World Academy of Science, Engineering and Technology
International Journal of Mechanical and Mechatronics Engineering
Vol:8, No:9, 2014
A Model for Analyzing the Startup Dynamics
of a Belt Transmission Driven by a DC Motor
Giovanni Incerti
AbstractIn this paper the vibration of a synchronous belt drive
during start-up is analyzed and discussed. Besides considering the
belt elasticity, the model here proposed also takes into consideration
the electromagnetic response of the DC motor. The solution of the
motion equations is obtained by means of the modal analysis in
state space, which allows to obtain the decoupling of all equations,
International Science Index, Mechanical and Mechatronics Engineering Vol:8, No:9, 2014 waset.org/Publication/9999272
without introducing the hypothesis of proportional damping. The
mathematical model of the transmission and the solution algorithms
have been implemented within a computing software that allows the
user to simulate the dynamics of the system and to evaluate the effects
due to the elasticity of the belt branches and to the electromagnetic
behavior of the DC motor. In order to show the details of the
calculation procedure, the paper presents a case study developed with
the aid of the above-mentioned software.
KeywordsBelt drive, Vibrations, Startup, DC motor.
I. I NTRODUCTION
I N many industrial devices the motion transmission between
parallel axes is obtained by means of belt drive systems,
which allow to obtain low noise operation, good mechanical
efciency and low cost design solutions.
However, for high dynamic loads, a belt transmission can
introduce unwanted vibrations on the devices to be operated;
as is well known, these phenomena largely depend on the
elasticity of the branches of the belt and occur especially for
multi-stage congurations arranged in series.
Based on these considerations, it is important to verify,
already at the design stage, the vibration of the transmission
in its real operating conditions. Furthermore, the study is more
accurate if the mathematical model of the system takes into
account, at the same time, the deformation of the belt and
the electromagnetic behavior of the motor. In this context, the
paper aims to provide a contribution on the subject, presenting
a model of the complete electromechanical system, in which
the motion equations related to the mechanical components
(belt drive ad pulleys) are solved together with the equation
that describes the electrical response of the motor.
In order to clarify the proposed approach we will analyze
a typical conguration, which consists of a DC motor, a belt
drive and a resistive load, paying particular attention to the
possibility of analytically solve the differential equations of
motion, without using a numerical integration method.
II. M ATHEMATICAL MODELING OF THE SYSTEM
To evaluate the dynamics of belt transmissions, taking
into account the axial deformability of the branches, lumped
parameter models are usually adopted; using this approach a
system of ordinary differential equations is written and solved.
G. Incerti is with the Department of Mechanical and Industrial Engineering,
University of Brescia, 25123 Brescia, Italy, (e-mail: giovanni.incerti@unibs.it).
International Scholarly and Scientific Research & Innovation 8(9) 2014
m Jm
m
z 1 1
J1 r1
1
k c c k k c
r
2
r2
2
2
J2
Fig. 1. Schematic representation of the electromechanical system analyzed
in the paper. Each branch of the belt is modeled using a viscoelastic model
(spring and damper in parallel).
This approach, which allows to limit the complexity of the
mathematical model, has been used in several works found in
the technical literature: for example, in [1] the authors analyze
belt transmissions with complex geometries and propose an
algorithm for the automatic deduction of the motion equations
of a system having an arbitrary number of belts and pulleys.
In [2] a lumped parameter model has been developed for a
belt-driven robot, in order to study some dynamic effects and
to optimize the control strategy of the machine.
A study of dynamic characteristics of a differential,
planetary, path-generating mechanism with a synchronous belt
is presented in [3]; by accounting for the belt elasticity
and employing a Lagrangian approach, a dynamic model of
the mechanism is obtained; the effects of belt drive speed
ratio, belt material damping and planet link balancing on the
mechanism behavior are investigated by the authors.
In general the motion equations for a mechanical system
can be derived by imposing conditions of dynamic equilibrium
or by using the Lagrange equations: usually the following
assumptions are considered:
the belt branches are modeled by linear springs in parallel
with viscous dampers;
the belt mass is negligible compared to the pulleys mass;
the motion transmission takes place in ideal kinematic
conditions (absence of slippage, due to the use of a
synchronous belt);
the belt has no exural stiffness.
The stiffness k of the belt branches can be determined by the
well-known relationship k = EA/l, where E is the Youngs
1537 scholar.waset.org/1307-6892/9999272
 World Academy of Science, Engineering and Technology
International Journal of Mechanical and Mechatronics Engineering
Vol:8, No:9, 2014

