MICROWAVE DESIGN PRINCIPLES

5
Impedance transformation, Impedance Matching, Microwave Filter Design, RF and
Microwave Amplifier Design, Microwave Power amplifier Design, Low Noise Amplifier
Design, Microwave Mixer Design, Microwave Oscillator Design

5.1 INTRODUCTION

In this chapter, the importance of impedance transformation in transmission lines is

discussed in the case of both lossless and lossy lines, followed by the importance of
impedance matching in transmission lines to achieve maximum power transfer by the use of
stubs and quarter-wave transformer. Then it is followed by the microwave design principles
such as filters, amplifiers, power amplifiers, low noise amplifiers, mixers, and oscillators
which are involved in several active research areas.

5.2 IMPEDANCE TRANSFORMATION

In the short-circuited transmission line, it was observed that the line became infinite at
a distance one-quarter wavelength from the short. The ability to change impedance by adding
a length of the transmission line is important to the microwave engineer. The general
impedance transformation equation involves hyperbolic functions with complex arguments.

5.2.1. Derivation of the impedance transformation equation

We know that the reflection coefficient is a function of position along the transmission
line

Γ =ΓL e-2γd (5.1)

Since the reflection coefficient is related to the impedance

Z= and Y=YO (5.2)

Impedance must also be a function of position.

5.2 Microwave Design Principles

This relationship is called the impedance transformation equation and will be derived
below.
From equations (5.1) and (5.2)

(5.3)

By using the equation

(5.4)

We can write

(5.5)

Rearranging the terms and use the trigonometric function we get the impedance
transformation equation as

or (5.6)

The transformation equation can be written in terms as admittance form as

or (5.7)

Where YO= 1/ZO, Y’=Y/YO, Y’L=YL/YO

The input impedance or admittance is obtained by merely setting d=1. In many

practical situations, the line attenuation can be neglected hence it is written for a lossless line
as =0, γd=jβd and tanhγd becomes jtanβd. Thus, for the lossless line

or (5.8)

and

or (5.9)

Several special cases of impedance transformation are examined below, all the results
obtained are directly applicable to a variety of microwave design problems. The cases are:
(i) Impedance transformation on a lossless line

(ii) Impedance transformation on a lossy line

 Antennas and Microwave Engineering 5.3

5.2.2. Impedance transformation on a lossless line

For =0, the impedance and admittance transformations are given by equations (5.8)
and (5.9). Let us now analyze some special cases of these equations

The Quarter wavelength line:

If the length of the line shown in fig.5.1 is an odd multiple of a quarter wavelength.
With l= (2n-1) λ/4, n being a positive integer, tanβl = infinite the equation (5.8) reduces to

or (5.10)

Fig 5.1 A transmission line with a load terminated in load ZL.

The above equation shows that a quarter-wave line transforms a large value of load
impedance to a small value at the input. Also, an inductive load is transformed into a
capacitive input impedance. Furthermore, it can be shown that if ZL is a series resonant circuit
Zin behaves like a parallel resonant circuit and vice versa. The quarter wavelength line is often
called as an impedance inverter since it inverts the normalized load impedance. It is also
referred to as a quarter-wave transformer since for ZL real it has all the properties of an ideal
transformer. The lossless quarter-wave line behaves like an ideal transformer with a turns
ratio of nt=5, the SWR value.

For ZL > ZO the line is equivalent to a transformer with nt<1 or N1< N2. The quarter-
wave transformer is useful in matching a resistive load to a generator which is a necessary
condition for delivering all the available generator power to the load. The circuit is shown
below in fig 5.2
5.4 Microwave Design Principles

Fig 5.2 A quarter-wave transformer and its equivalent circuit when ZL= RL.

If the generator impedance is real, then impedance Zin must be equal to RG for
maximum transfer of power. Solving equation (5.10) for ZO when the load is resistive (ZL=RL)
yields the required value of the characteristic impedance for the quarter-wave line.

ZO = √RGRL (5.11)

The short-circuited line – When a lossless line is shorted at the load end ZL=0 and thus the
input impedance is

Zin = jXin = jZOtanβl (5.12)

A plot of normalized input reactance versus βl is shown in fig.5.3. For βl<π/2 (l<λ/4)
the input impedance is inductive since Xin is positive. Thus

Xin=ωLeq and henries

Fig 5.3 Normalised input reactance versus βl for a short-circuited lossless transmission line
 Antennas and Microwave Engineering 5.5

For a fixed-length l, the characteristics represent the variation of a reactance with

frequency since βl=ωl/v. Near βl=π/2 and the behaviour of the shorted line is like that of the
resonant L-C circuits.

