression given back in (1.14) does not include the potential energy 1/2 k12(x1-x2)2 in
the coupling spring. (We assumed k12 was tiny.) Now we add it in to get a total PE function V(x1,x2).
V ( x 1, x 2 ) = PE osc 1 + PEosc 2 + PEcoupling 12
=
1 2
1
kx1 + kx22
2
2
+
(
)
(3.1)
2
1
k12 x1 x 2
2
Here individual spring constants k1=k=m 2 =k2 are assumed the same for each pendulum-oscillator. A
plot of this function in Fig. 3.1 reveals a beautiful bowl or elliptical paraboloid.
Imagine this is a bare-essentials ski bowl. (Ski Bare Valley!) A nearly overhead view such as Fig.
3.1(a) and a topographical map in Fig. 3.2 reveals elliptical topography contours, lines of equal altitude or
potential energy. If potential V(x1,x2) equals a constant E, that is a tipped ellipse equation.
V ( x 1, x 2 ) = E =
1 2
1
kx + kx 22
2 1
2
+
(
)
(
2
1
k
k12 x1 x 2 or:
x 2 + x 22
2
2E 1
)
(
)
2
k
+ 12 x1 x 2 = 1
2E
This is a quadratic form or ellipse matrix equation with a spring-force matrix
k + k12 2
k + k12 2
x1 +
x 2 k12 x1 x 2
2
2
k + k12
k12 x 1 1
= (x 1 x 2 ) • • k12
2
k + k12 x 2 k + k12
K =
k12
(3.2)
k12 k + k12 .
V ( x 1, x 2 ) =
or:
V=
The NW and SE corners have more crowded topo-lines and steeper slope with greater force
V ( x1 , x 2 )
= (k + k12 ) x1 k12 x 2
F1 1V k + k12
x1
or : = =
V (x 1, x 2 )
F2 2V k12
F2 =
= k12 x 1 + (k + k 12) x 2
x 2
F1 =
(3.3)
1
x • K •x
2
Fk = k12 x1 V
= K •x
• or :- F =
k + k12 x 2 x
V
x k
.
(3.4)
Force F=-K•x is along the fall-line and perpendicular to topo-lines. Vector-F is perpendicular to a Kellipse at point x as in Fig. 3.2 just like vector p=Q•r is perpendicular to r˙ in Step 2 of Fig. 2.2.
© 2008 W. G. Harter
Chapter 3. Strongly coupled oscillators and beats
Unit. 2
(a)
X2
(North)
-X1
(West)
X1
(East)
-X2
(South)
(b)
X1 axis
X2 axis
Fig. 3.1 Bare-Valley parabolic topography (k=0.3, k12=0.2). (a) Elliptic contours (b) Coordinates
34
© 2008 Harter
Chapter 2 - 35
Unit 2. Vibration, Resonance, and Wave Motion
(b) Gradients and modes
An important idea of calculus is being shown by the elliptical topo-lines. The idea involves a
vector whose x and y components ( and z component if we’re in 3D…and t component if we’re in 4D
spacetime…) are made of the x and y (and z, etc…) slopes or derivatives. There are several ways to write
this operation that is called the gradient
V =
V
x
of a function V=V(x).
V V x x V x V V =
= =
=
x V y V y V y (3.5)
I personally like the second way because it makes a nice shorthand for the old calculus formula
( x • K • x)
= 2K • x ,
x
( x • x)
= 2x
x
( ) = 2kx .
kx 2
x
(3.6)
(But be careful!) The main geometrical thing to remember about a gradient V is that it gives a vector
perpendicular to the V function it’s acting upon as we showed in (3.4). The Bare Valley topography map
in Fig. 3.2 shows this. And, minus the V vector is the applied force that the mass or “skier” at position x
will experience due to springs, gravity, or whatever is making things swing back and forth.
Bare Valley has its gentlest “beginner slopes” on the NE (45°) and SW (225°) sides and its steeper
“advanced slopes” on the NW (135°) and SE (315°) corners while “intermediate slopes” are found at
N(90°), W(180°), S(270°), and E(0°), unlike the usual compass headings with 0° for North at the top of a
map. The 45° beginner coordinate u+ and the 135° advanced coordinate u are related to the x and y or
East and North coordinates x1 and x2 by projecting on 45° or 135°. (cos 45°=1/2=sin 45°=sin135°= -cos135°)
u+ =
u =
x1 + x 2
x1 =
2
x1 + x 2
(3.7a)
x2 =
2
The energy equation (3.2) is one of an ellipse on (u+, u-) axes since
V=
(
k 2
x + x22
2 1
) + k2 ( x
12
1
x2
)
2
=
(
k 2
u + u2
2 +
) + k2 (
12
)
2
2u
=
u+ u
2
u+ + u
(3.7b)
2
u+2 + u2 = x12 + x22
and
k 2 k + 2k12 2
u +
u
2
2 +
x1 x 2 = 2u .
(3.2)redone
This means the beginner-slope force constant k is lower than the advanced slope constant k+2k12. So is
beginner oscillation frequency + less than that the frequency something sliding on the advanced slope.
+ =
k
m
(3.8a)
=
k + 2k12
m
(3.8b)
Fig. 3.2(a) shows the SLOW u+ mode that goes NE (45°) to SW (225°) so x1 and x2 are in phase. Fig.
3.2(b) shows the FAST u- mode that goes NW (135°) to SE (315°) so x1 and x2 are out of phase.
© 2008 W. G. Harter
Chapter 3. Strongly coupled oscillators and beats
Unit. 2
(a) PE Contours and gradients
36
SLOW axis
FAST axis
(c) Anti-symmetric
FAST Mode
Coordinate
U
+
Coordinate
(b) Symmetric
SLOW Mode
0
In-phase
mode
180
Out-of-phase
mode
is SLOW
phasor
is FAST
phasor
(90 turned)
(90 turned)
Fig. 3.2 (a) Force-potential of Bare Valley gives symmetric (+) or antisymmetric (-) vibration modes.
(b) SLOW (+) mode goes on minimum shallow slope. (c) FAST (-) mode goes on maximum steep slope.
Now let’s see what happens if we superimpose both these modes at once. We can do this and still
have a valid motion because the force equations are linear ones. The result is shown in Fig. 3.3.
© 2008 Harter
Chapter 2 - 37
Unit 2. Vibration, Resonance, and Wave Motion
(c) Superposition principle and beats
An important idea of modern physics of linear resonance dynamics is the superposition principle,
that you can simply add two valid motions together and get another one. Modes mix like paints that
together can paint anything that happens in the world! It’s an idea that is generally credited to Jean
Baptiste Fourier’s analysis, but it really goes back to Huygens and Leibenitz. It works well if mode
frequencies are the same for all amplitudes as they are for harmonic oscillations under linear force laws.
By adding each of the U+-SLOW-mode phasors in Fig. 3.2(a) to each of the U-FAST-mode
phasors in Fig. 3.2(b) we get the situation depicted in Fig. 3.3. The result, if the two modes were the same
frequency, is no different from the mode pictures of previous figures like Fig. 2.4. The elliptic orbit would
be drawn over and over because the phase lag = + of the SLOW U+-mode phasor behind the FAST
U-mode phasor must then stay fixed since the modes are turning at the same rate.
However, that is not the case here. The FAST U-mode phase soon runs ahead of the SLOW
U+-mode phase + . As their phase lag = + changes so does the (x1,x2)-ellipse drawn by the two
phasors. As the (x1 , x2)-ellipse changes so do the amplitude and phases of the x1 and x2 phasors.
This complicated motion is called beating after the effect by the same name that we hear all the
time in music and acoustics. The rate
Beat
of beating is simply the relative phase velocity
difference between the FAST and SLOW angular rates =
Beat = + =
d dt
d d d +
=
dt
dt
dt
and + =
d +
dt
d
dt
, that is, the
.
(3.9)
From the results of (3.8) we obtain the angular beat frequency.
Beat = + =
Its binomial approximation
k + 2k12
k
m
m
1
1
1
(a +b) 2 = a 2 + 1 a 2 b + ...
2
Beat 1 2k12 k12
=
2 km k
(3.10a)
, at first, grows linearly with coupling constant k12.
k
m
for: k12 << k
(3.10b)
At the moment shown in Fig. 3.3(a) the SLOW U+-mode phase is about =60° behind the FAST
U-mode phase. At a slightly later time in Fig. 3.3(b) the phase lag increases to =90° and continues
increasing at beat rate Beat (3.10). As rolls through 2 the (x1,x2)-phasors and their (x1,x2)-ellipse beat
or pulse in and out. The (x1,x2)-beat goes on while the U+-U-phasors each turn constantly.
© 2008 W. G. Harter
(a)
Chapter 3. Strongly coupled oscillators and beats
Unit. 2
U+
is SLOW
phasor
is FAST
phasor
-phasors have constant
amplitudes and frequencies
(b)
X1 and X2
phasors change
amplitudes
and
phases
U+
SLOW
falls 90
behind
FAST
Fig. 3.3 Superposition of the U+ and U- modes from Fig. 3.2 (a-b). (a)One time (b)Later time
38
© 2008 HarterSoft –LearnIt
Unit 2. Vibration, Resonance, and Wave Motion
2-
39
Chapter 4 Complex phasor analysis of oscillation, beats, and modes
The simplest oscillation is harmonic oscillation, for which anguolar frequency (wiggles per 2
second) is constant regardless of amplitude A as in an ideal pendulum or mass on a spring. To track
oscillation we use time plots of x versus t, that is, x(t), as well as phase plots of oscillating mass position
x=A cost versus its velocity v= -A sint. On a graph of x versus v/ in Fig. 4.1, the point moves on a
circle like a clock hand as in previous examples in Fig. 3.3 and first introduced in Fig. 10.5b of Unit 1.
Here we treat the phase plot in Fig. 4.1 as an actual 12-hour clock starting at 3 AM.
A day in the life of this oscillator begins at 3AM with it standing still at x=1. By 4AM it is at
x=0.866 (x is cos(-1·2/12)) going with a negative velocity toward the origin. By 5AM it is half way to
origin at x=0.5 (x is cos(-2·2/12)) going with even faster negative velocity given by sin(-2·2/12) in
units of frequency . Here the clock angular frequency is -2 radians per 12-hour day or a Hertz
frequency of 1-per-day, and a period =1/ of =1 day=12 hours =12(60) minutes =12(3600) seconds.
At 6AM it passes origin at x=0 (x is cos(-3·2/12)) going with maximum negative velocity (v is
sin(-3·2/12)=-1 in units of frequency . 6AM is halfway through the “compression” phase.
At 9AM it arrives on the other side of origin at x=-1 (x is cos(-6·2/12)) and comes to a stop at
zero velocity (v is sin(-6·2/12) in units of . This ends the “compression” phase in the Fig. 4.1(b).
The rest of the day from 9AM to 3PM (as the oscillator returns home) is its “expansion” phase in
which its velocity is positive. Then its phasor clock is in the upper half of the clock dial at the top of Fig.
4.1(b) that represents positive velocity. The lower half indicates negative velocity.
(a) Introducing complex phasors and wavefunctions: Real and imaginary axes
The phase-plot velocity-axis is called the imaginary axis by engineers and physicists who prefer to
use complex phasor clocks like the one shown in Fig. 4.1(c). It’s the same as the clock drawn in the upper
part of the figure, except it is based on what may be one of the most important mathematical results of all
time, the so called Euler identity. (The French call it the Theorem of DeMoivre.)
e-i = cos – i sin (i=-1)
(4.1a)
A wave function or oscillation function of a given angular frequency is written using Euler’s form.
= A e-it = A cost – i A sint (i=-1)
(4.1b)
The real part Re = A cost and the imaginary part Im = -Asint are the Cartesian coordinates of the
phasor clock in Fig. 4.1.(c). The absolute value ||= A and argument ARG=- t or phase angle are the
phasor polar coordinates. This is the main reason your calculator has Polar-Rect and Rect-Polar buttons!
Chapter 10 of Unit 1 introduced the complex phasor, Euler’s number e=2.7182818…, and
exponential functions in general. Here we give a quick description of applications of complex e-it
numbers to the beats described in the preceding Chapter 3 and relate them to ellipse geometry.
© 2008 W. G. Harter
Chapter 4. Phasor analysis of modes and beats
(a)
Velocity v
(in ω units)
Amplitude
A
A
12
11
10
Position x
−ωt
Expansion
1
2
3 x Position x
4
Compression
5
9
8
7
6
3AM
Phase angle
−ωt
k12
40
Velocity v
(in ω units)
(b)
F
v /ω
Unit. 2
Compression
period
5AM (Negative velocity
v<0)
4AM
6AM
7AM
(c)
Mathematical description using
Complex and Imaginary Numbers
i = -1
8AM Time
t
9AM
10AM
Complex
Phasor
Amplitude
A
ω
t
-i
Ψ=Ae
=Acosωt-iAsinωt
Imaginary part
Ψ=−Asinωt
F Im
−ωt
A
11AM
Expansion
period
(Positive velocity
v>0)
12 PM
1 PM
Real part
Re Ψ=Acosωt
2 PM
3 PM
Fig. 4.1 Single pendulum oscillation. (a) phasor plots. (b) Spacetime plot. (c) Complex phasor “clock.”
(b) Coupled oscillation between identical pendulums
The resonance process, by which two identical oscillators trade energy and momentum, may be
analyzed by first looking at the very simplest motions that the two can undergo. The two modes of the
oscillators are shown in Fig. 4.2. This is another and simpler way to look at the ones in Fig. 3.2
The slow mode in Fig. 4.2(a) is simply the two pendulums swinging together as though they were
one. The spring connecting them might as well not even exist. It does nothing since it is neither stretched
nor compressed by this motion. This is the same as Fig. 3.2(b).
The opposite situation holds in Fig. 4.2(b) or Fig. 3.2(c) where pendulums swing oppositely as
though each was connected to an immovable wall between them. Because both stretch and compress their
connecting spring at once, this fast mode motion will have a higher frequency than the slow mode in Fig.
4.2(a) effectively doubling the force of spring k12. This explains the 2k12 in the - frequency formula
(3.8b) for the (-)-mode and no k12 in the + frequency formula (3.8a) for the (+)-mode
© 2008 Harter
Chapter 4 - 41
Unit 2. Vibration, Resonance, and Wave Motion
For the sake of argument we will take the frequency of the fast mode to be twice that of the slow
mode. (That is k12=3k in (3.8).) This means that while the slow mode swings both masses on a one-way
trip to the other side between 3AM and 9AM, the fast mode brings them together and sends them back
home in that time and then repeats that whole two-way trip by 3PM in the afternoon.
(c) Summing phasors using Eulers identity: Beats
The frequency of a single harmonic oscillator is unaffected by changing amplitude A of
oscillation. This is because a harmonic spring is linear, that is, the force it exerts is proportional to the
distance x from origin or equilibrium. So acceleration is a constant 2 times the amplitude A , and this is
true no matter how big A is. (Recall the a(t) equation in the middle of (2.1).)
Similarly, the mode frequencies slow and fast of a double harmonic oscillator are not affected by
their amplitudes, either. That means neither Afast nor Aslow affect either slow or fast even if they are both
turned on at once! The system can swing at slow while vibrating at fast and neither frequency changes.
It’s a little like walking back and forth while chewing gum at the same time and not missing a beat!
This is no surprise if you followed the superposition (10.51) in Unit 1 or the analogy between two
1-dimensional pendulums and one 2-dimensional “starlet” oscillator-orbiter first introduced in Unit 1
Chapter 9. The two-1D to one-2D connection is a deep one that we use often. It’s appearance in quantum
theory is particularly interesting where one relates an m-particle n-dimensional system to an n-particle mdimensional system or to an 1-particle mn-dimensional one!
