NCTU Department of Electrical and Computer Engineering

Senior Course <Dynamic System Analysis and Simulation>

By Prof. Yon-Ping Chen

6. First-Order LTI Systems

A. First-Order LTI systems
B. First-Order Filter: RC Circuit
Linear time-invariant systems, or briefly called LTI systems, are the most
important systems in engineering even though they are ideal, not real.

A. First-Order LTI Systems

The simplest dynamic system is a first-order LTI system shown in Figure 6-1.
Mathematically, it is expressed by a state equation and an output equation, given as
below:
xt ax bu t ,

x(0)=x0

(61) y cx du
(6-2)

where x(t) is the state variable representing the system intrinsic feature and x(0)=x0 is
its initial state at the initial time t=0. Besides, the system is excited by the external
input u(t) to generate the output response y(t) accordingly.
d
u(t)

+ x(t)

x(t)

y(t)

Figure 6-1

From the state equation (1), the system state x(t) for t 0, starting from the initial
time t=0, can be obtained as
xt e x t e
at

and then the output equals to

a t

bu d

(6-3)

y t ce x c t e
at

a t

bu d du t

which is measurable.
6-1

(6-4)

NCTU Department of Electrical and Computer Engineering

Senior Course <Dynamic System Analysis and Simulation>
By Prof. Yon-Ping Chen

In addition to the model described by (1) and (2), an LTI system could be
expressed by a differential equation without the state variable x(t). First, lets calculate
the following equation:
y t ay t cx t du t acx t du t
c x& t ax du& t
adu t cbu du& adu

(6-5)

du& t cb ad
where the term xt axt is replaced by bu(t) according to (6-1). Clearly, the
system in Figure 6-1 can be changed into a first-order differential equation, called the
input-output equation and expressed by
y& a 0 y b u& t b0 u t ,

y(0)= y0

(6-6)

where a0 a, b1=d, b0=cb-ad and y0=cx0+du(0). It is known that if b1=d0 then (6-6)
is a proper system and if b1=d=0 then it is a strictly proper system, given as
y& a 0 y b0 u t ,

y(0)= y0

(6-7)

Figure 6-2 shows the block diagram of (6-7) correspondingly.

u(t)

b0

a0

y(t)

Figure 6-2

Obviously, the state equation and output equation of a first-order strictly proper
LTI system is described as
xt ax bu t ,

x(0)=x0

(68) y cxt
(6-9)

and the block diagram is shown in Figure 6-3. Since most of the systems in
engneering are strictly proper, we will focus on this knid of systems from now on.
u(t)

+ x(t)

x(t)

y(t)

6-2

Figure 6-3

NCTU Department of Electrical and Computer Engineering

Senior Course <Dynamic System Analysis and Simulation>
By Prof. Yon-Ping Chen

B. First-Order Filter: RC Circuit

In engineering, we often face problems related to signal processing, which
usually requires a system, called filter, to extract desired signals from the input u(t).
Figure-4 shows the simplest low-pass filter implemented by an RC circuit, which only
allows signals of low frequency to pass through.
R

iC(t)
+

vs(t)

vC(t)

Figure 6-4
Now, lets derive the state equation by choosing the capacitor voltage vC(t) as
the state variable due to its continuity in time domain. According to the Kirchhoffs
voltage law, we have
R iC t vC t v s ,

vC(0)=vC0

(6-10)

where vC(0)=vC0 is the initial voltage of the capacitor. The current through a capacitor
is related to the voltage across it as below:
iC t Cv C

(6-11)

By substituting (6-11) into (6-10), we have

v&C

RC

vC

RC

v s ,

vC(0)=vC0

(6-12)

Let the state and input be x(t)=vC(t) and u(t)=vs(t), then (6-12) is rewritten as
xt ax bu t ,
where a

1
R
C

and b

RC

x(0)=x0=vC0

(6-13)

. To be used as a low-pass filter, the capacitor voltage

is selected as the output, i.e.,

y vC t x

(6-14)

which is equal to the state variable. Hence, from (6-13) we can write its input-output
equation of the RC circuit as below:
y a 0 y t b0 u t ,

y(0)=y0=vC0

(6-15)

where a 0

RC

and b0

RC

. From (6-3), we have

6-3

NCTU Department of Electrical and Computer Engineering

Senior Course <Dynamic System Analysis and Simulation>
By Prof. Yon-Ping Chen

y t x t e a0tvC 0 t e a0 (t

)b0 u d

e RCtvC 0
e
RC t RC(t
where the function e

RC

)u d

(6-16)

is shown in Figure 6-5 and T=RC is called the time

constant. Obviously, if the time t increases by one time constant T, then the function
decreases in a ratio of e 1=0.3679. Moreover, the larger the time constant T is, the
slower the function decreases. Note that e

RC

0 as t

. Therefore, when the time

increases, the output (6-16) will be gradually approximated as below:

y t xt

1
RC

RC

)u d

(6-17)

which depends on the input u(t) only. As a result, the initial condition y0=vC0 is usually
neglected in filter design. Next, lets check the frequency response of the RC
circuit by the use of sinusoidal input u(t)=cos t.