1. At frequencies in the vicinity of βl = (2n-1) π/2 the short-circuited line behaves

like a parallel resonant circuit.

2. At frequencies in the vicinity of βl = nπ, the short-circuited line behaves

like a series resonant circuit.

The open-circuited line - When a lossless line is open-circuited at the load end YL=0
and thus the input admittance is

Yin = jBin = jYOtanβl (5.13)

Since the input impedance is reciprocal of input admittance we can write,

Zin = -jZocotβl (5.14)

A plot of normalized input susceptance versus βl is the tangent function and is shown in
fig.5.4. For βl<π/2 (l<λ/4) the input is capacitive since Bin is positive.

Thus Bin= ωCeq and farads. (5.15)

Fig 5.4 Normalised input reactance versus βl for an open-circuited lossless transmission line
5.6 Microwave Design Principles

For a fixed-length l, the characteristics represent the variation of a reactance with

frequency since βl=ωl/v. Near βl=π/2 and the behaviour of the open-circuited line is similar
to that of the resonant L-C circuits.

1. At frequencies in the vicinity of βl = (2n-1) π/2 the open-circuited line behaves

like a series resonant circuit.

2. At frequencies in the vicinity of βl = nπ, the open-circuited line behaves

like a parallel resonant circuit.

These properties are used in the design of microwave filters.

By using the short circuited and open circuited conditions we get

ZO = √Zsc*Zoc (5.16)

5.2.3 Impedance transformation on a lossy line

For ≠0, the line is called as lossy line. The two examples are

The short-circuited line- For ZL=0, the equation is

Z=Zotanh( d+jβd) (5.17)

The input impedance is written as

in (5.18)

If =0, and βl=nπ then Zin=0. However, if the line is lossy, the input impedance is
infinite. For ≠0, and the short-circuited line is an integral number of half wavelengths then

Zin=Zo tanh l= Zo l ohms. (5.19)

If the line is an odd number of quarter wavelengths, tanβl=∞, For a lossless line this
results in an open circuit at the input. For a lossy line, we get

in ohm (5.20)

The open circuited line- In this case YL=0, the input impedance is given by

Yin= Yotanh( l+jβl) (5.21)

For βl=nπ, the input admittance is finite and real and is given by
 Antennas and Microwave Engineering 5.7

(5.22)

On the other hand if βl=(2n-1)π/2, the input admittance is given by

mhos (5.23)

Thus, for a lossless line, the resistance is negligible.

5.3 IMPEDANCE MATCHING:

The basic idea of impedance matching is illustrated in Fig.5.5, which shows an

impedance matching network placed between a load impedance and a transmission line. The
matching network is ideally lossless, to avoid unnecessary loss of power, and is usually
designed so that the impedance seen looking into the matching network is Z0. Then reflections
will be eliminated on the transmission line to the left of the matching network, although there
will usually be multiple reflections between the matching network and the load. This
procedure is sometimes referred to as tuning.

Fig 5.5 A lossless matching network an arbitrary load impedance to a transmission line

Impedance matching or tuning is important for the following reasons:

1. Maximum power is delivered when the load is matched to the line (assuming the
generator is matched), and power loss in the feed line is minimized.

2. Impedance matching sensitive receiver components (antenna, low-noise amplifier,

etc.) may improve the signal-to-noise ratio of the system.

3. Impedance matching in a power distribution network (such as an antenna array

feed network) may reduce amplitude and phase errors.

If the load impedance, ZL, has a positive real part, a matching network can always be
found. Many choices are available, however, and we will discuss the design and performance
of several types of practical matching networks. Factors that may be important in the selection
of a matching network include the following:
5.8 Microwave Design Principles

1. Complexity - As with most engineering solutions, the simplest design that satisfies the
required specifications is generally preferable. A simpler matching network is usually
cheaper, smaller, more reliable, and less lossy than a more complex design.

2. Bandwidth- Any type of matching network can ideally give a perfect match (zero
reflection) at a single frequency. In many applications, however, it is desirable to
match a load over a band of frequencies. There are several ways of doing this, with, of
course, a corresponding increase in complexity.