Suppose we turn on both modes with equal amplitude, that is let Afast =1= Aslow. Then we get a
sum of two mode phasors describing mass-1 and difference of two mode phasors describing mass-2.
mass-1 = Afast exp(-ifastt) + A slow exp(-i slowt) = exp(-ifastt) + exp(-i slowt)
(4.2a)
mass-2 = Afast exp(-ifastt) - A slow exp(-i slowt) = exp(-ifastt) - exp(-i slowt)
(4.2b)
These sum and differences are done by geometry in the next several figures. However, for this case it is
quite easy to analyze them using Euler’s forms with a=fastt =2.0t and b= slowt=1.0t.
eia+ eib = ei(a+b)/2 (ei(a-b)/2+ e-i(a-b)/2)
eia- eib = ei(a+b)/2 (ei(a-b)/2- e-i(a-b)/2)
ix
-ix
ix
(4.3a)
(4.3b)
-ix
Now we use the cosine cos x = (e + e )/2 and sine sin x = (e - e )/2i in Euler form.
mass-1 = eia+ eib = 2 ei(a+b)/2 cos (a-b)/2= 2 ei3t/2 cos t/2
mass-2 = eia- eib = 2i ei(a+b)/2 sin (a-b)/2 = 2i ei3t/2 sin t/2
(4.4a)
(4.4b)
In Fig. 4.3 the mass-1 and mass-2 phasors are summed starting at 12PM (t=0) and then at 1 PM (t=1).
Transverse or Longitudinal?
This analysis applies equally to two pendulums that swing side-by-side like the ones in Fig. 4.3
(and demonstrated in class). These are called transverse oscillators as opposed to longitudinal oscillators
pictured in Fig. 4.2. The motions in Fig. 4.3 are transverse to their connection or coupling while the
masses in Fig. 4.2 move along a line connecting them.
© 2008 W. G. Harter
Chapter 4. Phasor analysis of modes and beats
(a) Slow mode ω=1(2
11 12 1
9
9
7
6
5
7
5
6
9
4
9
3
8
7
6
5
7
6
5
4
3AM
4AM
5AM
5AM
3AM Compression
period
4AM (Negative velocity
v<0)
6AM
6AM
7AM
Time
t
7AM
8AM
/12)
3
8
4
42
2
10
2
10
3
8
4
=2(2
11 12 1
11 12 1
2
10
3
8
(b) Fast mode
/12)
11 12 1
2
10
Unit. 2
9AM
Expansion
period
(Positive velocity
v>0)
8AM
9AM
10AM
10AM
11AM
11AM
12 PM
Longitudinal
Motion
12 PM
1 PM
1 PM
2 PM
2 PM
3 PM
3 PM
Fig. 4.2 Coupled oscillators. (a)Slow-=1 mode swings in phase and does not stretch connector.
(b) Fast-=2 mode swings out of phase and stretches connector by twice.
Time
12PM
11 12
2
9
3
11
9
6
12
9
7
mass-1
9
9
Mode
Mixed
Mode
mass-2 4
6
mass-2
Slow
Mode
6
12
Transverse
Motion
3
7
3
mass-1 4
4
2
mass-1 4 Fast
6
3
6
12
11
10
3
2
7
Mode
2
mass-2
Time
1 PM
12
2
mass-1 4
Slow
7
12
12
2
9
3
Fast
Mode
2
9
mass-14 (jumps
twice as fast)
Mixed
Mode
3
4
6
mass-2
Fig. 4.3 Coupled pendulum with mixed mode (sum of fast and slow modes) (a) 12 PM. (b) 1 PM
© 2008 Harter
Unit 2. Vibration, Resonance, and Wave Motion
Chapter 4 - 43
Here in Fig. 4.3 the pendulums swing transverse to their connectors so the real position x-axis is
vertical and the (imaginary) velocity v-axis points horizontally off to the left.
In Fig. 4.3 at 12PM mass-1 is standing still with double amplitude and mass-2 is stuck at origin
with no energy at all. But, by 1PM mass-1 is moving in a negative direction and about three quarters of
the way to origin. Its amplitude is slightly reduced to 2 cos(0.5·2/12) by eq. (3.4a) (See big lavender
circle in Fig. 4.3(right)) while mass-2 has acquired an amplitude of (2 sin(0.5·2/12) (See small lavender
circle in Fig. 4.3(extreme right)). Throughout this process the sum of the areas of the lavender circles will
remain constant. Each lavender area is proportional to the total energy of each oscillator. Recall why total
phasor area is constant in Fig. 1.11.
Note that the phasor arrow for mass-2 lags 90° or /2 radians behind that of mass-1. This checks
with the (i)-factor in equation (4.4b). An (i=ei/2)-factor represents a 90° rotation the is counter-clockwise.
So. it sets mass-2’s clock back a quarter turn or 3 hours.
A 90° phase lag is a very important thing to physics and engineering. It is the phase lag that is
most efficient at transferring energy from one oscillator to another. Then force and velocity are in phase so
that power, which is force times velocity, is as big as it can be.
In Fig. 4.4 the time sequence continues as the phasors rotate according to the phase factor ei3t/2 in
equations (3.4) but the (i=ei/2)-factor on mass-2 keeps it 90° or /2 radians behind that of mass-1 until
mass-1 runs out of energy and at 6PM mass-2 has it all. That marks the halfway point of the first beat.
At 7 PM the phasor for mass-1 comes out 90° or /2 radians behind that of mass-2, and the tables
are turned with mass-2 the energy donor and mass-1 soaking it up as fast as it can from 7PM to 12PM.
Finally, at 12 PM (not shown) the first beat completes itself and the whole sequence occurs again.
(c) Slow(er) Beats
The beat frequency in (4.4) is half the difference between the two mode frequencies in (4.1), in
this case that is (fast slow)/2=0.5, which means it will take 24 hours or two full slow-clock rotations. In
that time the fast clock laps the slow clock twice.
But, if you could only see the lavender circles and not the phase vectors inside, one beat appears to
go from mass-1 to mass-2 by 6 PM in Fig. 4.4 and will be back by 12 PM. All that is needed is for the fast
clock to lap the slow clock just once. The lavender-circle beat frequency is the relative angular velocity of
the fast versus slow clocks around the phasor track. Here (fast slow)=1.0 per 12-hr day.
beat =fast slow
(4.5)
If two modes have very nearly the same frequency, then a beat takes a longer time because the
faster phasor takes longer to lap the slower one. In Fig. 4.5 it takes a little over 4 phase oscillations to
complete one beat “lump” or group wave. In atomic laser optics this number is often more like 40 or 50
million. In nuclear physics it can be way over a million-million. That’s a lot of very patient persuasion! To
understand some of the subtler aspects of modern physics we must find ways to begin comprehending the
consequences of such an enormous amount of persuasion.
© 2008 W. G. Harter
t=0
12 PM
Chapter 4. Phasor analysis of modes and beats
t=1
t=2
2 PM
1 PM
t=3
3 PM
Unit. 2
t=4
t=5
4 PM
5 PM
t=6
6 PM
t=7
7 PM
Fig. 4.4 Coupled pendulum with mixed mode (sum of fast and slow modes) from 12 PM to 8 PM
(a) Standing Wave
Ratio
SWR = 0
(b) Standing Wave
Ratio
SWR =1/3
Time t=0
Time t=0
1-cos(ω+−ω−)t
ω+=11
ω−=14
2+cos(ω+−ω−)t
ω
sin(ω+−ω−)t
2
τBeat
2-cos(ω+−ω−)t
cos(ω+−ω−)t
2
1+cos(ω+−ω−)t
Beat
period
SWR
= 1
3
Fig. 4.5 Coupled pendulum with mixed mode and more nearly equal mode frequencies.
t=8
8 PM
44
© 2008 Harter
Unit 2. Vibration, Resonance, and Wave Motion
Chapter 4 - 45
(d) Geometry of resonance and phasor beats
Fig. 4.4 and Fig. 4.5 show some ways to plot the complex wave sums =A e i A t + B e i B t of
mode phasors in (4.2) over time. Now we consider the geometry of phasor vector sums and differences.
Head-to-tail addition of phasors in Fig. 4.5 gives nodes in Fig. 4.5(a) for equal amplitudes (A1=A2)
but not for unequal amplitudes in Fig. 4.5(b). The ratio of the nodal minimum amplitude (|A|-|B| at points
of destructive interference) to the maximum possible amplitude (|A+|B| at points of constructive
interference) is a standing wave ratio (SWR) . Its inverse is the standing wave quotient (SWQ).
SWR=(|A|-|B|)/(|A|+|B|) (4.6a)
SWQ=(|A|+|B|)/(|A|-|B|) (4.6b)
Zero SWR or infinite SWQ means equal “50-50” amplitudes as sketched in Fig. 4.5(a).
How a phasor is plotted depends upon what one chooses for the overall phase that is factored out.
This corresponds to choosing a reference plotting frame in phasor space. Fig. 4.6 displays three main
choices corresponding to the three factorizations below.
= Ae i A t + Be i B t
(4.7)
i t
+i t
i + A ) t / 2 Ae + i( B A ) t / 2 + Be i( B A ) t / 2 = e i A t A + Be ( B A ) = e i B t Ae ( B A ) + B = e ( B
A - phasor view
B - phasor view
A + B average - phasor view
The A-frame view rides with the A-phasor by ignoring its e i A t rotation as shown in Fig. 4.5(a)
for three different SWR values. In each case, a circle of radius B is traced around a fixed A-vector.
The B-frame view rides with the B-phasor by ignoring its e i B t rotation as shown in Fig. 4.5(b)
for three different SWR values. In each case, a circle of radius A is traced around the B-vector.
In each of these the angle between the sum (A+B) and difference (A-B) phasors should be noted
particularly for the (SWR=0)-case where that angle is either +90° or else –90°. We have noted the ±90°
phase between mass-1 and mass-2 in Fig. 4.4 and how it switches each time one of the phases loses all its
energy. Now we see that behavior has a simple geometric explanation based upon right triangles inscribed
in a circle with diameter on the hypotenuse.
The figures (a2) and particularly (b2) show a quite violent motion that the sum or difference
phasor must undergo each time it passes close to the origin and switches from nearly +90° to nearly –90°
or vice-versa. This galloping behavior is responsible for faster-than-light laser waves later.
i + t
The AB average-frame view factors out the average phase e ( B A )
/2
so that A and B rotate
with opposite but opposite angles, one centered on the other as shown in Fig. 4.5(c). The resultants
generate elliptical orbits of major radius a=A+B and minor radius b=A-B. The construction is quite the
same as the one in Fig. 2.2 and gives both the ellipse, by phasor sum A+B, and the inverse ellipse, by the
difference phasor A-B, whose vectors are the diagonal of the rectangle connecting the two ellipses.
© 2008 W. G. Harter
Chapter 4. Phasor analysis of modes and beats
(a1) A stationary as B rotates
(a2) Small SWR
A>B
A + B e-iωt with A>>B
A
ω
A - B e-i t
A
B
A
B
A
ω
A + B e-i t
B
A
ω
A + B e-i t
(b1) A rotates as B is stationary
ω
A + B e-i t
(b2) Small SWR
A>B
Ae+iωt + B with A>>B
Ae+i t + B
(b3) Zero SWR
A=B
ω
Ae+i t + B
A
A
A
B
A
A
B
B
ω
-Ae+i t + B
ω
Ae+i t + B
A
B
B
ω
-Ae+i t + B
ω
-Ae+i t + B
(c1) A and B counter-rotate
Ae+iωt/2+Be-iωt/2 with A>>B
(c2) Small SWR b/a= A - B
A+B
A>B
(c3) Zero SWR
A=B
ω
ω
Ae+i t/2-Be-i t/2
B
A
B
46
(a3) Zero SWR
A=B
ω
A - B e-i t
ω
A - B e-i t
A
Unit. 2
Ae+i t/2+Be-i t/2
A
A
A
b=A-B
a=A+B
Fig. 4.6 Three views of phasor sums A±B having three different SWR values.
Views (a1-a3) Phasor A fixed. Views (b1-b3) Phasor B fixed. Views (c1-c3) A and B counter rotate.
© 2008 Harter
Chapter 4 - 47
Unit 2. Vibration, Resonance, and Wave Motion
(e) Adding complex phasors
The best way known for adding up our phasors uses complex numbers, that is, numbers like 2+i
containing the imaginary square root i=-1 of minus one. Every physicist should be able to add (and
multiply) 2+i and 2+i quickly as well as find sin(2+i) and (2+i)2+i. Do complex numbers make you
queasy? Then you’ll be like a doctor that faints at the sight of blood. Let’s see how they work here.
The basic description of the 2-identical-oscillator is a complex 2D vector |X having complex
components z1=x1+iv1 and z2=x2+iv2 represented by X1 and X2 phasors in Fig. 4.7. Here we use (3.7b).
u+ (t) u (t) z1 (t) 2
= u (t) X =
= +
u (t) + u (t) z2 (t)
+
2
1 2
+ u (t) 1 2
1 2
= u+ (t) U + + u (t) U 1 2
(4.8)
The driving phasors are the SLOW U+-mode and FAST U--mode phasors whose complex components u+
and u- do nothing more than oscillate at their initially assigned amplitudes u±(0) and frequencies ±.
u+ ( t) = u+ ( 0)e i +t
u ( t) = u ( 0)e i t
(4.9a)
(4.9b)
The object of Fig. 4.7 is to construct the pulsing X1 and X2 -phasors for a single beat.
z (t) u (t) 1 u (t) 1 u (0) ei + t u (0) ei t X = 1 = + + = +
+
z2 (t)
2 1
2 1
2 ei + t 2 ei t (4.10a)
Fig. 4.7 has equal-but-opposite mode components u+(0)=1/2=-u-(0). To simplify the construction let
SLOW U+ be really slow (+=0) and only vary FAST U--phasor (or phase lag of U- behind U+). With
equal but opposite amplitudes (u+(0)=1/2= -u-(0)) the complex algebra simplifies a lot.
i t
i t i t
i t
z (t) 1e + + e 1 u+ (0)e + u (0)e = X = 1 =
z2 (t)
2 u (0)ei + t + u (0)ei t 2 ei + t ei t +
(4.10b)
We separate average or overall phase (++)t/2 from more easily observable relative phase (-+)t/2.
z (t) X = 1 =e
z2 (t)
(
)
i + + t
2
(
)
(
)
+ t cos
2
+ t i sin
2
(4.10c)
This is an example the sine-expo and cosine-expo identities introduced in (4.3).
eia eib = e
=
i
a+ b
2
a b i a b
i
e 2 e 2 a+ b
i
2ie 2
sin
ab
2
eia + eib = e
(4.10d)
=
i
a+ b
2
a b i a b
i
e 2 + e 2 a+ b
i
2e 2
cos
ab
2
(4.10e)
© 2008 W. G. Harter
Chapter 4. Phasor analysis of modes and beats
is FAST
phasor
180
Out-of-phase
mode
Unit. 2
Ratio of
Energy 1::Energy 2
Time t=0
ρ=180
phase lag
100%::0%
U
+
ρ=180
phase lag
0
In-phase
mode
Time t=0.083τ
Beat
ρ=150
phase lag
is SLOW
phasor
NOTE
In these examples
the phase of X1 is
90 ahead of
the phase of X2.
Time t=0.166τ
Beat
is FAST
phasor
180
Out-of-phase
mode
ρ=60
Time t=0.25τ
Beat
50%::50%
ρ=120
U
+
ρ=90
ρ=120
phase lag
Time t=0.333τ
Beat
0
In-phase
mode
is SLOW
phasor
Fig. 4.7 Oscillator phases for a symmetric beat. (Initial amplitudes u+(0)=1/2= -u-(0).)
ρ=90
phase lag
ρ=60
phase lag
48
© 2008 Harter
Chapter 4 - 49
Unit 2. Vibration, Resonance, and Wave Motion
Wave phasor-area |z*z| or envelope |z|=|z*z| oscillates at a slower relative frequency -+.
