EE2023 Signals & Systems Revision Notes: 1 Circuit Elements and Their Models
EE2023 Signals & Systems Revision Notes: 1 Circuit Elements and Their Models
Resistors
VR (t) i(t) R
Time Domain : i(t) =
R
Frequency Domain : I(s) =
VR (s)
+ VR(t) _
R
Capacitors
dVc (t)
Time Domain : i(t) = C
dt
∫ t i(t) +
1
Vc (t) = i(τ )dτ C Vc(t)
C −∞
_
I(s)
Frequency Domain : Vc (s) =
sC
Vc (s) 1
Zc (s) = =
I(s) sC
Inductors
di(t)
Time Domain : VL (t) = L
dt L
∫ t i(t)
1
i(t) = VL (τ )dτ
L −∞
+ VL(t) _
Frequency Domain : VL (s) = sLI(s)
VL (s)
ZL (s) = = sL
I(s)
AP Loh Page 1
EE2023 Signals & Systems Revision
Resistance : Rtotal = R1 + R2 + . . . + Rn
Series Combination : 1 1 1
Impedances : Zc,total (s) = + + ... +
sC1 sC2 sCn
ZL,total (s) = sL1 + sL2 + . . . + sLn
1 1 1 1
Resistance : = + + ... +
Rtotal R1 R2 Rn
Parallel Combination : 1
Impedances : Zc (s) =
sC1 + sC2 + . . . + sCn
1 1 1 1
= + + ... +
ZL (s) sL1 sL2 sLn
Zc (s) and ZL (s) are complex impedances ie they are complex numbers which behave like
resistances except that their impedance values change with frequency when s = jω where
ω is the frequency of operation. For example, in the case of a capacitor with capacitance
1
C Farad, its impedance value is Zc (s = jω) = jωC
Ω and therefore, this impedance
value, Zc (jω) will decrease as ω increases. At ω = 0 or at DC, the capacitor operates
like an open circuit since Zc (j0) = ∞. A similar behaviour can be said of an inductor
with inductance L Henry except that the impedance of an inductor is proportional to
ω. Hence an inductor has zero impedance at DC. This is in contrast to resistances, R,
which have real values which are not affected very much by frequencies of operation.
Applying Kirchoff’s Current and Voltage Laws in Figure 1, and working in the frequency
domain, we have :
V2 (s) = V3 (s)
I(s) = I1 (s) + I2 (s)
Z3 (s) V (s) = V1 (s) + V2 (s)
I1 (s) = I(s) Z1 (s)
Z2 (s) + Z3 (s) V1 (s) = V (s)
Z2 (s) Z1 (s) + ZT otal (s)
I2 (s) = I(s) ZT otal (s)
Z2 (s) + Z3 (s) V2 (s) = V (s)
Z1 (s) + ZT otal (s)
AP Loh Page 2
EE2023 Signals & Systems Revision
ZTotal
i(t) Z1
+ _ i1(t) i2(t)
+ V1(t) + +
_ _
_
1 1 1
ZTotal Z2 Z3
These principles are useful in analysing circuits containing resistors, inductors and ca-
pacitors. The above equations, however, are only valid in the frequency domain. In time
domain, differential equations are needed to model the currents and voltages in such circuits.
2 Complex Numbers
z = x + jy = Rejθ = R∠θ
where x and y are real quantities. R > 0 is also the magnitude of z, denoted as |z| = R and
θ is the phase of z in radians. Unless otherwise stated, x and y are assumed positive in the
following sections. I trust that you are able to prove the above relationships.
Complex numbers play a very important part in the analysis of many electrical systems
as you may already have seen in EG1108 and elsewhere. Hence it is very important that
the manipulation of complex numbers are done correctly. In particular, you have to be fully
familiar in converting between the different forms of the complex numbers. This is especially
crucial when you deal with frequency response plots involving complex numbers and other
calculations involving magnitudes and phases.
AP Loh Page 3
EE2023 Signals & Systems Revision
Some tips for ensuring the correct calculation of phase of a complex quantity.
