174 C H A P T E R 3: The Laplace Transform
the Laplace transform to test for convergence, we lets=σ+jand the term|ej|=1. Thus, all
regions of convergence will contain−∞< <∞.
If{σi}are the real parts of the poles ofF(s)=L[f(t)], the region of convergence corresponding to
different types of signals or impulse responses is determined from its poles as follows:
n For a causalf(t),f(t)=0 fort<0, the region of convergence of its Laplace transformF(s)is a
plane to the right of the poles,
Rc={(σ,):σ >max{σi},−∞< <∞}
n For an anti-causalf(t),f(t)=0 fort>0, the region of convergence of its Laplace transformF(s)
is a plane to the left of the poles,
Rac={(σ,):σ <min{σi},−∞< <∞}
n For a noncausalf(t)(i.e.,f(t)defined for−∞<t<∞), the region of convergence of its Laplace
transformF(s)is the intersection of the regions of convergence corresponding to the causal
component,Rc, andRaccorresponding to the anti-causal component:
Rc
⋂
Rac
See Figure 3.5 for an example illustrating how the ROCs connect with the poles and the type of signal.
Special case:The Laplace transform of a functionf(t)of finite supportt 1 ≤t≤t 2 , has the wholes-plane as
ROC.
FIGURE 3.5
ROC for (a) causal signal with poles with
σmax= 0 ; (b) causal signal with poles with
σmax< 0 ; (c) anti-causal signal with poles with
σmin> 0 ; (d) noncausal signal where ROC is
bounded by poles (poles on the left-hand
s-plane give causal component and poles on
the right-hands-plane give the anti-causal
component of the signal). The ROCs do not
contain poles, but they can contain zeros.
(a)
jΩ
o o
×
×
× σ
(b)
jΩ
o
×
×
σ
(c)
jΩ
o
×
o
o σ
(d)
jΩ
o
o
×
×
×
× σ