modulus of the belt material, A is the cross sectional area R L

I
of the branch and l its length; the damping coefcient c can
be experimentally determined through free vibration tests (for RI LI&
+
example using the logarithmic decrement method [4]).
As a case study, we consider the system shown in Fig. 1, V (t ) zk m & 1
consisting of a DC motor, a gear speed reducer, a synchronous
belt drive and a resistant load, which provides a torque linearly _
variable as a function of speed.
As regards the mechanical part, we observe that the system
has two degrees of freedom (DOF), represented by the
angular coordinates 1 e 2 ; the rotation m of the motor Fig. 2. Armature circuit of the DC motor.
is proportional to the rotation 1 according to the assigned
transmission ratio assigned z (m = z1 ) and therefore it
does not represent an additional degree of freedom. At this point we made explicit (6) with respect to the angular
International Science Index, Mechanical and Mechatronics Engineering Vol:8, No:9, 2014 waset.org/Publication/9999272

To analyze the dynamics of the system we use the Lagrange accelerations 1 e 2 and (7) with respect to the time derivative
equations, which, for the case under consideration, assume the of the armature current; nally we introduce the two identities
following form: 1 = 1 e 2 = 2 , obtaining in this way the following system
  of differential equations:
d T T D U W
+ + = i = 1, 2 (1)
dt i i i i i
1 = 1


where the symbols T , D, U and W respectively indicate
2 = 2


the kinetic energy of the system, the Rayleigh dissipation
1

1 = (2kr12 1 + 2kr1 r2 2 2cr12 1 +
function, the potential energy due to the elastic deformation
J1


of the belt branches and the work done by the external actions +2cr1 r2 2 + zkm I)
(8)
(motor torque and the load torque) acting on the system. The
1

= [2kr r 2kr 2
+
complete list of the symbols used in the equations is given in

2
J2
1 2 1 2 2


Appendix C.
+2cr1 r2 1 (2cr22 + )2 r0 ]


For a permanent magnet DC motor the torque m is
proportional to the current I that ows in the armature circuit; I = 1 [zkm 1 RI + V (t)]

L
the load torque r acting on the system can be considered as
the sum of a constant term r0 and linearly variable term with Using matrix notation, (8) can be rewritten in the form:
the angular speed 2 . Therefore will be valid the following
x(t) = Ax(t) + Bu(t) (9)
relationships:
m = k m I r = r0 + 2 (2) where:
0 0 1 0 0

The virtual work done by the motor and the load torque is: 0 0 0 1 0

W = zm 1 r 2 (3) 2kr2
1
2kr1 r2 2cr2
1
2cr1 r2 zkm
J1 J1 J1

J1 J1
Using the sign conventions shown in Fig. 1 the kinetic energy A= (10)
2kr1 r2 2kr22 2cr1 r2 2cr22 +
can be written as: 0
J2 J2 J2 J2
1 zkm R
T = (J1 21 + J2 22 ) (4) 0 0
L
0
L
2
where J1 = J1 +z 2 Jm . The potential energy and the Rayleigh
0 0
dissipation function assume the following expressions: 0 0
0 0
U = k(r1 1 r2 2 ) 2
D = c(r1 1 r2 2 ) 2
(5) B=

1
0

(11)
J2
Substituting the above expressions into (1) and calculating the 1
0
required derivatives we obtain the motion equations of the L
device shown in Fig. 1:  T
x(t) = 1 (t) 2 (t) 1 (t) 2 (t) I(t) (12)

J 1 + 2cr12 1 2cr1 r2 2 +
1 +2kr12 1 2kr1 r2 2 = zkm I
(6)  T

J 2cr1 r2 1 + 2cr22 2 2kr1 r2 1 + u(t) = r0 V (t) (13)
2 2
+2kr22 2 = (r0 + 2 )
As is known, A and x are respectively the state matrix and
As regards the electric behavior of the motor, the application
the state vector of the system, u is the input vector and B
of the Kirchhoffs voltage law to the armature circuit of the
the input matrix. The differential equations here obtained are
motor gives (see Fig. 2):
linear and for their solution the analytical method described
V (t) = RI + LI + zkm 1 (7) in the next section can be used.