• z = j × j = −1 : |z| = 1 and ∠z = −π or π.
√ −1 y
• z = x + jy : z = x2 + y 2 ej tan x . Thus
√ y
|z| = x2 + y 2 and ∠z = tan−1 .
x
Im
z
X jy
|z
|
z
x Re
• z= 1
x+jy
: |z| = 1
|x+jy|
=√1
and ∠z = ∠1 − ∠(x + jy) = 0 − tan−1 y
x
= − tan−1 xy .
x2 +y 2
AP Loh Page 4
EE2023 Signals & Systems Revision
√
• z = −x + jy : |z| = x2 + y 2 and ∠z = tan−1 y
−x
= − tan−1 xy . Calculators will
generally give arctan values between −0.5π < θ < 0.5π which means that the phase
angles given by calculators are only in the first or fourth quadrant. However, in this case
when z = −x + jy, the complex number is in the second quadrant since the real part
is negative while the imaginary part is positive. Therefore, the value obtained from the
calculator will not be the correct phase and has to be adjusted as follows :
y y
∠z = π − tan−1 or − (π + tan−1 ).
x x
√
• z = x − jy : |z| = x2 + y 2 and
−y y
∠z = tan−1 = − tan−1 (1)
x x
The phase given by the calculator will be correct since this complex number lies in the
fourth quadrant.
{ −y
}
• z= 1
x−jy
: |z| = √ 1
and ∠z = − tan−1 x
. It is not easy to see what the phase
x2 +y 2
should be without further manipulation as follows :
1
z =
x − jy
1 x + jy
= ×
x − jy x + jy
x + jy x y
= 2 2
= 2 2
+j 2 (2)
x +y x +y x + y2
1
Hence in this case, although z = x−jy
, z is actually a complex number in the first
quadrant, deducing from (2) where both the real and imaginary parts are positive. Thus
{ }
−1 −y y
∠z = − tan = tan−1 .
x x
1
z =
x − jy
1 1
= −jθ
= ejθ
Re R
where −θ is the angle in the fourth quadrant as in (1) above. Thus ∠z = θ which is in
the first quadrant : same conclusion earlier.
AP Loh Page 5
EE2023 Signals & Systems Revision
3 Laplace Transforms
Laplace Transforms (LT) are an integral part of systems and control. It is used widely in
solving ordinary differential equations (ODE) by transforming them into algebraic equations
involving the complex variable s. Via the LT, the time domain ODEs are converted into
algebraic equations in frequency domain where s is the frequency variable. This also gives
the fundamental link between time and frequency domains of signals and systems. Via the
frequency domain, many properties of linear time invariant systems can be described and
characterised without the need to solve the original ODEs. This leads to a generalization of
system behaviour for this class of systems.
where F (s) and f (t) are Laplace Transform pair. The lower limit of 0− is required in cases
where there are non-zero initial conditions. Example on how the Laplace integral is used :
∫ ∞
L{sin ωt} = sin ωte−st dt
−
∫0 ∞
1 jωt
= {e − e−jωt }e−st dt
− 2j
∫0 ∞
1 −(s−jω)t
= {e − e−(s+jωt) }dt
− 2j
0
[ ]∞
1 1 −(s−jω)t 1
−(s+jω)t
= e + e −
2j −s + jω s + jω 0
ω
= 2
s + ω2
Laplace transforms are linear operators which has the following properties. Note that F (s)
and G(s) are used to denote the Laplace Transforms of f (t) and g(t).
AP Loh Page 6
EE2023 Signals & Systems Revision
An example : } {
d2 f (t) − df
L = s F (s) − sf (0 ) −
2
dt2 dt t=0−
{∫ }
t
• Transform of an Integral : L 0 f (τ )dτ = F (s) s
signal, f(t),
1 starting at t=0
0.8
signal f(t)
0.6
0.2
0
−10 0 10 20 30 40 50 60 70 80 90 100
time
Notice how a shifted time function is written. Its Laplace transform has a simple rela-
tionship with the original LT of f (t). This can be shown as follows :
∫ ∞
L{f (t − t0 )} = f (t − t0 )e−st dt
∫0 ∞
= f (t − t0 )e−s(t−t0 ) e−st0 dt
0
∫ ∞
−st0
= e f (t − t0 )e−s(t−t0 ) dt
0
= e−st0 F (s) after making the substitution λ = t − t0
AP Loh Page 7
EE2023 Signals & Systems Revision
You should be able to prove this following the steps in the delayed signal case.
