Digital Signal Processing Reference
In-Depth Information
L
Mg
Figure 1.9
Simple pendulum.
rapidly is large. Hence, the differentiator amplifies any high-frequency noise in
However, some practical applications require the use of a differentiator. For
these applications, some type of high-frequency filtering is usually required before
the differentiation, to reduce high-frequency noise.
v
i
(t).
We now consider a differential-equation model of the simple pendulum, which is
illustrated in Figure 1.9. The angle of the pendulum is denoted as the mass of the
pendulum bob is
M
, and the length of the (weightless) arm from the axis of rotation
to the center of the bob is
L
.
The force acting on the bob of the pendulum is then
Mg
, where
g
is the gravi-
tational acceleration, as shown in Figure 1.9. From physics we recall the equation of
motion of the simple pendulum:
u,
ML
d
2
u(t)
dt
2
=-Mg sin u(t).
(1.13)
This model is a
second-order nonlinear differential equation;
the term
sin u(t)
is non-
linear. (Superposition does not apply.)
We have great difficulty in solving nonlinear differential equations; however,
we can
linearize
(1.13). The power-series expansion for
sin u
is given (from
Appendix D) by
u
3
3!
+
u
5
5!
-
Á
.
sin u = u -
(1.14)
For small, we can ignore all terms except the first one, resulting in when
is expressed in radians. The error in this approximation is less than 10 percent for
, is less than 1 percent for and
decreases as becomes smaller. We then express the model of the pendulum as,
from (1.13) and (1.14),
u
sin u L u
u
u = 45°
(p/4 radians)
u = 14°
(0.244 radians),
u
d
2
u(t)
dt
2
g
L
u(t) = 0
+
(1.15)