Lecture 1

Microwave Networks and Scattering

Matrix
Prof. Ratnajit Bhattacharjee
IIT Guwahati
 Contents
Impedance and Equivalent Voltages and Currents

N-port microwave networks

Impedance, admittance, and scattering matrix representations

Reciprocal and lossless networks

Transmission matrix
 Impedance and Equivalent Voltages and Currents
• At microwave frequencies, measurement of voltage or current is
not practical unless a clearly defined terminal pair is available.
• Such terminal pair may exist for TEM type lines but does not exist
in a strict sense for a non-TEM line
• Therefore, voltage and current as a measure of level of electrical
excitation of a circuit does not play a primary role at microwave
frequencies.
• However, introduction of equivalent voltages, currents and
impedances is helpful in extending circuit theory concepts to
microwave network.
 Impedance and Equivalent Voltages and Currents
For transmission line supporting TEM waves,
there exist a voltage and current uniquely related
to the transverse electric and magnetic field,
respectively.
𝐸
Let us illustrate the same by an example
The figure shows a co-axial transmission line 𝑎 𝐻

𝑉0 𝑒 −𝑗𝑘0 𝑧 𝑏
𝐸= 𝑎
ෞ𝜌
𝑏 𝜌
ln
𝑎
𝑘0 = 𝜔 𝜇0 𝜖0
𝑉0 is the potential difference between the inner
and outer conductor
 Impedance and Equivalent Voltages and Currents
Voltage wave associated with the electric field is
𝑉 = 𝑉0 𝑒 −𝑗𝑘0𝑧
The magnetic field is given by
𝑌0 𝑉0 𝑎 ෞ∅ −𝑗𝑘 𝑧
𝐻= 𝑒 0
𝑏 𝜌
ln
𝑎
Current wave associated with the magnetic field is
𝐼 = 𝐼0 𝑒 −𝑗𝑘0𝑧
𝑌0 𝑉0 2𝜋
where 𝐼0 = 𝑏
ln
𝑎
 Impedance and Equivalent Voltages and Currents
2
1 𝑏 2𝜋 ∗
𝜋𝑌 𝑉
0 0
𝑃 = න න 𝐸×𝐻 ∙𝑎 ෞ𝜌 𝑑𝜌𝑑∅ =
2 𝑎 0 ln 𝑏Τ𝑎
We find that
2
1 ∗
𝜋𝑌0 𝑉0
𝑅𝑒 𝑉𝐼 =
2 ln 𝑏Τ𝑎
and
𝑉0
𝑍0 =
𝐼0
 Impedance and Equivalent Voltages and Currents
• However, for waveguides there is
difficulty in defining such voltages and
𝑦
currents
• For the dominant 𝑇𝐸 10 mode of a
rectangular waveguide, the electric
field distribution is as shown
x
−𝑗𝜔𝜇𝑎 𝜋𝑥 −𝑗𝛽𝑧
• 𝑉= 𝐴 sin 𝑒 ‫𝑦׬‬ 𝑑𝑦
𝜋 𝑎
• Voltage depends on the position 𝑥
 Impedance and Equivalent Voltages and Currents
Propagating waveguide modes have the following properties:
• Power transmitted is given by an integral involving the transverse
electric and transverse magnetic fields only.
• In a loss-free guide supporting several propagating modes, Power
transmitted is the sum of individual modes
• Transverse fields vary with distance along the guide according to a
propagation factor 𝑒 ±𝑗𝛽𝑧
• Transverse magnetic field is related to transverse electric field as
𝑍𝑤 ℎത = 𝑎ෞ𝑧 × 𝑒ҧ
These properties suggest that equivalent voltage and current can be
introduced proportional to transverse electric and magnetic fields.
 Impedance and Equivalent Voltages and Currents

Equivalent voltage, current, and impedance for waveguides can

be defined in different ways as these quantities are not unique
for non-TEM lines.
The following considerations are usually used:
Voltage and current are defined only for a particular waveguide
mode.
These are defined so that the voltage is proportional to the
transverse electric field and the current is proportional to the
transverse magnetic field.
 Impedance and Equivalent Voltages and Currents

Equivalent voltages and currents should be defined in such a

way that their product gives the power flow of the waveguide
mode
The ratio of the voltage to the current for a single traveling
wave should be equal to the characteristic impedance of the
line.
This impedance may be chosen arbitrarily, but is usually
selected as equal to the wave impedance of the line, or else
normalized to unity.
 Impedance and Equivalent Voltages and Currents
Equivalent voltage and current waves can be written as:

𝑉 𝑧 = 𝑉 + 𝑒 −𝑗𝛽𝑧 + 𝑉 − 𝑒 𝑗𝛽𝑧
𝐼 𝑧 = 𝐼+ 𝑒 −𝑗𝛽𝑧 − 𝐼 − 𝑒 𝑗𝛽𝑧
𝑉+ 𝑉−
𝑍0 = + = −
𝐼 𝐼

(𝑍0 can be made equal to 𝑍𝑤 or normalized to unity)

 N-port Microwave networks
 Let us consider an arbitrary N-port microwave
network as shown.