2
z1 (t) = z1 (t)* z1 (t) = cos2
2
z2 (t) = z2 (t)* z2 (t) = sin 2
( + ) t
2
( + ) t
2
(4.11a)
z1 (t) =
(4.11a)
z2 (t) =
(
)
(4.11b)
)
(4.11b)
1 + cos + t
2
(
1 cos + t
2
Fig. 4.7 and Fig. 4.8(a) shows the time evolution of the amplitudes (4.10). It is easier to measure the
envelope beats that have a slower frequency Beat = + than the faster oscillation inside the envelope
that has the average or carrier frequency Average = ( + + ) / 2 = Carrier .
The same frequencies apply for more general initial amplitudes. Fig. 4.8(b) shows the result of
initial amplitudes (u+(0)= 2, u-(0)=-2/2) that give initial (x1(0)=1.5,x2(0)=0.5).
i t
i t i t
i t
z (t) 1 u+ (0)e + u (0)e 1 2e + + e = X = 1 =
z2 (t)
2 u (0)ei + t + u (0)ei t 2 2ei + t ei t +
(4.10b)
Envelope functions now need the complex sum or difference and complex square from (5.20).
(
)
(
)
1.5
(t = 0)
5
+ cos Beat t = 4
0.5 ( Beat t = )
(t = 0)
5
0.5
t =
cos Beat t = 1.5
(
4
Beat t = )
z1 (t) =
2
2
1
u+ (0) + u (0) 2u+ (0)u (0) cos + t =
2
z2 (t) =
2
2
1
u (0) + u (0) + 2u+ (0)u (0) cos +
2 +
(4.12a)
(4.12b)
The minimum-to-maximum amplitude ratio is the Standing Wave Ratio SWR introduced in (4.6).
z MIN
z MAX
=
| u+ (0) | | u (0) |
| u+ (0) | + | u (0) |
|1.5 | | 0.5 | 1
2
= for: u+ (0) = 2,u (0) =
=
2 |1.5 | + | 0.5 | 3
For the (u+(0)=1/2=-u-(0)) case in Fig. 4.8(a) SWR is zero. For Fig. 4.8(b) SWR=1/3.
(4.13)
© 2008 W. G. Harter
Chapter 4. Phasor analysis of modes and beats
(a) Standing Wave
Ratio
SWR = 0
(b) Standing Wave
Ratio
SWR =1/3
Time t=0
Time t=0
1-cos(ω+−ω−)t
2-cos(ω+−ω−)t
τBeat
ω+=11
ω−=14
2+cos(ω+−ω−)t
ω
sin(ω+−ω−)t
2
1+cos(ω+−ω−)t
Beat
period
SWR
= 1
3
cos(ω+−ω−)t
2
ω
Unit. 2
ω+=110
ω−=113
Fig. 4.8 Time evolution of symmetric oscillator beats. (a) Standing Wave Ratio SWR=0. (b) SWR=1/3.
50
© 2008 HarterSoft –LearnIt
Unit 2. Vibration, Resonance, and Wave Motion
2-
51
Chapter 5 Quaternion-Spinor Analysis of Oscillators and Beats
A fundamental equation of quantum evolution, namely Schrodinger's equation, is related to that of
classical coupled oscillators. Consider, the simplest quantum systems: a spin-1/2 particle (NMR) or a
two-level atom. The equation has abstract form (5.1a) that for 2-levels uses 2-D arrays (5.1b-c).
i
t =H t
t
()
()
(5.1a)
H is a 2-by-2 Hermitian (H † = H) matrix operator and Dirac ket is a 2-D complex vector.
A
H=
B + iC
x + ip1 a1 = 1 = 1
= (5.1c)
2 x 2 + ip2 a2 B iC (5.1b)
D Separating real and imaginary parts of amplitudes (5.1c) lets us convert the complex 2D equation (5.1a)
into twice as many real differential equations. The results are as follows.
x˙1 = Ap1 + Bp2 Cx2
p˙1 = Ax 1 Bx2 Cp2
(5.2a)
(5.2b)
x˙2 = Bp1 + Dp2 + Cx1
p˙ 2 = Bx1 Dx2 + Cp1
A classical analog of Schrodinger dynamics
The same equations arise from a classical oscillator Hamiltonian with coordinate and canonical
momentum pairs x j and p j , respectively. (Note: Canonical momentum p j = H
is not the usual mx j .)
x
j
(
)
(
)
A 2
D 2
(5.3a)
Hc =
p1 + x12 + B( x1 x 2 + p1 p2 ) + C( x1 p2 x2 p1 ) +
p + x 22
2
2 2
Hamilton’s equations of motion are identical to Schrodinger’s real equations (5.2).
H c
H
x˙1 =
= Ap1 + Bp1 Cx 2
p˙1 = c = ( Ax1 + Bx1 + Cp2 )
p1
x1
(5.3b)
(5.3c)
H c
H c
x˙2 =
p˙ 2 = = Bp1 + Dp2 + Cx1
= ( Ax1 + Dx2 Cp1 )
p1
x 2
Derivatives of x˙ j in (5.3b) go in p˙ j of (5.3c) to give 2nd order oscillator equations like (3.4).
x˙˙1 = Ap˙1 + B p˙2 Cx˙2
= A( Ax1 + Bx2 + Cp2 ) B( Bx1 + Dx2 Cp1 ) C( Bp1 + Dp2 + Cx1 )
(
)
= A2 + B 2 C 2 x1 ( AB + BD)x 2 C( A + D) p2
(5.4a)
x˙˙2 = Bp˙1 + Dp˙ 2 + Cx˙ 2
= B( Ax1 + Bx2 + Cp2 ) D( Bx1 + Dx2 Cp1 ) + C ( Ap1 + Bp2 Cx 2 )
(
)
= ( AB + BD)x1 B + D C x 2 + C( A + D) p1
2
2
2
(5.4b)
Setting Schrodinger parameter C to zero reduces (5.4) to coupled oscillator equations (3.4).
x1 = K11 x1 + K12 x 2
(5.5a)
x 2 = K 21 x1 + K 22 x 2
(5.5b)
Spring force matrix components Kmn are related below to H -matrix parameters A, B, and D.
K11 = A2 + B 2 ,
K12 = AB + BD,
K 21 = AB + BD,
K 22 = B 2 + D 2 .
(5.6)
© 2008 W. G. Harter
Chapter 5. Phasor analysis of modes and beats
Unit. 2
52
The eigenfrequencies for the Schrodinger equation (5.1) with (C 0) are squares of the
eigenvalues of the K-matrix in (5.5). This is quickly seen in the case A = D and C=0 where the quantum
Hamiltonian matrix (5.1b) has a super-symmetric form.
A B
H =
(5.7a)
B A
This Hamiltonian matrix has the following eigenvalues.
1 = A + B,
2 = A B.
(5.7b)
For the parameters A = D , B, and C=0, the classical K-matrix is super-symmetric, just like (3.4).
k + k12
K=
k12
k12 A2 + B 2
=
k + k12 2AB
2AB A B A B 2
=
B A = H
2
2
B
A
A +B (5.8a)
A is the matrix square of H so the classical acceleration eigenvalues are squares of quantum eigenvalues.
(
K1 = A + B
)2 = k + 2k12 ,
(
K2 = A B
)2 = k .
(5.8b)
Quantum dynamics differs from classical dynamics in important ways. First, classical equations
are second order differential equations so eigenvalues involve squared frequencies as in (5.8b) rather than
frequencies as in (5.7a). Also, quantum stimuli enter multiplicatively by varying components A, B, C, or
D of H . Exact quantum equation i = H is always homogeneous, i.e. i H = 0 is always zero
unlike classical x + K x = a . Finally, parameter C corresponds to classical cyclotron or Coriolis effects.
It gives circular cyclotron orbits if B and A-D are zero.
ABCD Symmetry operator analysis and U(2) spinors
Let us decompose the Hamiltonian operator H in (5.1) into four ABCD symmetry operators that are so
labeled to provide helpful dynamic mnemonics and symmetry names (as well as colorful analogies).
A
B + iC
1 0
0 1
0 i 0 0
B iC = A
+ B
+C
+ D
= Ae11 + B B + C C + De 22
0 0
1 0
i 0
0 1 D =
0 1
0 i A + D 1 0
A D 1 0 + B
+ C
+
2 0 1
2 0 1
1 0
i 0 (5.9a)
A+ D
A D
A + B B + C C
+
0
H=
2
2
The { 1, A , B , C } are best known as Pauli-spin operators { 1= 0 , B= X , C=Y , A= Z }but
ones quite like them were discovered a century earlier by Hamilton who was looking to generalize
complex numbers to 3-dimensional space. Hamilton’s quaternions {1, i, j, k} are related as follows to the
ABCD or ZXY0 operators. (He carved them into a bridge in Dublin in 1843, though, not in technicolor.)
{1=1= 0 , iB=i= iX , iC =j= iY, iA =k= iZ }
(5.9b)
Note: i 2 = j 2 = k 2 = -1. Each squares to negative-1 like imaginary i=-1 (i2=-1). Pauli’s ‘s drop the i
factor so each μ squares to positive 1 (X2 = Y 2 = Z 2 = +1) and each belongs to a cyclic C2 group.
© 2008 Harter
Chapter 5 - 53
Unit 2. Vibration, Resonance, and Wave Motion
We’ll consider each C2 symmetry C2A={1, A}, C2B={1, B}, and C2C={1, C} in turn. They are
labeled as A (Asymmetric-diagonal), B (Bilateral-balanced), or C (Chiral-circular) symmetry. Each is an
archetype of dynamics and symmetry. The systems in Sec. 4.3 belong to A-to-B cases in Fig. 5.1 below.
Fig. 5.1 Potentials for (a) C2A-asymmetric-diagonal, (ab) C2AB-mixed , (b) C2B-bilateral U(2)system.
A secret to Hamilton’s Hamiltonian decomposition (5.9) lies in how it can solve the fundamental 1st order
H = 0 equation (5.1) by evaluating and (most important!) visualizing matrix-exponent solutions.
i (t) = eiH·t (0)
(5.10a)
Hamilton generalized Euler’s expansion e it = cos t i sin t so a matrix exponential becomes powerful.
eiH·t = e
A
i B+iC
BiC ·t
D =e
i
0 1
0 i
A D 1 0 A+ D 1 0
·tiB
·tiC ·ti
·t
2 0 1
2 0 1
1 0
i 0
A A D
A+ D
i•·t /2 i0 ·t
=e
e
where: = t= B t = 2B t and: 0 =
2
2C C
(5.10b)
Each matrix H has a rotation crank vector = ·t that dots with quaternions to solve (5.1). is a 3D ABCor XYZ -space whirl rate like phasor described in Ch. 10. Hamilton generalized a 2D phasor rotation
e it = cos t i sin t and he did this by first generalizing the imaginary number i = 1 as described below.
(a) How spinors and quaternions work
Symmetry relations make spinors X , Y , and Z or quaternions i = i X , j = i Y , and k = i Z powerful.
Each X squares to one (unit matrix 1 = X X ) and each quaternion squares to minus-one (–1=i·i=j·j, etc.)
just like i = 1 . This is true even for spinor components based on any unit vector â = (aX , aY , aZ ) for which
â • â = 1 = a X 2 + aY 2 + aZ 2 . To see this just try it out on any â -component: a = • â = a X X + aY Y + aZ Z .
a 2 = ( • â)( • â) = (aX X + aY Y + aZ Z )(aX X + aY Y + aZ Z )
aX X aX X
+aX X aY Y
= +aY Y aX X
+aZ Z aX X
+aY Y aY Y
+aZ Z aY Y
+aX X aZ Z
aX aX X X
+aY Y aZ Z = +aY aX Y X
+aZ Z aZ Z +aZ aX Z X
+aX aY X Y
+aX aZ X Z
+aY aY Y Y
+aZ aY Z Y
+aY aZ Y Z
+aZ aZ Z Z
© 2008 W. G. Harter
Chapter 5. Phasor analysis of modes and beats
Unit. 2
54
To finish we need another symmetry property called anti-commutation: X Y = Y X , etc. (Check this!)
Put this together with unit squares 1 = X 2 , etc. Then all off-diagonal terms cancel so that 1 = a 2 , too.
a 2 = ( • â)( • â) = (a X X + aY Y + aZ Z )(a X X + aY Y + aZ Z )
aX 2 1
= a X aY X Y
+a X aY X Y
+a X aZ X Z
+aY 1
+aY aZ Y Z
a X aZ X Z
aY aZ Y Z
+aZ 2 1
2
(5.11)
= (a X 2 + aY 2 + aZ 2 )1 = 1
Finally, that anti-commutation relation is cyclic: X Y = i Z = Y X , Z X = i Y = X Z , and
Y Z = i X = Z Y . So, -products do dot • and cross products.
a b = ( • a)( • b) = (aX X + aY Y + aZ Z )(bX X + bY Y + bZ Z )
aX bX 1
+aX bY X Y
aX bZ Z X
+aY bY 1
+aY bZ Y Z
aZ bX Y Z
+aZ bZ 1
= aY bX X Y
+aZ bX Z X
+i(aY bZ aZ bY ) X
= (aX bX + aY bY + aZ bZ )1 +i(aZ bX aX bZ ) Y
+i(aX bY aY bX ) Z
(5.12a)
To see this we write the product in Gibbs notation. (Where do you think Gibbs got his {i,j,k} notation!)
a b = ( • a)( • b)=
(a • b)1
+ i(a b) • (5.12b)
Complex numbers A=Ax+iAy do a similar thing if you *-multiply them as follows.
A*B=(AX +iAY )*(BX +iBY )=(AX iAY )(BX +iBY )=(AX BX +AY BY )+i(AX BY AY BX )=(A • B)+i(A B)
(5.13)
Of course, the results are just the 2D versions of dot and cross products. Hamilton’s idea was to generalize
to three dimensions and even four dimensions. (Lorentz relativity transformations are done by spinors,
too!) So finally Hamilton is able to generalize Euler’s complex rotation operators e +i and e i .
e i = 1 + (i) +
1
1
1
(i)2 + (i)3 + (i)4 = [1
2!
3!
4!
1 2
2!
i(
+
+
1 4
] =
4!
1 3
3!
[ cos]
i( sin )
)
Euler’s series is the result of the binomial series definition of the exponential growth function ert.
2
e rt = limN (1 +
3
rt N
rt N(N 1) rt N(N 1)(N 2) rt ) = limN (1 + N +
+
+ …)
N N N
N
2!
3!
Euler’s identity works because even powers of (-i) are ±1 and odd powers of (-i) are ±i. That is
how i a works, too. Hamilton replaces (-i) with i a in an e i power series to get a sequence of terms
(i a )0 = +1, (i a )1 = i a , (i a )2 = 1, (i a )3 = +i a , (i a )4 = +1, (i a )5 = i a , etc.
This allows Hamilton to generalize Euler’s e i rotation to e i a for any a = ( • a) = aX X + aY Y + aZ Z .
ei = 1 cos i sin ei a = 1cos generalizes to:
i a sin Below are A= Z and C= Y rotations and a a-rotation around a general 3D whirl axis ˆ = ˆ a = â .
e
e
i(A )A
i 1
0
=
0 1 A
1
cosA i (A ) sin A = R(A )
1 0
cos i
= A
0 1
cos i sin A
A
= 0
e
1 0 0 1 sin A
(5.14a)
e iA
0
0
= i cosA i sin A 0
e A e
i a a
=e
i(• â) a
e
i(C )C
i 0
i
=
i 0 C
1 cosC i (C ) sin C = R(C )
1 0
0 i cos i = C
i 0 sin C
0 1
(5.14c)
cos sin C
C
= sin C
cosC = 1cos a i ( • â)sin a
(5.14b)
© 2008 Harter
Chapter 5 - 55
Unit 2. Vibration, Resonance, and Wave Motion
We now see that 3D (ABC) rotations are by an angle a=2a that is twice the angle a in 2D space{x1,x2}.
