Transfer Functions and Transfer Characteristics

This document was prepared as review material for students in EE 230

By: Randy Geiger
Last Updates: Jan 16, 2010
Electronic circuits and electronic systems are designed to perform a wide variety
of tasks. The performance requirements from task to task are often significantly
different. Although the performance requirements vary widely, there are considerable
benefits from both design and assessment viewpoints of having standard methods for
characterizing the performance of these systems. The concepts of transfer characteristics
and transfer functions are used extensively to characterize these circuits and systems.
An electronic system, an electronic circuit, or a more general system that may
have some or no electrical relationships are often characterized by the transfer functions
and the transfer characteristics as well and the concepts are universal. And within the
context of electronic circuits and systems, terminology such as circuit, network,
amplifier, filter, system, architecture, structure, along with several other terms are used
interchangeably. The difference between an amplifier, a circuit, a filter, a network, an
architecture, or even a system is often only in how the “entity” is intended to be used with
possible no difference in how the “entity” is characterized or in how it operates. In what
follows, these terms will be used interchangeably but the concepts of transfer functions
and transfer characteristics are universal.

Transfer Characteristics
A system with an input XIN and an output XOUT is shown in Fig. 1. In electrical
systems the input and output quantities, typically termed signals, are often dependent on a
single additional input variable, time. In this case the input and output signals would
time-dependent voltages or currents. In Fig. 2, three different methods of indicating a
system that serves as a voltage amplifier with time-dependent input and output voltages is
shown. For simplicity, the notation in part c) of this figure is most commonly used.

IN OUT

Fig. 1 Single-Input Single-Output System

Page 1 of 14
Fig. 2 Representations of a Voltage Amplifier

The transfer characteristics of a system is defined to be the pseudo-static relationship

between the input and output variable. If the system is a voltage-in, voltage-out system
we would term this pseudo-static relationship the dc transfer characteristics. Three
different transfer characteristics are shown in Fig. 3. In top part of the figure, the
relationship between the input and output is a straight line. In the middle part, it is
weakly nonlinear. And, in the bottom part it is highly nonlinear.

Page 2 of 14
XOUT

XIN

XOUT

XIN

XOUT

XIN

Fig. 3 Transfer characteristics of linear, weakly nonlinear, and highly nonlinear systems

Linear Systems
If the relationship between the input and output can be expressed as
XOUT = KXIN (1)
where K is a constant, we say the system is linear. Actually, the class of linear systems
extends somewhat beyond the case where K is constant. The following definition defines
a linear system:

Definition: A system is linear iff

XOUT ( a1XIN1+a2 XIN2 ) = XOUT (a1XIN1 )+XOUT (a2 XIN2 ) (2)
where a1 and a2 are any real numbers and XIN1 and XIN2 are any two inputs.

With this definition of a linear system, the relationship of (1) is satisfied. But the I/O
characteristics of many linear systems do not satisfy (1). For example, if sinusoidal inputs
at two different frequencies are applied to a linear system, the outputs due to each will
also be a sinusoid but the relative magnitude of the input and output sinusoidal signals
may be different. If a system is linear, the relationship of (1) is satisfied for dc inputs.
But, it is possible to have a system that is nonlinear where (1) is satisfied for dc inputs.

Transfer Functions
Any linear system is characterized by a transfer function. A linear system also
has transfer characteristics. But, if a system is not linear, the system does not have a
transfer function. The following definition will be used to define a transfer function.

Page 3 of 14
Definition: The transfer function, T(s), of a linear system with input XIN and
output XOUT is given by the expression
X
T ( s ) = OUT
( s) (3)
X IN( s )
where XOUT(s) and XIN(s) are the Laplace transforms of the output and input.

Although the input signal and correspondingly the output signal can take on arbitrary
values, the ratio of the Laplace Transforms of these two quantities is not dependent upon
the particular input that is applied provided the system is linear.
The input/output relationship defined by (3) is often termed a frequency domain
characterization of the circuit The relationship between the actual input/output and the
frequency domain input/output is depicted in Fig. 4.

Fig. 4 Time-domain and frequency-domain transformations

The frequency domain characterization of a linear system and correspondingly the

transfer function is of particular use in determining the sinusoidal steady state response of
the network. A key theorem, and one of the major reasons that the frequency domain was
studied in EE 201, follows.