 The ports may be any type of transmission line or

transmission line equivalent of a single propagating
mode.
𝑡2
 If the physical port of the network is a waveguide 𝑡3
supporting more than one propagating mode, such
modes can be accounted for by considering additional
electrical ports.
𝑡1
 At the nth port, we define a terminal plane 𝑡𝑛 as well
as equivalent voltages and currents for the incident 𝑡𝑁
waves (𝑉𝑛+ ,𝐼𝑛+ ) and (𝑉𝑛− ,−𝐼𝑛− ) for the reflected waves. 𝑉𝑁+ , 𝐼𝑁+ 𝑉𝑁− , −𝐼𝑁−
 The terminal planes are important for providing phase
reference for voltage and current phasors. Fig.1. An arbitrary N-port microwave network
Filter (2 port) Wilkinson power divider (3 port)

Magic T(4 port) Hybrid Coupler (4 port)

 N-port Microwave Networks
At the nth terminal plane,
𝑉𝑛 = 𝑉𝑛+ + 𝑉𝑛− Similarly the admittance matrix 𝑌 can
be defined as.
𝐼𝑛 = 𝐼𝑛+ − 𝐼𝑛−
𝐼1 𝑌11 𝑌12 ⋯ 𝑌1𝑁 𝑉1
The impedance matrix 𝑍 of a microwave ⋯ 𝑌2𝑁
𝐼2 𝑌 𝑌22 𝑉2
network relates this voltage and currents = 21 ⋮ ⋮
⋮ ⋮ ⋮ ⋮
𝐼𝑁 𝑌𝑁1 𝑌𝑁2 ⋯ 𝑌𝑁𝑁 𝑉𝑁
𝑉1 𝑍11 𝑍12 ⋯ 𝑍1𝑁 𝐼1
𝑉2 = 𝑍21 𝑍22 ⋯ 𝑍2𝑁 𝐼2 𝐼 = 𝑌 𝑉
⋮ ⋮ ⋮ ⋮ ⋮ ⋮
𝑉𝑁 𝑍𝑁1 𝑍𝑁2 ⋯ 𝑍𝑁𝑁 𝐼𝑁
Note that:
−1
𝑉 = 𝑍 𝐼 𝑌 = 𝑍
 N-port Microwave Networks
In the same manner,
We note that
𝑉𝑖
𝑍𝑖𝑖 = ቤ 𝐼𝑖
𝐼𝑖 𝐼 𝑌𝑖𝑗 = อ
𝑘 =0,𝑘≠𝑖 𝑉𝑗
Therefore, 𝑍𝑖𝑖 is the input impedance seen at port i 𝐼𝑘 =0,𝑘≠𝑗
when all other ports are open circuited. i.e. 𝑌𝑖𝑗 can be determined by driving port j with
voltage 𝑉𝑗 and short circuiting all other ports and
Similarly, measuring the short circuit current at port i.
𝑉𝑖
𝑍𝑖𝑗 = อ For a reciprocal network, 𝑍 and 𝑌 are symmetric
𝐼𝑗
𝐼𝑘 =0,𝑘≠𝑗 i.e.

is the transfer impedance between port i and j 𝑍 = 𝑍 𝑡

when all other ports are open circuited. and

𝑡
𝑌 = 𝑌

Note that a reciprocal network does not contain active devices, ferrites or plasmas.
 Lossless Network
If the network is lossless
𝑅𝑒 𝑃𝑎𝑣 = 0

1 1 1
𝑃𝑎𝑣 = 𝑉𝑡 𝐼 ∗ = 𝑍 𝐼 𝑡 𝐼 ∗ = 𝐼 𝑡 𝑍 𝐼 ∗
2 2 2
since, 𝑍 𝑡 = 𝑍 for a reciprocal network.

𝑍11 𝑍12 … 𝑍1𝑁 𝐼1∗

1 𝑍21 𝑍22 … 𝑍2𝑁 𝐼2∗
∴ 𝑃𝑎𝑣 = 𝐼1 𝐼2 … 𝐼𝑁 … ⋮
2 ⋮ ⋮ ⋮
𝑍𝑁1 𝑍𝑁2 … 𝑍𝑁𝑁 𝐼𝑁∗

𝑍11 𝐼1∗ + 𝑍12 𝐼2∗ + ⋯ +𝑍1𝑁 𝐼𝑁∗

1 𝑍21 𝐼1∗ + 𝑍22 𝐼2∗ + ⋯ +𝑍2𝑁 𝐼𝑁∗
𝑃𝑎𝑣 = 𝐼1 𝐼2 … 𝐼𝑁
2 ⋮
𝑍𝑁1 𝐼1∗ + 𝑍𝑁2 𝐼2∗ + ⋯ + 𝑍𝑁𝑁 𝐼𝑁∗
 Lossless Network
𝑍11 𝐼1∗ + 𝑍12 𝐼2∗ + ⋯ +𝑍1𝑁 𝐼𝑁∗
1 𝑍21 𝐼1∗ + 𝑍22 𝐼2∗ + ⋯ +𝑍2𝑁 𝐼𝑁∗
𝑃𝑎𝑣 = 𝐼1 𝐼2 … 𝐼𝑁
2 ⋮
𝑍𝑁1 𝐼1∗ + 𝑍𝑁2 𝐼2∗ + ⋯ + 𝑍𝑁𝑁 𝐼𝑁∗

1
= ሾ𝐼1 𝑍11 𝐼1∗ + 𝑍12 𝐼2∗ + ⋯ +𝑍1𝑁 𝐼𝑁∗ + 𝐼2 𝑍21 𝐼1∗ + 𝑍22 𝐼2∗ + ⋯ +𝑍2𝑁 𝐼𝑁
∗
2
 Lossless Network
Since 𝐼𝑛 s are independent, let us set all port Let us now consider that all port currents
currents to zero except for 𝑛𝑡ℎ port current. except 𝐼𝑚 and 𝐼𝑛 are zero. Then

∴ 𝑅𝑒 𝑃𝑎𝑣 = 0 𝑅𝑒 𝐼𝑚 𝑍𝑚𝑛 𝐼𝑛∗ + 𝐼𝑛 𝑍𝑛𝑚 𝐼𝑚

∗ =0

⇒ 𝑅𝑒 𝐼𝑚 𝑍𝑛𝑛 𝐼𝑛∗ =0 ⇒ 𝑅𝑒 𝐼𝑚 𝐼𝑛∗ + 𝐼𝑛 𝐼𝑚

∗
𝑍𝑚𝑛 = 0

∴ 𝐼𝑛 2 . 𝑅𝑒 𝑍𝑛𝑛 = 0 𝐼𝑚 𝐼𝑛∗ + 𝐼𝑛 𝐼𝑚
∗ is real and in general non zero.