The “mysterious” factors of 2
A factor of 2 or 21 relates a-rotation of 2D oscillator variables {x1,x2} to 3D vector rotation a in ZXY or
ABC-space. 3D vector â defines a combination a =aAA+ aBB+ aCC of operators A,B ,C to be rotated
by 2-by-2 matrices (5.15) acting twice, fore and aft-1(as operators do in (4.B.6)) by twice the 2D angle a.
R (C )
R1 (C )
A cos sin 1 0 cos
sin C C
C C
= sin cosC 0 1sin C cosC C
cos2 sin 2 2sin C cosC C
C
= 2sin cos
sin 2 C cos2 C C
C
0 1
1 0 cos 2 + = C
1 0 sin 2C
0 1
=
A cos 2C + B sin 2C
R (C )
B R1 (C )
cos sin 0 1 cos
sin C C
C C
= sin cosC 1 0sin C cosC C
2sin cos
cos2 C sin 2 C C
C
= 2
cos sin 2 2sin C cosC C
C
0 1
1 0
sin 2 + = C
1 0 cos 2C
0 1
= A sin 2C + B cos 2C
So the 2D rotation a angle must be exactly 21 the 3D angle a of rotation. When a spin axis goes from up-A
to down-A that is a rotation by C= or 180° in ZXY or ABC-space, but only by C=/2 or 90° in spinor
space {x1,x2}. State of spin up-Z and the state of spin down-Z are orthogonal kets 90° apart. This
analogy to a 2D{x1,x2}-oscillator underlies spin 21 . So, 3D crank vector and spin operator S are defined
for 3D ZXY or ABC-space with a ratio 21 or 2 between a and a= 21 a or between S and = 2S .
e
i• =e
i• / 2
=e
2D angle: = 21 (5.15b)
ˆ
ˆ ˆ )sin sin
(i
cos i
A
B
C
2
2 (5.15a)
2
ˆ
= 1cos i ( • )sin
=
2
2 ˆ
ˆ
ˆ
cos + i A sin (i B + C )sin
2
2
2
ˆ â=2 (5.15c)
2Dspin matrix: S= 21 (5.15d)
3D Crank vector: ==2
a
i S• Eighty years after Hamilton (1924) comes Pauli and Jordan spin S = 2 with its half-quantum factor 2 .
2D polarization and 3D Stokes vector S
In 1862 George Stokes found an application of Hamilton’s mathematics to optical polarization that has a
2D complex oscillator space {E1,E2} analogous to -space {a1,a2}={x1+ip1,x2+ip2} in (5.1c). He gave 3
Stokes vector components Sa that wonderfully define polarization ellipses, 2D HO orbits, and spin 21 states.
Asymmetry S A =
(
(
)
1
1
a A a = a1*
2
2
(
)
(
)
)
(
0 1 a1 1
a*2 = a1*a2 + a*2 a1 = p1 p2 + x1 x 2 2
1 0 a 2
(5.16b)
(
0 i a1 i *
a*2 a a a*2 a1 = x1 p2 x 2 p1 =
i 0 a 2 1 2
2
(5.16c)
SB =
1
1
a B a = a1*
2
2
Chirality SC =
1
1
a C a = a1*
2
2
Balance
1 0 a1 1 *
1
a*2 = a1 a1 a*2 a2 = x12 + p12 x 22 p22 (5.16a)
2
0 1
a 2 2
)
)
© 2008 W. G. Harter
Chapter 5. Phasor analysis of modes and beats
Unit. 2
56
Components of real 3D spin-vector S=(SA, SB, SC) label orbit states in 4D phase space (x1,p1,x2,p2) just as
real components of 3D whirl-vector =(A,B,C) label 4D matrix operators H that whirl these states.
ϑ= 60
A
Θ=Ω t
Ω
A
Ω
B
C
B
ϕ= 45
C
Fig. 5.2 Two views of Hamilton crank vector (,) whirling Stokes state vector S in ABC-space.
Matrix H cranks S around rotation axis = ·t according to (5.15) at whirl rate as Fig. 5.2 depicts.
=||= A2 + B 2 + C 2
(5.17)
Length of =(A,B,C)= ·t =(A·t,B·t,C·t) grows at a constant rate but its direction is fixed if the
constants ( A, B,C, D ) or ( A = A D, B = 2B, C = 2C ) in H are held constant. The -whirl direction is
given by polar coordinates (,) that we call Darboux angles after the inventor of -whirl vectors. The
spin S-vector has polar coordinates, too, so designated in the next section by Euler angles (,).
Fixed points: A port in the storm of action
It helps to look at Fig. 5.2 as analogous to phasor space Fig. 2.15.1. A point on H whirl vector is a stable
fixed point for state spin-S where it can rest and not be whirled. It still twists if S could do so. Such a twist
is in the 0th-overall average angular phase rate 0=(A+D)/2 in (5.10). S-states with S aligned (,)=(,) or
anti-aligned (,)=(,) to are called own-states or eigenstates of H. This lets us compute them.
The Sa-terms in (5.16) are the same as the parts of the classical Hamiltonian Hc in (5.3a).
A 2
D 2
(5.3a)repeated
Hc =
p1 + x12 + B( x1 x 2 + p1 p2 ) + C( x1 p2 x2 p1 ) +
p2 + x 22
2
2
Rearranging the A and D terms lends a classical action form q m pm = iJ = iS to the expression of Hc.
(
Hc =
)
(
)
2
2
2
2
A D x1 + p1 x 2 p2 1 A + D p 2 + x 2 + p 2 + x 2 + B p1 p2 + x1 x 2 +C x1 p2 x 2 p1 +
1
2
2 2 2
2 2 1
=
1
S 2 A A 1
+ B SB 2
1
+ C SC 2
1
+ 0
2
I (5.18a)
Contrast this with the quantum spin matrix operator form for H=•S+01 given by (5.9) thru (5.15).
© 2008 Harter
A
B + iC
Chapter 5 - 57
Unit 2. Vibration, Resonance, and Wave Motion
1
B iC = A D 2
0
D H=
A
0
0
+ 2B 1
21 2
0
+ 2C i
0 2
1
2
+ B SB
SA
2i A + D 1 0
+
2 0 1
0 + C S C
(5.18b)
+ 0 1
Classical Sa-magnitude is I / 2 = SA2 + SB 2 + SC 2 but the matrix forms give: S A2 + S B 2 + SC 2 = 43 1 .
(b) Oscillator states by spinor rotation
Sophus Lie sought to define classical dynamics in terms of transformation operators. The preceding (5.10)
and (5.15) let us do this for a 2D oscillator by relating it to a 3D spinning body. All states a = (aa ) of a 2D
1
2
1
0
oscillator or 3D body are defined by rotation R of an initial 2D state 1 = ( ) or 3D vector S(1) = (0, 0,1) .
()
i• a
a 1
a = 1 = R(a) 1 = e
0
a
2
22
(5.19a)
S A (a)
S(a) = S B (a) = R(a)iS(1) = ei S• S (a)
C
0
0 (5.19b)
33 1
3D rotation R has 3 parameters (A,B,C) or else 3 Euler angles (,,) as shown in Fig. 5.3. That
can define a 2D oscillator’s 4 phase variables (x1,p1,x2,p2) if energy is conserved. Else, we need to include
an intensity amplitude I 1/ 2 =A with the 3 rotation angles. Euler’s ZYZ or ACA rotation of 1st-state |1 gives
i
state |a=Ra|1 of spin S with 2 polar angles (,) and a phase factor e 2 with phase /2.
a = R ( ) 1
(Euler’s definition of any state a )
= R[ aboutZ ] R[ aboutY ] R[ aboutZ ] i cos 2
e
0 2
= i 0 e 2 sin
2
sin i 2
e
2 cos 0
2 0 i 2
e (5.20a)
(Using matrix R(/2) of (5.14a) and R(/2) of (5.14c))
+ A e i 2 cos 2
= i 0 e 2 sin
2
e
e
i
i
i 2
sin A
e cos i x + ip 2 = A 2 e 2 = 1
1
i
x
+
ip
e 2 sin 2
2
cos 0 2 2 2
+ 2
Real xk and imaginary pk parts of phasor variables ak=xk+ipk are functions of 3 Euler angles (,,) and A.
x1= Acos(+)/2·cos/2
x2= Acos(-)/2·sin /2
(5.20b)
p1=-Asin(+)/2·cos/2
p2= Asin(-)/2·sin /2
(5.20c)
But, the 3 components (5.16) of spin vector S depend on only 2 polar angles (,) and I as in Fig. 5.3.
1 2
I
2 2 x1 + p12 x22 p22 =
[cos
sin
]
2
2
2
2
+
+
SB = p1 p2 + x1 x2 =I sin
sin
+ cos
cos
2
2
2
2
+
+
SC = x1 p2 x2 p1 =I cos
sin
cos
sin
2
2
2
2
SA =
I
cos 2
I
cos 2 sin 2 = 2 cos sin I
cos 2 sin 2 = 2 sin sin =
(5.21a)
(5.21b)
(5.21c)
Intensity factor I=A2 is called a norm and is unity (I=1=A) for quantum states. Here it is the total action of
classical oscillators. Spin (SA,SB,SC) is independent of phase /2. Action I is independent of , , and .
Action = 2S0 = a 1 a = x12 + p12 + x 22 + p22 = I[cos2 + sin 2 ]= I = Intensity
2
2
(
)
(5.21d)
According to (5.18 a,b,c) action is twice the spin magnitude: I / 2 = SA2 + SB 2 + SC 2 .
Let a 2D elliptic orbit of frequency have amplitudes A1 and A2, and phase shifts 1 and 2 =1.
© 2008 W. G. Harter
Chapter 5. Phasor analysis of modes and beats
x1= A1cos( t +1)
x2= A2cos( t 1)
p1=-A1sin( t +1)
p2=-A2sin( t 1)
Unit. 2
58
(5.22a)
(5.22b)
This is a case of (5.20). Euler angles (,, ) and action amplitude A are set to match (5.22) as follows.
=2 1
tan/2=A2/A1
=2·t
A2=A12+A22
(5.22c)
This example is used to show how the Stokes-Hamilton formulas expose orbital geometry and dynamics.
Fig. 5.3 The operational definition of Euler ()-angle coordinates is applied to a unit spin-state.
© 2008 Harter
Chapter 5 - 59
Unit 2. Vibration, Resonance, and Wave Motion
The A-view in {x1,x2}-basis
The orbit (5.22) has angles A= 1-2 =2 1, A=2tan-1A2/A1, and A= t/2 with intensity I=A2=A12+A22.
(
(
)
)
cos( t + A ) i sin( t + A ) cos A e i A / 2 cos A a1 2
2
2 x1 + ip1 2
i t
=
= A
= A +i / 2 e
A
A
A x + ip e A sin A a2 2
2
cos(
t
)
i
sin(
t
)
sin
2
2
2
2 (5.23)
Fig. 5.3 shows an ellipse (5.22) next to its ABC space S-vector (5.21) for A or Z-axis Euler angles
=A= 1-2 =2 1=60° (5.24a)
=A=2tan-1A2/A1=60° (5.24b)
A= t/2.
(5.24c)
Cartesian-x1x2 axes in Fig. 5.3a map onto ± A or Z -axis in Fig. 5.3b. Azimuth angle-A off the Baxis is the phase lag between a1=e-i/2 |a1| and a2=e+i/2 |a2| in (5.23). Note projected A1 or A2 box contact
points in Fig. 5.2a. Contact points go to the box diagonal for A =0° or the other diagonal for A =180°,
or the box xk-axes for A =±90°. Polar A angle of S from A-or-z-axis is the angle between ellipse box
diagonals in Fig. 5.2a. An orbit with A=0°(180°) is x1(x2)-polarized. Circular orbits have A =±90°=A.
Fig. 5.3 Polarization described by (a) plane-x1x2 bases and (b) A-axis polar angles of Stokes vector.
Converting an A-based set (5.20) of Stokes parameters into a C-based one or into a B-based one is
simply a matter of cyclic permutation of A, B, and C polar formulas as follows.
I
I
I
cos A
= sin B sin B = cos C sin C
2
2
2
I
I
I
Balance
S B = cos A sin A = cos B
= sin C sin C
2
2
2
I
I
I
Circularity SC = sin A sin A = cos B sin B = cos C
2
2
2
Asymmetry SA =
To find the C-axis polar angle C in terms of A-axis angles A and A, we use (5.25 c).
(5.25a)
(5.25b)
(5.25c)
© 2008 W. G. Harter
Chapter 5. Phasor analysis of modes and beats
Unit. 2
60
The C-view in {xR,xL}-basis
The orbit can be expressed in right and left circular polarization {xR,xL}-bases using angles (C, C, C).
e i C / 2 cosC aR x + ipR 2 i2 C
= R
= A +i / 2 e
C
aL x R + ipR C
e
sin
2 (5.26)
Angles (C, C) are found quickly using (5.25). C-axial polar angle C uses (5.25c). See C in Fig. 5.4b.
sin A sin A = cos C
or: C = ACS(sin A sin A ) = ACS(
3 3
) = 41.4°
2 2
(5.27)
C-axis azimuth angle C relates to A-axis angles A and A by (5.25a) and (5.25b). See C in Fig. 5.4b.
cos A sin A
cos A
1 3 1
= tan C or: C = ATN (cos A sin A / cos A ) = ATN 2( , ) = 40.9°
2 2 2
(5.28)
Half the azimuth angle C is the ellipse tipping angle = C /2, and half –polar-elevation 2=/2C is the
ellipse diagonal half-angle as seen in Fig. 5.4a. Recall, 2D angles are 21 the corresponding 3D ones.
x2
(a) (x1,x2) Space
ψ
A1
A2
C-axis (b) (A,B,C) Space
C
polar
Stokes
elevation
b
ϕ
vector
S
2ψ =
ν
π/2−βC
2ν
x1
a
A
C-axis
azimuth
angle
αC=2ϕ
C-axis
polar
angle
βC
2ϑ
2ψ
B
2ϕ =
αC
I
Fig. 5.4 Polarization described by (a) circular-RL bases and (b) C-axis polar angles of Stokes vector.
A 90° B –rotation R( / 4) x1 = x R of axis A into C gets (C, C, C) from (A, A, A) all at once.
cos
4
i sin
4
i sin4 x1 + ip1 =
cos4 x 2 + ip2 2
2
i C / 2 cosC i / 2
x R + ipR 1 i Ae A cos 2 A i2 A Ae
2 i2 C
e
e
=
=
(5.29)
i 1
+i / 2 A +i C / 2 C x R + ipR A
Ae
sin
Ae
sin
2 2 Circular polarization is natural for things with chirality or “handedness” such as magnetic fields or
Coriolis rotational effects. Then the C-axis becomes the “special z-one” as indicated in Fig. 5.5.
© 2008 Harter
Chapter 5 - 61
Unit 2. Vibration, Resonance, and Wave Motion
C
(a)
Stokes Vector
ABC-Space
Ω
|R
|y
-B
A
|x
-A
β=2ψ
α=2ϕ
B
|L
I
(b)
Polarization
xy-Space
|y
ψ = β/2
b
ϕ=α/2
|x
a
Fig. 5.5 Polarization variables (a) Stokes real-vector space (ABC) (b) Complex xy-spinor-space (x1,x2).
© 2008 W. G. Harter
Chapter 5. Phasor analysis of modes and beats
Unit. 2
(c) How spinors give eigensolutions (Gone in 60 seconds!)
Can you write down all eigensolutions to the following H -matrix in 60 seconds?
