Analog Passive Filters Theory

Terminal Pairs or Ports:
Terminal Pairs or Ports – A pair of terminals at which an electrical signal may

enter or leave a network is called a port. The terminals or port is required for
connecting input excitation to the network. It is also required for connecting some
other networks such as load. The terminals are most useful for making
measurements. In general, the minimum number of terminals required is two.
A network having only one pair of terminals or one port is called one
port network. The Fig. 6.1 (a) shows a one port network.
A network consisting two pairs of terminals is called two port network as shown
in the Fig. 6.1 (b). The terminals are generally named as 1-1′ and 2-2′ In general, a
port designated 1-1′, is connected to the driving energy source while the other port
designated 2-2′ is connected to the load.
A port at which energy source is connected is called driving point of the network
or input port. A port at which load is connected is called output port.
Fig. 6.1 (c) represents a network with n-port called n-port network. In such
networks, generally one port is connected to energy source, one port is connected
to load and other ports may be connected to the different networks.
Two Port Network Parameters:
Consider a Two Port Network Parameters as shown in the Fig. 6.2. In all there are
four variables; two voltages and two currents. In general, any two port network has
one pair of voltage and current at each port.
To start with the discussion, we will first assume that other voltages and currents
inside the box are not available for the measurements. Secondly it is assumed that
the network consists only the linear elements along with dependent sources but
independent sources are not allowed.
Also it is assumed that the initial conditions on capacitor and inductor present in
the network are zero.
Assuming the variables at input and output ports as transformed quantities, the
voltage and current at input terminals are V 1 and I1 while at output terminals are
V2 and I2, as shown in the Fig. 6.2. The directions of both the currents I 1 and I2 are
assumed to be flowing into the network.
In order to describe the relationship between port voltages and currents one
requires the linear equations equal to the number of ports. So in two port network
analysis, we will require two linear equations interms of four above mentioned
variables.
We can obtain these equations by considering two variables as dependent variables

while other as independent variables. As the network consists only linear elements,
the linear relationship can be obtained by writing two variables in terms of other
two variables.
There are six possible ways of selecting two independent variables out of four
variables. Thus there are six different pairs of equations defining their own sets of
parameters such as impedance (Z), admittance (Y), hybrid (h), inverse hybrid (g),
transmission and inverse transmission parameters.
z Parameters of Two Port Network:
z Parameters of Two Port Network are also called impedance parameters. These
are obtained by expressing voltages at two ports in terms of currents at two ports.
Thus, currents I1 and I2 are independent variables; while V1 and V2 are dependent
variables. Thus, we have,
In equation form, above relations can be written as,
In matrix form, above equations can be written as,
The individual z-parameters are defined as follows.
From above equations it is clear that Z 11 and Z22 gives the relationship between
voltages and currents at port 1 – 1′ and port 2 – 2′ respectively. Such parameters
are the driving point impedances. While Z12 and Z21 gives the relationship between
voltage and currents at different ports. Such parameters are the transfer
impedances.
Moreover, these parameters are defined only when the current in one of the ports is
zero. This corresponds to the conditions that the one of the ports is open circuited.
Hence z-parameters are named as open circuit impedance parameters.
Hence we can write,
 Z11 = circuit driving point input impedance

 Z22 = Open circuit output impedance
 Z12 = Open circuit forward transfer impedance
 Z21 = Open circuit reverse transfer impedance.
The equivalent circuit of a two port network interms of open circuit impedance
parameters satisfying equations (1) and (2) is as shown in the Fig. 6.3.
FacebookTwitterPinterestTumblrLinkedInRedditBloggerShare
y Parameters of Two Port Network:

y Parameters of Two Port Network are also called admittance parameters. These
are obtained by expressing currents at two ports in terms of voltages at two ports.
Thus, voltages V1 and V2 are independent variables, while I1 and I2 are dependent
variables. Thus, we have

The individual y-parameters are defined as follows.
This is called driving point input admittance.
This is called forward transfer admittance.
This is called reverse transfer admittance
This is called output admittance.
These parameters are defined individually only when the voltage in any one of the
ports is zero. This corresponds to the condition that one of the ports is short
circuited. Hence y-parameters are also called short circuit admittance
parameters.
The equivalent circuit of a y Parameters of Two Port Network interms of short

circuit admittance parameters satisfying equations (1) and (2) is as shown in the
Fig. 6.4.
h Parameters of Two Port Network:
h Parameters of Two Port Network are also called hybrid parameters. These
parameters are very useful in constructing models for transistors. The transistor
parameters cannot be calculated using by either short circuit admittance parameter
or open circuit impedance parameter measurement.
These parameters are obtained by expressing voltage at input port and the current
at output port in terms of the current at the input port and the voltage at the output
port. Thus, the current I1 and voltage V2 are independent variables ; while current
I2 and voltage V1 are the dependent variables.
Thus, we have
In matrix form the above equations can be written as,

The individual h-parameters are defined as follows,
This is called short circuit input impedance.
This is called forward short circuit current gain.
This is called reverse open circuit voltage gain.
This is called open circuit output admittance.
All above parameters are having different units such as ohm for short circuit
impedance and mho for open circuit output admittance, the name of the parameter
is hybrid parameter.
The equivalent circuit of a two port network interms of hybrid parameters

satisfying equations (1) and (2) is as shown in the Fig. 6.5.
ABCD Parameters of Two Port Network or Transmission
Parameters or Chain Parameters:
ABCD Parameters of Two Port Network are known as transmission parameters.
These are generally used in the analysis of power transmission in which the input
port is referred as the sending end while the output port is referred as receiving
end. These are obtained by expressing voltage V1 and current I1 at input port
interms of voltage V2 and current I2 at output port. Thus, voltage V2 and current
I2 are independent variables while voltage V1 and current I1 are dependent variables.
Thus, we have
Generally we have considered the currents in both the ports are entering the port
and both are positive. The negative sign with I 2 indicates that, for the ABCD
parameters the current I2 is leaving the port 2.
In matrix form above equations can be written

The individual transmission parameters are defined as follows,
This is called open circuit reverse voltage gain.
This is called open circuit transfer admittance
This is called short circuit transfer impedance
This is called short circuit reverse current gain.
ABCD or Transmission parameters are also called chain parameters.
Condition for Symmetry in Two Port Network:

If the impedance measured at one port is equal to the impedance measured at the
other port with remaining port open circuited, the network is said to be Condition
for Symmetry.
The impedances at two ports can be measured as shown in Fig. 6.6 (a) and (b).
Condition of Symmetry for z Parameters:
The basic equations of z-parameters are as follows.
Consider Fig. 6.6 (a).
Hence equation (i) will become,
Consider Fig. 6.6 (b).
Equation (ii) can be written as,
For symmetrical networks, both the port impedances must be equal i.e.
Condition of Symmetry for y Parameters:
The basic equations of y-parameters are as follows
Consider Fig. 6.6 (a).
Equation (i) can be written as,
Substituting value of V2 in equation (iii), we have

Substituting value of V1 in equation (viii) from equation (vii), we have
Conditions of Symmetry for h Parameters:
The basic equations of h-parameters are as follows,

Consider Fig. 6.6 (a),
Equation (ii) can be written as
Substituting value of V2 in equation (iii), we have
Consider Fig. 6.6 (b),
Equation (i) can be written as

Condition of Symmetry for Transmission Parameters:
The basic equations for the transmission parameters are as following

For the symmetrical network, both the port impedances must be equal.
Condition for Reciprocal Two Port Network:

If the ratio of voltage at one port to the current at other port is same to the ratio if
the positions of voltage and current are interchanged, then the network is said to be
Condition for Reciprocal Two Port Network. The transfer impedances can be
measured as shown in Fig. 6.7 (a) and (b).
Condition of Reciprocity for z-Parameters:
The basic equations for z-parameters are as follows

Consider Fig. 6.7 (a)
Hence equation (i) can be written as,
Substituting value of I1 in equation (iii), we get
Consider the Fig. 6.7 (b).

Substituting value of I2 is equation (vii), we have,
For condition of reciprocity,
Condition of Reciprocity for y-Parameters:
The basic equations of y-parameters are as follows,

Consider Fig. 6.7 (b)
The network is said to be reciprocal if
Condition of Reciprocity for h-Parameters:
The basic equations of h-parameters are as follows,
Substituting value of I1 in equation (iii)

Consider Fig. 6.7 (b)
Equation (i) can be written as
The network is said to be reciprocal if,
Condition of Reciprocity for ABCD-Parameters:
The basic equations for ABCD parameters are as follows,

From equation (iii) we have,
Substituting this value in equation (iv) we have
The network is said to be reciprocal if,

Interconnection of Two Port Network:
Various types of Interconnection of Two Port Network such as series connection of
two ports, parallel connection of two ports, cascade connection of two ports etc.
Series Connection of Two Ports:
Consider two networks N’ and N” are connected in series as shown in Fig. 6.8 (a).
When two ports are connected in series, we can add their z parameters to get
overall z-parameter of the overall series connection.
Let the z-parameters of network N’ be z’ 11, z’12, z’21, z’22. Let the z parameters of
network N” be z”11, z”12, z”21, z”22. Let the overall z-parameters of series connection
be z11, z12, z21, z22.
For series connection we have,

For network N’ , z-parameter equations are,
For network N”, z-parameter equations are,
From equations (1) , (2) and (3) , (4) we can write,
In the Matrix form, above equations can be written as,
Thus, overall z-parameters are,
Hence, the z-parameters of the series connection are the sum of z-parameters of the
individual network connected in series.
Parallel Connection of Two Ports:

Consider two networks N’ and N” are connected in parallel as shown in Fig. 6.8
(b). When two ports are connected in parallel, we can add their y-parameters to get
overall y-parameters of the parallel connection.
Let the y-parameters of the network N’ by y’ 11, y’12, y’21, y’22. Let the y-parameters
of the network N” be y”11, y”12, y”21, y”22. Let the overall y-parameters of parallel
connection be y11, y12, y21, y22.
For parallel connection we have,
For network N’, the y-parameter equations are,
For network N” the y-parameter equations are,

From equations (1), (2) and (3), (4), we can write,
Thus, overall y-parameters are,
Hence, the y-parameters of the parallel connection are the sum of y-parameters of
the individual networks connected in parallel.
Cascade Connection of Two Ports:
The cascade connection is also called Tandem connection. Consider two networks
N’ and N” are connected in cascade as shown in Fig. 6.8 (c). When two ports are
connected in cascade, we can multiply their individual transmission parameters to
get overall transmission parameters of the cascade connection.
Let the transmission parameters of network N’ be A’ , B’ ,C’ , D’. Let the
transmission parameters of network N” be A”, B”, C”, D”. Let the overall
transmission parameters of cascade connection be A, B, C, D.
For cascade connection we have,
For the network N’, transmission parameter equations are,
For the network N”, transmission parameter equations are,
The overall transmission parameters of the cascade connection can be written as,
Hence, the transmission parameters for the cascaded two port network is simply
the matrix product of the transmission parameter matrix of each individual two
port network in cascade.
Series and Parallel Permissibility of Connection:
Before interconnecting two ports it is very important to check for the overall port if
the current entering one terminal is same as that leaving other terminal of same
port. This test is called as Validy test or Brune’s Test.
If the 2-port networks to be interconnected do not satisfy Brune’s test, an ideal

transformer of 1:1 turns ratio is connected at one of the ports. This transformer
provides necessary isolation and it also ensures that the characteristics of any
network does not alter due to the interconnection with other.
The permissibility of connection of two port network through the transformer is as

shown in the Fig. 6.9 (a) to (d).
T Network:
Consider two port T Network with assumed voltages V1 and V2 and currents I1 and
I2 as shown in Fig. 6.10.
Let the z-parameters of the above network are represented as follows
We know that, the z-parameters for any network can be calculated by open
circuiting input and output terminals ie. by making I1 = 0 and I2 = 0
With the port 2-2′ open circuited :
We have, I2 = 0
Total impedance looking into network from port 1-1′ is equal to ( z1+ z3).
Hence the current I1 in port 1 can be calculated as
The voltage at port 2 ie. V2 can be given as.