∵ 𝐼𝑛 ≠ 0 ∴ 𝑅𝑒 𝑍𝑚𝑛 = 0

𝑅𝑒 𝑍𝑛𝑛 = 0 Therefore, 𝑅𝑒 𝑍𝑚𝑛 is zero for any m and n.

 Therefore, for a lossless network elements of 𝑍 must be purely imaginary.

 In the same manner, elements of 𝑌 can also be shown to be imaginary.

 Examples of Z-matrix for different types of networks
𝑗2 0.2 𝑗3 𝑗2 𝑗0.2 𝑗3
𝑍 = 4 −2 0.5 𝑍 = 𝑗4 −𝑗2 𝑗0.5
𝑗 1 𝑗4 𝑗 𝑗 𝑗4
Neither lossless nor reciprocal Lossless but not reciprocal

𝑗2 4 𝑗3 𝑗2 𝑗4 𝑗3
𝑍 = 4 −𝑗2 𝑗0.5 𝑍 = 𝑗4 −𝑗2 𝑗0.5
𝑗3 𝑗0.5 4 𝑗3 𝑗0.5 𝑗4
Reciprocal but not lossless Both reciprocal and lossless
 Equivalent Voltage and Currents (Example 1)

Let us illustrate how we can find equivalent We have,

voltages and currents for TE10 mode in a 𝑒ҧ 𝑥, 𝑦
ℎത 𝑥, 𝑦 = 𝑎ො𝑧 ×
rectangular waveguide. 𝑍𝑤
where, ℎത and 𝑒ҧ are transverse field
For TE10 mode in a rectangular waveguide, when components and 𝑍𝑤 is the wave impedance.
waves travelling in both +z and –z direction are ∴ We can write
present, the transverse field components can be 1 + −𝑗𝛽𝑧 − 𝑗𝛽𝑧
𝜋𝑥
𝐻𝑥 = − 𝐴 𝑒 +𝐴 𝑒 sin
written as: 𝑍TE 𝑎
𝜋𝑥 −𝑗𝛽𝑧 𝜋𝑥
𝐸𝑦 = 𝐴+ sin 𝑒 + 𝐴− sin 𝑒 𝑗𝛽𝑧 The power for the incident wave is given by
𝑎 𝑎
𝑎 𝑏 + 2
𝜋𝑥 1 𝐴 𝜋𝑥
= 𝐴+ 𝑒 −𝑗𝛽𝑧 + 𝐴− 𝑒 𝑗𝛽𝑧 sin 𝑃+ = න න sin2 𝑑𝑥𝑑𝑦
𝑎 2 0 0 𝑍TE 𝑎

𝑎𝑏
= 𝐴+ 2
4𝑍TE
 Equivalent Voltage and Currents (Example 1)

Writing in the form of equivalent voltages

and currents. We have seen that
𝑉 𝑧 = 𝑉 + 𝑒 −𝑗𝛽𝑧 + 𝑉 − 𝑒 𝑗𝛽𝑧
𝐼 𝑧 = 𝐼 + 𝑒 −𝑗𝛽𝑧 − 𝐼 − 𝑒 𝑗𝛽𝑧 𝑃+ =
𝑎𝑏
𝐴+ 2 1
and 𝑃+ = 𝑉 + 𝐼+∗
1 4𝑍TE 2
= 𝑉 + 𝑒 −𝑗𝛽𝑧 − 𝑉 − 𝑒 𝑗𝛽𝑧 𝑎𝑏 1 + +∗
𝑍0 ∴ 𝐴+ 2 = 𝑉 𝐼
4𝑍TE 2
1 + +∗
∴ 𝑃+ = 𝑉 𝐼 1 +2
2 = 𝐴 𝐶1 𝐶2∗
Let 𝐶1 and 𝐶2 be two constants such that 2
Let 𝑍TE = 𝑍0
𝑉 + = 𝐶1 𝐴+ 𝐼+ = 𝐶2 𝐴+
𝑉 − = 𝐶1 𝐴− 𝐼− = 𝐶2 𝐴− 𝑉 + 𝐶1
∴ + = = 𝑍TE
𝐼 𝐶2
 Equivalent Voltage and Currents (Example 1)

For real 𝐶1 and 𝐶2 , we can solve

𝑎𝑏
𝐶1 =
2
and
1 𝑎𝑏
𝐶2 =
𝑍TE 2

With 𝐶1 and 𝐶2 fount out, we can find 𝑉 + , 𝐼+ , 𝑉 − and 𝐼 − in terms of amplitude 𝐴+ and 𝐴− .
 Equivalent Voltages and Currents (Example 2)
Let us consider a rectangular waveguide of dimensions 𝑎 = 2.5
cm and 𝑏 = 1.25 cm

Let the waveguide be filled with air for 𝑧 < 0 and a dielectric
with 𝜖𝑟 = 2.25 fills 𝑧 > 0. 𝑦

We apply the concept of equivalent transmission line model to

compute the reflection coefficient for 𝑇𝐸10 field incident on the
interface from 𝑧 < 0 . Let the frequency of operation 𝑓 = 7.5 𝑥
GHz.
𝑧
For the given guide dimensions, it may be verified that at 7.5
GHz, only 𝑇𝐸10 mode propagates in both the portions of the
waveguide
 Equivalent Voltages and Currents (Example 2)
Cutoff wavelength for 𝑇𝐸 10 mode in a rectangular waveguide is given by 2𝑎.

