Passive Filters:

Passive filters are mainly networks using inductors, capacitors and resistors. The classical theory
employed was based on the image parameter theory which in turn was based on the filter’s
characteristics and performance. Its component values were calculated by considering a source
having a specified source resistance and feeding it into a constant load impedance called the
termination impedance, resulting in constant — K type filter or prototype filter

We have also m-derived filters

Ideally a filter should have zero attenuation in the pass band. This condition can only be satisfied
if the elements of the filter are dissipation less, which cannot be realized in practice. Filters are
designed with an assumption that the elements of the filters are purely reactive. Filters are made
of symmetrical T, or π sections. T and π sections can be considered as combinations of
unsymmetrical L sections.

Types of Network Elements

We can classify the Network elements into various types based on some parameters. Following
are the types of Network elements −

• Active Elements and Passive Elements

• Linear Elements and Non-linear Elements
• Bilateral Elements and Unilateral Elements

Active Elements and Passive Elements

We can classify the Network elements into either active or passive based on the ability of
delivering power.

• Active Elements deliver power to other elements, which are present in an electric circuit.
Sometimes, they may absorb the power like passive elements. That means active
elements have the capability of both delivering and absorbing power. Examples: Voltage
sources and current sources.
• Passive Elements can’t deliver power (energy) to other elements, however they can
absorb power. That means these elements either dissipate power in the form of heat or
store energy in the form of either magnetic field or electric field. Examples: Resistors,
Inductors, and capacitors.

Linear Elements and Non-Linear Elements

We can classify the network elements as linear or non-linear based on their characteristic to
obey the property of linearity.

1
 • Linear Elements are the elements that show a linear relationship between voltage and
current. Examples: Resistors, Inductors, and capacitors.
• Non-Linear Elements are those that do not show a linear relation between voltage and
current. Examples: Voltage sources and current sources.

Bilateral Elements and Unilateral Elements

Network elements can also be classified as either bilateral or unilateral based on the direction
of current flows through the network elements.

Bilateral Elements are the elements that allow the current in both directions and offer the same
impedance in either direction of current flow. Examples: Resistors, Inductors and capacitors.

Filters as the name suggests, they filter the frequency components. That means, they allow
certain frequency components and / or reject some other frequency components.

Types of Filters

Filters are mainly classified into four types based on the band of frequencies that are allowing
and / or the band of frequencies that are rejecting. Following are the types of filters.

• Low Pass Filter

• High Pass Filter
• Band Pass Filter
• Band Stop Filter

Low Pass Filter

e.g A low pass filter as the name suggests, it allows (passes) only low frequency components.
That means, it rejects (blocks) all other high frequency components. The study of the behaviour
of any filter requires the calculation of its propagation constant , attenuation , phase shift and its
characteristic impedance 𝑍𝑂

IMAGE IMPEDANCE 𝒁𝑶

Image impedance is basically the concept which is used in the field of the network analysis and
design and also in the field of the filter design The term image impedance applies to the
impedance seen looking into a port of a network. Usually a two-port network is implied but the
concept can be extended to networks with more than two ports
One port network is a two terminal electrical network in which, current enters through one
terminal and leaves through another terminal. Resistors, inductors and capacitors are the
examples of one port network because each one has two terminals. One port network
representation is shown in the following figure.

2
Here, the pair of terminals, 1 & 1’ represents a port. In this case, we are having only one port
since it is a one port network.
Similarly, two port network is a pair of two terminal electrical network in which, current enters
through one terminal and leaves through another terminal of each port. Two port network
representation is shown in the following figure.

Here, one pair of terminals, 1 & 1’ represents one port, which is called as port1 and the other pair
of terminals, 2 & 2’ represents another port, which is called as port2.
There are four variables V1, V2, I1 and I2 in a two port network as shown in the figure. Out of
which, we can choose two variables as independent and another two variables as dependent. So,
we will get six possible pairs of equations. These equations represent the dependent variables in
terms of independent variables. The coefficients of independent variables are called as
parameters. So, each pair of equations will give a set of four parameters

The characteristics parameters of the two port network are given below:

Image Impedance of Two-port network

It is the impedance which when connected to the input and the output of the network, it will
make both the impedances equal at the input and the output terminal. It applies to the seen
impedance which is determined by looking through the ports of the network.

The Two-port network can be properly used to describe the concept of the image impedance in
the better way.

3
two port network

The impedance zi1 – when considered from the port 1

Zi2 –image impedance when considered from the port 2

The image impedance will not be equal until the network is the symmetrical network or anti-
symmetrical with respect to the ports.

