You are on page 1of 265

Mechanism Kinematics & Dynamics and

Vibrational Modeling


Dr. Robert L. Williams II
Mechanical Engineering, Ohio University


NotesBook Supplement for
ME 3011 Kinematics & Dynamics of Machines
2014 Dr. Bob Productions

williar4@ohio.edu
people.ohio.edu/williar4




These notes supplement the ME 3011 NotesBook by Dr. Bob

This document presents supplemental notes to accompany the ME 3011 NotesBook. The outline
given in the Table of Contents on the next page dovetails with and augments the ME 3011 NotesBook
outline and hence is incomplete here.

m
2
k
x(t)
1
k
x (t)
1
x (t)
2


2

ME 3011 Supplement Table of Contents

1. INTRODUCTION................................................................................................................................ 4
1.3 VECTORS. CARTESIAN RE-IM REPRESENTATION (PHASORS) ............................................................. 4
1.6 MATRICES .......................................................................................................................................... 6
2. KINEMATICS ANALYSIS .............................................................................................................. 15
2.1 POSITION KINEMATICS ANALYSIS .................................................................................................... 15
2.1.1 Four-Bar Mechanism Position Analysis .................................................................................. 15
2.1.1.1 Tangent Half-Angle Substitution Derivation and Alternate Solution Method .................. 15
2.1.1.3 Four-Bar Mechanism Solution Irregularities ..................................................................... 20
2.1.1.4 Grashofs Law and Four-Bar Mechanism J oint Limits ..................................................... 21
2.1.2 Slider-Crank Mechanism Position Analysis ............................................................................. 29
2.1.3 Inverted Slider-Crank Mechanism Position Analysis .............................................................. 32
2.1.4 Multi-Loop Mechanism Position Analysis ............................................................................... 38
2.2 VELOCITY KINEMATICS ANALYSIS .................................................................................................. 42
2.2.2 Three-Part Velocity Formula Moving Example ....................................................................... 42
2.2.5 Inverted Slider-Crank Mechanism Velocity Analysis ............................................................... 44
2.2.6 Multi-Loop Mechanism Velocity Analysis ................................................................................ 48
2.3 ACCELERATION KINEMATICS ANALYSIS .......................................................................................... 52
2.3.2 Five-Part Acceleration Formula Moving Example .................................................................. 52
2.3.4 Slider-Crank Mechanism Acceleration Analysis ...................................................................... 54
2.3.5 Inverted Slider-Crank Mechanism Acceleration Analysis ....................................................... 55
2.3.6 Multi-Loop Mechanism Acceleration Analysis ........................................................................ 61
2.4 OTHER KINEMATICS TOPICS ............................................................................................................ 65
2.4.1 Link Extensions Graphics ......................................................................................................... 65
2.5 J ERK KINEMATICS ANALYSIS ........................................................................................................... 67
2.5.1 Jerk Analysis Introduction ....................................................................................................... 67
2.5.2 Mechanism Jerk Analysis ......................................................................................................... 70
2.6 BRANCH SYMMETRY IN KINEMATICS ANALYSIS .............................................................................. 71
2.6.1 Four-Bar Mechanism ............................................................................................................... 71
2.6.2 Slider-Crank Mechanism.......................................................................................................... 73
2.7 KINEMATICS ANALYSIS EXAMPLES ................................................................................................. 75
2.7.1 Term Example 1: Four-Bar Mechanism .................................................................................. 75
2.7.2 Term Example 2: Slider-Crank Mechanism ............................................................................. 86
3. DYNAMICS ANALYSIS .................................................................................................................. 93
3.1 DYNAMICS INTRODUCTION .............................................................................................................. 93
3.2 MASS, CENTER OF GRAVITY, AND MASS MOMENT OF INERTIA ....................................................... 94
3.4 FOUR-BAR MECHANISM INVERSE DYNAMICS ANALYSIS ............................................................... 106
3.5 SLIDER-CRANK MECHANISM INVERSE DYNAMICS ANALYSIS ....................................................... 109
3.6 INVERTED SLIDER-CRANK MECHANISM INVERSE DYNAMICS ANALYSIS....................................... 111


3

3.7 MULTI-LOOP MECHANISM INVERSE DYNAMICS ANALYSIS ............................................................ 119
3.8 BALANCING OF ROTATING SHAFTS ................................................................................................ 124
3.9 INVERSE DYNAMICS ANALYSIS EXAMPLES .................................................................................... 128
3.9.1 Single Rotating Link Inverse Dynamics Example .................................................................. 128
3.9.2 Term Example 1: Four-Bar Mechanism ................................................................................ 131
3.9.3 Term Example 2: Slider-Crank Mechanism ........................................................................... 137
4. GEARS AND CAMS ....................................................................................................................... 143
4.1 GEARS ............................................................................................................................................ 143
4.1.1 Gear Introduction ................................................................................................................... 143
4.1.2 Gear Ratio .............................................................................................................................. 149
4.1.3 Gear Trains ............................................................................................................................ 152
4.1.4 Involute Spur Gear Standardization ...................................................................................... 154
4.1.5 Planetary Gear Trains ........................................................................................................... 163
4.2 CAMS ............................................................................................................................................. 171
4.2.1 Cam Introduction ................................................................................................................... 171
4.2.2 Cam Motion Profiles .............................................................................................................. 174
4.2.3 Analytical Cam Synthesis ....................................................................................................... 179
5. MECHANICAL VIBRATIONS INTRODUCTION .................................................................... 186
5.2 MECHANICAL VIBRATIONS DEFINITIONS ....................................................................................... 186
6. VIBRATIONAL SYSTEMS MODELING.................................................................................... 189
6.1 ZEROTH-ORDER SYSTEMS ............................................................................................................. 189
6.2 SECOND-ORDER SYSTEMS ............................................................................................................. 199
6.2.1 Translational m-c-k System Dynamics Model ........................................................................ 199
6.2.3 Pendulum System Dynamics Model ....................................................................................... 200
6.2.4 Uniform Circular Motion ....................................................................................................... 202
6.4 ADDITIONAL 1-DOF VIBRATIONAL SYSTEMS MODELS ................................................................... 203
6.5 ELECTRICAL CIRCUITS MODELING ................................................................................................. 234
6.6 MULTI-DOF VIBRATIONAL SYSTEMS MODELS ............................................................................... 242




4

1. Introduction

1.3 Vectors. Cartesian Re-Im Representation (Phasors)

Here is an alternate vector representation.

i
P Pe
u
=

The phasor
i
Pe
u
is a polar representation for vectors, where P is the length of vector P , e is the
natural logarithm base, 1 i = is the imaginary operator, and u is the angle of vector P .
i
e
u
gives the
direction of the length P, according to Eulers identity.

cos sin
i
e i
u
u u = +

i
e
u
is a unit vector in the direction of vector P .

Phasor Re-Im representation of a vector is equivalent to Cartesian XY representation, where the real (Re)
axis is along X (or

i ) and the imaginary (Im) axis is along Y (or

j ).

cos
(cos sin )
sin
cos

(cos sin )
sin
Re i
Im
X
Y
P P
P P i Pe
P P
P P
P P i j
P P
u
u
u u
u
u
u u
u

= + = = =
` `
) )

= + = =
` `
) )


A strength of Cartesian Re-Im representation using phasors is in taking time derivatives of vectors the
derivative of the exponential is easy ( ( )
s s
d ds e e = ).
2 2
2 2
2
2
2
2 2
2
2
2
2
2 2
2
( )
2
cos 2 sin sin cos
sin 2 cos cos
i
i i
i i i i i
i i i i
d P d Pe
dt dt
d P d
Pe iP e
dt dt
d P
Pe iP e iP e iP e i P e
dt
d P
Pe iP e iP e P e
dt
P P P P d P
dt P P P
u
u u
u u u u u
u u u u
u
u u u u
u u u
u u u u u u u
u u u u u
=
= +
= + + + +
= + +

=
+ +





2
sin Pu u

`
)




5

Where we had to use extensions of Eulers identity

2
2 2
cos sin
cos sin sin cos
sin cos cos sin
i
i
i
e i
ie i i i
i e i i i
u
u
u
u u
u u u u
u u u u
= +
= + = +
= + =



Compare this double-time-derivative with the XY approach.

2 2
2 2
2
2
2 2
2 2
2
2
cos
sin
cos sin
sin cos
cos sin sin sin cos
sin cos cos cos sin
cos 2 sin
P
d P d
P dt dt
P P d P d
dt dt P P
P P P P P d P
dt P P P P P
P P P d P
dt
u
u
u u u
u u u
u u u u u u u u u
u u u u u u u u u
u u u u

=
`
)

=
`
+
)

=
`
+ + +
)

=





2
2
sin cos
sin 2 cos cos sin
P
P P P P
u u u
u u u u u u u

`
+ +
)





We obtain the same result, but the Re-Im phasor time differentiation is made in compact vector notation
along the way.


Above we used the product and chain rules of time differentiation.

product rule
( ) ( )
( ) ( ) ( )
( )
( ( ) ) ( ) ( ) ( )
i t i t
i t i t i t
d dP t de de
P t e e P t P t e P t
dt dt dt dt
u u
u u u
= + = +



chain rule
( ) ( )
( )
( )
( )
( )
i t i t
i t
de de d t
ie t
dt d t dt
u u
u
u
u
u
= =



The result for this example is

( ) ( ) ( )
( ( ) ) ( ) ( ) ( )
i t i t i t
d
P t e P t e P t ie t
dt
u u u
u = +






6

1.6 Matrices

Matrix an m x n array of numbers, where m is the number of rows and n is the number of columns.
| |
11 12 1
21 22 2
1 2
n
n
m m mn
a a a
a a a
A
a a a
(
(
(
=
(
(



Matrices may be used to simplify and standardize the solution of n linear equations in n unknowns
(where m =n). Matrices are used in velocity, acceleration, and dynamics linear equations (matrices are
not used in position analysis which requires a non-linear solution).

Special Matrices
square matrix (m =n = 3)
| |
11 12 13
21 22 23
31 32 33
a a a
A a a a
a a a
(
(
=
(
(



diagonal matrix
| |
11
22
33
0 0
0 0
0 0
a
A a
a
(
(
=
(
(



identity matrix
| |
3
1 0 0
0 1 0
0 0 1
I
(
(
=
(
(



transpose matrix
| |
11 21 31
12 22 32
13 23 33
T
a a a
A a a a
a a a
(
(
=
(
(

(switch rows & columns)

symmetric matrix
| | | |
11 12 13
12 22 23
13 23 33
T
a a a
A A a a a
a a a
(
(
= =
(
(



column vector (3x1 matrix)
{ }
1
2
3
x
X x
x


=
`

)


row vector (1x3 matrix) { } { }
1 2 3
T
X x x x =



7

Matrix Addition add like terms and keep the results in place

a b e f a e b f
c d g h c g d h
+ + ( ( (
+ =
( ( (
+ +




Matrix Multiplication with Scalar multiply each term and keep the results in place

a b ka kb
k
c d kc kd
( (
=
( (




Matrix Multiplication
| | | || | C A B =

In general,
| || | | || | A B B A =

The row and column indices must line up as follows.

| | | || |
( x ) ( x )( x )
C A B
m n m p p n
=



That is, in a matrix multiplication product, the number of columns p in the left-hand matrix must equal
the number of rows p in the right-hand matrix. If this condition is not met, the matrix multiplication is
undefined and cannot be done.

The size of the resulting matrix [C] is from the number of rows m of the left-hand matrix and the
number of columns n of the right-hand matrix, m x n.

Multiplication proceeds by multiplying like terms and adding them, along the rows of the left-
hand matrix and down the columns of the right-hand matrix (use your index fingers from the left and
right hands).

Example
| |
(2x1) (2x3)(3x1)
g
a b c ag bh ci
C h
d e f dg eh fi
i
(
+ + ( (
(
= =
( (
(
+ +

(



Note the inner indices (p =3) must match, as stated above, and the dimension of the result is dictated by
the outer indices, i.e. m x n =2x1.



8

Matrix Multiplication Examples
| |
1 2 3
4 5 6
A
(
=
(


| |
7 8
9 8
7 6
B
(
(
=
(
(




| | | || |
7 8
1 2 3
9 8
4 5 6
7 6
7 18 21 8 16 18 46 42
28 45 42 32 40 36 115 108
C A B =
(
(
(
=
(
(

(

+ + + + ( (
= =
( (
+ + + +


(2x2) (2x3)(3x2)


| | | || |
7 8
1 2 3
9 8
4 5 6
7 6
7 32 14 40 21 48 39 54 69
9 32 18 40 27 48 41 58 75
7 24 14 30 21 36 31 44 57
D B A =
(
(
(
=
(
(

(

+ + + ( (
( (
= + + + =
( (
( ( + + +


(3x3) (3x2)(2x3)



9

Matrix Inversion

Since we cannot divide by a matrix, we multiply by the matrix inverse instead. Given
| | | || | C A B = , solve for [B].

| | | || | C A B =
| | | | | | | || |
| || |
| |
1 1
A C A A B
I B
B

=
=
=

| | | | | |
1
B A C

=

Matrix [A] must be square (m =n) to invert.

| || | | | | | | |
1 1
A A A A I

= =

where [I] is the identity matrix, the matrix 1 (ones on the diagonal and zeros everywhere else). To
calculate the matrix inverse use the following expression.

| |
| | 1
adjoint( ) A
A
A

=

where A is the determinant of [A].


| | | |
adjoint( ) cofactor( )
T
A A =

cofactor(A) ( 1)
i j
ij ij
a M
+
=

minor minor M
ij
is the determinant of the submatrix with row i and
column j removed.

For another example, given
| | | || | C A B = , solve for [A]

| | | || | C A B =
| || | | || || |
| || |
| |
1 1
C B A B B
A I
A

=
=
=

| | | || |
1
A C B

=

In general the order of matrix multiplication and inversion is crucial and cannot be changed.


10

Matrix Determinant
The determinant of a square n x n matrix is a scalar. The matrix determinant is undefined for a
non-square matrix. The determinant of a square matrix A is denoted det(A) or A . The determinant
notation should not be confused with the absolute-value symbol. The MATLAB function for matrix
determinant is det(A).

If a nonhomogeneous system of n linear equations in n unknowns is dependent, the coefficient
matrix A is singular, and the determinant of matrix A is zero. In this case no unique solution exists to
these equations. On the other hand, if the matrix determinant is non-zero, then the matrix is non-
singular, the system of equations is independent, and a unique solution exists.

The formula to calculate a 2 x 2 matrix determinant is straight-forward.

| |
a b
A
c d
(
=
(


A ad bc =

To calculate the determinant of 3 x 3 and larger square matrices, we can expand about any one
row or column, utilizing sub-matrix determinants. Each sub-determinant is formed by crossing out the
current row and its column and retaining the remaining terms as an n1 x n1 square matrix, each of
whose determinant must also be evaluated in the process. The pivot term (the entry in the cross-out row
and column) multiplies the sub-matrix determinants, and there is an alternating +/ / +/ etc. sign
pattern. Here is an explicit example for a 3 x 3 matrix, expanding about the first row (all other options
will yield identical results).
| |
a b c
A d e f
g h k
(
(
=
(
(


( ) ( ) ( )
e f d f d e
A a b c
h k g k g h
a ek hf b dk gf c dh ge
= + +
= + +


For a 3 x 3 matrix only, the determinant can alternatively be calculated as shown, by copying columns 1
and 2 outside the matrix, multiplying the downward diagonals with +signs and multiplying the upward
diagonals with signs (clearly the result is the same as in the above formula).

( ) ( ) ( )
a b c a b
A d e f d e
g h k g h
aek bfg cdh gec hfa kdb a ek hf b kd fg c dh ge
=
= + + = +


A common usage of the 3 x 3 matrix determinant is to calculate the cross product
1 2
P P .


1 2 1 2
1 1 1 1 1 1
1 2 1 1 1 1 2 1 2
2 2 2 2 2 2
2 2 2 1 2 1 2


y z z y
y z x y x z
x y z x z z x
y z x y x z
x y z x y y x
i j k p p p p
p p p p p p
P P p p p i j k p p p p
p p p p p p
p p p p p p p


= = + + =
`

)




11

System of Linear Equations

We can solve n linear equations in n unknowns with the help of a matrix. Below is an example
for n =3.

11 1 12 2 13 3 1
21 1 22 2 23 3 2
31 1 32 2 33 3 3
a x a x a x b
a x a x a x b
a x a x a x b
+ + =
+ + =
+ + =



Where a
ij
are the nine known numerical equation coefficients, x
i
are the three unknowns, and b
i
are the
three known right-hand-side terms. Using matrix multiplication backwards, this is written as
| |{ } { } A x b = .

11 12 13 1 1
21 22 23 2 2
31 32 33 3 3
a a a x b
a a a x b
a a a x b
(

(
=
` `
(

(
) )



where


| |
11 12 13
21 22 23
31 32 33
a a a
A a a a
a a a
(
(
=
(
(

is the matrix of known numerical coefficients


{ }
1
2
3
x
x x
x


=
`

)
is the vector of unknowns to be solved and


{ }
1
2
3
b
b b
b


=
`

)
is the vector of known numerical right-hand-side terms.


There is a unique solution { } | | { }
1
x A b

= only if [A] has full rank. If not, 0 A = (the determinant of


coefficient matrix [A] is zero) and the inverse of matrix [A] is undefined (since it would require dividing
by zero; in this case the rank is not full, it is less than 3, which means not all rows/columns of [A] are
linearly independent). Gaussian Elimination is more robust and more computationally efficient than
matrix inversion to solve the problem
| |{ } { } A x b = for {x}.


12

Matrix Example solve linear equations

Solution of 2x2 coupled linear equations.

1 2
1 2
2 5
6 4 14
x x
x x
+ =
+ =

1
2
1 2 5
6 4 14
x
x
(
=
` `
(
) )



| |
1 2
6 4
A
(
=
(


{ }
1
2
x
x
x

=
`
)
{ }
5
14
b

=
`
)


{ } | | { }
1
x A b

=

( ) ( ) 1 4 2 6 8 A = =
The determinant of [A] is non-zero so there is a unique solution.


| |
1
4 2 1/ 2 1/ 4
1
6 1 3/ 4 1/8
A
A

( (
= =
( (





check | || | | | | | | |
1 1
2
1 0
0 1
A A A A I
(
= = =
(




1
2
1/ 2 1/ 4 5 1
3/ 4 1/8 14 2
x
x
(
= =
` ` `
(

) ) )
answer.


Check this solution by substituting the answer {x} into the original equations
| |{ } { } A x b = and ensuring
the required original {b} results.

1 2 1 1(1) 2(2) 5
6 4 2 6(1) 4(2) 14
+
(
= =
` ` `
(
+
) ) )



13

Same Matrix Examples in MATLAB

%- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
%Mat r i ces. m- mat r i x exampl es
% Dr . Bob, ME 3011
%- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

cl ear ; cl c;

A1 = di ag( [ 1 2 3] ) %3x3 di agonal mat r i x
A2 = eye( 3) %3x3 i dent i t y mat r i x

A3 = [ 1 2; 3 4] ; %mat r i x addi t i on
A4 = [ 5 6; 7 8] ;
Add = A3 + A4

k = 10; %mat r i x- scal ar mul t i pl i cat i on
Mul t Sca = k*A3

Tr ans = A4' %mat r i x t r anspose ( swap r ows and col umns)

A5 = [ 1 2 3; 4 5 6] ; %def i ne t wo mat r i ces
A6 = [ 7 8; 9 8; 7 6] ;
A7 = A5*A6 %mat r i x- mat r i x mul t i pl i cat i on
A8 = A6*A5

A9 = [ 1 2; 6 4] ; %mat r i x f or l i near equat i ons sol ut i on
b = [ 5; 14] ; %def i ne RHS vect or
dA9 = det ( A9) %cal cul at e det er mi nant of A
i nvA9 = i nv( A9) %cal cul at e t he i nver se of A
x = i nvA9*b %sol ve l i near equat i ons
x1 = x( 1) ; %ext r act answer s
x2 = x( 2) ;
Check = A9*x %check answer shoul d be b
xG = A9\ b %Gaussi an el i mi nat i on i s mor e ef f i ci ent

who %di spl ay t he user - cr eat ed var i abl es
whos %user - cr eat ed var i abl es wi t h di mensi ons


The first solution of the linear equations above uses the matrix inverse. To solve linear
equations, Gaussian Elimination is more efficient (more on this in the dynamics notes later) and more
robust numerically; Gaussian elimination implementation is given in the third to the last line of the
above m-file (with the back-slash).

Since the equations are linear, there is a unique solution (assuming the equations are linearly
independent, i.e. the matrix is not near a singularity) and so both solution methods will yield the same
answer.



14

Output of Matrices.m

A1 =
1 0 0
0 2 0
0 0 3

A2 =
1 0 0
0 1 0
0 0 1

Add =
6 8
10 12

Mul t Sca =
10 20
30 40

Tr ans =
5 7
6 8

A7 =
46 42
115 108

A8 =
39 54 69
41 58 75
31 44 57

dA9 = - 8

i nvA9 =
- 0. 5000 0. 2500
0. 7500 - 0. 1250

x = 1
2

Check = 5
14

xG = 1
2




15

2. Kinematics Analysis

2.1 Position Kinematics Analysis

2.1.1 Four-Bar Mechanism Position Analysis

2.1.1.1 Tangent Half-Angle Substitution Derivation and Alternate Solution Method

Tangent half-angle substitution derivation

In this subsection we first derive the tangent half-angle substitution using an
analytical/trigonometric method. Defining parameter t to be

tan
2
t
|
| |
=
|
\ .


i.e. the tangent of half of the unknown angle |, we need to derive cos| and sin| as functions of
parameter t. This derivation requires the trigonometric sum of angles formulae.

cos( ) cos cos sin sin
sin( ) sin cos cos sin
a b a b a b
a b a b a b
=
=



To derive the cos| term as a function of t, we start with

cos cos
2 2
| |
|
| |
= +
|
\ .


The cosine sum of angles formula yields

2 2
cos cos sin
2 2
| |
|
| | | |
=
| |
\ . \ .


Multiplying by a 1, i.e.
2
cos
2
| | |
|
\ .
over itself yields

2 2
2 2 2
2
cos sin
2 2
cos cos 1 tan cos
2 2 2
cos
2
| |
| | |
|
|
| | | |

| |
( | | | | | |
\ . \ .
= =
| | | (
| |
\ . \ . \ .
|
\ .


The cosine squared term can be divided by another 1, i.e.
2 2
cos sin 1
2 2
| | | | | |
+ =
| |
\ . \ .
.



16

2
2
2 2
cos
2
cos 1 tan
2
cos sin
2 2
|
|
|
| |
( | |
|
(
( | |
\ .
( =
| (
| | | |
\ . (
+
| |
(
\ . \ .


Dividing top and bottom by
2
cos
2
| | |
|
\ .
yields

2
2
1
cos 1 tan
2
1 tan
2
|
|
|
(
(
( | |
( =
| (
| |
\ . (
+
|
(
\ .


Remembering the earlier definition for t, this result is the first derivation we need, i.e.

2
2
1
cos
1
t
t
|

=
+




To derive the sin| term as a function of t, we start with

sin sin
2 2
| |
|
| |
= +
|
\ .


The sine sum of angles formula yields

sin sin cos cos sin 2sin cos
2 2 2 2 2 2
| | | | | |
|
| | | | | | | | | | | |
= + =
| | | | | |
\ . \ . \ . \ . \ . \ .


Multiplying top and bottom by cosine yields

2 2
sin
2
sin 2 cos 2tan cos
2 2 2
cos
2
|
| | |
|
|
| |
|
| | | | | |
\ .
= =
| | |
| |
\ . \ . \ .
|
\ .


From the first derivation we learned

2
2
1
cos
2
1 tan
2
|
|
| |
=
|
| |
\ .
+
|
\ .





17

Substituting this term yields

2
1
sin 2tan
2
1 tan
2
|
|
|
(
(
| |
( =
|
| |
\ . (
+
|
(
\ .


Remembering the earlier definition for t, this result is the second derivation we need, i.e.

2
2
sin
1
t
t
| =
+



The tangent half-angle substitution can also be derived using a graphical method as in the figure
below.






18

Alternate solution method

The equation form

cos sin 0 E F G u u + + =

arises often in the position solutions for mechanisms and robots. It appeared in the u
4
solution for the
four-bar mechanism in the ME 3011 NotesBook and was solved using the tangent half-angle
substitution.

Next we present an alternative and simpler solution to this equation. We make two simple
trigonometric substitutions based on the figure below.



Clearly from this figure we have

2 2
cos
E
E F
=
+

2 2
sin
F
E F
=
+



In the original equation we divide by
2 2
E F + and rearrange.

2 2 2 2 2 2
cos sin
E F G
E F E F E F
u u

+ =
+ + +


The two simple trigonometric substitutions yield

2 2
cos cos sin sin
G
E F
u u

+ =
+


Applying the sum-of-angles formula cos( ) cos cos sin sin a b a b a b = yields

2 2
cos( )
G
E F
u

=
+



19


And so the solution for u is

1
1,2
2 2
cos
G
E F
u

(

=
(
+


where

1
tan
F
E


(
=
(



and the quadrant-specific inverse tangent function atan2 must be used in the above expression for .

There are two solutions for u, indicated by the subscripts 1,2, since the inverse cosine function is
double-valued. Both solutions are correct. We expected these two solutions from the tangent-half-angle
substitution approach. They correspond to the open- and crossed-branch solutions (the engineer must
determine which is which) to the four-bar mechanism position analysis problem.

For real solutions for u to exist, we must have

2 2
1 1
G
E F

s s
+
or
2 2
1 1
G
E F
> >
+


If this condition is violated for the four-bar mechanism, this means that the given input angle u
2
is
beyond its reachable limits (see Grashofs Law).




20

2.1.1.3 Four-Bar Mechanism Solution Irregularities

Four-bar mechanism position singularity 0 G E =
4 1 1 2 2
2 2 2 2
1 2 3 4 1 2 1 2
2 ( )
2 cos( )
E r r c r c
G r r r r r r u u
=
= + +


For simplicity, let u
1
=0 (just rotate the entire four-bar mechanism model for zero ground link angle).

2 2 2 2
1 2 3 4 1 4 2 4 1 2
2 2 ( ) 0 G E r r r r r r r r r c = + + + =

I have encountered two example four-bar mechanisms with this 0 G E = singularity.

Case 1
When
1 4
r r = and
2 3
r r = ,
2 2 2 2 2
1 2 2 1 1 2 1 1 2
2 2 ( ) 0 G E r r r r r r r r c = + + + = ALWAYS, regardless
of u
2
.

Example
Given
1 2 3 4
10, 6, 6, 10 r r r r = = = = ; this mechanism is ALWAYS singular. To fix this let
1 2 3 4
10, 5.9999, 6.0001, 10 r r r r = = = = and MATLAB will be able to calculate the position
analysis reliably at every input angle.


Case 2
When
1 3
2 r r = and
4 2
2 r r = , and furthermore
3 2
3 5 r r = ,

2 2 2 2
3 2 3 2 2 3 2 2 3 2
2 2 2 2 2 2
2 2 2 2 2 2 2
2
2 2
4 4 8 4 ( )c
100 25 40 8
4 c
9 9 3 3
8
c
3
G E r r r r r r r r r
r r r r r r
r
= + + +
= + +
=


This 0 G E = occurs only when
2
90 u =

. Case 2 is much less general than case 1.

Example
Given
1 2 3 4
10, 3, 5, 6 r r r r = = = = ; this mechanism is singular when
2
90 u =

. To fix this ignore
2
90 u =

or set your u
2
array to avoid these values.



21

2.1.1.4 Grashofs Law and Four-Bar Mechanism Joint Limits

Grashofs Law

Grashofs Law was presented in the ME 3011 NotesBook to determine the input and output link
rotatability in a four-bar mechanism. Applying Grashofs Law we determine if the input and output
links are a crank (C) or a rocker (R). A crank enjoys full 360 degree rotation while a rocker has a
rotation that is a subset of this full rotation. This section presents more information on Grashofs Law
and then the next subsection presents four-bar mechanism joint limits.

Grashof's condition states "For a four-bar mechanism, the sum of the shortest and longest link
lengths should not be greater than the sum of two remaining link lengths". With a given four-bar
mechanism, the Grashof Condition is satisfied if L S P Q + < + where S and L are the lengths of the
shortest and longest links, and P and Q are the lengths of the other two intermediate-sized links. If the
Grashof condition is satisfied, at least one link will be fully rotatable, i.e. can rotate 360 degrees.

For a four-bar mechanism, the following inequalities must be satisfied to avoid locking of the
mechanism for all motion.

2 1 3 4
4 1 2 3
r r r r
r r r r
+ >
+ >



With reference to the figure below, these inequalities are derived from the fact that the sum of
two sides of a triangle must be greater than the third side, for triangles
4 1 1
O AB and
2 2 2
O A B , respectively.
Note from our standard notation,
1 2 4
r O O ,
2 2
r O A ,
3
r AB , and
4 4
r O B .




A
B
A
O
2
2
O
4
A
1
B
1
B
2


22

Four-Bar Mechanism Joint Limits

If Grashof's Law predicts that the input link is a rocker, there will be rotation limits on the input
link. These joint limits occur when links 3 and 4 are aligned. As shown in the figure below, there will
be two joint limits, symmetric about the ground link.



To calculate the joint limits, we use the law of cosines.

2 2 2
3 4 1 2 1 2 2
2 2 2
1 1 2 3 4
2
1 2
( ) 2 cos
( )
cos
2
L
L
r r r r r r
r r r r
r r
u
u

+ = +
( + +
=
(





with symmetry about r
1
.


23

Joint Limit Example 1 Given
1 2 3 4
10, 6, 8, 7 r r r r = = = =

L S P Q + > + (10 6 8 7 + > + )

so we predict only double rockers from this Non-Grashof Mechanism.

| |
2 2 2
1 1
2
10 6 (8 7)
cos cos 0.742 137.9
2(10)(6)
L
u

( + +
= = =
(




This method can also be used to find angular limits on link 4 when it is a rocker. In this case links 2 and
3 align.

| |
2 2 2
1 1
4
10 7 (6 8)
cos cos 0.336 109.6
2(10)(7)
180 70.4
L
|
u |

( + +
= = =
(

= =




In this example, the allowable input and output angle ranges are:

2
137.9 137.9 u s s


4
70.4 289.6 u s s



This example is shown graphically in the ME 3011 NotesBook, in the Grashofs Law section (2. Non-
Grashof double rocker, first inversion).


Caution
The figure on the previous page does not apply in all joint limit cases. For Grashof
Mechanisms with a rocker input link, one link 2 limit occurs when links 3 and 4 fold upon each other
and the other link 2 limit occurs when links 3 and 4 stretch out in a straight line. See Example 4 (and
Example 3 for a similar situation with the output link 4 limits).



24

Joint Limit Example 2 Given
1 2 3 4
10, 4, 8, 7 r r r r = = = =

L S P Q + < + (10 4 8 7 + < + )

Since the S link is adjacent to the fixed link, we predict this Grashof Mechanism is a crank-rocker.
Therefore, there are no u
2
joint limits.

| |
2 2 2
1 1
2
10 4 (8 7)
cos cos 1.3625
2(10)(4)
L
u

( + +
= =
(



which is undefined, thus confirming there are no u
2
joint limits.


There are limits on link 4 since it is a rocker. For u
4min
, links 2 and 3 are stretched in a straight line
(their absolute angles are identical).

| |
2 2 2
1 1
4min
10 7 (4 8)
cos cos 0.036 88.0
2(10)(7)
180 92.0
|
u |

( + +
= + = =
(

= =





For u
4max
, links 2 and 3 are instead folded upon each other (their absolute angles are different by t).

| |
2 2 2
1 1
4min
10 7 ( 4 8)
cos cos 0.95 18.2
2(10)(7)
180 161.8
|
u |

( + +
= + = =
(

= =





In this example, the output angle range is

4
92.0 161.8 u s s



and u
2
is not limited. This example is shown graphically in the ME 3011 NotesBook, in the Grashofs
Law section (1a. Grashof crank-rocker).



25

Joint Limit Example 3 Given
1 2 3 4
11.18, 3, 8, 7 r r r r = = = = (in) and
1
10.3 u =



L S P Q + < + (11.18 3 8 7 + < + )

This is the four-bar mechanism from Term Example 1 and it is a four-bar crank-rocker Grashof
Mechanism. There are no limits on u
2
since link 2 is a crank.

The u
4
limits are

4
120.1
L
u =

(links 2 and 3 stretched in a line)

4
172.5
L
u =

(links 2 and 3 folded upon each other in a line)


The output angle range is

4
120.1 172.5 u s s



and u
2
is not limited. This example is NOT shown graphically in the ME 3011 NotesBook Grashofs
Law section. However, these u
4
limits are clearly seen in the F.R.O.M. plot for angle u
4
in Term
Example 1 in the ME 3011 NotesBook.



26

Joint Limit Example 4 Given
1 2 3 4
10, 8, 4, 7 r r r r = = = =


L S P Q + < + (10 4 8 7 + < + )

so we predict this Grashof Mechanism is a double-rocker (S opposite fixed link). The u
2
joint limits
are no longer symmetric about the ground link, as was the case in the Non-Grashof Mechanism double
rocker (Example 1). For u
2min
, links 3 and 4 are folded upon each other (their absolute angles are
identical).

| |
2 2 2
1 1
2min
10 8 (7 4)
cos cos 0.969 14.4
2(10)(8)
u

( +
= + = =
(




For u
2max
, links 3 and 4 are instead stretched in a straight line (their absolute angles are different by t as
in Example 1).

| |
2 2 2
1 1
2max
10 8 (7 4)
cos cos 0.269 74.4
2(10)(8)
u

( + +
= + = =
(





u
2min
Diagram u
2max
Diagram

0 1 2 3 4 5 6 7 8 9 10
0
1
2
3
4
5
6
7
8
X (m)
Y

(
m
)
0 1 2 3 4 5 6 7 8 9 10
0
1
2
3
4
5
6
7
8
X (m)
Y

(
m
)


27

This behavior reverses for the u
4
joint limits. For u
4min
, links 2 and 3 are stretched in a straight line
(their absolute angles are identical).

| |
2 2 2
1 1
4min
10 7 (8 4)
cos cos 0.036 88.0
2(10)(7)
180 92.0
|
u |

( + +
= + = =
(

= =




For u
4max
, links 2 and 3 are instead folded upon each other (their absolute angles are different by t).

| |
2 2 2
1 1
4min
10 7 (8 4)
cos cos 0.95 18.2
2(10)(7)
180 161.8
|
u |

( +
= + = =
(

= =





u
4min
Diagram u
4max
Diagram

0 1 2 3 4 5 6 7 8 9 10
0
1
2
3
4
5
6
7
8
X (m)
Y

(
m
)
0 1 2 3 4 5 6 7 8 9 10
0
1
2
3
4
5
6
7
8
X (m)
Y

(
m
)


28


In this plot we can see the minimum and maximum values we just calculated for links 2 and 4.

Note at
2min
14.4 u =

,
3
138.6 u =

and
3
221.4 u =

are the same angle.

Again, this example is NOT shown graphically in the ME 3011 NotesBook Grashofs Law section.
However, a similar case with the same dimensions, in different order, is shown in the ME 3011
NotesBook (
1 2 3 4
7, 10, 4, 8 r r r r = = = = , 1d. Grashof double rocker).


Grashofs Law only predicts the rotatability of the input and output links; it says nothing about
the rotatability of the coupler link 3 in this case, what is the rotatability of the coupler link? (In this
case the coupler link S rotates fully, proving that the relative motion is the same amongst all four-bar
mechanism inversions, though the absolute motion with respect to the possible 4 ground links is very
different.)


For more information, see:

R.L. Williams II and C.F. Reinholtz, 1987, Mechanism Link Rotatability and Limit Position
Analysis Using Polynomial Discriminants, Journal of Mechanisms, Transmissions, and Automation
in Design, Transactions of the ASME, 109(2): 178-182.

10 20 30 40 50 60 70 80
-250
-200
-150
-100
-50
0
50
100
150
200
u
2
(deg)
u

(
d
e
g
)


u
3
1
st
u
3
2
nd
u
4
1
st
u
4
2
nd


29

2.1.2 Slider-Crank Mechanism Position Analysis

Step 6. Solve for the unknowns alternate solution

Here are the same slider-crank mechanism position analysis XY component equations, rearranged
to isolate the u
3
terms.

3 3 2 2
3 3 2 2
r c x r c
r s h r s
=
=


We can square and add to eliminate u
3
, similar to the four-bar mechanism solution approach.

2 2 2 2 2
3 3 2 2 2 2
2 2 2 2 2
3 3 2 2 2 2
2
2
r c x xr c r c
r s h hr s r s
= +
= +


2 2 2 2
3 2 2 2 2 2
2 2 r x h r xr c hr s = + +

This quadratic equation in x has the following form:

2
0 ax bx c + + =
2 2
2 2 2
2 3 2 2
1
2
2
a
b r c
c r r h hr s
=
=
= +


There are two solutions for x, corresponding to the right and left branches.