International Scholarly and Scientific Research & Innovation 8(9) 2014 1538 scholar.waset.org/1307-6892/9999272
 World Academy of Science, Engineering and Technology
International Journal of Mechanical and Mechatronics Engineering
Vol:8, No:9, 2014

III. A NALYTICAL SOLUTION IN STATE SPACE Now we introduce the linear transformation:
An analytical solution for the system of equations (9) can be x(t) = X(t) (22)
found using the modal analysis in state space; this procedure
requires the calculation of two different types of eigenvectors where = [1 2 3 4 5 ]T is the vector containing the modal
[5] [6]. We proceed by considering rst the problem of free coordinates i (t). Substituting (22) into (9) and premultiplying
vibrations (or homogeneous problem), obtained by imposing by YT both sides of the equation, we obtain:
u(t) = 0; from (9) we get: YT X(t) = YT AX(t) + YT Bu(t) (23)
x(t) = Ax(t) (14)
Considering (20), we can rewrite the above equation in the
Equation (14) represents in matrix form a system of form:
ve rst-order ordinary differential equations with constant (t) = (t) + q(t) (24)
coefcients, whose solution is:
where:
x(t) = xet (15) q(t) = YT Bu(t) (25)
International Science Index, Mechanical and Mechatronics Engineering Vol:8, No:9, 2014 waset.org/Publication/9999272

where x is a constant vector and a constant scalar. By In this way we have obtained the complete decoupling of the
substituting (15) into (14) and simplifying the exponential term differential equations that describe the system dynamics, since
on both sides, we obtain: the matrix equation (24) contains ve independent equations,
which can be separately solved. Each of them is a rst
Ax = x (16) order nonhomogeneous linear differential equation, that can
Equation (16) represents the eigenvalue problem with a be written as:
nonsymmetric real matrix A. The solution of (16) gives
i (t) = i i (t) + qi (t) i = 1, . . . , 5 (26)
the eigenvalues i and the corresponding eigenvectors xi ,
(i = 1, . . . , 5). These eigenvalues can be real or complex. The solution of (26) can be expressed in the general form:
Since A is a real matrix, if i is a complex eigenvalue,  t
i t
its complex conjugate i will also be an eigenvalue of A; i (t) = i (0)e + qi ( )ei (t ) d i = 1, . . . , 5 (27)
moreover the eigenvectors xi and xi , corresponding to i and 0

i , will also be complex conjugates to one another.
Now let us consider the eigenvalues-eigenvectors problem
for the transposed matrix AT :
T
A y = y
Since the determinants of A and AT are equal, (16) and (17)
generates the same characteristic equation, i.e.:
|A I| |AT I| = 0
where I is the 5 5 identity matrix. Consequently the
eigenvalues of A and AT are identical; on the contrary the
eigenvectors of these matrices will be different.
The eigenvectors xi , i = 1, . . . , 5 of A are called right
eigenvectors, whereas the eigenvectors yj , j = 1, . . . , 5 of
AT are called left eigenvectors; more details about these
denominations are given in Appendix A.
The right eigenvalues xi and the left eigenvalues yi are the
columns of the matrices X and Y, which are dened by the
following relationships:
X = [x1 x2 x3 x4 x5 ] Y = [y1 y2 y3 y4 y5 ] (19)
For calculating the initial conditions i (0) in terms of modal
coordinates, it is sufcient to invert (22) and to set t = 0:
(0) = X1 x(0) (28)
(17) 1 T
From the rst of (20) we have X = Y and therefore
the calculation of the inverse of matrix X can be replaced by
the transposition of matrix Y; this simplies the computation
process. Therefore we can rewrite (28) in the form:
(18)
(0) = YT x(0) (29)
After determining the modal coordinates by means of (27),
the solution of the system (9) can be obtained using the linear
transformation given by (22).
IV. S IMULATION RESULTS
The previously described model has been used for
simulating the dynamic behavior of the belt transmission
shown in Fig. 1 during startup. To perform the simulations, it
was assumed that the DC motor was controlled with a voltage
signal having the following exponential form:

If the eigenvectors X and Y are properly normalized, the V (t) = Vmax (1 et/ ) (30)
following equations are valid (see Appendix B for details):
where Vmax and indicate, respectively, the maximum
YT X = I YT AX = (20) value of the voltage command and the time constant of the
exponential function. The complete list of parameters used
where is the 5 5 diagonal matrix of eigenvalues, i.e.: for the calculation is shown in Table I.

