Chapter 3

Synthesis of 1-port Dp functions

A network function must satisfy the necessary and sufficient conditions in order to be realized
using passive elements (resistors, capacitors, inductors and transformers). For 1-port network,
the network function (deriving-point impedance or admittance function) must be a positive real
function. Once the deriving point function (Z(s) or Y(s)) is checked for positive realness, it can
be realized as a one-port passive network by breaking it up in to smaller functions, and hence to
interconnection of smaller networks.

3.1. Elementary Synthesis procedures

Network synthesis is accomplished by breaking the given deriving-point function in to sum of

smaller functions. Consider a deriving-point impedance Z(s).

If Z(s) can be written as sum of two smaller functions:

Z(s) = Z1(s) + Z2(s) then the final network will be interconnection of two networks represented
by Z1 and Z2 as follows.

When one of the functions in the sum is simple so that it can be recognized as impedance of
some network (a single element or two kinds of elements in series or in parallel) then the process
continues by realizing that function and applying the same expansion to the remaining function
until all elements are identified. Now synthesis is completed by interconnecting smaller networks
obtained from the expansion.

By workneh .h Page 1
The following are the most common network functions that can be recognized as
impedances/admittances of known networks.

Network
Network Equivalent Admittance
function form Impedance (Z(s)) (Y(s))

The synthesis process is done by breaking the given deriving-point function in to sum of one of
the above functions and the remainder function. i.e

Z(s) = Z1(s) + Z2(s)

Where, Z1(s) is one of the above functions. Z2 is obtained by subtracting Z1 from Z. Synthesis
continues by applying the same process to Z2 until all elements are identified. In the above
process, Z1 is said to be removed from Z. Except the first case (where Z1 = H is a constant), all

By workneh .h Page 2
other forms of Z1 are associated with poles so that subtracting Z1 from Z means removing the
pole associated with Z1, and hence we say a pole is removed.

Therefore, synthesis is accomplished by successive removal of poles and constants. There are
four important removals in the synthesis of 1-port deriving-point functions.
Consider a deriving-point function, Z(s), (the same applies for Y(s))

a) Removal of pole at infinity

If Z(s) is positive real and has pole at infinity, i.e Deg N = Deg D + 1

By ordinary division,

K = an+1/bn

Now Z1(s) = Ks is impedance of an inductor where L = K.

Conclusion:
Removal of pole at infinity from deriving-point impedance function corresponds to removal of an
inductor from the network.

Similarly, if the function is Y(s),

By workneh .h Page 3
This corresponds to removal of a capacitor C = K

Conclusion:
Removal of pole at infinity from deriving-point admittance function corresponds to removal of a
capacitor from the network.

b) Removal of pole at zero

If Z(s) has pole at zero, i.e if D(s) is zero at s = 0

Ko = a0/b1

Now Z1(s) = Ko/s is impedance of a capacitor C = 1/Ko

Conclusion:
Removal of pole at zero from deriving-point impedance function corresponds to removal of a
capacitor from the network.

Similarly, if the function is Y(s),

By workneh .h Page 4
 Ko = a0/b1

Now Y1(s) = Ko/s is admittance of an inductor L = 1/Ko

Conclusion:
Removal of pole at zero from deriving-point admittance function corresponds to removal of an
inductor from the network.

c) Removal of conjugate imaginary poles

If Z(s) has pole on the imaginary axis at + jω, it can be written as follows:

Now

This is impedance of parallel combination of an inductor L = k/2 and a capacitor C = 1/k

Conclusion:
Removal of conjugate imaginary poles from a deriving-point impedance function corresponds to
removal of parallel combination of an inductor and a capacitor.

By workneh .h Page 5
Similarly, for admittance function,

This corresponds to removal of series combination of an inductor L = 1/k and a capacitor C =

k/2.

Conclusion:
Removal of conjugate imaginary poles from admittance deriving-point function corresponds to
removal of series combination of an inductor and a capacitor.

d) Removal of constant
This process is important when the deriving-point function has no imaginary poles (including at s
= 0 and s = ∞). If a constant K is subtracted from Z(s),

Z(s) = K + Z2(s) where Z2(s) = Z(s) – k

The difference of two positive real functions is not necessarily positive real. Therefore the
remainder function, Z2(s), must be positive real function to remove a constant from Z(s). In other
words, Z2(s) must satisfy the following condition.

For all real ω, Re[Z2(jω)] ≥ 0

Since Z(s) is positive real, Re[Z(jω)] ≥ 0 for all real ω

Where Re[Z(jω)] =K + Re[Z2(jω)]
 Re[Z2(jω)] = Re[Z(jω)] – K
 The maximum constant, K, that can be removed from Z(s) is the minimum of Z(s)
 K < Min(Re[Z(jω)])

Removal of a constant from deriving-point impedance/admittance function corresponds to

removal of a resistor from the network.

By workneh .h Page 6
The whole synthesis process involves repeated application of pole and constant removals from
the given deriving-point function.

Example 1
Realize the following 1-port deriving-point impedance function.

Solution:
s3 + 2s s3 + 2s
𝑍 𝑠 = 4 = 2
s + +9 s + 1 (s2 + 9)

1/2s 1/2s
= + Z2 (s) Z1 s = s2 +1
s 2 +1

1/2s 7/8s
Z2 s = Z s − s2 +1
= s 2 +9

Problem 1
Realize the following 1-port deriving-point impedance function.
s3 + s
𝑍 𝑠 = 3
s + 4s 2 + s + 2

3.2. LC, RC, RL and RLCM network synthesis

By workneh .h Page 7
LC Network Synthesis
If a network contains only inductors and capacitors, it is called a pure reactive network. In pure
reactive network, the average power dissipated is zero so that it is called lossless network.
Therefore, the real part of the impedance/admittance function is zero for pure imaginary
frequency, s = jω. Consider a deriving-point impedance function Z(s).

Re[Z(jω)] = 0

Let Z(s) be written as follows:

Where, M and N are even and odd parts respectively.

Re[Z(jω)] = 0

 M1(jω)M2(jω) – N1(jω)N2(jω) = 0

 M1 = 0 = N2 Or M2 = 0 = N1
N1 M1
 Z s = or
M2 N2

 Z(s) is always even to odd (Ne(s)/Do(s)) or odd to even (No(s)/De(s)) quotient of

polynomials.