3. Implementation- Depending on the type of transmission line or waveguide being

used, one type of matching network may be preferable to another. For example, tuning
stubs are much easier to implement in waveguide than are multisection quarter-wave
transformers.

4. Adjustability- In some applications, the matching network may require adjustment to

match a variable load impedance. Some types of matching networks are more
amenable than others in this regard.

5.3.1 Matching with lumped elements (L networks)

The simplest type of matching network is the L-section, which uses two reactive
elements to match an arbitrary load impedance to a transmission line. There are two possible
configurations for this network, as shown in Fig.5.6. If the normalized load impedance is Z L
=ZL / ZO is inside the 1+jx circle on the smith chart, then the circuit of fig 5.6(a) should be
used.

Fig 5.6 L-section matching networks (a) Network for ZL inside the 1+jx circle
(b) Network for ZL outside the 1+jx circle
If the normalized load impedance is ZL =ZL / ZO is outside the 1+jx circle on the smith
chart, then the circuit of fig.5.6(b) should be used. The 1 + j x circle is the resistance circle on
the impedance Smith chart for which r = 1.
 Antennas and Microwave Engineering 5.9

In either of the configurations of Fig.5.6, the reactive elements may be either inductors
or capacitors, depending on the load impedance. Thus, there are eight distinct possibilities for
the matching circuit for various load impedances. If the frequency is low enough and/or the
circuit size is small enough, actual lumped-element capacitors and inductors can be used. This
may be feasible for frequencies up to about 1 GHz or so, although modern microwave
integrated circuits may be small enough such that lumped elements can be used at higher
frequencies as well.

There is, however, a large range of frequencies and circuit sizes where lumped
elements may not be realizable. This is a limitation of the L-section matching technique. We
will first derive analytic expressions for the matching network elements of the two cases in
Fig.5.6, and then illustrate an alternative design procedure using the Smith chart.

Analytic Solutions

Although we will discuss a simple graphical solution using the Smith chart, it is also
useful to have simple expressions for the L-section matching network components. These
expressions can be used in a computer-aided design program for L-section matching, or when
it is necessary to have more accuracy than the Smith chart can provide.

Consider first the circuit of Fig.5.6(a) and let ZL = RL + j XL. We stated that this circuit
would be used when zL = ZL/Z0 is inside the 1 + j x circle on the Smith chart, which implies
that RL > Z0 for this case. The impedance seen looking into the matching network, followed
by the load impedance, must be equal to Z0 for an impedance-matched condition:

(5.24)

Rearranging and separating into real and imaginary parts give two equations for the
two unknowns, X and B:

(5.25)

(5.26)

Solving (5.25) for X and substituting into (5.26) gives a quadratic equation for B. The
solution is

√ √
(5.27)

Note that since RL > Z0, the argument of the second square root is always positive.
Then the series reactance can be found as
5.10 Microwave Design Principles

(5.28)

Equation (5.27) indicates that the two solutions are possible for B and X. Both
solutions are physically realizable since both positive and negative values of B and X are
possible (positive X implies an inductor and negative X implies a capacitor, while positive B
implies a capacitor and negative B implies an inductor). One solution, however, may result in
significantly smaller values for the reactive components or maybe the preferred solution if the
bandwidth of the match is better, or if the SWR on the line between the matching network and
the load is smaller.

Consider the circuit of Fig.5.6(b). This circuit is used when ZL is outside the 1 + j x
circle on the Smith chart, which implies that RL < Z0. The admittance seen looking into the
matching network, followed by the load impedance, must be equal to 1/Z0 for an impedance-
matched condition:

(5.29)

Rearranging and separating into real and imaginary parts gives two equations for the
two unknowns, X and B:

(5.30)

(5.31)

Solving for X and B gives

√ (5.32)

√
(5.33)

Because RL < Z0, the arguments of the square roots are always positive. Again, note
that the two solutions are possible.

To match an arbitrary complex load to a line of characteristic impedance Z0, the real
part of the input impedance to the matching network must be Z0, while the imaginary part
must be zero. This implies that a general matching network must have at least two degrees of
freedom; in the L-section matching circuit these two degrees of freedom are provided by the
values of the two reactive components.