1 0 0 0 0 In order to solve the differential equations, the
0 2 0 0 0 method of modal analysis in state space has been

= 0 0 3 0 0 (21) employed, using numerical techniques [7] for solving
0 0 0 4 0 the eigenvalue-eigenvector problems and for calculating the
0 0 0 0 5 integral in (27).

International Scholarly and Scientific Research & Innovation 8(9) 2014 1539 scholar.waset.org/1307-6892/9999272
 World Academy of Science, Engineering and Technology
International Journal of Mechanical and Mechatronics Engineering
Vol:8, No:9, 2014

TABLE I
S YSTEM PARAMETERS USED FOR NUMERICAL SIMULATIONS The results obtained by this method are in agreement with
those obtained by Runge-Kutta algorithm, which, although of
Jm = 4 104 kgm2 r0 = 0 Nm easier application, allows only a numerical evaluation of the
J1 = 0.018 kgm2 = 5 Nms/rad solution.
J2 = 0.25 kgm2 z = 15 The proposed method has been implemented in a software
k = 120 kN/m Vmax = 24 V package and it can be easily extended to the study of similar
c = 20 Ns/m R = 0.15 systems with lumped parameters, where the vibrations are due
r1 = 80 mm L = 0.2 mH to the compliance of the mechanical components (belt, joints,
r2 = 160 mm km = 0.09 Nm/A
etc.); in general, assuming that the system under considerations
has n degrees of freedom, its description in the state space
will require the use of 2n + 1 variables, since, in addition to
the mechanical coordinates, it is necessary to add the current
The exactness of the results has been veried by comparing owing in the armature circuit of the electric motor.
the solution obtained by modal approach with the solution
International Science Index, Mechanical and Mechatronics Engineering Vol:8, No:9, 2014 waset.org/Publication/9999272