Theorem 1: If a linear network has transfer function T(s) and input given by the
expression XIN(t)=XMsin(ωt + θ), then the steady state output is given by
XOUT ( t ) = XM T ( jω ) sin ( ωt + θ + ∠T ( jω ) ) (4)

This theorem states the steady state output is a sinusoid of the same frequency as the
excitation but scaled in magnitude by the magnitude of the transfer function evaluated at
s=jω and shifted in phase by the phase of the transfer function evaluated at s=jω.
Although this theorem is useful, there are still some challenges that need to be
overcome to simplify the sinusoidal steady state analysis of linear networks. The
challenge is in obtaining the transfer function T(s). The straightforward way to obtain
T(s) from (3) is to write a set of differential equations relating the input and output
variables of a circuit and then take the Laplace Transform of this set of equations to
obtain a set of transformed equations. These equations become algebraic and can be
solved to obtain T(s). But this process is still very tedious. A more practical approach is
to transform the time domain circuit itself to a frequency domain circuit and then analyze
the frequency domain circuit. This circumvents the need for writing any differential
equations.

Page 4 of 14
Two methods of obtaining the transfer function will be described here. One is a
transformation of the circuit into what is often termed an s-domain or “Laplace domain”
circuit. The other uses phasors. The methods are almost identical but circuits and
electronics textbooks typically treat these as different approaches so both will be
described.

s-domain circuit analysis

The concept of transforming the time-domain circuit to the s-domain circuit is

depicted in Fig. 5. The s-Domain Circuit is topologically identical to the

Fig. 5 Transformation from time-domain circuit to s-domain circuit

time-domain circuit, the only difference is how the circuit elements are modeled. All
elements in the two circuits are identical except for the capacitors and inductors. The
inductors and capacitors are simply replaced with impedances given in Fig 6.

1
sL
sC

Fig.6 Impedances of Inductors and Capacitors in s-domain network

Mathematically, these mappings of an element to an impedance can be expressed as

1
C→ (5)
sC
L → sL (6)
After the s-domain circuit is obtained, it can be analyzed using standard circuit analysis
techniques to obtain the transfer function T(s). Once T(s) is obtained, Theorem 1 can be
used to obtain the sinusoidal steady state response.

Phasor-domain circuit analysis

The concept of transforming a circuit to the phasor-domain is shown in Fig. 7. .

Page 5 of 14
Fig. 7 Transformation from time-domain circuit to phasor-domain circuit

The phasor-domain circuit is also topologically identical to the time-domain circuit, the
only difference being in how the circuit elements are modeled. All elements are the same
except for the inductors and capacitors. The inductors and capacitors are simply replaced
with impedances given in Fig.8.

1
jωL
jωC

Fig.8 Impedances of Inductors and Capacitors in phasor-domain network

Mathematically, these mappings of an element to an impedance can be expressed as

1
C→ (7)
jωC
L → jωL (8)
After the phasor-domain circuit is obtained, it can be analyzed using standard circuit
analysis techniques to obtain the phasor-domain transfer function TP(jω). Whereas the
transfer function T(s) includes the variable s and thus is neither a real or complex
quantity until further information about s is given, the phasor-domain transfer function is
a complex quantity as are all impedances in the phasor-domain circuit. As such,
manipulations of equations leading to TP(jω) can be made as part of the analysis process.
Unfortunately this additional flexibility, when exercised, often causes unnecessary
arithmetic calculations when calculating TP(jω).
Since the phasor-domain circuit and the s-domain circuits differ only in how the
energy storage elements are characterized and since this characterization is similar, it
follows that
T ( jω ) = T ( s ) s = jω = TP ( jω ) (9)
Thus, the sinusoidal steady state response can also be obtained from the phasor-domain
transfer function TP(jω). From (), it is apparent that TP(jω) can be readily obtained from
T(s). Obtaining T(s) from TP(jω) is not so easy to do but seldom would one want to do
this anyway.

Page 6 of 14
Relationship between time-domain, s-domain and phasor-domain
analysis

The process of obtaining the sinusoidal steady-state response has been

discussed using the concept of the transfer function, Theorem 1, and either the s-
domain circuit or the phasor-domain circuit. In this section, the relationship between
these two methods of analysis will be contrasted with using the time domain analysis.
The block diagram in Fig. depicts three methods for obtaining the sinusoidal steady
state response of a linear system. The central path shows the time-domain approach. It
involves writing the differential equations that characterize the actual circuit and solving
these equations to obtain the response. The left path uses the s-domain approach and the
right path uses the phasor-domain approach. The s-domain approach and the time-
domain approach can be used even if the input is not a sinusoid though the analysis
becomes more complicated. The phasor-domain approach is only applicable to when the
input is a sinusoidal function and the steady-state response is desired.

Fig.9 Time-domain, s-domain and phasor-domain analysis of linear circuits

It should be noted in the block diagram that the last blocks, the inverse transforms, were
not discussed in the previous section. The inverse transforms needed to obtain the
sinusoidal steady state response are provided by Theorem 1 and the inverse transforms
are given by (4).