1
RC

0.3679
0.1353
0.0498
T

2T

3T

t
0.0183
4T

Figure 6-5

One important tool of frequency response is the Laplace transform. Let U(s) and
Y(s) be the Laplace transform of u(t) and y(t), then (6-15) can be transformed as
sY s a0Y s b0U s

(6-18)

Y s

(6-19)

i.e.,

where H s

s a

s a

0 U s H s U s

0 is called the transfer function. If s is replaced by j , then

6-4

NCTU Department of Electrical and Computer Engineering

Senior Course <Dynamic System Analysis and Simulation>
By Prof. Yon-Ping Chen

H s s

H j

where the magnitude of H j

Hj

j Hj

H j e

(6-20)

is

1
1
b0

2
j RC 1
j a0
1 RC

(6-21)

and the phase is

Hj

tan

RC

(6-22)

It is well known that if the input is u(t)=cos t, then its output is

y t H j

cos t

Hj

(6-23)

which implies the output possesses the same frequency

Hj

magnitude and phase are changed by |H(j )| and

From (6-21), |H(j )| is decreased as the frequency

RC

as the input signal, but its

shown in Figure 6-6.

is increased. Bisides, when

, we have

Hj

1+

Hj

tan

c RC

(6-24)

c RC

45

(6-25)

and then
y t H j

c cos t

Hj

cos t

45

(6-26)

which implies the signals power, in a form concerning y t , at

2

to one half. Hence,

is called the half power frequency or the cutoff frequency.

To sum up, an input u(t)=cos t with lower frequency

will generate an output with

larger magnitude |H(j )|. In other words, the filter H s

s a

0 acts as a low-pass

filter since it allows input signals of lower frequency to pass through easier.
u(t)
cos t

is reduced

H(s)

y(t)
Hj

Figure 6-6

cos t

Hj

Lets take a numerical example to demonstrate the feature of the RC filter in

6-5

NCTU Department of Electrical and Computer Engineering

Senior Course <Dynamic System Analysis and Simulation>
By Prof. Yon-Ping Chen

Figure 6-4, where R=1k , C=10 F and the initial capacitor voltage is 1V. From
(6-12), we write the input-output equation as below:
v&C 100v C 100v s t ,

vC(0)=1

What are the output responses for u(t)=cos t, where

(6-26)
=1, 10, 100 and 100? The

numerical answer is obtained by MATLAB and given as below:

=========================================
Create m-file: RC1.m
function dy=RC1(t,y)
dy=-100*y+100*cos(t);
Create m-file: RC2.m
function dy=RC2(t,y)
dy=-100*y+100*cos(10*t);
Create m-file: RC3.m
function dy=RC3(t,y)
dy=-100*y+100*cos(100*t);
Create m-file: RC4.m
function dy=RC4(t,y)
dy=-100*y+100*cos(1000*t);
>> % key in the following instructions
>> [t,y]=ode23(@RC1,[0:0.01:6],1.0)
>> plot(t,y); xlabel(t); ylabel(y(t))

1.5
1
0.5
0
-0.5
-1
-1.5

3
t

>> % key in the following instructions

>> [t,y]=ode23(@RC2,[0:0.01:6],1)
>> plot(t,y); xlabel(t); ylabel(y(t))

1.5
1
0.5
0
-0.5
-1
-1.5

3
t

>> % key in the following instructions

6-6

NCTU Department of Electrical and Computer Engineering

Senior Course <Dynamic System Analysis and Simulation>
By Prof. Yon-Ping Chen

>> [t,y]=ode23(@RC3,[0:0.01:6],1)
>> plot(t,y); xlabel(t); ylabel(y(t))

1.5
1
0.5
0
-0.5
-1
-1.5

3
t

>> % key in the following instructions

>> [t,y]=ode23(@RC4,[0:0.01:6],1)
>> plot(t,y); xlabel(t); ylabel(y(t))

1.5
1
0.5
0
-0.5
-1
-1.5

3
t

=====================================================

The numerical result shows that the RC circuit indeed act as a low-pass filter
since the magnitude of output is decreased as the input frequency

is increased.

Problems
P.6-1 If the RC filter in Figure 6-4 contains R=16k , C=40 F and neglect the initial
capacitor voltage. What is the cutoff frequency
responses for u(t)=cos t for =0.1 c,

c,

10

and what are the output

and 100 c?

P.6-2 Replace the capacitor in Figure 6-4 by an inductor L. If R=16k L=10mH and
neglect the initial condition, then what is the cutoff frequency
the output responses for u(t)=cos t for =0.1 c,

6-7

c,

10

and what are

and 100 c?