Characteristic impedance

The characteristics impedance also known as the surge impedance is usually considered in the
case of the transmission line and is represented as Z0. The characteristics impedance is defined as
the ratio of the amplitude of the voltage and the current taking the consideration of the single
wave through the line. The surge impedance is usually allocated through the transmission line
with its geometry and the material. It is to be noted that this impedance is independent of the line
length.SI unit – ohm

Iterative Impedance

It is defined as the particular value of the load impedance which has the ability to produce an
input impedance with the value as same as the value of the load impedance. In the two port
system when it is connected at the one end then it produces equal impedance when looking at
each other.

Image transfer coefficient

It is usually considered for the linear passive type of the two-port network, such network must be
terminated with the image impedance of the network. Let

V1 – voltage at the input terminal

I1 – current at the input terminal

V2 – voltage at the output terminal

I2 – current at the input terminal

4
Hence, the image transfer coefficient can be calculated as half the logarithm of the product of V 1
andI1 divided by the product of the V2 and I2 .

Represented as,

P = ½ log {(V1 I1 ) / (V2 I2 )}

Propagation constant

This constant is usually considered for the wave and is defined as change in the phase angle with
respect to the per unit change in the distance travelled by the wave. In other words we can say as
the rate of the change in the phase of wave with distance. The constant is represented as the term
K.

Insertion Loss

The insertion loss can be explained as the loss in load power because of the insertion of a
particular component or device in a transmission system. It is represented in the ratio of the
decibels of the power received at the side of the load before the insertion of the component to the
power received at the load side after the insertion of the component or the device.

𝑃1
Insertion loss = 10 log db
𝑃2

Where 𝑃1 =Power output without network insertion and 𝑃2 = Power output without network
insertion

Attenuator

An Attenuator is a special type of electrical bidirectional circuit made up of entirely resistive

elements. It is nothing but a two port resistive network designed to weaken or attenuate the
power being supplied by a source to a level that is suitable for the connected load. The amount of
the power is reduced which is being delivered the load. These have the basic applications in
radio, communication and transmission line etc.

Attenuator Connection

The network of the attenuator can be designed to produce a fixed degree of attenuation or to
give a variable amount of attenuation. The standard attenuators networks generally known as an
attenuator pad are available from 0dB to more than 100dB. The Variable and switched
attenuator are the adjustable resistor type of the networks that show a calibrated increase in
attenuation for each switched step.

LADDER NETWORK

5
A ladder structure is one of the commonest forms of filter network. A cascade connection of
several T and π sections constitutes a ladder network. A common form of the ladder network
represents a T section ladder network or a π section ladder network. It can be observed that both
networks are identical except at the ends. It can either be symmetrical or asymmetric. Two
networks having the same electrical properties i.e. characteristic impedance (Z0) and propagation
constant (γ) are called equivalent networks or identical networks

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 (Z 0) 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.

6
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 Z 0,
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 Z0. In such network, the output impedance will
7
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
eγ 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

8
Note that input current is represented by sending end current, IS; 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.

9
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.

10
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 (e θ):

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 I1/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.
11
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

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 Z1 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 (Z0) and propagation constant (γ)
in terms of the network elements.

Characteristic Impedance (Z0): of T network:

(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.

12
The input impedance of a T network is given by,

(B) In terms of open and short circuit impedances

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.

Consider Fig. 8.9 (a),

Consider Fig. 8.9 (b),

13
Multiplying equations (3) and (4), we can write,

Hence in any symmetrical network, the characteristic impedance Z0 is the geometric mean of
open and short circuit impedances measured at any pair of terminals.

Propagation Constant (γ):of a T network

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,

14
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,

15
Adding equations (3) and (4), we get,

Symmetrical π- Network in Network Analysis:

The Symmetrical π-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 Z 1 and to have a total
shunt arm impedance of Z2, 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.

Characteristic Impedance (Z0): of π - network

(A) In terms of series and shunt arm impedances

16
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,

17
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).

Consider Fig. 8.17 (a),

Consider Fig. 8.17 (b),

18
Multiplying equations (3) and (4), we can write,

Thus the characteristic impedance of symmetrical π network is given by

Propagation Constant (γ): of a π - network

Consider correctly terminated symmetrical π network as shown in the Fig. 8.19.

As the network is symmetrical, by definition,

By potential divider rule,

19
Rearranging the terms,

where γπ is the propagation constant of symmetrical π network. Putting value of Z 0π

interms of Z0T,

Series and shunt arm impedances of symmetrical π network in terms of Z0π and γπ:

Lattice and bridged network

A network that is made up of four branches connected in series to form a mesh; two nonadjacent
junction points serve as input terminals, and the remaining two junction points serve as output
terminals. The lattice networks are being widely used in the areas like grid computing, sensor
networks and in many more areas. The main points which highlights the lattice networks are

20
1)Its optimal routing policies

2)limits on the capacities of its elements

3)Its performance with the finite amount of the buffers

A bridge-T network with a fourth branch connected between an input and an output terminal and
across two branches of the network.