• Final Value Theorem : For a time domain function which has a finite steady state value,
the final value theorem is given as :
Proof :
∫ ∞
df (t) df (t) −st
L{ } = e dt = sF (s) − f (0)
dt dt
∫ ∞ 0
df (t) −st
lim e dt = lim[sF (s) − f (0)]
s→0 0 dt s→0
∫ ∞
df (t)
lim dt = lim[sF (s) − f (0)]
s→0 0 dt s→0
This is a convenient formula to use when you want to determine the steady state value
of a system’s output response. You can obtain this steady state value from F (s) instead
of from f (t). Applicable only when the f (t) has a constant steady state value. Not
applicable when f (t) is always changing with time. Examples where the final value
theorem fails is when f (t) = sin ωt, f (t) = t, etc.
Fortunately, we do not need to evaluate the LT integral each time we want to find the Laplace
Transform of a function f (t), at least for the kind of f (t) that are generally encountered in
systems and control. We rely on tables to determine them. Some examples are given in
Table 1.
AP Loh Page 8
EE2023 Signals & Systems Revision
1
1 s
1
t s2
2
t2 s3
n!
tn sn+1
e−at 1
s+a
ω
sin ωt s2 +ω 2
s
cos ωt s2 +ω 2
ω
eat sin ωt (s−a)2 +ω 2
s−a
eat cos ωt (s−a)2 +ω 2
e−at t 1
(s+a)2
Inverse Laplace Transform is used to recover the time domain signal from its s-domain
equivalent. The inverse Laplace Transform is defined as :
∫ ∞
−1 1
L {F (s)} = f (t) = F (s)est ds
2jπ −∞
It can be seen that this is a complex integral. In most cases, we do not need to find the
inverse Laplace Transform using this integral. We should be able to use some well known
Laplace Transform pairs such as those in Table 1 to help us find the inverse.
1
Example 1 : Find the inverse Laplace Transform of .
s2 + 2s + 4
AP Loh Page 9
EE2023 Signals & Systems Revision
{ } { }
−1 1 −1 1
L = L
2
s + 2s + 4 (s + 1)2 + 4 − 1
{ }
−1 1
= L
(s + 1)2 + 3
{ √ }
1 −1 3
= √ L √
3 (s + 1)2 + ( 3)2
{ √ }
1 −1 3
= √ L √ where s1 = s + 1
3 s21 + ( 3)2
Since
ω { −at }
L{sin ωt} = and L e f (t) = F (s + a),
s2 + ω 2
it follows that { }
1 1 √
L 2
= √ e−t sin 3t
s + 2s + 4 3
Laplace Transforms which are encountered in linear systems theory are of the form :
N (s)
F (s) =
D(s)
where N (s) and D(s) are polynomials in the s-variable. Examples are F (s) = 1/s, F (s) =
s+1 1
(s+1)2 +3
, F (s) = s+2
, etc. For simple numerator and denominator polynomials of order 2 or
lower, the inverse Laplace Transform of F (s) can be obtained by inspection from Table 1.
However, when N (s) and D(s) are polynomials of higher order, the inverse may not be so
easily obtainable from the table.
The approach to obtain the inverse LT of a general F (s) is to first partial factorize F (s)
so that the partial factors consists of only polynomials of orders at most 2. The inverse LT
of F (s) is then the inverse LT of each partial factor which can be easily deduced from Table
1. This approach works because of the linearity property of the LT.
AP Loh Page 10
EE2023 Signals & Systems Revision
N (s) A1 A2 An
F (s) = = + + . . . +
(s + α)n s + α (s + α)2 (s + α)n
N (s) A1 s + B 1 A2 s + B 2
F (s) = 2 2
= 2 + 2
(s2 2
+ 2β1 s + γ1 )(s + 2β2 s + γ2 ) (s + 2β1 s + γ1 ) (s + 2β2 s + γ22 )
2
For such linear factors, the general form of the inverse LT is given by :
(√ ) (√ )
−β1 t −β2 t
f (t) = R1 e sin γ1 − β1 t + ϕ1 + R2 e
2 2
sin γ2 − β2 t + ϕ2 .