10 + 4 cos
3
H = 4 cos sin + i4 sin sin
4
3
4
3
4 cos
sin i4 sin sin 4
3
4
3 = 12
6(1 + i)
10 4 cos
3
6(1 i)
8
We’re not just asking for eigenvalues in 60 minutes, but all eigensolutions, vectors and values, in 60
seconds flat! If you know your spinors, it’s as easy as . Here they are.
eigenvalue 1
eigenvalue 2
2
12 8 +
= 10 + 2 ( 6) + ( 6)
2
2
= 10 + 4 = 14
eigenvector 1
i 1 i 8
cos
e
8
3
e
6 = = i
3
+i
2
e 4
e 8 sin 3
6
2
12 8 +
= 10 2 ( 6) + ( 6)
2
2
= 10 4 = 6
eigenvector 2
i 8
sin
e
i 4 3 i 8
3
e
6 = e
= 3
+i
2
e 8 cos 1 6 The trick: Get the H crank vector polar angles of azimuth , polar , and rate . Here = 8 .
= [(A D), 2B, 2C ] = [cos , cos sin , sin sin ]
A+D + cos 2
2
H = cos sin + i sin sin 2
2
where:
= (A D)2 + (2B)2 + (2C )2
cos sin i sin sin A
2
2
= B + iC
A+D cos 2
2
B iC D (5.30a)
(5.30b)
Eigenstate spin vector S has Euler angles of azimuth = , pole = , and any phase (let: = 0) .
i i i i 2
2
2
2
cos
cos
cos
sin
e
e
e
e
2 e i 2 = 2 2 e i 2 = 2 = =
+i
e +i 2 sin e +i 2 sin e +i 2 cos (5.30c)
e 2 sin 2
2 2 2 A+D A+D +
has eigenvalue: =
has eigenvalue: =
2
2
2
2
Eigenstate spin vector S has same azimuth = , flipped pole = ± , and any phase.
It doesn’t get much easier! You just line up state spin vector- S angles (, ) and (, ) of
Hamiltonian crank vector to get the first (spin-up) eigenstate, and then stick S the other way to get an
orthogonal (spin-down) eigenstate. But, don’t goof the H angles. Use atan2 or Rct Pol .
= atan2(C,B)
1
[tan (C / B) is unreliable]
= atan2(2 B 2 +C 2 ,A D)
62
© 2008 Harter
Chapter 5 - 63
Unit 2. Vibration, Resonance, and Wave Motion
Simple arc-functions like arctan(C/B) are unreliable due to their multi-valued quadrant ambiguity.
(d) How spinors give time evolution
Can you just as quickly write down the evolution operator of that Hamiltonian?
U(0,t) = e
iHt /
=e
12
i 6(1+i)
6(1i)
t /
8
? ?
= ? ?
A formal exponential is pretty useless unless you expand it using the H -crank in (5.15c).
ei( • t )/2 =ei S• t = 1 cos
t
t
ˆ
- i ( • )sin
2
2
This is the ei S• formula (5.15a) with whirl rate-times-time t = replacing turn-axis vector .
U(0,t) = e
i( • t )/2
A+D
i
t 2
The overall phase factor e
may be attached later. (Or ignored)
t
ˆ sin t
= 1 cos
- i ( • )
2
2
(5.31a)
ˆ .)
(The unit crank vector ˆ = [(A D), 2B, 2C] / = [cos , cos sin , sin sin ] is also unit t
t
- i ( A cos + B cos sin + C
sin sin )sin
2
2
1
0
1
0
0
1
0
i
t
t
= - i (
cos
cos + cos sin + sin sin )sin
2
2
0 1 0 1
1 0 i 0 U(0, t)=
1
cos
(5.31b)
Sum this all into a single matrix and you get a useful general evolution operator.
t
t
i cos sin
cos
2
2
U(0, t)= i ( cos i sin ) sin sin t
2
t
t
cos
i cos sin
2
2
=
t
ie+i sin sin
2
t 2 t
t
cos
+ i cos sin
2
2
t t
ˆ sin t
iei sin sin
i
cos
A
2 2
2
=
t
t t
ˆ + i
ˆ sin
cos
+ i cos sin
i B
C
2
2 2
i ( cos i sin ) sin sin
(
)
(
)
ˆ i
ˆ sin t i B
C
2 t
t
ˆ sin
cos
+ i
A
2
2 Our numerical crank rate is = 8. Our unit crank rate vector ˆ fills in the numbers for U(0, t) .
ˆ ,
ˆ ,
ˆ ]
ˆ = [(12 8), 2 6, 2 6 ] = 1 , 6 , 6 = [ cos , cos sin , sin sin ] =[
A
B
C
2
4
4
1
cos 4t i 2 sin 4t
U(0, t)= 6
sin 4t
i (1 + i )
4
1
6
sin 4t 4
1
cos 4t + i sin 4t 2
i (1 i )
Starting from any initial state like (0) = we compute what it will be at time t.
0
(5.31c)
© 2008 W. G. Harter
Chapter 5. Phasor analysis of modes and beats
1
cos 4t i 2 sin 4t
(t) = U(0, t) (0) = 6
sin 4t
i (1 + i )
4
1
6
cos 4t i sin 4t
sin 4t 1
2
4
=
6
0 1
sin 4t cos 4t + i sin 4t (1 i )
4
2
i (1 i )
Unit. 2
64
© 2008 Harter
Chapter 5 - 65
Unit 2. Vibration, Resonance, and Wave Motion
B-Type Oscillation: Simple examples of balanced beats
Resonant beats in Fig. 3.9 are super-positions of a 0°-in-phase mode ( ) of slower frequency slow
and a 180°-out-of-phase mode ( ) of faster frequency fast. Transverse 0°-in-phase modes and 180°-out-of-phase modes do the same. Each mode is denoted by complex amplitude vectors
e
i slow t
1
i fast t 1 1 and e
1 , respectively.
Let’s add them half-and-half or 50-50. (We let slow and fast phases be s=-slowt and f=-fastt.)
a5050 1
1 1 eis + eif 1 1
i t 1 1 i t 1 1
= e slow + e fast = eis + eif = a5050 2
1 2
1 2 1
1 2 eis eif 2
(5.32)
We could write the complex numbers in Cartesian form: eis = cos s + i sin s and eif = cos f + i sin f .
(
(
)
)
a5050 1 eis + eif 1 cos s + i sin s + cos f + i sin f 1 cos s + cos f + i sin s + sin
= 1
= =
a5050 2 eis eif 2 cos s + i sin s cos f i sin f 2 cos s cos f + i sin s sin
2
f X1 + iP1 =
f X 2 + iP2 Cartesian X1=Re a1, P1=Im a1, X2=Re a2, P2=Im a2 are then given. A factored polar form is better:
i s+ f
s f s f i s+ f i s f
i
a5050 1 eis + eif 1 e 2 (e 2 + e 2 ) e 2 cos
2 =
=
1
= s f s+ f
a5050 2 eis eif 2 i s+ f i s f
i
i
s f 2
e 2 (e 2 e 2 ) e 2 i sin
2 Note: this trick easily gives four sin-cos identities like cos
cos A=
eiA + e iA
2
i sin A=
eiA e iA
2
where:
s+ f
s f 1
cos
= (cos s + cos f ) . Now we factor
2
2
2
out overall phase and compute the (get real!) Stokes vector S=(SA, SB, SC) as functions of time.
s+ f
a5050 i
1
=e 2
a5050 2
s f s+ f
cos 2 i
=e 2
s
f
i sin
2 cos 2 gives the following where: = s f = fast slow t .
i sin 2
(
S A = cos
2
cos
1 0 2 i sin = cos
2 0 1
2
i sin 2
S B = cos
2
cos
0 1 2 i sin = cos
2 1 0
2
i sin 2
i sin
SC = cos
2
cos
0 i 2 i sin = cos
2 i
2
0 i sin 2 2 i sin = cos sin + sin cos = sin = sin(s f )
2
2
2
2
2
i cos 2
*
)
2 i sin 2
i sin 2
*
*
cos
i sin
2
2
cos 2
sin
= cos2
= i cos
sin 2 = cos = cos(s f )
2
2
sin i sin cos
2
2
2
2
=0
It is a B-axial rotation of Stokes vector S=(SA, SB, SC)=(cos,0,sin) past ±A and ±C-axes as in Fig. 5.6.
© 2008 W. G. Harter
Chapter 5. Phasor analysis of modes and beats
Unit. 2
66
The angular frequency ||=| fast slow | of the rotation is called a beat frequency . In quantum theory it is
the Rabi rotation frequency or the NMR precession frequency in spin resonance that Rabi pioneered. It is
simply the relative angular velocity between two phasors having a race around the clock. It is twice that of
the half-difference || /2 =| fast slow | /2 seen in phasor space. That mysterious factor of 1/2 discussed
after (5.14) appears again and makes us sharpen our view of space and time.
A
|x〉H crank-Ω vector
for negative B=-S
-C
Ω
-B
|L〉
|(−)〉
C
|(+)〉
B
|R〉
|y〉
-A
Fig. 5.6 Time evolution of a B-type beat. S-vector rotates from A to C to -A to -C and back to A..
For example if you follow the rotation in the X1=Re a1, P1=Im a1, X2=Re a2, P2=Im a2 space you
will see that one rotation by 360° in ABC-space is only half way back to the starting line, and another
360° rotation (either way) is needed in ABC-space to get a full 360° rotation in phasor xy-space. That is, a
rotation of 720° or 0° in the Stokes ABC-space is needed to return to where they started in xy-space.
Points that are separated by 180° in ABC-vector-space map onto phasor (2D-spinor) base vectors
that are only 90° apart since in xy-space x and y are orthogonal. So a 180° separation in xy-space, that
is, a 180° phase factor e ±i = 1 , maps onto the same Stokes-vector in ABC space. Another way to see
this strange phase shift is to look at a C-rotation that happens in Circular polarization environments, or in
a Cyclotron oscillator, or in the presence of a Coriolis or Chiral force of Earth rotation. (Notice all the C’s
including Complex that are used to describe the circular case.) The Fig. 5.7 shows how +x-polarization
returns first to –x-polarization, that is 180° out of phase and needs another 360° rotation around the C-axis
© 2008 Harter
Chapter 5 - 67
Unit 2. Vibration, Resonance, and Wave Motion
to really be back to +x-polarization. The xy-spinor rotates by half the angle its S-vector turns in ABCspace so xy-oscillation vector returns to the North pole “pointing backwards” with a 180° phase shift.
A
|x(15 )〉
|x〉
|x(150 )〉
-B
β=60
|x(30 )〉
β/2=30
|(−)〉
C
|x(45 )〉=|(+)〉
B
|x(60 )〉
H crank-Ω vector
for C=1
|y〉
-A
Ω
|x(120 )〉
Fig. 5.7 Time evolution of a C-type beat. S-vector rotates from A to B to -A to -B and back to A.
© 2008 W. G. Harter
Chapter 5. Phasor analysis of modes and beats
Unit. 2
68
© 2008 HarterSoft –LearnIt
Unit 2. Vibration, Resonance, and Wave Motion
2-
69
Chapter 6 Multiple Oscillators and Wave Motion
Two mass-m-spring-k oscillators coupled by spring-k12 feel a force vector F=(F1,F2) given by
F1 k + k 12 k12 x1 =
• or :-F = K • x
F2 k12 k + k12 x 2 (3.4)repeated
at position vector x=(x1,x2). (Recall discussion around (3.4).) Unless force F points down the vector x, as
it does for special directions u+ (beginner slope) or u- (advanced slope), complicated beating motion seen
in Fig. 5.5 results. The directions u± giving -F=K•u±=k±u± are simple harmonic mode directions.
k + k12
K • u+ = k12
k12 1
1
• = k+ k + k12 1
1
Eigenvalue: k + = k,
Eigenfrequency: + =
k+
m
(6.1a)
k + k12
K • u = k12
k12 1 1
• = k k + k12 1
1
Eigenvalue: k = k + 2k12 , Eigenfrequency: =
k
m
(6.1b)
The u±’s are called eigenvectors and the proportionality factors k± are called eigenvalues k± and lead to
the eigenfrquencies ± in (6.1). This solves two-oscillator motion. Now we consider N oscillators.
(a) The Shower Curtain Model
What if we hook up N oscillators in a ring? Let’s imagine rings of N masses in Fig. 6.1 like lead
weights at the bottom of a shower curtain. An N-dimensional K matrix like (3.4) gives forces Fm.
F0 K
F k
1 12
F2 .
F3 = .
F .
4 FN 1 k12
k12
K
k12
.
k12
.
.
.
.
k12
.
.
K
k12
.
.
.
.
K
k12
.
k12
K
.
k12
k12 . . . . k12 K x0 x 1 K = k + 2k 12
x2 Mg
• x 3 where: k =
x () = 0
4 x N 1 (6.2)
Each mass-M connects through a k12 spring to its two nearest neighbors. The mass at origin coordinate x0
connects to x1 on its right and xN-1 on its left. Then the first mass to the right of origin with coordinate x1
connects to x2 on its right and x1 on its left, and so on around the loop. If all in (6.2) are fixed except x0
(Let x0 vary but fix 0=x1=x2=…xN-1) then x0 has its pendulum oscillation frequency =
k
M
=
g
plus that
of two k12-springs connecting it to fixed neighbors with restoring force –F0=Kx0 = (k+2k12)x0 from (6.2).
This circular shower curtain model may be solved using complex arithmetic and symmetry
arguments. This seemingly silly system helps in Unit 3 to learn some things about the relativistic quantum
universe, another seemingly very silly system! Be prepared to learn a lot in a rather short time. If you have
understood what has gone before in this book then what follows will not be so difficult.
We now use symmetry to find eigenvectors um of the K-matrix in (6.2) for which K•um=kmum. The
um give directions in u-space that are invariant to spring-force matrix K and thus don’t change in time.
© 2008 W. G. Harter
N=2
Chapter 6. Multiple oscillators and wave dynamics
Unit. 2
N=6
N=4
p=3
p=4
p=2
p=5
p=1
p=6
N=3
N=7
N=5
70
p=12
p=0
N=12
p=11
p=7
p=8
p=10
p=9
Fig. 6.1 N-Coupled Pendulums. (Viewed from above.)
Nth Roots of unity
The eigenvectors and eigenvalues of the N-by-N-matrix K (6.2) are made of the complex Nth-roots
of unity, that are solutions to the equation xN=1. The 2-mass solutions (6.1) use 2nd or square roots ±1 of 1.
2i
Euler’s exponential leads to the Nth roots of 1=e
( )
1
2i
, that is, exactly N different solutions to xN= e .
2 i
x N = 1 = e 2 i implies : x = e 2 i N = e N
2 i
and: x m = e N
m
( )N = 1
satisfies: x m
(6.3)
Square, cubic, quartic, quintic(5 ), hexaic (6 ), and duodecaic (12 ) roots of 1 are plotted in Fig. 6.2.
2i/N
.
They form regular polygons whose vertices are powers (N)m of a fundamental Nth root N=e
(ψ12)3
4
(ψ12)2
(ψ12)
m=3
-1=e2i/2
i= e2i/4
e2i/6
e 2i/3
m=2 (ψ )1=e2 i/12
m=4
5
12
(ψ12)
m=5
+1
m=1
+1
+1 -1
-1
(ψ12)6 =-1
(ψ12)12=1
m=0
th
th
th
m=6
e 2i/3
-i=e -2i/4
e 2i/5
e4i/5
+1
+1
e -2i/3
e2i/7
+1
(ψ12)7
=(ψ12)
m=7
−5 m=8
(ψ12)8
=(ψ12)−4
m=12
m=11
m=9
m=10
(ψ12)9
(ψ12)11
=(ψ12)−1
(ψ12)10
=(ψ12)−2
=(ψ12)−3
e-4i/5 e -2i/5
e -2i/3
Fig. 6.2 Nth-Roots of Unity. Collections of Fourier coefficients for discrete N=2,3,4,5,6.7, and 12.