𝜋
𝑘𝑐 = 𝑎 = 125.66 m−1 𝑘0 = 2𝜋𝑓 𝜖0 𝜇0 = 157.1 m−1

2 𝜋 2 −1 2 𝜋 2
𝛽𝑎 = 𝑘0 − = 94.3 m and 𝛽𝑑 = 𝜖𝑟 𝑘0 − = 199.35 m−1
𝑎 𝑎

𝑘0 𝜂0 157.1 × 377
𝑍0𝑎 = = = 628.1 Ω
𝛽𝑎 94.3

𝜇
𝜖𝑟 𝑘0 𝜖 𝜖0
𝑘𝜂 𝑟 0 𝑘0 𝜂0 157.1 × 377
𝑍0𝑑 = = = = = 297.1 Ω
𝛽𝑑 𝛽𝑑 𝛽𝑑 199.35

𝑍0𝑑 − 𝑍0𝑎
Γ= = −0.36
𝑍0𝑑 + 𝑍0𝑎
Lecture 2
 Scattering matrix representation
Representation of microwave networks by
impedance or admittance matrix is not very 𝑧0𝑛 is the characteristic impedance of the port n.
convenient as at microwave frequency, the
voltage, current or impedances can not be Let us consider a two port network for which we
measured in a direct manner.
can write:

The quantities that may be measured easily are 𝑏1 = 𝑆11 𝑎1 +𝑆12 𝑎2

reflection coefficient and transmission coefficient. 𝑏2 = 𝑆21 𝑎1 +𝑆22 𝑎2

This forms the basis of scattering matrix 𝑏 = 𝑆 𝑎

formulation.
For a N port network, at the 𝑛𝑡ℎ port let us define
𝑉𝑛+ 𝑉𝑛− 𝑎1 𝑎2
𝑎𝑛 = & 𝑏𝑛 = Two Port
Two
𝑧0𝑛 𝑧0𝑛 𝑏1 𝑏2

Ref: K. C. Gupta, “Microwaves”, New Age Publishers 2002

 Scattering matrix representation
We find that
𝑏1
𝑆11 = ቚ
𝑎1 𝑎2 = 0

𝑉1−
𝑆11 = ฬ+ is the reflection coefficient at port 1 when no voltage is incident at port 2.
𝑉1+ 𝑉2 = 0

Similarly, 𝑆22 is the reflection coefficient at port 2.

𝑆12 & 𝑆21 are transmission coefficients.

The voltage at the 𝑛𝑡ℎ port is given by

𝑉𝑛 = 𝑉𝑛+ + 𝑉𝑛−
= 𝑧0𝑛 𝑎𝑛 + 𝑏𝑛
 Scattering matrix representation

The current at the 𝑛𝑡ℎ port

1
𝐼𝑛 = 𝑉𝑛+ − 𝑉𝑛−
𝑧0𝑛

1
= 𝑎𝑛 − 𝑏𝑛
𝑧0𝑛

Power flow at the 𝑛𝑡ℎ port is given by

1
𝑃𝑛 = Re 𝑉𝑛 𝐼𝑛∗
2

1 ∗
= Re 𝑎𝑛 + 𝑏𝑛 𝑎𝑛 − 𝑏𝑛
2

1
= Re 𝑎𝑛 𝑎𝑛∗ − 𝑏𝑛 𝑏𝑛∗ + 𝑎𝑛∗ 𝑏𝑛 − 𝑏𝑛∗ 𝑎𝑛
2
 Scattering matrix representation
Now , the term 𝑎𝑛∗ 𝑏𝑛 − 𝑏𝑛∗ 𝑎𝑛 is purely
imaginary
1
∴ 𝑃𝑛 = 𝑎𝑛 𝑎𝑛∗ − 𝑏𝑛 𝑏𝑛∗
2
2
1 𝑉𝑛+
𝑎𝑛 𝑎𝑛∗ = is the power carried to the 𝑛𝑡ℎ port by the incident
2 2𝑧0𝑛
wave

1 𝑉𝑛− 2
𝑏𝑛 𝑏𝑛∗ = is the power reflected back from the 𝑛𝑡ℎ port
2 2𝑧0𝑛
 Symmetry of S-matrix
We have seen that for a reciprocal network Z-matrix is symmetric.

∴ 𝑉 = 𝑍 𝐼

Let us assume that all ports have the same characteristics impedance 𝑍0 i.e. 𝑍0𝑛 = 𝑍0

Then we can write

1
𝑉+ + 𝑉− = 𝑍 𝑉+ − 𝑉−
𝑍0
= 𝑍′ 𝑉+ − 𝑉−
′ 𝑍
where 𝑍 =
𝑍0
 Symmetry of S-matrix
Rearranging the terms we can write
𝑍′ + 𝑈 𝑉 − = 𝑍′ − 𝑈 𝑉+
Here, 𝑈 is the unity matrix

when 𝑍0𝑛 = 𝑍0 , 𝑏 = 𝑆 𝑎 can be written as

𝑉− = 𝑆 𝑉+

Also from 𝑍 ′ + 𝑈 𝑉− = 𝑍′ + 𝑈 𝑉+
We can write

𝑉− = 𝑍′ + 𝑈 −1 𝑍′ − 𝑈 𝑉+

Comparing with 𝑉 − = 𝑆 𝑉 +

We can write 𝑆 = 𝑍′ + 𝑈 −1 𝑍′ − 𝑈
 Symmetry of S-matrix
An alternate form for the scattering matrix can be derived as
follows:

At the 𝑛𝑡ℎ port

𝑉𝑛 = 𝑉𝑛+ + 𝑉𝑛−
1
𝐼𝑛 = 𝑉𝑛+ − 𝑉𝑛−
𝑍0

1
∴𝑉𝑛+ = 𝑉𝑛 + 𝑍0 𝐼𝑛
2
−
1
𝑉𝑛 = 𝑉𝑛 − 𝑍0 𝐼𝑛
2
1 1
𝑉+ = 𝑉 + 𝑍0 𝐼 = 𝑍 + 𝑍0 𝑈 𝐼 and
2 2