 Since N and D are either even or odd polynomials, all poles and zeros of Z lie on the jω –
axis

 Z has pole at zero (when Z(s) = Ne(s)/Do(s)) or zero at zero (when Z(s) = No(s)/De(s))

 Degrees of N and D differ exactly by one.

 Z has pole at infinity (if Deg N > Deg D) or zero at infinity (if Deg N < Deg D)

In general, Z(s) can be written as follows:

H s 2 +ω 2z 1 s 2 +ω 2z 2 …
𝑍 𝑠 = ωz1 = 0 if Z has zero at s = 0
s s 2 +ω 2p 1 s 2 +ω 2p 2 …

The general partial fraction expansion will be:

k0 k1 s k2 s
Z s = + 2 + 2 + … + Hs
s s + ωp1 s + ω2p2
2

By workneh .h Page 8
The first term exists Z has pole at s = 0, and the last term exists if Z has pole at infinity. Since
Re[Z(jw)] = 0

Z(jω) = jX(ω) the imaginary part X(ω) is called the reactance function.

−k k1 ω k2 ω
X ω = + + + ⋯ + Hω
ω −ω + ωp1 −ω + ω2p2
2 2 2

X (ω) is an odd function.

dX ω k k1 ω2 + ω2p1 k 2 ω2 + ω2p2
= 2+ 2 + 2 + ⋯+ H ≥ 0
dω ω −ω2 + ω2p1 −ω2 + ω2p2

X (ω) has a positive slope where the minimum slope is H. Therefore, poles and zeros are
interlaced.

a) When Z has zero at s = 0

b) When Z has pole at s = 0

By workneh .h Page 9
Definition:
 Poles and zeros of Z(s) are collectively known as critical frequencies
 Poles and zeros located at zero and infinity are called external critical frequencies.
 All other poles and zeros are known as internal critical frequencies.

Summary:
Z(s)
 has only simple poles and zeros all interlaced on the jω – axis.
 is a quotient of even and odd polynomials. (Ne(s)/Do(s) or No(s)/De(s))
 has a zero (No/De) or pole (Ne/Do) at s = 0
 has a zero (Deg N < Deg D) or pole (Deg N > Deg D) at infinity
 can be completely and uniquely specified by its internal critical frequencies and H

All above properties also apply for the deriving-point admittance

function Y(s)

Realization of LC Networks
There are four canonical forms:
 Foster Forms
o Foster-1: partial fraction expansion at poles of Z(s)
o Foster-2: partial fraction expansion at poles of Y(s)
 Cauer Forms
o Cauer-1: Continued fraction expansion about infinity (successive removal of pole
at infinity)
o Cauer-2: Continued fraction expansion about the origin (successive removal of
pole at zero)

Foster Realizations

By workneh .h Page 10
Foster-1:

k0 k1 s k2 s
Z s = + 2 + + … + k∞ s
s s + ω2p1 s 2 + ω2p2

Foster – 2:

k0 k1 s k2 s
Y s = + 2 + 2 + … + k∞ s
s s + ωp1 s + ω2p2
2

Cauer Realizations

Removal of pole at zero or infinity leaves a remainder function that has zero at zero or infinity
respectively. This zero can be removed as a pole of the reciprocal function. Similarly, removal of
pole from the reciprocal leaves another remainder function that has zero at that frequency.
Repeated application of this process gives continued fraction expansion of the deriving point
impedance or admittance function. This expansion can be realized as a ladder network.
If the deriving-point function is H(s), its PFE will be:

If H(s) impedance function

By workneh .h Page 11
 If H(s) is admittance function

Cauer- 1: obtained by continued removal of pole at infinity. Cauer-1 form of H(s) will be:

Cauer- 2: obtained by continued removal of pole at zero. Cauer-2 form of H(s) will be:

Example 2
Given the following LC impedance function.

s 3 + 2s
𝑍 𝑠 =
s4 + 4s2 + 3

a) Plot X(ω)
b) Find the two Foster and two Cauer realizations of Z(s).

Solution:

a)
s(s2 + 2)
𝑍 𝑠 = 2
(s + 1)(s2 + 3)

1 ω(2 − ω2 )
𝑋 𝜔 = 𝑍 𝑗𝜔 =
𝑗 (1 − ω2 )(3 − ω2 )

Poles: ω = 1, √3

By workneh .h Page 12
 Zeros: ω = 0, √2, ∞
H=1

b)
 Foster – 1

s(s2 + 2)
Z s =
(s 2 + 1)(s2 + 3)

1 1
s s
= 22 + 22
s +1 s +3

1 1
= +
1 1
2s + 1 2s + 1
2s 6s

By workneh .h Page 13
  Foster – 2

1 (s 2 + 1)(s2 + 3)
Y s = =
Z(s) s(s 2 + 2)

3 1
s
= s + + 22
2
s s +2
1 1
= s+2 + 1
s 2s+1
3 s
4

 Cauer – 1 pole removal at infinity

s 3 + 2s
𝑍 𝑠 =
s4 + 4s2 + 3

Note: Z(s) does not have pole at infinity. Therefore, we take Y(s) since it will have pole at
infinity.

1 (s 2 + 1)(s2 + 3)
Y s = =
Z(s) s(s 2 + 2)

s4 + 4s2 + 3
=
s 3 + 2s

1
=s+1 1
s+ 1
2 4s +1
s
6

By workneh .h Page 14
  Cauer – 2 Pole removal at zero

s 3 + 2s
𝑍 𝑠 =
s4 + 4s2 + 3

Z(s) does not have pole at the origin. Therefore, we take Y(s) since it will have pole at zero.

1 3 + 4s2 + s 4
Y s = =
Z(s) 2s + s3

3 1
= +
2s 4 + 1
5s 25 1
2s + 1
5s

Problem 2
Realize an LC-network deriving admittance function with the following properties.
 Internal critical frequencies
o Pole at ω = 2
o Zero at ω = 1, 3
 H=1

RC and RL Networks
RC Networks
An RC-network is built from resistors and capacitors so that it can be taken as interconnection of
smaller networks shown below.