resulting from the numerical integration of the system

A PPENDIX A
equations (9); the numerical computation has been performed
R IGHT AND L EFT EIGENVECTORS
using the fourth order Runge-Kutta method, with step size
t = 0.1ms. The denominations right eigenvalues and left eigenvalues
Fig. 3 show the time histories of the following variables: are due to the position of the eigenvectors relative to the matrix
motor armature voltage (Fig. 3a), armature current (Fig. 3b), A. In fact, looking at the rst member of the equation (16)
electric and mechanical power (Fig. 3c, d), angular velocity we note that the eigenvector x is to the right of A;
of the pulleys (Fig. 3e, f), angular acceleration of the pulleys Considering instead the equation (17) and calculating the
(Fig. 3g, h). transpose of both members, we get: yT A = yT . The rst
The plots have been calculated for three different values of member of the equation thus obtained clearly shows that the
the time constant (1 = 20 ms, 2 = 40 ms, 3 = 60 ms), eigenvector y (in transposed form) is located to the left of the
using a maximum motor voltage Vmax = 24 V. matrix A. In Linear Algebra the eigenvalue problem for AT
The numerical simulations here presented show that the is called the adjoint eigenvalue problem and the eigenvectors
mathematical model of the belt transmission can be used yj , j = 1, . . . , 5 are known as adjoint eigenvectors of the
during the design stage to simulate with good reliability eigenvectors xi , i = 1, . . . , 5.
the dynamic phenomena occurring during the system startup;
clearly it is necessary to carry out the validation of the A PPENDIX B
model by means of experimental activity, especially as regards P ROOF OF E QUATIONS (20)
the determination of the damping parameters of the system. To give a proof of (20), we begin by rewriting (16) for
The identication of the system parameters represents an the generic right eigenvector xi and (17) for the generic left
interesting perspective of research that could be developed in eigenvector yj :
the future through the use of an instrumented test rig, which
allows the user to detect by experiments the most important Axi = i xi AT y j = j y j (31)
physical parameters (motor voltage and current, driving torque, Considering the second of (31) and transposing both members
angular velocity and angular acceleration of the pulleys), that we get:
are necessary for the model validation. yjT A = j yjT (32)
Pre-multiplying the rst of (31) to yjT and post-multiplying
V. C ONCLUSIONS
(32) to xi , we obtain:
The paper presented the dynamic analysis during startup
yjT Axi = i yjT xi yjT Axi = j yjT xi (33)
of a synchronous belt drive with non-negligible elasticity.
The study was carried out through the formulation of Subtracting the second of the (33) from the rst, we have:
a mathematical model which considers also the electrical
response of the DC motor. To simplify the system modeling, (i j )yjT xi = 0 (34)
a matrix approach in the state space was used. If i = j, the eigenvalues i and j are distinct and thus their
As regards the mechanical aspects, the analysis was carried difference is not null; therefore from (34) we deduce that:
out with the traditional Lagrangian method; for the 2 DOF
system under consideration, this method allowed to write a yjT xi = 0 i, j = 1, . . . , 5 (35)
system of two second order linear differential equations with
Equation (35) states that, for a given matrix A and for
constant coefcients. The addition of (7) required the insertion
i = j , the right eigenvectors xi are orthogonal to the left
of a row and a column in the state matrix.
eigenvectors yj .
To simulate the system dynamics the modal approach in the Substituting (35) into the rst or second of the (33) we get:
state space was used, which allowed to decouple the equations
and to obtain the solution of the problem in analytical terms. yjT Axi = 0 i, j = 1, . . . , 5 (36)

International Scholarly and Scientific Research & Innovation 8(9) 2014 1540 scholar.waset.org/1307-6892/9999272
 World Academy of Science, Engineering and Technology
International Journal of Mechanical and Mechatronics Engineering
Vol:8, No:9, 2014

(a) (b)
30 60
= 20 ms
= 40 ms
= 60 ms
Motor voltage [V]

Motor current [A]

20 40

10 20

= 20 ms
= 40 ms
= 60 ms

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5

time [s]
 time [s]
International Science Index, Mechanical and Mechatronics Engineering Vol:8, No:9, 2014 waset.org/Publication/9999272

(c) (d)
1000 800
= 20 ms = 20 ms
= 40 ms = 40 ms
800 = 60 ms = 60 ms

Mechanical power [W]

600
Electrical power [W]

600

400

400

200
200

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5

time [s]
 time [s]

(e) (f)
20 10
Ang. velocity - pulley (1) [rad/s]

Ang. velocity - pulley (2) [rad/s]

8
15

10

5
= 20 ms 2 = 20 ms
= 40 ms = 40 ms
= 60 ms = 60 ms

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5

time [s]
 time [s]

(g) (h)
400 300
= 20 ms = 20 ms
Ang. acceleration - pulley (1) [rad/s]

Ang. acceleration - pulley (2) [rad/s]

= 40 ms = 40 ms
300 = 60 ms = 60 ms
200

200

100

100

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5

100 100

time [s] time [s]

Fig. 3. Simulation results: a) Motor voltage (command signal); b) Armature current; c) Electric power at the motor terminals; d) Mechanical power at the
motor shaft; e) Angular velocity 1 ; f) Angular velocity 2 ; g) Angular acceleration 1 ; h) Angular acceleration 2 .

International Scholarly and Scientific Research & Innovation 8(9) 2014 1541 scholar.waset.org/1307-6892/9999272
 World Academy of Science, Engineering and Technology
International Journal of Mechanical and Mechatronics Engineering
Vol:8, No:9, 2014