1 1
𝑉− = 𝑉 − 𝑍0 𝐼 = 𝑍 − 𝑍0 𝑈 𝐼
2 2
 Symmetry of S-matrix
𝐼 = 2 𝑍 + 𝑍0 𝑈 −1 𝑉+

∴ 𝑉 − = 𝑍 − 𝑍0 𝑈 𝑍 + 𝑍0 𝑈 −1 𝑉+

𝑉 − = 𝑍′ − 𝑈 𝑍′ + 𝑈 −1 𝑉+

∴ 𝑆 = 𝑍′ − 𝑈 𝑍′ + 𝑈 −1

From our earlier derivation we have

𝑆 = 𝑍 ′ + 𝑈 −1 𝑍 ′ − 𝑈

Since 𝑍 ′ & 𝑈 are symmetrical matrices

𝑍′ + 𝑈 −1 𝑡 = 𝑍′ + 𝑈 −1 & 𝑍′ − 𝑈 𝑡 = 𝑍′ − 𝑈

∴ 𝑆 = 𝑆 𝑡
 Symmetry of S-matrix
When the port impedances 𝑍0𝑛 are all different we can write

𝑉ഥ𝑛 = 𝑉𝑛+ + 𝑉𝑛− Τ 𝑧0𝑛 = 𝑉𝑛 Τ 𝑧0𝑛 = 𝑎𝑛 + 𝑏𝑛

𝐼ഥ𝑛 = 𝑉 + − 𝑉 − Τ 𝑧0𝑛 = 𝐼𝑛 𝑧0𝑛 = 𝑎𝑛 − 𝑏𝑛

We define,

𝑧01 0 ⋯ 0
𝑧0𝑛 = ⋮ 𝑧02 ⋱ ⋮ and
0 ⋯ 𝑧0𝑁

1
0⋯ 0
𝑧01
1 1
= ⋮ ⋱ ⋮
𝑧0𝑛 𝑧02
1
0 ⋯
𝑧0𝑁
 Symmetry of S-matrix
Using these two matrices we can write
𝑉 = 𝑧0𝑛 𝑉ത

1
𝐼 = 𝐼ҧ
𝑧0𝑛

∴ 𝑉 = 𝑍 𝐼 can be written as

1
𝑧0𝑛 𝑉ത = 𝑍 𝐼ҧ
𝑧0𝑛

−1 1
∴ 𝑉ത = 𝑧0𝑛 𝑍 𝐼ҧ
𝑧0𝑛

= 𝑍ҧ 𝐼 ҧ

1
𝑍ҧ = 𝑧0𝑛 −1 𝑍 is symmetric
𝑧0𝑛
 Symmetry of S-matrix
𝑎 + 𝑏 = 𝑉ത
𝑎 − 𝑏 = 𝐼ҧ

𝑎 + 𝑏 = 𝑉ത = 𝑍ҧ 𝐼 ҧ
= 𝑍ҧ 𝑎 − 𝑏 Since 𝑈 and 𝑍ҧ are symmetric, 𝑆 is symmetric.

∴ 𝑏 + 𝑍ҧ 𝑏ത = 𝑍ҧ 𝑎 − 𝑎 Thus, for any linear, and reciprocal network (i.e.

where 𝑍 matrix is symmetric) , the scattering
matrix 𝑆 is symmetric.
∴ 𝑈 + 𝑍ҧ 𝑏ത = 𝑍ҧ − 𝑈 𝑎

∴ 𝑏ത = 𝑈 + 𝑍ҧ −1 𝑍ҧ − 𝑈 𝑎

∴ 𝑆 = 𝑍ҧ + 𝑈 −1 𝑍ҧ − 𝑈
 Scattering Matrix for a Lossless Junction
1
We have seen that 𝑃𝑛 = 𝑎𝑛 𝑎𝑛∗ − 𝑏𝑛 𝑏𝑛∗
2
For a lossless junction total power leaving the all 𝑁 ports must be equal to
sum of the incident powers.
Therefore, σ𝑁 𝑏𝑛 𝑏𝑛∗ = σ𝑁 𝑎𝑛 𝑎𝑛∗
Since, 𝑏 = 𝑆 𝑎
𝑁

𝑏𝑛 = ෍ 𝑆𝑛𝑖 𝑎𝑖
𝑖=1
Therefore, we can write
2
𝑁 𝑁 𝑁

෍ ෍ 𝑆𝑛𝑖 𝑎𝑖 = ෍ 𝑎𝑛 𝑎𝑛∗
𝑛=1 𝑖=1 𝑛=1
 Scattering Matrix for a Lossless Junction
Since 𝑎𝑛 parameters are independent, if we set all the 𝑎𝑛 except 𝑎𝑖 to be zero.
𝑁 𝑁 2
For this the equation σ𝑛=1 σ𝑖=1 𝑆𝑛𝑖 𝑎𝑖 = σ𝑁 𝑎 𝑎
𝑛=1 𝑛 𝑛
∗ reduces to:
𝑁

෍ 𝑆𝑛𝑖 𝑎𝑖 2 = 𝑎𝑖 𝑎𝑖∗ = 𝑎𝑖 2

𝑛=1
Therefore,
𝑁 𝑁
2 ∗
෍ 𝑆𝑛𝑖 = ෍ 𝑆𝑛𝑖 𝑆𝑛𝑖 =1
𝑛=1 𝑛=1
The dot product of any column of matrix 𝑆 with the conjugate of that same
column gives unity.
 Scattering Matrix for a Lossless Junction
Let us now consider another property of scattering parameters of a
lossless junction.
Let us set all the 𝑎𝑛 except 𝑎𝑠 and 𝑎𝑟 to be zero.
𝑁 2
For this the equation σ𝑁
𝑛=1 σ𝑖=1 𝑆𝑛𝑖 𝑎𝑖 = σ𝑁 𝑎 𝑎 ∗
𝑛=1 𝑛 𝑛 reduces to:
𝑁
2
෍ 𝑆𝑛𝑠 𝑎𝑠 + 𝑆𝑛𝑟 𝑎𝑟 = 𝑎𝑠 𝑎𝑠∗ + 𝑎𝑟 𝑎𝑟∗ = 𝑎𝑠 2
+ 𝑎𝑟 2

