Chapter 4

Synthesis of 2-port networks

Two-port networks are network having two pair’s terminals or ports (an input port and an output
port). Ex: Transformers, Amplifiers, Transistors, communication circuits, Power transmission &
distribution lines. Two port networks act as building blocks of electrical or electronic circuits.
Often the circuit between the two ports is highly complex. The two port parameters provide a
shorthand method for analyzing the input-output properties of two ports without having to deal
directly with the highly complex circuit internal to the two port network.

These networks are linear and passive and do not contain independent sources inside. In these
networks there are four variables V1, I1 and V2, I2. Two of them are expressed in terms of the
other two, to define two port parameters.

4.1. Deriving point and transfer network functions

Deriving point functions

These functions are obtained when the two variables, response and input, are from the same port.
1. Deriving point impedance at port-1

( )
]
( )

2. Deriving point impedance at port -2

( )
]
( )

3. Deriving point admittance at port -1

( )
]
( )
4. Deriving point admittance at port-2

By M & W :ECE 3309 Page 1

 ( )
]
( )

Transfer functions

These functions are obtained when the two variables, response and input, are from different
ports.

1. Transfer impedance

( )
]
( )

( )
]
( )

2. Transfer admittance

( )
]
( )

( )
]
( )

3. Voltage gain

( )
] ( )
( )

4. Current gain

( )
] ( )
( )

4.2. Two-port network parameters

By M & W :ECE 3309 Page 2

These parameters relate the four variables of a two-port network. Two of the variables are
expressed in terms of the other two variables using two-port parameters.

A. Impedance parameters

• Voltages V1 and V2 (dependant variables) are expressed in terms of I1 and I2 (independent

variables). Impedance parameters are very useful in designing impedance matching and
power distribution system. The input and output terminal voltage can be presented as
follows:

The voltage at port-1 is superposition of voltages due to currents I1 and I2

The voltage at port-2 is superposition of voltages due to currents I1 and I2

Rewriting the above equations

[ ] [ ][ ]

[ ]

( )
]
( )

( )
]
( )

By M & W :ECE 3309 Page 3

 ( )
]
( )

( )
]
( )
Impedance parameters are obtained by open circuiting one of the two ports, and hence they are
called open-circuit parameters.

[ ]

For reciprocal networks,

 Passive networks are reciprocal

B. Admittance parameters

• Voltages V1 and V2 are independent variables

• Currents I1 and I2 are dependent variables. The input and output terminal voltage can be
presented as follows:

The current through port-1 is superposition of currents due to voltages V1 and V2

The current at port-2 is superposition of currents due to voltages V1 and V2

Rewriting the above equations

[ ] [ ][ ]

[ ]

( )
]
( )

By M & W :ECE 3309 Page 4

 ( )
]
( )

( )
]
( )

( )
]
( )

Admittance parameters are obtained by short circuiting one of the two ports, and hence they are
called short-circuit parameters.

[ ]

For reciprocal networks,

Recall that

[ ] [ ][ ]
This can be written as:

But we know that

This implies

[ ] [ ]

Where

[ ]

By M & W :ECE 3309 Page 5

Therefore,

Similarly,

C. Transmission parameters

They are used to express the sending end voltage, V1, and current, I1, in terms of receiving end
voltage, V2, and current, I2. Voltage and current at port-1 are dependant variables and those at
port-2 are independent variables.

[ ] [ ][ ] [ ] [ ]

For reciprocal network AD – BC = 1

( )
]
( )

( )
]
( )

( )
]
( )

( )
]
( )

D. Hybrid parameters
(Reading assignment)

By M & W :ECE 3309 Page 6

Example 1
Find the open-circuit impedance parameters of the following T-network

Solution:
( )
]
( )

( )
]
( )

( )
]
( )

( )
]
( )

This implies that:

Therefore, a two port network can be represented in to its equivalent T-network from its
impedance parameters.

By M & W :ECE 3309 Page 7

Example 2
Find the short-circuit admittance parameters of the following Π-network

( )
]
( )

( )
]
( )

( )
]
( )

( )
]
( )

This implies that:

Therefore, a two port network can be represented in to its equivalent Π-network from its
admittance parameters.

By M & W :ECE 3309 Page 8

Problem 1

Find the open-circuit impedance and short-circuit admittance parameters for the following
networks.

Problem 2

Find the equivalent Π-network for the following T-network.