2 2 2 2
1,2 2 2 3 2 2 2 2
2 x r c r h r s hr s = +

Then u
3
is found from a ratio of the Y to X equations.

1,2
3 2 2 1,2 2 2
atan2( , ) h r s x r c u =


This alternate solution yields identical results as the earlier solution approach in the ME 3011
NotesBook for the right (
1
3 1
, x u ) and left (
2
3 2
, x u ) branches.



30

Slider-Crank Mechanism Snapshot and F.R.O.M. MATLAB m-files

No sample m-files are given in the ME 3011 NotesBook for the slider-crank mechanism since
you can readily adapt the snapshot and F.R.O.M. m-files given for the four-bar mechanism previously.

However, below we include a partial m-file to show how to draw the slider and fixed piston
walls for the slider-crank mechanism graphics, since this was not required for the four-bar mechanism.


Outside the loop:

Lp = put a number here; % length of piston (slider link)
Hp = put a number here; % height of piston
Xp = [-1 -1 1 1]*Lp/2;
Yp = [-1 1 1 -1]*Hp/2;

This establishes the rectangular corner coordinates for the slider link, centered at the origin of your
coordinate frame. It can be done once, outside the loop. Instead of typing numbers for Lp and Hp, I
scale them to a fraction of r
2
, for generality in different-sized slider-crank mechanisms. Note I only
included the four corner points MATLAB patch (below) closes the rectangular figure, i.e. back to
the starting point.


Inside the loop (right after the plot command where links 2 and 3 are drawn to the screen)

patch(Xp+x(i),Yp+h,'g'); % draw piston to screen


where x(i) is the variable horizontal slider displacement and h is the constant vertical offset. These
position parameters shift the piston coordinates from the origin to the correct location in each loop. You
can use any piston color you like (I show green here, 'g').

Further, to draw the horizontal lines representing the piston walls:

Outside the loop

Xpt = [-1000 1000]; % fixed piston walls
Ypt = [h+wall/2 h+wall/2];
Xpb = [-1000 1000];
Ypb = [h-wall/2 h-wall/2];

Inside the loop (right after the plot command where links 2 and 3 are drawn to the screen)

line(Xpt,Ypt,'LineWidth',2); line(Xpb,Ypb,'LineWidth',2);

Set the piston wall width wall to allow a small clearance between the piston and the walls. Again, it
can be scaled to a small fraction of r
2
for generality. The -1000 and 1000 coordinates used above are
to extend the piston wall lines off the screen to the left and to the right.



31

MATLAB subplot feature

In a slider-crank mechanism full-range-of-motion (F.R.O.M.) simulation you will need to plot
both u
3
and x vs. the independent variable u
2
. Since the units of u
3
(deg) and x (m) are dissimilar, they
may not fit clearly on the same plot. In this situation you should use a sub-plot arrangement.

Outside the F.R.O.M. loop you can do the subplot in this way:

subplot(211); % 2x1 arrangement of plots, first plot
plot(th2/DR,th3/DR);
subplot(212); % 2x1 arrangement of plots, second plot
plot(th2/DR,x);

Now, you can use the standard axis labels, linetypes, titles, axis limits, grid, etc., for each plot within a
subplot (repeat these formatting commands after each plot statement above to use similar formatting
for each). These options are not shown, for clarity.


The generalized usage of subplot is shown below.

subplot(mni); % m x n arrangement of plots, i
th
plot
plot( . . . );

As seen in the example syntax above, the integers need not be separated by spaces or commas.
However, I believe they may be so separated if you desire.




32

2.1.3 Inverted Slider-Crank Mechanism Position Analysis

This slider-crank mechanism inversion 2 is an inversion of the standard zero-offset slider-crank
mechanism where the sliding direction is no longer the ground link, but along the rotating link 4.
Ground link length r
1
and input link length r
2
are fixed; r
4
is a variable. The slider link 3 is attached to
the end of link 2 via an R joint and slides relative to link 4 via a P joint. This mechanism converts rotary
input to linear motion and rotary motion output. Practical applications include certain doors/windows
opening/damping mechanisms. The inverted slider-crank is also part of quick-return mechanisms.


Step 1. Draw the Kinematic Diagram




r
1
constant ground link length u
2
variable input angle
r
2
constant input link length u
4
variable output angle
r
4
variable output link length L
4
constant total output link length


Link 1 is the fixed ground link. Without loss of generality we may force the ground link to be
horizontal. If it is not so in the real world, merely rotate the entire inverted slider-crank mechanism so it
is horizontal. Both angles u
2
and u
4
are measured in a right-hand sense from the horizontal to the link.


Step 2. State the Problem

Given r
1
, u
1
=0, r
2
; plus 1-dof position input u
2


Find r
4
and u
4


2
1
4
3


33

Step 3. Draw the Vector Diagram. Define all angles in a positive sense, measured with the right hand
from the right horizontal to the link vector (tail-to-head; your right-hand thumb is located at the vector
tail).



Step 4. Derive the Vector-Loop-Closure Equation. Starting at one point, add vectors tail-to-head until
you reach a second point. Write the VLCE by starting and ending at the same points, but choosing a
different path.

2 1 4
r r r = +


Step 5. Write the XY Components for the Vector-Loop-Closure Equation. Separate the one vector
equation into its two X and Y scalar components.

2 2 1 4 4
2 2 4 4
r c r r c
r s r s
= +
=



Step 6. Solve for the Unknowns from the XY equations. There are two coupled nonlinear equations in
the two unknowns r
4
, u
4
. Unlike the standard slider-crank mechanism, there is no decoupling of X and
Y. However, unlike the four-bar mechanism, there is only one unknown angle so the solution is easier
than the four-bar mechanism. First rewrite the above XY equations to isolate the unknowns on one side.

4 4 2 2 1
4 4 2 2
r c r c r
r s r s
=
=


A ratio of the Y to X equations will cancel r
4
and solve for u
4
.

4 4 2 2
4 4 2 2 1
r s r s
r c r c r
=



4 2 2 2 2 1
atan2( , ) r s r c r u =

Then square and add the XY equations to eliminate u
4
and solve for r
4
.

2 2
4 1 2 1 2 2
2 r r r rr c =+ +

2
1
4


34

Note the same r
4
formula results from the cosine law. Alternatively, the same r
4
can be solved from
either the X or Y equations after is u
4
known.

X)
2 2 1
4
4
r c r
r
c

= Y)
2 2
4
4
r s
r
s
=

Both of these r
4
alternatives are valid; however, each is subject to a different artificial mathematical
singularity (
4
90 u =

and
4
0,180 u =

, respectively), so only the former square-root formula should be
used for r
4
, which has no artificial singularity. The X algorithmic singularity
4
90 u =

never occurs
unless
2 1
r r > , which is to be avoided (see below), but the Y algorithmic singularity occurs twice per full
range of motion.

Technically there are two solution sets the one above and
2 2
4 1 2 1 2 2
2 r r r rr c = + ,
4
u t + .
However, the negative r
4
is not practical and so only the one solution set (branch) exists, unlike most
planar mechanisms with two or more branches.


Full-rotation condition

For the inverted slider-crank mechanism to rotate fully, the fixed length of link 4, L
4
, must be
greater than the maximum value of the variable r
4
.


Slider Limits

The slider reaches its minimum and maximum displacements when u
2
=0 and t, respectively.
Therefore, the slider limits are
1 2 4 1 2
r r r r r s s + . Thus, the fixed length L
4
must be greater than
1 2
r r + .
In addition we require
1 2
r r > for full rotation.



35

Graphical Solution

The Inverted Slider-Crank mechanism position analysis may be solved graphically, by drawing
the mechanism, determining the mechanism closure, and measuring the unknowns. This is an excellent
method to validate your computer results at a given snapshot.

- Draw the known ground link (points O
2
and O
4
separated by r
1
at the fixed angle u
1
=0).
- Draw the given input link length r
2
at the given angle u
2
(this defines point A).
- Draw a line from O
4
to point A.
- Measure the unknown values of r
4
and u
4
.



36

Inverted Slider-Crank Mechanism Position Analysis: Term Example 3
Given:
1
2
4
1
0.20
0.10
0.32
0
r
r
L
u
=
=
=
=

m

Snapshot Analysis (one input angle)
Given this mechanism and
2
70 u =

, calculate u
4
and r
4
.

u
4
(deg) r
4
(m)
150.5 0.191

This Term Example 3 position solution is demonstrated in the figure below.

Term Example 3 Position Snapshot
-0.1 -0.05 0 0.05 0.1 0.15 0.2
-0.1
-0.05
0
0.05
0.1
0.15
0.2
X (m)
Y

(
m
)


37

Full-Range-Of-Motion (F.R.O.M.) Analysis: Term Example 3
A more meaningful result from position analysis is to report the position analysis unknowns for
the entire range of mechanism motion. The subplots below gives r
4
(m) and u
4
(deg), for all
2
0 360 u s s

, for Term Example 3.


Term Example 3 u
4
and r
4



u
4
varies symmetrically about 180

, being 180

at
2
0,180,360 u =

. r
4
varies like a negative cosine
function with minimum displacement
1 2
0.1 r r = at
2
0,360 u =

and maximum displacement
1 2
0.3 r r + =
. Since r
1
is twice r
2
in this example, whenever
2
60,300 u =

, a perfect 30 60 90

triangle is
formed; the relative angle between links 2 and 4 is 90

which corresponds to the max and min of


4
150,210 u =

, respectively. At these special points,
4
( 3 2)0.2 0.173 r = = m.

There is another right triangle that shows up for
2
90 u =

; in these cases
2 2
4
0.2 0.1 0.224 r = + = m and
4
153.4,206.6 u =

, respectively. Check all of these special values in the
F.R.O.M. plot results.


0 50 100 150 200 250 300 350
150
160
170
180
190
200
210
u
4

(
d
e
g
)
0 50 100 150 200 250 300 350
0.1
0.15
0.2
0.25
0.3
u
2
(deg)
r
4

(
m
)


38

2.1.4 Multi-Loop Mechanism Position Analysis

Thus far we have presented position analysis for the single-loop four-bar, slider-crank, and
inverted slider-crank mechanisms. The position analysis for mechanisms of more than one loop is
handled using the same general procedures developed for the single loop mechanisms. A good rule of
thumb is to look for four-bar (or slider-crank) parts of the multi-loop mechanism as we already know
how to solve the complete position analyses for these.

This section presents position analysis for the two-loop Stephenson I six-bar mechanism shown
below as an example multi-loop mechanism. This is one of the five possible six-bar mechanisms shown
in the on-line Mechanisms Atlas.



Stephenson I 6-Bar Mechanism

We immediately see that the bottom loop of the Stephenson I six-bar mechanism is identical to
our standard four-bar mechanism model. Since we number the links the same as in the four-bar, and if
we define the angles identically, the position analysis solution is identical to the four-bar presented
earlier. With the complete position analysis of the bottom loop thus solved, we see that points C and D
can be easily calculated. Then the solution for the top loop is essentially another four-bar solution:
graphically, the circle of radius r
5
about point C must intersect the circle of radius r
6
about point D to
form point E (yielding two possible intersections in general). The analytical solution is very similar to
the standard four-bar position solution, as we will show.

For multi-loop mechanisms, the number of solution branches for position analysis increases
compared to the single-loop mechanisms. Most single-loop mechanisms mathematically have two
solution branches. For multi-loop mechanisms composed of multiple single-loop mechanisms, the
number of solution branches is 2
n
, where n is the number of mechanism loops. For the two-loop
Stephenson I six-bar mechanism, the number of solution branches for the position analysis problem is 4,
two from the standard four-bar part, and two for each of these branches from the upper loop.


39

Now let us solve the position analysis problem for the two-loop Stephenson I six-bar mechanism
using the formal position analysis steps presented earlier. Assume link 2 is the input link.

Step 1. Draw the Kinematic Diagram this is done in the figure above.

r
1
constant ground link length u
1
constant ground link angle
r
2
constant input link length to point A u
2
variable input link angle
r
2a
constant input link length to point C o
2
constant angle on link 2
r
3
constant coupler link length, loop I u
3
variable coupler link angle, loop I
r
4
constant output link length, loop I u
4
variable output link angle, loop I
r
4a
constant input link length to point D o
4
constant angle on link 4
r
5
constant coupler link length, loop II u
5
variable coupler link angle, loop II
r
6
constant output link length, loop II u
6
variable output link angle, loop II

As usual, all angles are measured in a right-hand sense from the absolute horizontal to the link, as shown
in the kinematic diagram.


Step 2. State the Problem

Given r
1
, u
1
, r
2
, r
3
, r
4
, r
2a
, r
4a
, r
5
, r
6
, o
2
, o
4
; plus 1-dof position input u
2


Find u
3
, u
4
, u
5
, u
6



Step 3. Draw the Vector Diagram. Define all angles in a positive sense, measured with the right hand
from the right horizontal to the link vector (tail-to-head with the right-hand thumb at the vector tail and
right-hand fingers towards the arrow in the vector diagram below).




40

Step 4. Derive the Vector-Loop-Closure Equations. One VLCE is required for each mechanism loop.
Start at one point, add vectors tail-to-head until you reach a second point. Write each vector equation by
starting and ending at the same points, but choosing a different path.

2 3 1 4
2 5 1 4 6 a a
r r r r
r r r r r
+ = +
+ = + +



Note an alternative to the second vector loop equation is
2 5 3 4 6 b b
r r r r r + = + + . See if you can identify
2b
r and
4b
r , plus their angles u
2b
and u
4b
.


Step 5. Write the XY Components for each Vector-Loop-Closure Equation. Separate the two vector
equations into four XY scalar component equations.

2 2 3 3 1 1 4 4
2 2 3 3 1 1 4 4
r c r c rc r c
r s r s r s r s
+ = +
+ = +


2 2 5 5 1 1 4 4 6 6
2 2 5 5 1 1 4 4 6 6
a a a a
a a a a
r c r c rc r c r c
r s r s r s r s r s
+ = + +
+ = + +


where
2 2 2
4 4 4
a
a
u u o
u u o
= +
=



Step 6. Solve for the Unknowns from the four XY Equations. The four coupled nonlinear equations in
the four unknowns u
3
, u
4
, u
5
, u
6
can be solved in two stages, one for each mechanism loop.

Loop I. This solution is identical to the standard four-bar mechanism solution for u
3
, u
4
,
summarized here from earlier. From the first two XY scalar equations above, isolate u
3
terms, square
and add both equations to obtain one equation in one unknown u
4
. This equation has the form
4 4
cos sin 0 E F G u u + + = , where terms E, F, and G are known functions of constants and the input angle
u
2
. Solve this equation for two possible values of u
4
using the tangent half-angle substitution. The two
u
4
values correspond to the open and crossed branches. Then return to the original XY scalar equations
with u
3
terms isolated, divide the Y by the X equations, and solve for u
3
using the atan2 function,
substituting the solved values for u
4
. One unique u
3
will result for each of the two possible u
4
values.

Loop II. The method is analogous to the Loop I solution above. Since u
4
is now known, we also
know
4 4 4 a
u u o = . From the second two XY scalar equations above, isolate u
5
terms, square and add
both equations to obtain one equation in one unknown u
6
. This equation is of the form
2 6 2 6 2
cos sin 0 E F G u u + + = , where terms E
2
, F
2
, and G
2
are known functions of constants and the known
angles
2 2 2 a
u u o = + and
4 4 4 a
u u o = . Solve this equation for two possible values of u
6
using the tangent


41

half-angle substitution. The two u
6
values correspond to open and crossed branches, for each of the
Loop I open and crossed branches. Then return to the original XY scalar equations with u
5
terms
isolated, divide the Y by the X equations, and solve for u
5
using the atan2 function, substituting the
solved values for u
6
. One unique u
5
will result for each of the two possible u
6
values from each u
4
.

Branches. There are two u
5
, u
6
branches for each of the two u
3
, u
4
branches, so there are four
overall mechanism branches for the two-loop Stephenson I six-bar mechanism.


Full-rotation condition

The range of motion of a multi-loop mechanism may be more limited than that of single-loop
mechanisms. One can perform a compound Grashof analysis when four-bars are the component
mechanisms. For the two-loop Stephenson I six-bar mechanism, the second loop may constrain the first
loop (e.g. it may change an expected crank motion of the input link to a rocker). This is an important
issue in design of multi-loop mechanisms if the input link must still rotate fully.


Graphical Solution

The two-loop Stephenson I six-bar mechanism position analysis may readily be solved
graphically, by drawing the mechanism, determining the mechanism closure, and measuring the
unknowns. This is an excellent method to validate your computer results at a given snapshot.

Loop I (this part is identical to the standard four-bar graphical solution)
- Draw the known ground link (points O
2
and O
4
separated by r
1
at the fixed angle u
1
).
- Draw the given input link length r
2
at the given angle u
2
to yield point A.
- Draw a circle of radius r
3
, centered at point A.
- Draw a circle of radius r
4
, centered at point O
4
.
- These circles intersect in general in two places to yield two possible points B.
- Connect the two branches and measure the unknown angles u
3
and u
4
.

Loop II (this part is a modification of the standard four-bar graphical solution). Start with the end of the
procedure above, on the same drawing. For each solution branch above, perform the following steps.
- Draw the link r
2a
at angle
2 2 2 a
u u o = + from point O
2
.
- Draw the link r
4a
at angle
4 4 4 a
u u o = from point O
4
.
- Draw a circle of radius r
5
, centered at point C.
- Draw a circle of radius r
6
, centered at point D.
- These circles intersect in general in two places to yield two possible points E.
- Connect the two branches and measure the unknown values u
5
and u
6
.

In general, there are four overall position solution branches.




42

2.2 Velocity Kinematics Analysis

2.2.2 Three-Part Velocity Formula Moving Example

Given initial positions
{ } { }
0 0
0 0 0 0
x y
P P L u = (m, rad) and constant velocities
{ } { }
0 0
0.2 0.1 0.3 0.4
x y
V V V e = (m/s, rad/s), simulate this motion and determine
P
V at each
instant. t
f
=5 and 0.1 t A = sec was used. The initial and final snapshots, with their three-part velocity
diagrams, are shown below.


Initial Kinematic Diagram Initial Three-Part Velocity Diagram



Final Kinematic Diagram Final Three-Part Velocity Diagram
0 0.5 1 1.5 2 2.5 3 3.5 4
0
0.5
1
1.5
2
2.5
3
3.5
4
X (m)
Y

(
m
)
-0.5 0 0.5
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
V
x
(m/s)
V
y

(
m
/
s
)
0 0.5 1 1.5 2 2.5 3 3.5 4
0
0.5
1
1.5
2
2.5
3
3.5
4
X (m)
Y

(
m
)
-0.5 0 0.5
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
V
x
(m/s)
V
y

(
m
/
s
)
V
0
V e x L
V
P


43

Three-Part Velocity Moving Example Plots



Position Terms Velocity Terms

What is the relationship between these plots?




Point P Translational Velocity Sliding and Tangential Velocity Components


Constant velocity terms
{ } { }
0 0
0.2 0.1 0.3 0.4
x y
V V V e = lead to non-constant point P velocity
(due to the nonlinear position kinematics).


0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
time (sec)
P
o
s
it
io
n

T
e
r
m
s

(
m

a
n
d

r
a
d
)

P
0x
P
0y
L
u
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
time (sec)
V
e
lo
c
it
y

T
e
r
m
s

(
m
/
s

a
n
d

r
a
d
/
s
)

V
0x
V
0y
V
e
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
-0.5
-0.4
-0.3
-0.2
-0.1
0
0.1
0.2
0.3
0.4
0.5
time (sec)
P
o
in
t

P

V
e
lo
c
it
y

(
m
/
s
)

V
Px
V
Py
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
-0.5
-0.4
-0.3
-0.2
-0.1
0
0.1
0.2
0.3
0.4
0.5
time (sec)
S
lid
in
g

a
n
d

T
a
n
g
e
n
t
ia
l
V
e
lo
c
it
ie
s

(
m
/
s
)

V
sx
V
sy
V
tx
V
ty


44

2.2.5 Inverted Slider-Crank Mechanism Velocity Analysis

Again, link 2 (the crank) is the input and link 4 is the output. Remember r
4
is a variable so
4
0 r = in this problem.

Step 1. The inverted slider-crank mechanism Position Analysis must first be complete.

Given r
1
, u
1
=0, r
2
, and u
2
, we solved for r
4
and u
4



Step 2. Draw the inverted slider-crank mechanism Velocity Diagram.



where
i
e (i =2,4) is the absolute angular velocity of link i.
4
r is the slider velocity along link 4.
3 4
e e = since the slider cannot rotate relative to link 4.


Step 3. State the Problem

Given the mechanism r
1
, u
1
=0, r
2

the position analysis u
2
, r
4
, u
4

1-dof of velocity input e
2


Find the velocity unknowns
4
r and e
4


2
1
4
3


45

Step 4. Derive the velocity equations. Take the first time derivative of the vector loop closure
equations from position analysis, in XY component form.

Here are the inverted slider-crank mechanism position equations from earlier.
2 1 4
2 2 1 4 4
2 2 4 4
r r r
r c r r c
r s r s
= +
= +
=


The first time derivative of the position equations is given below.
2 2 2 4 4 4 4 4
2 2 2 4 4 4 4 4
r s r c r s
r c r s r c
e e
e e
=
= +



These two linear equations in two unknowns can be written in matrix form.
4 4 4 4 2 2 2
4 4 4 4 2 2 2
c r s r r s
s r c r c
e
e e
(
=
` `
(
) )




Step 5. Solve the velocity equations for the unknowns
4 4
, r e .

2 2 2 4 2 4
4 4 4 4 4 2 2 2
2 2 2 4 2 4
4 4 4 2 2 2 4
4
2 2 4 2
4
2 2 4 2
4
4
( )
1
( )
sin( )
cos( )
r s c c s
r r c r s r s
r s s c c
s c r c r
r
r
r
r
r
e
e
e
e e
e u u
e u u
e
+
(

= = +
` ` `
(

) )

)



=
` `
)

)




where we have used the trigonometric identities:
( )
( )
cos cos cos sin sin
sin sin cos cos sin
a b a b a b
a b a b a b
=
=


sin( ) sin( )
cos( ) cos( )
a a
a a
=
=


The units are all correct in the solution above, m/s and rad/s, respectively.



46

Inverted slider-crank mechanism singularity condition
When does the solution fail? This is an inverted slider-crank mechanism singularity, when the
determinant of the coefficient matrix goes to zero. The result is dividing by zero, resulting in infinite
4 4
, r e .

| |
4 4 4
4 4 4
c r s
A
s r c
(
=
(



2 2 2 2
4 4 4 4 4 4 4 4
( ) ( ) 0 A r c r s r c s r = = + = =

Physically, assuming
1 2
r r > as in the full rotation condition from the inverted slider-crank mechanism
position analysis, this is impossible, i.e. r
4
never goes to zero.
This matrix determinant
4
A r = was used in the solution of the previous page.




Inverted slider-crank mechanism velocity example Term Example 3 continued

Given
1
2
4
1
0.20
0.10
0.32
0
r
r
L
u
=
=
=
=

(m)
2
4
4
70
150.5
0.191 r
u
u
=
=
=

(deg and m)

Snapshot Analysis
Given this mechanism position analysis plus
2
25 e = rad/s, calculate
4 4
, r e for this snapshot.
4
4
0.870 0.094 2.349
0.493 0.166 0.855
r
e
(
=
` `
(

) )



4
4
2.47
2.18
r
e

=
` `
) )

(m/s and rad/s)



Both are positive, so the slider link 3 is currently traveling up link 4 and the link 4 is currently rotating in
the ccw direction, which makes sense from the physical problem.



47

Full-Range-Of-Motion (F.R.O.M.) Analysis Term Example 3 continued

A more meaningful result from velocity analysis is to report the velocity analysis results for the
entire range of mechanism motion. The subplot below gives e
4
(top, rad/s) and
4
r (bottom, m/s), for all
2
0 360 u s s

, for Term Example 3. Since e
2
is constant, we can plot the velocity results vs. u
2
(since it
is related to time t via
2 2
t u e = ).



Term Example 3 F.R.O.M., e
4
and
4
r

As expected, e
4
is zero at the max and min for u
4
(at
2
60 u =

); also, e
4
has a large range of nearly
constant positive velocity near the middle of motion this can be seen in a MATLAB animation.
4
r is
zero at the beginning, middle, and end of motion and is max at
2
60 u =

.


0 50 100 150 200 250 300 350
-25
-20
-15
-10
-5
0
5
10
e
4

(
r
a
d
/
s
)
0 50 100 150 200 250 300 350
-2
-1
0
1
2
u
2
(deg)
r
4
d

(
m
/
s
)


48

2.2.6 Multi-Loop Mechanism Velocity Analysis

Thus far we have presented velocity analysis for the single-loop four-bar, slider-crank, and
inverted slider-crank mechanisms. The velocity analysis for mechanisms of more than one loop is
handled using the same general procedures developed for the single loop mechanisms.

This section presents velocity analysis for the two-loop Stephenson I six-bar mechanism shown
below as an example multi-loop mechanism. It follows the position analysis for the same mechanism
presented earlier.



Stephenson I 6-Bar Mechanism

The bottom loop of the Stephenson I six-bar mechanism is identical to the standard four-bar
mechanism model and so the velocity analysis solution is identical to the four-bar presented earlier.
With the complete velocity analysis of the bottom loop thus solved, the solution for the top loop is
essentially another four-bar velocity solution.

As in all velocity analysis, the velocity solution for multi-loop mechanisms is a linear analysis
yielding a unique solution (assuming the given mechanism position is not singular) for each position
solution branch considered. The position analysis must be complete prior to the velocity solution.

Now let us solve the velocity analysis problem for the two-loop Stephenson I six-bar mechanism
using the formal velocity analysis steps presented earlier. Again, assume link 2 is the input link.


49

Step 1. The Stephenson I six-bar mechanism Position Analysis must first be complete.
Given r
1
, u
1
, r
2
, r
3
, r
4
, r
2a
, r
4a
, r
5
, r
6
, o
2
, o
4
, and u
2
we solved for u
3
, u
4
, u
5
, u
6
.


Step 2. Draw the Stephenson I six-bar mechanism Velocity Diagram.
This should include all the information from the position diagram, plus the new velocity
information. For clarity, we show only the new velocity information here. Refer to the previous
Stephenson I position kinematics diagram for complete information.



where
i
e (i =2,3,4,5,6) is the absolute angular velocity of link i. Triangular links 2 and 4 each have a
single angular velocity for the whole link. 0
i
r = for all links since all links are of fixed length (no
sliders).


Step 3. State the Problem

Given the mechanism r
1
, u
1
, r
2
, r
3
, r
4
, r
2a
, r
4a
, r
5
, r
6
, o
2
, o
4
,

the position analysis u
2
, u
3
, u
4
, u
5
, u
6
,

1-dof velocity input e
2

Find the velocity unknowns e
3
, e
4
, e
5
, e
6





50

Step 4. Derive the velocity equations. Take the first time derivative of each of the two vector loop
closure equations from position analysis, in XY component form.

Here are the Stephenson I six-bar mechanism position equations.

Vector equations
2 3 1 4
2 5 1 4 6 a a
r r r r
r r r r r
+ = +
+ = + +



XY scalar equations
2 2 3 3 1 1 4 4
2 2 3 3 1 1 4 4
r c r c rc r c
r s r s r s r s
+ = +
+ = +


2 2 5 5 1 1 4 4 6 6
2 2 5 5 1 1 4 4 6 6
a a a a
a a a a
r c r c rc r c r c
r s r s r s r s r s
+ = + +
+ = + +


where
2 2 2
4 4 4
a
a
u u o
u u o
= +
=



The first time derivatives of the Loop I position equations are identical to those for the standard
four-bar mechanism.
2 2 2 3 3 3 4 4 4
2 2 2 3 3 3 4 4 4
r s r s r s
r c r c r c
e e e
e e e
=
+ =


These equations can be written in matrix form.
3 3 4 4 3 2 2 2
3 3 4 4 4 2 2 2
r s r s r s
r c r c r c
e e
e e
(
=
` `
(

) )


The first time derivative of the Loop II position equations is
2 2 2 5 5 5 4 4 4 6 6 6
2 2 2 5 5 5 4 4 4 6 6 6
a a a a a a
a a a a a a
r s r s r s r s
r c r c r c r c
e e e e
e e e e
=
+ = +


These equations can be written in matrix form.
5 5 6 6 5 2 2 2 4 4 4
5 5 6 6 6 2 2 2 4 4 4
a a a a
a a a a
r s r s r s r s
r c r c r c r c
e e e
e e e
+ (
=
` `
(

) )


where we have used
2 2
4 4
a
a
e e
e e
=
=
since o
2
and o
4
are constant angles.


51

Step 5. Solve the velocity equations for the unknowns e
3
, e
4
, e
5
, e
6
.

The two mechanism loops decouple so we find e
3
and e
4
from Loop I first and then use e
4
to
find e
5
and e
6
from Loop II. The solutions are given below.

Loop I (identical to the standard four-bar mechanism)

1
3 3 4 4 3 2 2 2
3 3 4 4 4 2 2 2
r s r s r s
r c r c r c
e e
e e

(
=
` `
(

) )



Loop II (similar to the standard four-bar mechanism)

1
5 5 5 6 6 2 2 2 4 4 4
6 5 5 6 6 2 2 2 4 4 4
a a a a
a a a a
r s r s r s r s
r c r c r c r c
e e e
e e e

+ (
=
` `
(

) )



Remember, Gaussian elimination is more efficient and robust than the matrix inverse. Also, these
equations may be solved algebraically instead of using matrix methods to yield the same answers.



Stephenson I six-bar mechanism singularity condition

The velocity solution fails when the determinant of either coefficient matrix above goes to zero.
The result is dividing by zero, resulting in infinite angular velocities for the associated loop.

For the first loop, the singularity condition is identical to the singularity condition of the standard
four-bar mechanism, i.e. when links 3 and 4 either line up or fold upon each other, causing a link 2 joint
limit. For the second loop, the singularity condition is similar, occurring when links 5 and 6 either line
up or fold upon each other. These conditions also cause angle limit problems for the position analysis,
so the velocity singularities are known problems.




52

2.3 Acceleration Kinematics Analysis

2.3.2 Five-Part Acceleration Formula Moving Example

Given initial positions
{ } { }
0 0
0 0 0 0
x y
P P L u = (m, rad), initial velocities
{ } { }
0 0
0 0 0 0
x y
V V V e = (m/s, rad/s) and constant accelerations
{ }
0 0 x y
A A A o =
{ } 0.2 0.05 0.15 0.1 (m/s
2
, rad/s
2
), simulate this motion and determine
P
A at each instant. t
f
=5
and 0.1 t A = sec was used; the initial and final snapshots, with their five-part acceleration diagrams, are
shown below.


Initial Kinematic Diagram Initial Five-Part Acceleration Diagram


Final Kinematic Diagram Final Five-Part Acceleration Diagram

0 0.5 1 1.5 2 2.5 3 3.5 4
0
0.5
1
1.5
2
2.5
3
3.5
4
X (m)
Y

(
m
)
-0.8 -0.2 0.4
0
0.2
0.4
0.6
0.8
1
1.2
A (m/s
2
)
A
y

(
m
/
s
2
)
0 0.5 1 1.5 2 2.5 3 3.5 4
0
0.5
1
1.5
2
2.5
3
3.5
4
X (m)
Y

(
m
)
-0.8 -0.2 0.4
0
0.2
0.4
0.6
0.8
1
1.2
A
x
(m/s
2
)
A
y

(
m
/
s
2
)
A
0
A
2e x V
o x L
e x (e x L)
A
P


53

Five-Part Acceleration Moving Example Plots



Position Terms Velocity Terms




Acceleration Terms Point P Translational Acceleration


What is the relationship amongst the first three plots?


Constant acceleration terms
{ } { }
0 0
0.2 0.05 0.15 0.1
x y
A A A o = lead to non-constant point P
acceleration (due to nonlinear position kinematics and centripetal acceleration).



0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
0
0.5
1
1.5
2
2.5
time (sec)
P
o
s
it
io
n

T
e
r
m
s

(
m

a
n
d

r
a
d
)


P
0x
P
0y
L
u
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
time (sec)
V
e
lo
c
it
y

T
e
r
m
s

(
m
/
s

a
n
d

r
a
d
/
s
)


V
0x
V
0y
V
e
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
0
0.05
0.1
0.15
0.2
0.25
time (sec)
A
c
c
e
le
r
a
t
io
n

T
e
r
m
s

(
m
/
s
2

a
n
d

r
a
d
/
s
2
)


A
0x
A
0y
A
o
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
time (sec)
P
o
in
t

P

A
c
c
e
le
r
a
t
io
n

(
m
/
s
2
)


A
Px
A
Py


54

2.3.4 Slider-Crank Mechanism Acceleration Analysis

Derivative/Integral Relationships

When one variable is the derivative of another, recall the relationships from calculus. For example:
( )
( )
dx t
x t
dt
=
0
( ) ( ) x t x x t dt = +
}

( )
( )
dx t
x t
dt
=


0
( ) ( ) x t x x t dt = +
}


Term Example 2 F.R.O.M. Slider Results, x, x ,

x

The value of x at any point is the slope of the x curve at that point. The value of x at any point is
the integral of the x curve up to that point (the value of x at any point is the area under the x curve up
to that point plus the initial value x
0
). A similar relationship exists for x and x .
These graphs are plotted vs. u
2
, but the same type of relationships hold when plotted vs. time t
since e
2
is constant. This is the Term Example 2 F.R.O.M. result.
Note these curves should be plotted vs. time t instead of u
2
in order to see the true slope and area
values accurately.

0 50 100 150 200 250 300 350
0.1
0.2
x

(
m
)
0 50 100 150 200 250 300 350
-2
-1
0
1
2
3
x
d

(
m
/
s
)
0 50 100 150 200 250 300 350
-50
0
50
u
2
(deg)
x
d
d

(
m
/
s
2
)


55

2.3.5 Inverted Slider-Crank Mechanism Acceleration Analysis

Again, link 2 (the crank) is the input and link 4 is the output.

Step 1. The inverted slider-crank mechanism Position and Velocity Analyses must first be complete.

Given r
1
, u
1
=0, r
2
, and u
2
we solved for r
4
and u
4
; then given e
2
we solved for
4
r and e
4
.


Step 2. Draw the inverted slider-crank mechanism Acceleration Diagram.



where
i
o (i =2,4) is the absolute angular acceleration of link i.
4
r is the slider acceleration along link 4.
3 4
o o = since the slider cannot rotate relative to link 4.


Step 3. State the Problem

Given the mechanism r
1
, u
1
=0, r
2


the position analysis u
2
, r
4
, u
4


the velocity analysis e
2
,
4
r , e
4


1-dof acceleration input o
2



Find the acceleration unknowns
4
r and o
4


2
1
4
3


56

Step 4. Derive the acceleration equations. Take the first time derivative of the inverted slider-crank
mechanism velocity equations from velocity analysis, in XY component form.

Here are the inverted slider-crank mechanism velocity equations.

4 4 4 4 4 2 2 2
4 4 4 4 4 2 2 2
r c r s r s
r s r c r c
e e
e e
=
+ =



The first time derivative of the velocity equations is given below.
2 2
4 4 4 4 4 4 4 4 4 4 4 2 2 2 2 2 2
2 2
4 4 4 4 4 4 4 4 4 4 4 2 2 2 2 2 2
2
2
r c r s r s r c r s r c
r s r c r c r s r c r s
e o e o e
e o e o e
=
+ + =




These equations can be written in matrix form.
2 2
4 4 4 4 2 2 2 2 2 2 4 4 4 4 4 4
2 2
4 4 4 4 2 2 2 2 2 2 4 4 4 4 4 4
2
2
c r s r r s r c r s r c
s r c r c r s r c r s
o e e e
o o e e e
+ + (
=
` `
(
+
) )




Step 5. Solve the acceleration equations for the unknowns
4 4
, r o .

2 2
4 4 4 4 4 2 2 2 2 2 2 4 4 4 4 4 4
2 2
4 4 4 4 2 2 2 2 2 2 4 4 4 4 4 4
2 2
4 2 2 4 2 2 2 4 2 4 4
2
4
2 2 4 2 2 2 4 2 4 4
4
2 1
2
sin( ) cos( )
cos( ) sin( ) 2
r r c r s r s r c r s r c
s c r r c r s r c r s
r r r r
r r r
r
o e e e
o o e e e
o u u e u u e
o
o u u e u u e
+ + (
=
` `
(
+
) )

+


=
`

+
)


`



)


where we have again used the trigonometric identities:

( )
( )
cos cos cos sin sin
sin sin cos cos sin
a b a b a b
a b a b a b
=
=


sin( ) sin( )
cos( ) cos( )
a a
a a
=
=



A major amount of algebra and trigonometry is required to get the final analytical solution for
4 4
, r o above. Interestingly, the link 4 Coriolis term
4 4
2r e cancelled in
4
r while the link 4 centripetal
term
2
4 4
r e cancelled in o
4
. The units are all correct, m/s
2
and rad/s
2
, respectively.