With the port 1-1′ open circuited :
We have, I1 = 0
Total impedance looking back into network from port 2-2′ is equal to (z 2 + z3)
Hence the current I2 in port 2 can be calculated as,
The voltage at port 1 ie V1 can be given
Consider
Substituting in
Substituting z3 = z12 in above expression,
Hence, the circuit elements of the equivalent ‘T ‘ network are as follows,
π or pi Network in Network Analysis:

Consider following two port ’π’ or pi Network type network with assumed
voltages V1 and V2 and current I1 end I2 as shown in Fig. 6.11.
Let the y-parameters of the above network are represented as follows,

We know that, y-parameter for any network can be calculated by short circuiting
input and output ports ie by making V1 = 0 and V2 = 0.
With the port 2-2′ short circuited :
We have V2 = 0
Hence, y2 will be directly shorted.
Total admittance looking into network from port 1-1′ is equal to (y1 + y3).
Hence, current I1 in port 1-1′ can be calculated as follows,
The current I2 in port 2-2′ can be calculated as follows.
Note that current I2 will be negative
With the port 1-1′ short circuited :
We have V1 = 0. Hence, y1 will be directly shorted. Total admittance looking back

into network from port 2-2′ is equal to (y2 + y3).
Hence, current I2 in port 2-2′ can be calculated as follows

The current I1 in port 1-1′ can be calculated as follows.
Note that current I1 will be negative.
Consider the expression
substituting y3 = -y12 in first two expressions, we have,
Hence, the circuit elements of the equivalent ’π’ of pi network are as follows,
Lattice Network:
The lattice network is one of the common four terminal networks. These networks
are generally used in filter sections and also used as symmetrical attenuators.
Consider the lattice network shown in the Fig. 6.12 (a). The arms with impedances
Za and Zd are called series arms while the arms with impedances Z b and Zc are
called diagonal arms. The lattice network can be redrawn in the form of the bridge
network as shown in the Fig. 6.12 (b).
The lattice network is said to be symmetrical if Za = Zd and Zb = Zc
Consider I2 = 0 i.e. port 2 open circuited. The network can be redrawn as shown in
the Fig. 6.13 (a).
Voltage V1 can be expressed interms of current I1 as follows,

The voltage at port 2, V2, can be expressed in terms of current I2 as follows.
Putting value of V1 from equation (i), we get
Now consider I1 = 0 i.e. port 1 open circuited as shown in the Fig. 6.13 (b).
Similarly we can express voltage V1 in terms of voltage V2 as

Putting value of V2 from equation (v) in above equation, we get,
But for symmetrical networks,
Thus equations (ii), (iv), (vi) and (viii) can be written as follows.
Hence, we can express series arm and diagonal arm impedances in terms of z-
parameters of the network as follows,
Thus, we can express any symmetrical network interms of its lattice equivalent
network is the z-parameters of the symmetrical network are known. The lattice
network in terms of z-parameters is as shown in the Fig. 6.14.
Bridged T Network Analysis:
Consider Bridged T Network Analysis as shown in the Fig. 6.15 (a).
We can consider bridged T network as a parallel connection of the two networks as

shown in the Fig. 6.15 (b). The y-parameters of the overall network can be
obtained by adding y-parameters of networks connected in parallel.
The parameters of the network Na are given by
Consider the network Nb. As the network is symmetrical T network, we can write
Z-parameters of the network directly as follows,
From equations (3) and (4),
Hence y-parameters of the network Nb are given by
As both the networks Na and Nb are symmetrical so the overall bridged T network
is also symmetrical network. Hence overall y-parameters for the bridged T network
are obtained by adding y-parameters of the networks Na and Nb.
Ladder Network:
The Fig. 6.16 shows a ladder network. The Z1,Z3, ….Zn-1 are the impedances in
series arms while Y2,Y4 …..Yn are the admittances in shunt arms.
The Zin is the input impedance function of the ladder network.
Analysis of Ladder Network [Unit Output Method]:
The continued fraction method of finding driving point impedance of the network
is already discussed. Let us study another simple method of obtaining network
functions of a ladder network.
This method is based only on the relationships that exist between the branch
currents and the node voltages of the ladder. In this method we name the various
node voltages and branch currents and start at the output end, analyzing the
network back to the input end.
Consider a four branch ladder network as shown in the Fig. 6.17, to understand the
method.
The various node voltages and branch currents are indicated. Analyse the output
end.
It can be observed that each equation depends on the earlier two equations only.
Each new equation takes into account one new immitance (either impedance or
admittance). The each new equation is obtained by multiplying the equation just
preceding it by the immitance that is next down the line and then adding to this
product, an equation which is twice preceding it.
For example consider equation (2). The next equation is obtained by multiplying
this equation by Y2 which is next admittance down the line and adding to it an
equation (1) which is twice preceding the new equation.
Thus general procedure for the analysis is,
 Write alternately the node voltage and branch current equations.

 The next equation is obtained by multiplying the existing equation by the
next immitance, analyzing from port 2 to 1 i.e. from right to left and adding
to this the result of the previous equation.
Using these equations any network function such as driving point imittance,
voltage ratio transfer function and current ratio transfer function can be obtained.
The method discussed above is called unit output method as every equation contain
V2 factor which can be normalized to V2 = 1.
Transfer Function of Terminated Two Port Network:

A ratio of voltage or current at one port to the voltage or current at the other port is
called a Transfer Function of Terminated Two Port Network. In the last sections,
we have studied how to obtain two port network parameters. In this section, we
will study how to determine driving point and transfer functions of a two port by
use of two-port parameters.
First consider that the two port network is not terminated in source and load
impedances. Let us express the transfer functions in terms of z and y parameters.
Consider z parameter equations for a two port network.
When port-2 is open circuited, I2 = 0. Then we can write equations (1) and (2)
respectively as,
From equation (4), we can express current I1 interms of voltage V2 as,
Putting value of I1 in equation (3), we can write,
Hence, open circuit voltage ratio of two port network is given by,
Now consider y-parameter equations for a two port network.
When port 2 is open circuited, I2 = 0. Then equation (8) can be written as,
Hence, open circuit voltage ratio of two port network interms of y-parameters is
given by,
We can also determine short circuit current ratio of a two port network in terms of
z and y parameters with condition that V2 = 0.
When port 2 is short circuited, V2 = 0. Then equation (2) can be written as
Hence short circuit current ratio of a two port network interms of z parameters is
given by,
Similarly equations (7) and (8) respectively can be written as,
From equation (14), we can express voltage V1 interms of current I2 as,
Putting value of V1 in equation (13), we can write,
Hence, short circuit current ratio of a two port network interms of y-parameters is
given by,
Now consider a two port network expressed in terms of z parameters, terminated in
impedance ZL as shown in the Fig. 6.18.
From equations of z parameters we can write,
But from the network it is clear that, V2 = – I2 ZL
Putting value of I2 in above equation,
The transfer impedance Z21 can be obtained from above equation as,
Putting value of I1 in equation for V1 as follows,
The voltage ratio of transfer function of terminated two port network is given by,
Putting value of V2 in equation for V2, for we get,
The current ratio transfer function for terminated network is given by,
The driving point impedance of a terminated two port network can be obtained
from equations (17) and (18) as follows,
Consider a two port network terminated at both the ports of a network as shown in
the Fig. 6.19.
Applying KVL at input side, we can write,
Applying KVL at output side, we can write,
But V2 = -I2 R2
Putting values in equation (22), we get,

Putting value of I1 in equation (21) and writing equation in terms of voltages V g and
V2, we get,
Hence, the voltage ratio transfer function of terminated two port network
terminated at both the ports is given by,
Elements of Network Synthesis:

Elements of Network Synthesis – For any network, three things are associated with
it. These are network elements, input i.e. excitation to the network and output i.e.
response from the network.
In the network analysis, the network elements are known and excitation is also
known. Using number of methods, the networks are studied and the response is
obtained. Such a response is unique for a given network and known excitation. So
obtaining a response for a known network and known excitation is called network
analysis.
In the network synthesis, the procedure is exactly opposite to the analysis. The
excitation is known and the response requirements are known. It is necessary to
find the network satisfying the requirements. Thus obtaining a network for a
known excitation and known response requirements is called network synthesis.
The basic difference between network analysis and synthesis is shown in the Fig.
7.1 (a) and (b).
Another important difference between analysis and synthesis is that the analysis
gives the solution i.e. response which is always unique. But synthesis may give
different solutions i.e. networks satisfying the required specifications. The
synthesis does not give us unique solution. Sometimes it may not give the solution.
For example, the specification of the network is that its impedance Z(s) must be 10
Ω. Synthesizing of this specification may give the network with only one 10 Ω
resistance or two 5 Ω resistances in series or two 20 Ω resistances in parallel and so
on. So synthesis solution is not unique.
Similarly specification like Z(s) = —10 Ω can not be satisfied using passive
Elements of Network Synthesis. Thus synthesis may not give any solution also.
We have seen that driving point impedance and admittance functions for all the
networks, consisting of passive elements are rational and in the form of ratio of
two polynomials in ‘s’. But it is not possible that every ratio of two polynomials in
‘s’ can be synthesized with Elements of Network Synthesis.
Realisability of One Port Network:

Realisability of One Port Network – Consider the network function H(s) which is
the ratio of Laplace transform of the output R(s) to the Laplace transform of the
excitation E(s).
In the synthesis, we are going to obtain a network from the given network function
which may be admittance function or impedance function. But before synthesizing
H(s), it is necessary to check whether H(s) is physically realizable or not, using
passive elements. For this, there are two important conditions related to a network
which are causality and stability.
Now there can not be output without input i.e. there can not be current without any
voltage. So there exists a cause and effect relationship between input and output.
This is nothing but causality. To have the given One Port Network to be causal, the
impulse response of the network must be zero for t<0. The impulse response means
the response of the network to the unit impulse input.
The Laplace transform of unit impulse function is always unity.
Thus unit impulse response of the network is inverse Laplace transform of the
network function. For a network to be causal the condition is,
Impulse response = h(t) = 0, for t<0
This is the required condition for the network to be causal. Thus the impulse
response h(t) = e-2t u(t) is causal as u(t) = 0 for t< 0 thus h(t) is also zero for t<0.
While the response like h(t) = e-2|t| is not causal.
Practically in certain cases, the impulse response can be made causal by delaying
it. Consider the response h(t) shown in the Fig. 7.2 (a). It is not causal. But
delaying it by T, it becomes zero for all t< 0 and hence becomes causal. This is
shown in the Fig. 9.2 (b).
If the response is delayed by T then it becomes causal and realizable. For a system
function to be causal in frequency domain, it must satisfy the condition that h(t)
must possess a Fourier transform H(jω) and the square magnitude function |H(jω)|
2
must be integrable i.e.,
If the system function is to be causal then it must satisfy the condition in frequency
domain called Paley-Wiener criterion. The criterion states that it is necessary and
sufficient condition for an amplitude function |H(jω)| to be causal that
This means that the amplitude |H(jω)|of a realizable i.e. causal network must not be
zero over a finite band of frequencies.
Similar to causality, the network must be checked for stability. The One Port
Network is said to be stable if,
 It produces a bounded output for a bounded input.

 There is no output if input is reduced to zero.
Thus if e(t) is the excitation and |e(t)| < k1 for 0 ≤ t ≤ ∞ then output r(t) must satisfy
|r(t)| < k2 for 0 ≤ t ≤ ∞.
The k2 and k2 are real, positive, finite constants.
The stable network must satisfy the condition that its impulse response must be
absolutely integrable.
This gets satisfied when the impulse response h(t) approaches zero as t approaches
infinity,
In general for a stable network, the impulse response must be bounded for all t i.e.
where K is positive, real, finite number.
Now the functions like sin(ωt) and cos(ωt) are bounded but are not absolutely
integrable. Such systems have poles on the imaginary axis. These systems are not
stable from integration definition but stable from bounded definition. Hence such
systems are called marginally stable systems.
For the stability of network function H(s), it must satisfy following conditions,
1. H(s) should not have a pole in right half of s-plane.

2. The poles of H(s) on the imaginary axis must be nonrepeated. There should
not be multiple poles of H(s) on imaginary axis.
3. The degree of the numerator of H(s) can not exceed the degree of the
denominator by more than unity.
Thus first step in the synthesis is to obtain the information whether the given
function is realizable or not. For a given function to be realizable it is essential that
the One Port Network function should satisfy all the properties of Hurwitz
polynomials and those of positive real functions. Hence before the study of actual
synthesis, it is necessary to study the properties of Hurwitz polynomials and the
positive real functions.
Hurwitz Polynomial:
For a polynomial P(s) to be a Hurwitz polynomial, it has to satisfy following basic
properties,
 The polynomial P(s) is real when ‘s’ is real.

 The roots of the polynomial P(s) have real parts which are either zero or
negative.
In addition to these basic properties, Hurwitz polynomial has to satisfy few more
properties.
Properties of Hurwitz Polynomials:
Consider a polynomial P(s) of order ‘n’ as,
where an , an_1 , … a0 are the coefficients.
For the polynomial P(s) to be Hurwitz,
 All the coefficients ‘ai‘ are real and nonnegative. Thus all the coefficients
an to a0 must be positive.
 All the terms starting from highest power of s to lowest power of s must be
No power of s should be missing in between.
A corollary to this is that between highest order term in s and lowest order term in
s, none of the coefficients may be zero, unless the polynomial is either even or odd.
The polynomial is said to be even if it consists of only even powers of s. In such

case, all odd powers of s are missing but in such polynomial also no even power
should be missing in between. For example,
Similarly the polynomial is said to be odd if it consists of only odd powers of s. In

such case, all even powers of s are missing but in such polynomials also no odd
power should be missing in between. For example,
 If the polynomial is even or odd then all its roots must lie on imaginary jω
axis.
 Given Hurwitz polynomial can be separated into even and odd parts. The
odd part is denoted as n(s) while even by m(s) then,
Both these n(s) and m(s) parts must have roots on jω axis only.
If the ratio of odd to even parts of P(s) or even to odd parts of P(s) is expressed in
the continued fraction expansion then all the quotient terms must be positive.
Then the continued fraction expansion is,
All the quotients q1, q2 …. qn must be positive for the P(s) to be Hurwitz.
The process of obtaining the continued fraction expansion includes simple division
and then inversion. At each step we get a quotient qi and a remainder. We then
invert a remainder term and then divide to obtain a new quotient. The last
remainder must be zero. The continued fraction expansion must be finite in length.
To understand this, let us express P(s) in continued fraction expansion.
Ratio of n(s) to m(s) or m(s) to n (s) is dependent on their powers of s. The ratio of
higher power to lesser power is generally taken. Start with the division.
In the first step 3s is the remainder. The inversion means 2 s 2 +2 is then divided by
the remainder 3s. And then division is carried out. In this step, 2 is the remainder.
Again 3s is divided by 2 which is an inversion and then division is carried out.
Note that highest power polynomial out of n(s) and m(s) must be considered as a
numerator to start the division. Another important observation is that the number of
quotients obtained in the continued fraction expansion is equal to highest power of
s present in the polynomial P(s).
If the continued fraction expansion gives all the positive quotients then the
polynomial P(s) can be expressed as the product of the two factors which are
individually Hurwitz. Thus,
then P(s) is Hurwitz if W(s) and P1 (s) are individually Hurwitz.
Continued fraction expansion may terminate abruptly if there exists a common

factor between n(s) and m(s) which gets cancelled when the ratio is taken. Then
given P(s) can be expressed as the product of the two polynomials as W(s) P 1 (s)
where W(s) is the common factor present in n(s) and m(s)
Consider
The continued fraction terminates abruptly as the number of quotients are not equal
to highest power of s in P(s) i.e. 3. Then P(s) can be written as,
where s2 +5 is the common factor which is W(s) and (s + 4) is P 1 (s). Both W(s) and
P1(s) are Hurwitz and hence P(s) is also Hurwitz.
If P(s) is even or odd then it is not possible to express it in continued fraction form
as either n(s) or m(s) is absent. In such case, continued fraction expansion of P(s)
with its derivative P'(s) is obtained which has to satisfy all the above properties to
have P(s) to be Hurwitz.
Consider
The continued fraction expansion of P(s) / P'(s) is obtained.
It can be verified that all the quotients are not positive when P(s) / P'(s) is
expressed in continued fraction form, under this case. And hence P(s) is not
Hurwitz. The continued fraction expansion is also called Euclid’s algorithm for
testing a given polynomial to be Hurwitz or not.
Routh Hurwitz Array Method:
Another method of testing a given polynomial for Hurwitz is called Routh-

Hurwitz array method. In this method, an array is constructed using the
coefficients of given polynomial in a specific way. By the inspection of such an
array formed, the polynomial can be decided to be Hurwitz or not.
The method of finding Routh Hurwitz array is,
In this method, first row consists of all the coefficients of alternate powers of s
starting from n while the next row consists of all the coefficients of alternate
powers of s starting from n-1. Thus all the coefficients are accommodated in first
two rows. Then each next row is generated from the coefficients of the two
previous rows to it. So a row corresponding to sn-2 is generated from first two rows
as,
The row corresponding to sn-3 is generated from the two previous rows.
The procedure is continued till the row corresponding to s0 is obtained. The last
row always consists of a single coefficient which is equal to the constant term of
the given polynomial.
Once the array is completed, its first column is inspected. For the given polynomial
P(s) to be Hurwitz,
1. All the elements in the first column should be non zero.