5.3.2 Single stub tuning

 Antennas and Microwave Engineering 5.11

Another popular matching technique uses a single open-circuited or short-circuited

length of a transmission line (a stub) connected either in parallel or in series with the
transmission feed line at a certain distance from the load, as shown in fig. 5.7. Such a single-
stub tuning circuit is often very convenient because the stub can be fabricated as part of the
transmission line media of the circuit, and lumped elements are avoided. Shunt stubs are
preferred for microstrip line or stripline, while series stubs are preferred for slot line or
coplanar waveguide.

In single stub tuning the two adjustable parameters are the distance, d, from the load to
the stub position, and the value of susceptance or reactance provided by the stub. For the
shunt-stub case, the basic idea is to select d so that the admittance, Y, seen looking into the
line at distance d from the load is of the form Y0 + j B. Then the stub susceptance is chosen as
−j B, resulting in a matched condition. For the series-stub case, the distance d is selected so
that the impedance, Z, seen looking into the line at a distance d from the load is of the form Z0
+ j X. Then the stub reactance is chosen as −j X, resulting in a matched condition.

Fig 5.7 Single-stub tuning circuits (a) Shunt stub (b) Series stub

For transmission line media such as microstrip or stripline, open-circuited stubs are
easier to fabricate since a via hole through the substrate to the ground plane is not needed. For
lines like coax or waveguide, however, short-circuited stubs are usually preferred because the
cross-sectional area of such an open-circuited line may be large enough (electrically) to
radiate, in which case the stub is no longer purely reactive.

 Shunt Stub: Here, a complex load ZL = RL ± jXL can be matched with a line of
characteristic impedance Z0 = R0 using a single shunt stub as shown in Fig.5.8. The
shunt-stub matching is better explained in terms of normalized admittance parameter
yL = YL/Y0 = gL ± jbL. It is assumed that the generator end is matched to the
5.12 Microwave Design Principles

characteristic impedance of the line. The following section describes the matching
procedure for single stub matching. To avoid complicated mathematics, graphical
solutions using Smith chart are widely used in the matching design.

Fig 5.8 Single shunt stub matching

In general, when the only load is connected, a voltage standing wave is formed in the
line. If we move away from the complex load towards the generator along the line, there are
voltage minima at which the reflection coefficient is a negative real quantity and the input
admittance is pure conductance of value Yin = (1 – G)/(1 + G) = g = S, where S is the load
VSWR.

Similarly, there are voltage maxima at which the admittance is pure conductance of
value Yin = (1 + G)/ (1 – G) = 1/S. Therefore, in between these two points, there must be a
point at a certain distance d(say) towards the generator from the first voltage minimum where
the real part of the normalized admittance is unity, so that yin = 1 ± jb.

The reactive component ±jb of this admittance can be canceled by an equal and
opposite susceptance -+jb of a short-circuited shunt stub of length l, located at

that point and having characteristic impedance equal to that of the line, i.e. Z0. The total
normalized admittance of the terminated line and the stub looking towards the load is then

so that (5.34)

for matching. Here,

(5.35)

and
 Antennas and Microwave Engineering 5.13

√ (5.36)

The following methods can be used in the design of a single shunt stub with the help
of the Smith chart of Fig.5.9.

1. The load admittance point is located at a point A corresponding to VSWR S. Join the
center O with A.

2. The location of the stub is found by rotating OA clockwise towards the generator with
O as a center, to intersect the unit resistance circle at B, where the real part of the
normalized admittance is 1.0.

The distance d / l corresponding to the arc AB gives the stub position as per Fig.5.9.

Fig 5.9 Single shunt stub matching using Smith chart

3. The normalized susceptance at B is read and an equal and opposite pure susceptance
value of the stub is located at C on the outer circle (g = 0).

4. The stub length is found by rotating in a clockwise direction around the unit circle to
C, starting from the right-hand intersection between the unit circle and the real axis
which corresponds to the short-circuit admittance position on the stub.

 Series Stub At voltage minimum, the input impedance of the line is Zin = 1/S, and at
the maximum, Zin = S. Therefore, in between these two points, at a distance d from the
load towards the generator, normalized input impedance zin = 1 ± jx. A short-circuited
5.14 Microwave Design Principles

series stub of length l can cancel the reactive part to provide a matching condition
.

