7/3/2020 sparam

www.waves.utoronto.ca/prof/svhum/ece422/labs/sdr/sparam.html 1/8
7/3/2020 sparam

Introduction to s-parameters and ABCD Matrices Sub-

module
Representation of Two-Port Networks
Two-port networks are used to model the characteristics of any two-terminal device. In ECE422, we use them
to model 2-antenna links, and also, the components of a receiver which the present module is focused on. As
discussed in class, two-port networks can be represented using a variety of approaches. For example, the
impedance matrix is a common approach. As we will see, for characterizing RF circuits composing the RX-III
receiver, other matrix representations of two-port networks are more appropriate.

A general two-port network is shown in the figure below.

In class, we described the impedance matrix which allows the voltages V1 and V2 circuit to be found knowing
the currents I1 and I2 :

V1 z11 z12 I1
[ ] = [ ][ ]
V2 z21 z22 I2

Or simply,

[V ] = [Z ][I ] .

The behaviour of the circuit is completely described by the impedance matrix [Z ] . Elements of the matrix can
be found as follows. For example, by definition,

V1
z11 = ∣
I1
∣
I2 =0

That is, we leave port 2 open-circuit while injecting a fixed current I1 into port 1. We measure the
corresponding voltage V1 and take the ratio of V1 /I1 to obtain Z11 . Similarly,

V2
z21 = ∣
I1
∣
I2 =0

requires the same setup except that the voltage V2 is measured instead of V1 .

In practice, characterizing the z -parameters of a radio frequency circuit is difficult using this method. There are
a few reasons for this:

1. We need to inject currents into the network to excite it, and this is done by attaching transmission
lines to port 1 and port 2 connected to respective sources and loads. As we know, transmission lines
www.waves.utoronto.ca/prof/svhum/ece422/labs/sdr/sparam.html 2/8
7/3/2020 sparam

change the phase of the applied signal because of the finite time delay associated with the
transmission ine, and this phase is frequency-dependent. Hence, a way is needed to work out the
currents I1 and I2 when transmission lines are used.
2. Transmission lines transform load impedances into other impedances. Therefore, if we are trying to
impose an open circuit at one of the two ports by open circuiting a transmission line connected to
those ports, the open circuit at the end of the line is getting translated to some other reactance at the
terminals of the two-port network.
3. z -parameters do not directly tell us about the reflection and transmission characteristics of the two-
port network. If we are trying to characterize a filter, for example, we cannot infer the transfer function
of the filter very easily from the z -parameters of the filter.

For this reason, an alternative representation of two-port networks is desired. The matrix we will use instead is
a scattering matrix.

Scattering Parameters

www.waves.utoronto.ca/prof/svhum/ece422/labs/sdr/sparam.html 3/8
7/3/2020 sparam

We now use a scattering matrix to describe a two-port network instead of an impedance matrix. Instead of
using voltages and currents, we use travelling voltage waves when analyzing two-port networks in terms of
scattering parameters. Consider the following diagram:

Here, V1+ and V1− are the forward and reverse traveling waves of a voltage signal associated with port 1.
Likewise for port 2. There travelling waves are the same kinds of voltage traveling waves we have studied in
transmission lines in ECE320 / ECE357. We now describe the behaviour of the two-port network in terms of
scattering matrix as follows:
− +
V s11 s12 V
1 1
[ ] = [ ][ ]
− +
V s21 s22 V
2 2

or simply

[V
−
] = [S ][V
+
] .

Hence, the independent variables are the voltages entering the two-port while the dependent variables are
those leaving the two-port. A specific element of the scattering matrix can be determined as
−
V ∣
.
i
sij = +
∣
V
j ∣V +
=0 for k≠j
k

For example,
−
V
∣
s11 =
1

+
.
V ∣ +
1 V =0
2

Two establish V2+ = 0 requires that a matched load (Z = Z0 ) be placed at port 2 when port 1 is excited,
since such a load will prevent the voltage wave V2− leaving port 2 from being reflected back in and producing
≠ 0 . Similarly,
+
V
2

−
V
∣
s21 =
2

+
.
V ∣ +
1 V =0
2

also employs a matched load placed at port 2 when port 1 is excited.

Scattering parameters address all of the shortcomings identified with the impedance matrix identified earlier.
Consider the first two issues, which are related to the connection of transmission lines with characteristic
impedance Z0 to the ports of the two-port network. This situation is depicted in the figure below.

www.waves.utoronto.ca/prof/svhum/ece422/labs/sdr/sparam.html 4/8
7/3/2020 sparam

If the transmission lines are lossless, and matched to the source impedance (Zs = Z0 ), then we know that
the individual travelling waves are simply related through

Vn
′+
= Vn e
+ jθ n
,

,
′− + −jθ n
Vn = Vn e

where θn is the phase delay associated with the line,

θ n = β n ℓn .

βn is the phase constant of line n, and ℓn is the length of line n. From this, it is easy to find that

−jθ 1 −jθ 1
e 0 e 0
′
[S ] = [ ] [S ] [ ]
−jθ 2 −jθ 2
0 e 0 e

This means that we can attach sources and make measurements at the ends of the lines (effectively
measuring [S']) and then do this simple mathematical operation to extract [S ] . This process is called de-
embedding [S ] . It requires that the line lengths / phase delays be measured ahead of time; this process is
carried out during a calibration measurement.

The third issue with z -parameters was that we do not get a direct measure of the reflection and transmission
properties of the two-port network. By definition, s -parameters are these quantities. Specifically, sii
correspond to reflection coefficients seen looking into port i , while sij correspond to transmission coefficients
from port j to port i . Hence we see that s -parameters are very versatile for RF measurements.

Scattering parameters are measured by an instrument known as a vector network analyzer (VNA), which you will
employ in the lab. As the name suggests, it can be used to extract the behaviour of networks. Most network
analyzers are 2-port network analyzers, for characterizing 1- or 2-port networks, but there are variants with 4, 8,
12, and even more ports.

An important part of using a VNA is a calibration phase, whereby the transmission lines connected to the front
panel of the VNA, as well as other sources of error, are measured and calibrated out. This is done by connecting
known calibration standards to the ends of the transmission lines and carrying out measurements of those
standards. During this phase, the values of θ1 and θ2 are determined, among other things, allowing for the
transmission lines to be removed from the measurement and the two-port network characteristics properly de-
embedded from the measured s -parameters. You will learn more about the use and calibration of a VNA in a
separate note (vna.html).

www.waves.utoronto.ca/prof/svhum/ece422/labs/sdr/sparam.html 5/8
7/3/2020 sparam

Determining the Scattering Matrix for some Common

Two-Port Networks
While we can determine the scattering parameters of a two-port network from first principles using the theory
outline in the previous section, it is often easier to do this by using the so-called transmission or ABCD matrix of
the two-port network. This is yet another matrix representation of two-port networks. Consider the following
diagram of a two-port network:

Note that the direction of I2 has been reversed, for reasons that will be clear in a moment. The ABCD
parameters of the matrix are defined as follows:

V1 A B V2
[ ] = [ ][ ]
I1 C D I2

Consider a so-called cascade connection of two two-port networks as shown in the following figure.

Here we can see that I2 flows out of network 1 but into network 2, so these currents are one in the same. This is
the reason that the sign of the current on the output was reversed. Furthermore, we can write the following
relations. For network 1,

V1 A1 B1 V2
[ ] = [ ][ ]
I1 C1 D1 I2

For network 2,

V2 A2 B2 V3
[ ] = [ ][ ]
I2 C2 D2 I3

Substituting the second equation into the first yields

V1 A1 B1 A2 B2 V3
[ ] = [ ][ ][ ] ,
I1 C1 D1 C2 D2 I3

www.waves.utoronto.ca/prof/svhum/ece422/labs/sdr/sparam.html 6/8
7/3/2020 sparam

which shows that the total ABCD response of this cascaded connection of two networks can be simply found by
multiplying to two ABCD matrices associated with each constituent network. This is the main usefulness of the
ABCD matrix, because a complex two-port network can be built up from simpler ones where the ABCD matrices
are known.

The ABCD matrices of basic circuit elements can be easily derived. Here, we highlight two important results: the
ABCD matrix for a series impedance, and the ABCD matrix for a shunt admittance.

For the series impedance shown in the figure below, it can easily be derived that

A = 1 ,

B = Z ,

C = 0 ,

D = 1 .

Similarly, for the shunt admittance,

A = 1 ,

B = 0 ,

C = Y ,

D = 1 .

Knowing the ABCD matrices of these simple networks allows more complex networks to be formed. For
example, it is now easy to determine the response of a π network shown below, by cascading a shunt-Y network
with a series-Z network, followed by a shunt-Y network. The corresponding ABCD matrix of the cascade
connection is simply

1 Z 1 0 1 Z
[ ][ ][ ] .
0 1 Y 1 0 1

www.waves.utoronto.ca/prof/svhum/ece422/labs/sdr/sparam.html 7/8
7/3/2020 sparam

Computing S-parameters from ABCD Matrices

It is possible to convert between different matrix representations of two-port networks. For the present case, we
are interested in finding scattering parameters from ABCD matrices. The corresponding conversion formulae are
as follows:

A+B/Z −C Z −D
0 0
s11 =
A+B/Z +C Z +D
0 0

2(AD−BC )
s12 =
A+B/Z +C Z +D
0 0

2
s21 =
A+B/Z0 +C Z0 +D

−A+B/Z0 −C Z0 +D
s22 =
A+B/Z0 +C Z0 +D

Going Further
This has been a basic introduction to scattering parameters and various matrix representations of two-port
networks. This topic is covered in much more depth in RF electronics and microwave engineering courses,
namely, ECE424. Interested readers can also the following reference.

D. M. Pozar, "Microwave Engineering", 3rd edition, John Wiley and Sons, 2005, Chapter 4.

www.waves.utoronto.ca/prof/svhum/ece422/labs/sdr/sparam.html 8/8