57

Inverted slider-crank mechanism singularity condition

The acceleration problem has the same coefficient matrix [A] as the velocity problem, so the
singularity condition is identical (see the singularity discussion in the inverted slider-crank mechanism
velocity section the only singularity is when r
4
goes to zero; this will never to occur if
1 2
r r > ).



Inverted slider-crank mechanism acceleration example Term Example 3 continued

Given
1
2
4
1
0.20
0.10
0.32
0
r
r
L
u
=
=
=
=

(m)
2
4
4
70
150.5
0.191 r
u
u
=
=
=

(deg and m)
2
4
4
25
2.47
2.18
r
e
e
=
=
=
(rad/s and m/s)



Snapshot Analysis

Given this mechanism position and velocity analyses, plus
2
0 o = rad/s
2
, calculate
4 4
, r o for this
snapshot.

4
4
0.870 0.094 16.873
0.493 0.166 48.957
r
o
(
=
` `
(

) )



4
4
9.46
267.14
r
o

=
` `
) )

(m/s
2
and rad/s
2
)


4
r is negative, so the slider link 3 is currently slowing down its positive velocity up link 4 and o
4
is
positive, so the link 4 angular velocity is currently increasing in the ccw direction.


58

Full-Range-Of-Motion (F.R.O.M.) Analysis Term Example 3 continued

A more meaningful result from acceleration analysis is to report the acceleration analysis results
for the entire range of mechanism motion. The subplot below gives o
4
(top, rad/s
2
) and
4
r (bottom,
m/s
2
), for all
2
0 360 u s s

, for Term Example 3. Again, since e
2
is constant, we can plot the
acceleration results vs. u
2
(since it is related to time t via
2 2
t u e = ).



Term Example 3 F.R.O.M., o
4
and
4
r

As expected, o
4
is zero at the beginning, middle, and end since the e
4
curve flattens out at those
points. The maximum (and minimum) o
4
values correspond to the greatest slopes for e
4
.
4
r is
maximum (and minimum) at the beginning, middle, and end since the
4
r curve is steepest at those
points;
4
r is zero when the
4
r curve is flat, i.e.
2
60 u =

.
0 50 100 150 200 250 300 350
-1000
-500
0
500
1000
o
4

(
r
a
d
/
s
2
)
0 50 100 150 200 250 300 350
-50
0
50
100
u
2
(deg)
r
4
d
d

(
m
/
s
2
)


59

Derivative/Integral Relationships

When one variable is the derivative of another, recall the relationships from calculus (the
derivative is the slope of the above curve at each point; the integral is the area under the curve up to that
point, taking into account the initial value). For example:

4
4
( )
( )
d t
t
dt
u
e =
4 40 4
( ) ( ) t t dt u u e = +
}


4
4
( )
( )
d t
t
dt
e
o =
4 40 4
( ) ( ) t t dt e e o = +
}



0 50 100 150 200 250 300 350
160
180
200
u
4

(
d
e
g
)
0 50 100 150 200 250 300 350
-20
-10
0
10
e
4

(
r
a
d
/
s
)
0 50 100 150 200 250 300 350
-1000
-500
0
500
1000
u
2
(deg)
o
4

(
r
a
d
/
s
2
)


60

4
4
( )
( )
dr t
r t
dt
=

4 40 4
( ) ( ) r t r r t dt = +
}


4
4
( )
( )
dr t
r t
dt
=


4 40 4
( ) ( ) r t r r t dt = +
}




These plots are all from Term Example 3.


0 50 100 150 200 250 300 350
0.1
0.15
0.2
0.25
r
4

(
m
)
0 50 100 150 200 250 300 350
-2
-1
0
1
2
r
4
d

(
m
/
s
)
0 50 100 150 200 250 300 350
-50
0
50
100
u
2
(deg)
r
4
d
d

(
m
/
s
2
)


61

2.3.6 Multi-Loop Mechanism Acceleration Analysis

Thus far we have presented acceleration analysis for the single-loop four-bar, slider-crank, and
inverted slider-crank mechanisms. The acceleration analysis for mechanisms of more than one loop is
handled using the same general procedures developed for the single loop mechanisms.

This section presents acceleration analysis for the two-loop Stephenson I six-bar mechanism
shown below as an example multi-loop mechanism. It follows the position and velocity analyses for the
same mechanism presented earlier.



Stephenson I 6-Bar Mechanism

The bottom loop of the Stephenson I six-bar mechanism is identical to the standard four-bar
mechanism model and so the acceleration analysis solution is identical to the four-bar presented earlier.
With the complete acceleration analysis of the bottom loop thus solved, the solution for the top loop is
essentially another four-bar acceleration solution.

As in all acceleration analysis, the acceleration solution for multi-loop mechanisms is a linear
analysis yielding a unique solution (assuming the given mechanism position is not singular) for each
solution branch considered. The position and velocity analyses must be complete prior to the
acceleration solution.


62

Now let us solve the acceleration analysis problem for the two-loop Stephenson I six-bar
mechanism using the formal acceleration analysis steps presented earlier. Again, assume link 2 is the
input link.


Step 1. The Stephenson I six-bar mechanism Position and Velocity Analyses must first be complete.
Given r
1
, u
1
, r
2
, r
3
, r
4
, r
2a
, r
4a
, r
5
, r
6
, o
2
, o
4
, u
2
, and e
2
, we solved for u
3
, u
4
, u
5
, u
6
, e
3
, e
4
, e
5
, e
6
.



Step 2. Draw the Stephenson I six-bar mechanism Acceleration Diagram.
This should include all the information from the position and velocity diagrams, plus the new
acceleration information. For clarity, we show only the new acceleration information here. Refer to the
previous Stephenson I position and velocity kinematics diagrams for complete information.



where
i
o (i =2,3,4,5,6) is the absolute angular acceleration of link i. Triangular links 2 and 4 each have
a single angular acceleration for the whole link. 0
i
r = for all links since all links are of fixed length (no
sliders).


Step 3. State the Problem

Given the mechanism r
1
, u
1
, r
2
, r
3
, r
4
, r
2a
, r
4a
, r
5
, r
6
, o
2
, o
4
,
the position analysis u
2
, u
3
, u
4
, u
5
, u
6
,
the velocity analysis e
2
, e
3
, e
4
, e
5
, e
6
,
1-dof acceleration input o
2


Find the acceleration unknowns o
3
, o
4
, o
5
, o
6





63

Step 4. Derive the acceleration equations. Take the first time derivative of both sides of each of the
four scalar XY velocity equations.

The Stephenson I six-bar mechanism velocity equations are given below.

XY scalar velocity equations
2 2 2 3 3 3 4 4 4
2 2 2 3 3 3 4 4 4
r s r s r s
r c r c r c
e e e
e e e
=
+ =


2 2 2 5 5 5 4 4 4 6 6 6
2 2 2 5 5 5 4 4 4 6 6 6
a a a a a a
a a a a a a
r s r s r s r s
r c r c r c r c
e e e e
e e e e
=
+ = +


where
2 2 2
4 4 4
a
a
u u o
u u o
= +
=



The first time derivative of the Loop I velocity equations is identical to that for the standard four-
bar mechanism.

2 2 2
2 2 2 2 2 2 3 3 3 3 3 3 4 4 4 4 4 4
2 2 2
2 2 2 2 2 2 3 3 3 3 3 3 4 4 4 4 4 4
r s r c r s r c r s r c
r c r s r c r s r c r s
o e o e o e
o e o e o e
=
+ =


These equations can be written in matrix form.

2 2 2
3 3 4 4 3 2 2 2 2 2 2 3 3 3 4 4 4
2 2 2
3 3 4 4 4 2 2 2 2 2 2 3 3 3 4 4 4
r s r s r s r c r c r c
r c r c r c r s r s r s
o o e e e
o o e e e
+ (
=
` `
(
+
) )


The first time derivatives of the Loop II velocity equations are:

2 2 2 2
2 2 2 2 2 2 5 5 5 5 5 5 4 4 4 4 4 4 6 6 6 6 6 6
2 2 2 2
2 2 2 2 2 2 5 5 5 5 5 5 4 4 4 4 4 4 6 6 6 6 6 6
a a a a a a a a a a a a
a a a a a a a a a a a a
r s r c r s r c r s r c r s r c
r c r s r c r s r c r s r c r s
o e o e o e o e
o e o e o e o e
=
+ = +


These equations can be written in matrix form.

2 2 2 2
5 5 6 6 5 2 2 2 2 2 2 5 5 5 4 4 4 4 4 4 6 6 6
2 2 2 2
5 5 6 6 6 2 2 2 2 2 2 5 5 5 4 4 4 4 4 4 6 6 6
a a a a a a a a
a a a a a a a a
r s r s r s r c r c r s r c r c
r c r c r c r s r s r c r s r s
o o e e o e e
o o e e o e e
+ + + (
=
` `
(
+ +
) )


where we have used
2 2
4 4
a
a
e e
e e
=
=
and
2 2
4 4
a
a
o o
o o
=
=
since o
2
and o
4
are constant angles.



64

Step 5. Solve the acceleration equations for the unknowns o
3
, o
4
, o
5
, o
6
.

The two loops decouple so we find o
3
and o
4
from Loop I first and then use o
4
to find o
5
and o
6

from Loop II. The solutions are given below.

Loop I (identical to the standard four-bar mechanism)

1
2 2 2
3 3 4 4 3 2 2 2 2 2 2 3 3 3 4 4 4
2 2 2
3 3 4 4 4 2 2 2 2 2 2 3 3 3 4 4 4
r s r s r s r c r c r c
r c r c r c r s r s r s
o o e e e
o o e e e

+ (
=
` `
(
+
) )



Loop II (similar to the standard four-bar mechanism)

1
2 2 2 2
5 5 5 6 6 2 2 2 2 2 2 5 5 5 4 4 4 4 4 4 6 6 6
2 2 2 2
6 5 5 6 6 2 2 2 2 2 2 5 5 5 4 4 4 4 4 4 6 6 6
a a a a a a a a
a a a a a a a a
r s r s r s r c r c r s r c r c
r c r c r c r s r s r c r s r s
o o e e o e e
o o e e o e e

+ + + (
=
` `
(
+ +
) )



Remember, Gaussian elimination is more efficient and robust than the matrix inverse. Also, these
equations may easily be solved algebraically instead of using matrix methods.


Stephenson I six-bar mechanism singularity condition

The acceleration solution fails when the determinant of either coefficient matrix above goes to
zero. The result is dividing by zero, resulting in infinite angular accelerations for the associated loop.
Note the two coefficients matrices in the acceleration solutions are identical to those for the velocity
solutions. Therefore, the acceleration singularity conditions are identical to the velocity singularity
conditions.

For the first loop, the singularity condition is identical to the singularity condition of the standard
four-bar mechanism, i.e. when links 3 and 4 either line up or fold upon each other, causing a link 2 joint
limit. For the second loop, the singularity condition is similar, occurring when 5 and 6 either line up or
fold upon each other. These conditions also cause problems for the velocity and position analyses, so
the acceleration singularities are known problems.




65

2.4 Other Kinematics Topics

2.4.1 Link Extensions Graphics
Using the methods presented thus far, we can use MATLAB to animate mechanisms for the
entire range of motion. However, these methods have focused on the basic models shown below. What
if your term project requires animation of modifications of these basic mechanisms with extensions from
the existing rigid links?

Four-Bar Mechanism Offset Slider-Crank Mechanism

Four-bar mechanism link 3 extensions


Here are the kinematics equations (we already presented the point C kinematics equations in the ME
3011 NotesBook).
3 3
3 3
cos( )
sin( )
x x CA C
y y CA C
c a r
c a r
u o
u o
+ +
= =
` `
+ +
) )
C
3 3
3 3
cos( )
sin( )
x x DB D
y y DB D
d b r
d b r
u o
u o
+ +
= =
` `
+ +
) )
D

where
2 2
2 2
cos
sin
x
y
a r
a r
u
u

= =
` `
) )
A

1 1 4 4
1 1 4 4
cos cos
sin sin
x
y
b r r
b r r
u u
u u
+
= =
` `
+
) )
B

In this simple straight-line case, use
3
180
C
o =

and
3
0
D
o =

. Here is partial MATLAB code for link 3
extensions animation.
x2 = [0 ax(i)]; % coordinates of link 2
y2 = [0 ay(i)];
x3 = [cx(i) dx(i)]; % coordinates of link 3
y3 = [cy(i) dy(i)];
x4 = [r1x bx(i)]; % coordinates of link 4
y4 = [r1y by(i)];
figure; plot(x2,y2,'r',x3,y3,'g',x4,y4,'b');
C
D
A
B


66

Four-bar mechanism link 4 extensions


The kinematics equations are given below.

4 4
4 4
cos( )
sin( )
x x CB C
y y CB C
c b r
c b r
u o
u o
+ +
= =
` `
+ +
) )
C
4
4
4 4 4
4 4 4
cos( )
sin( )
x DO D
x
y y DO D
O r
d
d O r
u o
u o
+ +


= =
` `
+ +
)
)
D

In this simple straight-line case, use
4
0
C
o =

and
4
180
D
o =

. b
x
and b
y
were given above and
4
O is:
1 1
1 1
cos
sin
r
r
u
u

=
`
)
4
O

Here is partial MATLAB code for link 4 extensions animation.
x2 = [0 ax(i)]; % coordinates of link 2
y2 = [0 ay(i)];
x3 = [ax(i) bx(i)]; % coordinates of link 3
y3 = [ay(i) by(i)];
x4 = [cx(i) dx(i)]; % coordinates of link 4
y4 = [cy(i) dy(i)];
figure;
plot(x2,y2,'r',x3,y3,'g',x4,y4,'b');


Of course, one can use combinations of these graphics approaches as necessary. Also, use
patch for drawing solid polygonal links rather than straight lines. You can also use patch for
drawing solid circles. This method is also extendable to non-straight-line links by using the appropriate
o angles.

C
D
A
B
O
4


67

2.5 Jerk Kinematics Analysis

Jerk is the time rate of change of the acceleration (and hence the second and third time rates of
change of the velocity and position, respectively). Again, this jerk time rate of change may describe a
change in magnitude of acceleration, a change in direction of acceleration, or both. What names have
been given to the next three position derivatives after jerk? The answer is given somewhere in this
section. Jerk analysis is the fourth step in kinematics analysis. It is not required for standard Newton-
Euler dynamics analysis. However, it is useful for the following items.

- Input link motion specification
- Cam motion profiles and cam design
- Smooth motion control as in elevators

J erk can be important for kinematic motion analysis in general. Position, velocity, and
acceleration analyses must be completed first. J erk is the first time derivative of the acceleration, the
second time derivative of the velocity, and the third time derivative of the position. Like all the terms
preceding it, jerk is also a vector quantity. There are both translational and rotational jerk terms.


2 3
2 3
( ) ( ) ( )
( )
d A t d V t d P t
J t
dt dt dt
= = = SI units:
3
m
s



2 3
2 3
( ) ( ) ( )
( )
d t d t d t
t
dt dt dt
o e u
| = = = SI units:
3
rad
s



2.5.1 Jerk Analysis Introduction

In this section we will derive the n-part jerk formula, showing the most general jerk terms
possible for planar devices.

n-part Jerk Derivation Figure


This is a four-dof system consisting of a translating/rotating rigid rod with a slider. The same
system was used for the two-part position, three-part velocity, and five-part accelerations formula
derivations in the ME 3011 NotesBook. Find the total jerk of the point P, which is on the slider.
Y
L
X
u
L
,

L
,

L
P
O
P
O
e, o, |


68

Start with the five-part acceleration formula from before and take another time derivate of the XY
components. What is n? (Hint clearly n must be greater than 5.)

2
2
cos 2 sin sin cos
sin 2 cos cos sin
OX P
P
OY
A A V L L d A d
J
dt dt A A V L L
u e u o u e u
u e u o u e u
+
= =
`
+ + +
)


Recall the two-part position, three-part velocity, and five-part acceleration formula results below.

( )
2
2
cos
sin
cos sin
sin cos
2
cos 2 sin sin cos
sin 2 cos cos sin
P O
OX
OY
P O
OX
OY
P O
OX
OY
P P L
P L
P L
V V V L
V V L
V V L
A A A V L L
A A V L L
A A V L L
u
u
e
u e u
u e u
e o e e
u e u o u e u
u e u o u e u
= +
+
=
`
+
)
= + +
+
=
`
+ +
)
= + + + +
+
=
`
+ + +
)



The angle u, angular velocity e, angular acceleration o, and angular jerk | are all changing with respect
to time (only the planar case is this simple; the spatial rotation case is more complicated).

2 3
2 3
d d d
dt dt dt
o e u
| = = =


The rod length L, sliding velocity V, sliding acceleration A, and sliding jerk J are all changing with
respect to time.

2 3
2 3
d A d V d L
J
dt dt dt
= = =


Here are the same relationships, using the dot notation to indicate time differentiation.

2 3
2 3
d d d
dt dt dt
u u u
u = = =


2 3
2 3
d L d L d L
L
dt dt dt
= = =





69

Product and Chain Rules of Differentiation

Again, well need to use the product and chain rules repeatedly in jerk analysis derivations.

Product rule


( ) ( )
( ( ) ( )) ( ) ( )
d dx t dy t
x t y t y t x t
dt dt dt
= +

x, y are both functions of time


Chain rule


( ( )) ( )
( ( ( )))
d df x t dx t
f x t
dt dx dt
=

f is a function of x, which is an implicit function of t


Examples


2
( 2 sin ) 2 sin 2 sin 2 (sin )
(sin )
2 sin 2 sin 2
2 sin 2 sin 2 cos
d d
V A V V
dt dt
d d
A V V
d dt
A V V
e u e u o u e u
u u
e u o u e
u
e u o u e u
=
=
=


2 2 2 2
2
2 2
2 3
( cos ) cos ( )cos (cos )
( ) (cos )
cos cos
cos 2 cos sin
d d d
L V L L
dt dt dt
d d d d
V L L
d dt d dt
V L L
e u e u e u e u
e e u u
e u u e
e u
e u eo u e u
=
=
= +



Many terms will combine (like in the Coriolis acceleration case). Check the resulting units of all
components to check your results. The full n-part jerk derivation is left to the interested student.



70

2.5.2 Mechanism Jerk Analysis

Generic Mechanism Jerk Analysis Problem Statement

Given the mechanism, complete position, velocity, and acceleration analyses, and one-dof of jerk
input, calculate the jerk unknowns.

For a given branch of a known mechanism, this will yield linear equations so a matrix-vector
approach may be used to obtain the unique solution (assuming no singularity).

Since jerk kinematics is not required for dynamics analysis or machine design (generally), this
solution is beyond the scope of the class.

For our jerk needs, we may use the time differentiation approach of the previous subsection.




Snap, what a happy sound
Snap is the happiest sound I found
You may clap, rap, tap, slap,
But Snap makes the world go round
Snap, crackle, pop Rice Krispies!

I say its Crackle, the crispy sound
You gotta have Crackle or the clocks not wound
Geese cackle, feathers tickle, belts buckle, beets pickle,
But Crackle makes the world go round
Snap, crackle, pop Rice Krispies!

I insist that Pops the sound
The best is missed unless Pops around
You cant stop hoppin when the cereals poppin
Pop makes the world go round
Snap, crackle, pop Rice Krispies!

-Old Kelloggs Advertisement






71

2.6 Branch Symmetry in Kinematics Analysis

We have been doing F.R.O.M. analysis for the open-branch only (four-bar mechanism) and the
right-branch only (slider-crank mechanism). What do the kinematics results look like for the crossed
and left branches? Are there any relationships amongst the various analyses for the two branches? The
reader is left to draw their own conclusions.

2.6.1 Four-Bar Mechanism
Given:
1
2
3
4
10
4
8
7
r
r
r
r
=
=
=
=

1
3
2
0
4
0
10 (constant)
CA
r
u
o
e
=
=
=
=


Open Branch Crossed Branch






72

Open Branch Crossed Branch













73

2.6.2 Slider-Crank Mechanism
Given:
2 3 3 2
2, 6, 0, 3, 0, 10 (constant)
CA
r r h r o e = = = = = =

Right Branch Left Branch











74

We see a great deal of symmetry for both four-bar and slider crank mechanism kinematic
analyses. The coupler point curves and all plots have either horizontal midpoint
2
180 u =

or vertical
zero point flip-symmetry (some have both).

However, for the x and x slider-crank mechanism results, this symmetry is not immediately
evident. This special symmetry is revealed when first cutting the plots at
2
180 u =

, and then performing
the horizontal and/or vertical flipping.




75

2.7 Kinematics Analysis Examples

This section presents complete snapshot and full-range-of-motion (F.R.O.M.) examples to
demonstrate the rotational and translational position, velocity, and acceleration kinematics analyses in
the ME 3011 NotesBook. Term Example 1 is for the four-bar mechanism and Term Example 2 is for
the slider-crank mechanism.

2.7.1 Term Example 1: Four-Bar Mechanism

Four-Bar Mechanism Position Analysis Term Example 1

Given:
1
2
3
4
11.18
3
8
7
r
r
r
r
=
=
=
=
in
1
2
3
4
0.284
0.076
0.203
0.178
r
r
r
r
=
=
=
=
m

and
1
10.3 u =

(Ground link is 11" over and 2" up). Also given 5
CA
r = (in) and
3
36.9 o =

for the
coupler link point of interest.


Snapshot Analysis (for one given input angle u
2
)

Given this mechanism and
2
30 u =

, calculate
3 4
, , u u , and
C
P for both branches. Results:

0.076
0.005
0.036
E
F
G
=
=
=



branch


t


u
3



u
4






C
P


open


1.79


53.8




121.7




67.9




(0.06, 0.16)


crossed


1.57


313.0




245.1




67.9




(0.19, 0.02)



These two branch solutions are demonstrated in the figures below. The length units are m. Note
is identical for both branches due to the conventions presented earlier.



76


Open Branch


Crossed Branch

Term Example 1 Snapshot

-0.1 0 0.1 0.2 0.3
-0.1
-0.05
0
0.05
0.1
0.15
0.2
0.25
X (m)
Y

(
m
)
-0.1 0 0.1 0.2 0.3
-0.1
-0.05
0
0.05
0.1
0.15
0.2
0.25
X (m)
Y

(
m
)


77

Full-Range-Of-Motion (F.R.O.M.) Analysis Term Example 1

A more meaningful result from position analysis is to solve and plot the position analysis
unknowns for the entire range of mechanism motion.

The plot below gives u
3
(red), u
4
(green), and (blue), all deg, for all
2
0 360 u s s

, for Term
Example 1, the open branch only.



u
3
, u
4
, and



0 50 100 150 200 250 300 350
20
40
60
80
100
120
140
160
180
u
2
(deg)
u
i

(
d
e
g
)


u
3
u
4



78

The plot below gives the initial (and final) animation position, for
2
0,360 u =

. It also gives the
coupler curve for the open branch, plotting P
CY
vs. P
CX
in green circles.



C
P

Coupler Curve

Term Example 1 F.R.O.M. Position Results

-0.1 -0.05 0 0.05 0.1 0.15 0.2 0.25 0.3
-0.1
-0.05
0
0.05
0.1
0.15
0.2
0.25
X (m)
Y

(
m
)


79


Four-bar mechanism velocity example Term Example 1 continued

Given r
1
=0.284, r
2
=0.076, r
3
=0.203, r
4
=0.178, r
CA
=0.127 m, and
1
10.3 u =

,
2
30 u =

,
3
53.8 u =

,
4
121.7 u =

,
3
36.9 o =

. This is the open branch of the Term Example 1 four-bar mechanism.


Snapshot Analysis

Given this mechanism position analysis plus
2
20 e = rad/s (positive, which indicates ccw),
calculate
3 4
, e e , and
C
V for this instant in motion (snapshot).

3
4
0.164 0.151 0.76
0.120 0.093 1.32
e
e
(
=
` `
(

) )

3
4
8.09
3.73
e
e

=
` `

) )



Both e
3
and e
4
are negative, so they are in the cw direction for this snapshot. These results are the
absolute angular velocities of links 3 and 4 with respect to the ground link. The associated coupler point
translational velocity vector for this snapshot is:

0.27
1.33
C
V

=
`
)
(m/s)





80

Full-Range-Of-Motion (F.R.O.M.) Analysis: Term Example 1 continued

A more meaningful result from velocity analysis is to solve and plot the velocity analysis
unknowns for the entire range of mechanism motion.

The plot below gives e
3
(red) and e
4
(green) (rad/s) for all
2
0 360 u s s

, for Term Example 1,
for the open branch only. For all of Term Example 1, assume the e
2
given above is constant. Since e
2

is constant, we can plot the velocity results vs. u
2
(since it is related to time t via
2 2
t u e = ).



e
3
and e
4


0 50 100 150 200 250 300 350
-10
-8
-6
-4
-2
0
2
4
6
8
10
u
2
(deg)
e
i

(
r
a
d
/
s
)


e
3
e
4


81

The plot below gives the absolute translational coupler point C velocity for all
2
0 360 u s s

, for
Term Example 1, for the open branch only.



C
V

Term Example 1 F.R.O.M. Velocity Results


0 50 100 150 200 250 300 350
-1.5
-1
-0.5
0
0.5
1
1.5
2
u
2
(deg)
V
C

(
m
/
s
)


V
CX
V
CY


82

Four-bar mechanism acceleration example Term Example 1 continued

Given r
1
=0.284, r
2
=0.076, r
3
=0.203, r
4
=0.178, r
CA
=0.127 m, and
1
10.3 u =

,
2
30 u =

,
3
53.8 u =

,
4
121.7 u =

,
3
36.9 o =

;
2
20 e = (constant),
3
8.09 e = ,
4
3.73 e = rad/s. This is the open
branch of the position and velocity example (Term Example 1).


Snapshot Analysis

Given this mechanism position and velocity analysis, plus
2
0 o = rad/s
2
, calculate
3 4
, o o for this
instant in motion (snapshot).

3
4
3
4
0.164 0.151 35.56
0.120 0.093 23.87
8.65
244.40
o
o
o
o
(
=
` `
(

) )

=
` `
) )



Both are positive, so they are ccw in direction. These results are the absolute angular accelerations of
links 3 and 4 with respect to the ground link. The coupler point translational acceleration vector is:

27.40
23.57
C
A

=
`

)
m/s
2



83

Full-Range-Of-Motion (F.R.O.M.) Analysis Term Example 1 continued

A more meaningful result from acceleration analysis is to solve and plot the acceleration analysis
unknowns for the entire range of mechanism motion.

The plot below gives o
3
(red) and o
4
(green), (rad/s
2
), for all
2
0 360 u s s

, for Term Example
1, open branch. In the Term Example 1 velocity section it was assumed that the given e
2
is constant,
which means that the given o
2
is always zero. Since e
2
is constant, we can plot the acceleration results
vs. u
2
(since it is related to time t via
2 2
t u e = ).



o
3
and o
4


0 50 100 150 200 250 300 350
-250
-200
-150
-100
-50
0
50
100
150
200
250
u
2
(deg)
o
i

(
r
a
d
/
s
2
)


o
3
o
4


84

The plot below gives the translational coupler point acceleration for all
2
0 360 u s s

, for Term
Example 1, the open branch only.



C
A

Term Example 1 F.R.O.M. Acceleration Results

0 50 100 150 200 250 300 350
-30
-20
-10
0
10
20
30
u
2
(deg)
A
C

(
m
/
s
2
)


A
CX
A
CY


85

Derivative/Integral Relationships

When one variable is the derivative of another, recall the relationships from calculus. e.g.:
4
4
4
4
( )
( )
( )
( )
d t
t
dt
d t
t
dt
u
e
e
o
=
=

4 40 4
4 40 4
( ) ( )
( ) ( )
t t dt
t t dt
u u e
e e o
= +
= +
}
}




The value of e
4
at any point is the slope of the u
4
curve at that point. The value of u
4
at any point
is the integral of the e
4
curve up to that point (the value of u
4
at any point is the area under the e
4
curve
up to that point plus the initial value u
40
). A similar relationship exists for o
4
and e
4
.
These graphs are plotted vs. u
2
, but the same type of relationships hold when plotted vs. time t
since e
2
is constant. This is the Term Example 1 F.R.O.M. result, but u
4
was changed from deg to rad
for better comparison.
These curves should be plotted vs. time t instead of u
2
in order to see the true slope and area
values accurately.

0 50 100 150 200 250 300 350
2
2.5
3
u
4

(
r
a
d
)
0 50 100 150 200 250 300 350
-10
-5
0
5
10
e
4

(
r
a
d
/
s
)
0 50 100 150 200 250 300 350
-200
-100
0
100
200
300
o
4

(
r
a
d
/
s
2
)
u
2
(deg)


86

2.7.2 Term Example 2: Slider-Crank Mechanism

Slider-Crank Mechanism Position Analysis Term Example 2

Given:
2
3
4
8
3
r
r
h
=
=
=
in
2
3
0.102
0.203
0.076
r
r
h
=
=
=
m


Snapshot Analysis (one input angle)

Given this mechanism and
2
30 u =

, calculate x and u
3
for both branches. Results:

branch x (m) u
3
(deg)
right 0.290 7.2
left 0.114 172.8


These two branch solutions are demonstrated in the figures below (length units are m).



Right Branch Left Branch

Term Example 2 Snapshot

-0.1 0 0.1 0.2 0.3
-0.2
-0.1
0
0.1
0.2
X (m)
Y

(
m
)
-0.1 0 0.1 0.2 0.3
-0.2
-0.1
0
0.1
0.2
X (m)
Y

(
m
)


87

Full-Range-Of-Motion (F.R.O.M.) Analysis Term Example 2

A more meaningful result from position analysis is to solve and plot the position analysis
unknowns for the entire range of mechanism motion.

The top plot gives u
3
(deg) and the bottom plot gives x (m), for all
2
0 360 u s s

, for Term
Example 2, the right branch only.




u
3
and x


0 50 100 150 200 250 300 350
-10
0
10
20
30
40
50
60
70
u
3

(
d
e
g
)
0 50 100 150 200 250 300 350
0.05
0.1
0.15
0.2
0.25
0.3
u
2
(deg)
x

(
m
)


88

The plot below gives the initial (and final) animation position, for
2
0,360 u =

. It also gives the
coupler curve to scale for the right branch, plotting P
CY
vs. P
CX
in green. In this case the coupler point is
taken to be the midpoint of coupler link 3.



C
P

Coupler Curve

Term Example 2 F.R.O.M. Position Results

For Term Example 2, the slider translation limits are 0.067 0.295 x s s , as seen in the x plot
above, calculated from the equations in the 3011 NotesBook..


-0.05 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
-0.2
-0.15
-0.1
-0.05
0
0.05
0.1
0.15
0.2
X (m)
Y

(
m
)


89

Slider-crank mechanism velocity example Term Example 2 continued

Given r
2
=0.102, r
3
=0.203, h =0.076 m, and
2
30 u =

,
3
7.2 u =

, x =0.290 m. This is the right
branch of the slider-crank position example of Term Example 2.

Snapshot Analysis (one input angle)

Given this mechanism position analysis plus
2
15 e = rad/s (+so ccw), calculate
3
, x e for this
instant in time (snapshot).

3
3
0.025 1 0.762
0.202 0 1.320
6.55
0.60
x
x
e
e
(
=
` `
(

) )

=
` `

) )




These results are the absolute rotational and translational velocities of links 3 and 4 with respect to the
fixed ground link. Both are negative, so the coupler link 3 is currently rotating in the cw direction and
the slider 4 is currently moving to the left.



90

Full-Range-Of-Motion (F.R.O.M.) Analysis Term Example 2 continued

A more meaningful result from velocity analysis is to solve and plot the velocity analysis
unknowns for the entire range of mechanism motion. The subplot arrangement below gives e
3
(top,
rad/s) and x (bottom, m/s), for all
2
0 360 u s s

, for Term Example 2, the right branch only. For all of
Term Example 2, assume the e
2
given above is constant. Since e
2
is constant, we can plot the velocity
results vs. u
2
(since u
2
changes linearly, as it is related to time t via
2 2
t u e = ).



Term Example 2 F.R.O.M. Velocity Results, e
3
and x


0 50 100 150 200 250 300 350
-10
-5
0
5
10
e
3

(
r
a
d
/
s
)
0 50 100 150 200 250 300 350
-2
-1
0
1
2
3
u
2
(deg)
x
d

(
m
/
s
)


91

Slider-crank mechanism acceleration example Term Example 2 continued

Given r
2
=0.102, r
3
=0.203, h =0.076 m, and
2
30 u =

,
3
7.2 u =

, x =0.290 m; and
2
15 e = ,
3
6.55 e = rad/s, 0.60 x = m/s. This is the right branch of the position and velocity example for the
slider-crank mechanism of Term Example 2.


Snapshot Analysis (one input angle)

Given this mechanism position and velocity analysis plus
2
0 o = rad/s
2
, calculate
3
, x o for this
snapshot in time.

3
3
0.025 1 28.438
0.202 0 12.519
62.09
30.02
x
x
o
o
(
=
` `
(

) )

=
` `

) )




These results are the absolute angular and linear accelerations of links 3 and 4 with respect to the fixed
ground link.


92

Full-Range-Of-Motion (F.R.O.M.) Analysis Term Example 2 continued

A more meaningful result from acceleration analysis is to solve and plot the acceleration analysis
unknowns for the entire range of slider-crank mechanism motion. The top plot gives o
3
(rad/s
2
) and the
bottom plot gives x (m/s
2
), for all
2
0 360 u s s

, for Term Example 2, for the right branch only. In the
Term Example 2 velocity section it was assumed that the given e
2
is constant, which means that the
given o
2
is always zero. Since e
2
is constant, we can plot the velocity results vs. u
2
(since it is related to
time t via
2 2
t u e = ).



Term Example 2 F.R.O.M. Acceleration Results, o
3
and x


0 50 100 150 200 250 300 350
-250
-200
-150
-100
-50
0
50
100
150
o
3

(
r
a
d
/
s
2
)
0 50 100 150 200 250 300 350
-50
0
50
u
2
(deg)
x
d
d

(
m
/
s
2
)


93

3. Dynamics Analysis

3.1 Dynamics Introduction

DAlemberts Principle

We can convert dynamics problems into statics problems by the inclusion of a vector inertial
force
0 G
F mA = and a vector inertial moment
0 G
M I o = . Centrifugal force
2
mre , directed away
from the center of rotation, is an example of an inertial force vector. Its not really a force but a felt
effect of an inertia in acceleration. Using DAlemberts principle, the right-hand side of the translational
and rotational dynamics equations is subtracted to the other side of the equation. Then the forces and
moments balance to zero as in statics, when the inertial forces are included in the FBD.

We wont use this method, but it is mentioned for completeness. We would instead to prefer to
consider statics problems as a subset of dynamics problems, with zero accelerations.

0
0
G
O
R mA
R F
=
+ =



0
0
G
O
T r R I
T r R M
o + =
+ + =





Dynamics Humor from xkcd


xkcd.com




94

3.2 Mass, Center of Gravity, and Mass Moment of Inertia
This section presents a thorough review of mass, center of gravity, and mass moment of inertia,
for use in the translational and rotational dynamics equations for each FBD.
G
F mA =


GZ
G
M I o =


Newtons Second Law Eulers Rotational Dynamics Equation

Newtons Second Law requires m and CG, Eulers Rotational Dynamics Equation requires CG and I
GZ
.

m P
CG
I
GZ
translational mass center of gravity
rotational center of gravity mass moment of inertia

Mass
In Newtons Second Law
G
F mA =

, the mass m is the proportionality constant. Mass is


measure of translational inertia a resistance to change in motion according to Newtons First Law.
Mass is also a measure of the storage of translational kinetic energy
2
1
2
T
KE mv = and the units are kg.

Examples for m, CG, I
G


System of particles General rigid body

Rectangular rigid body

Mass calculation
System of particles General rigid body
1
N
i
i
m m
=
=


body
m dm =
}


Rectangular rigid body
using
m
V
= , dm dV = , so
body body
m dm dV = =
} }

/2 /2
/2 /2
b h
b h
m tdxdy tbh

= =
} }
m V = (an obvious result)


95

Center of Gravity (CG, G)
The CG is the point at which a body is balanced with respect to gravity. It is also the point at
which the body weight acts. The CG is also called the mass center, center of mass, and centroid. It is a
vector quantity and the units are length units, m.

CG calculation
System of particles

i i
CG
i
m r
P
m
=

Cartesian components
i i
CG
i
i i
CG
i
m x
X X
m
m y
Y Y
m
= =
= =



General rigid body

body
CG
rdm
P
dm
=
}
}
Cartesian components
x
CG
y
CG
xdm
X X
dm
ydm
Y Y
dm
= =
= =
}
}
}
}

Rectangular rigid body
Using an XY coordinate frame at the geometric center, the CG is calculated below.