2. There should not be any sign change in the first column.
Positive Real Function:
Let us study another class of functions called positive real functions. The
significance of positive real functions is that if the driving point immitance (i.e.
admittance or impedance) is a positive real function then only it is physically
realizable using passive R, L and C components. Hence immitance function must
be checked for positive realness before synthesizing.
For a function to be positive real function it has to satisfy the following basic
properties,
 The given function F(s) is real for real s.

 The real part of F(s) is greater than or equal to zero, when the real part
of s is greater than or equal to zero.
The function F(s) is rational function.


The positive real function is also called Brune function. In addition to the basic
properties, the positive real function has some more properties.
Properties of Positive Real Function:
Let
The function F(s) is positive real function having following properties,
1. The coefficients of the numerator and denominator polynomials N(s) and D(s) in
F(s) are real and positive.
Hence,
F(s) is real when s is real.


 The complex poles and zeros of F(s) occur in complex conjugate pairs.
 The scale factor, N = N = a0/b0 is real and positive.
2. The poles and zeros of F(s) are having negative or zero real parts.
3. The poles of F(s) on the imaginary (jω) axis must be simple. Their residues must
be real and positive. The same statement is true for the poles of 1/F(s).
4. The degrees of the numerator and denominator polynomials in F(s) differ at the
most by 1. So degrees m and n must differ at the most by 1. Thus the number of
finite poles and finite zeros of F(s) differ at the most by 1.
5. The terms of lowest degree in the numerator and denominator polynomials of
F(s) differ in degree at most by 1. So F(s) has neither multiple poles nor zeros at
the origin.
6. If F(s) is positive real function then 1/F(s) is also positive real function. Thus for
a network if driving point impedance function Z(s) is positive real then the driving
point admittance function Y(s) = 1/Z(s) is also positive real.
7. The sum of positive real functions is also positive real. If two impedances are in
series, the sum of the impedances is positive real. Similarly if two admittances are
in parallel, their addition gives positive real admittance. Note that the difference in
two positive real functions is not necessarily a positive real function.
These properties are nothing but necessary conditions for the given function to be
positive real but are not sufficient conditions.
Procedure for Testing a Function for Positive Realness:
Let F(s) be the function to be tested for the positive realness, which is a ratio of
two polynomials N(s) and D(s). Remove all the common factors in the numerator
and the denominator before testing for the positive realness.
The testing procedure can be divided as,
1. Testing for necessary conditions

2. Testing for necessary and sufficient conditions
Inspection Test for Necessary Conditions:
By inspecting the given function, the following requirements are tested,
1. All the coefficients of different polynomials must be real and positive.

2. The degrees of numerator and denominator polynomials differ at most
by 1.
3. Lowest degree in numerator and denominator differ at most by 1.
4. The imaginary axis poles and zeros of F(s) must be simple in nature. No
multiple pole or zero should lie on the imaginary axis.
5. There should be no missing terms in numerator and denominator
unless all even or all odd terms are missing.
6. The poles and zeros of F(s) must be located in the left half of s-plane.
7. There should not be multiple poles or zeros either at origin (s = 0)
and/or at infinity (s = ∞).
8. The simple poles on jω axis should have real and positive residues.
Test for Necessary and Sufficient Conditions:
The tests for necessary and sufficient conditions are,

1. The function F(s) must be real when s is real.
2. If F(s) = N(s)/D(s)
then N(s) + D(s) must be Hurwitz.
For N(s) + D(s) to be Hurwitz, all the quotients obtained by expressing it in

continued fraction expansion must be positive. Also the continued fraction
expansion should not terminate abruptly.
3. Re [F(jω)] ≥ 0 for all ω.
The real part of F (jω) must be greater than equal to zero for all ω.
To test this, separate the numerator and denominator polynomials into even and
odd parts.
Multiply numerator and denominator by m2 – n2
It is known that,
 Product of two even functions is an even function.

 Product of two odd functions is an even function.
 Product of an odd function and an even function is an odd function.
So as m1 and m2 are even, m1 m2 is even.
And n1 and n2 are odd, n1 n2 is even.
Also m1 n2 and m2 n1 is odd as the product of even and odd function.
Finally m22 and n22 are both even.

Hence in the above equation of F(s), first part is even while the second part is odd.
Hence substituting s =jω in the even part gives the real part of F(jω) while
substituting s =jω in odd part gives imaginary part of F(s).
In most of the cases, the condition can be verified by factorising A (ω 2). If

factorization is not sufficient for the required conclusion then A (ω 2) is plotted over
sufficiently large range of ω and it is ensured that it is not negative.
Thus the testing procedure for positive realness of a function can be

summarized as,
1. Check all the necessary conditions by inspection.
2. For all the poles and zeros of F(s) to be in the left half of s-plane or on the
imaginary axis, N(s) and D(s) polynomials must be Hurwitz.
3. If F(s) has poles on imaginary axis, the residues at the poles must be real and
positive.
Find the partial fractions of F(s) where s = ± jω 0 is the pair of poles on imaginary
axis.
So coefficients k and k· which are complex conjugates of each other must be real
and positive.
4. Finally test that
This can be done by factorising A (ω2). This also can be tested by plotting A (ω2)
graph against ω2 or using Sturm’s theorem.
Driving Point Immittance Function:

The immitance function must be positive real function so that its synthesis can be
done to obtain an electrical network, using passive elements. We have discussed
the tests to confirm the positive realness of a given function. Let us study now how
to synthesize the given Driving Point Immittance Function (impedance or
admittance) which is positive real.
There are three types of passive elements which are an inductor (L), a capaciter (C)
and a resistor (R). A given function can be synthesized using any two types of
passive elements. Thus synthesized network can be LC network, RC network or
RL network.
A network using any two types of passive elements can be synthesized generally in
two forms called,
1. Foster form and

2. Cauer form.
These forms are used for the network realization because the network is realized
using minimum number of passive elements using these basic forms. Hence these
forms are called canonical or simple forms of realization. In each of these forms
there are two subforms. The foster form is subdivided as Foster I form and Foster
II forms while the Cauer form is subdivided as Cauer I and Cauer II forms.
The following table is useful while realizing the network by any of the four forms
mentioned above
Foster I Form:
The Foster I form uses the partial fractions of the driving point impedance function
Z(s). The partial fraction expansion gives the equation of Z(s) as summation of the
various impedance functions.
Thus a network can be realized by connecting the impedances Z1,Z2, ……..Zn in

series.
Each impedance individually can be a series combination or parallel combination

of the elements, which is to be identified. The identification of various elements
existing in the different impedance function can be done by referring the Table 7.1.
While finding the partial fractions, make sure that the degree of numerator is less
than the degree of denominator. In this form, number of times, we get the functions
after partial fraction expansion as,
This can be synthesized by expressing it as
We know
The addition of two admittances indicate that there are two branches in parallel.
Referring to the Table 7.1, Y1(s) = s/A represents a capacitor of 1/A F while Y2(s)
= B/A represents conductance of B/A mho i.e. a resistance of A/B ohms. Hence
Z'(s) can be realized as parallel combination of R and C as shown in the Fig. 7.3.
Note that this example illustrates how to synthesize a particular function obtained
in the partial fraction form. It does not mean that every time Foster I form gives
RC networks.
Foster II Form:
The Foster II form uses the partial fractions of the driving point admittance
function Y(s) = 1/Z(s). The partial fraction expansion gives the equation of Y(s) as
summation of the various admittance functions.
Thus a network can be realized by connecting the admittances Y 1, Y2, …… Yn in

parallel.
Each admittance individually can be a series combination or parallel combination

of the various elements. These combinations are to identified to realize the
network. This can be easily done by referring the Table 7.1, as illustrated above.
Cauer I Form:
Basically Cauer form uses continued fraction expansion of the Driving Point
Immittance Function. The Cauer I uses the continued fraction expansion of the
driving point impedance function Z(s). In this form, the numerator and the
denominator are arranged in the descending powers of s, starting from highest to
lowest power of s.
This form gives the ladder structure of the network with alternate series and
shunt arms. Hence this form of realization is also called ladder realization.
The continued fraction expansion gives the form as,
The continued fraction expansion can be obtained by the division and inversion
procedure, used earlier for testing the Hurwitz polynomials.
In the continued fraction form of Z(s), the quotient of first division gives an
impedance which is a series arm and then the quotients give alternately Y(s) and
Z(s) terms representing shunt and series arms respectively.
The elements of each Z(s) and Y(s) obtained can be identified using the Table 7.1.
Remember that if the degree of numerator is same or less than the denominator,
there is possibility of getting negative coefficients in the continued fraction
expansion. In such case restart with the inversion and first quotient in such case
represents admittance Y(s) which is a shunt arm. The series arm is absent in such
case.
The Cauer I form is low pass structure of the network.
If the driving point admittance function Y(s) is expressed in continued fraction

expansion with division at start, first quotient represents admittance Y(s) in shunt
arm and then, alternately impedances and admittances in series and shunt arms
respectively. Starting with inversion in such case gives impedance Z(s) in series
arm as the first quotient, and then alternately admittances and impedances in shunt
and series arms.
Cauer-II Form:
The Cauer II form also uses continued fraction expansion of the driving point
immittance function either Z(s) or Y(s). In this form, the numerator and the
denominator are arranged in the ascending powers of s, starting from lowest to
highest power of s.
The continued fraction expansion gives the similar form as,
In this form also, there is a possibility of getting negative coefficients in the

expansion procedure. In such a case, restart with the inversion, which gives Y(s) as
the first quotient which indicates shunt arm, without a series arm in the circuit.
Similarly driving point admittance function Y(s) also can be synthesized using
Cauer II form.
The Cauer II form is basically high pass structure.
Remember that if Y(s) is to be synthesized and expansion starts with the division,
first quotient gives Y(s) in shunt arm and if starts with the inversion, first quotient
gives Z(s) in series arm and vice versa for Z(s).
LC Immittance Function:
The LC Immittance Function can be LC impedance functions denoted as Z LC(s) or
LC admittance functions denoted as YLC(s).
A LC network does not contain power dissipative components i.e. resistances and
only consists of reactive elements L and C components. Hence such network is
also called a reactance network or lossless network.
Consider a driving point impedance function of LC one port network represented

as the ratio of two polynomials in s.
Where m1, n1 are even and odd parts of N(s) while m2, n2 are the even and odd parts
of D(s) respectively.
For a purely reactive element jωL or 1/jωC, the real part is zero.
We have seen that the real part of F(jω) is its even part given by,
But real part must be zero for LC network.
Thus to satisfy this equation we must have,
So it can be concluded that the driving point impedance function of LC network is

the ratio of odd to even polynomial or even to odd polynomial. This is a very
important property of LC network.
For example,
So poles at s = 0, ± j 2
Zeros at s = ± j, ± j 3
The pole zero plot is shown in the Fig. 7.4.
From this impedance function of LC network, let us list the various other
properties of LC networks.
Properties of LC Immittance Function:

Referring to the pole zero plot shown in the Fig. 7.4, the properties of Z LC(s) and
YLC(s) functions can be stated as,
1. The LC Immittance Function is always a ratio of odd to even or even to odd

polynomials.
2. The poles and zeros are simple. There are no multiple poles or zeros either
at origin or infinity or at any point.
3. The poles and zeros are located on the jω axis only.
4. The poles and zeros interlace (alternate) each other on the jω axis. There are
no consecutive poles or zeros on the jω axis.
5. The imaginary poles and zeros occur in the form of complex conjugate
pairs.
6. The highest powers of numerator and denominator must differ by unity.
7. The lowest powers of numerator and denominator must differ by unity.
8. There must be either a pole or zero at the origin and infinity. As the function
is the ratio of even to odd polynomials, if the highest power of numerator is
2m then of denominator is 2m —1 which gives pole at ∞ or it can be 2m +1
which gives zero at ∞. And as lowest powers also differ by unity, there is
pole or zero at the origin.
9. Residues at the imaginary axis poles are real and positive.
10.The number of elements required in any of the four forms of realization is
equal to the highest power of s in the LC Immittance Function as a whole.
11.The slope of the graph of reactance against frequency is always positive.
We know that there is either pole or zero at the origin.
Consider
So there is pole at the origin. Let us obtain the graph of X(ω). At the starting pole
at s= 0 i.e. ω = 0 the X(ω) is infinity. As ω increases, X(ω) also increases and at
next critical frequency ω = ω2 where there is a zero, X(ω) becomes zero. As ω
further increases from ω = ω2, X(ω) increases and becomes infinity at ω = ω 3 where
there is a pole. At this frequency, X(ω) suddenly changes the sign and goes from +
∞ to — ∞, such that as we pass through ω = ω 3, slope of the graph always remains
positive. It becomes zero at ω = ω4 where there is a zero. The nature continues such
that the slope d/dω [X(ω)] always positive. The nature is shown in the Fig. 7.5 for
the Z(s) considered.
Between ω1 and ω2, nature of X(ω) is capacitive while between ω 2 and ω3 it is
inductive. At ω3 = 2, it changes from inductive to capacitive suddenly so that slope
of the graph remains always positive.
For a zero at the origin, the reactance curve takes the form as shown in the Fig. 7.6.
For a LC function,
From the nature of the reactance curves shown in the Fig. 7.5 and 7.6 with pole at
origin and zero at origin respectively, it can be seen that curve has to change its
sign at the poles to maintain the positive slope.
Hence to maintain the positive slope, the poles and zeros must separate each other
and in such a way that the poles and the zeros alternate along the real frequency
axis. This is called Separation Property or Foster Reactance Theorem for the
reactance functions.
Thus if
then H is scale factor. Now ωz1 can be zero which gives zero at the origin or has
finite existence if there is pole at the origin.
According to the Reactance theorem pole and zero frequencies must satisfy the
relation given by,
Realization of Immitance Function of LC Networks:
The realization of driving point immitance functions of LC networks can be done

by any of the four forms discussed earlier. The forms which are used are,
1. Foster I form
2. Foster II form
3. Cauer I form
4. Cauer II form
RC Driving Point Impedance function:

As the name indicates, the RC networks consist of only R and C components.
There is no inductor in RC networks. The RC Driving Point Impedance function is
denoted as ZRC(s). The properties of RC Driving Point Impedance function and the
properties of driving point admittance function of RL network are identical. Thus
the properties ZRC(s) and YRL(s) are same. There are no complex poles in RC
network function. The poles and zeros alternate each other and are located in left
half of s plane. To understand the properties of RC network function consider a
driving point impedance function of RC network as,
The poles are at s = 0, – 2
The zeros are at s = –1, – 4
The pole-zero plot is shown in the Fig. 7.11.

Properties of RC Driving Point Impedance Function:
Referring to the pole zero plot of ZRC(s) function considered, the various properties
of RC Driving Point Impedance function can be stated as,
1. The poles and zeros are simple. There are no multiple poles and zeros.
2. The poles and zeros are located on negative real axis.
3. The poles and zeros interlace (alternate) each other on the negative real
axis.
4. We know that the poles and zeros are called critical frequencies of the
The critical frequency nearest to the origin is always a pole. This may
be located at the origin.
5. The critical frequency at a greatest distance away from the origin is
always a zero, which may be located at ∞ also.
6. The partial fraction expansion of ZRC(s) gives the residues which are
always real and positive.
7. There is no pole located at infinity.
8. The slope of the graph of Z(σ) against σ is always negative.
9. There is no zero at the origin.
10.The value of ZRC(s) at s=0 is always greater than the value of ZRC(s) at s =
∞.
It can, be seen that for a ZRC considered ZRC(0) = ∞ while ZRC(∞) = 1.
Consider a simple RC network as shown in the Fig. 7.12.

Let us plot ZRC(σ) against σ where s = σ.
To find the slope of ZRC(σ) against σ find d ZRC(σ)/dσ
Thus the slope of ZRC(σ) against σ is always negative for any value of σ.
The graph of Z(σ) against σ for the RC network function is shown in the Fig. 7.13.
ω1,ω2,ω3 and ω4 are the critical frequencies. At the critical frequencies like ω 3, the
Z(σ) changes its sign suddenly such that the slope always remains negative. The
value of Z (∞) is constant so graph runs parallel to the σ axis, finally.
The nature of Z (σ) against σ graph when there is no pole at the origin is shown in
the Fig. 7.14.
All the properties of driving point admittance function of RL network [YRL(s)] are
exactly identical to the properties of driving point impedance function of RC
network [ZRC(s)].
Realization of Impedance Function of RC Network:
As mentioned earlier, the realization of ZRC(s) function can be achieved using

Foster I, Foster II, cauer I or cauer II form. Remember that the number of elements
are not equal to highest power of s in overall Z(s) for RC networks.
RL Driving Point Impedance:

The RL networks consist of only R and L components. There is no capacitor
present in such networks. The RL Driving Point Impedance of such networks is
denoted as ZRL(s). The properties of driving point impedance function of RL
networks [ZRL(s)] and the driving point admittance function of RC networks
[YRC (s)] are exactly identical.
The RL impedance function is dual of RC admittance function. There are no

complex poles in RL network functions and poles and zeros are located in left half
of s-plane.
Consider the RL Driving Point Impedance network as,

The poles are at s = – 2 and – 4 while the zeros are at s = – 1 and – 3.
The pole-zero plot is as shown in the Fig. 7.19.
Properties of RL Driving Point Impedance Functions:
Referring to the pole zero plot of ZRL(s) function considered above, the various
properties of RL, driving point impedance functions can be stated as,
1. The poles and zeros are simple. There are no multiple poles and zeros.
2. The poles and zeros are located on negative real axis.
3. The poles and zeros interlace each other on the negative real axis.
4. The poles and zeros are the critical frequencies. The critical frequency
nearest to the origin is always a zero, which may be located at the
origin.
5. The critical frequency at a greatest distance away from the origin is
always a pole, which may be located at infinity also.
6. Partial fraction expansion of ZRL(s) gives the residues which are negative
and real hence to obtain positive residues the expansion of Z RL(s)/s is
obtained.
7. There can not be a pole at the origin.
8. The slope of the graph of z (σ) against σ is always positive.
9. The value of ZRL(s) at s = 0 is always less than the value of ZRL(s)at s = ∞.
For the example considered ZRL(0) = 3/8 while ZRL(∞) = 1.

It can be easily verified from a simple RL network that the slope of the graph Z(σ)
against σ is always positive.
The graph of Z(σ) against σ for the RL network function without a zero at the
origin is shown in the Fig. 7.20.
ω1,ω2,ω3 and ω4 are the critical frequencies. At the critical frequencies

like ω2 and ω4 the sign of Z(σ) changes suddenly such that the slope always remains
positive.
The nature of Z(σ) against σ graph for the RL network function when there is a
zero at the origin is shown in the Fig. 7.21.
Realization of Impedance Function of RL Network:
The realization of ZRL(s) function can be obtained using Foster I, Foster II, Cauer I
and Cauer II forms. The number of elements are not equal to highest power of s in
the overall Z(s) for RL networks.
Elements of Transfer Function Synthesis:

A transfer function is generally defined to specify the properties of a two port
network. Hence Elements of Transfer Function Synthesis is also known as two port
synthesis.
Consider a two port network as shown in the Fig. 7.26.

V1 and I1 are voltage and current at port 1 while V2 and I2 are voltage and current at
port 2.
Thus there are four variables associated with the network, the two are voltage
variables V1, V2 and other two are current variables I1 and I2.
A ratio of voltage or current at one port to the voltage or current at the other port is
called a transfer function of the network.
The following transfer functions can be defined for the two port network,
A two port network is generally defined by its parameters such as z parameters and
y parameters.
In the transfer function synthesis, the two port network is terminated at port 2 in
resistance R. Let us express the transfer impedance and admittance functions of a
two port terminated network, in terms of z and y parameters.
Note that the transfer functions are denoted by capital letters while the network
parameters are denoted by small letters.
Transfer Functions of Two Port Terminated Network:
Consider a two port network, terminated in resistance R as shown in the Fig. 7.27.
Z21 = V2/V1 = Transfer impedance function
From the z parameters we can write,
Substituting in (1),
The transfer impedance function Z 21 is denoted as capital letter while z 21, z22 are the
z parameters, denoted by small letters.
Now
Y21 = I2/V1 = Transfer admittance function
From the y parameters we can write,
Substituting in (4),
Similar to impedance and admittance functions, the other transfer functions can
also be expressed in z and y parameters.
The results are helpful in the synthesis of two port networks
Properties of Transfer Functions:
The transfer function is the ratio of two polynomials in s.
It is denoted as T(s).
The T(s) is the transfer function of linear network consisting of lumped bilateral
passive elements. The properties of such T(s) are,
1. The T(s) is real when s is real.

When T(s) is rational with real coefficients, then this property gets satisfied.
2. The poles of T(s) are simple having zero or negative real parts. T(s) has no
poles in right half of s plane.
3. The T(s) has no multiple poles on the jω axis.
4. The degree of polynomial N(s) can not exceed the degree of the polynomial
D(s) by more than one.
5. The polynomial D(s) must be Hurwitz polynomial.
6. The T(s) can be split into odd and even parts as,
where m1, m2 are even parts and n1, n2 are odd parts of N(s) and D(s) respectively.
Replacing s by jω we get,
From T(jω) the amplitude and phase response of T(s) can be obtained. The
amplitude of T(jω) i.e. |T(jω)| is always even function of ω and amplitude response
is symmetrical about y-axis.
If ∠ T(jω) = 0 for ω = 0 then the phase response is always an odd function of ω and
the phase response is symmetrical about x-axis.
Properties of Open Circuit and Short Circuit Parameters:
The specific properties of z and y parameters can be stated as,
1. The poles of the parameter z21 are also the poles of the parameters
z11 and z22. But not all the poles of z11 and z22 are the poles of z21.
2. The poles of the parameter y21 are also the poles of the parameters
y11 and y22 , But not all the poles of y11 and y22 are the poles of y21.
3. Let y11, y22 and y12 all have a pole at s = s 1. Let k11 is the residue of y11, k22 is
the residue of y22 and k12 is the residue of y11 at the pole s = s1.
Then for all the two port networks,
This is called a residue condition.
Asymmetrical Network in Network Analysis:

An asymmetrical network has following electrical properties,
 Iterative impedance
 Image impedance
 Image transfer constant
Iterative impedance (Z′0):
Consider that infinite asymmetrical networks having identical electrical properties

are connected in cascade as shown in the Fig. 8.4 (a) and (b).
The iterative impedance is the impedance measured at one pair of terminals of the
network in the chain of infinite networks as shown in the Fig. 8.4 (a) and (b). This
is the impedance measured at any pair of terminals of the network when other pair
of terminals is terminated in the impedance of the same value as shown in the Fig.
8.4 (c) and (d).
The iterative impedances for any asymmetrical network are of different values
when measured at different ports of the network. The iterative impedances are
represented by Z′01 and Z’02 respectively at port 1 and port 2.
Image impedances (Zi):
Similar to the iterative impedances, the image impedances are also of different
values at different ports. Let the image impedances be denoted by Zi1 and Zi2.
Consider that the asymmetrical network is terminated with image impedance of
port 2 Zi2 at its output pair of terminals then the impedance measured at its input
pair of terminals will be image impedance of port 1 i.e. Zi1. Similarly if port one is
terminated in the image impedance of port 1 i.e. Zi1 then the impedance measured
at port two will be the image impedance of port 2 i.e. Zi2. These conditions are
illustrated by the Fig. 8.5 (a) and (b).
When an asymmetrical network is terminated in image impedances at both the

ports the network is called correctly terminated asymmetrical network as shown in
the Fig. 8.6.
Image transfer constant (eθ):
When an asymmetrical network is terminated in its image impedances at both

the ports as shown in the Fig. 8.6, then the ratio of currents I 1/I2 will be different
from the E1/E2.
Hence image transfer constant θ is defined as

The real part of image transfer constant is called image attenuation constant;
while the imaginary part is called image phase constant.
After detailed discussion of various properties of symmetrical and asymmetrical

networks let us study the properties of some of the important symmetrical and
asymmetrical networks including symmetrical T, π, lattice networks and
asymmetrical networks such as half sections, L sections.
Transmission Zero:
The Transmission Zero functions describe the transmission of signal from one end
to other end. The Fig. 7.28 shows the transfer of signal from one network to other
through a capacitor C.
There is a transmission through the capacitor producing finite output for finite
input, at all the frequencies. But at s = 0 i.e. at zero frequency, the capacitor
reactance becomes infinite and it acts as open circuit. There can not be any output,
though there is finite input at zero frequency. Thus s = 0 is called zero of
transmission for this network.
The frequencies at which, zero output results for a finite input are called zeros of
transmission. At these frequencies Transmission Zero results. The zeros of T.F.
T(s), which are the critical frequencies, are the zeros of transmission for the
network described by that transfer function T(s). So a zero of transmission is a zero
of transfer function.
Consider the network shown in the Fig. 7.29.