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 ZA 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.

Characteristic Impedance (Z0):

Consider Fig. 8.21 (b).

Consider closed path 1-2-2′-1′-1, applying KVL, we get

21
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

22
Putting value of E/Is in above equation,

(B) In terms of open and short circuit impedances

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).

Consider Fig. 8.22 (a),

Consider Fig. 8.22 (b),

Multiplying equations (6) and (7) , we can write,

23
Propagation Constant (γ):

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,

24
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.

25
.

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):

Consider Fig. 8.29 (a),

26
Finding roots, we can write,

Consider Fig. 8.29 (b),

Finding roots, we can write,

Image Impedances (Zi):

Consider Fig. 8.29 (a)

Thus, image impedance Zi1 at port 1 is given by

27
Similarly,

Thus, image impedance Zi2 at port 2 is given by

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).

28
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.

Consider Fig. 8.25 (a),

29
Finding roots of quadratic equation,

Consider Fig. 8.25 (b),

Find roots of above equation we can write,

30
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,

31
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

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 (Z0), 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.

32
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 Z0 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 Z0 is real resistive while in stop band it is purely imaginary or
reactive.

Thus, we can define cut off frequency fc 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 Z1 = 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.

33
Constant K Sections:

A T or π section in which series and shunt arm impedances Z1 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. R 0 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.

Design Equations:

Consider that the filter is terminated in design impedance R0 and the cut-off frequencies are f1
and f2.

Then from equation (7), at the lower cut-off frequency f1, we can write,

34
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,

35
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.

Low Pass Filter:

The prototype T and π low pass filter sections are as shown in the Fig. 9.3.

36
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.

37
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.

38
From above expression it is clear that, Z0T 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

39
Similarly we can write,

Hence

From equation (5), it is clear that as frequency increases from 0 to fc, 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,

40
In stop band, as frequency f increases above fc, 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 fc, β also increases from 0 to π radian. The variation of β with
frequency is as shown in the Fig. 9.7.

41
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.

Design Impedance (R0):

Total series arm impedance Z1 = -j/ωC

42
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.

43
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.

44
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,

45
Hence,

From equation (5), it is clear that as frequency f increases from f c 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 fc, attenuation decreases from ∞ to 0. The
variation of attenuation constant α with frequency is as shown in the Fig. 9.12.

46
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.

47
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 section

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).

48
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.

49
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).

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.

50
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 R0 and attenuation.

Let us find design equations for various Attenuator Network one by one.

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.

51
Consider a four terminal network between a generator and a load as shown in the Fig. 10.1.

Let input and output powers be P1 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,

52
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.

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.

53
If the power entering a network is Pin and that leaving the network is Pout, then the attenuation in
decibel is defined as,

If the current entering a network is Iin and that leaving the network is Iout, then the attenuation in
neper is defined as,

Under the condition of equal resistive components at generator and load side, we can write,

Similarly,

54
IMPORTANT EQUATIONS FOR FILTER NETWORKS DESIGNS,

1.Characteristic Impedances for T-type and π-type sections

2.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

55
3.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 section

Examples /Tutorial Questions

(1)(i) With the aid of a low pass π-filter section derive its characteristic impedance expression
(ii) Design a constant K low pass T-section filter having a cut-off frequency of 1kHz and an
impedance of 600Ω.

(2)(i) Show how a bridged T-section network can be reduced to a lattice section
(ii) A constant K T-section filter consists of two series elements each of 60mH and a shunt
capacitor of 0.1µF , Determine:-
I Its cut-off frequency
II Iterative impedance at 2kHz

3)(i) Sketch the relationship of the characteristic impedance and the frequency of the following
networks:

(I) 𝜋- filter
(II) T-filter

(ii) For the circuit of figure 3 determine from the first principles the iterative impedance of the
equivalent T-network for a frequency of 1kHz

56
 0.03H

0.1µF 0.1µF

figure 3

(4) (i) For the T-section of figure 4 define the characteristic impedance 𝑍𝑂𝑇 of the section and
show that:-

Z1
Cosh P = 1+
2Z2

𝑍1 𝑍1
2 2

Input 𝑍2 Output

Figure 4

(ii) Deduce the relationship between the impedance of a symmetrical π-network and that of the
equivalent T-network

(5)(i)With the aid of a low pass T filter section show that its characteristic impedance is given
by:
𝑍𝐼
ZOT= √{Z1Z2(1+4𝑍2)}

(ii) Design a constant K High pass T-section filter having a cut-off frequency of 2kHz and
an impedance of 800Ω.

(6)(i) Explain the following terms applied in filters:-

57
 I Image impedance

II Insertion loss

(ii) With the aid of a low pass π filter section derive its characteristic impedance expression

58