/2
/ 2
/2
/ 2
/2
2
/2
2 2
2
0
2 4 4
x
x
b
b
b
b
b
b
xdm
X
dm
xdV
m
xthdx
m
th
xdx
m
th x
m
th b b
m

=
=
=
=
=
| |
= =
|
\ .
}
}
}
}
}

/2
/2
/2
/2
/2
2
/2
2 2
2
0
2 4 4
y
y
h
h
h
h
h
h
ydm
Y
dm
ydV
m
ytbdy
m
tb
ydy
m
tb y
m
tb h h
m

=
=
=
=
=
| |
= =
|
\ .
}
}
}
}
}


0
0
CG
X
P
Y

= =
` `
) )


For a homogeneous, regular geometric body of uniform thickness, the CG is the geometric center.


96

Mass Moment of Inertia I
G

This is not the same as area moment of inertia (I
A
) for beam bending, which is recalled below.

2
Ax
y
I y dA =
}

2
Ay
x
I x dA =
}
units:
4
A
I m

In Eulers rotational dynamics equation
GZ
G
M I o =

, the mass moment of inertia I


GZ
is the
proportionality constant. I
GZ
is also a measure of rotational inertia, i.e. the resistance to change in
rotational motion according to Newtons First Law. Also, it is a measure of how hard it is to accelerate
about certain axes in rotation. I
GZ
is also a measure of the storage of rotational kinetic energy,
2
1
2
R GZ
KE I e = , and its units are kg-m
2
.

What is the mass moment of inertia, a scalar, vector, matrix, or something else? Answer it is a tensor.


Mass Moment of Inertia I
G
calculation

System of particles
2
axis i i
I m r =



where r
i
is the scalar perpendicular distance from the axis to the i
th
particle. With squaring, all terms
will be positive, so there can be no canceling like for the CG. If the first moment (CG) is balanced, the
second moment (I
GZ
) terms do not cancel since the squared terms are all positive.

General rigid body inertia tensor (symmetric)

XX XY XZ
axis XY YY YZ
XZ YZ ZZ
I I I
I I I I
I I I
(
(
=
(
(



2 2
( )
XX
body
I y z dm = +
}

2 2
( )
YY
body
I x z dm = +
}

2 2
( )
ZZ
body
I x y dm = +
}


What is the only term that matters for XY planar motion? Answer I
ZZ
.


In the yardstick example:
GZ GY GX
I I I > > also
OZ GZ
I I >




97

Rectangular rigid body

Using an XY coordinate frame at the CG, I
GZ
is calculated below.

/2 /2
2 2 2 2
/2 /2
/2
3
/2
2
/2
/2
3 3
/2
2
/2
/2
3 3 3
/2
2
/2
/2
3 3
( ) ( )
3
1
2 2 3 8 8
12 3 12
3 8
b h
GZ
b h
body
h
b
b
h
b
b
b
b
b
b
I x y dm x y tdxdy
y
t x y dx
h h h h
t x dx
h hx h x
t hx dx t
h b b
t

= + = +
| |
| = +
|
\ .
| |
| | | |
= +
|
| |
\ .
\ .
\ .
| |
| |
| = + = +
|
|
\ .
\ .

=
} } }
}
}
}
3
3 3
2 2
8 12 2 2
( )
12 12 12
h b b
b h bh tbh
t b h

| |
| | | |
+
|
| |
\ .
\ .
\ .
| |
= + = +
|
\ .



2 2
( )
12
GZ
m
I b h = + (using m V tbh = = )
units: mass times distance squared, kg-m
2


This formula agrees with the result given in tables.


How do we find mass, center of gravity, and mass moment of inertia in the real world?

- From tables for example, see the three tables at the end of this section.

- CAD packages (such as SolidEdge and AutoCAD) calculate m, CG, and I
GZ
automatically for
each link drawn, once material is associated is associated with the 3D model.


98

Parallel Axis Theorem
The mass moment of inertia through the CG is related to mass moments of inertia of parallel
axes through different points as follows.
2
ZZO ZZG
I I md = +

where d is the scalar distance separating the axis of interest O from the axis through the CG. Notice I
ZZG

is as small as possible. Any I
ZZO
must be greater, due to the term md
2
which is always positive.

Parallel axis theorem example

Rectangular rigid body (where axis O is the corner)

2
2 2
2 2
2 2 2 2
2 2
2 2
( )
12 4 4
12 4 12 4
3 3
( )
3
ZZO ZZG
I I md
m b h
b h m
b b h h
m
b h
m
m
b h
= +
| |
= + + +
|
\ .
| |
= + + +
|
\ .
| |
= +
|
\ .
= +


Combining multiple rigid bodies into a rigid link

To combine any number of bodies n of known material, shape, size, and location into one rigid
body, use the following equations for mass, center of gravity, and mass moment of inertia.
1 2
1
n
T n i
i
m m m m m
=
= + + + =


1 1 2 2 1
1 1 2 2
1
n
i i
n n i
T T T
CGT
n
n n T
i i
i T
T
m x
m x m x m x
m m X
P
m y m y m y Y
m y
m
m
=
=


+ + +




= = =
` ` `
+ + +
)

)

)


2 2 2 2
1 1 1 2 2 2
1
n
GZT GZ GZ GZn n n GZi i i
i
I I md I m d I m d I md
=
= + + + + + + = +



The subscript T indicates total (or combined) and d
i
is the distance between the combined CG (
T
CG
P )
and the CG of body i ( | |
i
T
CG i i
P x y = ). These equations are obtained from the mass, CG, and I
GZ

equations for particles, where now each particle is instead a rigid body.



99

Example 1. Two rectangles joined as shown below

Given b
1
=2.2 x h
1
=0.1 x t
1
=0.005 (m)
b
2
=1.0 x h
2
=0.8 x t
2
=0.005 (m)
the material is steel with a mass density of =7850 kg/m
3


The equations are:
1 2 T
m m m = +
1 1 2 2
1 1 2 2
T T
CGT
T
T
m x m x
m X
P
m y m y Y
m
+



= =
` `
+
)

)

2 2
1 1 1 2 2 2 GZT GZ GZ
I I m d I m d = + + +
where d
i
is the distance from CG
i
to the combined CG
CGT
P .
The results are:
m
1
= 8.64 m
2
= 31.40 m
T
= 40.04 (kg)

1
0.05
1.10
CG
P

=
`
)

2
0.60
1.80
CG
P

=
`
)

0.48
1.65
CGT
P

=
`
)
(m)
d
1
= 0.70 d
2
= 0.19 (m)
( )
2 2 1
1 1 1
3.49
12
GZ
m
I b h = + =

( )
2 2 2
2 2 2
4.29
12
GZ
m
I b h = + = I
GZT
= 13.15 (kgm
2
)


Example 1. Two Unequal Steel Rectangles (here b
1
is vertical and b
2
is horizontal)

0 0.5 1 1.5 2
0
0.5
1
1.5
2
X (m)
Y

(
m
)


100

Example 2. Two equal rectangles joined in the same manner

Given b =2.0 x h =0.2 x t =0.005 (m) (twice)
=7850 kg/m
3


The analytical equations for this special case are:
2
T
m m =
3
4
3
4
T
CGT
T
b h
X
P
b h Y
+



= =
` `

)

)

2 2
5
( )
12
GZT
I m b h = +
where
2 2
1 2
8
b h
d d
+
= = is the distance from CG
i
to
CGT
P .

The results are:
m
1
= m
2
= 15.70 m
T
= 31.40 (kg)
1
0.10
1.00
CG
P

=
`
)

2
1.20
1.90
CG
P

=
`
)

0.65
1.45
CGT
P

=
`
)
(m)
d
1
= d
2
= 0.71 (m)
2 2
1 2
( ) 5.29
12
GZ GZ
m
I I b h = = + = I
GZT
= 26.43 (kgm
2
)


Example 2. Two Equal Steel Rectangles
(again, b
1
is vertical and b
2
is horizontal, with b
1
=b
2
=b here)

0 0.5 1 1.5 2
0
0.5
1
1.5
2
X (m)
Y

(
m
)


101

Tables of Mass, Center of Gravity, and Mass Moment of Inertia

The tables below present the mass (kg), center of gravity (m), and mass moment of inertia (kg-
m
2
) for some common planar link shapes.

Note that mass moment of inertia (kg-m
2
) is not the same as area moment of inertia for beam
bending (m
4
). The former represents rotational inertia while the latter is a measure of resistance to beam
bending.

We assume that all link materials are homogeneous and uniformly distributed, with mass density
(kg/m
3
), all links have regular geometry, and all links have a constant thickness t in the Z direction.

The general equations for mass, center of gravity, and mass moment of inertia are given below
for general rigid bodies.

2 2
( )
body
body
CG
ZZ
body
m dm
rdm
P
dm
I x y dm
=
=
= +
}
}
}
}


All of these terms require double integrals over the rigid body in the XY plane. Both center of
gravity and mass moment of inertia depend on the origin of the chosen XYZ Cartesian coordinate
system.

For general 3D rigid bodies, mass moment of inertia is a 3x3 tensor. For planar mechanism
dynamics we only need one scalar term out of these 9 terms, I
ZZ
.

In the drawings below, the planar reference Cartesian coordinate system is shown, with origin O,
and the standard symbol is used for center of gravity, denoted as point G. The mass moment of inertia
about axis O is related to the mass moment of inertia about axis G is via the parallel axis theorem, where
d is the scalar distance (vector length) between axes G and O in the XY plane:

2
OZ GZ
I I md = +




102

Mass Properties for Planar Links


Name

Model Mass (kg)
Center of
Gravity (m)
Mass Moment of
Inertia (kg-m
2
)

point mass




m

0
0

`
)



0

point mass on
massless rod


m

0
L

`
)


2
0
GZ
OZ
I
I mL
=
=


slender rod



m

2
0
L

`

)

2
2
12
3
GZ
OZ
mL
I
mL
I
=
=


rectangular
parallelopiped


Lht
2
2
L
h



`


)

2 2
2 2
( )
12
( )
3
GZ
OZ
m L h
I
m L h
I
+
=
+
=


square




2
s t
2
2
s
s



`


)

2
2
6
2
3
GZ
OZ
ms
I
ms
I
=
=



X
Y
O
m
X
Y
m L
O
X
Y
L
O
X
Y
L
O
h
X
Y
O
s
s


103

Mass Properties for Planar Links (continued)


Name

Model Mass (kg)
Center of
Gravity
(m)
Mass Moment of Inertia
(kg-m
2
)


cylinder





2
R t t


R
R

`
)

2
2
2
5
2
GZ
OZ
mR
I
mR
I
=
=



hollow
cylinder




( )
2 2
o i
R R t t

0
0
R
R

`
)

2 2
2 2
( )
2
(5 )
2
o i
GZ
o i
OZ
m R R
I
m R R
I
+
=
+
=



thin ring





m

R
R

`
)


2
2
3
GZ
OZ
I mR
I mR
=
=



triangle



2
bht

3
3
a b
h
+


`


)

2 2 2
2 2 2
( )
18
( )
6
GZ
OZ
m a ab b h
I
m a ab b h
I
+ +
=
+ + +
=


The formula for the mass moment inertia of a triangle was derived via double integral over the
body by Ohio University Ph.D. student Elvedin Kljuno it could not be found in any sophomore-level
dynamics book, or in any other textbook, nor in any Internet search.


X
Y
O
R
X
Y
O
R
i
R
o
X
Y
O
R
X
Y
O
h
a
b


104

Mass Properties for Cylindrical Links

Now, the previous nine shapes for which the mass properties were summarized are all planar
shapes, with a constant thickness t in the Z direction (except for the point mass, point mass on massless
rod, and slender rod, whose Z dimensions are unimportant). Here we give one more shape, a cylinder
that is 3D but useful for planar slider-crank mechanisms and other planar mechanisms with a prismatic
joint and sliding cylindrical piston. The cylinder given on the previous page was arranged with the
circle in the XY plane. Now we need to rotate this so the rectangular projection of the cylinder is the XY
plane. The mass moment of inertia is quite different from that of the rectangular parallelepiped, due to
the effect of the radius R in this case, as opposed to the constant thickness t in the rectangular
parallelepiped case.

Not all pistons are solid, so we also include a similar model for the hollow piston cylinder, with
outer radius R
o
and inner radius R
i
.



Name

Model Mass (kg)
Center of
Gravity (m)
Mass Moment of Inertia
(kg-m
2
)


piston
cylinder




2
R L t

2
L
R


`

)

2 2
2 2
( 3 )
12
(4 15 )
12
GZ
OZ
m L R
I
m L R
I
+
=
+
=


hollow
piston
cylinder




( )
2 2
o i
R R L t


2
L
R


`

)

2 2 2
2 2 2
( 3 3 )
12
(4 15 3 )
12
o i
GZ
o i
OZ
m L R R
I
m L R R
I
+ +
=
+ +
=




X
Y
L
O
R
X
Y
L
O
R
o
R
i


105

English Units for Mass

ME 3011 uses SI units exclusively. However, many of you perform your capstone project work
using English units, which is fine, since we live in the U.S. In the 1970s the U.S. government mandated
a change to the SI system this failed spectacularly (why?).

One big benefit of the SI system is seen in the units for Newtons Second Law, F =ma. Using
standard SI units, this equation uses all ones (1s):

1 Newton accelerates 1 kg 1 m/s
2

Sadly the English units DO NOT behave with ones in Newtons Second Law:

1 lb
f
DOES NOT accelerate 1 lb
m
1 ft/s
2

1 lb
f
DOES NOT accelerate 1 lb
m
1 in/s
2


Further, the English system has another confusion which does not exist for the SI system. The
same unit, pound (lb), applies both to force (lb
f
) and mass (lb
m
), depending on the context. Please
always use the correct subscript for clarity. Happily, a mass of 1 lb
m
does weigh 1 lb
f
at standard gravity
(g =32.2 ft/s
2
or 386.1 in/s
2
).

Now we present the standard English mass units; there are two, depending on if you use feet or
inches for the length unit.

1 lb
f
accelerates 1 slug 1 ft/s
2

1 lb
f
accelerates 1 blob 1 in/s
2


WTF?!? slug? blob? I promise you I am not making this up. A slug is a rather large mass,
equivalent to 32.2 lb
m
(14.6 kg). A blob is even larger, equivalent to 12 slugs, 386.1 lb
m
(175.1 kg).

In conclusion, do not use lb
m
in dynamics equations for your project. Instead use slugs if you are
using feet or blobs if you are using inches. If you have estimated your masses in lb
m
, simply divide by
32.2 to get slugs, or divide by 386.1 to get blobs.

Finally, from the above we have the following units equivalents:

f 2
slug ft
1 lb 1
s

= so
2
f
lb s
1 slug 1
ft

=

f 2
blob in
1 lb 1
s

= so
2
f
lb s
1 blob 1
in

=





106

3.4 Four-Bar Mechanism Inverse Dynamics Analysis

Here are the link 2 and 4 details for the four-bar inverse dynamics matrix from the NotesBook.

Link 2 details Link 4 details




2 2 2 12
12
2 2 2 12
12 2 2
32 12 2
12 2 2
cos( )
sin( )
cos
sin
G X
G Y
X
Y
R r
r
R r
r r
r r r
r r
u o
u o
u
u
+
= =
` `
+
) )
+
= + =
`
+
)

4 4 4 14
14
4 4 4 14
14 4 4
34 14 4
14 4 4
cos( )
sin( )
cos
sin
G X
G Y
X
Y
R r
r
R r
r r
r r r
r r
u o
u o
u
u
+
= =
` `
+
) )
+
= + =
`
+
)


2
2 2 2 2 2 2 2 2 2
2
2
2 2 2 2 2 2 2 2 2
sin( ) cos( )
cos( ) sin( )
G X G G
G
G Y G G
A R R
A
A R R
o u o e u o
o u o e u o
+ +
= =
` `
+ +
) )


2
4 4 4 4 4 4 4 4 4
4
2
4 4 4 4 4 4 4 4 4
sin( ) cos( )
cos( ) sin( )
G X G G
G
G Y G G
A R R
A
A R R
o u o e u o
o u o e u o
+ +
= =
` `
+ +
) )


4
4
4 4 4
4
4 4 4
4
4
4
4 4
cos
sin
cos
sin
E
E
E X E E
E
E Y E E
E r
E X
E
E Y E r
F F
F
F F
r
r
r
r r
|
|
|
|

= =
` `
) )



= =
` `
)
)


4 4

E E
M M k =



107

Four-bar mechanism inverse dynamics matrix equation

21
21
12 12 32 32 32
32
43
23 23 43 43 43
14
14
34 34 14 14 2
1 0 1 0 0 0 0 0 0
0 1 0 1 0 0 0 0 0
0 0 0 0 1
0 0 1 0 1 0 0 0 0
0 0 0 1 0 1 0 0 0
0 0 0 0 0
0 0 0 0 1 0 1 0 0
0 0 0 0 0 1 0 1 0
0 0 0 0 0
X
Y
Y X Y X X
Y
X
Y X Y X Y
X
Y
Y X Y X
F
F
r r r r F
F
F
r r r r F
F
F
r r r r t
(
(

(
(
(


(

(
(

(
(


2 2
2 2
2 2
3 3 3
3 3 3
3 3 3 3 3 3 3
4 4 4
4 4 4
4 4 4 4 4 4 4
( )
( )
( )
G X
G Y
G Z
G X E X
G Y E Y
G Z E X E Y E Y E X E
G X E X
G Y E Y
G Z E X E Y E Y E X E
m A
m A g
I
m A F
m A g F
I r F r F M
m A F
m A g F
I r F r F M
o
o
o


+






= +
` `

+



+

+
) )

| |{ } { } A v b =

Step 6. Solve for the unknowns (alternate solution)
It is possible to partially decouple the solution to this problem
1
. If we consider the FBDs of only
links 3 and 4 first, this is 6 equations in 6 unknowns this is verified by looking at the original 9x9
matrix and noting three 6x1 columns of zeros (1, 2, 9) in rows 4 through 9. Here is a more efficient
solution. The reduced 6x6 set of equations for links 3 and 4 are given below.

3 3 3 32
3 3 3 32
3 3 3 3 3 3 3 23 23 43 43 43
4 4 4 43
14
34 34 14 14 14
1 0 1 0 0 0
( ) 0 1 0 1 0 0
0 0
0 0 1 0 1 0
0 0 0 1 0 1
0 0
G X E x X
G Y E y Y
G Z E x E y E y E x E Y X Y X X
G X E x Y
X
Y X Y X Y
m A F F
m A g F F
I r F r F M r r r r F
m A F F
F
r r r r F
o
(
(
+
(
+ (

=
( `

(
(

(

(
)
4 4 4
4 4 4 4 4 4 4
( )
G Y E y
G Z E x E y E y E x E
m A g F
I r F r F M o





`


+

+

)


| |{ } { }
34 34 34
A v b =

Solve for six unknowns { } | | { }
1
34 34 34
v A b

= and then use F


32X
and F
32Y
in the following 3x3 set of linear
equations, from the link 2 FBD, very similar to the single rotating link.

21 2 2 32
21 2 2 32
12 12 2 2 2 32 32 32 32
1 0 0
0 1 0 ( )
1
X G X X
Y G Y Y
Y X G Z Y X X Y
F m A F
F m A g F
r r I r F r F t o
(

(
= +
` `
(

( +
) )



1
R.L. Williams II, 2009, Partial Decoupling of the Matrix Method for Planar Mechanisms Inverse Dynamics, CD
Proceedings of the ASME International Design Technical Conferences, 33
rd
Mechanisms and Robotics Conference, Paper #
DETC2009-87054, San Diego CA, August 30-September 2.


108

We do not need a matrix solution here since the X and Y force equations are decoupled. The solution is:

21 2 2 32
21 2 2 32
2 2 2 32 32 32 32 12 21 12 21
( )
X G X X
Y G Y Y
G Z Y X X Y Y X X Y
F m A F
F m A g F
I r F r F r F r F t o
= +
= + +
= + +


Matrix inversion requires approximately
3
3
log
n
n
and Gaussian elimination requires approximately
2
2
( 1)
3
n n
n

+ multiplications/divisions
2
.


Number of Multiplications/Divisions for Four-bar Inverse Dynamics Solution

Method Inversion Gaussian Reduction
9x9 2292 321 86%
6x6 plus decoupled link 2 840 113 87%
Reduction in cost 63% 65%


There is a substantial 65% reduction in computational cost for Gaussian elimination with the
6x6 plus decoupled link 2 method. Also, the numerical accuracy may also improve with this method
since we neednt do unnecessary calculations with the three 6x1 columns of zeros.



2
E.D. Nering, 1974, Elementary Linear Algebra, W.B. Saunders Company, Philadelphia: 38-39.


109

3.5 Slider-Crank Mechanism Inverse Dynamics Analysis

Slider-crank mechanism inverse dynamics matrix equation

21 2 2
21 2 2
32 12 12 32 32
32
43
43 23 23 43 43
14
2
1 0 1 0 0 0 0 0
( ) 0 1 0 1 0 0 0 0
0 0 0 1
0 0 1 0 1 0 0 0
0 0 0 1 0 1 0 0
0 0 0 0
0 0 0 0 1 0 0
0 0 0 0 0 1 1 0
X G X
Y G Y
X G Y X Y X
Y
X
Y Y X Y X
Y
F m A
F m A g
F I r r r r
F
F
F r r r r
F
t
(
(
+
(
(
(


(
=
`
(


(


(

(


(

(
)
2 2
3 3 3
3 3 3
3 3 3 3 3 3 3
4 4 4
4 4
( )
Z
G X E X
G Y E Y
G Z E X E Y E Y E X E
G X E X
E Y
m A F
m A g F
I r F r F M
m A F
m g F
o
o





`
+


+

)


| |{ } { } A v b =

Step 6. Solve for the unknowns (alternate solution)

Like the four-bar mechanism, it is possible to partially decouple the solution to this problem
1
. If
we consider the FBDs of only links 3 and 4 first, this is 5 equations in 5 unknowns this is verified by
looking at the original 8x8 matrix and noting three 5x1 columns of zeros (1, 2, 8) in rows 4 through 8.
Here is a more efficient solution. The reduced 5x5 set of equations for links 3 and 4 are given below.

3 3 3 32
3 3 3 32
3 3 3 3 3 3 3 43 23 23 43 43
4 4 4 43
4 4 14
1 0 1 0 0
( ) 0 1 0 1 0
0
0 0 1 0
0 0 0 1 1
G X E x X
G Y E y Y
G Z E x E y E y E x E X Y X Y X
G X E x Y
E y Y
m A F F
m A g F F
I r F r F M F r r r r
m A F F
m g F F
o

(
(
+
(

+ ( =
` `
(


(

(

) )


| |{ } { }
34 34 34
A v b =

Solve for five unknowns { } | | { }
1
34 34 34
v A b

= and then use F


32X
and F
32Y
in the following 3x3 set of
linear equations, from the link 2 FBD.

21 2 2 32
21 2 2 32
12 12 2 2 2 32 32 32 32
1 0 0
0 1 0 ( )
1
X G X X
Y G Y Y
Y X G Z Y X X Y
F m A F
F m A g F
r r I r F r F t o
(

(
= +
` `
(

( +
) )





110

We do not need a matrix solution here since the X and Y force equations are decoupled. The solution is
identical to that of the four-bar mechanism.

21 2 2 32
21 2 2 32
2 2 2 32 32 32 32 12 21 12 21
( )
X G X X
Y G Y Y
G Z Y X X Y Y X X Y
F m A F
F m A g F
I r F r F r F r F t o
= +
= + +
= + +


Matrix inversion requires approximately
3
3
log
n
n
and Gaussian elimination requires approximately
2
2
( 1)
3
n n
n

+ multiplications/divisions (Nering, 1974).



Number of Multiplications/Divisions for Slider-Crank Inverse Dynamics Solution

Method Inversion Gaussian Reduction
8x8 1701 232 86%
5x5 plus decoupled link 2 544 72 87%
Reduction in cost 68% 69%


There is a substantial 69% reduction in computational cost for Gaussian elimination with the
5x5 plus decoupled link 2 method. Also, the numerical accuracy may also improve with this method
since we neednt do unnecessary calculations with the three 5x1 columns of zeros.




111

3.6 Inverted Slider-Crank Mechanism Inverse Dynamics Analysis

This problem can be solved with a 9x9 matrix, after eliminating a redundant equation. Lets try
a simpler approach assume the link 3 mass is small and use only FBDs for links 2 and 4. We further
assume zero external forces and moments.

Step 1. The inverted slider-crank Position, Velocity, and Acceleration Analyses must be complete.


Step 2. Draw the Inverted Slider-Crank Mechanism Free-Body Diagrams







ij
F unknown vector internal joint force of link i acting on link j.
ij
r known vector moment arm pointing to the joint connection with link i from the CG of link j.



112

Step 3. State the Problem

Given r
1
, u
1
=0, r
2

u
2
, r
4
, u
4

e
2
,
4
r , e
4

o
2
,
4
r , o
4
assume zero external forces and moments

Find
21
F ,
42
F ,
14
F and t
2



Step 4. Derive the Newton-Euler Dynamics Equations.

Newton's Second Law

Link 2
2 42 21 2 2 2 G
F F F W m A = + =




Link 4
4 14 42 4 4 4 G
F F F W m A = + =




Euler's Rotational Dynamics Equation

Link 2
2 42 42 12 21 2 2 2 G z G Z
M r F r F I t o = + =




Link 4
4 14 14 24 42 4 4 G z G Z
M r F r F I o = =




Count the number of unknowns and the number of equations: 6 scalar equations and 7 scalar unknowns.
We need an additional equation; let us assume zero friction between links 2 and 4. Therefore,
42
F is
always perpendicular to link 4 and there is only one unknown from this vector, the magnitude F
42
.

42 42 4
42
42 42 4
cos( 2)
sin( 2)
X
Y
F F
F
F F
u t
u t
+
= =
` `
+
) )



113

Step 5. Derive the XYZ scalar dynamics equations from the vector dynamics equations.

Link 2

42 21 2 2
42 21 2 2
2 42 42 42 42 12 21 12 21 2 2
( )
( ) ( )
X X G X
Y Y G Y
X Y Y X X Y Y X G Z
F F m A
F F m A g
r F r F r F r F I t o
=
= +
+ =


Link 4

( )
( ) ( )
14 42 4 4
14 42 4 4
14 14 14 14 24 42 24 42 4 4
X X G X
Y Y G Y
X Y Y X X Y Y X G Z
F F m A
F F m A g
r F r F r F r F I o
=
= +
=



Express these scalar equations in matrix/vector form. The simplified inverted slider-crank mechanism
inverse dynamics matrix equation is given below.

2 2 21
2 2 21
2 2 12 12 42 42 42
4 4 14
4 4 14
4 4 24 24 14 14 2
1 0 0 0 0
( ) 0 1 0 0 0
0 0 1
0 0 1 0 0
( ) 0 0 0 1 0
0 0 0
G X X
G Y Y
G Z Y X X Y
G X X
G Y Y
G Z X Y Y X
m A c F
m A g s F
I r r r s r c F
m A c F
m A g s F
I r s r c r r
o
o t
(
(
+
(
(

=
( ` `

(
(
+
(
+
(
)

)


where:
4
4
cos( 2)
sin( 2)
c
s
u t
u t
+
=
` `
+
) )


| |{ } { } A v b =



114

Step 6. Solve for the unknowns
The coefficient matrix [A] is dependent on the mechanism geometry (i.e. the angles from the
position kinematics solution). The right-hand-side vector {b} is dependent on inertial terms and gravity.

Matrix/vector solution { } | | { }
1
v A b

=

MATLAB v = inv(A)*b; % Solution via matrix inverse

Using Gaussian elimination is more efficient and robust.

MATLAB v = A\b; % Solution via Gaussian elimination

The solution to the internal forces and input torque are contained in the components of v. To save these
values for later plotting, use the following MATLAB code, inside the i loop.

f21x(i) = v(1);
f21y(i) = v(2);

tau2(i) = v(6);

Like the four-bar and slider-crank mechanisms, it is possible to partially decouple the solution to
this problem. If we consider the FBDs of only 4 first, this is 3 equations in 3 unknowns this is verified
by looking at the original 6x6 matrix and noting three 3x1 columns of zeros (1, 2, 6) in rows 4 through
6. Here is a more efficient solution. The reduced 3x3 set of equations for link 4 is given below.

42 4 4
14 4 4
24 24 14 14 14 4 4
1 0
0 1 ( )
G X
X G Y
X Y Y X Y G Z
c F m A
s F m A g
r s r c r r F I o
(

(
= +
` `
(

( +
) )


| |{ } { }
4 4 4
A v b =

Solve for three unknowns { } | | { }
1
4 4 4
v A b

= and then use F


42
in the following 3x3 set of linear equations,
from the link 2 FBD.

21 2 2 42
21 2 2 42
12 12 2 2 2 42 42 42
1 0 0
0 1 0 ( )
1 ( )
X G X
Y G Y
Y X G Z X Y
F m A cF
F m A g sF
r r I r s r c F t o
(

(
= +
` `
(

(
) )


We do not need a matrix solution here since the X and Y force equations are decoupled. The solution is
identical to that of the four-bar mechanism.

21 2 2 42
21 2 2 42
2 2 2 12 21 12 21 42 42 42
( )
( )
X G X
Y G Y
G Z Y X X Y X Y
F m A cF
F m A g sF
I r F r F r s r c F t o
= +
= + +
= +



115

Matrix inversion requires approximately
3
3
log
n
n
and Gaussian elimination requires approximately
2
2
( 1)
3
n n
n

+ multiplications/divisions (Nering, 1974).



Number of Multiplications/Divisions
Method Inversion Gaussian Reduction
6x6 833 106 87%
3x3 plus decoupled link 2 177 24 86%
Reduction in cost 79% 77%

There is a substantial 77% reduction in computational cost for Gaussian elimination with the 3x3 plus
decoupled link 2 method. Also, the numerical accuracy may also improve with this method since we
neednt do unnecessary calculations with the three 3x1 columns of zeros.


Step 7. Calculate Shaking Force and Moment

After the basic inverse dynamics problem is solved, we can calculate the vector shaking force
and vector shaking moment, which is the force/moment reaction on the ground link due to the motion,
inertia, weight, and external loads (which we assumed to be zero in this problem). The shaking force
and moment for the inverted slider-crank mechanism is identical to the four-bar in notation and terms.

Ground link force/moment diagram







Shaking force
21 14
21 41 21 14
21 14
X X
S
Y Y
F F
F F F F F
F F

= + = =
`

)



Shaking moment
2 21 21 41 14
2 21 21 21 21 41 14 41 14
S
X Y Y X X Y Y X
M r F r F
r F r F r F r F
t
t
= +
= + +





116

Inverted slider-crank mechanism inverse dynamics example Term Example 3 continued

Given
1
2
4
1
0.20
0.10
0.32
0
r
r
L
u
=
=
=
=

(m)
2
4
4
70
150.5
0.19 r
u
u
=
=
=

(deg and m)

2
4
4
25
2.47
2.18
r
e
e
=
=
=
(rad/s and m/s)
2
4
4
0
9.46
267.14
r
o
o
=
=
=
(rad/s
2
and m/s
2
)

In this problem the external forces and moments are zero for both links 2 and 4. In inverse
dynamics we ignore the slider link mass and inertia. The mechanism links 2 and 4 are uniform,
homogeneous rectangular solids made of steel ( =7850 kg/m
3
) with a constant thickness of 2 cm and
link widths of 3 cm. The CGs are in the geometric center of each link. This yields the following fixed
dynamics parameters.
m
2
=0.47 and m
4
=1.51 (kg) I
GZ2
=0.0004 and I
GZ4
=0.013 (kg-m
2
)

Snapshot Analysis
Given the previous mechanism position, velocity, and acceleration analyses, solve the inverse
dynamics problem for this snapshot (
2
70 u =

). The matrix-vector equation to solve is given below.

21
21
42
14
14
2
1 0 0.49 0 0 0 5.03
0 1 0.87 0 0 0 9.21
0.05 0.02 0.01 0 0 1 0
0 0 0.49 1 0 0 30.77
0 0 0.87 0 1 0 41.82
0 0 0.03 0.08 0.14 0 3.47
X
Y
X
Y
F
F
F
F
F
t
(
(

(
(

=
( ` `

(
(


) )


The answer is:
21
21
42
14
14
2
35.35
62.69
61.47
0.46
11.65
1.10
X
Y
X
Y
F
F
F
F
F
t





=
` `





) )



The associated vector shaking force and moment are
35.81
51.03
SX
S
SY
F
F
F

= =
` `
) )
(N)

8.53
S
M k = (Nm)


117

Full-Range-Of-Motion (F.R.O.M.) Analysis Term Example 3 continued

A more meaningful result is to solve and plot the inverse dynamics analysis results for the entire
range of mechanism motion. The plots below give the inverse dynamics results for all
2
0 360 u s s

,
for Term Example 3. Since e
2
is constant, we can plot the velocity results vs. u
2
(since it is related to
time t via
2 2
t u e = ).



Input Torque t
2

0 50 100 150 200 250 300 350
-20
-10
0
10
20
30
u
2
(deg)
t
2

(
N
m
)


t
2
t
AVG
t
RMS


118


Shaking Force


Shaking Moment


0 50 100 150 200 250 300 350
-200
-150
-100
-50
0
50
100
150
u
2
(deg)
F
S

(
N
)


F
SX
F
SY
0 50 100 150 200 250 300 350
-20
-15
-10
-5
0
5
10
15
20
u
2
(deg)
M
S

(
N
m
)


119

3.7 Multi-loop Mechanism Inverse Dynamics Analysis

The Matrix Method can be applied to any planar mechanism inverse dynamics problem. Here
are the five two-loop six-bar mechanisms from Dr. Bobs on-line Mechanism Atlas.


Stephenson I 6-Bar Mechanism


Stephenson II 6-Bar Mechanism Stephenson III 6-Bar Mechanism


Watt I 6-Bar Mechanism Watt II 6-Bar Mechanism




120


For example, here are the Watt II six-bar mechanism FBDs, ignoring external forces and moments.



121

The Watt II six-bar mechanism inverse dynamics 15x15 matrix-vector equation is given below.

12 12 32 32
23 23 43 43
34 34 14 14 54 54
1 0 1 0 0 0 0 0 0 0 0 0 0 0 0
0 1 0 1 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 1
0 0 1 0 1 0 0 0 0 0 0 0 0 0 0
0 0 0 1 0 1 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 1 0 1 0 1 0 0 0 0 0 0
0 0 0 0 0 1 0 1 0 1 0 0 0 0 0
0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 1
Y X Y X
Y X Y X
Y X Y X Y X
r r r r
r r r r
r r r r r r

21
21
32
32
43
43
14
14
54
54
45 45 65 65
56 56 16 16
0 1 0 0 0 0
0 0 0 0 0 0 0 0 0 1 0 1 0 0 0
0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 1 0 1 0 0
0 0 0 0 0 0 0 0 0 0 0 1 0 1 0
0 0 0 0 0 0 0 0 0 0 0
X
Y
X
Y
X
Y
X
Y
X
Y
Y X Y X
Y X Y X
F
F
F
F
F
F
F
F
F
F
r r r r
r r r r
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(
(

(
(

(

(
(

(

(

2 2
2 2
2 2
3 3
3 3
3 3
4 4
4 4
4 4
5 5
5 5 65
5 5 65
6 6 16
6 6 16
6 6 2
( )
( )
( )
( )
( )
G X
G Y
G Z
G X
G Y
G Z
G X
G Y
G Z
G X
G Y X
G Z Y
G X X
G Y Y
G Z
m A
m A g
I
m A
m A g
I
m A
m A g
I
m A
m A g F
I F
m A F
m A g F
I
o
o
o
o
o t


+





+





= +
`




+





+


)





)


| |{ } { } A v b =



122

Step 6. Solve for the unknowns (continued)
Like the four-bar mechanism, it is possible to partially decouple the solution to this problem. If
we consider the FBDs of only links 5 and 6 first, this is 6 equations in 6 unknowns this is verified by
looking at the original 15x15 matrix and noting nine 6x1 columns of zeros (1, 2, 3, 4, 5, 6, 7, 8, 15) in
the six rows 10 through 15. Here is a more efficient solution. The reduced 6x6 set of equations for links
5 and 6 is given below.
54 5 5
54 5 5
45 45 65 65 65 5 5
65 6 6
16 6 6
56 56 16 16 16 6 6
1 0 1 0 0 0
0 1 0 1 0 0 ( )
0 0
0 0 1 0 1 0
0 0 0 1 0 1 ( )
0 0
X G X
Y G Y
Y X Y X X G Z
Y G X
X G Y
Y X Y X Y G Z
F m A
F m A g
r r r r F I
F m A
F m A g
r r r r F I
o
o
(
(
+
(
(

=
( ` `

(
(
+
(

(
) )



| |{ } { }
56 56 56
A v b =

Solve for the six unknowns { } | | { }
1
56 56 56
v A b

= . Second, consider the FBDs of only links 3 and 4: this


is 6 equations in 6 unknowns this is verified by looking at the original 15x15 matrix and noting seven
6x1 columns of zeros (1, 2, 11, 12, 13, 14, 15) in the six rows 4 through 9. Recognizing that now F
54X

and F
54Y
are now known from the above 6x6 partial solution, here is a more efficient solution. The
reduced 6x6 set of equations for links 3 and 4 is given below.
3 3 32
3 3 32
23 23 43 43 3 3 43
4 4 54 43
4 4 54 14
34 34 14 14 4 4 54 54 14
1 0 1 0 0 0
0 1 0 1 0 0 ( )
0 0
0 0 1 0 1 0
0 0 0 1 0 1 ( )
0 0
G X X
G Y Y
Y X Y X G Z X
G X X Y
G Y Y X
Y X Y X G Z Y X Y
m A F
m A g F
r r r r I F
m A F F
m A g F F
r r r r I r F r F
o
o
(
(
+
(
(

=
( `

(
(
+
(
+
(
) 54 54 X Y
F





`




)


| |{ } { }
34 34 34
A v b =

Solve for six unknowns { } | | { }
1
34 34 34
v A b

= and then use F


32X
and F
32Y
in the following 3x3 set of linear
equations, from the link 2 FBD.
21 2 2 32
21 2 2 32
12 12 2 2 2 32 32 32 32
1 0 0
0 1 0 ( )
1
X G X X
Y G Y Y
Y X G Z Y X X Y
F m A F
F m A g F
r r I r F r F t o
(

(
= +
` `
(

( +
) )


We do not need a matrix solution here since the X and Y force equations are decoupled. The solution is
identical to that of the four-bar mechanism.
21 2 2 32
21 2 2 32
2 2 2 32 32 32 32 12 21 12 21
( )
X G X X
Y G Y Y
G Z Y X X Y Y X X Y
F m A F
F m A g F
I r F r F r F r F t o
= +
= + +
= + +




123

Matrix inversion requires approximately
3
3
log
n
n
and Gaussian elimination requires approximately
2
2
( 1)
3
n n
n

+ multiplications/divisions (Nering, 1974).