We’ll use N=12 since it’s easy to picture clock numerals. Note that 0 and 12 o’clock share a point
(12)0=ei0=(ei2/12)12=(12)12 as do (12)-1=(12)11, and (12)-2=(12)10, and so on down to the 6 o’clock point
(12)-6=-1=(12)6. Real axis (6 and 12 o’clock) is horizontal in Fig. 6.2 but vertical in later Fig. 6.3.
(b) Solving shower curtain models by symmetry
Eigenvectors of rotate-shuffle-ops r and r -1 are also eigenvectors of K-matrix (6.2).
© 2008 Harter
Chapter 6 - 71
Unit 2. Vibration, Resonance, and Wave Motion
K = K 1 k12 r k12 r 1
. . . . .
1 . . . .
. 1 . . .
r• x = . . 1 . .
. . . 1 .
. . . . .
1
.
.
.
.
.
1 .
where: 1= unit matrix , and :
x 0 x N 1 x x 1 0 x2 x1 • x3 = x2 ,
x x 4 3 x N 1 x N 2 . 1 . . . .
. . 1 . . .
. . . 1 . .
1
r • x = . . . . 1 .
. . . . . .
1
1 . . . . . .
(6.4a)
x 0 x1 x x 1 2
x2 x3 • x3 = x4 x x 4 5
x N 1 x 0 (6.4b)
So let’s find vectors um=(x0,x1,x2,x3,…) that are the same after a rotate-shuffle except for a factor fm.
r -1•um=r -1(x0,x1,x2,x3,…) =( x1,x2,x3,x4,…) = fm (x0,x1,x2,x3,…) = fmum
(6.4c)
Factors fm will be the desired eigenvalues and we’ll have solved r -1, r and K!
The N=2 eigenvector u+=(1,1) in (6.1a) gives us a clue for N=12. Fig. 6.3(g) shows all the
pendulums swinging together in the (m12)=012 wave (All phasors set to 0th-root (12)=1) or -out-ofphase in a (m12)=612 wave (xp set to the pth power of 6th-root (12)=-1) like (6.1b) vector u-=(1,-1).
u0 = (1,1,1,1,1,1,…) (6.5a)
u6 = (1,-1,1,-1,1,-1,…)
(6.5b)
To understand general um eigenvectors consider the (m12)=112 and 212 vectors plotted in Fig. 6.3(g).
x p:
u1 =
u2 =
x0
(1
(1
x1
x2
x3
x4
x5
x6
ik1 1
ik1 2
ik1 3
ik1 4
ik1 5
ik1 6
e
e ik1 2
e
e ik1 4
e
e ik1 6
e
e ik1 8
e
e ik1 10
e
1
x7
ik1 5
e
e ik1 2
x8
ik1 4
e
e ik1 4
x9
ik1 3
e
e ik1 6
x 10
x11
x12
e
e ik1 8
e
e ik1 10
1)
1)
ik1 2
ik1 1
e ik1 = e 2i/12
e i12 k1 = 1
Each phasor in wave u1 leads the phasor to the left of it by 1-hour. Clocks are set 12 o’clock, 1 o’clock, 2
o’clock, 3 o’clock, 4 o’clock, 5 o’clock, and so on for one complete 12 hr. day going around the “world.”
Each phasor in wave u2 leads the phasor to the left of it by 2-hours. Clocks are set 12 o’clock, 2 o’clock, 4
o’clock, 6 o’clock, 8 o’clock, 10 o’clock, and so on for two complete 12 hr. days going around the loop.
This means the u±m eigenvalues of the shuffle-ops r -1 and r are the single-step phase-factor fm = e ±ik1 m .
r 1 • u m = e ik1 m u m
r • u m = e ik1 m u m
(6.6a)
(6.6b)
Here k1=2/N=2/12. The K-matrix has the same eigenvalue and frequency for a u+m wave as for u-m.
[
(
)]
[
)]
(
K • u m = K1 k12 r 1 + r • u m = K k12 e ik1 m + e ik1 m u m = [ K 2k12 cos k1m ]u m
(6.6c)
This gives the wave or spectral dispersion function. An (k) is a “bottom-line” for wave theory.
m =
K 2k12 cos k1m
=
M
(
k + 2k12 2k12 cos 2m / N
M
)
(6.7)
Linear wave motions, velocity, and spreading (dispersion) are ruled by dispersion functions of some kind.
The K-pendulum dispersion function m above is a good example and is plotted with others in Fig. 6.6.
An (k) plot is a graph of per-time (frequency ) versus per-space (wavevector k). In other words
it is a per-space-time graph. Let’s see how these “winks-vs-kinks” functions determine wave physics.
© 2008 W. G. Harter
(a)
N=2
Chapter 6. Multiple oscillators and wave dynamics
r0 r1
(b)
N=4
r0 r1 r2 r3
Unit. 2
(c)
N=6
72
r0 r1 r2 r3 r4 r5
06
04
16
14
(d) 0
1
2
26
24
N=3 r r r
36
34
03
(e)
0
1
2
3
4
N=5 r r r r r -26=4 6
13
-16=5 6
05
23
(f) 0
1
2
3
4
5
6
15
r
r
r
r
r
r
r
N=7
25
07
Re Ψ
-25= 3 5
17
ImΨ
-15= 4 5
27
37
(g)
47
N=12
C12 r 0 r 1 r 2 r 3 r 4 r 5 r 6 r 7 r 8 r 9 r10 r11 5 7
012
67
02
12
112
212
312
412
512
612
712
812
912
1012
1112
Entry in
-912= +312 Row-mN and Column-rp
-812= +412 of a Fourier Wave Table
is a plane wave phasor
-712= +512
Ψm(xp)=e-ikmxp
-612= +612
-1012= +212
*
-512= +712
-412= +812
-312= +912
-212= +1012
-112= +1112
with
wavevector:
km=m2 /N
position point:
xp=p
Fig. 6.3 Discrete wave phasor or Fourier transform tables for N=2,3,4,5,6,7, and 12.
© 2008 Harter
Chapter 6 - 73
Unit 2. Vibration, Resonance, and Wave Motion
(c) Wave structure and dynamics
When watching phasor waves in motion (You should view WaveIt and BohrIt animations in order
to see what’s going on here!) we are struck by the impression that some “spirit” is going through the
phasors. That, like many spooky notions, is an illusion because only the individual phasors determine the
motion of their respective oscillator. But their synchrony begs an explanation and our mind provides one.
To honor our spooky illusion we construct what’s called a wavefunction. We simply “fill in” the
space between each xp-phasor point p=0,1,2… with the continuous plane wave function e
ikx
of a
continuous coordinate x. That is easy to since we are given discrete functions m ( x p ) = e ikm x p in Fig. 6.3
and it is easy to replace xp with x and plot the resulting m ( x ) = e ikm x . Wavevector km is still discrete.
m ( x ) = e ikm x = cos km x + i sin k m x where: k m = m
2
.
N
(6.8a)
This is done in Fig. 6.4 with the real part of the wave plotted darkly and the imaginary part shaded.
Distinguishing and *: Conjugation and time reversal
OK! It’s time to tell about some symmetry conventions. The wave phasors plotted in Fig. 5.3 are
for what’s called the conjugate * or time-reversed-wave. Its phasors run backwards like the engineers’
phasors. Conjugation (*), if you recall from (5.19a), means reverse imaginary part ±sign.
m * ( x ) = e ikm x = cos k m x i sin k m x where: k m = m
2
.
N
(6.8b)
ikx
Wave tables with wavevector km or m-number as a row label must contain the conjugate (e )*=e
-ikx
.
Since we’re plotting real Re in the vertical-up direction and Im in the horizontal-left direction, the
act of conjugation (*) reflects the horizontal direction of phasor arrows. (This will make more sense later
on! It has to do with Dirac’s bra-ket notation x|m=m (x)= m |x * which demands that “bras” or
“row-vectors” carry a star (*) and sounds like sex-discrimination!)
It helps to remember the sine-phase-lag rule from (1.10): A leading phasor feeds its follower.
That’s the wave propagation direction: from leader to follower. A main idea of plane wave eigenvectors is
that they are eigenvectors of the shuffle-op r in (6.4). So each phasor is equally a leader over one side and
a follower by the same lag of the other side. Whatever is taken gets passed on with no chance to swallow
and get fatter! Thus all phasors in a K-matrix eigenvector stay the same area forever. Thus each m (x) is
what we call a stationary state or eigenstate wave. Recall the mnemonic stated after (10.19) in Unit 1:
Imagination precedes reality by one quarter. The imaginary wave (gray) precedes the real one by a 1/4wavelength. Im is the “gonna’be” and Re “is.”
© 2008 W. G. Harter
Chapter 6. Multiple oscillators and wave dynamics
Unit. 2
Fig. 6.4 Continuous “spirit” wavefunctions for N=2-7, and 12. NOTE:Gray Im is for not *.
74
© 2008 Harter
Chapter 6 - 75
Unit 2. Vibration, Resonance, and Wave Motion
(d) Wave superposition
Suppose we add a m=1(xp) wave to a m=2(xp) wave of the same amplitude. The construction in
Steps 1-5 of Fig. 6.5 does this phasor-by-phasor at time t=0, showing both their real and imaginary
components. A vector sum is performed to obtain a resultant phasor at each of 12 points x0, x1, x2, … x11,
where starting point is x0=x12. The algebraic sum is done by the expo-sine-cosine identities (4.10)
( )
( )
imp
( )
( )
ip
m =1 x p + m = 2 x p = e
m =1 x p + m = 2 x p = e
The overall phase factor e
ip
4
=
6
2
12
+e
+e
imp
i2p
3 2
i p
e 2 12
2
12
6
=e
=e
ip
m + m m m ip
2 6 e
2 6
ip
1+ 2 12 ip
2 6 e
2 6
+e
+e
ip
ip
m m 2 6
1 2 2 6
= 2e
ip
= 2e
m +m 2 6
ip
turns 1.5 times while the envelope factor cos
4
cos
cos p
m m
12
p
12
(6.9)
p
1 p2
= cos
12
2 12
does
half a turn in the space between p=0 and p=12.
Fig. 6.5 reveals a beat that appears at first glance to be 100% complete, but the cos
p
12
envelope is
twisted -out of phase at p=12, a kind of half-beat like the example in Fig. 4.8(a). The real wave inside the
envelope does 1/2 phasor turn by p=4, which is 1/3 of the ring circumference from p=0 to p=12. So it’s
doing 3/2 turns per 12 units, but the twist of the envelope makes that look like two full turns.
Wave phase velocity
Each wave-km phasor turns clockwise with time in Step 6 according to its frequency m.
(
)
(
) i m t = e ikm x p e i m t = e i( km x p m t )
m x p ,t = m x p ,0 e
(6.10a)
The wave appears to move. The point where phase km x p mt=0 moves at phase velocity Vphase.
V phase (1 plane wave) =
xp
t
=
m
km
(6.10b)
Phase velocity is time wink rate over spatial kink setting, over k. The wink rate m is given in terms of
kink setting or wavevector km=k1m by a dispersion function like (6.7). For low km, m is linear: m=C·km.
m =
2k12 2k12 cos k m
k
k
= 2 12 sin m Ck m , where: C=Vphase =
M
M
2
(Note: gravity g and
k=
Mg
are
k12
2
for: k m = m
<< 1
M
N
(6.10c)
zero here.) Short waves, like dogs with short legs, must walk faster to keep
up a speed C. If long wave-km=1 advances all its phasor by one tick, then twice-as-kinky and half-as-short
wave-km=2 must do two ticks to keep up in Step 6-8 of Fig. 6.5 and make the beat pattern move rigidly.
Otherwise, wave patterns disperse as we’ll see. That’s why m(km) is called a dispersion function.
© 2008 W. G. Harter
Chapter 6. Multiple oscillators and wave dynamics
Unit. 2
76
Step-1 Construct 4 rows of 13 radius-2unit phasor circles with center-to-center distance of 5-units
(For sideways paper 5-units=1 inch. For normal vertical paper: 5-units=1/2 inch)
Label points p=0, 1, 2, 3, ....11, 12
p=0
p=1
p=2
p=3
p=4
p=5
p=6
p=7
p=8
p=9
p=10
p=11
p=12
(startover)
Step-2 Construct 12 equal time tics on first ( p=0) and last ( p=12). Connect them with light horizontal lines
Step-3 Construct and label the m12=112 or k=1 wave by setting phasors back 1 hr. for each p to 0, -1, -2, -3, ...
(12, 11, 10, 9 ..)
Sketch real and imaginary “spirit“ waves Re Ψm and Im Ψm
p=0
p=1
p=2
p=3
p=4
p=5
p=6
p=7
p=8
p=9
p=10
p=11
0 1
11
10
2
9
3
4
8
7
5
6
Step-4 Construct and label the m12=212 or k=2 wave by setting phasors back 2 hr. for each p to 0, -2, -4, -6, ...
(12, 10, 8, 6 ..)
Sketch real and imaginary “spirit“ waves Re Ψm and Im Ψm
p=0
p=1
p=2
p=3
p=4
p=5
p=6
p=7
p=8
p=9
p=10
p=11
p=6
p=7
p=8
p=9
p=10
p=11
0 1
11
10
2
9
3
4
8
7
5
6
Step-5 Add phasors of k=1 wave to k=2 wave
p=0
p=1
p=2
p=3
p=4
p=5
0 1
11
10
2
9
3
4
8
7
5
6
Step-6 Advanceall m12=112 or k=1 phasors by 1 hr.
Sketch real and imaginary “spirit“ waves Re Ψm and Im Ψm
p=0
11
p=1
p=2
p=3
p=4
p=5
p=6
p=7
p=8
p=9
p=10
p=11
0 1
10
2
9
3
4
8
7
6
5
Step-7 Advance all m12=212 or k=2 phasors by 2 hr.
Sketch real and imaginary “spirit“ waves Re Ψm and Im Ψm
p=0
11
p=1
p=2
p=3
p=4
p=5
p=6
p=7
p=8
p=9
p=10
p=11
p=6
p=7
p=8
p=9
p=10
p=11
0 1
10
2
9
3
4
8
7
6
5
Step-8 Add phasors of k=1 wave to k=2 wave
p=0
11
p=1
p=2
p=3
p=4
p=5
0 1
10
2
9
3
4
8
7
6
5
Fig. 6.5 Adding (m=1) and (m=2) waves. Step 1-5 is initial time t=0. Step 6-8 is later time t=1 tick.
© 2008 Harter
Chapter 6 - 77
Unit 2. Vibration, Resonance, and Wave Motion
Group velocity and mean phase velocity
A sum of two waves forms patterns that don’t resemble either wave and the patterns may change
or disperse over time. If an m-wave and n-wave have equal amplitude their sum is easy by (4.10e).
(
)
(
km x p ,t + kn x p ,t
)
(
=e
i km x p m t
km + kn
+ n
xp m
2
2
) + e i( kn x p n t) = 2e i
t
k kn
n cos m
xp m
t
2
2
(6.11)
The sum wave has two velocities, one for the cosine envelope, and one for the exponential phase inside.
Zeroing the cos(.) phase gives exterior envelope velocity Venvelope or group velocity Vgroup.
km k n
n n
xp m
t = 0 implies: x p = m
t
2
2
k m kn
or: Venvelope =
m n
= Vgroup
k m kn
(6.12)
Zeroing the ei(.) phase gives an interior wave velocity Vinterior or mean phase velocity Vmean phase.
km + k n
+ n
xp m
2
2
+n
t = 0 implies: x p = m
t
k m + kn
or: Vinterior =
m +n
= Vmean phase
km + k n
(6.13)
A linear dispersion (m =Ckm as in Fig. 3.5) makes a sum wave move rigidly, that is, its interior and
envelope wave go the same speed C. By (6.10c), individual wavespeed is C, too. (m/km=C=const.)
Vinterior =
Ckm + Ckn
Ck Ckn
=C= m
= Venvelope if: m = Ck m for all m.
km + kn
k m kn
(6.14)
For nonlinear dispersion, speed Vinterior of a wave’s “guts” differs from speed Venvelope of its “skin.”