Example – The three approaches to sinusoidal steady state

analysis

Page 7 of 14
Consider the circuit if Fig. 10. We will obtain the sinusoidal steady state response
using the three methods of analysis, the s-domain approach, the phasor-domain approach,
and the time-domain approach. In all cases, it will be assumed that
VIN = VMsin ( ωt+θ ) (10)
and the goal is to obtain the steady state response VOUT(t).

Fig. 10 Example time-domain circuit

s-domain analysis
From Fig.6 the s-domain equivalent of the circuit of Fig.10 can be obtained. This
is shown in Fig. 11.

1
sC

Fig. 11 s-domain circhit

It follows from KCL by summing currents at the output node that

⎛ 1⎞ ⎛ 1⎞
V OUT ⎜ sC+ ⎟ = VIN ⎜ ⎟ (11)
⎝ R⎠ ⎝R⎠
Solving for VOUT in terms of VIN, we obtain the transfer function
1
T ( s) = (12)
1+RCs

Page 8 of 14
Evaluating at s= jω we obtain
1
T ( jω ) = (13)
1+jωRC
The magnitude and phase of T(jω) are given by
1
T ( jω ) = ∠T ( jω ) = − tan −1 ( ωRC ) (14)
1+ ( ωRC )
2

It thus follows from (10) and Theorem 1 that the steady state output is given by
⎛ ⎞
1 ⎟ sin ( ωt + θ − tan −1 ( ωRC ) )
VOUT ( t ) = VM ⎜ (15)
⎜ 1+ ωRC 2 ⎟
⎝ ( ) ⎠

phasor-domain analysis
From Fig.8, the phasor-domain equivalent of the circuit of Fig.10 can be
obtained. This is shown in Fig. 12

1
jωC

Fig.12 Phasor-domain circuit

It follows from KCL by summing currents at the output node that

K ⎛ 1⎞ K ⎛ 1⎞
V OUT⎜ jωC+ ⎟ = VIN ⎜ ⎟ (16)
⎝ R⎠ ⎝R⎠
K K
Solving for V OUT in terms of VIN , we obtain the phasor-domain transfer function
1
TP ( jω ) = (17)
1+jωRC
The magnitude and phase of TP(jω) are given by
1
TP ( jω ) = ∠TP ( jω ) = − tan −1 ( ωRC ) (18)
1+ ( ωRC )
2

It thus follows from (10), (9) and Theorem 1 that the steady state output is given by

Page 9 of 14
⎛ ⎞
1 ⎟ sin ( ωt + θ − tan −1 ( ωRC ) )
VOUT ( t ) = VM ⎜ (19)
⎜ 1+ ωRC 2 ⎟
⎝ ( ) ⎠
As expected, it can be seen by comparing the results in (15) and (19) that the phasor-
domain and s-domain analysis results are identical.

time-domain analysis
The time-domain analysis involves writing and solving the differential equations that
describe the circuit. From Fig.8, we obtain the following three equations.

d VOUT ⎫
i(t) = C ⎪
dt
⎪⎪
VIN -VOUT = i ( t ) R ⎬ (20)
⎪
⎪
VIN = VMsin ( ωt+θ ) ⎪⎭
It remains to solve this set of three equations. One way to solve this simultaneous set of
differential equations is to use Laplace transform methods. Taking the Laplace
Transform of these three equations, we obtain the new set of three equations

⎫
I = sCVOUT ⎪
⎪⎪
V IN -V OUT = IR ⎬ (21)
( sin θ ) s + ωcosθ ⎪⎪
VIN = VM
s2 + ω2 ⎭⎪

Solving this set of equations, we obtain the VOUT as given by

⎡ ( sin θ ) s + ωcosθ ⎤ ⎛ 1 ⎞
V OUT = ⎢ VM ⎥⎜ ⎟ (22)
⎣ s +ω
2 2
⎦ ⎝ 1 + sRC ⎠
It remains to take the inverse Laplace transform of VOUT to obtain V OUT ( t ) where the
“~” operator is shown to indicate that the inverse Laplace transform will contain both a
forced and natural response. The steady state response, VOUT(t), is the forced response.
From a straightforward but tedious calculation, it follows that

⎡ ⎡ ⎛ ⎞ ⎤
sinθ ⎛ ωRC ⎞ - t RC ⎤ ⎢ ⎜ 1 ⎟ sin ( ωt + θ − tan −1 ( ωRC ) ) ⎥ (23)
V OUT ( t ) = ⎢ VM 2 ⎜
1- ⎟ e ⎥ + VM
⎢ ⎜ ⎟ ⎥
⎣⎢ (RC ) ⎝ tanθ ⎠ ⎦⎥ ⎣⎢ ⎝ 1+ ( ωRC )
2
⎠ ⎦⎥

Page 10 of 14
The first term in [] on in this expression is the natural response and vanishes at t → ∞ .
The forced response is the second term in []. It thus follows that the steady state response
is given by
⎛ ⎞
⎜ 1 ⎟ sin ( ωt + θ − tan −1 ( ωRC ) )
VOUT ( t ) = VM (24)
⎜ 1+ ωRC ⎟
( )
2
⎝ ⎠
As expected, this is the same solution as was obtained from the s-domain analysis and the
phasor-domain analysis.