The transmission is through LC parallel circuit. For ideal L and C the impedance of
the parallel resonant circuit is ∞. It acts as open circuit at a resonant frequency of
1/√LC
Now if finite input is applied at ωr = 1/√LC, then LC resonant circuit acts as open
circuit which results into Transmission Zero. Thus the zeros of transmission for
this networks are s = ± j 1/√LC.
Consider another network shown in the. Fig. 7.30.
The impedance of the reactance is sL. At zero frequency, it acts as short circuit.
Hence zero output results for finite input at zero frequency. Hence zero of
transmission is at s =0 for an inductor in shunt arm.
Consider a network with LC series network as transmitting network as shown in
the Fig. 7.31. For ideal L and C, as resonant frequency ω r = 1/√LC, the impedance
of LC series resonant circuit is zero. It acts as short circuit, producing zero output.
For such a network, the zeros of transmission are at s = ± j 1/√LC.
Consider the networks shown in the Fig. 7.32 and Fig. 7.33 with inductor in series
arm and capacitor in shunt arm respectively.
At s = ∞, the inductor acts as open circuit and capacitor acts as short producing
zero output. Hence the zero of transmission is at s = co for both these networks.
In general, zeros of transmission are zeros of the transfer function. All the transfer
functions Z12,Z21,Y21 etc have the same zeros of transmission except in some special
cases.
The table 7.2 gives possible combinations of elements and the corresponding
Transmission Zero.
Minimum phase functions : The transfer function having zeros of transmission

only on the jω axis or in the left half of s plane, is called minimum phase
function.
Non-minimum phase functions : The transfer function having one or more zeros
of transmission in right half of s plane is called non-minimum phase function.
Any transfer function of a passive ladder network is always minimum phase

function.
Symmetrical Network in Network Analysis:

A symmetrical network has two important properties as
 Characteristic impedance (Z0)

 Propagation constant (γ)
The two networks having the same electrical properties i.e. characteristic
impedance (Z0) and propagation constant (γ) are called equivalent
networks or identical networks.
Characteristic Impedance (Z0):
Consider that infinite number of identical symmetrical networks are connected

in cascade or tandem as shown in the Fig. 8.1 (a). The input impedance measured
at the input terminals of the first network in the chain of infinite networks will have
some finite value which depends on the network composition. This impedance is
the important property of a symmetrical network. Thus the characteristic
impedance of a symmetrical network is the impedance measured at the input
terminals of the first network in the chain of infinite networks in cascade and it is
represented by Z0.
If first network is disconnected from the chain as shown in the Fig. 8.1 (b), then
also the input impedance measured at the input terminals of second network will be
Z0 again as number of networks in the chain are still infinite. That means we can
replace this chain by impedance Z0 at the output port of the first network as shown
in the Fig. 8.1 (c). Then the impedance at input terminals of the first network will
be still Z0.
Thus in general when any symmetrical network is terminated in its characteristic

impedance Z0, the input impedance will also be Z0.
This property is true for output impedance if the symmetrical network terminated
in Z0 is driven by a generator with internal impedance equal to Z 0. In such network,
the output impedance will be Z0 only. The network terminated in characteristic
impedance at input as well as output terminals is said to be correctly
terminated or properly terminated symmetrical network as shown in the Fig.
8.2.
Propagation Constant (γ):
Consider a chain of identical symmetrical networks connected in cascade as shown

in the Fig. 8.3.
The current leaving any section will be definite proportion of that entering section
and in general will be out of phase with it. Thus the relationship between the
currents entering and leaving the section is a vector quantity with both modulus
and angle. This quantity is represented in the form e γ for convinence where γ is a
complex number given by γ = α + jβ
Let the ratio of input to output current be given by
Since all the sections are identical, we can write

Hence for n identical sections connected is cascade the ratio of input to output
current is given by
Note that input current is represented by sending end current, I S; while the output
current is represented by receiving end current, IR. Above equation can be written
as
when enα gives ratio of absolute magnitudes of sending end current to receiving end
current and nβ gives the phase angle between these two currents.
If the network is correctly terminated, then we can write,
The real part α of the propagation constant γ is called attenuation constant and it
is measured in nepers.
Similarly for n-sections,

The imaginary part β of propagation constant γ is called phase constant and is
equal to the angle in radians by which output current leaving section lags that input
current entering section. For n-sections, the phase constant will be nβ radians.
Symmetrical pi Network in Network Analysis:

The Symmetrical pi Network in Network Analysis is another important network in
line transmission fulfilling the conditions of total series and shunt arm impedances
as Z1 and Z2 respectively. Thus the series arm impedance of a it network is selected
as Z1 and to have a total shunt arm impedance of Z 2, each shunt arm impedance is
selected as 2Z2 as shown in the Fig. 8.15.
Similar to the symmetrical T network, let us derive the expressions for the
characteristic impedance (Z0) and propagation constant (γ) of the Symmetrical pi
Network in Network Analysis.
(A) In terms of series and shunt arm impedances
Consider a symmetrical π network terminated at its output terminals with its

characteristic impedance Z0 as shown in the Fig. 8.16.
By the property of the symmetrical network, the input impedance of such network
terminated with Z0 at other port is equal to Z0.
The input impedance of a symmetrical π network is given by

Multiplying numerator and denominator by the factor Z1/4,
Here the characteristic impedance of it section is indicated by Z0π and that of T

section by Z0T.
Taking square root on both the sides,
(B) In terms of open and short circuit impedances
Consider Symmetrical pi Network in Network Analysis shown in the Fig. 8.17 (a)
and Fig. 8.17 (b).
Multiplying equations (3) and (4), we can write,

Thus the characteristic impedance of symmetrical π network is given by
Consider correctly terminated symmetrical π network as shown in the Fig. 8.19.
As the network is symmetrical, by definition,
By potential divider rule,

Rearranging the terms,
where γπ is the propagation constant of symmetrical π network. Putting value of

Z0π interms of Z0T,
Series and shunt arm impedances of symmetrical π network in terms of

Z0π and γπ:
For symmetrical T network, each series arm is given by,

Similarly shunt arm of symmetrical T network is given by
Hence π network with series and shunt arms expressed in terms of Z0π and γπ, is as
shown in the Fig. 8.20.
Symmetrical T Network in Network Analysis:
In line transmission theory, the symmetrical T network is the most frequently used
network. The condition in the symmetrical T network is that the total series arm
impedance and shunt arm impedance must be Z 1 and Z2 respectively. To have a
total series arm impedance of Z1 the two series arm impedances must be selected as
,
Z1/2 each as shown in the Fig. 8.7.
Let us derive the expressions for the characteristic impedance (Z 0) and propagation
constant (γ) in terms of the network elements.
(A) In terms of series and shunt arm impedances
Consider a symmetrical T network terminated at its output terminal with its

characteristic impedance as shown in the Fig. 8.8.
By the property of the symmetrical network, the input impedance of such network
terminated in Z0 at other port is equal to Z0.
The input impedance of a T network is given by,
For symmetrical networks, impedances measured at any pair of terminals with

other pair of terminals either open circuit or short circuit are of same value.
Consider symmetrical T network with terminals 2-2′ either open circuit or short
circuit as shown in the Fig. 8.9.

Multiplying equations (3) and (4), we can write,
Hence in any symmetrical network, the characteristic impedance Z 0 is the

geometric mean of open and short circuit impedances measured at any pair of
terminals.
Consider properly terminated symmetrical T network as shown in the Fig. 8.12.

The currents entering and leaving network are assumed to be IS and IR respectively.
Applying KVL to the outer loop,
Hence
The propagation constant of T section is therefore given by
Putting value of Z0 in equation (1), we can write,
From above equation we can write,

Adding equations (3) and (4), we get,
Subtracting equation (4) from (3), we get,
Dividing equation (6) by (5) we get,

Series and shunt arm impedances in terms of characteristic impedance (Z 0) and
propagation constant (γ)
Consider equation (5),
Consider equation (6),
Dividing equation (8) and (9),
Thus each series arm impedance, in terms of the characteristic impedance and
propagation constant is given by
Similarly, shunt arm impedance in terms of the characteristic impedance and
propagation constant is given from equation (6) as
Hence T network with components expressed in terms of characteristic impedance

and propagation constant is as shown in the Fig. 8.13.
Lattice Network in Network Analysis:

The Lattice Network in Network Analysis is a symmetrical and balanced with four
arms as shown in the Fig. 8.21 (a). The arms consisting impedance Z A are
called series arms of the lattice network. The arms consisting impedance Z B are
called shunt or diagonal arms. The lattice network can be rearranged in the
bridge structure as shown in the Fig. 8.21 (b) which is very suitable for the analysis
of the lattice network.
As lattice network is a symmetrical network, let us derive expressions for the
characteristic impedance (Z0) and propagation constant (γ). It is very convenient to
use bridge structure of the lattice network for the calculation of propagation
constant.
Consider closed path 1-2-2′-1′-1, applying KVL, we get
Consider closed path 1-2′-2-1′-1, applying KVL, we get
From equation (1),
From equation (2),

Equating equations (3) and (4),
But by the property of symmetrical network, the input impedance of the network
terminated in its characteristic impedance is equal to Z0.
Let
Putting value of E/Is in above equation,
For the calculation of open and short circuit impedances arranging bridge structure
of the lattice network as shown in the Fig. 8.22 (a) and (b).
Multiplying equations (6) and (7) , we can write,
For any symmetrical network, propagation constant can be expressed as,
Consider equations for current IR given by equation (3) and (4) derived in section
8.6.1.
But we know that E = Is . Z0
Impedances ZA and ZB in terms of characteristic impedance (Z0) and

propagation constant (γ)
Consider equation,
Similarly,
Hence lattice network with impedances expressed in terms of characteristic

impedance and propagation constant is as shown in the Fig. 8.23.
After detail analysis of some of the important symmetrical networks now consider
analysis of typical asymmetrical networks.
Half Section in Four Terminal Network:

The Half Section in Four Terminal Network can be obtained by splitting
symmetrical T and π networks into two identical networks. The half sections
obtained by splitting shunt arm of symmetrical T network in two equal impedances
and series arm of symmetrical π network in two equal impedances are identical as
shown in the Fig. 8.24 (a) and (b).
Consider a half section as shown in the Fig. 8.24. As half section is of
asymmetrical type, we have to analyze this network on the basis of iterative
impedances and image impedances separately.
Iterative Impedances (Z′0):
By the property of asymmetrical network, the iterative impedances at the two ports
of the network are of different value and they are represented by Z′01 and Z′02.

Finding roots of quadratic equation,
Find roots of above equation we can write,

Image Impedances (Zi):
For any asymmetrical network, the image impedance at any port is the geometric
mean of open and short circuit impedances at that port.
If Zi1 is the image impedance at port 1, then it is given by
Similarly if Zi2 is the image impedance at port-2, then it is given by
Consider a half section without any termination as shown in the Fig. 8.26.
The impedances at terminals 1 – 1′ with terminals 2 – 2′ open and short circuited

are represented by Z1OC and Z1SC respectively. These impedances are given by
Thus, image impedance Zi1 at port 1 is given by,

The impedances at terminals 2 – 2′ with terminals 1 – 1′ open and short circuited
are represented by Z2OC and Z2SC respectively. These impedances are given by
Thus, image impedance Zi2 at port 2 is given by,
From equations (3) and (6) it is clear that the image impedances at port 1 and port
2 are the characteristic impedances of symmetrical T and π networks having total
series and shunt arm impedances as Z1 and Z2 respectively. Because of this property
of the half section, it can be used to match a T section with a π section if both are
having same series and shunt arm impedances.
L Section in Four Terminal Network:
L Section in Four Terminal Network is very much similar to half section consisting
series and shunt arm impedances Z1 and Z2 respectively as shown in the Fig. 8.28.
The ladder network can be analyzed into a series of L sections rather than T and π
sections. The main application of L section is in matching of impedances between
two sections.
Interative Impedance (Z′0):

Finding roots, we can write,
Finding roots, we can write,
Image Impedances (Zi):

Thus, image impedance Zi1 at port 1 is given by
Similarly,
Thus, image impedance Zi2 at port 2 is given by
Ideal Filter Characteristics:

Ideal Filter Characteristics – The range of frequencies over which attenuation by
filter is zero is called pass band.
The range of frequencies over which attenuation is infinite is called stop

band or attenuation band of the filter.
The frequencies which separate the pass band from attenuation or stop band are
called cut-off frequencies of the filter represented as fc.
Here we will realize filters using symmetrical unbalanced T or π sections as shown
in the Fig. 9.1 (a) and Fig. 9.1 (b).
Earlier we have studied two important properties of symmetrical networks,

namely, characteristic impedance (Zo) and propagation constant (γ).
The characteristic impedance of symmetrical T network is given by,
The propagation constant of symmetrical T network is given by,
Similarly the characteristic impedance and propagation constant of symmetrical π

network are given by following expressions.
Basic Filter Sections and Ideal Filter Characteristics:
If filter passes all frequencies upto the cut off frequency and attenuates all
frequencies above it, then it is called low pass filter.
If filter attenuates all frequencies upto the cut off frequency and passes all
frequencies above it, then it is called high pass filter.
Any simplest type of filter has only one pass band, one stop band and a single cut
off frequency.
One can design a filter with two cut off frequencies to get two more filter sections.
If filter passes all the frequencies between the two cut off frequencies and
attenuates all other frequencies, then it is called band pass filter.
On the other hand, if filter attenuates all the frequencies between the two cut off
frequencies and passes all other frequencies, then it is called band stop filter or
band elimination filter.
An ideal filter would have zero attenuation in the pass band and infinite attenuation
in the stop band. An ideal filter characteristics are as shown in the Fig. 9.2 (a) to
(d).
Filter Fundamentals in Network Analysis:

Filter Fundamentals in Network Analysis – The complete study of behavior of any
filter section needs calculations of its characteristic impedance (Z 0), propagation
constant (γ), attenuation constant (α) and phase constant (β), using advanced
mathematical calculations at any frequency. However we can easily predict pass
band and stop band of a filter from an elementary consideration of variation of
Z0 with frequency.
An important consideration for all Filter Fundamentals in Network Analysis is that

they are constructed from purely reactive elements, otherwise the attenuation could
never become zero. From the expressions of the characteristic impedances of T and
π sections it is clear that, the characteristic impedance depends on the reactances
Z1 and Z2 offered by purely reactive elements used in series and shunt arms of a
filter. Hence the characteristic impedance Z 0 varies with frequency as Z1 and
Z2 both vary with frequency.
Hence in a filter, over the range of the frequencies Z 0 may be either real or
imaginary. Over the range of frequencies if Z0 is real, the filter and its terminating
impedance will absorb power from any generator connected to it. Since filter is
composed of reactive elements, it cannot itself absorb power. Hence all the power
delivered by generator is passed to the load. Thus there is no attenuation i.e. α = 0.
This indicates pass band.
On the other hand, if Z0 is imaginary or purely reactive, the filter and its
termination cannot absorb any power. Thus, no power is passed to the load. Thus
attenuation is very high, ideally attenuation is infinite. This indicates stop band.
Above discussion is also useful in determining the cut off frequency of any filter.
We have already seen that in pass band Z 0 is real resistive while in stop band it is
purely imaginary or reactive.
Thus, we can define cut off frequency f c is the frequency at which Z0 changes from
being real to being imaginary.
For a T section, the characteristic impedance is given by
For purely reactive T section, let Z 1 = jX1 and Z2 = jX2. Substituting values of
Z1 and Z2 in above formula, we can write,
Thus, Z0 is purely imaginary if X1 and (X1/4 + X2) have the same sign. This gives
stop band. We get Z0 purely resistive if X1 and (X1/4 + X2) have the opposite signs.
This gives pass band.
By drawing the reactance sketches for X 1 and X1/4 + X2 against frequency we can
easily get cut off frequency. To get cut off frequency, the rule is as follows :
“Band of frequencies for which the curves lie on opposite side of frequency axis is
pass band while the band of frequencies for which the curves lie on same side of
frequency axis is stop band.” The change over point gives cut off frequency.
Constant K Sections:
A T or π section in which series and shunt arm impedances Z 1 and Z2 satisfy the
relationship → Z1. Z2 = R20 where R0 is a real constant is called constant K section.
R0 is real resistance which is frequency independent. R0 is known as design

impedance of the section.
For the same series and shunt arm impedances the characteristics impedances of T
and π sections can be related with each other as follows,
For a constant K section we can write,
The constant K sections either T or π, of any type of filter are known as prototype
sections as other more complex sections may be derived from prototypes.
Low Pass Filter:

The prototype T and π low pass filter sections are as shown in the Fig. 9.3.
Design Impedance (R0):
Here in low pass filter sections,
Total series arm impedance Z1 = jωL
Total shunt arm impedance Z2 = -j/ωC
Hence, Z1 . Z2 = (jωL) (-j/ωC) = L/C which is real and constant. Hence sections are
constant K sections so we can write,
Reactance Curves and Cut-off Frequency Expression:
As both T and π sections have same cut-off frequency, it is sufficient to calculate

fc for the ‘T’ section only.
The reactance curves are as shown in the Fig. 9.4.