2 2
The Ri are functions of βi and γi . In this quadratic form, it should be presumed that
each quadratic factor admits complex roots.
Each of these different forms of partial factors require a different approach to determine Ai .
These methods are briefly given as follows :
N (s) A1 A2 An
F (s) = = + + ... +
(s + α1 )(s + α2 ) . . . (s + αn ) s + α1 s + α2 s + αn
2
Example : Find f (t) from its LT F (s) = (s+1)(s+2)
.
2 A1 A2
F (s) = = +
(s + 1)(s + 2) s+1 s+2
AP Loh Page 11
EE2023 Signals & Systems Revision
2 2
F (s) = −
s+1 s+2
f (t) = L−1 {F (s)} = 2e−t − 2e−2t
N (s) A1 A2 An
F (s) = n
= + 2
+ ... +
(s + α) s + α (s + α) (s + α)n
dn−k
where dsn−k
(.) denotes differentiating with respect to s, (n − k) times.
2
Example : Find f (t) from its LT F (s) = (s+1)(s+2)2
.
2 A1 A2 A3
F (s) = 2
= + 2
+
(s + 1)(s + 2) s + 1 (s + 2) (s + 2)
N (s) A1 s + B 1 A2 s + B 2
F (s) = 2 2
= 2 + 2
(s2 2
+ 2β1 s + γ1 )(s + 2β2 s + γ2 ) (s + 2β1 s + γ1 ) (s + 2β2 s + γ22 )
2
AP Loh Page 12
EE2023 Signals & Systems Revision
As + B As + B
= √ √
(s2 2
+ 2βs + γ ) (s + β + j γ 2 − β 2 )(s + β − j γ 2 − β 2 )
P Q
= √ + √
(s + β + j γ − β ) (s + β − j γ 2 − β 2 )
2 2
P and Q can be obtained in the same way as those with distinct linear factors.
10 1
△ Find the inverse Laplace Transform of 2
and 2 .
s(s + 2)(s + 3) s +s+1
Consider the first order RC circuit given in Figure 4. Assume a general input voltage v(t)
i(t) R
v(t) C vc(t)
t=0
and an output voltage vc (t) across the capacitor. Deriving the model in time domain using
differential equation and by applying circuit laws :
Equation (3) may be solved using mathematical techniques and it can also be solved using
Laplace Transform. Using the LT approach, since the input voltage is a general input v(t),
the LT of v(t) is written generally as V (s) while the LT of the output voltage vc (t) is written
AP Loh Page 13
EE2023 Signals & Systems Revision
V RCvc (0)
Vc (s) = +
s(RCs + 1) RCs + 1
= V + [vc (0) − V ] e− RC
t
(5)
From (5), it is clear that limt→∞ vc (t) = V which is the final value which the capacitor will
be charged up to when a DC voltage of V is applied. This same result can be obtained by
applying the Final Value Theorem to Vc (s) as follows :
{ }
V RCvc (0)
lim vc (t) = lim s + = V.
t→∞ s→0 s(RCs + 1) RCs + 1
AP Loh Page 14
EE2023 Signals & Systems Revision
v (0)
c
V V
v (0)e−t/RC+V(1−e−t/RC)
total output
c
vc(0) e−t/RC
+ t/RC
V(1−e )
=
0
0
time time
time
Notice how the initial condition vc (0) decays over time. The plot in the middle is the
typical charging characteristic of a capacitor with zero initial condition. Finally, the sum
total of the first two plots gives the actual charging trajectory of the capacitor, including
the non-zero initial condition.
AP Loh Page 15