Number of Multiplications/Divisions

Method Inversion Gaussian Reduction
15x15 8609 1345 84%
6x6 twice, plus decoupled link 2 1672 183 89%
Reduction in cost 81% 86%

There is an astonishing 86% reduction in computational cost for Gaussian elimination with the
methods using 6x6 inversion twice, plus the decoupled link 2 solution. Also, the numerical accuracy
may also improve with this method since we neednt do unnecessary calculations with the sixteen 6x1
columns of zeros.


Any mechanism with a dyad of binary links may be decoupled in this manner. Thus, the method
is similar and the computational complexity identical for the Stephenson I, Stephenson III, Watt I, and
Watt II six-bar mechanisms.


The Stephenson II six-bar mechanism does not include a dyad of binary links and so it cannot
be solved like the other 4 six-bar mechanisms (first links 5 and 6, then links 3 and 4 with one unknown
vector force from 5 and 6, then link 2 independently). But links 3, 4, 5, and 6 can be solved first
independently of link 2: a 12x12 solution followed by the standard link 2 solution. The computational
savings is not as impressive as in the former six-bar cases.


Number of Multiplications/Divisions, Stephenson II Six-Bar

Method Inversion Gaussian Reduction
15x15 8609 1345 84%
12x12 plus decoupled link 2 4811 723 85%
Reduction in cost 44% 46%


There is a 46% reduction in computational cost for Gaussian elimination with the 12x12 plus
decoupled link 2 method. Also, the numerical accuracy may also improve with this method since we
neednt do unnecessary calculations with the three 12x1 columns of zeros.




124

3.8 Balancing of Rotating Shafts

If high-speed shafts are unbalanced, this can lead to the following problems.

- unwanted vibrations
- shaking forces
- wear
- noise
- safety concerns
- comfort of users/riders
- less efficient
- shorter service life


Let us start with the balancing of a single idealized point mass.

1) Static Balance Moments about the rotating shaft must be balanced statically.




0
Z
M =

cos cos 0
B B
mgr m gr u u + =
B B
mr m r =


2) Dynamic Balance Inertial forces due to motion must also be balanced.

Inertial forces are not actual forces but are effects of acceleration, e.g. centripetal force. Assuming
constant input angular velocity e, the inertial force is directed outward, opposite to the centripetal
acceleration directed inward.

( ( ))
I C
F mA m r e e = = the vector magnitude is
2
I
F mre =

For dynamic balance, we again add a balance mass. The dynamic balance condition is:

0
I IB
F F + =

The original and balance inertial forces must be equal in magnitude and opposite in direction. This
vector balance condition is equivalent to the two scalar equations below.

0
0
X
Y
F
F
=
=


2 2
2 2
cos cos 0
sin sin 0
B B
B B
mr m r
mr m r
e u e u
e u e u
=
=





125

Both equations yield the same condition as the static balance case, namely:
B B
mr m r =

So, if a single mass is balanced statically, it is also balanced dynamically.



Now let us include a system of idealized point masses attached to the same rotating shaft.




1) Static Balance
The rotating shaft will be balanced statically if the system CG lies on the axis of rotation.

0
0
CG
X
P
Y

= =
` `
) )


Recall the scalar CG equations for a system of point masses:

i i
CG
i
m x
X X
m
= =


i i
CG
i
m y
Y Y
m
= =




Let us consider a system of four point masses. There are two equations to satisfy:

4
1
cos 0
i i i
i
mr u
=
=


4
1
sin 0
i i i
i
mr u
=
=



where cos
i i i
x r u = , sin
i i i
y r u = , and the
4
1
i
i
m
=

term in the denominator cancels out.


If we fix each m
i
r
i
, these are 2 equations in the four unknowns u
i
. Therefore, arbitrarily fix u
1
, u
2

and solve for u
3
, u
4
. This is equivalent to solving the four-bar linkage position problem.

1 1 1 2 2 2 3 3 3 4 4 4
1 1 1 2 2 2 3 3 3 4 4 4
0
0
mrc m r c m r c m r c
mr s m r s m r s m r s
+ + + =
+ + + =

2 2 3 3 1 1 4 4
2 2 3 3 1 1 4 4
r c r c rc r c
r s r s r s r s
+ = +
+ = +




126

Let R
i
=m
i
r
i
; the vectors 1 and 4 are reversed compared to the four-bar mechanism. So we know how to
solve these equations. The shaft will be statically balanced for any shaft angle u. Here is the associated
figure.


2) Dynamic Balance
For dynamic balance, we must consider the front view in addition to the previously-shown side
view.


If the following conditions are satisfied assuming constant e, we will have dynamic balance.
0
0
0
X
Y
Z
F
F
F
=
=
=


0
0
0
X
Y
Z
M
M
M
=
=
=



Let us consider each in turn.

4 4
2
1 1
4 4
2
1 1
cos cos 0
sin sin 0
0 0
X i i i i i i
i i
Y i i i i i i
i i
Z
F mr mr
F mr mr
F
e u u
e u u
= =
= =
= = =
= = =
=



The
X
F

and
Y
F

equations are already satisfied by the static balancing, because e


2
divides out.
The
Z
F

equation yields nothing because all inertial forces are in the XY plane.




127




4
2
1
4
2
1
sin 0
cos 0
0 0
X i i i i
i
Y i i i i
i
Z
M mr L
M mr L
M
e u
e u
=
=
= =
= =
=




The
X
M

and
Y
M

equations must be solved to satisfy dynamic balancing. The


Z
M

equation
yields nothing because all inertial forces pass through the axis of rotation.

We fixed each m
i
r
i
, we previously determined u
i
, and the e
2
term divides out. Therefore, we
have two equations in the four unknowns L
i
; arbitrarily fix L
1
, L
2
and solve for L
3
, L
4
. The result is two
linear equations in the two unknowns.

1 1 1 1 2 2 2 2 3 3 3 3 4 4 4 4
1 1 1 1 2 2 2 2 3 3 3 3 4 4 4 4
sin sin sin sin 0
cos cos cos cos 0
mr L m r L m r L m r L
mr L m r L m r L m r L
u u u u
u u u u
+ + + =
+ + + =



3 3 3 4 4 4 3 1 1 1 1 2 2 2 2
3 3 3 4 4 4 4 1 1 1 1 2 2 2 2
sin sin sin sin
cos cos cos cos
m r m r L m r L m r L
m r m r L m r L m r L
u u u u
u u u u
(
=
` `
(

) )


Solve for L
3
, L
4
and the system will be balanced statically and dynamically.




128

3.9 Inverse Dynamics Analysis Examples

This section presents complete snapshot and full-range-of-motion (F.R.O.M.) examples to
demonstrate the inverse dynamics analyses in the ME 3011 NotesBook. The first example is for the
single rotating link inverse dynamics (whose kinematics were not yet presented). Term Example 1 is
for the four-bar mechanism and Term Example 2 is for the slider-crank mechanism. The latter two
examples are continuations for the Term Example 1 and 2 kinematics examples presented earlier in this
ME 3011 NotesBook Supplement.

3.9.1 Single Rotating Link Inverse Dynamics Example

Given: L =1, h =0.1 m, m =2 kg, e =100 rad/s, o =0, F
E
=150 N, |
E
=0, M
E
=0 Nm.

Calculated terms
12
0.5
E
r r = = m I
GZ
= 0.17 kgm
2


Snapshot inverse dynamics analysis

At 150 u =

, given this link, motion, and external force, calculate
12 12
, ,
X Y
F F t and ,
S S
F M .

4330
2500
Gx
Gy
A
A
=
=

2
m
s

12
12
1 0 0 8510
0 1 0 4980
0.250 0.433 1 37.5
X
Y
F
F
t
(

(
=
` `
(

(
) )

12
12
8510
4980
66.5
X
Y
F
F
t


=
` `

) )
N, Nm

21 12
8510
4980
S
F F F

= = =
`
)
N

66.5
S
M k t = = Nm


Full-Range-Of-Motion (F.R.O.M.) Analysis

A more meaningful result from inverse dynamics analysis is to solve and plot the dynamics
unknowns for the entire range of mechanism motion. For the same example as the snapshot we specify
that the given e is constant. Prior to solving the inverse dynamics problem, the CG translational
acceleration results for all 0 360 u s s

are given in the plot below. The X components are red and the
Y green. Is the static link weight (mg) significant in this problem?




129





The plot above gives the Shaking Force
S
F for all 0 360 u s s

. The X component is red and the Y
green. The Shaking Moment
S
M is simply the negative of the driving torque t plot shown next.


0 50 100 150 200 250 300 350
-5000
0
5000
u
(deg)
A
G
X

a
n
d

A
G
Y

(
m
/
s
2
)
CG Acceleration, X (red) and Y (green)
0 50 100 150 200 250 300 350
-1
-0.5
0
0.5
1
x 10
4
u
(deg)
F
S
X

a
n
d

F
S
Y

(
N
)
Shaking Force, X (red) and Y (green)


130

The plot below gives the required driving torque t (Nm, red) for all 0 360 u s s

, assuming
the given e is constant, for the same example as the snapshot. This shows the torque that must be
supplied by an external DC servomotor to cause the specified motion. Also plotted is the average torque
(green) t
AVG
=0 and the root-mean-square (RMS) torque value (blue) t
RMS
=106.1 Nm.


Single Rotating Link Torque t Results

Here are the calculations for the average and root-mean-square torques.

0 1 2
1
k
AVG
k
t t t t
t
+ + + +
=
+


2 2 2 2
0 1 2
1
k
RMS
k
t t t t
t
+ + + +
=
+



where k+1 is the total number of elements in the t array (since the counting index k starts at zero).
MATLAB can be used to calculate and plot the average and root-mean-square torques on the plot of t
for easy comparison.

tauAVG = mean(tau); % after the for loop
tauRMS = norm(tau)/sqrt(k+1);
wuns = ones(1,length(th)); % to plot a constant line
plot(th/DR,tau,'r',th/DR,tauAVG*wuns,'g',th/DR,tauRMS*wuns,'b');

0 50 100 150 200 250 300 350
-150
-100
-50
0
50
100
150
u
(deg)
t

(
N
m
)
Tau (red) with average (green) and rms (blue)


131

3.9.2 Term Example 1: Four-Bar Mechanism

Figure for Term Example 1 Inverse Dynamics

The coupler link 3 is a rectangle of dimensions 0.203 x 0.152 (m). The triangle tip we have been
using all along in Term Example 1 (previously called point C) is the CG of the rectangular link shown
below for inverse dynamics.



-0.1 0 0.1 0.2 0.3
-0.1
-0.05
0
0.05
0.1
0.15
0.2
0.25
0.3
X (m)
Y

(
m
)


132

Four-bar mechanism inverse dynamics example Term Example 1 continued
This is the mechanism from Term Example 1 (open branch), with kinematics solutions as
presented before. Given r
1
=0.284, r
2
=0.076, r
3
=0.203, r
4
=0.178,
1
10.3 u =

,
2
30 u =

,
3
53.8 u =

,
4
121.7 u =

, R
G2
=0.038, R
G3
=0.127, R
G4
=0.089 (m),
2
0 o = ,
3
36.9 o =

,
4
0 o = ,
2
20 e = ,
3
8.09 e = ,
4
3.73 e = (rad/s),
2
0 o = ,
3
8.65 o = , and
4
244.4 o = (rad/s
2
).
All moving links are wood, with mass density 830.4 = kg/m
3
. Links 2 and 4 have rectangular
dimensions r
i
x 0.019 x 0.013 thick (m; i=2,4); link 3 has rectangular dimensions 0.203 x 0.152 x 0.013
thick (m), as shown on the previous page. The calculated mass and inertia parameters are
2
0.015 m = ,
3
0.327 m = ,
4
0.036 m = (kg) and
6
2
7.9 10
G Z
I

= ,
3
3
1.8 10
G Z
I

= ,
5
4
9.5 10
G Z
I

= (kgm
2
). All
external forces and moments are zero but gravity, g =9.81 m/s
2
, is included.

Snapshot Analysis (one input angle)
At
2
30 u =

, given this mechanism and motion, calculate the four vector internal joint forces, the
driving torque
2
t , and the shaking force and moment ,
S S
F M for this snapshot.

21
21
32
32
43
43
14
14
1 0 1 0 0 0 0 0 0
0 1 0 1 0 0 0 0 0
0.019 0.033 0.019 0.033 0 0 0 0 1
0 0 1 0 1 0 0 0 0
0 0 0 1 0 1 0 0 0
0 0 0.127 0.002 0.037 0.122 0 0 0
0 0 0 0 1 0 1 0 0
0 0 0 0 0 1 0 1 0
0 0 0 0 0.076 0.047 0.076 0.047 0
x
y
x
y
x
y
x
y
F
F
F
F
F
F
F
F
(
(

(
(
(

(
(

(

(
(

(
(
2
0.202
0.034
0
8.955
4.497
0.015
0.638
0.095
0.023 t






=
` `





) )


The solution is accomplished by Gaussian elimination, or { } | | { }
1
v A b

= , or by the reduced 6x6


plus decoupled link 2 method (see the on-line ME 3011 NotesBook Supplement). All methods yield the
same results. Snapshot answers:

{ }
21
21
32
32
43
43
14
14
2
6.20
10.08
5.99
10.11
2.96
5.61
3.60
5.52
0.43
x
y
x
y
x
y
x
y
F
F
F
F
v F
F
F
F
t







= =
` `





) )
(N, Nm)
9.80
4.56
S
F

=
`
)
(N)

1.68
S
M k = (Nm)




133

Full-Range-Of-Motion (F.R.O.M.) Analysis Term Example 1 continued

A more meaningful result from inverse dynamics analysis is to solve and plot the dynamics
unknowns for the entire range of four-bar mechanism motion. Prior to solving the inverse dynamics
problem, the plot below shows the CG translational acceleration results for link 3 for all
2
0 360 u s s

.
The X component is red and the Y component is green.



Term Example 1,
3 G
A

0 50 100 150 200 250 300 350
-30
-20
-10
0
10
20
30
u
2
(deg)
A
G
3

(
m
/
s
2
)


A
G3X
A
G3Y


134



Term Example 1 Inverse Dynamics, t
2

The plot above gives the required driving torque
2
t (Nm) for all
2
0 360 u s s

, for the Term
Example 1 mechanism, assuming the given
2
20 e = rad/s is constant. This plot shows the torque (red)
that must be supplied in all configurations by an external DC servomotor to cause the specified motion.
Also plotted is the average torque (green)
2
0.003
AVG
t = and the root-mean-square torque value (blue)
2
0.354
RMS
t = Nm. The root-mean-square (RMS) torque is more meaningful than the average torque
since its terms do not cancel each other (k+1 is the number of elements in the t
2
array).

0 1 2
2 2 2 2
2 2 2 2
2
1
k
RMS
k
t t t t
t
+ + + +
=
+




0 50 100 150 200 250 300 350
-0.4
-0.2
0
0.2
0.4
0.6
0.8
u
2
(deg)
t
2

(
N
m
)


135

The plot below gives the shaking force
S
F (N) results for all
2
0 360 u s s

. The X component
is red and the Y component is green.



Term Example 1 Inverse Dynamics,
S
F


0 50 100 150 200 250 300 350
-10
-5
0
5
10
u
2
(deg)
F
S

(
N
)

F
SX
F
SY


136

The plot below gives the shaking moment
S
M (Nm) results for all
2
0 360 u s s

. There is only
the Z component since a planar moment is a

k

vector.



Term Example 1 Inverse Dynamics,
S
M

0 50 100 150 200 250 300 350
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
u
2
(deg)
M
S

(
N
m
)


137

3.9.3 Term Example 2: Slider-Crank Mechanism

Figure for Term Example 2 Inverse Dynamics

The Term Example 2 slider-crank mechanism is shown below at the starting (or ending) position,
with zero (or 360

) input angle u
2
.


-0.1 0 0.1 0.2 0.3
-0.25
-0.2
-0.15
-0.1
-0.05
0
0.05
0.1
0.15
0.2
X (m)
Y

(
m
)


138

Slider-crank mechanism inverse dynamics example Term Example 2 continued

This is the mechanism from Term Example 2 (right branch), with kinematics solutions as
presented before. Given r
2
=0.102, r
3
=0.203, h =0.076 m,
2
30 u =

,
3
7.1 u =

, x =0.29 m,
2
15 e =
rad/s (constant),
3
6.58 e = rad/s, 0.60 x = m/s,
2
0 o = ,
3
62.33 o = rad/s
2
, and 30.15 x = m/s
2
.
All moving links are wood, with mass density 830.4 = kg/m
3
. Links 2 and 3 have rectangular
dimensions r
i
x 0.019 x 0.013 thick (m; i=2,3); link 4 has rectangular dimensions 0.076 x 0.019 x 0.013
thick (m), as shown on the previous page. The calculated inertia parameters are m
2
=0.020, m
3
=0.041,
m
4
=0.015 (kg) and I
G2Z
=1.819e-005, I
G3Z
=1.418e-004 (kgm
2
). The CGs all lie at their respective link
centers. There is a constant external force of 1 N acting at the center of the piston end, directed
horizontally to the left; gravity is included but all other external forces and moments are zero. We
assume the coefficient of kinetic friction between the piston and the fixed wall is =0.2.


Snapshot Analysis (one input angle)

At
2
30 u =

, given this mechanism and motion, calculate the four vector internal joint forces, the
driving torque
2
t , and the shaking force and moment ,
S S
F M for this instant (snapshot).

21
21
32
32
43
43
14
2
1 0 1 0 0 0 0 0 0.202
0 1 0 1 0 0 0 0 0.0
-0.025 0.044 -0.025 0.044 0 0 0 1
0 0 1 0 1 0 0 0
0 0 0 1 0 1 0 0
0 0 -0.013 0.101 -0.013 0.101 0 0
0 0 0 0 1 0 0.2 0
0 0 0 0 0 1 1 0
X
Y
X
Y
X
Y
Y
F
F
F
F
F
F
F
t
(
(

(
(
(


(
=
`
(


(

(

(


(

(
)
84
0
1.018
0.168
0.009
0.540
0.150





`





)


The solution is accomplished by Gaussian elimination, or { } | | { }
1
v A b

= , or by the reduced 5x5 plus


decoupled link 2 method (see the on-line ME 3011 NotesBook Supplement). All methods yield the
same results. Snapshot answers:

{ }
21
21
32
32
43
43
14
2
0.736
0.121
0.534
0.037
0.484
0.131
0.281
0.039
X
Y
X
Y
X
Y
Y
F
F
F
F
v
F
F
F
t


= =
` `






) )
(N, Nm)
0.680
0.401
S
F

=
`

)
(N)

0.116
S
M k = (Nm)



139

Full-Range-Of-Motion (F.R.O.M.) Analysis Term Example 2 continued

A more meaningful result from inverse dynamics analysis is to solve and plot the dynamics
unknowns for the entire range of mechanism motion. Prior to solving the inverse dynamics problem, the
plot below gives the CG translational acceleration results for link 3 for all
2
0 360 u s s

. Here CG
3
is
taken as the midpoint of link 3. The X component is red and the Y component is green.



Term Example 2,
3 G
A

0 50 100 150 200 250 300 350
-30
-25
-20
-15
-10
-5
0
5
10
15
20
25
u
2
(deg)
A
G
3

(
m
/
s
2
)


A
G3X
A
G3Y


140



Term Example 2 Inverse Dynamics, t
2

The plot above gives the required driving torque
2
t (Nm) for all
2
0 360 u s s

, for the Term
Example 2 slider-crank mechanism, right branch only, assuming the given
2
15 e = rad/s is constant.
This plot shows the torque (red) that must be supplied in all configurations by an external DC
servomotor to cause the specified motion. Also plotted is the average torque (green)
2
0.004
AVG
t =

and the root-mean-square torque value (blue)
2
0.099
RMS
t = Nm. The root-mean-square (RMS) torque
is more meaningful than the average torque since its terms do not cancel each other (k+1 is the number
of elements in the t
2
array).

0 1 2
2 2 2 2
2 2 2 2
2
1
k
RMS
k
t t t t
t
+ + + +
=
+


0 50 100 150 200 250 300 350
-0.1
-0.05
0
0.05
0.1
0.15
0.2
0.25
u
2
(deg)
t
2

(
N
m
)


141

The plot below gives the shaking force
S
F (N) results for all
2
0 360 u s s

. The X component
is red and the Y component is green.



Term Example 2 Inverse Dynamics,
S
F


0 50 100 150 200 250 300 350
-3
-2.5
-2
-1.5
-1
-0.5
0
0.5
1
1.5
u
2
(deg)
F
S

(
N
)


F
SX
F
SY


142

The plot below gives the shaking moment
S
M (Nm) result for all
2
0 360 u s s

. There is only
the Z component since a planar moment is a

k

vector.



Term Example 2 Inverse Dynamics,
S
M

0 50 100 150 200 250 300 350
-0.1
-0.05
0
0.05
0.1
0.15
u
2
(deg)
M
S

(
N
m
)


143

4. Gears and Cams

4.1 Gears

4.1.1 Gear Introduction

Gears are used to transfer motion between rotating shafts in machinery, mechanisms, robots,
vehicles, toys, and other electromechanical systems. Gears cause changes in angular velocity, torque,
and direction. Gears are used in various applications, from can openers to aircraft carriers. Belt and
chain drives are related to gear mechanisms.

Robot joint example















Gear Classification



Externally-meshing Spur Gears Internally-meshing Spur Gears


144


Helical (Parallel Shaft) Helical (Crossed Shaft)


Rack & Pinion Worm and Gear


Straight Bevel Gears Spiral Bevel Gears



145


V-Belt Drive Chain Drive



Toothed Belt Drive Bicycle Sprockets



Herringbone Gears Automotive Hypoid Gears



146


Automotive Gear Train Automotive Differential


Planetary Gear Train Planetary Gear Train Aircraft


Non-Circular Gears Harmonic Gearing





147

Harmonic Gearing

The harmonic gear allows high reduction ratios with concentric shafts and with very low
backlash and vibration. It is based on a very simple construction utilizing metals elasto-mechanical
property.

Harmonic drive transmissions are noted for their ability to reduce backlash in a motion control
system. How they work is through the use of a thin-walled flexible cup with external splines on its lip,
placed inside a circular thick-walled rigid ring machined with internal splines. The external flexible
spline has two fewer teeth than the internal circular spline. An elliptical cam enclosed in an antifriction
ball bearing assembly is mounted inside the flexible cup and forces the flexible cup splines to push
deeply into the rigid ring at two opposite points while rotating. The two contact points rotate at a speed
governed by the difference in the number of teeth on the two splines. This method basically preloads
the teeth, which reduces backlash.

roymech.co.uk



Harmonic Gear Sketch

roymech.co.uk




148

The wave generator is attached to the input shaft, the flexible spline is attached to the output
shaft, and the circular spline is fixed.



Harmonic Gear Diagram

roymech.co.uk


For harmonic gearing, the gear ratio is also calculated from the numbers of teeth in each gear.

IN WG FS
OUT FS FS CS
N
n
N N
e e
e e
= = =




where WG stands for wave generator, FS stands for flexible spline, and CS stands for circular spline.
For example, if N
FS
=200 and N
CS
=202, the gear ratio is

IN FS
OUT FS CS
200
100
200 202
N
n
N N
e
e
= = = =



which means that the output shaft rotates 100 times slower than the input shaft, but the output shaft
carries 100 times more torque than the input shaft. Therefore, this example would be good for the robot
joint case, i.e. reducing speed and increasing torque, with n >>1. The negative sign indicates the
angular velocity and torque of the output shaft are in the opposite direction of the angular velocity and
torque of the input shaft.



149

4.1.2 Gear Ratio


Classification of gear ratios


If 1 n >
2 1
2 1
e e
t t
<
>

The output has reduced speed and increased torque.

This is the electric motor / robot joint case, where n >>1.


If 1 n <
2 1
2 1
e e
t t
>
<

The output has increased speed and reduced torque.

This is the bicycle transmission case, except for some granny gears where n can be as
high as 1.5.


If 1 n =
2 1
2 1
e e
t t
=
=


This case is called an idler, where the output speed and torque are unchanged, but the
direction reverses (for external spur gears)




150

Gear ratio examples Bicycle Transmissions

Gear Ratio:
OUT R F R
IN F R F
N N
n
N N
e t
e t
= = = =

Cannondale M400 mountain bike
front teeth
N
i
44 32 22
11 0.25 0.34 0.50
12 0.27 0.38 0.55
14 0.32 0.44 0.64
16 0.36 0.50 0.73
rear teeth 18 0.41 0.56 0.82
21 0.48 0.66 0.95
24 0.55 0.75 1.09
28 0.64 0.88 1.27
32 0.73 1.00 1.45

The Cannondale mountain bike has a traditional front/rear derailleur transmission with three
gears in the front and nine gears in the rear, for a total combination of 27 gears. The standard BikeE
recumbent bike has a traditional derailleur transmission with seven gears only in the rear. Instead of a
traditional derailleur transmission in the front with three gears, the BikeE has an internal hub planetary
gear arrangement with three selectable ratios of 1.2913 : 1 (high), 1 : 1 (medium), and 0.7 : 1 (low). The
BikeE then has a single chain ring in front to drive the rear chain rings.

The standard BikeE original front single chain ring had 46 teeth I changed this to a smaller
front chain ring of 34 teeth for more granny gear in order to climb Mulligan Hill; the cost of this is loss
of high gear for the bike path. I designed the lowest gear to mimic the lowest Cannondale gear ratio
since I knew that granny gear climbed well. The Cannondale mountain bike has a standard wheel size
of 26 diameter and the BikeE has a rear wheel size of 20 diameter. We must consider this difference
in wheel sizes to compute the overall effective BikeE gear ratios. The table below reflects this
calculation.




151

BikeE standard recumbent bike
front teeth
34 34 34
N
i
1.2913 : 1 1 : 1 0.7 : 1
11 0.33 0.42 0.60
13 0.38 0.50 0.71
15 0.44 0.57 0.82
rear teeth 18 0.53 0.69 0.98
21 0.62 0.80 1.15
24 0.71 0.92 1.31
28 0.83 1.07 1.53


I was able to obtain a used deluxe BikeE recumbent bike from noted luthier Dan Erlewine. I
decided to keep the front chain ring of 46 teeth (numbers of teeth in the rear chain ring and the planetary
gear ratios in the rear hub are identical between the standard and deluxe BikeE models). This means my
new deluxe BikeE doesnt climb as well as my modified standard BikeE, but it flies much faster on the
bike path in high gear than the standard BikeE! Again, the difference in wheel diameter is taken into
account in the table below.

BikeE deluxe recumbent bike
front teeth
46 46 46
N
i
1.2913 1 0.7
11 0.24 0.31 0.44
13 0.28 0.37 0.52
15 0.33 0.42 0.61
rear teeth 18 0.39 0.51 0.73
21 0.46 0.59 0.85
24 0.53 0.68 0.97
28 0.61 0.79 1.13


We see that the standard BikeE that was designed to equal the granny gear of the Cannondale (it
was exceeded, 1.53 vs. 1.45). However, the mountain bike still climbs better in granny gear, since your
legs are positioned above the pedals in the mountain bike case, and your legs are positioned straight out
in front of you in the recumbent bike case.


Unlike the robot joint example, bicycle gearing generally has n <1 and so the transmission
- increases angular velocity
- decreases torque
by the gear ratio n. The exception is the granny gears with n >1.



152

4.1.3 Gear Trains

To obtain a higher gear ratio than practical with a single pair of standard involute spur gears, one
can mate any number of spur gears in a gearbox, or gear train. The leftmost gear is the driving gear and
the rightmost is the output gear. All intermediate gears are first the driven gear and then the driving gear
as we proceed from left to right. Let us calculate the overall gear ratio n
GT
.
IN
GT
OUT
n
e
e
=
Example







We can find the overall gear ratio by canceling neighboring intermediate angular velocities.






Each term in the above product may be replaced by its known number of teeth ratio.






All intermediate ratios cancel, so





We could have done the same with pitch radii instead of number of teeth because they are in direct
proportion.





So, the intermediate gears are idlers. Their number of teeth effect cancels out, but they do change
direction. We should have included the +/ signs, by inspection. For gear trains composed of
externally-meshing spur gear pairs:
odd number of gears the output is in the same direction as the input
even number of gears the output is in the opposite direction as the input


153

Improved gear train
That gear train concept did not work. Now let us mate any number of spur gears, where the
driving and driven gears are distinct, because each pair is rigidly attached to the same shaft. Again, let
us calculate the overall gear ratio.
IN
GT
OUT
n
e
e
=
Example









Again, we use the equation






But now the gears rigidly attached to the same shaft have the same angular velocity ratio, so






The general formula for this case is






Again, we must consider direction





For gear trains composed of externally-meshing spur gear pairs:
odd number of pairs the output is in the opposite direction as the input
even number of pairs the output is in the same direction as the input



154

4.1.4 Involute Spur Gear Standardization

Rolling Cylinders

Mating spur gears are based on two pitch circles rolling without slip. These are fictitious circles,
i.e. you cannot look on a gear to see them. The actual gear teeth both roll and slide with respect to each
other (via the two-dof gear joint).


Fundamental Law of Gearing

The angular velocity between the gears of a gearset must remain constant throughout the mesh.

From our study of linkage velocity, we know this is no easy feat. Velocity ratios in a linkage
vary wildly over the range of motion.

Velocity Ratio
1
VR
OUT IN
IN OUT
r
r n
e
e
= = =

Torque Ratio TR
IN OUT
OUT IN
r
n
r
e
e
= = =

The velocity ratio is the inverse of the gear ratio n and the torque ratio is the same as the gear
ratio n defined previously. The torque ratio is also called Mechanical Advantage (MA).


Involute Function

Standard spur gears have an involute tooth shape. If the gears center distance is not perfect
(tolerances, thermal expansion, wear in design the center distance is increased slightly by the engineer
to allow for these effects; this is called clearance), the angular velocity ratio will still be constant to
satisfy the Fundamental Law of Gearing.



155

The involute of a circle is a curve generated by unwrapping a taut sting from the circumference of a so-
called base circle, always keeping it tangent to the circle. The figure below has th = [0:5:60]*DR
and r
b
=1 m. The red circle is the gear base circle, the blue lines are the taut tangent construction lines,
and the green curve is the involute function.


Involute Function Construction

In polar coordinates (u, r), the parametric equations for the involute of a circle are given below.
tan
sec
b
t t
r r t
u =
=


where t is the independent parameter and r
b
is the base circle radius. In Cartesian coordinates (x
I
, y
I
):
cos sec cos(tan )
sin sec sin(tan )
I b
I b
x r r t t t
y r r t t t
u
u
= =
= =

-1 -0.5 0 0.5 1 1.5
-1
-0.5
0
0.5
1
X (m)
Y

(
m
)


156

Involute Function Example

The plot below shows the involute function for a circle of r
b
=1 m. The parameter range is t =
[0:1:80]*DR; near the circle the involute points are very close to each other and farther away the
step size increases dramatically.
The involute function is symmetric (try t = [0:5:360]*DR) but watch out for those
intermediate steps). Gear teeth only require the involute near the base circle, with two symmetric sides.



Involute Function of a Circle

-6 -5 -4 -3 -2 -1 0 1 2
-4
-3
-2
-1
0
1
2
3
4
X (m)
Y

(
m
)


157

The next three figures from Norton (2008)
3
show important gear geometry for planar spur gear
design and standardization. In these figures, the pink pinion is the driving or input gear, and the gray
gear is the driven or output gear.



Norton (2008)

Base Circle The involute function starts from this circle.

Pitch Circle A fictitious circle (you cannot see it on a spur gear) with theoretical pure
rolling in contact between two cylinders of the pinion and gear.

Pitch Point The contact point between the two pitch circles.

Pressure Angle The angle between the common normal (also called axis of transmission)
of the two meshing teeth and the velocity of the pitch point (the tangent to
both pitch circles). The point of contact slides along this line. A similar
angle is defined for cams and followers.


The relationship among the base circle radius r
b
, pitch circle radius r
p
, pressure angle | is

cos
b p
r r | =


3
R. Norton, 2008, Design of Machinery: An Introduction to the Synthesis and Analysis of Mechanisms and Machines, McGraw-Hill.


158


Norton (2008)

The length of contact is measured along the axis of transmission. The beginning of contact is
when the tip of the driven gear tooth intersects the axis of transmission. The end of contact is when the
tip of driving gear tooth intersects the axis of transmission, as shown in the figure above. Only one or
two teeth are in contact at any one time for standard spur gears.

For harmonic gearing, many teeth are in contact at any one time, which provides a higher gear
ratio in a smaller package.


159


Norton (2008)

Increasing the center distance increases the pressure angle and increases the pitch circle
radii, but doesnt change the base circles (obviously the gears are made based on their own constant
r
b
). Thanks to the involute tooth shape, increasing the center distance does not affect the angular
velocity ratio. This is why the involute function is so widely used in spur gears.

The relationship from before still applies with an increase in center distance.

cos
b p
r r | =

Again, r
b
is fixed, and r
p
and | both increase the cosine function maintains the constant r
b
.




160

Spur Gear Standardization

Gear standardization is used to allow interchangeability in manufacturing and to allow meshing
of different size gears (different pitch radii and number of teeth) to achieve desired gear ratios. For two
spur gears to mesh, they must have

1) the same pressure angle (see previous figures and definition)

2) the same diametral pitch (see the equation below)

3) standard tooth proportions (see the figure below)


diametral pitch
d
N
p
d
=

Where N is the number of teeth and d is the pitch diameter, both for each gear.


module
1
d
d
m
N p
= =

Module is the SI version of diametral pitch (it is the inverse). SI gears are not
interchangeable with English system gears because of different tooth proportion
standards.


circular pitch
c
d
p
N
t
=

Circular pitch is the circumferential distance (arc length) between teeth along the pitch
circle of a spur gear.



161


from Norton (2008)


Standard involute tooth proportions

- Addendum is radial distance from pitch circle to top land of tooth.
- Dedendum is radial distance from pitch circle to bottom land of tooth (not to the base circle).
- Clearance is radial distance from bottom land to mating gear top land (radial backlash).
- Face width is thickness of tooth and gear (mating widths neednt be the same).
- Tooth thickness t is the circumferential arc length of each tooth. It is related to the circular
pitch p
c
and backlash (next page) b by

2
c
p t b = +





162

Backlash

Backlash B is the distance between mating teeth measured along the pitch circle circumference.
Backlash can be thought of as circumferential clearance. All real-world gears must have some backlash
in order to still function despite real-world problems of manufacturing tolerances, thermal expansion,
wear, etc. However, one must minimize backlash for smooth operation. For example, robot joints must
be driven both directions. Upon changing direction, nothing happens until the backlash is passed, and
then an impact occurs, which is bad for gear dynamics. This is a non-linear effect in robotics. On earth
gravity tends to load the backlash for predictable effects. In space however, the backlash is less
predictable.








163

4.1.5 Planetary Gear Trains

Planetary gear trains are also called epicyclic gear trains. The sun gear rotates about a fixed axis.
Each planet gear rotates about its own axis and also orbits the sun gear. This happens with direct
meshing of teeth, unlike celestial planetary motion. The arm link (which is a rigid body with no teeth)
carries the planet(s) around the sun. The arm has a revolute joint to the sun gear on one side and another
revolute joint to the planet gear(s) on the other side. Planetary gear trains have two-dof, so two inputs
must be given to control the mechanism. For instance, one can drive the sun gear and the arm link with
independent external motors. Alternatively, the sun gear may be fixed and the arm link driven. Using
planetary gear trains, one can obtain a higher gear ratio in a smaller package, compared to non-planetary
gear trains.