[Hint: first find the admittance parameters from impedance parameters of the network. Then find
the equivalent Π-network.]

Answer

Problem 3
Find the open-circuit impedance parameters for the following network.

By M & W :ECE 3309 Page 9

4.3. Series, parallel and cascade interconnections

Complex large network can be modeled with interconnected two-port networks. This simplifies
the analysis /synthesis and design process. Parameters of interconnected two-port networks can
be obtained easily: depending on the type of parameters and type of connections:

• Series connected networks: using z parameters

• Parallel connected networks: using y parameters
• Cascaded networks: using transmission parameters

A. Series Connection

[ ] [ ] [ ] [ ]

[ ] [ ] ( )[ ]

B. Parallel Connection

By M & W :ECE 3309 Page 10

 [ ] [ ] [ ] [ ]

[ ] [ ] ( )[ ]

C. Cascade Connection

[ ] [ ] [ ] ( [ ]) ( [ ])

4.4. Properties of Impedance (Z) and Admittance (Y) parameters

By M & W :ECE 3309 Page 11

We have seen that a two port network can be represented by to its Π and T equivalent networks
from the admittance and impedance parameters respectively.

Consider the impedance parameters.

 Poles of Z12 are poles of Z11 and Z22 if there is no pole cancelation in the sum ZA +ZB and
ZB + ZC.
 Since synthesis specifications are usually given in the form of transfer functions AV ( =
Z12/Z11) or AI ( = -Z12/Z22) we have freedom in selecting Z12 so that such cancelation will
not occur.
 There can be poles of Z11 and Z22 due to ZA and ZC respectively that do not exist in Z12.
These poles of deriving-point functions are called private poles.

Consider a more general network.

By M & W :ECE 3309 Page 12

This shows that private poles of deriving point functions Z11 and Z22 can be removed by partial
fraction expansion that gives ZA and ZB respectively. The same is true for admittance parameters.
Therefore, the first step is to remove private poles of deriving point functions/parameters. After
private pole removal, impedance parameters, and hence admittance parameters, can be written
with the same denominator function since they have similar poles.

( )

( )

( )
The equivalent T-network elements

( ) ( ) ( )
( )

( ) ( ) ( )
( )

Since all coefficients must be positive, Z12 has the following properties.


 Therefore Z12 has to satisfy the above coefficient conditions and consequently the degree
of its numerator should be less than or equal to the minimum of degrees of numerators of
Z11 and Z22.

By M & W :ECE 3309 Page 13

In this chapter, we will see synthesis procedures of two port lossless networks (made up of
inductors and capacitors). From the above conditions, properties of LC two-port networks can be
derived as follows:

1. The deriving point parameters Z11 and Z22 (similarly Y11 and Y22) are positive real odd
functions (even/odd or odd/even). (the same properties as one-port LC network deriving
point functions)
2. Z12 has the following characteristics:
 Has simple imaginary axis poles (poles of Z12 are poles of Z11 and Z22. The same
is true for admittance parameters)
 Is an odd function due to coefficient condition! (all missing terms in numerators
of Z11 and Z22 should also be missing from numerator of Z12, and all have the
same denominator)
 degree of its numerator should not exceed the minimum of degrees of numerators
of Z11 and Z22
 zeros can be anywhere

4.5. Synthesis of Lossless two-port networks

Two-port networks are treated as black boxes which are represented by their parameter functions
that tell the relation between voltages and currents at the two ports. The network is usually
needed to have required characteristics which are expressed using network functions. The most
common network functions that are used to characterize a two-port network are transfer
functions shown below.

1. Voltage gain

( )
( ) ]
( )
2. Current gain

( )
( ) ]
( )

3. Transfer impedance

Assuming that all private poles of deriving point parameters are removed, the three impedance
parameters can be written with the same denominator functions as follows. Note that for
passive networks Z12 = Z21.

( ) ( ) ( )

By M & W :ECE 3309 Page 14

Remember that due to coefficient conditions, Deg N12 < Deg N11 and Deg N12 < Deg N22

Now the transfer functions can be written as:

( ) ( )

Therefore, AV (port-2 open circuited) and AI (port-2 short circuited) have the following
properties.

 Deg N < Deg D

 No pole at infinity
 Even rational functions of s
 No pole at zero (all terms in N12 are present in N11 and N22)
 Poles lie on the jω – axis and are finite and non-zero
 Zeros can be anywhere

The following are the two most common two-port network configurations.