From above characteristic it is clear that all the reactance curves have positive
slope as all curves slope upward to the right side with increasing ω.
The curves are on opposite sides of the frequency axis upto point A; while on the
same side, from point A on wards. Hence all the frequencies upto point A give pass
band and above point A give stop band. Thus point A marks cut-off frequency
given by ω = ωc.
At point A, ω = ωc, the curve for (X1/4 + X2) crosses the frequency axis, hence we
can write,
The algebraic approach to calculate cut-off frequency is as follows.

From above expression it is clear that, Z 0T is real if ω2LC/4 < 1 and imaginary if
ω2LC/4 > 1. Hence condition ω2LC/4 -1 = 0 gives expression,
Thus, above prototype section passes all frequencies below ω = 2/√LC while
attenuates all frequencies above this value. Therefore cut-off frequency of low pass
filter is given by
Above frequency comes out to be the same as calculated by reactance sketch

method.
Variation of Z0T and Z0π with Frequency:
Consider expression
From equation (2), we can write

Similarly we can write,
Hence
From equation (5), it is clear that as frequency increases from 0 to f c, Z0T decreases
from R0 to 0 in passband. For π section, from equation (6), it is clear that in pass
band as frequency increases for 0 to fc, Z0π increases from R0 to ∞.
The variation of Z0T and Z0π with frequency is as shown in the Fig. 9.5.
Variation of Attenuation Constant α with Frequency:
In pass band attenuation is zero. In stop band attenuation is given by,
In stop band, as frequency f increases above f c, attenuation also increases. The

variation of α with frequency is as shown in the Fig. 9.6.
Variation of Phase Constant β with Frequency:

In stop band, phase constant β is always equal to π radian. In pass band where α =
0, the phase constant β is given by
As frequency increases from 0 to f c, β also increases from 0 to π radian. The

variation of β with frequency is as shown in the Fig. 9.7.
Design Equations of Prototype Low Pass Filter:
The design impedance R0 and cut-off frequency fc can be given in terms of L and C
as follows.
Dividing equation for R0 by fc, we get,

Multiplying equation for R0 and fc we get,
Equations (9) and (10) are called design equations for prototype low pass filter
sections.
High Pass Filter:

The prototype high pass filter T and π sections are as shown in the Fig. 9.9.
Total series arm impedance Z1 = -j/ωC
Total shunt arm impedance Z2 = jωL
Hence, Z1 . Z2 =(-j/ωC) (jωL) = L/C which is real and constant. Hence above
sections are constant K sections. So we can write,
Reactance Curves and Expression for Cut-off Frequency:
As both T and π sections have same cut-off frequency, it is sufficient to calculate

the cut-off frequency for the T section only.
The reactance curves are as shown in the Fig. 9.10.
From above characteristics it is clear that all the reactance curves have positive
slope as all curves slope upward to the right side with increasing ω.
Here the curves are on the same side of the horizontal axis up to the point B, giving
a stop band. For frequencies above point B, the curves are on opposite sides of the
axis, giving pass band. Thus, point B gives cut-off frequency, represented as ω =
ωc .
At point B, ω = ωc, the curve for (X1/4 + X2) crosses the frequency axis, hence we
can write,
The algebraic approach to calculate cut-off frequency is as follows.
From above expression it is clear that, Z0T is real if 1/4ω2LC < 1 and imaginary if
1/4ω2LC > 1. Hence condition 1 – 1/4ω2LC = 0 gives expression,
Thus, above prototype section passes all frequencies above ω = 1/2√LC while
attenuates all frequencies below this value. Therefore cut-off frequency of high
pass filter is given by
Above frequency comes out to be same as frequency calculated by reactance

sketch method.
Variation of Z0T and Z0π with Frequency:
Consider expression for Z0T as
From equation (2) we can write,
Similarly we can write,
Hence,
From equation (5), it is clear that as frequency f increases from fc to ∞ in pass band,
Z0T also increases from 0 to R0. For π section, from equation (6), it is clear that as
frequency increases from fc to ∞, Z0 decreases from ∞ to R0 in pass band. The
variation of Z0T and Z0π with frequency is as shown in Fig. 9.11.
Variation of Attenuation Constant (α) with Frequency:
In pass band, attenuation is zero (α = 0). In stop band attenuation is given by
In stop band, as frequency f increases from 0 to f c, attenuation decreases from ∞ to

0. The variation of attenuation constant α with frequency is as shown in the Fig.
9.12.
Variation of Phase Constant β with Frequency:
In stop band, phase constant β is always π radian. In pass band where α = 0, the
phase angle β is given by
From the above equation it is clear that in pass band when frequency f increases
from fc to ∞, β decreases to 0. The variation of phase constant β with frequency is
as shown in the Fig. 9.13.
Design Equations of Prototype High Pass Filter:
The design impedance R0 and cut-off frequency fc for high pass filter section can be
given in terms of L and C as follows
Dividing equation for R0 by fc, we get,

Multiplying equation for R0 and fc, we get,
Equation (9) and (10) are called design equations of prototype high pass filter
sections.
Band Pass Filter:

Band pass filter pass a certain range of frequencies (called as pass band) while
attenuate all other frequencies. Such band pass filters can be obtained by
connecting low pass filter sections in cascade with high pass filter sections as
shown in Fig. 9.15.
In above type of connection, the cut-off frequency of low pass filter section must
be selected higher than that of high pass filter section.
Although cascade connection of low pass filter and high pass filter sections
functions properly as band pass filter, it is more economical to combine both
sections in one single filter section. An alternative form of band pass filter can be
obtained either as a T or π section if series arm contains a series resonant circuit
while the shunt arm contains a parallel resonant circuit as shown in the Fig. 9.16
(a) and (b).
The band pass filter characteristics can be obtained by using conventional band
pass filter (either T or π type) as shown in the Fig. 9.16, if the series resonant
frequency of the series arm is selected same as anti resonant frequency of the shunt
arm. Consider T type band pass filter section as shown in the Fig. 9.16 (a). Let the
frequency of series and shunt arm be ω0 rad/sec.
Then, for series arm, frequency of resonance is given by,
Similarly for shunt arm, frequency of anti resonance is given by,
From equations (1) and (2), for same resonant frequencies of series and shunt arms
we can write,
Total series arm impedance Z1 is given by
Similarly, total shunt arm impedance Z2 is given by

Hence, Z1 Z2 = L2/C1 = L1/C2 which is real and constant. Hence above sections
are constant k sections. So we can write,
Reactance Curves and Expressions for Cut-off Frequencies:
To verify the band pass characteristics, let Z 1 = j X1 and Z2 =j X2. Similar to the
reactance curves drawn for low pass filter section and high pass filter section,
sketching reactances X1 and (X1/4 + X2) against frequency f as shown in the Fig.
9.17.
From the above characteristics it is clear that the reactance curves for X 1 and (X1/4
+ X2) are on the same sides the axis below f 1 and above f2. At the same time, the
reactance curves between f1 and f2 are on opposite sides of frequency axis. Thus
frequencies between f1 and f2 constitute a pass band ; while the frequencies below
f1 and above f2 give stop band. Hence the section considered shows band pass filter
characteristics where f1 and f2 are lower and upper cut-off frequencies of the filter.
In band pass filter, condition for cut-off frequency is,
But from the condition of constant-k filter section, Z1 Z2 = R20
From above equation (7) it is clear that the value of the series arm impedance
Z1 can be obtained at two different cut-off frequencies namely f 1 and f2. So at f = f1,
Z1 = – j(2 R0) and at f = f2 , Z1 = + j(2 R0). Thus impedance Z1 at f1, i.e. lower cut-off
frequency, is negative of the impedance Z1 at f2 i.e. upper cut-off frequency. Hence
we can write,
But from equation (1) we can write,
Substituting value of (L1 C1) in above equation (8), we can write,
Simplifying above equation,
Hence, above equation (9) indicates that frequency of resonance of the individual
arms is the geometric mean of two cut-off frequencies:
Variation of Z0T and Z0π, Attenuation Constant (α) and Phase Constant (β) with Frequency:
The variations of Z0T and Z0π, attenuation constant (α) and phase shift (β) with
frequency are as shown in the Fig. 9.18. (a), (b) and (c). Consider that the design
impedance of band pass filter is R0 and cut-off frequencies are f1 and f2.
Design Equations:
Consider that the filter is terminated in design impedance R 0 and the cut-off
frequencies are f1 and f2.
Then from equation (7), at the lower cut-off frequency f1, we can write,
But for band pass filter constant k section

Substituting the value of C1 from equation (10),
As f20 = f1 f2, we get
From equation (6), we can write,
Substituting value of L1 from equation (11),
From equation.(6), we can write,

Substituting value of C1 from equation (10),
Equations (10) to (13) are called design equations of prototype band pass filter
sections.
Band Stop Filter:

Band Stop Filter stop a range of frequencies between two cut-off frequencies f 1 and
f2 while pass all the frequencies below f 1 and above f2. Thus range of frequencies
between f1 and f2 constitutes a stop band in which attenuation to the frequencies is
infinite ideally. The frequencies below f 1 and above f2 constitute two separate pass
bands in which attenuation to the frequencies is zero ideally.
The Band Stop Filter can be obtained by connecting low pass filter and high pass
filter sections in parallel where cut-off frequency of the low pass filter section is
less than that of the high pass filter section. But the economial form of the band
elimination filter can be obtained by combining the low pass and high pass filter
section if series, arm contains parallel resonant circuit while shunt arm contains
series resonant circuit as shown in the Fig. 9.20 (a) and (b).
The band elimination characteristics can be obtained by using conventional Band
Stop Filter (either T or π type) as shown in the Fig. 9.20, if the series resonant
frequency of the shunt arm is selected same as the parallel resonant frequency of
the series arm. Consider ‘T’ type band elimination filter section as shown in the
Fig. 9.20(a).
Let the frequency of the series and shunt arm be ω 0 rad/sec. Then for series arm,
frequency of anti resonance is given by,
Similarly, for shunt arm, frequency of resonance is given by,
From equations (1) and (2), for same resonant frequencies of series and shunt arm
resonant circuit we can write,
Total series arm impedance is given by,

Similarly, total shunt arm impedance Z2 is given by,
Hence Z1Z2 = L2/C1 = L2/C2 which is real and constant. Hence above sections are
constant k sections. So we can write,
Reactance Curves and Expressions for Cut-off Frequencies:
To verify the band elimination characteristics, let Z1 =j X1 and Z2 =j X2. The

reactance curves of X1 and X1/4 + X2 against frequency are as shown in the Fig.
9.21.
From the above characteristics it is clear that the reactance curves for X1 and (X1/4
+ X2) are on the same sides of the frequency axis between f1 and f2 which indicates
stop band. These curves are on opposite sides of the axis below f1 and above
f2 which indicates two pass band. Hence for the given section, the characteristics
are of band elimination filter where f1 and f2 are the cut-off frequencies.
In Band Stop Filter, the condition for cut-off frequencies is given by
But from the condition of constant K filter section, Z1Z2 = R20
From above equation it is clear that the value of the series arm impedance Z 1 can be
obtained at two different cut-off frequencies namely f 1 and f2. So at f = f1, Z1 = + j(2
R0) and at f = f2, Z1 = -j(2 R0). Thus impedance Z1 at f1, i.e. at lower cut-off
frequency, is negative of the impedance Z1 at f2 i.e. upper cut-off frequency. Hence
we can write,
But from equation (1), frequency of resonance is given by
Substituting value of (L1 C1 ) in above equation, we can write,