Determining which analysis method is most useful

It should be apparent from this simple example that the total effort in obtaining
the steady state response using either of the frequency domain approaches is much less
than following the time domain approach and for more complicated circuits, the
difference in effort will be even much larger. The issue of whether to use the s-domain or
the phasor-domain analysis remains. The electronics community almost exclusively uses
the s-domain analysis. There are probably several reasons for this but one of the most
important, though not apparent from this simple example, is the amount of manipulative
detail required with the two approaches. Since magnitude and phase operators are not
defined in the s-domain analysis, one can not make intermediate “simplifications”
whereas all equations present in a phasor-domain analysis are complex quantities and
thus intermediate manipulations are possible and often implemented. Although this may
appear to simplify the analysis, in actuality, unnecessary manipulations are often made
and these intermediate manipulations thus often increase the effort required to solve the
system of equations. A second is simply the number of characters that must be written
when analyzing a circuit. It requires two characters to represent either an inductor or a
capacitor in the s-domain and it requires three characters to represent these elements in
the phasor domain. This actually results in considerably longer expressions when doing
hand calculations in the phasor domain. A third reason also favors the s-domain analysis.
By multiplying the transfer function T(s) by the Laplace transform of the excitation, the
Laplace transform of the output can be obtained for arbitrary inputs and thus the s-
domain approach can be used to obtain more than just the sinusoidal steady state
response. In the latter case, there are some additional considerations that will not be
discussed in detail here but they are associated with requiring the appropriate provisions
for the initial conditions on all energy storage elements, specifically the inductors and
capacitors. The s-domain analysis method, which is dominant in the electronics
community, has been extracted from Fig. 9 and appears in Fig. 13.

Page 11 of 14
Fig. 13 s-domain approach to analysis of linear networks

Magnitude and Phase Plots and Amplifier/Filter Nomenclature

If a linear circuit has a sinusoidal excitation, it follows form Theorem that the
magnitude and phase of the transfer function evaluated at the frequency of the excitation
give the magnitude of the gain and the phase shift of the output relative to the input.
Plots of the magnitude and phase are often used to visually show how gain and phase
shift vary with frequency. These magnitude and phase plots are often termed Bode plots.
From the magnitude and phase plots, the sinusoidal steady state response can be readily
obtained for any frequency of the input.
The transfer function for the circuit of Fig. 10 was given in equation (12). The
corresponding magnitude and phase for this transfer function are given by equation (14).
The magnitude and phase are plotted in Fig.14. From this magnitude plot, it can be

Page 12 of 14
T ( jω )

1
2

ω
1
∠T ( jω ) RC

0o

-45o

ω
-90o
1
RC
Fig. 14 Transfer function magnitude and phase plots for circuit of Fig. 10

seen that the magnitude of the gain is constant and nearly equal to 1 at low frequencies
and drops at the frequency increases. The region where the gain is large and nearly
constant is termed the pass band of the amplifier and the region where the gain is very
low is termed the stop band of the filter. The transition between the passband and the
stop band is not real abrupt in this circuit but it occurs around ω = (RC)-1. A circuit that
passes sinusoids at low frequencies but blocks sinusoids at high frequency is termed a
lowpass circuit. Some may term this a lowpass amplifier or a lowpass filter.
Several transfer function characteristics that vary intentionally with frequency
that describe specific magnitude shapes are the lowpass, bandpass, highpass, band-reject,
and notch functions. When circuits are designed to intentionally have these types of
magnitude characteristics, the circuits are typically called filters. The transfer function
magnitude for representative filters in each of these classes are shown in Fig. 15 . It
should be noted that each of these classes of filters is large and that the magnitude
responses shown in Fig.15 are simply representative characteristics. The phase response
for these classical filter shapes is also of interest but is not shown in the figure.

Page 13 of 14
T ( jω ) T ( jω ) T ( jω )

T ( jω ) T ( jω ) T ( jω )

Fig. 15 Representative filter characteristics for lowpass, bandpass, highpass,

bandreject and notch filters

Page 14 of 14