1. Unterminated (port – 2 open circuited)

[ ]

[ ]

[ ]

Transfer functions are usually given in normalized form. When ZT is normalized to 1Ω resistor R
(impedance of elements reduced by a factor R. i.e Rnormalized = Ractual /R , Lnormalized = Lactual /R ,
Cnormalized = Cactual *R)

[ ]

By M & W :ECE 3309 Page 15

 ( )
( ) ( )

N11 and q(s) (or D11) are normalized to R = 1.

2. Singly terminated (passive termination at port – 2)

This is commonly used passive filter configuration where the load (R) is purely resistive. The
source can be either voltage or current; the transfer function can be AV (Voltage gain), AI
(Current gain) or ZT.

[ ]

When normalized to 1Ω resistor R.

[ ]

( )
( )
( )

By M & W :ECE 3309 Page 16

When normalized to 1Ω resistor R.

( )

[ ]

( )
( )
( )

Definition:
Transmission zeros – are frequencies at which zero output occurs for finite input. These
frequencies are zeros of the transfer functions Z12, Y12, AV and AI.

Consider the following ladder network.

By M & W :ECE 3309 Page 17

For such ladder networks, transmission zero (zero output) occurs when the series impedance
elements are open-circuited (infinite impedance) or when shunt admittance elements are short-
circuited (infinite admittance). In other words, transmission zeros occur at poles of series
impedance elements or at poles of shunt admittance elements. These poles can be realized by
continued fraction expansion (Cauer realizations). Therefore, transmission zeros can be realized
by continued fraction expansion of deriving-point impedance/admittance parameter functions.

Transmission zeros can occur at zero, infinity or finite non-zero frequencies. Realization of such
zeros gives the following networks.

Location of Network function

transmission
zero Removed pole Z11 or Z22 Y11 or Y22

At s = 0 K0/s

At infinity Ks

At finite non-zero
frequency

Two – port LC Ladder Synthesis

One of the most important classes of passive filters uses LC ladder networks. This is especially
true when networks are generated by pole removal.

In this section, we will see ladder realization of lossless two port networks by removal of poles
from impedance and admittance functions. Any time poles are removed completely from a

By M & W :ECE 3309 Page 18

driving-point function of a two port network, transmission zeros at the locations of the poles are
created in the transfer function of the two port network. Since two port network specifications
are given in the form of transfer functions, zeros of the given transfer function (transmission
zeros) have to be realized by removing poles at frequencies where transmission zeros occur from
the deriving – point functions.

When transmission zeros are at the boundaries (s = 0 and s = ∞) only, the two port network can
be realized by Continued Fraction Expansion since Cauer realizations remove poles at zero and
infinity. However, if there are transmission zeros at finite non-zero frequencies, zeros should be
created in the deriving – point functions before they are removed as poles; this is accomplished
by zero shifting or partial removal of poles (we will discuss this technique later)

I. Transmission zeros at S = 0 or S = ∞

Transmission zeros at zero and at infinity can be realized using Cauer-2 and Cauer-1 forms
respectively.

Example 3
Realize the following LC two port network which satisfies the given transfer function.

Solution:

N11 and D11 are numerator and denominator of Z11 (deriving point function of LC network)
respectively.

( ) ( )

By M & W :ECE 3309 Page 19

Assuming no pole cancelation, we select D11 and N11 as follows:

Hence,

Where

Now, the network is realized by removing poles at infinity (Cauer -1) from Z11.

Note that impedance values are normalized.

The scale factor k cannot be controlled by Cauer realization (only transmission zeros are
realized). Transfer function of realized network is just proportional to the required transfer
function.

The quality factor km is defined as:

k of realized network can be found by computing the value of ZT at some arbitrary value of s.
s=0, ∞ or 1 are commonly used. ZT should not have pole or zero at that frequency.

By M & W :ECE 3309 Page 20

For example, in the above network, we can take s = 0 (capacitors are open-circuited and
inductors are short-circuited).

Example 4
Realize the following LC two port network which satisfies the given transfer function.

Solution:

( )

By M & W :ECE 3309 Page 21

Where

Hence apply Cauer-2 twice to Y22.

Problem 4
Realize the following network from the given transfer impedance function, and find the value of
K.