By workneh .h Page 15
Where R = 0 (short circuited) if the sub network contains only a capacitor and c = ∞ (short
circuited) if the sub network contains only a resistor. Therefore, the general impedance function
of a sub network can be written as:

ZRC(s) = Ri + 1/Cis = A + 1/Bs

Similarly, an LC-network is built from sub networks of the form shown below.

Where L = 0 (short circuited) when there is no inductor and C = ∞ when there is no capacitor.
The general impedance function of such sub network can be written as:

ZLC(s) = Lis + 1/Cis = αs + 1/βs

Let us substitute s with p

ZLC(p) = αp + 1/βp

1 1
 Z p = α + βp 2
p LC

1 1
 [p ZLC p ]p 2 =s = α + βs

This transformed function has the same form as the RC-network impedance function so that the
general RC-network impedance function can be obtained from the general LC-network
impedance function.

1
ZRC s = [ ZLC p ]p 2 =s
p
Similarly, the admittance function can be obtained from reciprocal of the impedance function.

1
YRC (s) =
ZRC (s)

1
= [p ]p 2 =s
ZLC p

YRC s = [pYLC (p)]p 2 =s

Case 1: ZLC has pole at s = 0

By workneh .h Page 16
 N(s) H s2 + ω2z1 s2 + ω2z2 …
ZLC s = = where 0 < ωz1 < ωp1 < ωz2 < ωp2 < ⋯
D(s) s s2 + ω2p1 s2 + ω2p2 …

a) When Deg N = Deg D + 1

 Pole at ∞
 Highest critical frequency is a pole at infinity
 Highest internal critical frequency is a zero

b) When Deg N +1 = Deg D

 Highest critical frequency is a zero at infinity

 Highest internal critical frequency is a pole

1 N2 (s) H s + ω2z1 s + ω2z2 …

ZRC s = [ ZLC p ]p 2 =s = =
p D2 (s) s s + ω2p1 s + ω2p2 …

𝐻 𝑠 + 𝛼1 𝑠 + 𝛼2 …
= 𝑤𝑕𝑒𝑟𝑒 0 < 𝛼1 < 𝛽1 < 𝛼2 < 𝛽2 < ⋯
𝑠 𝑠 + 𝛽1 𝑠 + 𝛽2 …

The first critical frequency is a pole at zero

►Lowest critical frequency is a pole
Poles and zeros are simple and lie on the negative real axis
►Poles and zeros interlace

a) When Deg N = Deg D + 1

 Deg N2 = Deg D2
 No pole at infinity
 Highest critical frequency for ZRC is a zero

b) When Deg N +1 = Deg D

 Deg N2 < Deg D2

 A zero at infinity
 Highest critical frequency is a zero.

Case 2: ZLC has zero at s = 0

N(s) Hs s2 + ω2z1 s2 + ω2z2 …

ZLC s = = where 0 < ωp1 < ωz1 < ωp2 < ωz2 < ⋯
D(s) s 2 + ω2p1 s 2 + ω2p2 …

By workneh .h Page 17
 1 N2 s H s + ω2z1 s + ω2z2 …
ZRC s = [ ZLC p ]p 2 =s = =
p D2 s s + ω2p1 s + ω2p2 …

H s + α1 s + α2 …
= where 0 < β1 < α1 < β2 < α2 < ⋯
s + β1 s + β2 …

 The first critical frequency is a pole at β1

 Lowest critical frequency is a pole
 Poles and zeros are simple and lie on the negative real axis
 Poles and zeros interlace

a) When Deg N = Deg D + 1

 Deg N2 = Deg D2
 No pole at infinity
 Highest critical frequency for ZRC is a zero

b) When Deg N +1 = Deg D

 Deg N2 < Deg D2

 A zero at infinity
 Highest critical frequency is a zero.

Note that poles and zeros of ZRC(s) are zeros and poles of YRC(s) respectively.

 Summary:
Poles and zeros are simple and lie on the negative real frequency axis.
Poles and zeros interlace
Highest critical frequency is a zero for ZRC and a pole for YRC
Lowest critical frequency is a pole for ZRC and a zero YRC.
Deg N < Deg D for Z and Deg N > Deg D for Y.
In general,

H s + α1 s + α2 …
ZRC s =
s s + β1 s + β2 …

k0 k1 k2
= + + + … + k ∞ ← if Deg N = Deg D
s s + β1 s + β2

𝑑𝑍𝑅𝐶 (𝛿) −k0 −k1 −k2

= 2 + + + … ≤0
𝑑𝛿 𝛿 (𝛿 + 𝛿1 ) 2
(𝛿 + 𝛿2 )2

By workneh .h Page 18
  Z(𝛿) has a negative slope along the real frequency axis. Moreover, ZRC(s) has no poles and
zeros on the positive real frequency axis as it is a positive real function.
 Z(𝛿) is monotically decreasing from 𝛿 = 0 to 𝛿 = ∞.
 Z(0) > Z(∞)
 Y(0) < Y(∞).

Recall:

H s + ω2z1 s + ω2z2 … k0 k1 s k2 s
YLC s = = + 2 + + … + Hs
s s + ω2p1 s + ω2p2 … s s + ω2p1 s 2 + ω2p2

k1 s k2 s
YRC s = pYLC p p 2 =s = k0 + 2 + + … + Hs
s + ωp1 s + ω2p2

1 k0 k1 k2
( )YRC s = + + + …+ H
s s s + ωp1 s + ω2p2
2

1 k0 k1 k2
YRC s = + + + … + k∞
s s s + β1 s + β2

The residues at poles of ZRc and YRc /s are real and positive.

RL Networks

ZRL(s) = Ri +Lis = A + Bs
Similarly, for LC-network

ZLC(s) = Lis + 1/Cis = αs + 1/βs

Let us s with p

ZLC(p) = αp + 1/βp

By workneh .h Page 19
 1
 (p)ZLC p = αp2 +
β

1
 [(p)ZLC p ]p 2 =s = β + αs = X + Ys

This transformed function has the same form as the RL-network impedance function so that the
general RL-network impedance function can be obtained from the general LC-network
impedance function.

ZRL s = [pZLC p ]p 2 =s

Since the impedance and admittance functions of LC networks have similar

characteristics, and hence similar forms, we can substitute ZLC with YLC in
the above transformation.