𝑛=1
𝑁
∗ 2 2
෍ 𝑆𝑛𝑠 𝑎𝑠 + 𝑆𝑛𝑟 𝑎𝑟 𝑆𝑛𝑠 𝑎𝑠 + 𝑆𝑛𝑟 𝑎𝑟 = 𝑎𝑠 + 𝑎𝑟
𝑛=1
 Scattering Matrix for a Lossless Junction
Since 𝑎𝑠 and 𝑎𝑟 are independent, if We get
we chose 𝑎𝑠 = 𝑎𝑟 from 𝑁
𝑁
∗ ∗
∗ 𝑎∗ ∗ 𝑎∗ ෍ 𝑆𝑛𝑠 𝑆𝑛𝑟 − 𝑆𝑛𝑟 𝑆𝑛𝑠 =0
෍ 𝑆𝑛𝑟 𝑎𝑟 𝑆𝑛𝑠 𝑠 + 𝑆𝑛𝑠 𝑎𝑠 𝑆𝑛𝑟 𝑟 =0
𝑛=1
𝑛=1
Since neither 𝑎𝑠 nor 𝑎𝑟 is zero, the
we get above two conditions can be
𝑁
satisfied only if
∗ ∗
෍ 𝑆𝑛𝑠 𝑆𝑛𝑟 + 𝑆𝑛𝑟 𝑆𝑛𝑠 =0 𝑁
∗ =0
𝑛=1 ෍ 𝑆𝑛𝑠 𝑆𝑛𝑟 𝑠≠𝑟
If, instead, we chose 𝑎𝑠 = 𝑗𝑎𝑟 with 𝑛=1
𝑎𝑟 real, from
𝑁
∗ 𝑎∗ ∗ 𝑎∗
Therefor, the dot product of a column
෍ 𝑆𝑛𝑟 𝑎𝑟 𝑆𝑛𝑠 𝑠 + 𝑆𝑛𝑠 𝑎𝑠 𝑆𝑛𝑟 𝑟 =0 of the scattering matrix of a lossless
𝑛=1 junction with the complex conjugate
of any other column is zero.
 Scattering Matrix for a Lossless Junction
We can also prove the properties of the S matrix for a lossless
junction following a different approach as follows:
We have,
෍ 𝑏𝑛 𝑏𝑛∗ = ෍ 𝑎𝑛 𝑎𝑛∗
𝑁 𝑁
Therefore we can write
𝑏𝑡 𝑏∗= 𝑎𝑡 𝑎∗
𝑡
𝑎 𝑆𝑡 𝑆∗𝑎∗= 𝑎𝑡 𝑎 ∗

𝑆𝑡 𝑆∗= 𝑈
 Scattering Matrix for a Lossless Junction
𝑁
∗ 2 2
෍ 𝑆𝑛𝑠 𝑎𝑠 + 𝑆𝑛𝑟 𝑎𝑟 𝑆𝑛𝑠 𝑎𝑠 + 𝑆𝑛𝑟 𝑎𝑟 = 𝑎𝑠 + 𝑎𝑟
𝑛=1

𝑁 𝑁 𝑁
2 ∗ ∗ ∗ ∗ ∗ 2 ∗ 2 2
𝑎𝑠 ෍ 𝑆𝑛𝑠 𝑆𝑛𝑠 + ෍ 𝑆𝑛𝑟 𝑎𝑟 𝑆𝑛𝑠 𝑎𝑠 + 𝑆𝑛𝑠 𝑎𝑠 𝑆𝑛𝑟 𝑎𝑟 + 𝑎𝑟 ෍ 𝑆𝑛𝑟 𝑆𝑛𝑟 = 𝑎𝑠 + 𝑎𝑟
𝑛=1 𝑛=1 𝑛=1

𝑁
∗ 𝑎 ∗ + 𝑆 𝑎 𝑆 ∗ 𝑎 ∗ =0
෍ 𝑆𝑛𝑟 𝑎𝑟 𝑆𝑛𝑠 𝑠 𝑛𝑠 𝑠 𝑛𝑟 𝑟
𝑛=1
 Scattering Matrix for a Lossless Junction
𝑡 ∗
𝑆 𝑆 = 𝑈 can be written as
𝑁
∗
෍ 𝑆𝑘𝑖 𝑆𝑘𝑗 = 𝛿𝑖𝑗
𝑘=1
𝛿𝑖𝑗 = 1 for 𝑖 = 𝑗 and 𝛿𝑖𝑗 = 0 for 𝑖 ≠ 𝑗
Therefore,
𝑁
∗
෍ 𝑆𝑘𝑖 𝑆𝑘𝑖 =1
𝑘=1
and

𝑁
∗
෍ 𝑆𝑘𝑖 𝑆𝑘𝑗 =0
𝑘=1
for 𝑖 ≠ 𝑗
 Example-1
Example 1: Scattering matrix for the junction of two transmission
lines
𝑆21
𝑏1 𝑆11 𝑆12 𝑎1
We have =
𝑏2 𝑆21 𝑆22 𝑎2
𝑍1 𝑆12 𝑍2
𝑏1
𝑆11 = ቚ Since 𝑎2 = 0, we have 𝑉2+ = 0
𝑎1 𝑎 =0
2 𝑆11 𝑆22
𝑉1−
𝑆11 = +ቤ
𝑉1 𝑉 + =0
2
Therefore, at the junction we can write
1 𝑉 −
𝑉1+ + 𝑉1− = 𝑉2− and 𝑉1+ − 𝑉1− = 2
𝑍1 𝑍2
 Example-1
𝑍2
Therefore, 𝑉1+ + 𝑉1− = 𝑉1+ − 𝑉1− 𝑉1+ + 𝑉1− = 𝑉2−
𝑍1
𝑍1 −
𝑉1− 𝑍2 −𝑍1 + −
𝑉1 − 𝑉1 = 𝑉2
and 𝑆11 = ฬ = 𝑍2
𝑉1+ 𝑉 + =0 𝑍2 +𝑍1
2
𝑍1
In the same manner, Therefore, 2𝑉1+ = 1+ 𝑉2−
𝑍2
𝑉2− 𝑍1 −𝑍2 𝑉2− 2𝑍2
𝑆22 = ฬ = = −𝑆11 +ቤ =
𝑉2+ + 𝑍2 +𝑍1 𝑉1 𝑉 +=0 𝑍1 + 𝑍2
𝑉1 =0
2
𝑏2 𝑍1 2𝑍2 2 𝑍1 𝑍2
𝑆21 = ቤ 𝑆21 = =
𝑎1 𝑎
2 =0 𝑍2 𝑍1 + 𝑍2 𝑍1 + 𝑍2
𝑍1 𝑉2− 𝑆12 = 𝑆21
𝑆21 = +ቤ
𝑍2 𝑉1 𝑉 + =0 2
 Example 2
In this example we find the reflection coefficient at the port 1 of a two-port
for which the port 2 is terminated to a short circuit
Suppose the S parameters for the given two-port be
𝑏1 𝑆11 𝑆12 𝑎1
=
𝑏2 𝑆21 𝑆22 𝑎2
When port two is short circuited, 𝑉2+ + 𝑉2− =0 . Therefore,
𝑎2 = −𝑏2
𝑏2 = 𝑆21 𝑎1 + 𝑆22 𝑎2 = −𝑎2
𝑆21
𝑎2 = − 𝑎1
1 + 𝑆22
𝑉1− 𝑏1 𝑆21
Γ= += = 𝑆11 −
𝑉1 𝑎1 1 + 𝑆22
Lecture 3
 Properties of 3-port
Using properties of S-parameters, we show that a three port network cannot be lossless, reciprocal and
matched at all three ports.
If possible, let a three port network be lossless, reciprocal and matched at all three ports.