(Imagine a slithering boa constrictor swallowing live rabbits that continuously hop along inside it!)
Absolute square | |2= * of equi-amplitude sum =m+ n (6.11) is just a cosine squared.
2
(
= e
i km x p m t
) + ei ( k n x p n t )
2
k kn
n = cos2 m
xp m
t
2
2
(6.15a)
The mean phase of ei(.) cancels (ei(.)*ei(.) = e-i(.)ei(.)=1) leaving envelope wave | |=( * ). Envelope | |
is more complicated for a sum of waves with unequal amplitudes (Am An). (Recall (4.12).)
= =
*
Am
(
e
i km x p m t
) + A ei ( k n x p n t )
n
2
=
k kn
n
2
2
Am + An + 2 Am An cos2 m
xp m
2
2
t
(6.15b)
Both (6.15a) and (6.15b) have the same cos(.) and envelope velocity Vgroup but internal phase behavior is
very different and “gallops” at a rate that depends on SWR (4.6). (The rabbits try to hop out of the boa!)
Fig. 6.6 plots archetypical dispersion functions (km). Constant dispersion (=K=Mg/l) describes
uncoupled (k12=0) pendulums. Weak coupling but no gravity (g=0) is approximated by linear dispersion
(m=Ckm)(6.10c), and quadratic dispersion (m=B+Ckm2) approximates weak coupling with gravity for
low wavevector km<<. Next we will see how any per-space-time (k) dispersion graph is related to a
space-time x(t) graph of paths of wave peaks and zeros.
© 2008 W. G. Harter
Chapter 6. Multiple oscillators and wave dynamics
Unit. 2
78
Archetypical Examples of Dispersion Functions
(a) Constant dispersion (b) Linear dispersion
ωm
ωm
-6-5-4-3-2-1
(c) Quadratic dispersion (d) Phonon dispersion
ωm
ωm
(e) Exciton dispersion
ωm
1
1
1
1
1
0.5
0.5
0.5
0.5
0.5
-6-5-4-3-2-1
12345678
12345678
-6-5-4-3-2-1
-6-5-4-3-2-1
12345678
km=m k1
km=m k1
km=m k1
12345678
-6-5-4-3-2-1
km=m k1
12345678
km=m k1
Applications:
Uncoupled
pendulums
Weakly coupled pendulums (No gravity)
Weakly coupled pendu- Strongly coupled pendu- Strongly coupled pendulums (With gravity)
lums (With gravity)
lums (No gravity)
Movie marquis
Xmas lights
Light in vacuum (Exactly) Light in fiber (Approx) Acoustic mode in solids
Non-relativistic
Sound (Approximately)
Schrodinger matter wave
Optical mode in solids
Relativistic matter
(If exact hyperbola)
Reading Wave Velocity From Dispersion Function by (k,ω) Vectors
slope
(ω-5-ω-2)/(k-5-k-2)
is
(-5,-2)- group velocity
slope
(ω-5+ω-2)/(k-5+k-2)
is
(-5,-2)mean phase velocity
ωm
1
slope (ω-8/k-8)
is (-8)-phase velocity
0.5
slope
(ω8-ω3)/(k8-k3)
is (8,3)- group velocity
slope
(ω8+ω3)/(k8+k3)
is (8,3)-mean phase velocity
slope (ω7/k7)
is (7)- phase velocity
-8 -7 -6-5-4-3-2-1
12345678
km=m k1
Fig. 6.6 Types of dispersion (k) functions and geometry of phase and group velocity.
© 2008 Harter
Chapter 6 - 79
Unit 2. Vibration, Resonance, and Wave Motion
Wave-zero (WZ) and pulse-peak (PP) space-time coordinate grids
Fig. 6.7 and 6.8 compare and superimpose time-vs-space (x,t)-plots of group and phase waves and
their inverse per-time-space or reciprocal space-time plots of frequency-vs-wavevector (,k).
The plots apply to all waves and not just to light. The example in Fig. 6.7 begins by picking four
random numbers, say, 1,2,4, and 4 to insert into frequency-wavevector K2= (2,k2)=(1,2) of a mythical
source-2 and frequency-wavevector K4= (4,k4)=(4,4) of another mythical source-4. Velocity c2=2 /k2
=1/2 of source-2 and c4=4 /k4 =1 of source-4 are unequal. For light waves in Fig. 6.8, c2 equals c4.
Let the continuous waves (CW) from the two sources interfere in a 2-CW sum.
2CW = (e (
i k4 x 4 t
) + ei ( k2 x 2 t ) ) / 2
(6.16a)
To solve for zeros of this sum we first factor it into a phase-wave eip and a group-wave cos g factor.
2CW
=e
=
k +k
+ 2 i 4 2 x 4
t
2
2
(
e
i k p x p t
)
(e
k k
2 i 4 2 x 4
t
2
2
cos
(
+e
gt
kg x
k k
2 i 4 2 x 4
t
2
2
)e
ip
)/2
(6.16b))
cos g
Phase factor eip uses the half-sum (,k)-vector Kphase=(K4+K2)/2 in its argument p = k p x pt . Group
factor cos g has the half-difference (,k)-vector Kgroup=(K4K2)/2 in its argument g = k g x g t .
K phase =
K4 + K2 1 4 + 2 = 2
2 k4 + k2 p 1 4 + 1 2.5
=
= = k p 2 4 + 2 3.0 K4 K2 1 4 2 = 2
2 k4 k2 K group =
(6.16c)
g 1 4 1 1.5
=
= = k g 2 4 2 1.0 (6.16d)
The (,k)-vectors Kn define paths and coordinate lattices for pulse peaks and wave zeros in Fig. 6.7a.
Real zeros (Re=0) have velocity Vphase on Kphase paths. Group zeros (||=0) move at Vgroup on Kgroup.
Vphase =
p
kp
=
4 + 2 2.5
=
= 0.83
k4 + k2 3.0
Vgroup =
(6.17a)
Zeros of phase factor real part Re eip = Re e (
i k p x pt
g
kg
=
4 2 1.5
=
= 1.5
k4 k2 1.0
(6.17b)
) = cos p lie on phase-zero paths where angle p is N(odd)·/2.
(
)
k p x pt = p = N p / 2 N p = ±1, ±3... .
Zeros of group amp-factor cos g=cos ( k g x g t ) lie on group-zero or nodal paths where angle g is N(odd)·/2.
kg x gt = g = N g / 2
(N
g
)
= ±1, ±3... .
Both factors are zero at wave zero (WZ) lattice points (x,t). This defines the lattice vectors in Fig. 6.7a.
kp
k g
p x p
=
=
g t g Np N g 2
(6.18a)
© 2008 W. G. Harter
Chapter 6. Multiple oscillators and wave dynamics
Unit. 2
80
Solving gives spacetime (x,t) zero-path lattice that are white lines in Fig. 6.7a. Each lattice intersection
point is an odd-integer (N p , N g ) combination of wave-vectors P = K phase / 2D and G = K group / 2D .
g p p
g p - p +g k g k p x k g k p g =
=
(N pK group +N g K phase )
=
t 2D
p kg g k p
p kg g k p
(a) Spacetime (x,t)
(6.18b)
(b)Per-spacetime (ω,k)
Wave phase vectors
Wave phase
zero-paths
Wavevector k
Time t
Wave group vectors
Kgroup
=(K4-K2)/2
Kphase=(2.5, 3.0)
K2
Wave
group
nodepaths
Kgroup=(1.5, 1.0)
K4
Kphase
=(K4+K2)/2
Frequency ω
Space x
CW
lattice
PW
lattice
Kphase
source 2
K2=(ω2,k2)
=(1, 2)
K4
Kgroup
K2
source 4
K4=(ω4,k4)
=(4, 4)
Fig. 6.7 “Mythical” sources and their wave coordinate lattices in (a) Spacetime and (b) Per-spacetime.
CW lattices of phase-zero and group-node paths intermesh with PW lattices of pulse, packet, or “particle” paths.
© 2008 Harter
Chapter 6 - 81
Unit 2. Vibration, Resonance, and Wave Motion
Scaling factor 2D/ = 2( p k g g k p ) / converts (per-time, per-space) vectors Kgroup or Kphase into (space,
time) vectors P = (tx ) p or G = (tx )g . (Plot units are set so 2D/ =1 or D=/2. This works only if D is non-zero.)
Fig. 6.7b is a lattice of source vectors K2 and K4 (the difference and sum of Kgroup and Kphase).
1
K 2 = K phase K group = 2 = k 2 2
(6.19a)
4
K 4 = K phase + K group = 4 = (6.19b)
k4 4 Source-2 has phase speed c2 on K2 paths of slope c2. Source-4 has speed c4 on K4 paths of slope c4.
c2 =
2 1
= = 0.5
k2 2
(6.20a)
c4 =
4 1
= = 1.0
k4 1
(6.20b)
One may view the K2 and K4 paths from a classical or semi-classical viewpoint if pulse waves
(PW) were wave packets (WP) that mimic particles. Newton took a hard-line view of nature and ascribed
reality to “corpuscles” but viewed waves as illusory. He misunderstood light if it exhibited interference
phenomena and complained that its particles or “corpuscles” were having “fits.”
Newtonian corpuscular views are parodied here by imagining that frequency 2 = 2 /2
(or 4 = 4 /2 ) is the rate at which source-2 (or 4) emits “corpuscles” of velocity c2 (or c4). Then the
wavelengths 2 =2 / k2 (or 4 =2 / k4 ) are just inter-particle spacing of K2 (or K4) lines in Fig. 6.7a.
Since wavelength 2 (4 ) separates K2 (K4) lattice lines in Fig. 6.7b, one can imagine them as “corpuscle
paths.” The paths are diagonals of the Kgroup(Kphase) wave-zero lattice in time vs space (x,t) of Fig. 6.7a.
This development shows wave-particle, wave-pulse, and CW-PW duality in the cells of each CWPW wave lattice. Each (K2 ,K4)-cell of a PW lattice has a CW vector 2P or 2G on each diagonal, and each
(P,G)-cell of the CW lattice has a PW vector K2 or K4 on each diagonal. This is due to sum and difference
relations (6.16d) or (6.19b) between (P,G)=(Kphase, Kgroup) and (K2, K4).
In order that space-time (x,t)-plots can be superimposed on frequency-wavevector (,k)-plots or
(,)-plots, it is necessary to switch axes for one of them. The space-time t(x)-plots in Fig. 6.7a follow the
convention adopted by most relativity literature for a vertical time ordinate (t-axis) and horizontal space
abscissa (x-axis) that is quite the opposite of Newtonian calculus texts that plot x(t) horizontally. However,
the frequency-wavevector k()-plots in Fig. 6.7b switch axes from the usual (k) convention so that t(x)
slope due to space-time velocity x/t or x/t (meter/second) in Fig. 6.7a matches that of equal per-time-perspace wave velocity /k or /k (per-second/per-meter) in Fig.6.7b.
Superimposing t(x)-plots onto k()-plots also requires that the latter be rescaled by the scale factor
/2D derived in (6.18b), but rescaling fails if cell-area determinant factor D is zero.
© 2008 W. G. Harter
Chapter 6. Multiple oscillators and wave dynamics
Unit. 2
D = p k g g k p = K phase K group
82
(6.21)
Co-propagating light beams K2= (2,k2)=(2c,2) and K4= (4,k4)=(4c,4) in Fig. 6.8b have D=0 since all Kvectors including Kphase=(p,kp)=(3c,3) and Kgroup= (g,kg)=(c,1) lie on one c-baseline of speed c that has
unit slope (/ck=1) if we rescale (,k)-plots to (,ck) and (x,t)-plots to (x,ct).
(a) Spacetime (x,ct)
(b) Per-spacetime (ω,ck)
Wave zero-paths all the same speed c
Time ct
Wavevector ck κ
Kgroup
=(K4-K2)/2
K2
K4
Kphase
=(K4+K2)/2
Frequency ω
Space x
Infrared laser
source 2
K2=(ω2,k2)
=(2c, 2)
Replaced by:
source 4
Krypton laser
K4=(ω4,k4)
=(4c, 4)
Fig. 6.8 Co-propagating laser beams produce a collapsed wave lattice since all parts have same speed c.
In Unit 3 counter-propagating (right-left) light wave vectors (R,L)= (K2,-K4) are used to make CW
bases (P= Kphase, G= Kgroup) with a non-zero value for area D =|GxP|. Opposing PW base vectors are sum
and difference (R,L)=(P+G,P-G) of CW bases so a PW cell area |RxL| is twice that of CW cell |GxP|.
|RxL|= |(P+G) x (P-G)|=2|GxP|
Wave cell areas are key geometric invariants for relativity and quantum mechanics.
(6.22)
© 2008 Harter
Chapter 6 - 83
Unit 2. Vibration, Resonance, and Wave Motion
(e) Counter-propagation: Standing and galloping waves
Counter-propagating light waves include standing waves (Fig. 6.9) or galloping waves (Fig. 6.10). The
latter are the rule rather than the exception. You get galloping unless you turn off one of the lasers!
i kx t )
that traces 45° wave zero lines going
Turning off the right laser gives a pure right-moving wave e (
at a constant lightspeed c as shown in the space-time plot of Fig. 6.10(a). Turning on even a small amount
i kx t )
of left-moving wave e (
results in galloping paths as shown in Fig. 6.10(b-e). The real wave Re
gallops faster than light once each half-cycle as it leap-frogs a similarly galloping Im.
As the relative amount of left moving wave increases, galloping becomes more pronounced, and
then, for exactly equal left and right amplitudes, the zeros of the real standing wave gallop infinitely fast at
each moment Re is zero everywhere. (Being everywhere is tantamount to going infinitely fast.) This is
the special case that gives a Cartesian (x,ct) grid shown in Fig. 6.9. Finally, for dominant left-moving
amplitudes, the galloping reverses sign and then subsides as in Fig. 6.10(e-f).
Counter-propagating laser waves in Fig. 6.10 have the following wave zeros of Re.
(
)
(
)
0 = Re ( x, t) = Re Ae i k 0 x 0 t + A ei k 0 x 0 t = A [ cos k 0 x cos 0 t + sin k 0 x sin 0 t ] + A [ cos k 0 x cos 0t sin k0 x sin 0 t]
= ( A + A )[ cos k 0 x cos 0t ] + (A A )[ sin k 0 x sin 0 t]
Galloping varies according to a Standing Wave Quotient SWQ or its inverse Standing Wave Ratio SWR.
tank 0 x = SWQ cot 0 t
(6.23a) where: SWQ =
A + A
1
=
A A SWR
(6.23b)
The time derivative gives upper and lower speed limits in terms of V phase =
0
= c and SWQ or SWR.
k0
c SWQ for: t = 0, , 2 ...
csc 2 0t
c SWQ
dx
= c SWQ
=
=
2
2
2
2
dt
sec k 0 x cos 0 t + SWQ sin 0 t c SWR t = / 2, 3 / 2,...
(6.23c)
These are seen in Fig. 6.10, but the fastest gallop is ± as seen for a “standing” wave in Fig. 6.9.
Time
ct
Dye Laser
600 THz
Dye Laser
600 THz
1.67 fs
1.0 fs
1 femtosecond
1.0 fs=10-15s
Wavelength
λ =1/2 μm
Period
τ=5/3 fs
0
0.5 μm
Space
1.0 μm x
1 micron
1.0 μm=10-6meter
Fig. 6.9 Standing wave (SWR=0) due to counter-propagating 600Thz (green) laser waves.