ZRL s = [pYLC p ]p 2 =s This is similar to YRC

Similarly, the admittance function can be obtained from reciprocal of the impedance function.

1
YRL (s) =
ZRL (s)

1 1
=[ ] 2
p YLC p p =s
1
YRL p = [p ZLC (p)]p 2 =s This is similar to ZRC

Conclusion:
o Deriving-point RC-impedance and RL-admittance functions have similar forms,
and hence similar characteristics.
o Deriving-point RC-admittance and RL-impedance functions have similar forms,
and hence similar characteristics.

Foster realization of RC networks

Foster – 1
𝑁(𝑠)
𝑍𝑅𝐶 𝑠 = 𝐷𝑒𝑔 𝑁 𝑠 ≤ 𝐷𝑒𝑔 𝐷(𝑠)
𝐷(𝑠)

By workneh .h Page 20
 H s + α1 s + α2 …
ZRC s =
s s + β1 s + β2 …

k0 k1 k2
= + + + … + k ∞ ← if Deg N = Deg D
s s + β1 s + β2

1 1 1
= + + + … + k∞
1 1 1 1 1
s s+ s+
k0 k1 k1 /β1 k2 k 2 /β2

Foster – 2

H s + α1 s + α2 …
YRC s = where 0 < α1 < β1 < α2 < β2 < ⋯
s + β1 s + β2 …

1 k0 k1 k2
YRC s = + + + … + k∞
s s s + β1 s + β2

k1 s k2 s
YRC s = k0 + + + … + k∞ s
s + β1 s + β2

1 1
= k0 + + + … + k∞ s
1 1 1 1
+ +
k1 k1 /β1 s k 2 k 2 /β2 s

By workneh .h Page 21
Foster realization of RL networks

Recall that
o Deriving-point RC-impedance and RL-admittance functions have similar forms,
and hence similar characteristics.
o Deriving-point RC-admittance and RL-impedance functions have similar forms,
and hence similar characteristics.

Foster – 1

H s + α1 s + α2 …
ZRL s = where 0 < α1 < β1 < α2 < β2 < ⋯
s + β1 s + β2 …

1 k0 k1 k2
ZRL s = + + + … + k∞
s s s + β1 s + β2

k1 s k2 s
ZRL s = k0 + + + … + k∞ s
s + β1 s + β2

1 1
= k0 + + + … + k∞ s
1 1 1 1
+ +
k1 k1 /β1 s k 2 k 2 /β2 s

Foster – 2

By workneh .h Page 22
 𝑁(𝑠)
𝑌𝑅𝐿 𝑠 = 𝐷𝑒𝑔 𝑁 𝑠 ≤ 𝐷𝑒𝑔 𝐷(𝑠)
𝐷(𝑠)

H s + α1 s + α2 …
YRL s =
s s + β1 s + β2 …

k0 k1 k2
= + + + … + k ∞ ← if Deg N = Deg D
s s + β1 s + β2

1 1 1
𝑌𝑅𝐿 𝑠 = + + + … + k∞
1 1 β 1 β
s s+ 1 s+ 2
k0 k1 k1 k2 k2

Example 3
Find Foster – 1 and Foster – 2 realization of the following deriving-point function.

𝑠 + 1 𝑠 + 3 (𝑠 + 5)
𝑍𝑅𝐶 𝑠 =
𝑠 𝑠 + 2 𝑠 + 4 (𝑠 + 6)

Solution:

F–1

𝐷𝑒𝑔 𝑁 𝑠 < 𝐷𝑒𝑔 𝐷 𝑠 𝑡𝑕𝑒𝑟𝑒𝑓𝑜𝑟𝑒 𝑘∞ = 0

𝑘0 𝑘1 𝑘2 𝑘3
𝑍𝑅𝐶 𝑠 = + + +
𝑠 𝑠+2 𝑠+4 𝑠+6
5 3
𝑘0 = ; 𝑘1 = = 𝑘2 ; 𝑘3 = 5/16
16 16
1 1 1 1
𝑍𝑅𝐶 𝑠 = + + +
16/5𝑠 16/3𝑠 + 1 1 1
16/3𝑠 + 16/5𝑠 +
3/32 3/64 5/96

By workneh .h Page 23
F–2

𝑠 𝑠 + 2 𝑠 + 4 (𝑠 + 6)
𝑌𝑅𝐶 𝑠 =
𝑠 + 1 𝑠 + 3 (𝑠 + 5)

𝑌𝑅𝐶 𝑠 𝑠 + 2 𝑠 + 4 (𝑠 + 6)
=
𝑠 𝑠 + 1 𝑠 + 3 (𝑠 + 5)

𝑘1 𝑘2 𝑘3
= + + + 𝑘∞
𝑠+1 𝑠+3 𝑠+5

𝑘1 = 15/8; 𝑘2 = 3/4 ; 𝑘3 = 3/8; 𝑘∞ = 1

𝑌𝑅𝐶 𝑠 15/8 3/4 3/8

= + + + 1
𝑠 𝑠+1 𝑠+3 𝑠+5

15/8𝑠 3/4𝑠 3/8𝑠

𝑌𝑅𝐶 𝑠 = + + + 𝑠
𝑠+1 𝑠+3 𝑠+5
1 1 1
= + + + 𝑠
1 1 1
8/15 + 4/3 + 8/3 +
15/8𝑠 1/4𝑠 3/40𝑠

Example 4
Consider the following one-port RL network deriving-point function.

By workneh .h Page 24
 𝑠+1 𝑠+3
𝐹 𝑠 =
𝑠+2 𝑠+4

a) Identify whether the function is impedance or admittance function.

b) Realize it using Foster – 1 method.

Solution:

 Poles: s = -2, -4
 Zeros: s = -1, -3
 Least critical frequency (*** the one nearest to the origin) is a zero
 Highest critical frequency (*** the farthest from the origin) is a pole
 This is property of RC admittance and RL impedance functions.
 Therefore, the function is impedance function since it is RL network deriving-point
function.