Conventional Gear Pair Planetary Gear Train

3(3 1) 2(2) 1(1) 1 M = = 3(4 1) 2(3) 1(1) 2 M = =


IN OUT OUT
OUT IN IN
r N
n
r N
e
e
= = = ?
IN
OUT
n
e
e
= =

We present the tabular method below to determine the gear ratio for various planetary gear trains.




164

Planetary Gear Train Applications



Airplane Propeller Transmission

Planetary gear train applications include airplane propellers, some automotive transmissions and
differentials, machine tools, hoists, and a hub-enclosed multi-speed bicycle transmission.




Old-Fashioned Pencil Sharpener

appauto.files.wordpress.com




165



Planetary Gear Box in Stages

This is a commercial planetary gear set with 8 possible ratios (4:1, 5:1, 16:1, 20:1, 25:1, 80:1,
100:1, 400:1).

servocity.com


166

Table Method Analysis

First, consider a simplified planetary gear system with 0
S
e = . With the sun gear as the fixed
link, we have a 1-dof system. Given the input e
A
, calculate the absolute value of e
P
. The Table
Method is based on the following relative velocity equation.

/ G A G A
e e e = +

This is a vector equation, but since all rotations are about the Z axis, we just use the magnitudes and +
for CCW and for CW, according to the right-hand-rule. This equation is written for each gear in the
system (replace G with the appropriate index). A stands for arm. ,
G A
e e are the absolute angular
velocities of a gear and the arm link.
/ G A
e is the relative velocity of a gear with respect to the moving
arm link. Construct a table as below; each row is the relative equation written for a different gear.

G e
G

=
e
A
+
e
G/A

S
P


First, fill in the given information.

G e
G

=
e
A
+
e
G/A

S 0
e
A

P
e
A



Each row must add up according to the relative velocity equation

G e
G

=
e
A
+
e
G/A

S 0
e
A

e
A

P
e
A



Now we can fill in down the relative column, using a simple gear ratio (relative to the arm).
/
/
P A S
S A P
N
N
e
e
=
/ /
S S
P A S A A
P P
N N
N N
e e e = =

G e
G

=
e
A
+
e
G/A

S 0
e
A

e
A

P
e
A


S
A
P
N
N
e



167

The last row must add up according to the relative equation, to finish the table.

G e
G

=
e
A
+
e
G/A

S 0
e
A

e
A

P
1
S
A
P
N
N
e
| |
+
|
\ .


e
A


S
A
P
N
N
e

Therefore, the absolute angular velocity of the planetary gear is 1
S
P A
P
N
N
e e
| |
= +
|
\ .
. Since the sign is
positive, it has the same direction as e
A
(CCW).


Calculate Effective Gear Ratio

1
1
IN A P
S
OUT P P S
P
N
n
N
N N
N
e e
e e
= = = =
+
+



Example

10
S
N = 40
P
N = 0
S
e = 100
A
e = + rpm, CCW

10
1 100 125
40
P
e
| |
= + =
|
\ .
rpm, CCW

40
0.8
40 10
P
P S
N
n
N N
= = =
+ +


Check
A
P
n
e
e
=
100
125
0.8
A
P
n
e
e = = =


So we see that with 0
S
e = , the gear ratio is not higher than the conventional gear train.

40
10
P
S
N
n
N
= = .

Let us include 0
S
e = next.



168

Table Method Analysis Now we present a more general system with 0
S
e = .

1) The given information is starred (*).

2) The S equation must add up; therefore
/ S A S A
e e e = .

3) Fill down the right column using a simple gear train, relative to arm A.
/
/
P A S
S A P
N
N
e
e
=
/ /
( ) ( )
S S S
P A S A S A A S
P P P
N N N
N N N
e e e e e e = = =

4) The P equation must add up; therefore:
( ) 1
S S S
P A A S A S
P P P
N N N
N N N
e e e e e e
| |
= + = +
|
\ .
.

G e
G

=
e
A
+
e
G/A

S * e
S
* e
A

S A
e e

P
1
S S
A S
P P
N N
N N
e e
| |
+
|
\ .

* e
A

( )
S
A S
P
N
N
e e



Examples 0
S
e =

1) 10
S
N = 40
P
N = 100
S
e = rpm, CW 100
A
e = + rpm, CCW
( )
10 10
1 100 100 125 25 150
40 40
P
e
| |
= + = + = +
|
\ .
rpm, CCW Still not a big ratio.


2) 40
S
N = 10
P
N = 100
S
e = rpm, CW 100
A
e = + rpm, CCW
40 40
1 100 ( 100) 500 400 900
10 10
P
e
| |
= + = + = +
|
\ .
rpm, CCW Thats a big ratio.


3) 40
S
N = 10
P
N = 125
S
e = + rpm, CCW 100
A
e = + rpm, CCW
40 40
1 100 (125) 500 500 0
10 10
P
e
| |
= + = =
|
\ .

In Example 3, the Sun and Arm rotational velocities cancel so the absolute angular velocity of the Planet
is zero.


169

Table Method Analysis

Now let us consider an even more general system with 0
S
e = , also adding an internal-teeth ring
gear.


1) The given information is starred (*).

2) The first two rows are identical to the case above.

3) Fill down the right column using a simple gear train, relative to arm A.
/
/
R A P
P A R
N
N
e
e
= +
/ /
( ) ( )
S S P P
R A P A A S A S
R R P R
N N N N
N N N N
e e e e e e = + = + =

4) The R equation must add up; therefore:
( ) 1
S S S
R A A S A S
R R R
N N N
N N N
e e e e e e
| |
= + = +
|
\ .
.


G e
G

=
e
A
+
e
G/A

S * e
S
* e
A

S A
e e

P
1
S S
A S
P P
N N
N N
e e
| |
+
|
\ .

* e
A

( )
S
A S
P
N
N
e e

R
1
S S
A S
R R
N N
N N
e e
| |
+
|
\ .

* e
A

( )
S
A S
R
N
N
e e





170

Examples 0
S
e = plus ring gear

1) 10
S
N = 40
P
N = 100
R
N = 100
S
e = rpm, CW 100
A
e = + rpm, CCW

10 10
1 100 ( 100) 110 10 120
100 100
R
e
| |
= + = + = +
|
\ .
rpm, CCW


2) 40
S
N = 20
P
N = 80
R
N = 100
S
e = rpm, CW 200
A
e = rpm, CW

1
S S
P A S
P P
N N
N N
e e e
| |
= +
|
\ .


40 40
1 ( 200) ( 100) 600 200 400
20 20
P
e
| |
= + = + =
|
\ .
rpm, CW


1
S S
R A S
R R
N N
N N
e e e
| |
= +
|
\ .


40 40
1 ( 200) ( 100) 300 50 250
80 80
R
e
| |
= + = + =
|
\ .
rpm, CW




171

4.2 Cams

4.2.1 Cam Introduction

Applications




Compared to linkages, it is easier to design desired motion with cams, but it is more expensive and
difficult to produce. Also, the cam contact and wear properties are worse than for linkages.

Cam Classification

Mabie & Reinholtz (1987
4
)

a) disk cam with knife-edge
follower
b) disk cam with radial roller
follower
c) disk cam with offset radial
roller follower
d) disk cam with oscillating
roller follower
e) disk cam with radial flat-faced
follower
f) disk cam with oscillating flat-
faced follower


4
H.H. Mabie and C.F. Reinholtz, 1987, Mechanisms and Dynamics of Machinery, Wiley.


172

Disk cams with followers



Norton (2008)





Norton (2008)




173



Norton (2008)

Two cam/follower systems are shown above with equivalent four-bar and slider-crank
mechanism models.

Caution
These are instantaneous equivalents only, i.e. the virtual link lengths for both the four-bar and
slider-crank models change with cam mechanism configuration.


Degrees of Freedom (mobility)
A cam joint is a J
2
, i.e. it has two-dof since it allows both rolling and sliding, like a gear joint.









Function Generation
In function generation, the output parameter is a continuous function of the input parameter in a
mechanism. With linkages, we can only satisfy a function exactly at a finite number of points: 3, 4, or 5,
usually. For example, for a four-bar linkage with
( )
4 2
f u u = , this function is only exact at a few points.

With a cam and follower mechanism, however, we can satisfy function generation at infinite
points. u is the cam input angle and the output is S for a reciprocating (translating) follower and the
output is | for an oscillating (rotating) follower.
( ) S f u = ( ) f | u =




174

4.2.2 Cam Motion Profiles

Up to this point, we have been mostly concerned with mechanism analysis: given a mechanism
design and its input parameters, determine the position, velocity, acceleration, and dynamics behavior.
With cams we must consider mechanism synthesis for the first time: given the motion requirements
(follower motion and timing with the input cam angle), design the cam. The first step is to determine a
smooth cam follower motion profile. In general a cam follower has 4 motion zones (rise, dwell, fall,
dwell), as shown below.


















When the motion transitions between different motion functions, we must ensure smooth motion.


Fundamental Law of Cam Design

The cam function must be continuous through the first and second derivatives of displacement
across the entire motion interval.

Which means:

Position, velocity, and acceleration must be continuous for the entire 360

of cam rotation. The


jerk function (the derivative of the acceleration) must be finite, but need not be continuous.


If the Fundamental law of Cam Design is satisfied, the resulting dynamic performance will be
acceptable for high-speed cam/follower operation. If not, there will be performance degradation due to
noise, vibrations, high wear, etc. There is a cyclical impulse hammering at each point in the cam cycle
when acceleration is not continuous (even worse if position or velocity is not continuous).



175

S V A J Diagrams

In cam synthesis (design), we are only given the total motion range and perhaps some timing
requirements. It is the engineers job to determine the position curves and to match the velocity and
acceleration across junctions. Position is automatically matched by shifting the dependent function
axes. Draw S V A J diagrams vs. time to graphically see if the Fundamental Law of Cam Design is
satisfied for candidate curves. We can plot vs. time or vs. input cam angle u (assuming constant angular
velocity, t u e = ).

The slope of a function is the value of its derivative at a point in time (or u). Therefore, for
continuous velocity and acceleration curves, the slopes of the position and velocity curves must match
across all junctions. The slope of the acceleration can be discontinuous (leading to finite jumps in jerk),
but the acceleration itself must be continuous.

Cam motion curves are very much like the input link motion curves discussed earlier, for input
links that start and stop at zero velocity and acceleration. In fact, I adapted the input motion curves from
cam motion curve design.

Generic Cam-Follower Motion Profile Figure














Define each separate function so the value is zero at the initial angle, which is zero. Then to put the
whole cam motion profile together, just shift the u and S axes.


Match S easy, just shift the S axis.


Match V slope of S must match across junctions.
1 1
( ) ( 0)
i i i i i
v v u | u
+ +
= = =
apply to all functions / junctions.

Match A slope of V must match across junctions.
1 1
( ) ( 0)
i i i i i
a a u | u
+ +
= = =
apply to all functions / junctions.



176

Cam Follower Motion Profile Examples Example 1

rise dwell portion. Specify parabolic (constant acceleration) to straight line (constant velocity) rise,
followed by a dwell.


parabolic function constant velocity function dwell

S:
2
1 1 0 1
1
( )
2
f A u u =
2 2 0 2
( ) f V u u =
3 3
( ) 0 f u =


V:
1 1 0 1
( ) v A u u =
2 2 0
( ) v V u =
3 3
( ) 0 v u =


A:
1 1 0
( ) a A u =
2 2
( ) 0 a u =
3 3
( ) 0 a u =


J:
1 1
( ) 0 j u =
2 2
( ) 0 j u =
3 3
( ) 0 j u =


Match S at junction B just shift the vertical axis up.


Match V at junction B

1 1 1 2 2
( ) ( 0) v v u | u = = =
0 1 0
A V | = so
0 0 1
V A | =


Try to match A at junction B:

1 1 1 2 2
( ) ( 0) a a u | u = = =
0
0 A =


0
0 A = is impossible, or else the parabola is degenerate, which we cannot allow. This case violates the
Fundamental Law of Cam Design since the acceleration function cannot be made to be continuous at
junction B. Therefore, this cam motion profile example cannot be used for cam design.

We have a bigger problem at junction C, between functions 2 and 3: the velocity function cannot
be made to be continuous at junction C. Discontinuous velocity is one level worse than discontinuous
acceleration; either renders the resulting cam motion profile unacceptable.

Example 1 Plots



177

Cam Follower Motion Profile Examples Example 2
Let us fix the rise portion only, at junction B. Then the problem at junction C can be fixed using
symmetry. We specify a half-cycloidal function (sinusoidal in cam angle) to a straight line (constant
velocity) rise.
half-cycloidal function constant velocity function
S
1 1
1 1 1
1 1
1
( ) sin f L
u tu
u
| t |
| |
=
|
\ .

2 2 0 2
( ) f V u u =
V
1 1
1 1
1 1
( ) 1 cos
L
v
tu
u
| |
| |
=
|
\ .

2 2 0
( ) v V u =
A
1 1
1 1 2
1 1
( ) sin
L
a
t tu
u
| |
| |
=
|
\ .

2 2
( ) 0 a u =
J
2
1 1
1 1 3
1 1
( ) cos
L
j
t tu
u
| |
| |
=
|
\ .

2 2
( ) 0 j u =

Match S at junction B just shift the vertical axis up
Match V at junction B

1 1 1 2 2
( ) ( 0) v v u | u = = =
1 1
0
1 1
1 cos
L
V
t|
| |
| |
=
|
\ .
so
1
0
1
2L
V
|
=
Match A at junction B

1 1 1 2 2
( ) ( 0) a a u | u = = =
1 1
2
1 1
sin 0
L t t|
| |
| |
=
|
\ .
0 =0
In this case the acceleration function is continuous because the half-cycloidal function ensures
that the acceleration is zero at the end of the function range. This case obeys the Fundamental Law of
Cam Design and so this cam motion profile example portion can be used for cam design.

Half-Cycloidal Rise (to connect with constant velocity)
0 20 40 60 80
0
1
2
S

(
m
)
Half Cycloid
0 20 40 60 80
0
0.05
V

(
m
/
d
e
g
)
0 20 40 60 80
0
5
x 10
-4
A

(
m
/
d
e
g
2
)
0 20 40 60 80
-2
0
2
x 10
-5
u (deg)
J

(
m
/
d
e
g
3
)


178

Cam Follower Motion Profile Examples Example 3
We now specify a full-cycloidal function (sinusoidal in cam angle). This will rise all the way to
meet a dwell smoothly; it satisfies the Fundamental Law of Cam Design. This is the same function
used in term project input link motion specification earlier, when starting at stopping at rest.
S
1 1
1 1 1
1 1
2 1
( ) sin
2
f L
u tu
u
| t |
| |
=
|
\ .

2 2
( ) 0 f u =
V
1 1
1 1
1 1
2
( ) 1 cos
L
v
tu
u
| |
| |
=
|
\ .

2 2
( ) 0 v u =
A
1 1
1 1 2
1 1
2 2
( ) sin
L
a
t tu
u
| |
| |
=
|
\ .

2 2
( ) 0 a u =
J
2
1 1
1 1 3
1 1
4 2
( ) cos
L
j
t tu
u
| |
| |
=
|
\ .

2 2
( ) 0 j u =
The full-cycloidal function plots are shown below, rising through the displacement to connect with a
dwell.

Full-Cycloidal Rise (to connect with a dwell)

Note that the derivatives above are with respect to u (deg). To find the time derivatives, use the chain
rule (e.g., for velocity, multiplying by ( ) ( ) d t dt t u u e = =

, a constant). However, for a constant e, it is


customary to use the derivatives with respect to u for cam design. The full-cycloidal function matches
the ensuing dwell: the displacement functions are made to match, and the velocity and acceleration are
zero at the end of the full-cycloidal function and the start of the ensuing dwell. The jerk does not match,
but the discontinuity in jerk is finite, which satisfies the Fundamental Law of Cam Design.

0 20 40 60 80
0
1
2
S

(
m
)
Full Cycloid
0 20 40 60 80
0
0.05
V

(
m
/
d
e
g
)
0 20 40 60 80
-2
0
2
x 10
-3
A

(
m
/
d
e
g
2
)
0 20 40 60 80
-2
0
2
x 10
-4
u (deg)
J

(
m
/
d
e
g
3
)


179

4.2.3 Analytical Cam Synthesis

Disk Cam with Radial Flat-Faced Follower

Assume a valid cam motion profile has been designed according to the Fundamental Law of
Cam Design; i.e. we now have continuous S, V, A curves. Given the motion profile found by the
engineer, now we must determine the cam contour.

Is it as simple as polar-plotting S =f(u) vs. cam angle u ? No that approach would not account
for the face width of the cam follower, i.e. the contact points are not along radial lines in general. We
will use kinematic inversion to simplify the synthesis process.

DCRFFF Figure



As seen in the figure, the radius R out to the flat-faced follower (not to the point of contact x, y) is:
( ) R C f u = +

where C is the minimum cam radius, a design variable, and S =f(u) is the given cam motion profile.
The radius R and the flat-faced follower length L can be related to the contact point x, y and the cam
angle through geometry.


180

cos sin
sin cos
R x y
L x y
u u
u u
= +
= +


Notice that
sin cos
dR
x y L
d
u u
u
= + =

( ( ))
d df
L C f
d d
u
u u
= + =


To calculate the follower flat-face length, double the maximum of L from above. It is doubled because
by symmetry the contact point will change to the other side at 180 u =

.

To summarize thus far:
( ) R C f u = +
df
L
du
=


This is sufficient to manufacture the cam since it is machined with u, R, L coordinates. If we want to
know the cam contour in Cartesian coordinates, we must solve the relationships for x, y. In matrix form:

cos sin
sin cos
x R
y L
u u
u u
(
=
` `
(

) )



This special coefficient matrix [A] is orthonormal, which means both columns and rows are
perpendicular to each other and both columns and rows are unit vectors. One unique property of
orthonormal matrices is | | | |
1 T
A A

= . The Cartesian cam contour solution is thus:



cos sin cos sin
sin cos sin cos
x R R L
y L R L
u u u u
u u u u

(
= =
` ` `
(
+
) ) )



( )
( )
( )cos sin
( )sin cos
df
x C f
d
df
y C f
d
u u u
u
u u u
u
= +
= + +




181

Minimum Cam radius to Avoid Cusps

A cusp is when the cam becomes pointed or undercut. Clearly, this must be avoided for good
cam motion. The cusp condition is that for a finite u A , there is no change in x, y.

0
dx dy
d d u u
= = will cause a cusp.


( )
( )
2
2
2
2
( )sin cos cos sin
( )cos sin sin cos
dx df df d f
C f
d d d d
dy df df d f
C f
d d d d
u u u u u
u u u u
u u u u u
u u u u
= + +
= + + +



( )
( )
2
2
2
2
sin
cos
dx d f
C f
d d
dy d f
C f
d d
u u
u u
u u
u u
| |
= + +
|
\ .
| |
= + +
|
\ .



0
dx dy
d d u u
= = occurs simultaneously only when
2
2
( ) 0
d f
C f
d
u
u
+ + =

Therefore, to avoid cusps on the entire cam contour, we must ensure that
2
2
( ) 0
d f
C f
d
u
u
+ + >

Note that C is always positive and f(u) starts and ends at zero and never goes negative. So the sum of
these positive terms and the sometimes-negative second derivative of the cam motion profile must
always be greater than zero to avoid cusps or undercutting in the practical cam you are designing.



182

Disk Cam with Radial Flat-Faced Follower Design Example

Specify a full-cycloidal rise with a total lift of 50 mm, followed by a high dwell, a symmetric
full-cycloidal return with a total fall of 50 mm, and then a low dwell. Each of these four motion steps
occurs for 90

of cam shaft rotation.



The full-cycloidal rise and fall cam motion profile associated with this specification is shown
below. Clearly, this satisfies the Fundamental Law of Cam Design because the position, velocity, and
acceleration curves are continuous. The jerk is not continuous, but it remains finite over all cam angles.



0 50 100 150 200 250 300 350
0
10
20
30
40
50
S
0 50 100 150 200 250 300 350
-50
0
50
V
0 50 100 150 200 250 300 350
-100
-50
0
50
100
A
0 50 100 150 200 250 300 350
-500
0
500
u
J


183

Choosing a minimum cam radius of C =100 mm, the resulting cam contour is shown below.



Cam Cartesian Contour

-150 -100 -50 0 50 100
-100
-50
0
50
100
150
X (mm)
Y

(
m
m
)


184

Let us check the cusp avoidance plot. To avoid cusps in this cam, we require that

2
2
( ) ( ) ( ) 0
d f
C S A C f
d
u u u
u
+ + = + + >

As seen in the plot below, this inequality is satisfied for the entire range of motion, so this cam design is
acceptable with respect to avoiding cusps and undercutting.



Cam Cusp Avoidance Plot

0 50 100 150 200 250 300 350
0
50
100
150
200
250
u
C
+
S
+
A


185



Cam/Follower Animation Snapshot for
2
60 u =

(CW)


-200 -150 -100 -50 0 50 100 150 200
-200
-150
-100
-50
0
50
100
150
200
X (m)
Y

(
m
)


186

5. Mechanical Vibrations Introduction

5.2 Mechanical Vibrations Definitions

This section presents some important definitions for mechanical vibrations concepts, used
throughout the ME 3011 and ME 3012 NotesBooks. You may already be familiar with some these
terms if some are unfamiliar, dont freak out, we will discuss them later.

mechanical vibrations Continual, periodic motions about an equilibrium point by a machine or
machine component, in some cases desired but in others unwanted.

degrees of freedom dof the number of independent parameters required to fully specify the
location of a device. The number of different ways something can move.

statics The study of forces/torques without regard to motion. The study of
force/moment equilibrium in structures.

kinematics The study of motion without regard to forces/torques.

dynamics The study of motion with regard to forces/torques.

free-body diagram FBD, a diagram drawn out of context for each separate mass or inertia
with all external and internal forces and moments shown to give context.

cycle A repeating unit of the vibration.

period T, time for one cycle, sec.

cyclical frequency f, number of cycles per second, the inverse of time period, 1 / T, Hz.

circular frequency 2 f e t = , scalar measure of rotational rate, i.e. the magnitude of the
angular velocity vector, rad/sec.

natural frequency The rate at which an unforced, undamped vibratory system tends to
vibrate, either cyclical frequency f
n
, or circular frequency e
n
.

spring Idealized massless mechanical element that provides the oscillatory
motion in a vibrational system.

dashpot Idealized massless mechanical element that provides the energy
dissipation in a vibrational system.

mass Idealized mechanical element that provides the inertia in a vibrational
system.

mass moment of inertia Idealized mechanical element that provides the rotational inertia in a
vibrational system.



187

damping A linear model for energy loss due to friction and other energy dissipaters,
for either translational or rotational motion. Damping is provided by a
virtual viscous dashpot (an automotive shock system is a real example).

amplitude The maximum magnitude of a sinusoidal oscillation.

phase angle Angle in a single sinusoidal function that indicates the amount of lead or
lag of a cosine or sine wave relative to the standard cosine or sine wave.

differential equation An equation in which the unknown appears with its time derivatives.

initial conditions Given values at time t =0 for the output and its time derivatives. There
must be as many initial conditions as the order of the differential equation.

input The external forcing function that drives a dynamic system.

output The variable of interest in motion of a dynamic system.

free vibrations Vibrations in the absence of a forcing function.

homogeneous solution Transient solution due to initial conditions and the external input forcing
function, yielding zero in the ODE.

transient vibrations Vibrations that disappear given enough time.

forced vibrations Vibrations due to an external input forcing function.

particular solution Steady-state solution due to the external input forcing function. There is
also a transient solution due to the external input forcing function.

steady-state vibrations Vibrations that do not disappear as time increases.

forcing function The input actuation provided by an external source.

driving frequency The rate at which an external input forces a vibratory system, either
cyclical f, or circular e.

damped frequency The rate at which an unforced, damped vibratory system tends to vibrate,
either cyclical f
d
, or circular e
d
.

logarithmic decrement The natural logarithm of the ratio of two consecutive points, separated by
the damped time period T
d
, on a unforced underdamped vibratory wave.

beat frequency Vibrational phenomenon wherein two frequencies in a dynamic system are
close in value, causing a vibration within a vibrational envelope.

resonance This condition occurs when the driving frequency e is the same as, or
close to, the systems natural frequency e
n
. In this case the resulting
amplitude of the steady-state vibration becomes very large.


188


frequency response The plots of system amplitude and phase angle vs. independent variable
input frequency.

critical shaft speed The angular shaft velocity that excites the natural frequency of a rotating
dynamic system. When driven at the critical speed, resonance occurs, a
condition generally to be avoided at all costs.

transmissibility The ratio of the maximum force transmitted to the base over the maximum
input force magnitude in a vibratory dynamic system.

vibration mode Characteristic shape at which a dynamic system will vibrate. Multi-dof
systems have multiple modes that sum to yield the total time response.


A linear system is one in which all governing equations (differential, algebraic) are linear.
Linear systems satisfy the principles of linear superposition and homogeneity. Let u(t) be the input
and y(t) be the output. Further, ( ) ( ) u t y t indicates a system yielding output y(t) given input u(t).

1) linear superposition

if
1 1
( ) ( ) u t y t and
2 2
( ) ( ) u t y t then
1 2 1 2
( ) ( ) ( ) ( ) u t u t y t y t + +


2) homogeneity

if ( ) ( ) u t y t then ( ) ( ) u t y t | |

where | is any constant.



189

6. Vibrational Systems Modeling

6.1 Zeroth-Order Systems

Calculation of Spring Constants

Helical Coiled Spring. The jagged springs indicated so far represent physical translational
springs coiled from an elastic wire into a helical shape about a central axis. For this physical spring,
from Strength of Materials, the linear spring constant k for small displacements is given below.



4
3
8
Gd
k
nD
=

where
G material shear modulus N/m
2
d wire diameter m
n number of spring coils unitless
D diameter of spring coiling m


This type of coiled physical spring is common; however, there are many mechanical systems
without a clear spring such as above. Instead, their elastic properties are provided by distributed
material in the system. Here we present several such distributed springs that are modeled as linear
springs for small displacements, as follows.


For the following 4 translational springs, the following method is used. For each given physical
situation, we find the displacement o under an applied force F, from Strength of Materials. Then the
spring stiffness constant k is found from Hookes Law F =ko: k =F / o. Below we assume that the force
is applied and the displacement measured at the end of the cantilevered beam and in the center of the
other beams.



f(t)
d
D
f(t)


190

Cylindrical Rod Spring. An elastic rod has an equivalent spring constant along the longitudinal
axis of the rod. From Strength of Materials, the linear spring constant k for small displacements is given
below.


EA
k
L
=
where
E material Youngs modulus N/m
2
A rod cross-sectional area,
2
r t
m
2

r radius of rod m
L length of rod m

So the cylindrical rod spring constant for a constant circular cross section becomes
2
E r
k
L
t
=

If the elastic rod is tapered with different radii r
1
and r
2
on opposite ends of the rod, the equivalent
spring constant is
1 2
E rr
k
L
t
=

Cantilever Beam Spring. A rectangular cantilevered beam has an equivalent spring constant
for small vertical displacements. From Strength of Materials, the linear spring constant k for small
displacements is:

3
3EI
k
L
=
where
E material Youngs modulus N/m
2
I beam area moment of inertia,
3
12 bh
m
4

b beam width m
h beam height m
L length of beam m

So the cantilevered beam spring constant becomes
3
3
4
Ebh
k
L
=

f(t)
r
L
f(t)
f(t)
r
L
f(t)
1
r
2
m
L x(t)
h


191

Doubly-Simply-Supported Beam Spring. A rectangular beam simply-supported at both ends
has an equivalent spring constant for small vertical displacements. From Strength of Materials, the
linear spring constant k for small displacements is given below.

3
48EI
k
L
=

where
E material Youngs modulus N/m
2
b beam width m
h beam height m
L length of beam m

So the doubly-simply-supported beam spring constant becomes
3
3
4Ebh
k
L
=

This doubly-simply-supported beam is 16 times stiffer in vertical displacement than the cantilever beam.

Doubly-Fixed Beam Spring. A rectangular beam rigidly fixed at both ends has an equivalent
spring constant for small vertical displacements. From Strength of Materials, the linear spring constant
k for small displacements is given below.


3
192EI
k
L
=

where
E material Youngs modulus N/m
2
b beam width m
h beam height m
L length of beam m

So the doubly-fixed beam spring constant becomes
3
3
16Ebh
k
L
=

This doubly-fixed beam is 64 times stiffer in vertical displacement than the cantilever beam, and 4 times
stiffer than the simply-supported beam presented earlier.

m
L
x(t)
h
m
L
x(t)
h


192

Air Spring. An enclosed volume of air has an equivalent spring constant for small
displacements. From Fluid Mechanics for compressible air, the linear spring constant k for small
displacements is given below.



2
PA
k
V

=

where
ratio of specific heats unitless
P enclosed air pressure N/m
2
A enclosed air cross-sectional area m
2
V enclosed air volume m
3




These six spring constant examples are all for translational springs, so their spring constant k
units should all be N/m. The material Youngs Modulus E and Shear Modulus G for steel are:

9 2
200 GPa = 200 10 N/ m E ~

9 2
79 GPa = 79 10 N/ m G ~


Let us now check the units for the previous six spring constants.

4 4
3 2 3
Nm N
8 mm m
Gd
nD
=

2
2
Nm N
mm m
EA
L
=

4
3 2 3
3 Nm N
mm m
EI
L
=


3 3
3 2 3
4 N(m)m N
mm m
Ebh
L
=


3 3
3 2 3
16 N(m)m N
mm m
Ebh
L
=


2 4
2 3
Nm N
mm m
PA
V

=

All units are correct for a translational spring, i.e. N/m.

We have also used torsional springs in our mechanical systems models. Torsional springs can be
constructed of a coiled elastic material, spiraling in one plane. An elastic cylindrical rod also provides a
torsional spring effect when torque about the longitudinal axis of the rod. Additionally, a fixed-simply-
supported beam also provides a torsional spring effect.


f(t)


193

Torsional Rod Spring. An elastic rod has an equivalent spring constant about the longitudinal
axis of the rod. From Strength of Materials, the torsional spring constant k
R
for small angular
displacements is given below.


4
2
R
G r
k
L
t
=

where
G material shear modulus N/m
2
r radius of rod m
L length of rod m

If the torsional shaft is hollow, the torsional spring stiffness constant is

4 4
( )
2
o i
R
G r r
k
L
t
=

where r
o
and r
i
are the outer and inner shaft radii, respectively.


Fixed-Simply-Supported Torsional Beam Spring. A rectangular beam rigidly fixed at one end
and simply supported on the other has an equivalent torsional spring constant for small angular
displacements. From Strength of Materials, the torsional spring constant k
R
for small displacements is
given below.


4
R
EI
k
L
=

where
E material Youngs modulus N/m
2
b beam width m
h beam height or thickness m
L length of beam m

So the fixed-simply-supported torsional beam spring constant becomes
3
3
R
Ebh
k
L
=

r
L
(t) t (t) t
m
L
(t)
h
u


194

Let us now check the units for these torsional spring constants.

4 4
2
Nm Nm Nm
2 mm 1 rad
G r
L
t
= =

3 4
2
N m Nm Nm
= =
3 m m 1 rad
Ebh
L


Recall that rad doesnt count as a unit so we are free to include it above, yielding the correct SI units N-
m/rad for torsional spring constant.



Non-massless Spring. For translational mechanical systems in which the distributed mass of the
spring is significant and cannot be ignored, we use the following formula for the lumped system mass.

3
S
E
m
m m = + kg

Where m is the modeled point mass, m
S
is the spring mass, and m
E
is the calculated equivalent point
mass to include in the model with a massless spring. The spring constant k does not change and is
calculated as shown earlier in this subsection.

For the cantilever and simply-supported beam spring models, the equivalent mass is given below
(left and right below, respectively, and m
B
is the distributed beam mass, all kg).

0.23
E B
m m m = +

2
B
E
m
m m = +




195

Calculation of Damping Constants

In general the damping coefficient for a physical system is harder to identify and calculate than
the mass and spring constants. This subsection presents 4 equivalent damping coefficient cases. Many
physical systems have no identifiable damping dashpot, but a viscous damping coefficient c is included
to account for friction losses in the system, using a simple linear model (the law for linear dashpots
where the friction force is proportional to the velocity).

Also see the later section on logarithmic decrement, where the damping coefficient c can be
estimated from experimental observations.

Dashpot. The piston-based dampers indicated so far represent physical dashpots with a piston
moving (translating in a reciprocating fashion) in a viscous fluid to remove energy from the system,
such as in an automotive shock absorber. For this physical dashpot, from Fluid Mechanics, the linear
viscous damping coefficient c for small velocities is given below.



3
3
6
1
r h w
c
w r
tq | |
= +
|
\ .


where
q fluid viscosity N-s/m
2
r piston radius m
h piston height m
w wall thickness r
o
r m


f(t)
h
r
w


196

Coulomb damping. Coulomb damping arises from the so-called dry friction of sliding. This is
a nonlinear effect that can be approximated by the following equivalent linear viscous damping
coefficient c
eq
.



4
4
f
eq
f
N
c
B B

et et
= =


where
f
f
friction force N N
coefficient of friction unitless
N normal force N
B amplitude of vibration m
e frequency of vibration rad/s

Flat plates. When two flat plates move relative to each other, separated by a layer of viscous
fluid, damping arises. From Fluid Mechanics, the linear viscous damping coefficient c for small
velocities is given below.

eq
A
c
w
q
=
where
q fluid viscosity N-s/m
2
A area of smaller flat plate m
2
w viscous fluid layer thickness m

These three damping examples are all for translational dampers, so their viscous damping
coefficient c units should all be N-s/m. The SI units for fluid viscosity are N-s/m
2
and the Coulomb
friction coefficient is unitless. Let us now check the units for the previous three damping coefficients.

3 3
3 2 3
6 Ns mm m Ns
1
m m m m
r h w
w r
tq | | | |
+ =
| |
\ . \ .

4 N Ns
rad
m
m
s
N
B

et
=

2
2
Nsm Ns
m m m
A
w
q
=


All units are correct for a damping coefficient, i.e. N-s/m.

m
v(t)
f
f
f(t)
w


197

Rotational dashpot. We have also used torsional dashpots in our mechanical systems models.
Torsional dashpots can be constructed in a similar manner to a translational dashpot, but the
reciprocating motion is rotary about the center axis, instead of translational. From Fluid Mechanics, the
linear rotational viscous damping coefficient c
R
for small angular velocities is given below.



3
0
2
8
R
h r
c r
w h
tq
| |
= +
|
\ .


where
q fluid viscosity N-s/m
2

r piston radius m
h piston height m
w wall thickness r
o
r m
h
0
height from base to rotating piston m



(t)
h
r
w
t
h
0


198

Journal bearing. A journal bearing also has a fluid providing rotational damping. From Fluid
Mechanics, the linear rotational viscous damping coefficient c
R
for small angular velocities is given
below.



3
2
R
wr
c
tq
o
=

where
q fluid viscosity N-s/m
2

w base thickness m
r bearing radius m
o journal bearing fluid clearance m



Let us now check the units for these torsional damping coefficients.

3 3
2
0
Ns m Nms
2 m Nms
8 m m rad
h r
r
w h
tq
| |
| |
+ = =
| |
\ .
\ .

3 4
2
2 Nsm Nms
Nms
m m rad
wr tq
o
= =

Recall that rad doesnt count as a unit so we are free to include it above, yielding the correct SI units N-
m-s/rad for torsional damping coefficient.


r
w
o


199

6.2 Second-Order Systems

6.2.1 Translational m-c-k System Dynamics Model

Equivalent Springs

There is an alternate parallel springs case, shown in the diagram below.


1-dof Translational Mechanical System with Parallel Springs, Alternate


From the FBD (not shown), again there is a common displacement x(t). Assuming a small
positive displacement x(t), both spring restoring forces are negative (to the left). The equivalent spring
derivation is now presented.

1 2
1 2 1 2
( )
f f f
f k x k x k k x
= +
= + = +


Therefore, the equivalent spring constant for parallel springs case is again:

par 1 2
f
k k k
x
= = +

This result is identical to the original parallel springs case presented in the ME 3011
NotesBook.


The equivalent-spring formulas for 2, 3, and 4 springs in series are:

1 2
ser2
1 2
k k
k
k k
=
+

1 2 3
ser3
1 2 1 3 2 3
k k k
k
k k k k k k
=
+ +

1 2 3 4
ser4
1 2 3 1 2 4 1 3 4 2 3 4
k k k k
k
k k k k k k k k k k k k
=
+ + +


Actually, the original general formula below yields the same answers, with less computation:

ser
1
1
1
n
i i
k
k
=
=



x(t)
m
k k
1 2


200

6.2.3 Pendulum System Dynamics Model

Other Pendulum Models

Linearized ODE models can be derived that are very similar to the simple pendulum in the ME
3011 NotesBook, for the two pendulum models shown below. The figure on the left shows a pendulum
consisting of a slender rod of length L, with distributed mass m. The figure on the right shows a
pendulum consisting of a slender rod of length L, with distributed mass m
1
, plus an end-mounted point
mass of m
2
.