Here (5.37)

√ (5.38)

The series stub is convenient for waveguide and stripline circuits. The shunt stub is
easy to construct for coaxial lines (also for two-wire lines).

5.3.3 Double Stub Matching

The single-stub tuner of the previous section can match any load impedance (having a
positive real part) to a transmission line but suffers from the disadvantage of requiring a
variable length of line between the load and the stub. This may not be a problem for a fixed
matching circuit but would probably pose some difficulty if an adjustable tuner was desired.

In this case, the double-stub tuner, which uses two tuning stubs in fixed positions, can
be used. Such tuners are often fabricated in coaxial lines with adjustable stubs connected in
shunt to the main coaxial line. We will see, however, that a double-stub tuner cannot match all
load impedances.

The double-stub tuner circuit is shown in Fig.5.12(a), where the load may be an
arbitrary distance from the first stub. Although this is more representative of a practical
situation, the circuit of Fig.5.12(b), where the load YL has been transformed back to the
position of the first stub, is easier to deal with and does not lose any generality.

The shunt stubs shown in Fig.5.7 can be conveniently implemented for some types of
transmission lines, while series stubs are more appropriate for other types of lines. In either
case, the stubs can be open-circuited or short-circuited.
 Antennas and Microwave Engineering 5.15

Fig.5.12 Double-stub tuning. (a) Original circuit with the load an arbitrary distance
from the first stub. (b) Equivalent circuit with the load transformed to the first stub.

Double-stub matching is conveniently designed and positioned using the Smith chart
and from the knowledge of the load ZL and its VSWR. The illustration of double-stub
matching is shown in Fig.5.13 where it is desired that the input admittance at the second stub
position C should be Y0 corresponding to a point on the unit circle (g = 1) of the Smith chart
having its center at O’.

The following steps are involved.

1. The normalized load admittance yL is located at A, diametrically opposite to load

impedance point ZL = ZL/Z0.

2. From the load admittance point A, a rotation towards generator on the load VSWR
circle by an arbitrary amount d/λg (= 0.06 say), gives the position B of the first stub
where the normalized admittance towards load for the line alone may be expressed as
yB = gB + jbB.

3. The unit conductance circle (g = 1) is rotated around the chart center O towards the
load by 3λg/8, where the new center of the circle becomes O’’ on the line passing +j
and –j susceptance points.

4. The point of intersection B’ between gB circle and the rotated unit conductance circle
with centre O’’ gives the total normalized admittance yBS of the combination of the line
at the right of the point B and the short-circuited first stub of length l1:

yBS = yB + jbS1 = gB + jbB + jbS1.

Here, jbS1 is the susceptance of the first stub alone which can be determined by
knowing yBS and yB from the Smith chart, and length l1/λg is determined from the
short circuit admittance point of stub 1.

5. From the point B’, rotates towards generator by 3λg/8 to reach point C on the original
g = 1 circle. The total normalized admittance at C for the original line and first stub is
then yC = 1 + jbC

6. For matching, the total admittance at C with the combination of (i) line, (ii) first stub,
and (iii) second stub should be yCS = 1.0 where the reactive point is canceled. This is
achieved by adjusting the length l2 of the second stub to provide an equal and opposite
susceptance of jbS2 = –jbC corresponding to the point C’ on the Smith chart.
5.16 Microwave Design Principles

7. The stub lengths are found by rotating towards the generator around the unit reflection
coefficient outermost circle, starting from the right-hand intersection of the unit circle
and the real axis, which corresponds to the origin of infinite susceptance (y = ∞) of the
short termination of the second stub, up to the point of corresponding stub susceptance
–jbC corresponding to the point C’.

Thus, finally, the impedance matching is done from the values of d, l1, and l2 read from
the Smith chart.

Fig 5.13 Double stub design using Smith chart

Limitation

The double-stub system cannot match all impedances. If the normalized load
conductance exceeds 1 / Sin2βs, the double stub cannot be used, since the matching condition
will not be satisfied for those values of load conductance, where s is the separation distance
between the stubs. Also, alternatively, if gL > 2, there will be no point of intersection between
the rotated circle and g = 1 circle, and no solution will be possible.
 Antennas and Microwave Engineering 5.17

5.3.4 Quarter Wave Transformer

The quarter-wave transformer is a simple and useful circuit for matching a real load
impedance to a transmission line. An additional feature of the quarter-wave transformer is that
it can be methodically extended to multisection designs to provide broader bandwidth. If only
a narrow band impedance match is required, a single-section transformer may suffice.
However, as we will see in the next few sections, multisection quarter-wave transformer
designs can be synthesized to yield optimum matching characteristics over the desired
frequency band.