Hence, above equation (9) indicates that in band elimination filter, the frequency of
resonance of the individual arms is the geometric mean of two cut-off frequencies.
Variation of Z0T and Z0π, Attenuation Constant (α), Phase Constant (β) with Frequency:
The variations of Z0T and Z0π, attenuation constant (α) and phase shift (β) are as
shown in the Fig. 9.22 (a), (b) and (c) respectively. Consider that f 1 and f2 are two
cut-off frequencies and R0 is the design impedance of the Band Stop Filter.
Design Equations:
Consider that a band elimination filter with two cut-off frequencies f1 and f2 is
terminated in design impedance R0. Then, from equation (7), at lower cut-off
frequency f1, we can write,
For band elimination filter constant K section, frequency of resonance in series

arms is given by,
Substituting value of L1 in above equation,

Substituting value of L1 from equation (10),
Similarly from equation (6) we can write,
Substituting value of C1 from equation (11),
Equations (10) to (13) are called design equations of prototype band elimination
filter sections.
m Derived Filters:
m Derived Filters – The first disadvantage of prototype filter sections can be
overcome by connecting two or more prototype sections of same type (either all T
type or all π type) in cascade. In such a cascade connection, attenuation to the
frequencies in pass band remains zero ideally, but attenuation to the frequencies in
entire attenuation band considerably increases. e.g. If two sections of same type are
cascaded, the attenuation in the attenuation band gets doubled giving much sharper
cut-off characteristics than that obtained by using only a single section.
But due to the resistance in the components used in cascade connection, the
attenuation in pass band slightly increases, instead of being zero. Thus the curve
becomes rounded off at the cut off frequency in pass band as shown in the Fig.
9.24.
To fulfill all the requirements discussed above, it is necessary to design a new

section having same cut-off frequency as that of the prototype section but different
attenuation characteristics in the attenuation band. Also to maintain same cut-off
frequency, both the sections must have same characteristic impedance Z 0. It is
possible to derive a new section from a prototype constant K section. Thus, a new
section derived is called m-derived section.
Derivation of m-derived Sections:
Consider any general T section and a new section derived from it as shown in the
Fig. 9.25 (a) and (b).
In the m-derived section, the series arm is of same type as that in prototype section
but having different value i.e. m Z 1/2 where m is a constant. Now Z 2 of prototype
section will change to Z′2 in m-derived section such that the value of Z 0 for both the
sections is same.
For a prototype section, the characteristic impedance is given by,
Similarly for a m-derived section, the characteristic impedance is given by,
To maintain same Z0, equating equations (1) and (2) by squaring,
From equation (3) it is dear that, the shunt arm of m-derived section is a series
connection of two impedances (Z2/m) and (1-m2/4m) Z1 if 0<m<1 as shown in the
Fig. 9.26.
The same technique can be used to obtain m-derived π section. Consider any
general π section with shunt arm of same type but having different value i.e. Z2/m
as shown in the Fig. 9.27.
For a prototype π section, the characteristic impedance Z0π is given by,
Similarly for a m-derived π section, the characteristic impedance is given by,

Squaring equations (4) and (5) and equating,
Multiplying numerator and denominator by the factor m/1-m2 we get

From equation (6) it is clear that, the series arm Z′ 1 of m-derived section is a
parallel combination of two impedances (m Z1) and (4m/1-m2) Z2 with condition 0<
m< 1. The m-derived π section is as shown in the Fig. 9.28.
In m-derived filters (either T or π), attenuation characteristic can be improved in

stop band by using series resonant circuit formed in the shunt arm of m-derived T
section and anti resonant circuit formed in the series arm of m-derived π section.
We can select a frequency of resonance of these resonant circuits such that

attenuation increases very rapidly at that frequency upto ∞. Consider m-derived
low pass filter. Here if frequency of resonance is selected just above
the cut-off frequency of the low pass filter, f c, the shunt arm in T section acts as a
short circuit at this frequency being series resonant circuit. It short circuits the
transmission path. Hence output becomes zero and attenuation increases to ∞. In
π section, at frequency which is just above f c, series arm becomes open circuit
being anti resonant circuit. Hence output becomes zero and attenuation increases to
∞. The frequency at which the attenuation increases to ∞ is called frequency of
infinite attenuation and it is denoted by f ∞. In in-derived high pass filter section, the
frequency of infinite attenuation is selected just below f c, the cut-off frequency of
prototype high pass filter section.
Consider the m-derived T and π sections as shown in the Fig. 9.28 and Fig. 9.30. If
we put m=1, both sections reduce to corresponding prototype sections. As m is a
constant, for different values of m, we can design infinite m-derived sections for
same fc and z0 specifications. The shunt arm of T section is a series combination of
impedances interms of Z1 and Z2 ; while series arm of π section is a parallel
combination of impedances in terms of Z 1 and Z2. But we know that Z 1 and Z2 must
be of different types. So to maintain this relationship in the shunt arm of T section
and the series arm of π section, factor 1-m2/4m must be positive which gives m
always positive. As we have seen earlier, for m=1, we get original prototype filter
sections. Thus the value of m must be always selected as
m Derived Low Pass Filter:

The m Derived Low Pass Filter T and π sections are as shown in the Fig. 9.29 (a)
and (b) respectively.
Consider that the shunt arm of T section resonates at the frequency of infinite
attenuation i.e. f∞, which is selected just above cut-off frequency fc. The frequency
of resonance is given by,
But for low pass filter, cut-off frequency is given by,

Substituting above value in equation for f∞,
Simplifying equation (1),
Thus, equation (2) clearly indicates that if the cut-off frequency and frequency of
infinite attenuation are specified the value of m can be easily evaluated.
Variations of Characteristic Impedance (Z0) , Attenuation Constant (α) and Phase Constant
(β) with Frequency:
The variations of characteristic impedance (Z 0), attenuation constant (α) and phase
constant (β) with frequency are as shown in the Fig. 9.30 (a), (b) and (c)
respectively.
As the characteristic impedance Z0 for prototype filter and m-derived section is
same, the variation of Z0 in m Derived Low Pass Filter section is similar to that in
prototype filter section.
The variation of attenuation constant over the attenuation band depends on types of
the reactances. The attenuation constant a in attenuation band is given by,
The phase shift β in pass band is given by,
As studied earlier, in stop band phase shift is always π c. But in m Derived Low
Pass Filter, above f∞, the phase shift drops to 0 value as shunt arm becomes
inductive above resonant frequency.
m Derived High Pass Filter:

The m Derived High Pass Filter T and π sections are as shown in the Fig. 9.32 (a)
and (b).
Consider that the shunt arm of the T section resonates at a frequency of infinite
attenuation i.e. f∞ which is selected just below cut-off frequency fc. The frequency
of resonance is given by,
But for high pass filter, cut-off frequency is given by
Substituting value of fc in the equation for f∞,
Simplifying equation (1) further,

Variations of Characteristic Impedance (Z0), Attenuation Constant (α) and Phase Shift (β)
with Frequency:
shift (β) with frequency are as shown in the Fig. 9.33 (a), (b) and (c) respectively.
As the characteristic impedance for both the sections i.e. prototype section and m
Derived High Pass Filter section, the variation of Z 0 in m-derived section is similar
to that in the prototype filter section.
The attenuation constant α in the attenuation band is given by,
The phase shift β in the pass band is given by
m Derived Band Pass Filter:

We can obtain m Derived Band Pass Filter if the prototype band pass filter is
simplified according to the network in the Fig. 9.35 which has been used to obtain
m-derived low pass and high pass sections.
The T section in each case will have a shunt impedance, Z 2/m + (1-m2/4m) Z1 where
Z1 and Z2 are the impedances of the prototype sections. We can get m-derived band
pass section if we substitute values of Z 1 and Z2 of prototype band pass filter
section in equation,
The m Derived Band Pass Filter section is as shown in the Fig. 9.35.
If we substitute Z1 and Z2 in equation (1), we can get two values of frequencies.

These frequencies are frequencies of infinite attenuation namely f1∞ and f2∞. If the
frequency of resonance is f0, then the relationship between cut-off frequencies, and
frequencies of infinite attenuation is given by,
Also we can write

Variations of Characteristic Impedance (Z0) Attenuation Constant (α) and Phase Shift (β)
with Frequency:
shift (β) with frequency are as shown in the Fig. 9.36 (a), (b) and (c).
m Derived Band Stop Filter:
The m Derived Band Stop Filter can be derived from the prototype band
elimination filter section in the exactly same way as the m-derived band pass filter.
The m Derived Band Stop Filter section is as shown in the Fig. 9.37.
The relationship between frequency of infinite attenuation (f 1∞ , f2∞) and cut-off

frequencies (f1, f2) is given by,
If the frequency of resonance is f0 then it is the geometric mean of two cut-off

frequencies as well as of two frequencies of infinite attenuation.

Variations of Characteristic Impedance (Z0) , Attenuation Constant (α) and Phase Constant
(β) with Frequency:
The variations of characteristic impedance Z 0, attenuation constant α and phase

shift β with frequency is as shown in the Fig. 9.38 (a), (b) and (c)
Impedance Matching using Half Sections:
While connecting number of different sections in the filter, it is very important to
match the impedances of the sections at the junction points. Thus a T section
should not be connected to a π section directly as both the sections have different
impedances. Hence it is necessary to use matching sections to join T and π sections
together. Such impedance matching section is half section. The half section is
derived from either symmetrical T section or symmetrical π section. In
symmetrical network, the impedance at both the ports is same. But half section is
asymmetrical network and it has two different impedances at the two ports, termed
as image impedances. The prototype half section is as shown in the Fig. 9.39.
Above prototype half section can be obtained from symmetrical T and π sections
by dividing these sections centrally into two half sections as shown in the Fig. 9.40
(a) and (b).
Terminating m-derived Half Sections:
The m-derived half sections can be obtained by splitting m-derived T or π section

centrally. When m-derived T section is splitted centrally, the image impedances of
the resulting m-derived half sections are Z0T and Z0πm as shown in the Fig. 9.41.
Similarly when m-derived π section is splitted centrally, the image impedances of

the resulting m-derived half sections are Z0π and Z0Tm as shown in the Fig. 9.42.
From above half sections, it is clear that the image impedances Z 0Tm and Z0πm depend
on the value of m. So for different values of m, we can get infinite number of half
sections. But it is observed that for m=0.6, the characteristic impedance remains
almost constant over entire pass band at value R 0. This enables filter to be
terminated properly in a pure resistance equal to design impedance R 0. The
variations of Z0Tm and Z0πm with frequency over a pass band in low pass filter section
are as shown in the Fig. 9.43.
Composite Filters in Network Analysis:

Composite Filters in Network Analysis – In prototype filter sections, the
attenuation characteristic is not very sharp in the attenuation band as it is expected.
This drawback can be overcome by using m-derived filter sections which are
derived from respective prototype filter sections. But it is observed that in the stop
band attenuation drastically reduces after f ∞ in low pass section and before f ∞ in
high pass section. This drawback of m-derived filter can be overcome by
connecting number of sections including prototype sections and m-derived sections
with terminating half sections. Such a combination of different sections is
called composite filter.
The comparison of attenuation constant variations in low pass prototype section,

m-derived section and composite section is as shown in the Fig. 9.44.
In Composite Filters in Network Analysis, cut off frequency and the design
impedance are the two important design specifications. The number of various
sections in the composite filter totally depends on the attenuation characteristics
required. If it is required that the attenuation should rise sharply in the attenuation
band, we must select at least one m-derived section with low value of m.
In general, for lower values of m, attenuation at cut off rises rapidly. The typical
value of m for such attenuation at cut off is m= 0.3 to 0.35. If it is required to
maintain this attenuation at a high value in attenuation band, we must connect
either a prototype section or another m-derived section with comparatively larger
value of m.
If required both the sections can be used in the composite filter. To have proper
impedance matching and constant characteristic impedance throughout pass band,
we must connect the terminating sections with m = 0.6.
The general block schematic of the composite filter is as shown in the Fig. 9.45.
Normalized Low Pass Filter Characteristics:

Normalized Low Pass Filter Characteristics – The passive filters are the filters
consisting only passive components such as resistors, inductors, capacitors. These
classical filters are designed starting with filter networks such as T type, Π type
etc. Low Pass Filter Characteristics is obtained by connecting inductor in the series
arm and capacitor in the shunt arm of the network, while high pass filter is
obtained by connecting capacitor in the series arm and inductor in the shunt arm of
the network. These filters are designed for given cut-off frequency and
characteristic impedance.
There are few disadvantages of these filters as follows.
1. Attenuation does not rise sharply with frequency in the stop band.
2. In pass band, the characteristic impedance Z0 varies with frequency so
proper termination of filter is not possible.
In comparison with classical filter design, modern filter design approach is exactly
other-way. In modern filter design we start with required characteristics and then
try to synthesize a network giving desired characteristics. Let us study modern
filter design approach based on approximation theory.
Low Pass Filter (LPF) can be converted in any other type of filter. So let us study
low pass filter characteristics first. Consider ideal LPF characteristics as shown in
Fig. 9.46.
The response shown in Fig. 9.46 is normalized response in which we assume ω c =
1 rad/sec and response in pass band |M(jω)| = 1. Then by using proper scale factors
for magnitude and frequency, we can get desired filter characteristics.
But this ideal response is not possible with any of the practical filter network.
Hence it is desired to have a characteristics which approximates the ideal
characteristics shown in Fig. 9.46. Many scientists have suggested many
approximations to the ideal characteristics.
Butterworth Approximation:
Butterworth Approximation – In low-pass filter design, we have to assume that all
transmission zeros of the system function are at infinity. Then the magnitude
function in general form can be written as,
Where k is called as dc gain constant as it is the magnitude at ω = 0. f (ω 2) is the

polynomial which is selected to give the desired amplitude response.
Butterworth suggested that, let
where n is the order of filter.
Then amplitude function can be written as

Above approximation is called as Butterworth approximation. It is also known
as maximally flat response.
To get normalized Butterworth response assuming k = 1. The magnitude function