𝑠+1 𝑠+3
𝑍𝑅𝐿 𝑠 =
𝑠+2 𝑠+4

Foster – 1

𝑍𝑅𝐿 𝑠 𝑠+1 𝑠+3

=
𝑠 𝑠 𝑠+2 𝑠+4

𝑘0 𝑘1 𝑘2
= + +
𝑠 𝑠+2 𝑠+4

𝑘0 = 3/8; 𝑘1 = 1/4; 𝑘2 = 3/8

3/8𝑠 1/4𝑠 3/8𝑠

𝑍𝑅𝐿 𝑠 = + +
𝑠 𝑠+2 𝑠+4

3 1 1
= + +
8 4+ 1 8/3 +
1
1/8𝑠 3/32𝑠

By workneh .h Page 25
Cauer Realization of RC and RL deriving-point impedance function
Cauer realization of a deriving – point function is obtained by Continued Fraction Expansion
of the function about the highest or lowest degrees of both the numerator and the denominator
polynomials.

For LC deriving-point functions, highest and lowest degrees of the numerator are always
deferent (by one) from the highest and lowest degrees of the denominator respectively so that
division about the highest or lowest degrees always gives a quotient polynomial of the form αs
or 1/αs that removes a pole at infinity or at zero respectively. Therefore, Cauer realization of LC
deriving – point functions is obtained by continued removal of poles.

On the other hand, RC and RL deriving – point functions (N(s)/D(s)) can have both numerator
and denominator polynomials of the same degree. In this case, division about the highest or
lowest degrees gives a constant quotient that removes a constant from the deriving - point
function. But to remove a constant from the deriving – point function, the constant must be less
than or equal to the minimum of the function. Therefore, it should be known whether from the
impedance or admittance a constant can be removed.

Recall the following properties.

ZRC(0) > ZRC(∞) ► YRL(0) > YRL(∞) since ZRC and YRL have similar properties
YRC(0) < YRC(∞) ► ZRL(0) < ZRL(∞) since YRC and ZRL have similar properties

This shows that:

ZRC and YRL have minimum values at infinity (∞) so that a constant can be removed during
Cauer-1 realization (continued fraction expansion about infinity).
YRC and ZRL have minimum values at zero (0) so that a constant can be removed during Cauer-2
realization (continued fraction expansion about zero).

Cauer – 1 realization

For RC networks use the impedance function, ZRC, and for RL network use the admittance
function, YRL.

N s
ZRC s = = YRL s Deg N(s) < Deg D(s) ► No pole at infinity
D s

Case 1: Deg N = Deg D

By workneh .h Page 26
Since these functions have minimum values at infinity, a constant can be removed by division
about infinity (highest degrees). This leaves a remainder function whose numerator has a degree
less that the denominator, and hence has zero at infinity. This can be removed as a pole from the
reciprocal. Now the second remainder function will have both numerators and denominators of
the same degree. Now a constant (at infinity) cannot be removed from these functions! But a
constant (at infinity) can be removed from the reciprocal since it has a minimum at infinity.

Let us assume the function is ZRC s (the same conclusion can be reached for YRL s )

A constant at infinity can be removed from ZRC not YRC. Now after a constant is removed from
Z, the remainder Z2 will have a numerator with its degree less than degree of the denominator so
that Z2 will have zero at infinity. This implies that Y2 = 1/Z2 has a pole at infinity. After a pole is
removed from Y2, the remainder Y3 will again have numerator and denominator with the same
degree. Since a constant (at infinity) cannot be removed from YRC (and ZRL) we take the
reciprocal of Y3 (Z3) and remove a constant. The expansion continues by applying the same
process.

Cauer – 1 forms of ZRC and YRL will be as shown below.

1
𝑍𝑅𝐶 𝑠 = 𝛽1 + = 𝑌𝑅𝐿 (𝑠)
1
𝛽2 𝑠 + 1
𝛽3 +
𝛽4 𝑠 + ⋯

Case 2: Deg N < Deg D zero at infinity

In this case there is no constant removal at infinity (the function becomes zero at infinity) so that
the expansion starts from the reciprocal function. 𝜷𝟏 = 𝟎

1
𝑍𝑅𝐶 𝑠 = = 𝑌𝑅𝐿 (𝑠)
1
𝛽2 𝑠 + 1
𝛽3 +
𝛽4 𝑠 + ⋯

Cauer – 2 realizations

For RC networks use the admittance function, YRC, and for RL network use the impedance
function, ZRL.

N s
YRC s = = ZRL s Deg N > Deg D
D s

These functions have minimum values at zero so that a constant about the origin (at zero) can be
removed by dividing about the lowest degrees.

By workneh .h Page 27
Case 1: zero at s = 0

No constant removal at zero since the function becomes zero at s = 0. However, the reciprocal
will have pole at zero. This pole is removed first and the remainder will have no pole at zero
anymore. Now a constant at zero is removed from the reciprocal. Removal of a constant at zero
leaves another remainder which has zero at s = 0 (since the constant is already removed).
Therefore, a pole at zero is removed from reciprocal of this function. The expansion continues by
following the same procedure.

Finally, Caurer-1 form of such function will be as shown below.

1
𝑌𝑅𝐶 𝑠 = = 𝑍𝑅𝐿 (𝑠)
1 1
𝛼1 𝑠 1 1
+ 1 1
𝛼2
𝛼3 𝑠 + 1
𝛼4 + ⋯
Case 2: no zero at s = 0
In this case a constant at zero is removed first.

1
𝑌𝑅𝐶 𝑠 = 𝛼 + = 𝑍𝑅𝐿 (𝑠)
1 1
𝛼1 𝑠 1 1
𝛼2 + 1 1
+
𝛼3 𝑠 1
𝛼4 + ⋯

Example 5

Realize the following RL deriving-point function using Cauer – 1 and Cauer – 2

𝑠 2 + 5𝑠 + 4
𝑍𝑅𝐿 𝑠 =
3𝑠 + 4

Solution:
ZRL(0) < ZRL(∞)
YRL(∞) < YRL(0)

For Cauer – 1 (Continued Fraction Expansion about infinity), use YRL

For Cauer – 2 (Continued Fraction Expansion about the origin) use ZRL

Cauer – 1

By workneh .h Page 28
 3𝑠 + 4
𝑌𝑅𝐿 𝑠 =
𝑠 2 + 5𝑠 + 4

Deg N < Deg D

 No pole or constant removal at infinity
 Start from the reciprocal function (ZRL)