By M & W :ECE 3309 Page 22

 II. Transmission zeros at finite non-zero frequencies

Recall that two port network transfer functions are synthesized by removal of poles of deriving
point functions where transmission zeros occur. When all transmission zeros are at the
boundaries ( at zero and at infinity), realization becomes easy for lossless networks since
deriving point functions of LC networks have either poles or zeros at the boundaries. In the later
case, the reciprocal will have a pole at that frequency. Therefore, two port lossless network
transfer functions with all transmission zeros at the boundaries can be completely realized by
continued fraction expansion (Cauer realizations). However, if there are finite non-zero
transmission zeros, poles at these frequencies have to be removed from the deriving point
functions even though the deriving point function may not have poles or zeros at these
frequencies. In this case, poles have to be created in the deriving point function before they are
removed to realize the transmission zeros.

Realization of finite non-zero transmission zeros is a two step process. First, a pole is partially
removed from the deriving point function. Since the pole is not completely removed, the
remainder function will have the same poles, and since poles and zeros (of LC deriving point
function) are interlaced, the remainder will also have equal number of zeros. However, the
locations zeros are affected or more precisely they are shifted. The amount by which zeros are
shifted can be tuned by varying fraction of the pole removed. Hence, fraction of the pole
removed can be adjusted so that one of the zeros is shifted to the frequency where transmission

By M & W :ECE 3309 Page 23

zero occurs. Therefore, the first step creates a deriving point function that has zero at the
transmission zero. The second step will be removal of pole at that frequency from reciprocal of
the deriving point function. This realizes the corresponding transmission zero.

Consider the following deriving point function.

( )( )
( )
( )

Poles: ω = 0, √2
Zeros: ω = 1, 2

( ) ( ) ( ) ( )

( ) ( ) ( )

A. Partial removal of pole at s = 0

( )

( )

By M & W :ECE 3309 Page 24

Zeros are shifted towards the affected pole (pole at the origin)

B. Partial removal of pole at infinity

( )

( )

Zeros are shifted towards the affected pole (pole infinity)

C. Partial removal of pole at s = + j√2

( )

( )

By M & W :ECE 3309 Page 25

Zeros are shifted towards the affected pole

Conclusion:

1. Partial removal of pole shifts all zeros (except zeros at the origin and at infinity) towards
the affected pole.
2. Since the pole is not completely removed, zeros cannot be shifted beyond the poles
adjacent to them.
3. The larger the fraction of a pole removed, the more the zeros are shifted.
4. By selecting appropriate fraction of a pole, a zero can be shifted towards the location of a
finite non-zero transmission zero. This is an important step to realize such kind of
transmission zeros as the reciprocal function will have a pole at that frequency.

Example 5
Design a reactance ladder for the following circuit.

Solution:

Transmission zeros:
S = + j, ∞

By M & W :ECE 3309 Page 26

 ( ) ( )

To realize the transmission zero at s = j, we have create a zero at this frequency on the
admittance function Y22 by partial removal of zeros. Y22 has two zeros one of which lies at
infinity. This zero cannot be shifted (boundary zero). Therefore, we need to shift the zero at ω
=√1/2. Since the zero has to be shifted to the right, we need to partially remove a pole that lies to
the left of the zero we are shifting. The only pole on the left side of the zero is the pole at ω = 2.

K is selected so that Y2 has a zero at s = j or s2 = -1 (transmission zero).

( ) |

By M & W :ECE 3309 Page 27

Removal of the pole at s =+ j realizes the first transmission zero (at s = + j)

( ) |

The only remaining transmission zero is the zero at infinity which can be realized from Z3 as an
inductor connected in series.

Note that Z3 has a pole at infinity. If the transmission zero were at the origin, it would be realized
by removing pole at s = 0 which would be obtained from the reciprocal of Z3 (Y3 = 1/Z3 = 1/s).
In this case, the inductor would be connected in parallel (the pole is removed from the
admittance function).

Problem 5
Realize the following network from the given transfer impedance function, and find the value of
K.

By M & W :ECE 3309 Page 28

 ( )

By M & W :ECE 3309 Page 29

Chapter 5
Active networks

5.1. Active network components

Linear, lumped, and time-invariant active network components can be classified in to two
categories: basic building blocks and secondary building blocks. This classification is based on
the observation that every element in the secondary building block category can be realized by
interconnecting elements in the basic building block category. Secondary building blocks are not
fundamental; they are useful to model some network components.

