Module H: Laplace Transform

x(t) = est y(t) = H(s)est

h(t)

Z 1
st
H(s) = h(t)e dt.
1

For any CT signal x(t), it’s Laplace-transform is the function:

Z 1
X(s) = x(t)e st dt = L{x(t)}.
1

The Region of Convergence of X(s) is defined to be

Z 1 Z 1
{s | |x(t)||e st |dt = |x(t)|e Re(s)t dt < 1}.
1 1

Letting ↵ = Re(s) and ! = Im(s), we have:

Z 1 Z 1
st
X(s) = x(t)e dt = x(t)e ↵t e !t
dt
1 1
Z 1
= (x(t)e ↵t )e !t dt = FT(x(t)e ↵t
).
1

↵t
Therefore, if s = ↵ + !j 2 S, X(s) = FT(x(t)e ).
If ↵ = 0, X(s) = FT(x(t)).
↵t
Even when x(t) is not absolutely integrable, it may be possible for x(t)e
to be absolutely integrable if ↵ is sufficiently large.
Notation: x(t) ! X(s)

1
 Laplace Transform

at
Example 1: What is the Laplace Transform and ROC of x(t) = e u(t)
where a 2 C?

Example 2: What is the Laplace Transform and ROC for x(t) = eat u( t)?

Moral: Like z-transform, Laplace-Transform without ROC is meaningless!

Example 3: What is the Laplace Transform and ROC for x(t) = cos(t) +
e2t u(t)?

2
 Laplace Transform and Z-Transform

Properties of ROC of Laplace transform for CT signals is closely related to

those of Z-transform:
1. Replace circles with Re(s) = constant lines
2. Replace outside circle with right-half plane
3. Replace inside circle with left-half plane
4. Likewise z-Transform, we almost exclusively work with rational Laplace
transforms:
a0 (s z1 ) · · · (s zm )
X(s) = ,
(s p1 ) · · · (s pn )
where z1 , . . . , zm 2 C are the zeros and p1 , . . . , pn 2 C are the poles.
5. Location of the poles are important for ROC and stability
2s 1
Example: What are the poles and zeros of X(s) = s2 +2s 1 ?

3
 Properties of ROC for Laplace Transform

Property 1: ROC does not contain any poles.

Property 2: If x(t) is of finite duration and x(t) is absolutely integrable, i.e.,
Rb
a |x(t)|dt < 1, then ROC is the entire C.

– To show this, let s 2 C with Re(s) = ↵. Then:

Z 1 Z b
↵t
|x(t)|e dt = |x(t)|e ↵t dt
1 a
Z b
↵t
 max(e : t 2 [a, b]) |x(t)|dt < 1.
a

Property 3: If x(t) is right-sided and s0 2 ROC, then the whole half-plane

{s : Re(s) > Re(s0 )} is in ROC.
Property 4: If x(t) is left-sided and s0 2 ROC, then the half-plane s :
Re(s) < Re(s0 ) is in ROC.
Property 5: If x(t) is two-sided and s0 is in ROC, then there exist a <
Re(s0 ) < b such that the strip s : a < Re(s) < b is in ROC.
Property 6: If X(s) is rational, then ROC is either bounded between two
poles or extends to infinity.
Property 7: If X(s) is rational and right-handed, then the ROC is right-half
plane to the right-most pole.
Property 8: If X(s) is rational and left-handed, then the ROC is left-half
plane to the left-most pole.

4
 Inverse Laplace Transform

Suppose that the line {s|Re(s) = 1} 2 ROC

imag

real
-4 -3 -2 -1 0 1 2 3 4

+j! t
We know that along this line X(e ) = FT(e x(t)).
Therefore,
Z 1
1
x(t) = t IFT(X(e +j! )) = t X(e( +j!)t
)ej!t d!
2⇡ 1
Z +j1
1
= X(s)est ds.
2⇡j j1

5
 Inverse Laplace Transform

q(s)
For rational transfer functions X(s) = p(s) , we proceed as in the case of
z-transform:
a. We perform partial fraction on
q(s) A B
X(s) = = + + ···
p(s) s p1 s p2
pi t 1
b. If pole pi is to the left side of ROC, we use e u(t) ! s pi
pi t 1
c. If pole pi is to the right side of ROC, we use e u(t) ! s pi
s
Example: What is the Laplace Transform inverse of X(s) = s2 +5s+6 given
that ROC= {s | 3 < Re(s) < 2}?

6
 Causality and Stability

For causality, we know that the impulse response h(t) is right-sided, there-
fore, ROC is to the right-side of the right-most pole
R1
For stability of such a system, we we need 1 |h(t)|dt < 1
This means that the j! axis, i.e., {s | Re(s) = 0}, be in the ROC of H(s).
A causal system with rational H(s) is stable if and only if all the poles are
to the left of the j! axis.
Determine causality and stability of
s 1
H(s) = .
(s + 1)(s 2)

7
 Properties of Laplace Transform

Linearity: ax1 (t) + bx2 (t) ! aX1 (s) + bX2 (s), ROC contains the inter-
section.
st0
Time-Shift: x(t t0 ) !e X(s). ROC is not a↵ected.
Frequency Shift: es0 t x(t) ! X(s s0 ). ROC now includes Re(s0 ).
Convolution: x1 (t) ⇤ x2 (t) ! X1 (s)X2 (s). ROC contains the intersection.
dx(t)
Di↵erentiation: dt ! sX(s).

Di↵erentiation: tx(t) ! dX(s)

ds .
R0
Integration: 1 x(t)dt ! 1s X(s).
Initial Value Theorem: For a causal signal x(t) which does not contain an
impulse at the origin, we have

x(0+ ) = lim sX(s).

s!1

Final value theorem: For a causal signal x(t) for which limt!1 x(t) is finite,
we have
lim x(t) = lim sX(s).
t!1 s!0

8
Finding Impulse Response of Di↵erential Equations

9
Block Diagram Representation

10
Feedback System

11