𝑠 2 + 5𝑠 + 4
𝑍𝑅𝐿 𝑠 =
3𝑠 + 4

1 1
= 𝑠+
3 9 1
11 + 121 1
24 𝑠 + 2/11

Cauer – 2

𝑠 2 + 5𝑠 + 4
𝑍𝑅𝐿 𝑠 =
3𝑠 + 4

 No zero at zero
 A constant can be removed at zero

1 1
𝑍𝑅𝐿 𝑠 = 𝑠+
3 9 1
11 + 121 1
𝑠+
4 2/11

1
=1+
2 1
𝑠+ 1
2+
1/𝑠

By workneh .h Page 29
By workneh .h Page 30
1-port RLCM Network Synthesis
Synthesis procedures:
1. Test Z(s) (or Y(s)) for jω – axis poles and zeros, and if any remove and realize. Zeros are
removed as poles of the reciprocal function.
2. Test the remainder function for jω – axis poles and zeros, and if any remove and realize
3. Repeat steps (1) & (2) until the remainder has no more poles and zeros.
4. Find minimum value of real part of the remainder function along the jω – axis. Remove
this constant as a resistor. Then go to step (1).

Example 6
Realize the following deriving-point impedance function.

2𝑠 3 + 3𝑠 2 + 2𝑠 + 3
𝑍 𝑠 = 3
𝑠 + 3𝑠 2 + 4𝑠 + 1

Solution:

jω – axis zero:

𝑃𝑘 𝑠 = 3(𝑠 2 + 1)

 Z(s) has zero at s = + j

(𝑠 2 + 1)(2𝑠 + 3)
𝑍 𝑠 =
𝑠 3 + 3𝑠 2 + 4𝑠 + 1

This zero can be removed as pole of Y(s).

𝑠 3 + 3𝑠 2 + 4𝑠 + 1
𝑌 𝑠 =
(𝑠 2 + 1)(2𝑠 + 3)

𝑘1 𝑠
= + 𝑌1 𝑠 ; 𝑘1 = 1
𝑠2+1
By workneh .h Page 31
 𝑠 1
=
𝑠2 +1 1
𝑠+𝑠

𝑠+1
𝑌1 𝑠 =
2𝑠 + 3

 No jω – axis poles and zeros

 Constant removal at minimum of Re[Y1(jω)]

3 + 2ω2
R e Y1 jω =
9 + 4ω2

Let ω2 = x

3 + 2x
9 + 4x

The function will have minimum value at one of the following points
 At the boundaries (at 0 or at ∞)
 At one of the points where derivative of the function becomes zero (points where the
slope becomes zero)

 At the boundaries
o R e Y1 j0 = 1/3

o R e Y1 j∞ = 1/2

 At zeros of the derivative

𝑑 3 + 2x
=0
𝑑𝑥 9 + 4x


6
=0
(9 + 4x)2
No finite root

By workneh .h Page 32
Therefore, the minimum is 1/3 at ∞

𝑠+1 1
𝑌1 𝑠 = = + 𝑌2 𝑠
2𝑠 + 3 3

𝑠
𝑌2 𝑠 =
6𝑠 + 9
1
=
9
6𝑠 + 𝑠

Synthesis is complete!

Example 7
Realize the following deriving-point impedance function.

5𝑠 2 + 18𝑠 + 8
𝑍 𝑠 =
𝑠 2 + 𝑠 + 10

Solution:

 No jω – axis poles and zeros

5𝜔4 − 40𝜔2 + 80
𝑅𝑒 [𝑍 𝑗𝜔 ] =
𝜔 4 − 19𝜔 2 + 10

5 𝜔2 − 4 2
= 4
𝜔 − 19𝜔 2 + 10
By workneh .h Page 33
 𝑅𝑒 𝑍 𝑗𝜔 = 0 𝑎𝑡 = 2

 𝑅𝑒 𝑍 𝑗2 = 0
 𝑅𝑒 𝑍 𝑗0 = 0.8
 𝑅𝑒 𝑍 𝑗∞ = 5

 No constant Removal

Important Definitions:
1. An impedance function, Z(s), with no jω – axis poles and zeros is called minimum
reactance function.
2. An admittance functions, Y(s), with no jω – axis poles and zeros is called minimum
suspectance function.
3. An impedance function whose real part vanishes at some real finite non-zero frequency,
ω, is called minimum resistance function.
4. An admittance function whose real part vanishes at some real, finite non-zero frequency,
ω, is called minimum conductance function.
5. A positive real function which is simultaneously minimum reactance/suspectance and
minimum resistance/conductance is called minimum function.

Minimum functions have the following properties:

i) No jω – axis poles and zeros
ii) Finite, real and positive values at the boundaries (at zero and at infinity)
iii) Real part vanishes at least at one finite, non-zero frequency, ω1, such that

Z(jω1) = 0 + jX1

Brune’s RLCM synthesis method

Consider Z(s) is a minimum function whose real part vanishes at s = jω1

Z(jω1) = jX1 = jω1L1 L1 = X1/ω1

By workneh .h Page 34
Brune’s synthesis cycle begins with removal of an inductor, L1, real number that can be either
positive or negative.

Case 1: L1 is negative (L1 < 0 )

Step 1: Removal of L1

𝑍2 𝑠 = 𝑍 𝑠 − 𝐿1 𝑠

 L1s is not positive real function (L1 < 0)

 -L1s is positive real function
 Z2 = Z + (-L1s) which is sum of two positive real functions is positive real

𝑍2 𝑗𝜔1 = 𝑍 𝑗𝜔1 − 𝑗𝜔1 𝐿1

= 𝑗𝑋1 − 𝑗𝜔1 𝐿1

=0

 Z2 has zero at s = jω1

 Y2 (= 1/Z2) has pole at s = jω1.