Two Additional Pendulum System Diagrams

Only the mass moment of inertia I
OZ
changes; for the cases on the left and right above, respectively, the
formulae are given below.
2
3
OZ
mL
I =

2
2 1
2
3
OZ
m L
I m L = +

The associated models are:
3
( ) ( ) 0
2
g
t t
L
u u + =


1 2
1 2
3( 2 )
( ) ( ) 0
2( 3 )
m m g
t t
m m L
u u
+
+ =
+



Where, for both slender rod cases, in the sum of moments we used the fact that the slender rod weight
(mg and m
1
g, respectively), acts at the slender rod CG, at a radius of L/2 from point O.

It is possible to account for friction in the simple pendulum system by including a rotational
viscous damping coefficient at the rotational bearing.

2
2
( ) ( ) ( ) 0
( ) ( ) ( ) 0
R
R
mL t c t mgL t
c g
t t t
mL L
u u u
u u u
+ + =
+ + =




And while not common, it is also possible to include an input forcing function, input torque t(t)
at the bearing.
g
m
L
O
(t) u
2
1
g
m
m
(t)
O
L
u


201

2 2
1
( ) ( ) ( ) ( )
R
c g
t t t t
mL L mL
u u u t + + =



Finally, if we model a horizontal pendulum, the restoring torque of gravity disappears and we
have the dynamic model of a gravity-neutral single rotating link.

2 2
1
( ) ( ) ( )
R
c
t t t
mL mL
u u t + =






202

6.2.4 Uniform Circular Motion
Uniform circular motion (UCM), wherein a point P at the end of a fixed-length link of length r
rotating in the XY plane about a fixed point O with a constant angular velocity e (see figure below)
generates Simple Harmonic Motion, like an unforced, undamped m-k translational mechanical system.


Uniform Circular Motion Diagram

In this UCM case no differential equations or solutions are required. Simultaneously simple
harmonic motion (SHM) is generated in the X with a pure cosine wave (with no phase angle) and in the
Y with a pure sine wave (with no phase angle); see the figure below, generated by UCM.m. Since
angular velocity e is constant, the angle linearly increases with time, ( ) t t u e = . With u
0
=0 at t =0, this
system will generate SHM in the X and Y directions indefinitely, repeating every 360 deg or 2t rad.

( ) cos ( ) cos
( ) sin ( ) sin
x t r t r t
y t r t r t
u e
u e
= =
= =




MATLAB UCM Animation

u
r
(t)
e
P
O
X
Y


203

6.4 Additional 1-dof Vibrational Systems Models

1-dof Vibrational Systems

Derive the vibrational equation of motion for each given mechanical system below.


1. m-c-k translational systems with parallel and series springs and dampers

The ME 3011 NotesBook presented the derivation of equivalent translational springs for the
parallel springs and series springs cases. Translational dashpots also follow the same formulas for their
respective parallel and series cases.

Considering the parallel-case figure shown below,


1-dof Translational Mechanical System with Parallel Springs and Dashpots

the system model is:

1 2 1 2
( ) ( ) ( ) ( ) ( ) ( ) mx t c c x t k k x t f t + + + + =




Considering the series-case figure shown below,


1-dof Translational Mechanical System with Series Springs and Dashpots

the system model is:

1 2 1 2
1 2 1 2
( ) ( ) ( ) ( )
c c k k
mx t x t x t f t
c c k k
| | | |
+ + =
| |
+ +
\ . \ .



x(t)
m
k k
1 2
c
c
1
2
f(t)
x(t)
m
k k
1 2
c
1
f(t)
c
2


204

2. J-c
R
-k
R
rotational systems with parallel and series springs and dampers

The same type of second-order models result for the rotational systems as in the translational
systems presented on the last page, with regard to parallel and series arrangements of rotational springs
and dashpots.

Considering the parallel case, the system model is given below.

1 2 1 2
( ) ( ) ( ) ( ) ( ) ( )
R R R R
J t c c t k k t t u u u t + + + + =





Considering the series case, the system model is given below.

1 2 1 2
1 2 1 2
( ) ( ) ( ) ( )
R R R R
R R R R
c c k k
J t t t t
c c k k
u u u t
| | | |
+ + =
| |
+ +
\ . \ .






205

Compound m-c-k translational systems with parallel and series springs and dampers

Certain real-world systems may require more complicated models with combinations of springs
and dampers in parallel and in series. There are infinite possible combinations of such models, dictated
by the real-world component arrangement. The next two sections present two such possibilities.

3. Parallel springs and dampers in series

The figure below shows two sets of parallel springs arranged in series. For clarity, the dashpots
are not shown but they have the same arrangement.



The second figure shows a simplified diagram for this case.



Using the parallel- and then serial-springs equivalent spring formulas derived in the ME 3011
NotesBook:

1 3
2 4
a
b
k k k
k k k
= +
= +

1 2 2 3 1 4 3 4
1 2 3 4
a b
eq
a b
k k k k k k k k k k
k
k k k k k k
+ + +
= =
+ + + +


The system model is now given.

( ) ( ) ( ) ( )
eq eq
mx t c x t k x t f t + + =


where
1 2 2 3 1 4 3 4
1 2 3 4
eq
c c c c c c c c
c
c c c c
+ + +
=
+ + +


This example is for a translational system, but equivalent results exist for an analogous rotational
system.


1
3
k
x(t)
m
k
2
4
k
k
f(t)
k
x(t)
m
k
f(t)
a b


206

4. Series springs and dampers in parallel

The figure below shows two sets of series springs arranged in parallel. For clarity, the dashpots
are not shown but they have the same arrangement.



The second figure shows a simplified diagram for this case.



Using the series- and then parallel-springs equivalent spring formulas derived in the ME 3011
NotesBook:

1 2
1 2
3 4
3 4
a
b
k k
k
k k
k k
k
k k
=
+
=
+

3 4 1 2 3 1 2 4 1 3 4 2 3 4 1 2
1 2 3 4 1 3 2 4 1 4 2 3
eq a b
k k k k k k k k k k k k k k k k
k k k
k k k k k k k k k k k k
+ + +
= + = + =
+ + + + +


The system model is given below.

( ) ( ) ( ) ( )
eq eq
mx t c x t k x t f t + + =


Where
1 2 3 1 2 4 1 3 4 2 3 4
1 3 2 4 1 4 2 3
eq
c c c c c c c c c c c c
c
c c c c c c c c
+ + +
=
+ + +


This example is for a translational system, but equivalent results exist for an analogous rotational
system.


1
3
k
x(t)
m
k
2
4
k
k
f(t)
k
x(t)
m
k
f(t)
a
b


207

5. Electric motor with vibration isolation FBD



Often a vibration isolation system will be designed to minimize the transference of shaking
forces and shaking moments from rotating machinery to the ground. The rotating electric motor shown
above, whose moment of inertia about the motor shaft axis is J, is supported by four identical springs (2
shown, 2 symmetric into the page on the back side of the motor) of spring constant k. We measure
angular displacement u(t) according to the right-hand-rule (positive CCW) such that u(t) is zero when
the system is horizontal. First, draw the FBD.

Using Eulers Rotational Dynamics Law for the motor shaft, in words we can say that the
excitation torque caused by the inertia J and angular acceleration ( ) t u

of the motor shaft must be


balanced by the restoring torque due to the four translational springs through the distance a. Assuming
positive angular acceleration, the two left translational springs are in compression and the two right
translational spring are in tension in applying the restoring torques.

( )
4 ( ) ( )
M J t
kax t J t
u
u
=
=



Where x(t) is the total displacement of each spring. Now substitute ( ) ( ) x t a t u = (approximation
for small displacement) into the above equation to yield the final 1-dof model for this system.

2
( ) 4 ( ) 0 J t ka t u u + =




J
k
a
u(t)
k


208

6. Sprung mass/cylinder FBDs



In this dynamic system, m
1
is the point mass suspended by cable about a cylindrical pulley of
mass m
2
and radius r. The other end of the cable is attached to a linear translational spring with
constant k. We measure displacement x(t) (positive downward) from the neutral position of the spring.
We also measure angular displacement u(t) with the right-hand-rule (positive CCW) such that it is zero
when x(t) is zero. First, draw the FBDs.

Using Newtons Second Law for the point mass:
1 1
( )
X
F m g T m x t = =



Using Eulers Rotational Dynamics Law for the cylinder:
( ( )) ( )
Z
M Tr kr x t J t o u = + =




Where ( ) x t o + is the total displacement of the spring (o is the static deflection of the spring due
to gravity). Also, the mass moment of inertia of the cylindrical pulley is
2
2
2 J m r = . Now substitute
( ) ( ) x t r t u = and ( ) ( ) x t r t u =

into the above equations.
1 1
2
2
( )
( ( )) ( )
2
m g T m r t
m r
Tr kr r t t
u
o u u
=
+ =



We can combine the above two equations to eliminate the cable tension T.
1 1
2
2 2
1 1
( )
( ) ( ( )) ( )
2
T m g m r t
m r
m gr m r t kr r t t
u
u o u u
=
+ =




At static equilibrium we have
1
m gr kro = , so the final 1-dof model for this system is
2
2 2 2
1
2
1
( ) ( ) ( )
2
( ) ( ) 0
2
m r
m r t kr t t
m
m t k t
u u u
u u
=
| |
+ + =
|
\ .



m
1
m
2
k
x(t)
r
u(t)


209

7. Doubly-sprung torsional rod FBD



In this dynamic system, m is the distributed mass of the slender rigid rod of length L. The rods
left end is connected to ground by a torsional spring k
R
and the rods right end is supported by two
identical linear translational springs k as shown. We measure angular displacement u(t) in a right-
handed sense from the horizontal rod position. To linearize we will make a small angle assumption; we
further ignore the small horizontal motion and gravity loading of the springs. First, draw the FBD.

Using Eulers Rotational Dynamics Law for the slender rod:

( ) 2 ( ) ( )
Z R
M k t kLy t J t u u = =




The mass moment of inertia of the slender rod about its end is
2
3 J mL = . Also, ( ) sin ( ) y t L t u = ,
leading to:

2
2
( ) ( ) 2 sin ( ) 0
3
R
mL
t k t kL t u u u + + =



Applying the small angle assumption, we have sin ( ) ( ) t t u u ~ and the final 1-dof model for this system
is

( )
2
2
( ) 2 ( ) 0
3
R
mL
t k kL t u u + + =






L
k
y(t)
u(t)
k
k
R


210

8. Sprung rolling cylinder FBD



In this dynamic system, a cylinder of mass m and radius r rolls without slipping on a horizontal
surface. The cylinder center is attached to the ground via a linear translational spring with constant k.
We measure displacement x(t) (positive to the right) from the neutral position of the spring. We also
measure angular displacement u(t) with the right-hand-rule (positive CCW) such that it is zero when x(t)
is zero. First, draw the FBD.

Using Newtons Second Law for the cylinder (F
f
is the force due to friction):

( ) ( )
X f
F F kx t mx t = =



Using Eulers Rotational Dynamics Law for the cylinder:

( )
Z f
M F r J t u = =




We can combine the above two equations to eliminate the friction force F
f
:

( ) ( )
( ( ) ( )) ( )
f
F mx t kx t
mx t kx t r J t u
= +
+ =




The mass moment of inertia of the cylindrical pulley is
2
2 J mr = . Now substitute
( ) ( ) t x t r u =

into the above equation to express the motion in terms of x(t).



2
( )
( ( ) ( ))
2
mr x t
mx t kx t r
r
| |
+ =
|
\ .



The final 1-dof model for this system is


3
( ) ( ) 0
2
mx t kx t + =

m
k
x(t)
r
u(t)


211

9. Doubly-sprung rotating cylinder FBD



In this dynamic system, a rotating cylindrical pulley of mass m and radius r is suspended from
the ground link as shown. A massless inextensible cable is passed over the pulley and connected to two
grounded linear translational springs of equal spring constant k. At static equilibrium both springs are
pre-tensioned by a vertical displacement o. We measure angular displacement u(t) with the right-hand-
rule (positive CCW). First, draw the FBD.

Using Eulers Rotational Dynamics Law for the cylinder:

1 2
( )
Z
M T r T r J t u = =




Where the left cable tension is
1
( ( )) T k r t o u = and the right cable tension is
2
( ( )) T k r t o u = + .
Also, the mass moment of inertia of the cylindrical pulley is
2
2 J mr = .

2
( ( )) ( ( )) ( )
2
mr
kr r t kr r t t o u o u u + =




The final 1-dof model for this system is

( ) 4 ( ) 0 m t k t u u + =






m
k
r
u(t)
k


212

10. Sprung pendulum FBD



In this dynamic system, a standard point-mass m pendulum is tied to the ground through a linear
translational spring with constant k, connected a distance L
0
along the length L of the massless pendulum
rod. We measure angular displacement u(t) with the right-hand-rule (positive CCW) such that it is zero
when the pendulum is vertical (down). First, draw the FBD.

Using Eulers Rotational Dynamics Law for the pendulum:

sin ( ) cos ( ) sin ( ) ( )
Z
M mgL t a t ka t J t u u u u = =




Where sin ( ) ka t u is the amount the spring stretches. The mass moment of inertia of the point-
mass pendulum about the point of rotation is
2
J mL = .
2 2
( ) ( cos ( ))sin ( ) 0 mL t mgL ka t t u u u + + =



This is a nonlinear model; in addition to the sine of the regular pendulum, now we also have a cosine
term of the unknown angular displacement u(t). To linearize let us assume small angle motion.

sin ( ) ( ) t t u u ~

cos ( ) 1 t u ~

We are implicitly ignoring the vertical motion of the spring. The final 1-dof model for this system is

2 2
( ) ( ) ( ) 0 mL t mgL ka t u u + + =





m
k
L
u(t)
L
0


213

11. Doubly-sprung rolling cylinder FBD



In this dynamic system, a cylinder of mass m and radius r rolls without slipping on a horizontal
surface. The cylinder is attached off-center (distance a from the center to the springs connection point)
to the ground via two identical linear translational springs with constant k. We measure displacement
x(t) (positive to the left) from the equilibrium position of the springs. At static equilibrium both springs
are pre-tensioned by an equal horizontal displacement o. We also measure angular displacement u(t)
with the right-hand-rule (positive CCW) such that it is zero when the springs are in their equilibrium
position. First, draw the FBD.

Using Newtons Second Law for the cylinder (F
f
is the force due to friction):

( ( ) ( )) ( ( ) ( )) ( ) ( )
X f
F F k r a t k r a t mx t mr t o u o u u = + + + + = =




Using Eulers Rotational Dynamics Law for the cylinder:

( ( ) ( )) ( ( ) ( )) ( )
Z f
M F r ka r a t ka r a t J t o u o u u = + + + + =




Where the left spring force tension is ( ( ) ( )) k r a t o u + and the right cable tension is
( ( ) ( )) k r a t o u + + + . Also, the mass moment of inertia of the cylinder is
2
2 J mr = . We can combine the
above two equations to eliminate the friction force F
f
:

2
( ) 2 ( ) ( )
( ( ) 2 ( ) ( )) 2 ( ) ( ) ( )
2
f
F mr t k r a t
mr
mr t k r a t r ka r a t t
u u
u u u u
= +
+ + =





The final 1-dof model for this system is

2 2
3
( ) 2 ( ) ( ) 0
2
mr t k r a t u u + + =






m
k
x(t)
r
u(t)
k
a


214

12. Hung-sprung mass/cylinder FBDs



In this dynamic system, m
1
is the point mass suspended by an inextensible cable about a
cylindrical pulley of mass m
2
and radius r. The other end of the cable is rigidly attached to the ground.
The pulley is attached to ground via a linear translational spring with constant k. We measure
displacements x
1
(t) and x
2
(t) (both positive downward) from the neutral position of the spring. We also
measure angular displacement u(t) with the right-hand-rule (positive CCW) such that it is zero when x(t)
is zero. First, draw the FBDs.

Using Newtons Second Law for the point mass:

1 1 1 1
( )
X
F m g T m x t = =



Using Newtons Second Law for the cylinder:

1 2 2 2 2 2
( ( )) ( )
X
F T T m g k x t m x t o = + + + =



Where o is the static deflection of the spring due to gravity. Using Eulers Rotational Dynamics Law for
the cylinder:

1 2
( )
Z
M T r T r J t u = =




The mass moment of inertia of the cylindrical pulley is
2
2
2 J m r = . Substituting the mass moment of
inertia and eliminating the cable tensions T
1
and T
2
amongst these three equations yields

m
1
m
2
k
x (t)
r
u(t)
1
x (t)
2


215

1 1 1 1
2 2 2 2 2 1
2
1 1 1 2 2 2 2
( )
( ) ( ( ))
2( ( )) ( ) ( ( )) ( )
2
T m g m x t
T m x t m g k x t T
m r
m g m x t m x t m g k x t t
o
o u
=
= + +
+ + =




At static equilibrium we have

1 2 2
1 2
(2 )
T T m g k
m m g k
o
o
+ + =
+ =


So the model simplifies to

2
1 1 2 2 2
( ) 2 ( ) ( ) ( ) 0
2
m r
t m x t m x t kx t u + + + =



Note that from kinematics we know
1 2
( ) 2 ( ) x t x t = and
2
( ) ( ) x t r t u = . Substituting the kinematic
equations yields the final 1-dof model for this system.

1 2 1 1
3
4 ( ) ( ) 0
2
m m x t kx t
| |
+ + =
|
\ .


Interestingly, the same exact equation form results, regardless of whether the model variable is x
1
(t), as
above, or x
2
(t), or u(t).




216

13. Horizontally-sprung mass/cylinder FBDs



In this dynamic system, m
1
is the point mass suspended by cable about a cylindrical pulley of
mass m
2
and radius r. The pulley is grounded via a linear translational spring with constant k, attached
distance a from the rotational center of the pulley. We measure displacement x(t) (positive downward)
from the neutral position of the spring. We also measure angular displacement u(t) with the right-hand-
rule (positive CCW) such that it is zero when x(t) is zero. First, draw the FBDs.

Using Newtons Second Law for the point mass:
1 1
( )
X
F m g T m x t = =



Using Eulers Rotational Dynamics Law for the cylinder:
( ( )) ( )
Z
M Tr ka a t J t o u u = + =




Where ( ) a t o u + is the total displacement of the spring (o is the static deflection of the spring due
to gravity). Also, the mass moment of inertia of the cylindrical pulley is
2
2
2 J m r = . We can combine
the above two equations to eliminate the cable tension T.

1 1
2
2
1 1
( )
( ( )) ( ( )) ( )
2
T m g m x t
m r
m g m x t r ka a t t o u u
=
+ =



At static equilibrium we have
1
m gr k a o = , so the model becomes
2
2 2
1
( ) ( ) ( )
2
m r
mrx t ka t t u u =



Now using the kinematic relationships ( ) ( ) x t r t u = and ( ) ( ) x t r t u =

in the above equation, the final 1-
dof model for this system is
2 2
2
1
( ) ( ) 0
2
m
m r t ka t u u
| |
+ + =
|
\ .



m
1
m
2
k
x(t)
r
u(t)
a


217

14. Cylinder rolling in a big cylinder FBD


In this dynamic system, a cylinder of mass m and radius r rolls without slipping in a concave
cylindrical surface of radius R as shown. u
1
(t) is the angular displacement of the cylinder and u
2
(t) is the
angular displacement of the cylinder CG from the vertical. Both angles are measured with the right-
hand-rule (positive CCW) such that they are zero when the cylinder is at static equilibrium at the bottom
of the well. We define the moving X axis such that it is perpendicular to the radial direction, pointing to
the left. First, draw the FBD.

Using Newtons Second Law for the cylinder:
2
sin ( ) ( )
X f
F mg t F mx t u = =



Using Eulers Rotational Dynamics Law for the cylinder:
1 2
( ( ) ( ))
Z f
M F r J t t u u = = +




Note we must use the absolute angular acceleration
1 2
( ) ( ) t t u u +

in the above equation. The mass
moment of inertia of the cylindrical pulley is
2
2 J mr = . Substituting the mass moment of inertia and
eliminating the friction force F
f
between these two equations yields

2
2 1 2
sin ( ) ( )
sin ( ) ( ) ( ( ) ( ))
2
f
F mg t mx t
mr
mg t mx t t t
u
u u u
=
= +



Note that from kinematics we know
2
( ) ( ) ( ) x t R r t u =

and
1 2
( ) ( ) t R t r u u = . Substituting the
kinematic equations yields
2 2
3( )
( ) sin ( ) 0
2
R r
t g t u u

+ =



We apply a small angle assumption such that
2 2
sin ( ) ( ) t t u u ~ . The final linearized 1-dof model for this
system is
2 2
3( )
( ) ( ) 0
2
R r
t g t u u

+ =



m
r
u (t)
R
1
u (t)
2


218

15. Doubly-sprung pulleys/mass FBDs



In this dynamic system, m is the point mass suspended by an inextensible cable about two
pulleys of negligible mass as shown. Each pulley is connected to the ground via linear translational
springs with constants k
1
and k
2
, respectively. We measure displacements x(t) (positive downward),
x
1
(t), and x
2
(t) as shown.

The kinematics relationship is:
1 2
( ) 2 ( ) 2 ( ) x t x t x t = +

From a FBD for pulley 1 and pulley 2:
1 1
2 2
2 ( ) ( )
2 ( ) ( )
T t k x t
T t k x t
=
=


Substituting the above relationships into the kinematics relationship yields
1 2
4 ( ) 4 ( )
( )
T t T t
x t
k k
= +

Looking at the point mass we can write
( ) ( )
EQ
k x t T t =

Where k
EQ
is the equivalent spring constant.
( )
1 2
1 2
( )
( ) 4
EQ
k k T t
k
x t k k
= =
+


The final 1-dof model for this system is
( )
1 2
1 2
( ) ( ) 0
( ) ( ) 0
4
EQ
mx t k x t
k k
mx t x t
k k
+ =
+ =
+



m
2
k
x(t)
1
k
x (t)
1
x (t)
2


219

16. Rolling cylinder/pendulum FBD


In this dynamic system, a standard point-mass pendulum is rigidly connected to a cylinder that is
free to rotate without slipping on a horizontal surface. M
1
is the point mass and the cylinder has mass m
2

and radius r. The length of the massless pendulum rod is L. We measure angular displacement u(t) with
the right-hand-rule (positive CCW) such that it is zero when the pendulum is vertical (down). First,
draw the FBD.
At any instant of the motion, the rotation is about (moving) point C, the instantaneous center of
zero velocity. Therefore let us take point C as the moment center for deriving the dynamic model
( )
CZ C
M J t u =


. The mass moment of inertia of the compound rigid link, using the parallel axis
theorem, is:
2
2 2 2 2 2
1 2 1 2
3
2 2
C
m r
J m R m r m R m r = + + = +

Where, by use of a vector loop-closure equation,
2 2 2
2 cos ( ) R L Lr t r u = + . Using Eulers
Rotational Dynamics Law for the cylinder:
1
2 2
1 2 1
2 2 2
1 2 1
sin ( ) ( )
3
( ) sin ( ) 0
2
3
( 2 cos ( ) ) ( ) sin ( ) 0
2
CZ C
M m gL t J t
m R m r t m gL t
m L Lr t r m r t m gL t
u u
u u
u u u
= =
| |
+ + =
|
\ .
| |
+ + + =
|
\ .



This is a nonlinear model; in addition to the sine of the regular pendulum, now we also have a cosine
term of the unknown angular displacement u(t). To linearize let us assume small angle motion.
sin ( ) ( ) t t u u ~

cos ( ) 1 t u ~

The final 1-dof model for this system is
2 2
1 2 1
3
( ) ( ) ( ) 0
2
m L r m r t mgL t u u
| |
+ + =
|
\ .



m
L
u(t)
1
m
2
r


220

17. Rigid-body pendula
Earlier we derived the linearized dynamic model for a simple pendulum consisting of a massless
rigid rod of length L and a point mass m.
( ) ( ) 0
g
t t
L
u u + =



We also derived the dynamic model for a slender-rod pendulum with distributed mass and also
for a slender-rod pendulum with distributed mass plus point mass. Here we present the dynamic models
for pendula of various other rigid-body shapes, including for the general case. The pendulum mass is m,
O is the fixed point of rotation, G is the pendulum center of mass, and r
G
is the distance between these
two points. As before, we measure angular displacement u(t) with the right-hand-rule (positive CCW)
such that it is zero when the pendulum is vertical (down). Draw the FBD for the general rigid-body
pendulum below.








General Rigid-Body Pendulum FBD
2
G
O G G
J
J J mr = +

( ) ( ) 0
G
O
mgr
t t
J
u u + =



Note that if we define a virtual point mass pendulum length of
EQ O G
L J mr = , we could use the simple
point-mass pendulum model for any rigid body pendulum. Now we can apply these results to specific
rigid-body pendula below.


Semi-circular Circular Rectangular Square
2
4
3
2
G
O
r
r
mr
J
t
=
=

2
2
2
3
2
G
O
mr
J
mr
J
=
=

2 2
2 2
( )
12
(4 )
12
G
O
m L h
J
m L h
J
+
=
+
=

2
2
6
5
12
G
O
ms
J
ms
J
=
=


8
( ) ( ) 0
3
g
t t
r
u u
t
+ =


2
( ) ( ) 0
3
g
t t
r
u u + =


2 2
6
( ) ( ) 0
4
gL
t t
L h
u u + =
+


6
( ) ( ) 0
5
g
t t
s
u u + =



u
r
g
(t)
m
u
r
g
(t)
m
u
h
g
(t)
m
L
u
g
(t)
m
s
s


221

18. Beams with a point mass

Earlier we derived the equivalent spring constants k
EQ
for three massless beams with point
masses attached as shown, the cantilevered beam, the simply-supported beam, and the rigidly-clamped
beam. For all three cases the general model is ( ) ( ) 0
EQ
mx t k x t + = .


Cantilevered beam Simply-supported beam

3
3
( ) ( ) 0
4
Ebh
mx t x t
L
+ =

3
3
4
( ) ( ) 0
Ebh
mx t x t
L
+ =




Doubly-fixed beam
3
3
16
( ) ( ) 0
Ebh
mx t x t
L
+ =

Where, for all three cases:
E material Youngs modulus
b beam width
h beam height
L length of beam



m
L x(t)
h
m
L
x(t)
h
m
L
x(t)
h


222

19. Rotating imbalance

Machines with rotating parts are common in industry. The figure below represents a simplified
idealized machine with an offset mass m
1
that is unbalanced in rotation. The total mass of the machine
is m
2
, supported by equivalent spring and damping coefficients k and c, respectively, to the ground
(there may be four machine mounts, one on each corner of the machine, in parallel, so their individual k
i

and c
i
coefficients add up to yield the overall k and c). The machine is constrained to move in the
vertical direction only. We measure displacement x(t) in the vertical (up) direction, from the gravity-
loaded spring location.



The model is the same we have derived earlier for the standard m-c-k translational mechanical
system.

2
( ) ( ) ( ) ( ) m x t cx t kx t f t + + =

Where now the forcing function f(t) is supplied by the rotating imbalance, mass m
1
at an offset of e, as
shown. Assume the rotation is at a constant angular velocity of e. Then the centripetal acceleration of
the rotating mass is
2
ee (directed inward) and thus the inertial force of the rotating mass is
2
1
m ee
(directed outward). The angle of rotation is t e and so the vertical component of the rotating imbalance
is:

2
1
( ) sin f t me t e e =

And so the final 1-dof model for this system is

2
2 1
( ) ( ) ( ) sin m x t cx t kx t me t e e + + =


m
1
m
2
k
c
x(t)
e
t e


223

20. m-c-k system with displacement input FBD



Some machines can be modeled as the standard linearized, lumped-parameter, constant
coefficient, second-order m-c-k system, but with a base displacement input in place of the standard force
f(t) input (see the figure above).

In this case, the base is not fixed, but harmonically-moved via base displacement u(t) in units of
meters. We measure displacement x(t) in the same manner (positive to the right, from the neutral
position of the spring, this time with u(t) =0). There is no input forcing function f(t). The moving base
portion is massless.

The model is similar to what we have previously derived. Assume that displacements x(t) and
u(t) are small and positive at this instant in time, with ( ) ( ) x t u t > . We know that the model we derive
based on these assumptions will hold good for all small displacement motions, as both x(t) and u(t)
vibrate at different phases in different directions. With this assumption, the spring is in tension, and the
damper behaves similarly. First, draw the FBD. Using Newtons Second Law for mass m:
( ( ) ( )) ( ( ) ( )) ( )
( ) ( ) ( ) ( ) ( )
X
F k x t u t c x t u t mx t
mx t cx t kx t cu t ku t
= =
+ + = +





We can genericize this model similar to previous derivations by using the dimensionless
damping ratio
2
c
km
=

and natural frequency
n
k
m
e = as follows.
2 2
( ) 2 ( ) ( ) 2 ( ) ( )
n n n n
x t x t x t u t u t e e e e + + = +


It may be convenient in some cases to use a new variable in the above model, the relative
difference in displacements, ( ) ( ) ( ) z t x t u t = . Then we have
( ) ( ) ( ) ( ) mz t cz t kz t mu t + + =


Where now ( ) mu t plays the role of input forcing function. If we specify a sinusoidal input
displacement ( ) sin u t A t e = then the model becomes
2
( ) ( ) ( ) sin mz t cz t kz t mA t e e + + =

Note that this version of the model is identical in form to the forced, damped m-c-k system where the
force amplitude F has been replaced by
2
mAe .

k
c
m
x(t) u(t)


224

21. Horizontal sprung pendulum FBD



In the figure above, a horizontal simple pendulum has a point mass on the end of a massless rigid
rod of length a. The rod extends by length b to the other side to connect to the ground through a linear
translational spring with constant k. We measure angular displacement u(t) with the right-hand-rule
(positive CCW) such that it is zero when the pendulum is horizontal. We further measure displacement
variable x(t) as the extension/compression of the spring as shown. First, draw the FBD.

Using Eulers Rotational Dynamics Law for the pendulum:

cos ( ) ( ( )) ( )
Z
M mga t bk x t J t u o u = + =




Where ( ) x t o + is the total displacement of the spring (o is the static deflection of the spring in
tension due to gravity). Also, the mass moment of inertia of the pendulum about the axis of rotation is
2
J ma = .

2
( ) ( ( )) cos ( ) 0 ma t bk x t mga t u o u + + =



This is a nonlinear model, due to the cosine term of the unknown angular displacement u(t). To
linearize let us assume small angle motion cos ( ) 1 t u ~ .

2
( ) ( ( )) 0 ma t bk x t mga u o + + =




At static equilibrium we have mga kbo = , so the simplified model for this system is

2
( ) ( ) 0 ma t bkx t u + =



For a small displacement x(t) the approximation ( ) sin ( ) x t b t u ~ holds good. Once again assuming
small angle motion, to linearize we use sin ( ) ( ) t t u u ~ , so the final 1-dof model for this system is

2 2
( ) ( ) 0 ma t kb t u u + =



m
k
a
u(t)
x(t)
b


225

22. Sprung cylinder FBD



In this dynamic system, a cylinder of mass m and radius r is suspended via a cable on one side
and connected to ground via a linear translational spring with constant k on the other side. We measure
displacement x(t) (positive downward) from the neutral position of the spring and y(t) is the
displacement of the cylinder center. We also measure angular displacement u(t) with the right-hand-rule
(positive CCW) such that it is zero when x(t) is zero. First, draw the FBD. Using Newtons Second
Law for the cylinder:
( ( )) ( ) ( )
X
F mg k x t T t my t o = + =



Where o is the static deflection of the spring due to gravity. Using Eulers Rotational Dynamics Law for
the cylinder:
( ) ( ( )) ( )
Z
M T t r k x t r J t o u = + =




The mass moment of inertia of the cylindrical pulley is
2
2 J mr = . Substituting the mass moment of
inertia and eliminating the cable tension T(t) between these two equations yields

( ) ( ( )) ( )
2 ( ( )) ( ) ( )
2
T t mg k x t my t
mr
mg k x t my t t
o
o u
= +
+ =



At static equilibrium we have
2
mg
k T o = =


So the model simplifies to
( ) ( ) 2 ( ) 0
2
mr
t my t kx t u + + =



Note that from kinematics we know ( ) 2 ( ) x t y t = and ( ) ( ) y t r t u = . Substituting the kinematic equations
yields the final 1-dof model for this system.
3
( ) 4 ( ) 0
2
m t k t u u + =



k
x(t)
r
u(t)
y(t)


226

23. Vertical sprung pendulum with damping FBD



In the figure above, a simple pendulum rotating about point O has a point mass on the end of a
massless rigid rod of length L. The rod extends by length b to the other side of point O to connect to the
ground through a horizontal linear translational spring with constant k. We measure angular
displacement u(t) with the right-hand-rule (positive CCW) such that it is zero when the pendulum is
vertical. We further measure displacement variable ( ) ( )
b
x t b t u = as the compression/extension of the
spring as shown. A viscous dashpot is connected by length a down along the rod to provide damping,
with velocity variable ( ) ( )
a
x t a t u =

. First, draw the FBD.

Using Eulers Rotational Dynamics Law for the pendulum:

sin ( ) ( ) cos ( ) ( ) cos ( ) ( )
OZ
M mgL t kb t b t ca t a t J t u u u u u u = =




The mass moment of inertia of the pendulum about the axis of rotation is
2
J mL = .

2 2 2
( ) sin ( ) ( )cos ( ) ( )cos ( ) 0 mL t mgL t kb t t ca t t u u u u u u + + + =



This is a nonlinear model; in addition to the sine of the regular pendulum, now we also have a cosine
term of the unknown angular displacement u(t). To linearize let us assume small angle motion.

sin ( ) ( ) t t u u ~

cos ( ) 1 t u ~

Then the final 1-dof model for this system is

2 2 2
( ) ( ) ( ) ( ) 0 mL t ca t mgL kb t u u u + + + =




m
k
L
u(t)
c
a
b
x (t)
b


227

24. Mass suspended via tensed string



In this system a point mass m is suspended between two strings of tension T. Assume that for
small motions the tension can be considered to be constant. Also assume there is only vertical
displacement, i.e. ignore horizontal motion of the mass. We measure displacement x(t) positive
vertically (up).


Using Newtons Second Law for the point mass:

( )
X
F f mx t = =



Where f is the restoring force due to both strings:

1 1 2 2 1 2
sin ( ) sin ( ) (sin ( ) sin ( )) f T t T t T t t u u u u = + = +

Assuming small displacement we also have small angular motions. For small angular motions
sin ( ) tan ( ) t t u u ~ and then we have

1 2 1 1
( ) ( ) ( ) ( )
( ) ( )
x t x t x t x t
f T T
L L L L L
= + = +



The final 1-dof model for the tensed-string system is

1 1
1 1
1 1
( ) ( ) ( ) 0
( ) ( ) ( ) 0
( )
mx t T x t
L L L
L
mx t T x t
L L L
+ + =

+ =




m
L
L
1
L
2
L
1
L
2
x(t)
T
T


228

25. Manometer



A manometer is used for measuring pressure. The manometer shown has a uniform cross-
sectional area A, a liquid column of length L, and a liquid density of . We measure displacement x(t)
positive vertically as shown, relative to atmospheric pressure, where the two columns have equal height.

Using Newtons Second Law for the fluid mass:

( ) ( )
X
F w t mx t = =



Where w(t) is the weight acting due to the difference in height in the manometer. Therefore:

( ) ( ) 2 ( ) wt gV t gAx t = =

Since the total liquid mass m in the manometer is AL , the final 1-dof manometer dynamics
model is

( ) 2 ( ) 0
( ) 2 ( ) 0
( ) 2 ( ) 0
mx t gAx t
ALx t gAx t
Lx t gx t


+ =
+ =
+ =




x(t)
x(t)


229

26. Cylinder in water bath



A solid buoyant cylinder is immersed in a water bath. Derive the vibratory dynamic equation of
motion assuming purely vertical motion with no rotations. The cylinder has a height h and radius r and
thus a mass of
2
m V r h t = = , where is the cylinder material density. Further let
W
be the water
density. Displacement x(t) is the amount of the cylinder height under water at any instant, as shown.
From Archimedes principle, the time-varying restoring force f
R
due to the water is the weight of water
displaced.

2
( ) ( )
R W W W
f t V g g r x t t = =

The vibratory equation of motion is

2 2
( ) ( ) 0
( ) ( ) 0
( ) ( ) 0
W
W
mx t kx t
r hx t g r x t
hx t gx t
t t

+ =
+ =
+ =




x(t)
h
r


230

27. Human skeletal muscle model

This section presents a common single skeletal muscle dynamics model, including the figure,
equations, state-space form, and simulations. This is a vibrational model, included for completeness, to
show one type of engineering mechanics modeling for muscles. In this class we will generally treat the
effect of muscles more simply, i.e. as simple cables in tension with no mass, spring constants, or
dynamics of their own.