M(jω) decreases monotonically with ω.
For any value of n i.e. for any order of the filter, at ω = 0.
Similarly, for any value of n i.e. for any order of the filter, at ω = 1, 12n = 1.
Thus at ω = 1, we get normalized cut-off frequency for any order of the filter. As
the order of the filter increases approximate characteristics approaches ideal one.
This indicates that for higher n, approximation is better. Thus order of the filter, i.e.
n controls the closeness of approximation to the ideal response in pass band as well
as in stop band. The Butterworth Approximation responses for different orders of
filters are as shown in Fig.9.47.
From binomial series expansion, we can write,
When above expression is differentiated with respect to ω, at ω = 0, it is observed

that first (2n — 1) derivatives are zero. The derivative of system function is zero
means the slope of graph or curve is zero. Thus graph is horizontal or flat, hence
approximation is called maximally flat response.
Consider amplitude function given by,
In pass band of filter, f (ω2) should

be very small i.e. very less than 1. Then we can neglect f (ω 2) and the amplitude
function can be written as,
In stop band, f (ω2) should be very high i.e. much greater than 1. Then
the amplitude function decreases as ω increases. Thus, in stop band, as frequency
increases, response falls down giving low pass filter characteristics.
Let us consider attenuation in stop band. The magnitude response is given by
In stop band, ω2n is very large as compared to 1, hence the magnitude response is
given by
The attenuation or loss expressed in dB is given by
Chebyshev Approximation:
Chebyshev Approximation – In the earlier section, we have studied that the
Butterworth approximation is the best at ω = 0. But as we move towards cut-off
frequency, ωc = 1, approximation becomes poorer. It departs from ideal
characteristics.
Let us consider an approximation which “ripples” about the normalized magnitude,

unity, in the pass band and magnitude decreases very rapidly beyond cut-off
frequency in stop band. This Chebyshev approximation is equally good at ω = 0
and ω = 1. So the approximation is also called equal ripple approximation.
Equal ripple property is obtained by using Chebyshev cosine polynomials defined

as,
Let us find Chebyshev cosine polynomials for different values of n.
Higher order Chebyshev polynomials are obtained by using recursive formula,
Thus for n = 2, we get C2(ω)
Substituting values of C1(ω) and C0(ω) from equation (4) and (3) respectively, we
get
For n = 3, we get C3(ω) as
Substituting values of C1(ω) and C2(ω) from equations (4) and (6) respectively, we
get,
The Chebyshev polynomials for different values of n are as given in the Table
9.11.
The behaviour of Chebyshev polynomials with the variation of ω is as shown in
Fig. 9.48.
Some important properties of Chebyshev polynomials useful in low-pass filter
approximation are as follows :
1. From the plots of Chebyshev polynomials, it is clear that the zeros of the
polynomials are located in the interval ω = – 1 to ω = 1, that is, |ω| ≤ 1.
2. Within the interval from ω = -1 to ω = 1, the magnitude of Chebyshev
polynomial is always less than or equal to unity, that is, |Cn(ω)|≤1 for |ω|≤1.
3. Outside interval lei |ω|≤1, magnitude of Chebyshev polynomial rapidly
increases with increase in ω.
After the summary of few properties of Chebyshev polynomials, let us study how
to use Chebyshev polynomials in low-pass filter approximation. Consider the
function ε2 C2n(ω) where ε is the real number which is very small compared to
unity. As seen from above properties ε2 C2n(ω) will vary between 0 and ε2 is the
interval |ω|≤1 as C1(ω) = 1. Let us add 1 to this function making it 1 + ε 2 C2n(ω).
This new function will vary between 1 and 1+ε 2 for |ω|≤1 where (1 + ε2) is slightly
greater than 1. Inverting this new function and relating it to the square of the
magnitude we get
In the interval fro |ω|≤1 i.e. in the pass band, the square of the magnitude will
oscillate or ripple about a unity such that maximum value is 1 while minimum is
1/(1+ε2) Outside interval |ω|≤1 i.e. stop band, C 2n(ω) becomes very large such that |
M(jω)|2 approaches zero very rapidly with increase in co. Hence it is most suitable
for ideal low pass filter characteristics.
The Chebyshev approximation to the ideal low-pass filter is as shown in Fig. 6.49.
Within the pass band 0 ≤ ω ≤ 1, the magnitude M (jω) oscillates or ripples between
1 and √1+ε2. The distance between maximum and minimum in the pass band is
called ripple height and it is given by,
In the stop band, that is, for |ω|≥1, as ω increases, ε 2C2n(ω) becomes very large as
compared to 1 then
In the stop band, the loss or attenuation expressed in dB is given by
But for large ω, Cn(ω) can be expressed as,
Substituting value of Cn(ω) from equation (12) in the expression for loss,
From above expression it is observed that Chebyshev response falls of at the rate
of 20 n dB/decade after initial loss of [20 log10 ε + 6 (n —1)] dB.
In most of the applications ε is very small, which gives 20 log ε term negative.
Thus it is necessary to compensate this decrease in the loss in the stop band. To
achieve this, a filter of sufficiently higher order is selected.
Basically Chebyshev approximation depends upon two variables namely ε and n.

These two values can be determined directly from the specifications given for the
filter. The value of ε is determined from maximum permissible ripple. Once the
value of ε is fixed, value of n can be found from the required attenuation in stop
band.
Attenuators in Network Analysis:

In various transmission equipments, it is many times required to supress or reduce
the levels of the currents and voltages at certain points. To fulfill the need of
attenuation, a four terminal resistive network called attenuator is used. Attenuators
in Network Analysis are designed to provide a known amount of attenuation
between input and output terminals without changing the matching of the
impedance at a given value. Attenuators are resistive networks, so all frequencies
are attenuated by same degree of amount preventing attenuation distortion. As all
the components are resistive in the attenuator networks, no phase shift will be
introduced by such networks. Hence the phase constant (β) will be zero and the
propagation constant (γ) will be equal to only the attenuation constant (α).
Attenuators in Network Analysis are either symmetrical or asymmetrical networks.

Attenuators are either of fixed value or adjustable value type. Generally fixed
attenuators providing constant attenuation are called pads. The variable attenuators
are generally used in radio broadcasting stations as volume controls. Attenuation is
usually expressed in neper or decibel.
Power Ratios, Voltage Ratios and Current Ratios:
In line communication, when AC power from one point (generally sending end) to
another point (generally receiving end) is considered, various elements in the
communication system introduce gains or losses of power at various points.
Consider a four terminal network between a generator and a load as shown in the
Fig. 10.1.
Let input and output powers be P 1 and P2 respectively. The network introduced in
between generator and load may provide loss or gain in power. Let the ratio of
input power to output power i.e. P1/P2 be M.
If M is greater than unity, then the network introduces loss. If M is less than unity,
then the network introduces gain.
If ‘n’ number of such networks are connected in cascade or tandem, then the
overall power ratio of cascade connection can be obtained by multiplying
individual power ratios of networks.
Let M1 to Mn-1 be individual input power to output power ratios for (n —1) number
of networks respectively connected in cascade. Then overall power ratio M can be
written as,
From above expression it is clear that in the complex systems, calculation of

overall power ratio is very tedious. To simplify this calculation the individual
power ratios are expressed in logarithmic scale. This enables addition of power
ratios in the logarithmic scale instead of multiplication. The logarithmic unit
employed is “BELL”.
Hence, we can express power ratio in bell as follows,
But in practice, it is observed that the unit bell is too large. Hence instead of bell a
smaller unit called as decibel is introduced, where,
Thus, if ratio (P1/P2) is greater than unity, then D will be positive which indicates
power loss. Similarly if ratio (P1/P2) less than unity, then D will be negative which
indicates power gain.
The power ratio can be expressed in decibel as follows,

Thus using this conversion we can express any power (either input or output) in
watt.
Consider again the network shown in the Fig. 10.1. Assume that two equal
resistors of value R are connected at generator and load side. Let the current and
voltage at source side be I1 and E1 and current and voltage at load side be I2 and E2.
The power developed across resistors at source and load side can be written as,
Then input to output power ratio can be given by
Expressing all the ratios in decibel as follows,
Expression of Attenuation in Neper and Decibel:
Attenuation is defined as loss of power in a transmission line or an electrical

network. Attenuation is expressed either in neper (N) or decibel (dB) notations.
Usually power ratios are expressed in decibel. But it is possible to express voltage
and current ratios in decibel if the resistive components of generator and load
impedances are equal.
Similarly neper is usually used to express current ratios. But it is possible to

express power ratio in neper if the resistive components of generator and load
impedances are equal.
Consider a four terminal network as shown in the Fig. 10.2 between generator and
load.
If the power entering a network is P in and that leaving the network is P out, then the
attenuation in decibel is defined as,
If the current entering a network is Iin and that leaving the network is I out, then the
attenuation in neper is defined as,
Under the condition of equal resistive components at generator and load side, we
can write,
Similarly,
If the resistive components of the impedances at input (i.e. generator) and output
(i.e. load) of the network are equal, we can convert attenuation in any notation as
follows
Attenuator Network:
An attenuator network must fulfil following conditions.
 It must give correct input impedance,

 It must give correct output impedance and
 It must provide specified attenuation.
In general, attenuation is expressed in decibel as follows,
where D is the attenuation in decibel.
But we can express attenuation in neper as follows,
where N is the attenuation in neper.
In this topic, we shall study symmetrical attenuators such as symmetrical T type,

symmetrical π type, lattice type and bridged T type, alongwith asymmetrical
attenuator such as ‘L’ type attenuator.
Any Attenuator Network is designed for specified characteristic resistance R 0 and

attenuation.
Let us find design equations for various Attenuator Network one by one.
Symmetrical T Type Attenuator:
Consider properly terminated symmetrical T network as shown in the Fig. 10.3.
According to current divider rule,

But for symmetrical networks,
For properly terminated network, input impedance Rin is given by,
From equation (2),
From equation (2), we can write,

Equations (A) and (B) are called design equations of symmetrical T attenuator.
Symmetrical π Type Attenuator:
Consider properly terminated symmetrical π network as shown in the Fig. 10.5.
Let the
characteristic impedance be pure resistive i.e. R 0 and propagation constant γ = α.
Then according to the theory of symmetrical network, the shunt arm and series arm
can be expressed interms of R0 and α as follows,

Multiplying numerator and denominator by factor e α/2 on right handside of the

equation,
Equations (A) and (B) are called design equations of symmetrical π attenuator.
Symmetrical Lattice Type Attenuator:

Consider properly terminated lattice Attenuator Network as shown in the Fig. 10.7
(a).
According to the theory of the symmetrical networks, characteristic impedance is

the geometric mean of open and short circuit impedance.
Consider lattice network shown in the Fig. 10.8 (a) and (b) with open and short
circuit output terminals respectively.
By definition, characteristic impedance is given by,

Consider Fig. 10.7, applying KVL to closed path A-1-2-2′-1′-2-A, we can write,
Thus, we can write,
Applying KVL to closed path A-1-2′-2–1′-B-A, we can write,
Equations (A) and (B) are called design equations of symmetrical lattice attenuator.
Bridged T Attenuator:
Consider properly terminated bridged T network as shown in the Fig. 10.10.

Assuming 3 loop currents in clockwise direction as shown.
Consider closed path 1-A-B-1′-1, applying KVL,
Consider closed path A-2-2′-B-A, applying KVL,
Consider closed path C-D-2-A-1-C, applying KVL,
Adding equations (1) and (2),

Substituting value of I3 in
Substituting value of RB from equation (A) we can write,
Equations (A) and (B) are called design equations of bridged T attenuator.
L Type Asymmetrical Attenuator:
An asymmetrical L type attenuator is as shown in the Fig. 10.12.
Input resistance looking into network from terminals 1-1′ is
Putting value of R2 from equation (A),

equations (A) and (B) are called design equations of asymmetrical L type
attenuator.

Analog Passive Filters Theory

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Analog Passive Filters Theory

Uploaded by

Copyright:

Available Formats

Terminal Pairs or Ports:

Terminal Pairs or Ports – A pair of terminals at which an electrical signal may

We can obtain these equations by considering two variables as dependent variables

In equation form, above relations can be written as,

In matrix form, above equations can be written as,

The individual z-parameters are defined as follows.

Hence we can write,

 Z11 = circuit driving point input impedance

y Parameters of Two Port Network:

In equation form, above relations can be written as,

The individual y-parameters are defined as follows.

This is called driving point input admittance.

This is called forward transfer admittance.

This is called reverse transfer admittance

This is called output admittance.

The equivalent circuit of a y Parameters of Two Port Network interms of short

In equation form, above relations can be written as,

In matrix form the above equations can be written as,

This is called short circuit input impedance.

This is called forward short circuit current gain.

This is called reverse open circuit voltage gain.

This is called open circuit output admittance.

The equivalent circuit of a two port network interms of hybrid parameters

In equation form, above relations can be written as,

In matrix form above equations can be written

This is called open circuit reverse voltage gain.

This is called open circuit transfer admittance

This is called short circuit transfer impedance

This is called short circuit reverse current gain.

ABCD or Transmission parameters are also called chain parameters.

Condition for Symmetry in Two Port Network:

The basic equations of z-parameters are as follows.

Consider Fig. 6.6 (a).

Hence equation (i) will become,

Consider Fig. 6.6 (b).

Equation (ii) can be written as,

The basic equations of y-parameters are as follows

Consider Fig. 6.6 (a).

Equation (i) can be written as,

Equation (ii) can be written as,

Substituting value of V2 in equation (iii), we have

Consider Fig. 6.6 (b).

Equation (ii) can be written as,

Substituting value of V1 in equation (viii) from equation (vii), we have

Conditions of Symmetry for h Parameters:

The basic equations of h-parameters are as follows,

Equation (i) can be written as,

Equation (ii) can be written as

Substituting value of V2 in equation (iii), we have

Consider Fig. 6.6 (b),

Equation (i) can be written as

Equation (ii) can be written as

Condition of Symmetry for Transmission Parameters:

The basic equations for the transmission parameters are as following

Consider Fig. 6.6 (a),

Equation (i) can be written as,

Equation (ii) can be written as,

Consider Fig. 6.6 (b),

Equation (i) can be written as,

Condition for Reciprocal Two Port Network:

Condition of Reciprocity for z-Parameters:

The basic equations for z-parameters are as follows

Hence equation (i) can be written as,

Equation (ii) can be written as,