𝑘1 𝑠
𝑌2 𝑠 = + 𝑌3 𝑠
𝑠2 + 𝜔12

𝑘1 𝑠 1 1 𝟏
2 = 1 1
=
1
; 𝒌𝟏 =
𝑳𝟐
𝑠2 + 𝜔1 𝑠+ 𝐿2 𝑠 + 𝐶 𝑠
𝑘1 𝑘1 2
𝑠
𝜔12

By workneh .h Page 35
 𝑘1 𝑠
𝑌3 𝑠 = 𝑌2 𝑠 −
𝑠2 + 𝜔12

1 𝑘1 𝑠
= − 2
𝑍2 𝑠 𝑠 + 𝜔12

1 𝑘1 𝑠
= − 2
𝑍 𝑠 − 𝐿1 𝑠 𝑠 + 𝜔12

1 + 𝑘1 𝐿1 𝑠 2 + 𝑘1 𝑠𝑍 𝑠 + 𝜔12
=
−𝐿1 𝑠 3 + 𝑠 2 𝑍 𝑠 − 𝐿1 𝜔12 𝑠 + 𝜔12 𝑍(𝑠)

−𝐿1 𝑠 3 + 𝑠 2 𝑍 𝑠 − 𝐿1 𝜔12 𝑠 + 𝜔12 𝑍(𝑠)

𝑍3 𝑠 =
1 + 𝑘1 𝐿1 𝑠 2 + 𝑘1 𝑠𝑍 𝑠 + 𝜔12

 Z3(s) has pole at infinity which can be removed as an inductor

 Z3(s) = L3s + Z4(s)

−𝐿1 1
𝐿3 = 𝑤𝑕𝑒𝑟𝑒 𝑘1 =
1 + 𝑘1 𝐿1 𝐿2

−𝐿1
𝐿3 =
𝐿
1 + 𝐿1
2

 The above operations make one Brune cycle.

 Z4 again may be realized using pole and constant removal until the remainder function
becomes a minimum function. In this case a second Brune cycle follows.
 Brune cycle removes one capacitor and three inductors among which one has negative
value. These inductors cannot be realized directly, but rather the three inductors will be
replaced by an equivalent transformer.

Consider the following networks.

By workneh .h Page 36
 𝑉1 = 𝐿1 𝑠𝐼1 + 𝐿2 𝑠 𝐼1 + 𝐼2

𝑉1 = (𝐿1 + 𝐿2 )𝑠𝐼1 + 𝐿2 𝑠𝐼2

Similarly,
𝑉2 = 𝐿2 𝑠𝐼1 + (𝐿2 + 𝐿3 )𝑠𝐼2

𝑉1 (𝐿1 + 𝐿2 )𝑠 𝐿2 𝑠 𝐼1
=
𝑉2 𝐿2 𝑠 (𝐿2 + 𝐿3 )𝑠 𝐼2

𝑉1 = 𝐿𝑝 𝑠𝐼1 + 𝑀𝑠𝐼2
Similarly,
𝑉2 = 𝑀𝑠𝐼1 + 𝐿𝑠 𝑠𝐼2

𝑉1 𝐿𝑝 𝑠 𝑀𝑠 𝐼1
=
𝑉2 𝑀𝑠 𝐿𝑠 𝑠 𝐼2

The two networks are equivalent if they have the same VI relations. i.e if they have the same
impedance matrices.

𝐿𝑝 = 𝐿1 + 𝐿2 ; 𝐿𝑠 = 𝐿2 + 𝐿3 ; 𝑀 = 𝐿2

Coupling coefficient

𝑀 𝐿2 𝐿2 𝐿2
𝑘= = = = =1
𝐿𝑝 𝐿𝑠 𝐿1 + 𝐿2 (𝐿2 + 𝐿3 ) −𝐿1 𝐿2 2
𝐿1 + 𝐿2 (𝐿2 + 𝐿1 )
1+𝐿
2
For the first case of Brune cycle

𝐿1 < 0 ; 𝐿2 > 0 ; 𝐿3 > 0

This implies

By workneh .h Page 37
 𝑀 = 𝐿2 > 0 ; 𝐿𝑠 = 𝐿2 + 𝐿3 > 0 ; 𝐿𝑝 𝐿𝑠 = 𝐿2 2 > 0 → 𝐿𝑝 > 0

Conclusion:
The Brune’s T-coupled 3 inductors (L1, L2, & L3) can be replaced by the equivalent realizable
transformer whose coupling coefficient k =1.

𝐿𝑝 = 𝐿1 + 𝐿2 ; 𝐿𝑠 = 𝐿2 + 𝐿3 ; 𝑀 = 𝐿2

Case 2: L1 is positive (L1 > 0 )

Step 1: Removal of L1
𝑍2 𝑠 = 𝑍 𝑠 − 𝐿1 𝑠

 Z(s) is finite at ∞ (no pole or zero at infinity)

 𝑍2 ∞ = 𝑍 ∞ − ∞ = −∞ → 𝑝𝑜𝑙𝑒 𝑎𝑡 𝑖𝑛𝑓𝑖𝑛𝑖𝑡𝑦
 𝑍2 𝑚𝑢𝑠𝑡 𝑕𝑎𝑣𝑒 𝑜𝑛𝑒 𝑧𝑒𝑟𝑜 𝑖𝑛 𝑡𝑕𝑒 𝑅𝐻 𝑠 − 𝑝𝑙𝑎𝑛𝑒 𝑑𝑢𝑒 𝑡𝑜 𝑎𝑑𝑑𝑖𝑡𝑖𝑜𝑛 𝑜𝑓 𝑝𝑜𝑙𝑒 𝑎𝑡 𝑖𝑛𝑓𝑖𝑛𝑖𝑡𝑦
 Z2 is not positive real function

𝑍2 𝑗𝜔1 = 𝑍 𝑗𝜔1 − 𝑗𝜔1 𝐿1

= 𝑗𝑋1 − 𝑗𝜔1 𝐿1

=0

 Z2 has zero at s = jω1

 Y2 (= 1/Z2) has pole at s = jω1.

