Introduction to S-Parameters

Maulik Mehta

Abstract: This paper mainly talks about the S- node and current in the branches. Each
(scattering) parameters. Microwave transistor parameter set is related to a set of four variables
technology is continually pushing maximum associated with the two-port model. Two of
operating frequencies upward [1]. Experiencing these variables represent the excitation of the
the practical difficulty in finding the AHYZ network (independent variables), and the
parameters at high frequencies to know about remaining two represent the response of the
network, S parameters are introduced. As a network to the excitation (dependent variables)
[2]
result, manufacturers of transistors are .The only difference in the parameter sets is
specifying their transistors in terms of S- the choice of independent and dependent
parameters. Many electrical properties of variables[1]. All parameter sets contain the same
networks or components may be expressed using Information about a network and it is always
S-parameters such as gain, return loss, voltage possible to calculate any set in terms of any
standing wave ratio (VSWR), reflection other set. All parameters (A, Z, Y, H
coefficient and amplifier stability. parameters) use lumped element model in which
voltages and currents are assumed to be uniform
Two Port Network: in each node and in each branch accordingly.
Figure 1 shows a two-port network. You can And connecting wires are assumed to be perfect
treat any circuit as a two-port network if you can conductors. The lumped element model is valid
select two pairs of nodes (for example, an input whenever Lc < < λ, where Lc denotes the
pair and an output pair) such that the current into circuit's characteristic length, and λ denotes the
the positive node of a pair equals the current out circuit's operating wavelength [3]. Measurement
of the negative node of the same pair—that is, i1 of all the parameters require opening and
must equal i A and i2 must equal iB in Figure 1. shorting of the input and output port. All these
parameters give accurate result of the network.
At higher frequency lumped element model is
not that accurate and connection wires are not
assumed to be pure conductors. Attributes of the
circuit like resistance, capacitance and
inductance are distributed across the whole
geometry so voltages at each node and currents
in each branch are assumed to be non uniform as
they positive relative to wavelength [3]. All other
Figure 1. Two port network. parameters just relate voltages and currents at
each node.
A two-port network can be represented by a 2-
by-2 matrix. (An n-port network, having n pairs Moving to higher and higher frequencies, some
of nodes, can be represented by an n-by-n Problems arise [1]:
matrix). This matrix can be replaced by any type 1. Equipment is not readily available to measure
of parameter. In Figure 1a, the xij terms (where total voltage and total current at the ports of the
xij represents the value at row i column j) stand network.
in for impedance parameters (Z-parameters), 2. Short and open circuits are difficult to achieve
admittance parameters (Y-parameters), hybrid over a broad band of frequencies as we have to
parameters (h-parameters), chain parameters (A- tune the stub at particular measurement
, B-, C-, and D-parameters), or S-parameters. frequency each time. And tuning stub shunting
the input or output may cause a transistor to
Chain parameters, Z parameters, Y parameters, oscillate, making the measurement invalid.
H parameters are deal with the voltages at the
3. At higher frequencies, active device have to
be stable when open circuit and short circuit is
done.
To overcome these problems distributed element
model is used and transmission line is the
common example of the distributed element
model. The logical variables to use at these
The linear equations describing the two port
frequencies are traveling waves rather than total
network are,
voltages and currents.
(1)
S – Parameters: (2)
S-parameters are measured by sending a single
frequency signal into the input port and
The S-matrix for an N-port contains N2
detecting what waves exit from each port.
coefficients (S-parameters), each one
Power, voltage and current can be considered to
representing a possible input-output path. The S
be in the form of waves travelling in both the
parameters,
directions as shown in Fig 2.
.
=Input reflection coefficient
with the output port terminated by a matched
load (ZL=Z0 and a2=0). This is also called as
Return loss. It represents how much of the signal
is reflected back due to any impedance
mismatch. Three special cases are,
Figure 2 1). Zero Input Impedance (short circuit): S11=-1
because the entire signal is reflected and the
The incident waves are designated by the letter reflected voltage wave must be in anti phase in
an, where n is the port number of the network. orders to make the voltage at short circuit zero.
For each port, the incident and reflected wave 2). Infinite Input Impedance (Open Circuit):
are measured. The reflected wave is designed by S11=+1 because the entire signal is reflected and
bn, where n is the port number. voltage is maximum at open circuit.
Knowing the reflective and incident waves, any 3). Matched Input: This means that input
system can be fully characterized. impedance is equal to the system impedance and
it leads to S11=0 and entire signal goes to the
network.

= Reverse transmission
(insertion) gain with the input port terminated
Figure 3 [2]. Two port network terminated with in a matched load (a1=0). It is commonly
resistive load. referred to as the reverse isolation factor because
it represents the portion of the voltage waveform
The variables ai and bi are normalized complex applied at port-2 that appears leaving port-1.
voltage waves incident on and reflected from the
ith port of the network. They are defined in terms = Forward transmission
of the terminal voltage Vi. Voltage source that
has a characteristic impedance, ZS= 50 Ω is (insertion) gain or loss with the output port
applied to a linear two-port network which is terminated in a matched load (a2=0). It tells the
driving a load with impedance of ZL= 50 Ω. how much of the signal is lost from transferring
energy from one port to another. This is
essentially the definition of gain, and therefore
S21 represents the gain of the two-port network.
 = Output reflection coefficient reduces the difficulties involved in obtaining the
broadband open and short circuit conditions
with the input port terminated by a matched load required for the H-, Y-, and Z-parameters.
(ZS=Z0 and a1=0). Parasitic oscillations in active devices are
minimized when these devices are terminated in
When the incident wave travels through the resistive loads. Equipment is available for
network, its value is multiplied (i.e. is gain and determining S-parameters since only incident
phase are changed) by the scattering parameter and reflected voltages need to be measured. Any
(which is analogous to the gain of the network), set of parameters can be obtained from S
thus giving the resulting output value. So S - parameters.
parameters are complex values. S-parameters
depend upon the network and the characteristic References:
impedances of the source and load used to 1. Agilent AN 154 S-Parameter Design
measure it, and the measurement frequency (Application Note).
2. S-parameters techniques for faster, more
Advantages of S-parameter: accurate network design.
Although applicable at any frequency, S- http://www.hp.com/go/tmappnotes
parameters are mostly used for networks 3. www.wikipedia.org
operating at radio frequency (RF) and 4. S-parameters An introduction
microwave frequencies. S-parameters are www.polarinstruments.com
determined with resistive terminations. This