 Basic building blocks

This category includes passive elements such as resistors, capacitors, inductors, and active
elements such as operational amplifiers (OpAmps).

 Secondary building blocks

o Negative impedance converter (NIC)

It is a two-port device that has input impedance equal to the negative of impedance of the load
scaled by a constant 1/k.

Zin = (-1/k)ZL

o Generalized impedance converter (GIC)

By M & W :ECE 3309 Page 30

 It is a two-port device capable of making the input impedance of one of its two ports the
product of impedance of the load and some internal impedance.

V1/I1 = ZIN = q(s)Zout = V2/-I2

( )
( )

( )
( )
For example, if Z2 = R2, Z3 = R3, Z4 = R4 (resistors), Z5 = 1/Cs (Capacitor) and the termination
load Z6 = R6

Inductors can be realized from resistors. This is very important in filters to substitute inductors
with active element especially in extreme frequency applications since they require physically
large inductors which will be difficult to fabricate integrated circuits.
o Frequency dependant negative resistance (FDNR)

By M & W :ECE 3309 Page 31

An FDNR is a 1-port device with an impedance 1/(s2D), where D is a positive constant and has
the unit Farad squared, F2. Under sinusoidal stead-state operations, the impedance Z(s) of an
FDNR becomes:

⁄
( )

It is equivalent to a resistor whose negative resistance depends on the operating frequency.

A circuit realization of an FDNR can be obtained by terminating port 1 of the GIC with a
capacitor as shown below.

The deriving-point impedance of the resulting 1-port device is given by

( ⁄ )

Operational amplifiers

By M & W :ECE 3309 Page 32

 • An op amp is a high voltage gain, DC amplifier with high input impedance, low output
impedance, and differential inputs.
• Positive input at the non-inverting input produces positive output; positive input at the
inverting input produces negative output.
• Practically, opamps are not used in open-loop manner; a feedback is included to reduce
the gain and get a more precise and predictable characteristic.
• If one of the terminals is grounded in feedback configuration, the other will be virtually
grounded.

5.2. Operational amplifier circuits

The following are some of most common opamp feedback configurations.

Inverting feedback

The impedance elements Zs and ZF are impedances of a one-port network containing one or more
elements.

By M & W :ECE 3309 Page 33

 iS  iF  iin  0
iS  iF
vS  v  vout  v 

ZS ZF
vS  0 v 0
  out
ZS ZF
ZF
vout   vS
ZS
When the two impedance elements are resistors, then the circuit becomes inverting amplifier.

Non-inverting feedback

 Z 
vout  1  1 v 
 Z2 
 Z 
vout  1  1 vS
 Z2 
v Z
Gain  out  1  1
vS Z2

When the two impedance elements are resistors, then the circuit becomes non-inverting
amplifier.

Buffer (voltage follower)

Vo = V IN

By M & W :ECE 3309 Page 34

5.3. Realization of Active Networks

Transfer functions can be realized using active elements (especially operational amplifiers) and
RC elements. Inductors are not common in active networks because of their large size and since
they can be simulated using active circuits. Transfer functions of simple feedback circuits have
orders not more than 2. However, in most applications such as filters and control systems, higher
order transfer functions are required. The most common and easiest way to realize these function
is to break the given transfer function in to product of smaller 1st and 2nd ordered transfer
functions so that cascaded interconnections of sub networks obtained from these functions will
realize the given transfer function if there is no loading effect. If the transfer function is voltage
gain function, cascaded opamp feedback circuits can be used.

Operational amplifiers have high input impedance and low output impedance so that there will
not be loading effect during cascaded interconnections.

Example 1
Realize the following transfer function using opamps and RC elements.

( )( )
( )
( )( )
Solution:
( ) ( )
( )
( ) ( )

1. using inverting feedback

( )
( )

( )
( )

Note, however, that taking impedance values directly will require inductors.
For example

By M & W :ECE 3309 Page 35

To avoid the need for inductors, we use another technique.

Now,

Similarly

2. using non-inverting feedback

( ) ( )
( ) ( )

( ) ( )
( ) ( )

The same procedures are followed.

Problem 1
Realize the following transfer function with opamps and RC elements.

a) Using inverting feedback configuration

b) Using non-inverting feedback configuration

( )( )
( )
( )( )

By M & W :ECE 3309 Page 36