By workneh .h Page 38
 𝑘1 𝑠
𝑌2 𝑠 = + 𝑌3 𝑠
𝑠2 + 𝜔12

𝑘1 𝑠 1 1 𝟏
= = ; 𝒌𝟏 =
𝑠 2 + 𝜔12 1 1 1 𝑳𝟐
𝑠+ 𝐿2 𝑠 + 𝐶 𝑠
𝑘1 𝑘1 2
𝑠
𝜔12

𝑘1 𝑠
𝑌3 𝑠 = 𝑌2 𝑠 −
𝑠 2 + 𝜔12

1 𝑘1 𝑠
= − 2
𝑍2 𝑠 𝑠 + 𝜔12

1 𝑘1 𝑠
= − 2
𝑍 𝑠 − 𝐿1 𝑠 𝑠 + 𝜔12

1 + 𝑘1 𝐿1 𝑠 2 + 𝑘1 𝑠𝑍 𝑠 + 𝜔12
=
−𝐿1 𝑠 3 + 𝑠 2 𝑍 𝑠 − 𝐿1 𝜔12 𝑠 + 𝜔12 𝑍(𝑠)

−𝐿1 𝑠 3 + 𝑠 2 𝑍 𝑠 − 𝐿1 𝜔12 𝑠 + 𝜔12 𝑍(𝑠)

𝑍3 𝑠 =
1 + 𝑘1 𝐿1 𝑠 2 + 𝑘1 𝑠𝑍 𝑠 + 𝜔12

 Z3(s) has pole at infinity which can be removed as an inductor

 𝑍4 = 𝑍3 𝑠 − 𝐿3 𝑠

−𝐿1 1
𝐿3 = < 0 𝑤𝑕𝑒𝑟𝑒 𝑘1 =
1 + 𝑘1 𝐿1 𝐿2

∗ 𝑅𝑒𝑚𝑜𝑣𝑎𝑙 𝑜𝑓 𝑡𝑕𝑖𝑠 𝑝𝑜𝑙𝑒 𝑎𝑡 𝑖𝑛𝑓𝑖𝑛𝑖𝑡𝑦 𝑟𝑒𝑚𝑜𝑣𝑒𝑠 𝑡𝑕𝑒 𝑅𝐻 𝑠 − 𝑝𝑙𝑎𝑛𝑒 𝑧𝑒𝑟𝑜 𝑠𝑜 𝑡𝑕𝑎𝑡 𝑍4

= 𝑍3 𝑠 − 𝐿3 𝑠 𝑖𝑠 𝑝𝑜𝑠𝑖𝑡𝑖𝑣𝑒 𝑟𝑒𝑎𝑙
−𝐿1
𝐿3 =
𝐿
1 + 𝐿1
2

By workneh .h Page 39
 𝐿𝑝 = 𝐿1 + 𝐿2 ; 𝐿𝑠 = 𝐿2 + 𝐿3 ; 𝑀 = 𝐿2

For the second case of Brune cycle

𝐿1 > 0 ; 𝐿2 > 0 ; 𝐿3 < 0

This implies

𝑀 = 𝐿2 > 0 ; 𝐿𝑃 = 𝐿1 + 𝐿2 > 0 ; 𝐿𝑝 𝐿𝑠 = 𝐿2 2 > 0 → 𝐿𝑠 > 0

Example 7
Realize the following deriving-point impedance function.

6𝑠 2 + 19𝑠 + 18
𝑍 𝑠 =
𝑠 2 + 𝑠 + 10

Solution:

 No jω – axis poles and zeros

 Minimum of Re[Z(jω)]

6𝜔4 − 59𝜔2 + 180

𝑅𝑒 𝑍(𝑗𝜔) =
𝜔 4 − 19𝜔 2 + 100

𝐿𝑒𝑡 𝑥 = 𝜔2

By workneh .h Page 40
 6𝑥 2 − 59𝑥 + 180
= 2
𝑥 − 19𝑥 + 100

𝑑 6𝑥 2 − 59𝑥 + 180
= 0 → 𝑥 = 4 𝑜𝑟 𝑥 = 11.27 → 𝜔 = 2 𝑜𝑟 𝜔 = 11.27
𝑑𝑥 𝑥 2 − 19𝑥 + 100

 𝑅𝑒 𝑍 𝑗2 = 1
 𝑅𝑒 𝑍 𝑗 11.27 = 21.56
At the boundaries
 𝑅𝑒 𝑍 𝑗0 = 1.8
 𝑅𝑒 𝑍 𝑗∞ = 6

 𝑀𝑖𝑛[𝑅𝑒 𝑍 𝑗𝜔 ] = 1 𝑎𝑛𝑑 𝜔1 = 2

𝑍1 𝑠 = 𝑍 𝑠 − 1

5𝑠 2 + 18𝑠 + 8
=
𝑠 2 + 𝑠 + 10

𝑁𝑜𝑤 𝑍1 𝑠 𝑖𝑠 𝑎 𝑚𝑖𝑛𝑖𝑚𝑢𝑚 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛 𝑎𝑛𝑑 𝜔1 = 2

𝑍1 𝑗2 = 𝑗6 = 𝑗𝜔1 𝐿1 → 𝐿1 = 3 > 0 [𝑐𝑎𝑠𝑒 2]

𝑍2 𝑠 = 𝑍1 𝑠 − 𝐿1 𝑠

5𝑠 2 + 18𝑠 + 8
= − 3𝑠
𝑠 2 + 𝑠 + 10

−3𝑠 3 + 2𝑠 2 − 12𝑠 + 8 (−3𝑠 + 2)(𝑠 2 + 4)

𝑍2 𝑠 = =
𝑠 2 + 𝑠 + 10 𝑠 2 + 𝑠 + 10

By workneh .h Page 41
 𝑠 2 + 𝑠 + 10
𝑌2 𝑠 = 𝑕𝑎𝑠 𝑝𝑜𝑙𝑒 𝑎𝑡 𝜔1 = 2
(−3𝑠 + 2)(𝑠 2 + 4)

𝑘1 𝑠 1
𝑌3 𝑠 = 𝑌2 𝑠 − 𝑘1 =
𝑠2+4 2

1
2𝑠 = 1
𝐿2 = 2 𝐶2 = 1/8
𝑠2 + 4 1
2𝑠 + 1
8𝑠

5
𝑌3 𝑠 = 2
−3𝑠 + 2
6 4 6 4
𝑍3 𝑠 = − 𝑠 + 𝐿3 = − 𝑍4 =
5 5 5 5

By workneh .h Page 42
 𝑀 = 𝐿2 = 2 ; 𝐿𝑃 = 𝐿1 + 𝐿2 = 5 ; 𝐿𝑠 = 𝐿2 + 𝐿3 = 4/5

Problem 2
Realize the following impedance function

𝑠2 + 𝑠 + 2
𝑍 𝑠 =
2𝑠 2 + 𝑠 + 1

By workneh .h Page 43
By workneh .h Page 44