𝑆11 𝑆12 𝑆13 0 𝑆12 𝑆13

𝑆 = 𝑆21 𝑆22 𝑆23
When ports are matched
𝑆 = 𝑆21 0 𝑆23
𝑆31 𝑆32 𝑆33 𝑆31 𝑆32 0

0 𝑆12 𝑆13
𝑆 = 𝑆12 0 𝑆23
Reciprocal Network
𝑆13 𝑆23 0
 Properties of 3-port

By applying the properties of S-parameters, we can write

∗
𝑆13 𝑆23 =0
∗
To satisfy this either 𝑆13 = 0 or 𝑆23 = 0 or both equal to zero
Let 𝑆13 ≠ 0. Then, 𝑆23 = 0
 Properties of 3-port
Further, we have 𝑆12 2 = 1 − 𝑆13 2 = 1 − 𝑆23 2
To satisfy this, 𝑆13 = 𝑆23 . Therefore, 𝑆13 = 0 which contradicts our
earlier assumption that 𝑆13 ≠ 0.
Therefore, we do not obtain consistent solution from the given 𝑆 .
0 𝑆12 𝑆13
Therefore, 𝑆 = 𝑆12 0 𝑆23
𝑆13 𝑆23 0

is not valid and we cannot have a three port network which is

lossless, reciprocal and at the same time matched at all three ports.
Let us now drop the condition that 𝑆 is reciprocal
 Properties of 3-port
0 𝑆12 𝑆13 ∗
𝑆31 𝑆32 = 0
∗
𝑆 = 𝑆21 0 𝑆23 𝑆12 𝑆13 = 0
𝑆31 𝑆32 0 ∗
𝑆23 𝑆21 = 0

2 2 These sets of equations are satisfied by:

𝑆21 + 𝑆31 =1
2 2
𝑆12 + 𝑆32 = 1 𝑆 = 𝑆 = 𝑆 = 0 and 𝑆 = 𝑆 = 𝑆 = 1
2 2 12 23 31 21 32 13
𝑆13 + 𝑆23 =1
and
𝑆13 = 𝑆32 = 𝑆21 = 0 and 𝑆31 = 𝑆23 = 𝑆12 = 1
 Properties of 3-port
The three port satisfying the condition The three port satisfying the condition
𝑆12 = 𝑆23 = 𝑆31 = 0 and 𝑆13 = 𝑆32 = 𝑆21 = 0 and
𝑆21 = 𝑆32 = 𝑆13 = 1 𝑆31 = 𝑆23 = 𝑆12 = 1
will give rise to a circulator which will provide will give rise to a circulator which will provide
clockwise circulation. With proper choice of phase counterclockwise circulation. With proper choice of
references phase references

0 0 1 0 1 0
𝑆 = 1 0 0 𝑆 = 0 0 1
0 1 0 1 0 0

2
2
1 1

3
3
 Change of reference plane
𝑆′
𝑒 −𝑗𝜃1 … 0 𝑒 −𝑗𝜃1 … 0
= ⋮ ⋱ ⋮ 𝑆 ⋮ ⋱ ⋮
0 … 𝑒 −𝑗𝜃𝑁 0 … 𝑒 −𝑗𝜃𝑁 Port-1

𝜃𝑛 = 𝛽𝑛 𝑙𝑛 𝑧1 =𝑙1 𝑧1 =0 N-port
network

Port-N

𝑧𝑁 =𝑙𝑁 𝑧𝑁 =0
 Transmission Matrix (ABCD Parameters)
For a two-port network, ABCD
parameters are defined as
𝑉1 𝐴 𝐵 𝑉2 𝐼1 𝐼2
=
𝐼1 𝐶 𝐷 𝐼2
+
𝐴 𝐵
+
𝑉1 𝑉2
𝐶 𝐷
The parameter 𝐴 can be evaluated as : - -
𝑉1
𝐴= ቤ Port-1 Port-2
𝑉2 𝐼 =0
2
𝐵, 𝐶 & 𝐷 can be found in the same
manner.
 Transmission Matrix for a cascade
𝐼1 𝐼2 𝐼3
+ + +
𝐴1 𝐵1 𝐴2 𝐵2
𝑉1 𝑉2 𝑉3
𝐶1 𝐷1 𝐶2 𝐷2
- - -