The model in this section was suggested by Dr. Scott Hooper (OU Biological Sciences) and the
equations were derived and the simulations performed by Elvedin Kljuno (OU Fulbright Scholar from
Bosnia). Dr. Hooper uses this model to study neural muscular control in the stomachs of lobsters, but he
says it can be scaled to adequately model skeletal muscle in many animals, including humans.

The skeletal muscle vibrational dynamics model is shown below. The lumped mass m represents
the load the muscle is lifting vs. gravity g. The muscle itself is represented by linear elastic spring
stiffness k
1
, in parallel with linear elastic spring stiffness k
2
that is in series with linear dissipative
dashpot b. The absolute displacement of the dashpot end and the muscle end are measured by
coordinates x and y, respectively. Lengths L
1R
and L
2R
(not shown) are the resting lengths (not stretched
by gravity) of springs 1 and 2, respectively. In vibrations it is more common to measure change in
displacements relative to the gravity-stretched position, but biologists need to include the absolute spring
lengths as well. The actuator F
A
represents the contractile element of the muscle and F
m
is the force
generated by the muscle. Note: this derivation assumes zero pennation angle, i.e. zero angle o between
the muscle fibers and the tendons.


Dynamic Skeletal Muscle Diagram


231

The parallel elastic component k
1
, provided by the muscle membranes, gives spring resistance
when passively stretched. The series elastic component k
2
represents the tendons, storing elastic
energy when a muscle is stretched. The tension-generating contractile effect of the sarcomeres,
modeled by actuator force F
A
, is in parallel with the membranes and in series with the tendons.
From the left figure below, the equation for the left spring is:

1 1R 1
( ) k y L F = (1)



From the right figure above, the equations for the right spring and dashpot are:

2 2R 2
2 A
( ) k y x L F
bx F F
=
= +
(2, 3)

From the FBD shown below, the dynamics equation for the mass is obtained using Newtons Second
law:

1 2 m
vert vert
F ma
F F F mg my
=
+ + =

(4)






232

Substituting (1) and (2) into (4), and also substituting (2) into (3) yields the following coupled
second/first order linear vibrational dynamics model for skeletal muscle.

1 1R 2 2R m
2 2R A
( ) ( )
( )
my k y L k y x L mg F
bx k y x L F
+ + = +
=

(5, 6)


Equations (5, 6) can be written in the following form:

2 m
1 2 1 1R 2 2R
2 2 A 2
2R
1 1
( ) ( )
k F
y k k y x k L k L g
m m m m
k k F k
x x y L
b b b b
+ + = + + +
+ =

(7, 8)

As mentioned above, (7, 8) can be simplified by choosing more adequate coordinates that will
measure the deviation from the static position only. The form shown above is formed intentionally for
biological systems, using absolute coordinates and initial spring lengths.

The system of differential equations (7, 8) can be solved for coordinates x and y as functions of
time given time functions for F
A
and F
m
. In the case of a simple form of F
A
and F
m
the solution could
be found using reduction and decoupling by a formation of a single third-order linear ordinary
differential equation.

2 1 2 1 2 2
1 1
( ) ( )
( ) ( ) ( ) ( ) ( ( ) ( ) )
m
A m R
F t k k k k k k
y t y t y t y t F t F t mg k L
b m mb m mb
+
+ + + = + + + +

(9)

Equation (9) can be solved for y(t) which is then substituted into (8) to directly solve for x(t).






233

Common Model Features


Considering all vibrational models we have derived, including the many cases in this section,
some general comments can be made about all models.


1. When gathering the unknown and its time derivatives on one side of the ODE, there will
never be a change in sign amongst these terms; generally they will all be positive.


2. The effect of gravity always cancels out. That is, any mg static weight terms will load the
spring with a static deflection and these terms always subtract out of the final ODE.


3. As always, the resulting units must be consistent. This is a good check to perform on your
resulting model.





234

6.5 Electrical Circuits Modeling
This section presents modeling of electrical circuits. Like mechanical translational and rotational
systems, electrical circuits have dynamics described by IVP ODEs, including input and output variables
that are functions of time. We will make analogies between the mechanical and electrical systems, i.e.
the models they yield are very similar. Hence, we can use the same techniques for solving these
mechanical and electrical ODEs (in Chapters 3 and 4).

Many electrical circuits can exhibit electrical vibrations, i.e. oscillation of the electrical output
variable of interest, which is why their modeling is included here.

Kirchoffs Laws
Instead of Newtons and Eulers dynamics laws for mechanical systems, we use Kirchoffs Laws
to derive the models for electrical circuits.

Kirchoffs Current Law (KCL) states that the sum of currents flowing into any circuit node is
zero.

Kirchoffs Voltage Law (KVL) states that the sum of voltages around any circuit loop is zero.

1
( ) 0
n
j
j
i t
=
=


1
( ) 0
m
j
k
v t
=
=








KCL Diagram KVL Diagram

Electrical Circuit Elements
In order to use Kirchoffs Laws to derive models for electrical circuits, we need the following
relationships. They relate the current i(t) flowing through and the voltage v(t) across the three standard
electrical elements. A capacitor C stores electrical energy, a resistor R dissipates electrical energy, and
an inductor L oscillates electrical energy. Note the equations in the last two columns of each row are
equivalent, solved for the current i(t) or the voltage v(t).

element notation units i(t) v(t)
capacitor C Farads

( )
( )
C
C
dv t
i t C
dt
=
1
( ) ( )
C C
v t i t dt
C
=
}

resistor R Ohms

( )
( )
R
R
v t
i t
R
= ( ) ( )
R R
v t Ri t =
inductor L Henrys

1
( ) ( )
L L
i t v t dt
L
=
}

( )
( )
L
L
di t
v t L
dt
=




235

Additional Electrical Circuit Variables

In some electrical circuit modeling you may encounter two more variable types, electrical charge
and electrical flux. As shown in Dr. Bobs on-line Atlas of Models and Transfer Functions,

www.ohio.edu/people/williar4/html/PDF/ModelTFAtlas.pdf

electrical charge q(t) is the time integral of current i(t) and electrical flux |(t) is the time integral of
voltage v(t).

( ) ( ) q t i t dt =
}

( ) ( ) t v t dt | =
}




Convention for Electron Flow

When early electricity experimenters and early electrical engineer pioneers established the
convention for the flow of electrons composing current, they got it backwards. This error persists to this
day, because it really doesnt matter in the equations.




xkcd.com




236

Current-Driven Parallel R-L-C Electrical Circuit Model

The system diagram for the current-driven parallel R-L-C electrical circuit is shown below. The
input is current i(t), and the output is voltage v(t). The voltage across the current generator, the resistor
R, the inductor L, and the capacitor C are all identical, v(t).



Parallel R-L-C Circuit Diagram

Using Kirchoffs Current Law (the sum of currents into any circuit node is zero), we have:

( ) ( ) ( ) ( ) 0
( ) ( ) ( ) ( )
R L C
C R L
i t i t i t i t
i t i t i t i t
=
+ + =


Substituting the current through each element as a function of voltage v(t) from the previous table, we
obtain the model for the current-driven parallel R-L-C circuit.

( ) 1 1
( ) ( ) ( )
dv t
C v t v t dt i t
dt R L
+ + =
}


This model is a linear, lumped-parameter, constant-coefficient, integro-differential equation. It is
second-order, since, although the highest derivative order is first-order, there are two time differentiation
steps between the integral and the first derivative. This equation is written in standard form, where the
output variable v(t) appears with its derivative and integral on the left-hand-side and the input forcing
function i(t) appears on the right-hand-side of the equation. The external current input is i(t) (amps) and
the output is voltage v(t) (volts).

i(t)
R L C
v(t)
+
-
i (t)
L
i (t)
R
i (t)
C


237

Force-Current Analogy
We can make an analogy between the mechanical and electrical modeling worlds by comparing
the translational (and rotational) mechanical system with the current-driven parallel R-L-C circuit model.
This is called the force-current analogy.
We first rewrite the left-hand-side of the translational mechanical system model in terms of
velocity v(t) instead of displacement x(t). We can also compare to the rotational mechanical system,
rewritten in terms of angular velocity e(t) instead of angular displacement u(t).

( ) ( ) ( ) ( )
( ) ( ) ( ) ( )
mx t cx t kx t f t
mv t cv t k v t dt f t
+ + =
+ + =
}


( ) ( ) ( ) ( )
( ) ( ) ( ) ( )
R R
R R
J t c t k t t
J t c t k t dt t
u u u t
e e e t
+ + =
+ + =
}



( ) 1 1
( ) ( ) ( )
dv t
C v t v t dt i t
dt R L
+ + =
}


Compare the mechanical systems ODEs to the parallel R-L-C circuit model (repeated above).

Energy Role
Translational
Mechanical
Rotational
Mechanical
Electrical
store energy
( )
( )
S
dv t
f t m
dt
=
( )
( )
S
d t
t J
dt
e
t =
( )
( )
S
dv t
i t C
dt
=
dissipate energy
( ) ( )
D
f t cv t = ( ) ( )
D R
t c t t e =
1
( ) ( )
D
i t v t
R
=
oscillate energy
( ) ( )
O
f t k v t dt =
}
( ) ( )
O R
t k t dt t e =
}

1
( ) ( )
O
i t v t dt
L
=
}


Force-Current Analogy

Variable Type

Translational
Mechanical System

Rotational
Mechanical System

R-L-C Parallel
Circuit

input (through) f(t) t(t) i(t)
output (across) v(t) (velocity) e(t) v(t) (voltage)
inertia m J C
damping c c
R
1 / R
stiffness k k
R
1 / L

Caution
The Force-Voltage Analogy, presented next, based on the voltage-driven series R-L-C circuit
model, is different.



238

Voltage-Driven Series R-L-C Electrical Circuit Model

The system diagram for the voltage-driven series R-L-C electrical circuit is shown below. The
input is voltage v(t), and the output is current i(t). The current i(t) flowing through all elements is the
same. The input voltage source v(t) increases as shown negative to positive; the voltage drops across
each of the circuit elements, positive to negative in each case.



Series R-L-C Circuit Diagram

Using Kirchoffs Voltage Law (the sum of voltages around the circuit loop is zero), we have:

( ) ( ) ( ) ( ) 0
( ) ( ) ( ) ( )
R L C
L R C
v t v t v t v t
v t v t v t v t
=
+ + =


Substituting the voltage drop across each element as a function of current i(t) from the first table in this
section, we obtain the model for the voltage-driven series R-L-C circuit.

( ) 1
( ) ( ) ( )
di t
L Ri t i t dt v t
dt C
+ + =
}


This model is a linear, lumped-parameter, constant-coefficient, integro-differential equation. Again, it is
second-order since there are two time differentiation steps between the integral and the first derivative.
This equation is written in standard form, where the output variable i(t) appears with its derivative and
integral on the left-hand-side and the input forcing function v(t) appears on the right-hand-side of the
equation. The external input is voltage v(t) (volts) and the output is current i(t) (amps).


v(t)
L
R
+
-
i(t)
C


239

Force-Voltage Analogy

We can make a different analogy between the mechanical and electrical modeling worlds by
comparing the translational (and rotational) mechanical system with the voltage-driven series R-L-C
circuit model. This is called the force-voltage analogy.

We again use the translational and rotational mechanical system models, rewritten in terms of
velocity v(t) instead of displacement x(t) and rewritten in terms of angular velocity e(t) instead of
angular displacement u(t).

( ) ( ) ( ) ( )
( ) ( ) ( ) ( )
mx t cx t kx t f t
mv t cv t k v t dt f t
+ + =
+ + =
}


( ) ( ) ( ) ( )
( ) ( ) ( ) ( )
R R
R R
J t c t k t t
J t c t k t dt t
u u u t
e e e t
+ + =
+ + =
}



( ) 1
( ) ( ) ( )
di t
L Ri t i t dt v t
dt C
+ + =
}



Force-Voltage Analogy

Variable Type

Translational
Mechanical System

Rotational
Mechanical System

R-L-C Series
Circuit

input (through) f(t) t(t) v(t) (voltage)
output (across) v(t) (velocity) e(t) i(t)
inertia m J L
damping c c
R
R
stiffness k k
R
1/C




240

Subset Electrical Circuit Models

This subsection presents subset models derived from the standard current-driven parallel R-L-C
electrical circuit and the voltage-driven series R-L-C electrical circuit models.


Parallel R-L-C Circuit Series R-L-C Circuit

( ) 1 1
( ) ( ) ( )
dv t
C v t v t dt i t
dt R L
+ + =
}

( ) 1
( ) ( ) ( )
di t
L Ri t i t dt v t
dt C
+ + =
}




Parallel R-L Circuit Series R-L Circuit

1 1
( ) ( ) ( ) v t v t dt i t
R L
+ =
}

( )
( ) ( )
di t
L Ri t v t
dt
+ =



Parallel R-C Circuit Series R-C Circuit

( ) 1
( ) ( )
dv t
C v t i t
dt R
+ =

1
( ) ( ) ( ) Ri t i t dt v t
C
+ =
}




Parallel L-C Circuit Series L-C Circuit

( ) 1
( ) ( )
dv t
C v t dt i t
dt L
+ =
}

( ) 1
( ) ( )
di t
L i t dt v t
dt C
+ =
}





241

Parallel R-L-C models based on flux |(t) and series R-L-C models based on charge q(t)

Recall that, in an electrical circuit, electrical flux |(t) is the time integral of voltage v(t) and
electrical charge q(t) is the time integral of current i(t). The subset models for the current-driven parallel
R-L-C electrical circuit and voltage-driven series R-L-C electrical circuit can thus be rewritten as
follows, using flux |(t) as the output variable in place of voltage v(t) and using charge q(t) as the output
variable in place of current i(t), respectively. Here are the flux/voltage and charge/current substitution
relationships.

( ) ( )
( ) ( )
( )
( )
t v t dt
t v t
dv t
t
dt
|
|
|
=
=
=
}


( ) ( )
( ) ( )
( )
( )
q t i t dt
q t i t
di t
q t
dt
=
=
=
}




Parallel R-L-C Circuit Series R-L-C Circuit

1 1
( ) ( ) ( ) ( ) C t t t i t
R L
| | | + + =


1
( ) ( ) ( ) ( ) Lq t Rq t q t v t
C
+ + =



Parallel R-L Circuit Series R-L Circuit

1 1
( ) ( ) ( ) t t i t
R L
| | + =


( ) ( ) ( ) Lq t Rq t v t + =



Parallel R-C Circuit Series R-C Circuit

1
( ) ( ) ( ) C t t i t
R
| | + =


1
( ) ( ) ( ) Rq t q t v t
C
+ =



Parallel L-C Circuit Series L-C Circuit

1
( ) ( ) ( ) C t t i t
L
| | + =


1
( ) ( ) ( ) Lq t q t v t
C
+ =


In the Force-Current Analogy the new output variable would be flux |(t) instead of voltage v(t)
and in the Force-Voltage Analogy the new output variable would be charge q(t) instead of current i(t).
The rest of these analogies are unchanged.



242

6.6 Multi-dof Vibrational Systems Models

This chapter presents some mechanical system models for systems with greater than one degree-
of-freedom (dof). For multiple independent masses, draw a free-body diagram (FBD) for each mass and
derive an ODE using Newtons Second Law or Eulers Rotational Dynamics Law. Generally there will
be one second-order ODE for each independent mass. Normally the resulting set of n ODEs will be
coupled, where n is the number of degrees-of-freedom.

1. Two-Mass Vibrational System




Two-mass system free-body diagrams


( ) ( )
1 1 1 2 1 1 2 1 2 2 2 2 1
2 2 2 2 2 2 2 1 2 1 2
( ) ( ) ( ) ( ) ( ) ( )
( ) ( ) ( ) ( ) ( ) ( )
m y t c c y t k k y t c y t k y t u t
m y t c y t k y t c y t k y t u t
+ + + + =
+ + =





( )
( )
( )
( )
( )
( )
( )
( )
1 1 1 1 1 1 2 2 1 2 2
2 2 2 2 2 2 2 2 2
0
0
y t y t y t u t m c c c k k k
y t y t y t u t m c c k k
+ + ( ( (

+ + =
` ` ` `
( ( (


) ) ) )





m
y (t)
1
y (t)
2
u (t)
1
u (t)
2
2
m
1
c
2
c
1
k
2
k
1
m
1
m
2
1 2 1
2
u (t)
1
u (t)
2
k y (t)
1
1
c y (t)
1
k (y (t) - y (t))
2 1 2
c (y (t) - y (t))


243

2. Automotive Shock Model

This is a car suspension model. M
2
is a point mass representing one-fourth of the car mass
(excluding wheels/tires). Spring constant k
2
and viscous dashpot coefficient c
2
represent the shock
spring and damping fluid friction dissipation elements. Point mass m
1
represents one wheel/tire mass
and spring constant k
1
and viscous dashpot coefficient c
1
represent the spring effect and energy
dissipation quality of the tire, respectively. The input is the road displacement z(t), and there are two
displacement outputs y
1
(t) and y
2
(t). The output y
2
(t) relates to passenger comfort. Since this is a
passive system, there are no force inputs.




( ) ( )
1 1 1 2 1 1 2 1 2 2 2 2 1 1
2 2 2 2 2 2 2 1 2 1
( ) ( ) ( ) ( ) ( ) ( ) ( )
( ) ( ) ( ) ( ) ( ) 0
m y t c c y t k k y t c y t k y t c z t k z t
m y t c y t k y t c y t k y t
+ + + + = +
+ + =





( )
( )
( )
( )
( )
( )
1 1 1 1 1 2 2 1 2 2 1 1
2 2 2 2 2 2 2 2
0 ( ) ( )
0 0
y t y t y t m c c c k k k c z t k z t
y t y t y t m c c k k
+ + + ( ( (
+ + =
` ` ` `
( ( (

)
) ) )





y (t)
m
1
1
1
c
m
2
1
k
y (t)
2
z(t)
2
c
2
k


244

3. Another Two-Mass Vibrational System




1 1 1 1 2 2 1
2 2 2 2 2 1
( ) ( ) [ ( ) ( )] ( )
( ) ( ) [ ( ) ( )] 0
m y t k y t k y t y t f t
m y t cy t k y t y t
+ =
+ + =





( )
( )
( )
( )
( )
( )
( )
1 1 1 1 1 2 2
2 2 2 2 2 2
0 0 0
0 0 0
y t y t y t m k k k f t
y t y t y t m k k c
+ ( ( (
+ + =
` ` ` `
( ( (

) ) ) )





y (t)
f(t)
1
y (t)
2
m
1
m
2
k
1
k
2
c


245

4. Inverted Pendulum




Coupled nonlinear differential equations

2
1 2 2 2
2
2 2 2
( ) ( ) cos ( ) ( ) sin ( ) ( ) ( )
( ) cos ( ) ( ) sin ( ) 0
m m w t m L t t m L t t f t
m L t m L t w t m gL t
u u u u
u u u
+ + =
=





Coupled linearized differential equations

1 2 2
2
2 2 2
( ) ( ) ( ) ( )
( ) ( ) ( ) 0
m m w t m L t f t
m Lw t m L t m gL t
u
u u
+ =
+ =




1 2 2
2
2 2 2
0 0 ( ) ( ) ( )
0 ( ) ( ) 0
m m m L w t w t f t
m L m L m gL t t u u
+ ( (
+ =
` ` `
( (

) ) )




Y
f(t)
X
g
L
m
w(t)
1
m
2
(t) u


246

5. Ball and Beam




Coupled nonlinear differential equations

2
2
2
( ) sin ( ) ( ) ( ) 0
( ) ( ) 2 ( ) ( ) ( ) ( )cos ( ) ( )
b
b
J
m p t m g t m p t t
r
m p t J J t m p t p t t m g p t t t
u u
u u u t
(
+ + =
(

( + + + + =




Coupled linearized differential equations

2
2
( ) ( ) 0
( ) ( ) ( ) ( )
b
b
J
m p t m g t
r
m p t J J t m g p t t
u
u t
(
+ + =
(

( + + + =




2
2
( ) 0 ( ) 0 0
( ) 0 ( ) ( ) 0 ( )
b
b
p t mg p t J r m
t mg t t m p t J J u u t
( + (
+ =
` ` ` ( (
+ +
) ) )




u
t
p
(
t
)
(t)
(t)


247

6. Proof-Mass Actuator




Coupled nonlinear differential equations

2
2
( ) ( ) ( ) ( ( )cos ( ) ( )sin ( )) 0
( ) ( ) ( )cos ( ) ( )
M m q t kq t me t t t t
J me t meq t t n t
u u u u
u u
+ + + =
+ + =





Coupled linearized differential equations

2
( ) ( ) ( ) ( ) 0
( ) ( ) ( ) ( )
M m q t kq t me t
J me t meq t n t
u
u
+ + + =
+ + =





2
( ) 0 ( ) 0
( ) 0 0 ( ) ( )
M m me q t k q t
me J me t t n t u u
+ ( (
+ =
` ` `
( (
+
) ) )





m
J
k
M
e
f(t)
q(t)
n(t)
(t) u


248

7. Two Masses suspended via Tensed Strings




1 1 1 2
2 2 2 1
2
( ) ( ) ( ) 0
2
( ) ( ) ( ) 0
T T
m x t x t x t
L L
T T
m x t x t x t
L L
+ =
+ =



( )
( )
( )
( )
1 1 1
2 2 2
0 2 0
0 2 0
x t x t m T L T L
x t x t m T L T L
( (
+ =
` ` `
( (

)
) )




m
L
x (t)
T
T
1
m
L
2
L
1
x (t)
2
L L L
T
T


249

8. Double Pendulum



Coupled nonlinear differential equations

2
1 2 1 1 2 1 2 2 1 2 1 1
2
2 1 2 1 2 2 2 2 2 2
( ) ( ) ( ) ( ) sin ( ) 0
( ) ( ) sin ( ) 0
m m L t m L L t m m gL t
m L L t m L t m gL t
u u u
u u u
+ + + + =
+ + =





Coupled linearized differential equations

2
1 2 1 1 2 1 2 2 1 2 1 1
2
2 1 2 1 2 2 2 2 2 2
( ) ( ) ( ) ( ) ( ) 0
( ) ( ) ( ) 0
m m L t m L L t m m gL t
m L L t m L t m gL t
u u u
u u u
+ + + + =
+ + =





2
1 2 1 1 1 2 1 2 1 2 1
2
2 2 2 2 1 2 2 2 2
( ) 0 ( ) 0 ( ) ( )
0 ( ) 0 ( )
m m gL t m m L m L L t
m gL t m L L m L t
u u
u u
+ ( + (
+ =
` ` ` ( (
) ) )




If m
1
=m
2
and L
1
=L
2


1 1
2 2
( ) 2 1 2 0 0 ( )
( ) 1 1 0 0 ( )
t g L t
t g L t
u u
u u
( (
+ =
` ` `
( (
) ) )



L
u
1
1
(t)
m
1
L
u
2
2
(t)
m
2
g


250

9. Sprung Double Pendulum



Coupled nonlinear differential equations

2 2
1 1 1 1 1 1 0 1 2
2 2
2 2 2 2 2 2 0 2 1
( ) sin ( ) ( ( ) ( )) 0
( ) sin ( ) ( ( ) ( )) 0
m L t m gL t kL t t
m L t m gL t kL t t
u u u u
u u u u
+ + =
+ + =




Coupled linearized differential equations

2 2
1 1 1 1 1 1 0 1 2
2 2
2 2 2 2 2 2 0 2 1
( ) ( ) ( ( ) ( )) 0
( ) ( ) ( ( ) ( )) 0
m L t m gL t kL t t
m L t m gL t kL t t
u u u u
u u u u
+ + =
+ + =




2 2 2
1 1 1 0 0 1 1 1
2 2 2
2 0 2 2 0 2 2 2
( ) 0 0 ( )
( ) 0 0 ( )
t m gL kL kL m L t
t kL m gL kL m L t
u u
u u
( ( +
+ =
` ` ` ( (
+
) ) )




If m
1
=m
2
and L
1
=L
2


2 2
0 0
1 1
2 2
2 2
0 0
( ) 1 0 0 ( )
( ) 0 1 0 ( )
L L g k k
t L m L m L t
t t
L L k g k
m L L m L
u u
u u
(
| | | |
+
(
| |
(
\ . \ .
(
+ =
` ` `
(
(
) ) ) | | | |
(
+
| |
(
\ . \ .




m
k
L
L
0
m
L
u
(t)
L
0
1
u
(t)
2


251

10. Doubly-sprung mass/cylinder




1 1 1
2
2 2
1 2 1
( ) ( ) ( ) 0
( ) ( ) ( ) ( ) 0
2
m x t k x t k r t
m r
t k k r t k rx t
u
u u
+ =
+ + =



1 1 1
2 2
2 1 1 2
0 ( ) ( ) 0
0 2 ( ) ( ) ( ) 0
m k k r x t x t
m r k r k k r t t u u
( (
+ =
` ` `
( (
+
) ) )




m
2
m
2
k
x(t)
r
u
(t)
1
k


252

11. Sprung Dual Cylinders



2
2 1 1
1 1 2 1 1 2 1 2 2
2
2 2 2
2 2 3 2 2 2 1 2 1
( ) ( ) ( ) ( ) 0
2
( ) ( ) ( ) ( ) 0
2
m r
t k k r t k r r t
m r
t k k r t k r r t
u u u
u u u
+ + =
+ + =




2
1 1
2
1 1 2 1 2 1 2 1
2 2
2 2 1 2 2 3 2 2 2 2
0
( ) 0 ( ) ( )
2
( ) 0 ( ) ( )
0
2
m r
t k k r k r r t
t k r r k k r t m r
u u
u u
(
(
( +
( + =
` ` ` (
+
(
) ) )
(




If m
1
=m
2
, and r
1
=r
2
, and k
1
=k
2
=k
3


1 1
2 2
4 2
( ) 1 0 0 ( )
( ) 0 1 2 4 0 ( )
k k
t t
m m
t k k t
m m
u u
u u
(

(
(
+ = (
` ` `
(
) ) ( )

(




3 m
2
k
r
u (t)
2
2
2
k
m
1
r
u (t)
1
1
1
k


253

12. Sprung Cart Pendulum




1 2 2
2
2 2 2
( ) ( ) ( ) 2 ( ) 0
( ) ( ) ( ) 0
m m x t m L t kx t
m L t m Lx t m Lg t
u
u u
+ + + =
+ + =





1 2 2
2
2 2 2
2 0 ( ) ( ) 0
0 ( ) ( ) 0
m m m L k x t x t
m L m L m Lg t t u u
+ ( (
+ =
` ` `
( (
) ) )




m
k
x(t)
u(t)
k
1
m
2


254

13. Translating rotating rigid body

This is a 2-dof system, still constrained to vibrate in the vertical direction, with small-angle
rotational vibrations as well.




1 2 1 1 2 2
2 2
1 1 2 2 1 1 2 2
( ) ( ) ( ) ( ) ( ) 0
( ) ( ) ( ) ( ) ( ) 0
mx t k k x t k L k L t
J t k L k L x t k L k L t
u
u u
+ + =
+ + =




1 2 1 1 2 2
2 2
1 1 2 2 1 1 2 2
0 ( ) ( ) 0
0 ( ) ( ) 0
k k k L k L m x t x t
k L k L k L k L J t t u u
+ + ( (
+ =
` ` `
( (
+ +
) ) )




m, J
k
u(t)
k
1 2
x(t)
L L
1 2


255

14. Doubly sprung cylinder mass




2 2
1 1 2 1 2 1 2 2
2 2
1 2 2 1 2 2
( ) ( ) ( ) ( ) ( ) ( ) 0
2 2
3
( ) ( ) ( ) ( ) 0
2 2
m m
m x t x t k k x t k x t
m m
x t x t k x t k x t
+ + + =
+ + =





2 2
1
1 1 2 2 1
2 2 2 2 2 2
( ) ( ) 0
2 2
( ) ( ) 3 0
2 2
m m
m
x t k k k x t
x t k k x t m m
(
+
(
+ (
+ = (
` ` `
(

) ) ) (




m
k
x (t)
r
u(t)
2
2
k
1
m
1
2
x (t)
1


256

15. Cylinder rolling in a sprung big cylinder




Coupled nonlinear differential equations

1 2 2 1 2
1 2
( ) ( ) ( ) ( )sin ( ) ( ) 0
3
( )sin ( ) ( ) ( ) sin ( ) 0
2
m m x t m r r t t kx t
x t t r r t g t
u u
u u u
+ + + =
+ + =




The linearized differential equations become decoupled.

1 2
1 2
( ) ( ) ( ) 0
3
( ) ( ) ( ) 0
2
m m x t kx t
r r t g t u u
+ + =
+ =




m
r
u(t)
2
2
r
1
k
m
1
x(t)


257

16. Dual torsional disks coupled by a spring




2
2 2 1 1
1 1 1 2
2
2 2 2 2
2 2 2 1
( ) ( ) ( ) ( ) 0
2
( ) ( ) ( ) ( ) 0
2
R
R
m r
t k ka t ka t
m r
t k ka t ka t
u u u
u u u
+ + =
+ + =




2
1 1
2 2
1 1 1
2 2 2
2 2 2 2 2
0
( ) 0 ( )
2
( ) 0 ( )
0
2
R
R
m r
t k ka ka t
t ka k ka t m r
u u
u u
(
(
( +
( + =
` ` ` (
+
(
) ) )
(




m
k
r
u (t)
a
1
1
k
R1
1
m
r
u (t)
a
2
2
k
R2
2


258

17. Dual point masses




1 1 1 2 1 2 2
2 2 2 2 2 2 1
( ) ( ) ( ) 2 ( ) ( )
( ) ( ) 4 ( ) 2 ( ) 0
m x t k k x t k x t f t
m x t cx t k x t k x t
+ + =
+ + =





1 1 1 1 2 2 1
2 2 2 2 2 2
0 ( ) ( ) 2 ( ) 0 0 ( )
0 ( ) ( ) 2 4 ( ) 0 0
m x t x t k k k x t f t
m x t x t k k x t c
+ ( ( (
+ + =
` ` ` `
( ( (

) ) ) )





k
c
m
x (t)
1
k
2
1
m
2
L L
1
x (t)
2
f(t)


259

18. Rotating sprung point masses




2 2
( ) 2 ( ) ( ) 0
2 ( ) 2 ( ) ( ) ( )
mx t kx t kL t
mL t kL t kLx t t
u
u u t
+ =
+ =




2 2
0 ( ) 2 ( ) 0
0 2 ( ) 2 ( ) ( )
m x t k kL x t
mL t kL kL t t u u t
( (
+ =
` ` `
( (

) ) )




k
m
x(t)
L L
k
m m
k
(t) t
(t) u


260

19. Three-Mass Vibrational System



1 1 1 2 1 1 2 1 2 2 2 2 1
2 2 2 3 2 2 3 2 2 1 2 1 3 3 3 3 2
3 3 3 4 3 3 4 3 3 2 3 2 3
( ) ( ) ( ) ( ) ( ) ( ) ( ) ( )
( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( )
( ) ( ) ( ) ( ) ( ) ( ) ( ) ( )
m y t c c y t k k y t c y t k y t u t
m y t c c y t k k y t c y t k y t c y t k y t u t
m y t c c y t k k y t c y t k y t u t
+ + + + =
+ + + + =
+ + + + =





1 1 1 2 2 1 1 2 2 1 1
2 2 2 2 3 3 2 2 2 3 3 2 2
3 3 3 3 4 3 3 3 4 3 3
0 0 ( ) 0 ( ) 0 ( ) ( )
0 0 ( ) ( ) ( ) ( )
0 0 ( ) 0 ( ) 0 ( ) (
m y t c c c y t k k k y t u t
m y t c c c c y t k k k k y t u t
m y t c c c y t k k k y t u
+ + ( ( (

( ( (
+ + + + =
` ` `
( ( (

( ( ( + +
) ) )


) t


`

)





y (t)
m
1
m
2
m
3
k
1
k
2
k
3
c
1
c
2
c
3
k
4
c
4
1
y (t)
2
y (t)
3
u (t)
1
u (t)
2
u (t)
3


261

20. Three Masses suspended via Tensed Strings




1 1 1 2
2 2 2 1 3
3 3 3 2
2
( ) ( ) ( ) 0
2
( ) ( ) ( ) ( ) 0
2
( ) ( ) ( ) 0
T T
m x t x t x t
L L
T T T
m x t x t x t x t
L L L
T T
m x t x t x t
L L
+ =
+ =
+ =




( )
( )
( )
( )
( )
( )
1 1 1
2 2 2
3 3 3
0 0 2 0 0
0 0 2 0
0 0 0 2 0
m x t T L T L x t
m x t T L T L T L x t
m x t T L T L x t
( (

( (
+ =
` ` `
( (

( (
)
) )




m
L
x (t)
T
T
1
m
L
2
L
2
x (t)
3
L L L
m
3
L
T
T
T
T
L
x (t)
1


262

21. Triple Pendulum




Let m
1
=m
2
=m
3
and L
1
=L
2
=L
3



1 2 3 1
1 2 3 2
1 2 3 3
3 ( ) 2 ( ) ( ) 3 ( ) 0
2 ( ) 2 ( ) ( ) 2 ( ) 0
( ) ( ) ( ) ( ) 0
L t L t L t g t
L t L t L t g t
L t L t L t g t
u u u u
u u u u
u u u u
+ + + =
+ + + =
+ + + =






1 1
2 2
3 3
3
0 0
3 2 1 ( ) ( ) 0
2
2 2 1 ( ) 0 0 ( ) 0
1 1 1 ( ) ( ) 0
0 0
g
L
t t
g
t t
L
t t
g
L
u u
u u
u u
(
(

(
(

(
(
+ =
` ` `
(
(

(
(
) )
)
(
(




L
u
1
1
(t)
m
1
L
u
2
2
(t)
m
2
L
u
3
3
(t)
m
3
g


263

22. Sprung Triple Pendulum




where L
0
=L / 2.


2 2
2
1 1 2
2 2 2
2
2 2 1 3
2 2
2
3 3 2
( ) ( ) ( ) ( ) 0
4 4
( ) ( ) ( ) ( ) ( ) 0
2 4 4
( ) ( ) ( ) ( ) 0
4 4
kL kL
mL t mgL t t
kL kL kL
mL t mgL t t t
kL kL
mL t mgL t t
u u u
u u u u
u u u
+ + =
+ + =
+ + =




1 1
2 2
3 3
0
4 4
1 0 0 ( ) ( ) 0
0 1 0 ( ) ( ) 0
4 2 4
0 0 1 ( ) ( ) 0
0
4 4
g k k
L m m
t t
k g k k
t t
m L m m
t t
k g k
m L m
u u
u u
u u
(
+
(

(
(

(
(
+ + =
` ` `
(
(

(
(
) )
)
(
+
(




m
k
L
L
0
m
L
u
(t)
L
0
m
k
L
L
0
1
u
(t)
2
u
(t)
3


264

23. Sprung Cylinder with Point Masses




1 1 1 2 1 1 2 2 2 2 3
1 2 2 2 2 1 1 2 1 2 3
3 3 2 3 2 1 2 2
3 ( ) ( 9 ) ( ) 2 ( ) 6 ( ) 3 ( ) 0
(2 ) ( ) 4 ( ) 2 ( ) 6 ( ) 2 ( ) 0
( ) ( ) 3 ( ) 2 ( ) 0
m x t k k x t m x t k x t k x t
m m x t k x t m x t k x t k x t
m x t k x t k x t k x t
+ + =
+ + + =
+ + =



( )
( )
( )
( )
( )
( )
1 1 1 1 2 2 2 1
1 1 2 2 2 2 2 2
3 3 2 2 2 3
3 2 0 9 6 3 0
2 2 0 6 4 2 0
0 0 3 2 0
m m x t k k k k x t
m m m x t k k k x t
m x t k k k x t
+ ( (

( (
+ + =
` ` `
( (

( (
)
) )




m
1
m
1
k
x (t)
r
u(t)
2
k
r/3
3
3
m
x (t)
2
2
x (t)
1


265

Common Model Features


Considering all multi-dof vibrational models we have derived, some general comments can be
made about the models.


1. When gathering the unknown i and its time derivatives on one side of the ODE i, there will
never be a change in sign amongst these terms; generally they will all be positive. Generally
the unknown j appears on the ODE i with a negative sign.


2. The effect of gravity always cancels out. That is, any mg static weight terms will load the
springs with static deflections and these terms always subtract out of the final ODE.


3. As always, the resulting units must be consistent. This is a good check to perform on your
resulting model.


4. Considering the matrix-vector forms of the coupled ODEs

a. The matrix multiplying the vector of second-order derivatives is called the mass
matrix. It is square, symmetric, and positive-definite.

b. The matrix multiplying the vector of first-order derivatives is called the damping
matrix. It is square and symmetric.

c. The matrix multiplying the vector of zeroth-order unknowns is called the stiffness
matrix. It is square and symmetric.