This relationship shows that the eigenvectors xi and yj are R EFERENCES

orthogonal also with respect to the state matrix A.
For i = j we obtain from (33) the following result:
yiT Axi = i yiT xi
Generally yiT xi = 1, but, for each pair of eigenvectors yi e xi
we can always nd a normalization coefcient i such that:
(i yiT )(i xi ) = 1
According to this equation the coefcient i will be:
1 Traduzione italiana a cura di A. Barozzi, Pitagora Editrice (1995).
i = 
yiT xi
Multiplying by i the components of the vectors yi and xi , we
International Science Index, Mechanical and Mechatronics Engineering Vol:8, No:9, 2014 waset.org/Publication/9999272
[1] Iwatsubo T., Hasegawa K., Arii S., Shiohata K., The Formulation and
Dynamic Analysis of a Multiple Belt System, Journal of Sound and
Vibration, 205(3), 293-307 (1997).
i = 1, . . . , 5 (37) [2] Lee K.M., Rutherford C., Frequency reshaped quadratic control of a
low-cost human level performance belt-driven robot, Mechatronics, 9,
95-110 (1999).
[3] Al-Dwairi A.F., Al-Lubani S.E., Modeling and dynamic analysis of
a planetary mechanism with an elastic belt, Mechanism and Machine
(38) Theory 39, 343-355 (2004).
[4] Kelly S.G., Mechanical Vibrations - Theory and applications, Cengage
Learning (2012).
[5] Kramer E., Dinamica delle macchine - Calcolo delle Vibrazioni,
i = 1, . . . , 5 (39) [6] Meirovich L., Fundamentals of Vibrations, McGraw-Hill Higher
Education (2001).
[7] Press W.H., Teukolsky S.A., Vetterling W.T., Flannery B.P., Numerical
Recipes in C, Cambridge University Press (1992).

get new normalized eigenvectors which satisfy the relation1 :

yiT xi = 1 i = 1, . . . , 5 (40)
According to (40), we rewrite (37) in the form:
yiT Axi = i i = 1, . . . , 5 (41)
Using the matrices dened by (19), considering the
orthogonality property (35) and taking into account the
normalization condition (40), we can write:
YT X = I (42)
Finally, from (41) we have:
YT AX = (43)
where is dened by (21).

A PPENDIX C
L IST OF SYMBOLS Giovanni Incerti graduated in Mechanical
Engineering in 1990 at the University of Brescia,
Pulley (1) rotation angle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Italy. At the same university he received the Ph.D
Pulley (2) rotation angle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 in Applied Mechanics in 1995. Actually he is
Motor shaft rotation angle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . m associate professor at the University of Brescia
and his teaching activity is related to the courses
Kinetic energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . T of Mechanical Vibrations, Applied Mechanics and
Potential energy due to system elasticity . . . . . . . . . . . . . . . . . . . . . . . . . U Vibration Control. He is author of papers dealing
Rayleigh dissipation function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . D with the dynamic analysis of cam systems, the
Work done by motor and resistant torques . . . . . . . . . . . . . . . . . . . . . . . W mathematical modeling of devices for industrial
automation, the mechanism design by optimization
Motor torque . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . m techniques and the study of robots and servomechanisms. He took part in
Load torque . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . r researches funded by the Italian Council of Researches (CNR) and by the
Constant component of the load torque . . . . . . . . . . . . . . . . . . . . . . . . . r0 Italian Ministry of University and Research. Recently he has been scientic
Viscous friction coefcient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . coordinator for the University of Brescia inside the National research project
Design and experimental validation of cam transmissions.
Moment of inertia of the motor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Jm
Moment of inertia of pulley (1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . J1
Moment of inertia of pulley (2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . J2
Axial stiffness of the belt branches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . k
Damping constants of the belt branches . . . . . . . . . . . . . . . . . . . . . . . . . . . c
Pulley (1) radius . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . r1
Pulley (2) radius . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . r2
Gear ratio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . z

Armature voltage of the DC servomotor . . . . . . . . . . . . . . . . . . . . . . . . . . V

Maximum armature voltage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Vmax
Armature current . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . I
Armature resistance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . R
Armature inductance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . L
Torque/back e.m.f. costante fo the DC motor . . . . . . . . . . . . . . . . . . . . km
Time constant of the exp. function (used as command voltage) . . . . .

1 For convenience, the new normalized eigenvectors are indicated with the
same symbols yi e xi used for the non-normalized eigenvectors.

International Scholarly and Scientific Research & Innovation 8(9) 2014 1542 scholar.waset.org/1307-6892/9999272