For the cascade shown above

𝑉1 𝐴1 𝐵1 𝐴2 𝐵2 𝑉3
=
𝐼1 𝐶1 𝐷1 𝐶2 𝐷2 𝐼3
 Examples: ABCD Parameters for some 2-ports
For the two port shown, when 𝐼2 = 0,
𝐼1 = 0 and 𝑉2 = 𝑉1
𝑉1 𝐼1
𝐴= ቚ = 1 and 𝐶 = ቚ =0
𝑉2 𝐼 =0 𝑉2 𝐼 =0
2 2 𝑍
When 𝑉2 = 0, 𝐼1 = 𝐼2
𝑉1 𝑉1 𝐼1 𝐼1
𝐵= ቚ = = 𝑍 and 𝐷 = ቚ = =1
𝐼2 𝑉 =0 𝐼1 𝐼2 𝑉 =0 𝐼1
2 2
Therefore,
𝐴 𝐵 1 𝑍
=
𝐶 𝐷 0 1
 Examples: ABCD Parameters for some 2-ports
In this example our two-port is a section of
transmission line as shown
For a transmission line, we have Port-1 𝑍0 , 𝛽 Port-2

𝑉 𝑍 = 𝑉 + 𝑒 −𝑗𝛽𝑧 + 𝑉 − 𝑒 𝑗𝛽𝑧 𝑙
1
𝐼 𝑍 = 𝑉 + 𝑒 −𝑗𝛽𝑧 − 𝑉 − 𝑒 𝑗𝛽𝑧
𝑍0
Let us set 𝑧 = 0 at port-2. Therefore, 𝑉2 = Further,
+ −
𝑉 + 𝑉 and 𝐼2 =
𝑍0
1
𝑉+ − 𝑉− 𝑉1 = 𝑉 + 𝑒 𝑗𝛽𝑙 + 𝑉 + 𝑒 −𝑗𝛽𝑙
𝑉1 Therefore,
𝐴= ቤ
𝑉2 𝐼
2 =0
When 𝐼2 =0, 𝑉2 = 2𝑉 + 𝐴 = cos 𝛽𝑙
 Examples: ABCD Parameters for some 2-ports
𝐼1 1
𝐶= ቚ We have, 𝐼1 = 𝑉 + 𝑒 𝑗𝛽𝑙 − 𝑉 + 𝑒 −𝑗𝛽𝑙
𝑉2 𝐼 =0 𝑍0
2
1
Therefore, 𝐶 = 𝑒 𝑗𝛽𝑙 − 𝑒 −𝑗𝛽𝑙 = 𝑗𝑌0 sin 𝛽𝑙
2𝑍0
In the same manner, it can be shown that
𝐵 = 𝑗𝑍0 sin 𝛽𝑙 and 𝐷 = cos 𝛽𝑙
Therefore,
𝐴 𝐵 cos 𝛽𝑙 𝑗𝑍0 sin 𝛽𝑙
=
𝐶 𝐷 𝑗𝑌0 sin 𝛽𝑙 cos 𝛽𝑙
Examples: ABCD Parameters for some 2-ports

𝐴 𝐵 1 0
= Y
𝐶 𝐷 𝑌 1

N:1

𝑁 0
𝐴 𝐵 1
=
𝐶 𝐷 0
𝑁
 ABCD Parameters’ relation with S-parameters
S-parameters of a two-port can be 𝐼1 𝐼2
related to ABCD parameters. We + +
𝐴 𝐵
assume that both ports have 𝑉1 𝑉2
𝐶 𝐷
characteristics impedance 𝑍0 . - -
Port-1 Port-2
𝑉1−
𝑆11 = +ቤ
𝑉1 𝑉 + =0 𝑉2 = 𝑉2+ + 𝑉2−
2
1 +
𝑉1 = 𝑉1+ + 𝑉1− −𝐼2 = 𝑉2 − 𝑉2−
1 𝑍0
𝐼1 = 𝑉1+ − 𝑉1− 𝑉1 = 𝐴𝑉2 + 𝐵𝐼2
𝑍0
When, 𝑉2+ =0, 𝑉2 = 𝑉2−
 ABCD Parameters’ relation with S-parameters
𝐷
𝐼1 = 𝐶𝑉2 + 𝐷𝐼2 = 𝐶 + 𝑉2−
𝑍0
𝐵
𝑉1 = 𝐴𝑉2 + 𝐵𝐼2 . Therefore, 𝑉1+ + 𝑉1− = 𝐴 + 𝑉2−
𝑍0

𝐵 𝐵 𝐵
𝐴+𝑍 𝐴+𝑍 1 𝐴+𝑍
𝑉1+ + 𝑉1− = 𝐷
0
𝐼1 = 𝐷
0
𝑉1+ − 𝑉1− = 0
𝑉1+ − 𝑉1−
𝐶+𝑍 𝐶+𝑍 𝑍0 𝐶𝑍0 +𝐷
0 0

𝐵 𝐵 𝐵
𝑉1− 𝐴+ 𝐴+ 𝐴 + − 𝐶𝑍0 − 𝐷
𝑍0 𝑍0 𝑍0
𝑆11 = +ቤ = −1 ൚ +1 =
𝑉1 𝑉 + =0 𝐶𝑍0 + 𝐷 𝐶𝑍0 + 𝐷 𝐵
𝐴 + + 𝐶𝑍0 + 𝐷
2
𝑍0
 ABCD Parameters’ relation with S-parameters

All the four S parameters may be related to ABCD parameters

as follows:

𝐵
𝐴+ −𝐶𝑍0 −𝐷 2 𝐴𝐷−𝐵𝐶
𝑍0
𝑆11 = 𝐵 𝑆12 = 𝐵
𝐴+ +𝐶𝑍0 +𝐷 𝐴+ +𝐶𝑍0 +𝐷
𝑍0 𝑍0
𝐵
2 −𝐴+ −𝐶𝑍0 +𝐷
𝑍0
𝑆21 = 𝐵 𝑆22 = 𝐵
𝐴+ +𝐶𝑍0 +𝐷 𝐴+ +𝐶𝑍0 +𝐷
𝑍0 𝑍0