© 2008 W. G. Harter
Chapter 6. Multiple oscillators and wave dynamics
(a) Right moving SWR=+1
Unit. 2
(f) Left moving SWR=-1
Ψ=1.0 ei(2x -2t)+0.0 ei(-2x -2t)
Ψ=0.0ei(2x -2t)+1.00 ei(-2x -2t)
Laser
time
ct
Laser space x
(b) Right galloping SWR=+1/2
Ψ=1.5ei(2x -2t)+0.5 ei(-2x -2t)
(c) Right galloping SWR=+1/10
Ψ=1.1ei(2x -2t)+0.9 ei(-2x -2t)
(e) Left galloping SWR=-2/3
Ψ=0.20ei(2x -2t)+1.00 ei(-2x -2t)
(d) Left galloping SWR=-1/8
Ψ=0.84ei(2x -2t)+1.08 ei(-2x -2t)
Fig. 6.10 Spacetime plots of monochromatic waves of varying Standing Wave Ratio.
84
© 2008 Harter
Unit 2. Vibration, Resonance, and Wave Motion
Chapter 6 - 85
Kepler’s Law for galloping
Galloping wave velocity (6.23c) is related to Kepler’s Law for isotropic force field orbits, such as
in a 2D oscillator orbit constructed by Fig. 6.11 or Fig. 2.1. (Recall Fig. 1.9 in Unit 1.) Clockwise polar
angle (t) of ellipse-orbiting point P=(x=a sin t, y=a cos t) and orbital phase t, relate as follows.
y
b
tan ( t ) = = cot t .
(6.24)
x
a
This resembles the galloping wave equation (6.23a) with the ellipse aspect ratio b/a replacing a standing
wave ratio. To conserve orbital angular momentum rxv in the absence of torque, the orbital velocity v(r)
gallops to a faster v(b) at perigee (r=b) and a slower v(a) at apogee (r=a). In the same way waves in Fig.
6.10 gallop faster through smaller parts of their envelope and slow down as their amplitudes grow.
Analogy of laser wave dynamics (6.23) to classical orbital mechanics (6.24) has physical as well
as historical use. Wave galloping shown in Fig. 6.10 happens equally in systems with open or infinite
boundaries as it does in closed or periodic (ring laser or Bohr-ring) systems. In fact, Fig. 6.11 are pictures
of 2nd lowest (km=±2)-modes of a micro-ring-laser or the 2nd excited Schrodinger (m=±2)-waves on a Bohrring. Exactly two wavelengths fit in each space frame and two periods fit in each time frame.
Analogy with polarization ellipsometry
If a frame in Fig. 6.10 were drawn instead for 1st or fundamental (m=±1)-waves or (km=±1)-modes
of either ring system it would just be a 1/4-area square section with only one sine wave per frame. (Recall
Fig. 3.1.1(c).) Such a wave has an average dipole moment p=p that orbits an ellipse like radius r in Fig.
6.11 and is analogous to a polarization figures used to depict states in optical ellipsometry.
In the polarization analogy, purely right-moving (m=+1) or purely left-moving (m=-1) wave states
+ikx
-ikx
e
and e are analogous, respectively, to right or left circular polarization states. Equal combinations
+ikx -ikx
+ikx -ikx
e +e =2cos kx or e -e =2isin kx are analogous, respectively, to x-plane or y-plane polarization.
+ikx
-ikx
The vast majority of arbitrary combinations ae +be are analogous to general elliptical polarization.
A polarization vector for elliptic states enjoys the same Kepler galloping described by Fig. 6.11.
Perhaps the simplest explanation of wave galloping in Fig. 6.10 uses an analogy with the elliptical
polarization states as in Fig. 6.12. The uniformly spaced ticks on the circular polarization circles are
crowded into a traffic jam at the long axes of their elliptic orbits as the aspect ratio b/a or SWR approaches
zero. The ticks near the short axes maintain their spacing in the Kepler geometry. Like a uniformly
turning lighthouse beacon viewed edge-on, the beam is seen to gallop by quickly and then slow to a crawl
as it swings perpendicular to the line of sight.
An animated simulation of Fig. 6.11 provides an optical illusion of a uniformly rotating circular
disc tipped out of the XY plane by an angle of sin-1(b/a), that is, a perspective projection.
© 2008 W. G. Harter
Chapter 6. Multiple oscillators and wave dynamics
SWR = b/a = 1/5
ω.t
Y
y
A
B
a
86
b cos ω t
tanφ(t)= x = a sin ω t
x=a sin ω t
Highest speed v=5
at perigee r=b=1
Unit. 2
φ(t)
P
O
y=b cos ω t
b
X
Lowest speed v=1
at apogee r=a=5
ω
Fig. 6.11 Elliptical oscillator orbit and a Kepler construction
y
b/a = 1/1
b/a = 1/2
right
circular
polarization
r-elliptical
polarization
y
b/a = -1/1
x
left
circular
polarization
b/a = -1/2
b/a = -1/8
b/a = 1/10
b/a = 0
x-plane polarization
Fig. 6.12 Elliptical polarization states of varying aspect ratio a/b or standing wave ratio SWR. This
figure is analogous to Fig. 6.10 according to the Keplerian geometry of Fig. 6.11.
x
© 2008 Harter
Chapter 6 - 87
Unit 2. Vibration, Resonance, and Wave Motion
(e) Taming the phase: Wavepackets and pulse trains
In each of Fig. 6.9 or Fig. 6.10 real wave Re(x,t) is plotted in spacetime. If the intensity * or
envelope || is plotted, the part of the wave having the fast and wild phase velocity disappears leaving
only its envelope moving constantly at the slow and tame group velocity.
For example, the complicated dynamics of the (SWR=-1/8) galloping in Fig. 6.10(d) is reduced to
parallel grooves by a ||-plot in Fig. 6.13(a). The grooves follow the group envelope motion that has only
a steady group velocity. The lower part of Fig. 6.10 is thus tamed. Pure plane wave states Fig. 6.10 (a)
and Fig. 6.10 (f) are tamed even more in a ||-plot to become featureless and flat like their envelopes.
One gets a glimpse of phase behavior in an envelope or * plot by adding the lowest scalar DC
fundamental (m=0)-wave 0=1 to a galloping combination wave such as = a +4 + b -1. The result
*
* *
* *
*
*
= (1 + a+4 + b1) (1 + a+4 + b1) = 1 + a+4 + b1 + a +4 + b 1 + (6.25)
*
= 1 + 2 Re + is plotted in the upper part of Fig. 6.13(b). The DC bias keeps the phase part from canceling itself, and the
probability distribution shows signs of, at least half-heartedly, following the fast phase motion of the Re
wave plotted underneath it. (Dashed lines showing phase and group paths are superimposed on *.)
Indeed, (6.25) shows that if the (1+* ) background could be subtracted, then the real wave Re
(
)
plots of Fig. 6.13 would emerge double-strength! However, such a subtraction, while easy for the theorist,
is more problematic for the experimentalist. More often one must be content with results more like the
upper than the lower portions of Fig. 6.13. It’s a bit like watching an orgy going on under a thick rug.
However, such censorship can be a welcome feature. As more participating Fourier components
enter the fray, a simpler view can help to sort out important effects that might otherwise be hidden in the
milieu. We consider examples of this with regard to sharper wavepackets such as pulse trains and
Gaussian wavepackets.
Continuous Wave (CW) versus Optical Pulse Trains (OPT): colorful versus colorless
Optical pulse train (OPT) waves are Fourier series of N continuous wave (CW) 1-harmonics.
N ( x,t ) = 1+ e (
i k1 x 1t )
= 1+
+e (
i k1 x 1t )
2e i1 t cosk 1x
+e (
i k 2 x 2 t )
i k x 2 t)
+ e( 2
+
2e i 2 t cos k2 x
+ …+ e (
i k N x N t )
i k x N t)
+ e( N
+… +
2e i N t cos k N x
(6.26a)
The fundamental OPT or (N=0-1) beat wave in Fig. 6.14(b) is a rest-frame view of Fig. 6.13(b)
1( x, t) = 1 +
1
2e i 1t cos k1 x
(6.26b)
should be compared to the pure or unbiased fundamental (m=±1)-standing wave 1 in Fig. 6.14(a).
1( x, t) = 2e i1 t cos k1x
(6.26c)
© 2008 Harter
Unit 2. Vibration, Resonance, and Wave Motion
Chapter 6 - 89
The real part Re 1 is discussed in connection with the Cartesian spacetime wave grid in Fig.
3.2.3(a). The modulus | 1|, unlike | 1|, is constant in time as indicated by the vertical time-grooves at the
extreme upper right of Fig. 6.14(a). In contrast, the magnitude | 1| of the DC-biased beat wave makes an
“H” or “X” in its spacetime plot of Fig. 6.14(b) thereby showing the beats. The width of the fundamental
(0-1) beat is one fundamental wavelength x=2/k0 of space and one fundamental period t=2/ 0 of time.
Including N=2,3,… terms in (6.26) reduces the pulse width by a factor of 1/N as seen in Fig. 6.14
(c-e) below. The spatial
sinN x
/x wave shape is the same as is had by adding N=2,3,… slits to an elementary
optical diffraction experiment. Adding more frequency harmonics makes the pulse narrower in time, as
well as space. Using 12 terms with 11 harmonics reduces the pulse width to 1/11 of a fundamental period.
A pico-period pulse would have a trillion harmonics!
Reducing pulse width or spatial uncertainty x and temporal duration t of each pulse requires
increased wavevector and frequency bandwidth k and . The widths obey Heisenberg relations xk~2
or t~2. The sharper the pulses the more white or colorless they become. Finally, the spacetime plots
will simplify to simple equilateral diamonds or 45°-tipped squares shown in the N=11 plots of Fig.
6.14(e). (They resemble the sketched pulse paths of Fig. 6.7(b).)
As N increases there is much less distinction between the Re and | | plots than there is for the
cases of N=1,2, or 3 shown in the preceding plots of Fig. 6.14(a-d). However, as in Fig. 6.13, there is a
considerable distinction between Re and | | plots with all Re plots being sharper than | | plots in all
cases including the high-N cases. Having phase information increases precision particularly for low N.
The sharpest set of zeros, somewhat paradoxically, are found in the N=1 case of Fig. 6.14(a) and
for the unbiased Re plot of the Cartesian wave grid. However, plotting zeros by graphics shading is one
thing. Finding experimental phase zeros using * or || squared, is quite another thing.
A close look at the center of the Re plot for N=11 in Fig. 6.14(e) reveals a tiny Cartesian
spacetime grid. It is surrounded by “gallop-scallops” similar to the faster-and-slower-than-light paths
shown in Fig. 6.7. It is due to the interference of counter-propagating ringing wavelets that surround each
counter propagating
sinN x
/x –pulse. In contrast the | | plots for which the ringing leaves only vertical
grooves like those that occupy the entire N=1 plot of || in Fig. 6.14(a).
© 2008 W. G. Harter
Chapter 6. Multiple oscillators and wave dynamics
(a)Unbiased Ψ=ψ-1 + ψ+4
(b)DC biased Φ=ψ0 + Ψ
Fig. 6.13 Examples of group envelope plots of galloping waves (a) Unbiased. (b) DC biased.
Unit. 2
88
© 2008 W. G. Harter
Chapter 6. Multiple oscillators and wave dynamics
Unit. 2
(a) 1-Fourier Component Continuous Wave (CW Stationary State)
k= 1
(b) (0 to 1)-Fourier Component Train of τ-Wide-Pulses (Beats)
k= 0
plus:
k= 1
equals:
Δt=τ
space x or time ct
(c) (0 to 2)-Fourier Component Train of 1/2-τ-Wide-Pulses
plus:
plus:
Δt=τ/2
k= 0
k= 1
k= 2
equals:
space x or time ct
(d) (0 to 3)-Fourier Component Train of 1/3-τ-Wide-Pulses
plus:
k= 0
k= 1
plus:
k= 2
plus:
k= 3
equals:
Δt=τ/3
space x or time ct
(e) (0 to 11)-Fourier Component Train of 1/11-τ-Wide-Pulses
k=
k=
k=
k=
k=
k=
k=
k=
k=
k=
k=
k=
0
1
2
3
4
5
6
7
8
9
10
11
Δt=τ/11
space x or time ct
Fig. 6.14 Optical Pulse Trains (OPT) and Continuous Wave (CW) Fourier components
90
© 2008 Harter
Chapter 6 - 91
Unit 2. Vibration, Resonance, and Wave Motion
Wave ringing: mMax-term cutofft effects
An analysis of wave pulse ringing reveals it may be blamed on the last Fourier component added.
There are 11 zeros in the ringing wave envelope in Fig. 6.14(e), the same number as in the 11th and last
Fourier component. An integral over k = m2 / N approximatea a Fourier sum S(mMax) up to a maximum
mMax=11. The unit sum interval m=1 is replaced by a smaller k-differential dk multiplied by
S(m Max ) =
m Max
m =m Max
e im ( ) =
m Max
m = m Max
k
k
N
N
i Max ( )
i Max ( )
2
2
e
e
i ( )
m e im( ) .
N
k Max
k Max
m N
=
2
dk
dk
m ik 2 ( )
e
dk
(6.27)
k
N( )
sin Max
sin mMax ( )
2
=2
=2
( )
This geometric sum verifies our suspect’s culpability. The sum rings according to the highest mMax-terms
while lesser m-terms seem to experience an interference cancellation. The last-one-in is what shows.
Ringing supressed: mMax-term Gaussian packets
Ringing is reduced by tapering off higher-m waves so they cancel each other’s ringing and no
single wave dominates. A Gaussian e-(m/m)2 taper makes cleaner “particle-like” pulses in Fig. 6.15.
SGuass (m Max ) =
m2
1
1
2
e m e im =
e
2 m = 2 m =
m
m Max
2
im e
, where: m =
mMax
(6.28a)
Completing the square of the exponents extracts a Gaussian -angle wavefunction e (m /2) with an
2
angular uncertainty that is twice the inverse of the momentum quanta uncertainty m. ( =2/m).
m
m 2 m 2
m 2
1 m i 2 2 A( m, ) 2 (6.28b)
SGuass (m Max ) =
e
=
e
2
2 m = Definition m=mMax/ of momentum uncertainity relates half-width-(1/e)th-maximum m to the value
m=mMax for which the taper e-(m/m)2 is e-. (e- = 0.04321 is an easy-to-recall number near 4%. Waves
e im
beyond e im Max have e-(m/m)2 amplitudes below e-.) Amplitude A(m,) becomes an integral for
large mMax as does (6.27). Then A(m,) approaches a Gaussian integral whose value itself is mMax.
A( m, ) =
e
m 2
m
i
2 m
m = m
dk e
>>1 k 2
m = m = mMax
(6.28c)
The resulting Gaussian wave e ( / ) has angular uncertainty =Max/ defined analogously to m.
2
m 2
m 2
2
1 m Max
m
m
2 SGuass (m Max ) e m e im = Max e
= Max e where: = Max
2 m = m Max
2
2
(6.28d)
Uncertainty relations in Fig. 6.15 are stated using m and or in terms of 4% limits mMax and Max.
m· = 2
(6.29a)
mMax·Max = 2
(6.29b)
In Fig. 6.14, the number of pulse widths in interval 2 is the the number mMax of (>4%)-Fourier terms.
© 2008 W. G. Harter
Chapter 6. Multiple oscillators and wave dynamics
Unit. 2
-2 -1 0 1 2 3
(a)
Narrow
spectral
distribution
Δm=mMax/
=2/ =1.13
Wide
wavepacket
φMax=2 /mMax
= 2 /2=3.1
-2 -1 0 1 2 3 4 5
(b)
Medium
spectral
distribution
Δm=mMax/
=6/ =3.4
Medium
wavepacket
φMax=2 /mMax
= 2 /6=1.05
-15
-10
-5
0
5
10
15
(c)
Broad
spectral
distribution
Δm=mMax/
=15/ =8.5
Narrow
wavepacket
φMax=2 /mMax
= 2 /15=0.4
Fig. 6.15 Gaussian wavepackets. (Ringing is reduced compared to Fig. 6.14.)
92
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement