Chapter 8

2-port Scattering (S)

Parameters
8.1

Introduction and Denition of S-parameters

Recall from Chapter 7 that each parameter set discussed therein has a reference impedance
associated with each port. In the cases discussed so far, the reference impedance is either
a short circuit or an open circuit. The four complex numbers that make up the various
parameter sets include two parameters that characterize the input and output impedance
or admittance of the 2-port when each port is terminated in its reference impedance and
two additional parameters that describe the forward and reverse transfer characteristics
of the 2-port when terminated with the reference impedance. In each case, the reference
impedance is simply the termination that forces one of the independent parameters (e.g.
voltage or current) to zero. For example, the reference impedance for the Y-parameters is a
short circuit because terminating a port with a short circuit forces the voltage at that port
to zero.
At high frequencies it is dicult to implement terminations that accurately represent shortor open-circuits over a broad frequency range. Thus, it is desirable to dene a parameter
set that is based on a nite reference impedance that is easily realized in practice. In most
applications, a reference impedance that is purely resistive and having a moderate value is
relatively easy to implement. In order to retain the feature that terminating a port with the
reference impedance will force the independent variable corresponding to that port to zero,
any parameter set the employs a nite reference impedance must be based on independent
variables that are linear combinations of voltage and current rather than just voltage or
current. For example, consider the fact that terminating the output port of a 2-port with
some resistance R forces V2 /I2 = R, by Ohms Law (the minus sign arises because positive
current is dened to ow in to the port). Thus, when the output port is terminated with
R, we have V2 + RI2 = 0. Any constant times the quantity (V2 + RI2 ) could be chosen as
the independent variable in a parameter set based on a reference impedance R.
Scattering parameters (or S-parameters) can be viewed as a parameter set that is based on a
nite reference impedance. In most cases the reference impedance is either 50 or 75 and
229

230

CHAPTER 8. 2-PORT SCATTERING (S) PARAMETERS

is the same for both ports, although in some cases it is advantageous to allow the reference
impedance to be dierent at port 1 and port 2. In addition to being easier to implement
than short- or open-circuits, at HF, VHF, and microwave frequencies 2-ports are much less
likely to oscillate (become unstable) when terminated in a nite resistance or conductance.
Dene Zo (= Ro ) to be the reference termination for the S-parameter set. Throughout our
discussion we will assume that the reference termination is purely resistive and that the
same reference termination is used at both ports. Both of these assumptions can be relaxed
in a more advanced treatment of scattering parameters. Figure 8.1 shows the voltage and
current conventions that we will use for the S-parameter denitions:

I1

I2
2-port

V1
-

I1

V2
I2

Figure 8.1: Voltage and current conventions for S-parameters.

The S-parameter set is dened in terms of variables (a1 , b1 ) (a2 , b2 ) which are simply linear
combinations of the voltage and current at a particular port, i.e.,

a1

V1 + Zo I1

2 Zo

(8.1)

a2

V2 + Zo I2

2 Zo

(8.2)

b1

V1 Zo I1

2 Zo

(8.3)

b2

V2 Zo I2

2 Zo

(8.4)

a1 + b 1


V1 / Zo

(8.5)

a2 + b 2


V2 / Zo

(8.6)

a1 b 1


Zo I1

(8.7)

a2 b 2


Zo I2

(8.8)

The independent variables are a1 and a2 and the dependent variables are b1 and b2 . Remember, the reference impedance for a parameter set is that impedance which will force one

8.2. INTERPRETATION OF S-PARAMETERS

231

of the independent variables to zero when it is used to terminate the 2-port. The choice of
a nite reference impedance, Zo , forces us to choose the independent variables to be proportional to Vi + Zo Ii , as dened in Equations 8.1 and 8.2, so that they will be forced to
zero when the 2-port is terminated in the reference impedance, Zo .
The variables dened in Equations 8.1-8.4 have been normalized with respect to the square
root of the reference impedance, Zo . With this normalization (and assuming that the voltages and currents are rms values) the time-averaged power delivered to a given port (denoted by i) can be written as P = Re[Vi Ii ] = |ai |2 |bi |2 . Thus, the variables have units of
[Watts]1/2 .
Given the denitions in Equations 8.1-8.8, the scattering parameters are dened as follows:
b1

= S11 a1 + S12 a2

(8.9)

b2

= S21 a1 + S22 a2

(8.10)

Thus

8.2

S11 =

b1
|a =0
a1 2

(8.11)

S12 =

b1
|a =0
a2 1

(8.12)

S21 =

b2
|a =0
a1 2

(8.13)

S22 =

b2
|a =0
a2 1

(8.14)

Interpretation of S-parameters

Consider S11 :
S11 =

b1
|a =0
a1 2

(8.15)

Setting a2 = 0 is equivalent to setting

V2

= Zo I2

V2
I2

= Zo

(8.16)

That is, to set a2 = 0, we terminate port 2 with the reference impedance, Zo . Thus
S11 =

b1
V1 Zo I1
|a =0 =
|
a1 2
V1 + Zo I1 output port terminated in
=

V1 /I1 Zo
|
V1 /I1 + Zo output port terminated in

Zo

Zo

(8.17)

CHAPTER 8. 2-PORT SCATTERING (S) PARAMETERS

232
or,
S11 =

Zin Zo
|
Zin + Zo output port terminated in

(8.18)

Zo

where Zin is the input impedance of the 2-port when the output is terminated in Zo .
Recall from transmission line theory that S11 has the same form as the reection coecient,
, that relates the incident and reected voltage waves on a transmission line
=

reected voltage wave

V
=
+
V
incident voltage wave

(8.19)

where the total voltage on the line, V, is given by

V =V + V+

(8.20)

Comparison of Equations 8.19 and 8.20 with

S11 =

b1
a1

and
V1 = (a1 + b1 )

(8.21)

Zo

(8.22)

leads to the following correspondence between the variables a1 and b1 and the voltage waves
on a transmission line:
1
(Voltage wave incident on port i)
Zo

(8.23)

1
bi = (Voltage wave reected from port i)
Zo

(8.24)

ai

This is an interesting observation, but where is the transmission line? Although there
has been no mention of transmission lines in the system so far, it is helpful to imagine
that the source and load are connected to the 2-port through sections of transmission line
with characteristic impedance Zo . Well assume that the sections of transmission line have
innitesimal length, so that they do not aect the electrical characteristics of the system,
as in Figure 8.2.
ZS
Vs

I1

I2

a1
b1

V1
-

Zo

I1

2-port

a2
b2

V2
-

I2

ZL

Zo

Figure 8.2: Source and load connected to 2-port through transmission line with impedance
Zo
Employing this conceptual model, the a and b variables may be interpreted as representing
the (normalized) voltage waves that would exist on the sections of transmission line. When
working with S parameters, the model can be used to visualize the a and b variables as

8.2. INTERPRETATION OF S-PARAMETERS

233

incident and reected (normalized) voltage waves, although in practice the terminations may
be connected directly to the 2-port without any intervening transmission lines. (Actually,
in practice you will nd that it is virtually impossible to get signals into and out of a 2-port
without employing transmission lines between the source and the input port and the output
port and the load.)
The model shown in Figure 8.2 makes it clear that setting ZL = Zo will make a2 = 0, since
there will be no reection from the load termination. Notice that setting Zs = Zo would not
make a1 = 0 in the circuit shown in Figure 8.2, because the source would cause a non-zero
incident wave (a1 ) to exist on the input line.
Now consider S21 :
S21 =

b2
|a =0
a1 2

(8.25)

As discussed earlier, to set a2 = 0, terminate port 2 with Zo . To aid in obtaining an intuitive

feel for the signicance of S21 , it is helpful to consider the special circuit shown in Figure 8.3
where the source impedance is assumed to be equal to the reference impedance.
Zo

I1
+

a1
b1

Vs

V1 2-port
-

I2
+

V2
-

I1

b2

Zo

I2

Figure 8.3: A 2-port embedded in a system with ZS = ZL = Zo .

Note that the circuit has been drawn without showing any intervening lengths of transmission line between the 2-port and the terminations. The as and bs can be thought of
as existing at the node where the the 2-port is connected to the terminations. To solve
for S21 , rst relate b2 to the output voltage using Equation 8.4 and the auxiliary relation
V2 /I2 = Zo :
b2

1
(V2 Zo I2 )
2 Zo

= V2 / Zo

(8.26)

Then a1 can be related to the open circuit voltage of the source

1
a1 = (V1 + Zo I1 )
2 Zo

(8.27)

I1 = (Vs V1 )/Zo

(8.28)

using
Thus
a1

=
=

1
(V1 + Vs V1 )
2 Zo

Vs /2 Zo

(8.29)

234

CHAPTER 8. 2-PORT SCATTERING (S) PARAMETERS

Finally,
S21

b2
V2
|Z =Z
|a =0 =
a1 2
Vs /2 L o

(8.30)

So for the circuit shown in Figure 8.3, S21 is the voltage across the load divided by 1/2 of
the open circuit source voltage. Notice that if the 2-port is removed and the load connected
directly to the source, then the voltage across the load would be Vs /2. In other words, S21 is
the insertion voltage gain in a system where the source and load impedances are both Zo .
Remember that this result is not the denition of S21 ; it is a special result that was derived
from the denition for the case when both the source and load impedances are equal to Zo .
This result is often useful when it is necessary to derive an expression for S21 starting from
the circuit model of a 2-port. S12 has a similar interpretation - it is the reverse insertion
voltage gain, i.e. the insertion gain that would be measured if the two port was inserted
into the system backwards, i.e. with the output port connected to the source and input port
connected to the load. The insertion gain interpretion is useful to remember when trying to
get an intuitive feel for published or measured values of S21 or S12 .
Example 8.1. Computing S-parameters Given a Circuit Model
The 2-port in Figure 8.4 consists of a series impedance. To nd S11 , terminate the
ZS

Figure 8.4: 2-port consisting of a series impedance.

2-port in Zo and nd Zin as in Figure 8.5.
ZS
Zin

Zo

Figure 8.5: Setup for nding S11 .

Zin

= Zs + Zo

S11

Zin Zo
Zs
=
Zin + Zo
Zs + 2Zo

Because the circuit is not changed if the two ports are reversed, S22 = S11 .
To nd S21 , refer to Figure 8.6.

(8.31)
(8.32)

8.2. INTERPRETATION OF S-PARAMETERS

Zo

ZS

Vs

235

V2
-

Zo

Figure 8.6: Setup for nding S21 . Embed the 2-port in a Zo system, and then S21 = 2V2 /Vs .
Zo
2Zo + Zs

V2

Vs

V2
Vs

Zo
2Zo + Zs

S21

2Zo
V2
=
Vs /2
2Zo + Zs

(8.33)

(8.34)

Again, because the circuit is not changed if port 1 and port 2 are reversed, S12 =
S21 .

The procedure illustrated in the preceeding example can be applied to any 2-port, and
provides a general method for calculating the S-parameters when a circuit model is available.

8.2.1

Summary of 2-port S-parameters:

S11

Zin Zo
Zin +Zo |ZL =Zo

Input reection coecient with output terminated in Zo

S22

Zout Zo
Zout +Zo |ZS =Zo

Output reection coecient with input terminated in Zo

S21

Forward insertion voltage gain when source and load impedances are Zo

S12

Reverse insertion voltage gain when source and load impedances are Zo

CHAPTER 8. 2-PORT SCATTERING (S) PARAMETERS

236

8.3

Applications of Scattering Parameters

Consider in Figure 8.7 a 2-port with arbitrary source and load terminations (ZS , ZL ).
ZS
Vs

ZL

[S]Zo
S in

out L

Figure 8.7: 2-port with arbitrary source and load terminations

The reference impedance is assumed to be Zo . Dene in and out to be the input and
output reection coecients, i.e.,
in

Zin Zo
Zin + Zo

(8.35)

out

Zout Zo
Zout + Zo

(8.36)

The source reection coecient, S , is dened as

S =

ZS Zo
ZS + Zo

(8.37)

and the load reection coecient, L , is dened as

L =

8.3.1

ZL Zo
ZL + Zo

(8.38)

Derivation of Input and Output Reection Coecients

We will now derive expressions for in and out in terms of the source and load reection
coecients and the S-parameters of the 2-port. Consider the input reection coecient,
in = b1 /a1 . We expect in to be a function of the four S-parameters as well as the
load reection coecient, L . The relationship can be derived by noting that the load
termination imposes the following relationship between a2 and b2 :
L =

a2
b2

(8.39)

Using this relationship and Equation 8.10 we can write

a2 = a1

S21 L
1 S22 L

(8.40)

Inserting Equation 8.40 into Equation 8.9 yields

in =

b1
S12 S21 L
= S11 +
a1
1 S22 L

(8.41)

8.3. APPLICATIONS OF SCATTERING PARAMETERS

237

Similarly, it can be shown that the output reection coecient is given by

out =

b2
S12 S21 S
= S22 +
a2
1 S11 S

(8.42)

Inspection of Equation 8.41 shows that the input reection coecient reduces to S11 if
the load reection coecient is 0 (L = 0). This is consistent with the discussion in the
previous section regarding the interpretation of S11 . Similarly, out = S22 if S = 0. In
general, however, the input and output reection coecients depend on the way the 2-port
is terminated. In the special case of a 2-port with S12 = 0, the input and output reection
coecients are independent of the terminations. Such a 2-port is said to be unilateral, since
the device does not exhibit any reverse transmission. In other words, if a unilateral 2-port
(S12 = 0) is excited at port 2, no response will result at port 1. Although it is usually not
possible to construct a 2-port that is perfectly unilateral, in some cases the S12 coecient
is small enough that the 2-port can be considered to be approximately unilateral. The case
where S21 = 0 is also unilateral but not useful, since this 2-port would not give a response
at the output when excited at the input.

8.3.2

Stability of 2-ports

Before detailed design calculations based on a particular 2-port are made, it is usually
necessary to investigate whether the 2-port is potentially unstable. Of course, stability is
only a concern when the 2-port contains active elements such as a transistor or a negative
resistance device. The questions that need to be answered are:
1. Is the 2-port unconditionally stable? That is, is there any combination of passive
source and load terminations (L , S ) for which the 2-port will oscillate? If not, the
2-port is said to be unconditionally stable.
2. If the 2-port is not unconditionally stable, i.e., if it is potentially unstable, then we
would like to be able to determine which source and load terminations make it unstable.
If the 2-port is to be used as an amplier, we would avoid the unstable terminations.
If the goal is to design an oscillator, the designer would deliberately choose source and
load terminations to cause oscillation.
The stability question can be studied using the negative resistance concept. Considering
Figure 8.8, suppose that one of the impedances represents the input or output impedance
of the 2-port.

Z1

Z2

Figure 8.8: Z1 or Z2 represents input or output impedance of 2-port

CHAPTER 8. 2-PORT SCATTERING (S) PARAMETERS

238

Remember that the circuit will oscillate when the two impedances are connected if Re
[Z1 + Z2 ] 0 and Im [Z1 + Z2 ] = 0. Thus, to make an oscillator, one of the Zi s must
have a negative real part.
For a particular 2-port, if there is some load impedance that makes the real part of Zin 0,
or if there is some source impedance that makes the real part of Zout 0, then the 2-port
is potentially unstable because it is possible to choose a passive source or load termination
that will make the system oscillate.
An impedance with a negative resistive part corresponds to a reection coecient with a
magnitude greater than 1. For example, consider Zin and write Zin = Rin + j Xin , then
in

Zin Zo
Zin + Zo

(Rin Ro ) + jXin
(Rin + Ro ) + jXin


|in |

(Zo = Ro (real))

2
(Rin Ro )2 + Xin
2
(Rin + Ro )2 + Xin

(8.43)

1/2
(8.44)

Examination of Equation 8.44 leads to the following observations:

If Rin > 0
If Rin 0

|in | < 1
|in | 1

(8.45)
(8.46)

To determine whether a 2-port is potentially unstable, we check to see if either |in | or |out |
can be larger than or equal to unity.
First consider |in |:
|in | =
=

|S11 +
|

S12 S21 L
|
1 S22 L

(8.47)

S11 L (S11 S22 S12 S21 )

|
1 S22 L

Dene the determinant of the S-parameter matrix, D, as follows:

D = S11 S22 S12 S21

(8.48)

S11 L D
|
1 S22 L

(8.49)

Then Equation 8.47 becomes

|in | = |

We can now set |in | = 1 and solve for the corresponding locus of points in the L plane,
i.e., we can solve for the values of L that make |in | = 1. Setting |in | = 1:
|S11 L D| = |1 S22 L |

(8.50)

The absolute value signs can be eliminated by squaring both sides of Equation 8.50:

L D ) = (1 S22 L )(1 S22
L )
(S11 L D)(S11

(8.51)

8.3. APPLICATIONS OF SCATTERING PARAMETERS

239

Expanding both sides of the equation and collecting terms yields

|L |2 +

))
2 Re(L(S22 DS11
|S11 |2 1
+
|D|2 |S22 |2
|D|2 |S22 |2

(8.52)

In deriving Equation 8.52, use has been made of the following identity
z + z = 2Re(z)

(8.53)

where the operator Re() extracts the real part of its argument. Also note that
z z = 2Im(z)

(8.54)

where the operator Im() extracts the imaginary part of its argument. Now, it is convenient
to rewrite Equation 8.52 in terms of the real and imaginary parts of L .
Let L = UL + jVL and substitute Equation 8.52. After some fairly extensive algebraic
manipulation, Equation 8.52 can be written in the form
(UL UCL )2 + (VL VCL )2 = rL 2

(8.55)

where
UCL

Re(DS11
S22 )
2
|D| |S22 |2

(8.56)

VCL

)
Im(S22 DS11
2
2
|D| |S22 |

(8.57)

rL

|S12 S21 |
||S22 |2 |D|2 |

(8.58)

Equation 8.55 indicates that the locus of points in the L plane that correspond to input
reection coecients with unit magnitude is a circle with center at the complex point CL =
UCL + jVCL and radius rL . It is convenient to write the coordinates of the center of the
stability circle as a complex number, i.e. dene CL = UCL + jVCL . Then
CL =

D S11
S22
|S22 |2 |D|2

(8.59)

The circle, when plotted on the L plane, is referred to as the L -plane stability circle.
Consider the stability circle shown in Figure 8.9.
Figure 8.9 contains two circles. The circle that is centered on the origin is a unit circle
(circle with radius = 1) and represents the outer boundary of the Smith Chart. Points
within that circle correspond to load reection coecients (L ) with magnitudes less than 1
or, equivalently, load impedances with a positive real part. The other circle is the stability
circle, and values of L which lie on that circle will map to input reection coecients
with magnitude equal to 1 or, equivalently, input impedances that are purely reactive.
The stability circle represents the boundary between the region of the L plane that maps
to input reection coecients with magnitude less than one (|in | < 1) and the region
that maps to (|in | > 1). The region of the L plane that maps to (|in | > 1) is called

240

CHAPTER 8. 2-PORT SCATTERING (S) PARAMETERS

Figure 8.9: L -plane stability circle

the unstable region of the L plane. The two regions (stable and unstable) correspond to
the regions inside and outside the stability circle. To decide whether the unstable region
corresponds to the region inside or outside of the stability circle, it is sucient to examine
the value of S11 for the 2-port under consideration. Note carefully that the origin of the L
plane (the point L = 0) maps to S11 , i.e., recall
in = S11 +
so that when L = 0, we have

S12 S21 L
1 S22 L

in = S11

(8.60)

(8.61)

Now consider Figure 8.9 and suppose that it is known that |S11 | < 1. The origin in the L
plane is outside the stability circle and this point maps to S11 . This leads us to conclude:
(i) the region of the L plane that lies outside the stability circle maps to |in | < 1 and (ii)
the region inside the stability circle maps to |in | > 1. The region inside the stability circle
would therefore be referred to as the unstable region of the L plane.
Thus far we have treated only the so-called L -plane stability circle. A complete characterization of the 2-ports stability requires that the S -plane stability circle also be studied.
The S -plane stability circle corresponds to the values of S which map to output reection
coecients (out ) with magnitude equal to one. TheS -plane stability circle is described by
(US UCS )2 + (VS VCS )2 = rS 2

(8.62)

where
CS

UCS + jVCS

(8.63)

8.3. APPLICATIONS OF SCATTERING PARAMETERS

rS

241

D S22
S11
|S11 |2 |D|2

(8.64)

|S12 S21 |
||S11 |2 |D|2 |

(8.65)

A decision regarding whether the unstable region in the S -plane lies inside or outside of
the S -plane stability circle is made by examining the magnitude of S22 and noting that the
point S = 0 (the origin of the S -plane) maps to S22 .
It is now appropriate to dene the concept of unconditional stability. An unconditionally
stable 2-port has the property that no choice of passive source and load terminations will
make the 2-port oscillate. In other words, an unconditionally stable 2-port has |in | < 1 and
|out | < 1 for any choice of passive source and load terminations. The restriction to passive
sources and loads means that we limit our attention to sources with |S | 1 and loads with
|L | 1, i.e., those regions of the S - and L -planes that lie inside and on the unit circle
centered on the origin. A 2-port is unconditionally stable if the unstable regions of the S and L -planes lie completely outside of the unit circles. Once the center coordinates and
the radii of the stability circles have been computed, the circles can be plotted and, utilizing
the known magnitudes of S11 and S22 , we can determine whether the unstable regions lie
outside of the unit circles.
As an alternative to plotting the stability circles, it is possible to derive a relatively simple
algebraic criterion (or set of criteria) that must be satised in order for a 2-port to be
unconditionally stable. The derivation of one such criterion is given in section 8.3.3. A
summary of the various sets of criteria that have been derived is given here. In each case, if
the criterion (or set of criteria) is satised, then the 2-port is unconditionally stable:
K > 1, |S12 S21 | < 1 |S11 |2
2

(8.66)

K > 1, |S12 S21 | < 1 |S22 |

(8.67)

K > 1, B1 > 0

(8.68)

K > 1, B2 > 0

(8.69)

K > 1, |D| < 1

(8.70)

ES =

1 |S11 |2
D| + |S S | > 1
|S22 S11
12 21

(8.71)

ES =

1 |S22 |2
D| + |S S | > 1
|S11 S22
12 21

(8.72)

1 |S11 |2 |S22 |2 + |D|2

>1
2 |S12 | |S21 |

(8.73)

where
K=

D = S11 S22 S12 S21

(8.74)

CHAPTER 8. 2-PORT SCATTERING (S) PARAMETERS

242

B1 = 1 + |S11 |2 |D|2 |S22 |2

(8.75)

B2 = 1 + |S22 |2 |D|2 |S11 |2

(8.76)

Any one set of criteria given by 8.66-8.72 is a necessary and sucient set of criteria for a
2-port to be unconditionally stable. (It should be noted that some textbooks state that the
union of the criteria listed in 8.66 and 8.67 are necessary and sucient criteria. This results
in 3 criteria, but it can be shown that either 8.66 or 8.67 are necessary and sucient, i.e.
at most only 2 inequalities must be checked.)
If any one of the sets of stability criteria are satised, then no choice of passive source and
load terminations will make the 2-port oscillate. This statement applies only at the frequency
where the S-parameters were measured. It is possible for a 2-port to be unconditionally
stable in a certain frequency band but potentially unstable in some other frequency band.
In practice, stability should be checked at many frequencies in the bandwidth within which
the 2-port has appreciable gain. This requires that the S-parameters be measured at many
frequencies. Checking the one of the sets of conditions given in equations 8.66-8.72 is easier
than plotting the stability circles, especially when one wishes to quickly determine the
stability status over a wide frequency range; however if the 2-port turns out to be potentially
unstable, then the stability circles provide detailed information on which terminations make
the 2-port unstable.
Finally, note that the parameter K dened in Equation Fac8.73 and used in criteria 8.66
through 8.70 is called the Rollet Stability Factor. While the condition K>1, by itself, does
not guarantee that a 2-port is unconditionally stable, it turns out that K plays an important
role in determining whether or not a 2-port can be simultaneously matched with passive
terminations at both the input and output ports.
Example 8.2. 2-port stability analysis.
A 2-port has S-parameters (Zo = 50 )
S11

0.4 20

(8.77)

S12

0.140

(8.78)

S21

7.5150

(8.79)

S22

0.6 30

(8.80)

The Rollett Stability factor for this 2-port is K = 0.853 < 1, so the 2-port is
potentially unstable. The coordinates of the stability circles are
UCS

0.163

VCS

0.589

rS

= 1.17

UCL

0.401

VCL

0.913

rL

= 1.70

(8.81)
The stability circles for this 2-port are plotted in Figures 8.10 and 8.11. The stable
region of the S - and L -planes is shaded. For the S -plane stability circle, the

8.3. APPLICATIONS OF SCATTERING PARAMETERS

243

1.

0.5

-1.

-0.5

0.5

1.

-0.5

-1.

-1.5

Figure 8.10: S plane stability circle.

1.
0.5
-2. -1.5 -1. -0.5

0.5

1.

-0.5
-1.
-1.5
-2.
-2.5

Figure 8.11: L plane stability circle.

244

CHAPTER 8. 2-PORT SCATTERING (S) PARAMETERS

stable region was identied by noting that the origin of the S -plane (S = 0) maps
to an output reection coecient out = S22 . Since the origin of the S -plane is
in the interior of the stability circle and |S22 | < 1, we conclude that the interior
of the stability circle maps to output reection coecients |out | < 1; hence, the
interior of the stability circle is the stable region. Similar reasoning was used to
determine that the interior of the L -plane stability circle was stable. To verify
that the stable and unstable regions have been correctly identied, consider the
two points labeled A and B in the L -plane plot. The point A is in the unstable
region and represents a load reection coecient of L = 0.5 + j0.7 = 0.8654.
Using this value in Equation 8.41 yields in = 1.3 76.4 which represents an
input impedance with a negative real part, as expected. Point B, which is inside
the stable region, represents a load reection coecient L = 0.4 + j0.4 = .5745.
The input reection coecient is in = 0.70 83 .

The following example is identical to the previous one, except that the magnitude of S12
has been reduced from 0.1 to 0.01.
Example 8.3. 2-port stability analysis. A 2-port has S-parameters (Zo = 50 ):
S11

0.4 20

(8.82)

S12

0.0140

(8.83)

S21

7.5150

(8.84)

S22

0.6 30

(8.85)

The stability factor, K, is 3.74 (> 1), and we also have

1 |S11 |2

>

|S12 S21 |

(8.86)

1 |S22 |2

>

|S12 S21 |

(8.87)

so the 2-port is unconditionally stable. (Note, only one of 8.86 and 8.87 needs to be
checked.) Stability circles for this example are shown in Figures 8.12 and 8.13. As
in the rst example, we have made use of the fact that the origin of the S - and L planes map to output and input reection coecients, respectively, with magnitudes
less than 1. Hence the origin of each plane is in the stable region. Since the origin
is outside the stability circle, the region outside the circle is the stable region and
the region inside is the unstable region. Since the unstable regions do not include
any passive source or load terminations, the 2-port is said to be unconditionally
stable.

In the following section, a simple test is derived that can predict whether or not a particular
2-port is unconditionally stable.

8.3. APPLICATIONS OF SCATTERING PARAMETERS

245

2.5
2.
1.5
1.
0.5
-1.

1.

2.

3.

-0.5
-1.
Figure 8.12: S plane stability circle

1.

0.5

-1.

-0.5

0.5

1.

1.5

-0.5

-1.
Figure 8.13: L plane stability circle

CHAPTER 8. 2-PORT SCATTERING (S) PARAMETERS

246

8.3.3

Derivation of a Criterion for Unconditional Stability of 2ports

All of the essential information necessary to investigate the stability of a 2-port (at a particular frequency) is contained in the 4 parameters that dene the stability circles in the
S and L planes. In this section, we derive a criterion that can be used to quickly check
to see if a 2-port is unconditionally stable. As already pointed out, numerous sets of necessary and sucient criteria have been derived that can be used to determine whether or
not a 2-port is unconditionally stable.1 The criterion derived in this section is particularly
interesting because it involves only a single inequality and because the numerical value of
the stability parameter has the useful property that its magnitude represents the smallest
distance between the origin of the reection coecient plane and the edge of the unstable
region.
In order for a 2-port to be unconditionally stable, it is necessary for the unstable region
in the L and S planes to lie outside of the unit circle. Consider the L plane rst, and
assume that |S11 | < 1 (this is a necessary condition for unconditional stability, because if
|S11 | 1 a negative, or zero, real part of ZIN is obtained by setting ZL = Zo . In this case,
the 2-port can oscillate with a passive source termination and hence is potentially unstable.)
There are two cases to consider.
1. First, suppose that the stability circle does not enclose the origin of the L plane
(i.e. |CL | > rL ). Then the inside of the stability circle is the unstable region. If the
2-port is unconditionally stable it is necessary for the minimum distance between the
origin and the edge of the stability circle to be larger than 1, i.e. it is necessary that
|CL | rL > 1. Inserting the equations for |CL | and rL we nd that this amounts to:

or

|S12 S21 |
D S11 |
|S22

>1
||S22 |2 |D|2 | ||S22 |2 |D|2 |

(8.88)

D S11 | |S12 S21 |

|S22
>1
||S22 |2 |D|2 |

(8.89)

2. Now consider the other possibility, namely that the stability circle encloses the origin,
i.e. that |CL | < rL . In this case, the necessary condition |S11 | < 1 implies that the
exterior of the stability circle is the unstable region. In order to have the unstable
region lie completely outside of the unit circle, we require that the stability circle
completely encloses the unit circle centered at L = 0, i.e.:
rL |CL | > 1
or

1 Lombardi,

|S12 S21 | |S22

DS11
|
>1
2
2
||S22 | |D| |

(8.90)

G. and B. Neri, Criteria for the evaluation of unconditional stability of microwave linear
two-ports: a critical review and new proof, IEEE Trans. MTT, Vol. 47, No. 6, June, 1999, p746.

8.3. APPLICATIONS OF SCATTERING PARAMETERS

247

Note carefully that equation 8.89 guarantees that the unstable region is outside of the unit
circle provided that |CL | > rL whereas equation 8.90 guarantees that the unstable region
is outside of the unit circle when |CL | < rL . Now consider what |CL | > rL (or |CL | < rL )
implies in terms of the S parameters. Consider |CL | > rL - then:

|S22
D S11 |
|S12 S21 |
>
||S22 |2 |D|2 |
||S22 |2 |D|2 |

or

|S22
D S11 | > |S12 S21 |

Squaring both sides (note that squaring both sides does not lose any generality, because
both sides are always positive) yields

(S22
D S11 )(S22 DS11
) > |S12 S21 |2

and expanding the LHS yields

DS11
S22 D S11 + |D|2 |S11 |2 > |S12 S21 |2
|S22 |2 S22

Expand the second and third terms on the LHS using D = S11 S22 S12 S21 :

|S22 |2 S11
S22
(S11 S22 S12 S21 ) S11 S22 (S11
S22 S12
S21 ) + |D|2 |S11 |2 > |S12 S21 |2

and group terms on the LHS:

|S22 |2 2|S11 |2 |S22 |2 + S11
S22 S12 S21 + S11 S22 S12
S21 + |D|2 |S11 |2 > |S12 S21 |2

(8.91)

Now, noting that

|D|2 = |S11 |2 |S22 |2 S11 S22 S12

S21
S11
S22
S12 S21 + |S12 |2 |S21 |2

(8.92)

Equation 8.92 can be used to eliminate the complex terms (3rd and 4th term on LHS) in
equation 8.91. The result is:
|S22 |2 |S11 |2 |S22 |2 + |S12 |2 |S21 |2 |D|2 + |D|2 |S11 |2 > |S12 |2 |S21 |2
or, after grouping terms, we determine that |CL | > rL amounts to:
(|S22 |2 |D|2 )(1 |S11 |2 ) > 0

(8.93)

Since we have already assumed that |S11 | < 1, the second term on the LHS is always positive
so the combined requirements that the stability circle not enclose the origin ( |CL | > rL )
and that the inside of the stability circle represent the unstable region (|S11 | < 1) means
that:
(8.94)
|S22 |2 |D|2 > 0
Similarly, the requirements that the stability circle encloses the origin (|CL | < rL ) and that
the inside of the stability circle represents the unstable region (|S11 | < 1) amounts to:
|S22 |2 |D|2 < 0

(8.95)

We now know that criterion 8.66 applies only when |S22 |2 |D|2 > 0 and criterion 8.67
applies when |S22 |2 |D|2 < 0. Note that criteria 8.66 and 8.67 dier only in the sign of the

CHAPTER 8. 2-PORT SCATTERING (S) PARAMETERS

248

numerator on the LHS, and that the denominator is the magnitude of |S22 |2 |D|2 . Thus,
these criteria can be combined into into a single inequality which is valid for either case:

|S22
D S11 | |S12 S21 |
>1
|S22 |2 |D|2

(8.96)

There is an apparent singularity in equation 8.96 when |S22 |2 |D|2 . This corresponds to the
situation where the radius of the stability circle approaches innity, i.e. when the stability
circle degenerates into a straight line. The singularity can be removed by noting that:
|S22 |2 |D|2 =

D|2 |S12 S21 |2

|S22 S11
1 |S11 |2

D| |S12 S21 |)(|S22 S11

D| + |S12 S21 |)
(|S22 S11
2
1 |S11 |

(8.97)

Use equation 8.97 in equation 8.96 to write:

ES =

|S22

1 |S11 |2
>1
D S11 | + |S12 S21 |

(8.98)

The parameter ES is referred to as the Edwards-Sinsky stability criterion,2 and has a useful
geometric interpretation: ES is the minimum distance between the origin of the L plane
and the unstable region. If the parameter ES is negative, it means that the unstable region
includes the origin of the L plane.
If the identical analysis is carried out for the S plane stability circles, we derive the dual
constraint:

1 |S22 |2
>1
(8.99)
ES =
|S11 D S22 | + |S12 S21 |


The geometric interpretation of ES is the same as that of ES , but applied to the S plane.
It turns out that either criterion 8.98 or criterion 8.99 are necessary and sucient for unconditional stability of a 2-port, even though each one was derived by considering only one
of the L plane or S plane stability circles. To see why this is so, we need to show that if a
2-port satises the conditions for oscillation at one port, then it must satisfy the conditions
for oscillation at the other port, and vice versa.
Suppose that the L plane analysis shows that there exists a passive load termination, call it
Lu , that causes |in | to be 1 , i.e. |in (Lu )| 1. Then, the negative resistance criterion
for steady-state oscillation will be satised at the input port if the source reection coecient
is chosen to be Su = in (Lu )1 , which causes Su in (Lu ) = 1. (As an exercise, you may
wish to verify that the condition Z1 + Z2 = 0 is equivalent to 1 2 = 1.) Note that the
source reection coecient is passive, since |Su | = 1/|in (Lu )| 1. Now, it is left as an
exercise to show that out (Su ) = 1
Lu . This implies
out (Su )Lu = 1
2 Edwards,

M. L. and J. H. Sinsky, A new criterion for linear 2-port stability using a single geometrically
derived parameter, IEEE Trans. MTT, Vol. 40., No. 12, December 1992, p2303.

8.3. APPLICATIONS OF SCATTERING PARAMETERS

249

which means that if the oscillation condition is satisifed at the input port, then it is also
satised at the output port. The converse can be proven using the same approach by
changing subscripts.
Since |Lu | < 1, this result implies that |out (Su )| > 1 and hence we conclude that if a
passive load termination exists such that |in (Lu )| > 1 then there is always a corresponding
passive source termination, given by Su = in (Lu ), that will make |out | > 1.
So, we can conclude that if the oscillation conditions can be satised at the input port with
a particular set of passive source and load terminations, then the oscillation conditions must
also be satised at the output port with the same terminations, and vice versa. This means
that if the oscillation condition cannot be satised at the input port with any passive source
and load, then it is also impossible for the oscillation condition to be satised at the output
port. This is why a necessary and sucient stability criterion can be derived by considering
only the L plane or S plane stability circles.

8.3.4

Terminations for Simultaneous Conjugate Match

Whenever K > 1 it is possible to nd a combination of passive source and load terminations for which both the input and output of the 2-port are conjugately matched. For a
given available source power, a simultaneous conjugate match condition leads to the highest
possible power delivered to the load. The particular source and load reection coecients
which result in a simultaneous conjugate match at both ports are denoted by ms and ml .
Equations for ms and ml can be derived by solving Equations 8.100 and 8.101, which
enforce the conjugate match relationship at the input and output of the 2-port, i.e.,
in = ms = S11 +

out = ml

S12 S21 ml
1 S22 ml

S22 +

S12 S21 ms
1 S11 ms

(8.100)

(8.101)

The solution to Equations 8.100 and 8.101 is given by

B1 B12 4|C1 |2
ms =
2C1
B1

1 + |S11 |2 |D|2 |S22 |2

C1

S11 DS22

ml

B2

1 + |S22 |2 |D|2 |S11 |2

C2

S22 DS11

B2

(8.102)

B22 4|C2 |2
2C2

(8.103)

The proper choice of sign in Equations 8.102 and 8.103 is determined by the requirement
that |ms | < 1 and |ml | < 1. In other words, the correct solution in each case is the

CHAPTER 8. 2-PORT SCATTERING (S) PARAMETERS

250

one with a magnitude less than 1. Equations 8.102 and 8.103 will yield valid solutions
whenever K > 1. Since this is only a necessary condition for stability, it is possible to nd
a simultaneous conjugate match solution for some potentially unstable 2-ports (those with
K > 1).
Suppose ms and ml are known for a particular 2-port (with K > 1). Then the procedure
for designing an amplier which maximizes the power delivered to the load consists of
designing matching networks that transform the actual source and load reection coecients
into ms and ml , as in Figure 8.14.
ZS
Lossless
Matching
Network

Vs
S S

Lossless
Matching
Network

[S]Zo
K>1
ms ms

ml ml

ZL
L L

Figure 8.14: An amplier that is simultaneously matched to source and load terminations.
A conjugate match exists at all ports.
If a 2-port does not satisfy the conditions for unconditional stability, then it is said to be
potentially unstable. Useful and stable ampliers can be designed with potentially unstable
2-ports, however it is necessary to carefully choose source and load terminations to avoid the
unstable regions of the S - and L - planes. If K > 1, then one such choice of terminations
is S = ms and L = ml . If K 1, then it will not be possible to achieve a simultaneous
conjugate match, and the terminations will have to be chosen according to a carefully
measured tradeo between transducer gain and relative stability.
Of course, potentially unstable 2-ports can also be used to design oscillators. For example,
a basic oscillator design could be accomplished by choosing a S in the unstable region, i.e.,
a S that maps to |out | > 1. Using this source termination, the output impedance of the
2-port will have a negative real part. The load termination is then chosen such that the
negative resistance criterion for oscillation is satised at the output port. As shown earlier,
this will automatically cause the negative resistance criterion to be satised at the input
port.

8.3.5

Power Gains

1. Operating Power Gain (or just power gain)

G

Pout
Pin

|S21 |2 (1|L |2 )
(1|S11 |2 )+|L |2 (|S22 |2 |D|2 )2Re(L C2 )

with
D = S11 S22 S12 S21

C2 = S22 DS11

Note that G depends only on the load termination, L .

(8.104)

8.3. APPLICATIONS OF SCATTERING PARAMETERS

251

2. Transducer Power Gain

GT

Pout
Pavs

|S21 |2 (1|S |2 )(1|L |2 )

|(1S11 S )(1S22 L )S12 S21 L S |2

(8.105)

Notice that GT depends on both terminations, S and L .

3. Available Power Gain
GA

Pavo
Pavs

|S21 |2 (1|S |2 )
(1|S22 |2 )+|S |2 (|S11 |2 |D|2 )2Re(S C1 )

(8.106)

with
D = S11 S22 S12 S21

C1 = S11 DS22

Note that GA depends solely on the source impedance. It tells us how much power is
potentially available from the output of the 2-port.
4. Unilateral Transducer Power Gain
If the internal feedback within the device is small (S12 0), then the device can be
considered unilateral. This approximation yields a particularly simple form for the
transducer gain:
GT U =
GT |S12 =0
(8.107)
2
1|S |2
2 1|L |
GT U = |1S
|S
|
2
21
2
|1S22 L |
11 S |
This form clearly shows the eects of input and output mismatch which are contained
in the rst and last terms of the expression.
5. Maximum Available Gain
A system that is conjugately matched at both ports (as in Figure 8.14) will have
G = GA = GT = GA,max where:



 S21

2

[K K 1]
GA,max = 
(8.108)
S12
The upper sign applies when B1 < 0 and the lower sign applies when B1 > 0. The
maximum available gain could also be calculated by using ml in the formula for G,
or ms in the formula for GA , or (ms , ml ) in the formula for GT .

CHAPTER 8. 2-PORT SCATTERING (S) PARAMETERS

252

8.4

Summary of useful S-parameter formulas

I1

I2

a1
V1
b1

a2
V
b2 2

2-port

I1

I2

Relationship between impedance ZX (X can be in, out, S, L) and reection coecient X :

X =

ZX Zo
ZX + Zo

ZX = Zo

1 + X
1 X

Input reection coecient with arbitrary ZL :

in = S11 +

S12 S21 L
1 S22 L

Output reection coecient with arbitrary ZS :

out = S22 +

S12 S21 S
1 S11 S

Voltage gain with arbitrary ZL (voltage gain does not depend on ZS ):

AV =

V2
S21 (1 + L )
=
V1
(1 + S11 )(1 S22 L ) + S12 S21 L

Current gain with arbitrary ZL (current gain does not depend on ZS ):

AI =

I2
S21 (L 1)
=
I1
(1 S11 )(1 S22 L ) S12 S21 L

Denitions of some commonly used quantities:

D

S11 S22 S12 S21

1|S11 |2 |S22 |2 +|D|2

2|S12 S21 |

B1

1 + |S11 |2 |D|2 |S22 |2

C1

S11 DS22

B2

1 + |S22 |2 |D|2 |S11 |2

C2

S22 DS11

Operating Power Gain:

G
Available Power Gain:
GA

|S21 |2 1 |L |2
Pout
=
Pin
(1 |S11 |2 ) + |L |2 (|S22 |2 |D|2 ) 2Re(L C2 )
`

|S21 |2 1 |S |2
Pavo
=
Pavs
(1 |S22 |2 ) + |S |2 (|S11 |2 |D|2 ) 2Re(S C1 )

8.4. SUMMARY OF USEFUL S-PARAMETER FORMULAS

Transducer Power Gain:
GT

253

`
`

|S21 |2 1 |S |2 1 |L |2
Pout
=
Pavs
|(1 S11 S )(1 S22 L ) S12 S21 L S |2

A 2-port is unconditionally stable if:

K > 1 and 1 |S11 |2 > |S12 S21 |
or, if:

K > 1 and 1 |S22 |2 > |S12 S21 |

or, if:
K > 1 and B1 > 0
or, if:
K > 1 and B2 > 0
or, if:
K > 1 and |D| < 1
or, if:
ES =

1 |S11 |2
>1
D| + |S S |
|S22 S11
12 21

ES =

1 |S22 |2
>1
D| + |S S |
S22
12 21

or, if:

Stability circles in the L plane:

Center is at :

|S11

CL =

D S
S22
11
|S22 |2 |D|2

radius :

rL =

|S12 S21 |
||S22 |2 |D|2 |

CS =

D S
S11
22
|S11 |2 |D|2

radius :

rS =

|S12 S21 |
||S11 |2 |D|2 |

Stability circles in the S plane:

Center is at :

Source and load reection coecient for simultaneous conjugate match (K must be > 1) - in each case,
choose the sign that results in a reection coecient with magnitude less than 1:
q
q
B1 B12 4|C1 |2
B2 B22 4|C2 |2
ms =
ml =
2C1
2C2
Maximum available power gain (only dened for K > 1):

p
S21

GA,max =
[K K 2 1]
S12
GA,max is dened only for 2-ports that can be conjugately matched at both ports (K > 1). For unconditionally stable 2-ports, B1 > 0, use the lower (negative) sign. For potentially unstable 2-ports, B1 0, use
the upper (positive) sign.

CHAPTER 8. 2-PORT SCATTERING (S) PARAMETERS

254

8.5

References

1. Carson, Ralph S., High Frequency Ampliers, John Wiley & Sons, New York, 1975.
2. Gonzalez, Guillermo, Microwave Transistor Ampliers: Analysis and Design, PrenticeHall, New Jersey, 1984.
3. Liao, Samuel Y., Microwave Circuit Analysis and Amplier Design, Prentice-Hall,
New Jersey, 1987.
4. Pozar, David M., Microwave Engineering, Addison-Wesley Publishing Company, 1990.
5. Vendelin, George D., Design of Ampliers and Oscillators by the S-parameter Method,
John Wiley & Sons, New York, 1982.
6. Vendelin, George D., Anthony M. Pavio, Ulrich L. Rohde, Microwave Circuit Design
Using Linear and Nonlinear Techniques, John Wiley & Sons, 1990.

8.6. HOMEWORK PROBLEMS

8.6

255

Homework Problems

1. Find the S-parameters for the T-network in Figure 8.15. Port 1 is on the left:

Figure 8.15: T-network

2. Find expressions for the S-parameters of the 2-port in Figure 8.16. Denote the reference
impedance by Zo .

Figure 8.16: 2-port

3. Consider an ideal transformer as shown in Figure 8.17:

I1

1:N

V1

I2

V2

Figure 8.17: Ideal transformer

This 2-port is characterized by the following relations:
V2
I2

= N V1
= I1 /N

(8.109)

Find expressions for the S-parameters of this 2-port. Denote the reference impedance
by Zo .

CHAPTER 8. 2-PORT SCATTERING (S) PARAMETERS

256

4. Suppose that two 2-ports are cascaded. Show that the S21 parameter for the overall
network is given by
(1)

S21 =
(1)

(2)

S21 S21
(2)

(1)

1 S11 S22

(8.110)

(2)

where Sij and Sij refer to the S-parameters of the rst and second 2-ports, respectively.
5. A 2-port has the following S-parameters (Z0 = 50 ):
S11

0.5 96

(8.111)

S12
S21

=
=

0.350
5.245

S22

0.4 120

(a) What is the input impedance when the 2-port is terminated with a short circuit,
i.e., ZL = 0?
(b) Is this 2-port unconditionally stable? (Check one of the necessary and sucient
sets of criteria.)
(c) Find the coordinates of the center and the radius of the stability circles in the S
and L planes. Sketch the stability circles and shade the regions that correspond
to those values of S and L that make |out | > 1 or |in | > 1.
6. Suppose that S12 in Problem 5 is changed to S12 = 0.0550. Find the coordinates of
the center and the radius of the stability circles in the S and L planes. Sketch the
stability circles and shade the regions that correspond to those values of S and L
that make |out | > 1 or |in | > 1. Is the 2-port unconditionally stable?
7. The stability circles in the output (L ) plane for cases (a) and (b) are shown in Figures
8.18 and 8.19.

Figure 8.18: Case (a)

8.6. HOMEWORK PROBLEMS

257

Figure 8.19: Case (b)

Suppose that in both cases it is known that |S11 | < 1. For each case indicate what
region of the L plane corresponds to the stable region. The stable region of the L
plane corresponds to those values of L that will cause the input reection coecient
to have a magnitude less than 1.
8. For a particular 2-port with |S11 | < 1 and |S22 | < 1, the center and radius of the
stability circles in the S and L planes are known to be
CL
rL

=
=

4.0 j2.0
4.0

CS
rS

=
=

3.0 j4.0
7

(8.112)

(a) Make two sketches, one each for the S and L planes, and show the stability
circles. Shade the unstable region.
(b) Is the 2-port under consideration unconditionally stable?
9. Consider the system in Figure 8.20 where RS = 300 and RL = 100 . Suppose that
the 2-port is known to be unilateral and that S22 = 0.8 (Zo = 50 ). The transducer
gain of the 2-port is GT = 12 dB. Find the available gain of the 2-port in this system.
10. Consider the system in Figure 8.21 where RS = RL = 300 . The voltage gain,
Av = VVoi , in this system is found to be Av = 645. In addition, it is known that the
2-port is unilateral, and that S11 = 0.2 (Zo = 50 ).
(a) Find the operating power gain. Express your result in dB.
(b) Find the transducer powergain. Express your result in dB.
11. A 2-port has the following S-parameters (Z0 = 50 ):
S11

0.6 100

(8.113)

258

CHAPTER 8. 2-PORT SCATTERING (S) PARAMETERS

Figure 8.20: Unilateral 2-port

Figure 8.21: Unilateral 2-port

S12
S21

=
=

0.0545
3.5060

S22

0.1 30

(a) Is this 2-port unconditionally stable?

(b) Can this 2-port be conjugately matched simultaneously at the input and output?
If so, nd the source and load impedances (Zms and Zml ) that will make this
condition occur.
(c) If this 2-port is used in a system with the source and load impedances found in
part 11b, nd the operating, transducer, and available gains (G, GT , and GA ).
Express your results in dB.
12. The 2-port from Problem 11 is used in Figure 8.22:
(a) Compute the power available from the source, Pavs . The peak open circuit voltage
of the source is 3.0 V. Express your result in dBm.
(b) Find G, GT , and GA for the 2-port when used in this system. Express your
results in dB
(c) Find the power delivered to the 2-port and the power delivered to the load.
Express your result in dBm.
(d) How much power will be delivered to the load if a lossless matching network
is added between the output of the 2-port and the 500 load? The matching

8.6. HOMEWORK PROBLEMS

259

Figure 8.22: System with amplier equivalent circuit

network would be designed to transform 500 into Zout where Zout is the
output impedance of the 2-port with the 100 source termination. Express your
result in dBm.
13. Consider the 2-port from Problem 11. Suppose the reference impedance, Zo , is changed



, S21
, S12
to 500 . Denote the new S-parameters for this reference impedance by S11

and S22 .

.
(a) Find S11

. Hint: The voltage gain Av = Vo /Vi in a system with arbitrary load
(b) Find S21
and source impedance is given by the following formula:

Av =

S21 (1 + L )
(1 S22 L )(1 + IN )

(8.114)

where IN is the input reection coecient and Vo , Vi are the voltages measured
across the output and input terminals of the 2-port, respectively.
14. A 2-port has 50 S-parameters:
S11

0.3

S12
S21

=
=

0.01
10.0

S22

0.1

(8.115)

The Rollett Stability Factor for this 2-port is K = 4.525, and the 2-port is unconditionally stable.
(a) The 2-port is used with a source having impedance Zs = 300 and a load
ZL = 50 . The power available from the source is -3 dBm. Find the power
delivered to the load. Express your result in dBm.
(b) Now suppose a lossless matching network is used between the source and the
2-port. Find the power delivered to the load. Express your result in dBm.
(c) Suppose a lossless matching network is used between the 2-port and the load (no
matching network at the input). Find the power delivered to the load. Express
your result in dBm.

260

CHAPTER 8. 2-PORT SCATTERING (S) PARAMETERS

(d) Suppose both input and output matching networks are used so that both ports
are conjugately matched. Find the power delivered to the load. Express your
result in dBm.

15. Show that the maximum available gain GA,max for a unilateral 2-port is
GA,max =

|S21 |2
(1 |S11 |2 )(1 |S22 |2 )

(8.116)

Hint: When the input and output of the 2-port are conjugately matched, G = GT =
GA,max .
16. Consider a system with source and load reection coecients S and L as shown in
Figure 8.23.

Figure 8.23: System with source and load reection coecients S and L
Dene the insertion gain
GI =

Pdelivered to load with network

Pdelivered to load without network

(8.117)

where Pdelivered to load without network is the power delivered to the load when the
source is connected directly to the load, and Pdelivered to load with network is the
power delivered to the load when the 2-port is used between the source and the load.
(a) Find an expression for the insertion gain of an arbitrary 2-port.
(b) Find an expression for the insertion gain of a lossless matching network.
17. The T-parameters (sometimes called the chain scattering parameters) are dened
in terms of the same variables as the S-parameters. They are useful when cascading
2-ports, because the T-parameter matrix for a cascade is simply the product of the
T-matrices for the individual 2-ports. (Can you show that this is true?) The denition
of the T-parameters follows from choosing a1 and b1 to be the independent variables,
i.e.,

8.6. HOMEWORK PROBLEMS

261

a1 = T11 b2 + T12 a2

(8.118)

and
b1 = T21 b2 + T22 a2
where the as and bs are dened
ai =

Vi + Zo Ii

2 Zo

(8.119)

bi =

Vi Zo Ii

2 Zo

(8.120)

Suppose the S-parameters for a 2-port are known. Find the T-parameters in terms of
the S-parameters.
18. Consider a cascade of two identical 2-ports. The cascade is used in a system with
source and load as shown in Figure 8.24:

Figure 8.24: Cascade of two identical 2-ports

The available source power is -10dBm and the 50 S-parameters of the 2-ports are
S11
S12

=
=

0.35
0.1

S21
S22

=
=

3.0
0.50

(8.121)

The 2-ports are unconditionally stable. Answer the following questions. For parts 18a
- 18g, express all results in dB or dBm, whichever is appropriate.
(a) The power delivered to the rst 2-port.
(b) The power delivered to the second 2-port.
(c) The power delivered to the load.
(d) The transducer gain for the cascaded 2-ports.
(e) The available gain of the cascaded 2-ports.
(f) The operating power gain of the cascaded 2-ports.

262

CHAPTER 8. 2-PORT SCATTERING (S) PARAMETERS

(g) Suppose three lossless matching networks are used between the source and the
rst 2-port, the rst and second 2-ports, and the second 2-port and the load.
Find the power delivered to the load.
(h) Find S11 (Zo = 50) for the cascaded 2-ports. The answer is NOT 0.35!

19. Consider the system shown in Figure 8.25.

50

Amplier

[S]
Pavs = 0dBm

Lossless

Zo = 50
Z

Figure 8.25: 2-port.

The 2-port has the following S-parameters (Zo = 50 ):
S11 = 0.2, S21 = 5.0
S12 = 0.1, S22 = 0.5
(a) Find the impedance Z that will maximize the power delivered to the load, ZL .
(b) Find the power that will be delivered to the load, assuming that the matching
network has been designed to present the Z found in part 19a to the output of
the 2-port. Express your result in dBm.
(c) Find the impedance Z that will maximize the power delivered to the input of the
2-port.
(d) Find the power that will be delivered to the load, assuming that the matching
network has been designed to present the Z found in part 19c to the output of
the 2-port. Express your result in dBm.
20. In this problem, we prove that a particular set of passive terminations that cause the
negative resistance criterion for oscillation to be satised at one port will automatically
cause the conditions for oscillation to be satised at the other port if the 2 port is
bilateral (i.e. if S12
= 0 and S21
= 0).
(a) Show that the negative resistance criterion for steady-state oscillation (Z1 + Z2 =
0) can be written in terms of the corresponding reection coecients as 1 2 = 0.
(b) Now, suppose that L plane stability circle analysis shows that at least part of
unstable region lies within the unit circle in the L plane. We choose a load
reection coecient within the unstable region and also inside of the unit circle.
Denote this load termination by Lu . Find the reection coecient of the passive
source termination, Su , that will cause the conditions for steady state oscillation
to be satised at the input port. Write your answer in terms of the S parameters
and Lu .

8.6. HOMEWORK PROBLEMS

263

(c) Now, show that the conditions for oscillation are satised at the output port
when the 2-port is terminated with Lu and Su .
21. A 2-port has S parameters ( Zo = 50 ): S11 = 0.2, S12 = 0.0, S21 = 2.0, S22 = 0.8
(a) Suppose that the 2-port is used in a system with source impedance ZS = 50
and load impedance ZL = 50. The power available from the source Pavs = 0
dBm. Find the power that will be delivered to the load. Express your answer in
dBm.
(b) Calculate the input impedance of the 2-port when it is used in the system described in part a.
(c) This 2 port is unconditionally stable. Find the power delivered to the load when
lossless matching networks are used at the input and the output of the 2 port so
that the 2 port is simultaneously matched at both ports. Express your answer in
dBm.
(d) Find the power delivered to the load if 2 of these 2 ports are cascaded and the
cascade is used in between the source and load specied in part a. Express your
answer in dBm.
22. A 2 port has the following Z parameters (all given in ):
Z11 = 20 Z22 = 300 Z12 = 0.0 Z21 = 1000
Find S11 and S21 (Zo = 50 ). Hint: the input impedance and voltage gain of a 2
port can be written in terms of Z parameters and the load impedance as follows:
ZIN = Z11
Av =

Z12 Z21
ZL + Z22

V2
Z21 ZL
=
.
V1
Z11 ZL + Z11 Z22 Z12 Z21

23. In a particular system it is found that the operating, transducer, and available gains
of a 2-port are G = 10 dB, GT = 8 dB, GA = 14 dB. The 2-port is unilateral and is
unconditionally stable. The source impedance ZS = 100 and the power available
from the source is 3 dBm.
(a) Find the power delivered to the load. Express your result in Watts (NOT in
dBm! I want to see that you know the relationship between power in Watts and
dBm.)
(b) Find the power delivered to the 2-port by the source. Express your result in
dBm.
(c) Find the power that would be delivered to the load if a single lossless matching
network is used between the source and the 2-port. Express your result in dBm.
(d) Find the power that would be delivered to the load if a single lossless matching
network is used between the 2-port and the load. Express your result in dBm.
(e) Find the power that would be delivered to the load if lossless matching networks
are used at both the input and the output of the 2-port. Express your result in
dBm.

264

CHAPTER 8. 2-PORT SCATTERING (S) PARAMETERS

(f) Suppose it is known that the 2-port has S22 = 0 (Zo = 100 ) and calculate
the power that would be delivered to the load if a cascade of 2 of these identical
2-ports is used between the source and the load. Express your result in dBm.

24. In a system with Pavs = 0 dBm, ZS = 100 and unknown, but constant, ZL ,
it is empirically determined that the operating, transducer, and available gains of
a particular 2-port are G = 12 dB, GT = 8 dB, GA = 14 dB when this single 2port is used to couple the source to the load. The 2-port is known to be unilateral,
unconditionally stable, and to have S22 = 0 (Zo = 100).
(a) Calculate the power that would be delivered to the load if a cascade of 2 of these
identical 2-ports is used between the source and load described above. Express
your result in dBm.
(b) Suppose that a lossless matching network is added between the load and the
ouput of the second 2-port in the system described in part a. Determine the
power delivered to the load. Express your result in dBm.