One drawback of the quarter-wave transformer is that it can only match a real load
impedance. A complex load impedance can always be transformed into a real impedance,
however, by using an appropriate length of transmission line between the load and the
transformer, or an appropriate series or shunt reactive element. These techniques will usually
alter the frequency dependence of the load, and this often has the effect of reducing the
bandwidth of the match.

Here we will concentrate on the bandwidth performance of the transformer as a

function of the load mismatch. The single-section quarter-wave matching transformer circuit
is shown in Fig.5.15, with the characteristic impedance of the matching section given as

Z1 = √Z0ZL (5.39)

Fig 5.15 A single-section quarter-wave matching transformer l = λ0/4

at the design frequency f0.

At the design frequency, f0, the electrical length of the matching section is λ0/4, but at
other frequencies the length is different, so a perfect match is no longer obtained. We will
derive an approximate expression for the resulting impedance mismatch versus frequency.

The input impedance seen looking into the matching section is

5.18 Microwave Design Principles

in (5.40)

where t = tan βl = tan θ, and βl = θ = π/2 at the design frequency fo. The resulting
reflection coefficient is

(5.41)

Because Z12 = ZOZL the above reduces to

(5.42)
√

The reflection coefficient magnitude is

(5.43)

If we assume that the operating frequency is near the design frequency f0, then L=λ0/4
and θ = π/2. Then sec2 θ = 1, and (5.43) simplifies to

for near (5.44)

√

This result gives the approximate mismatch of the quarter-wave transformer near the
design frequency, as sketched in Fig.5.16.
 Antennas and Microwave Engineering 5.19

Fig 5.16 Approximate behavior of the reflection coefficient magnitude for a

single-section quarter-wave transformer operating near its design frequency.

The above results are strictly valid only for TEM lines. When non-TEM lines (such as
waveguides) are used, the propagation constant is no longer a linear function of frequency,
and the wave impedance will be frequency dependent. These factors serve to complicate the
general behavior of quarter-wave transformers for non-TEM lines, but in practice, the
bandwidth of the transformer is often small enough that these complications do not
substantially affect the result.

Another factor ignored in the above analysis is the effect of reactance’s associated
with discontinuities when there is a step-change in the dimensions of a transmission line. This
can often be compensated by making a small adjustment in the length of the matching section.
Fig.5.17 shows a plot of the reflection coefficient magnitude versus normalized frequency for
various mismatched loads. Note the trend of increased bandwidth for smaller load
mismatches.

The fractional BW is given by

√ (5.45)
√

Fractional bandwidth is usually expressed as a percentage, 100∆f/ f0%. Note that the
bandwidth of the transformer increases as ZL becomes closer to Z0 (a less mismatched load).
5.20 Microwave Design Principles

Fig 5.17 Reflection coefficient magnitude versus frequency for a single section
Quarter wave matching transformer with various load mismatches.

Example 5.1

Design a single-section quarter-wave matching transformer to match a 10 Ω load

to a 50 Ω transmission line at f0 = 3 GHz. Determine the percent bandwidth for which
the SWR ≤ 1.5.

Solution

The characteristic impedance of the matching section is

√ √

and the length of the matching section is λ/4 at 3 GHz (the physical length depends on the
dielectric constant of the line). An SWR of 1.5 corresponds to a reflection coefficient
magnitude of

The fractional bandwidth is computed as

√
√
 Antennas and Microwave Engineering 5.21

√
√

or

5.4 MICROWAVE FILTER DESIGN

Microwave filters are two-port, reciprocal, passive, linear devices that heavily
attenuate the unwanted signal frequencies while permitting transmission of wanted
frequencies. There are three types of construction of filters -a reflective filter which consists
of capacitive and inductive elements producing ideally zero reflection loss in the passband and
very high attenuation in the stopband, absorptive filters which dissipate the unwanted signal
internally and pass the wanted signal, and a lossy filter which uses lossy material in the filter
to heavily attenuate the rejected signal but permit passage with low loss to the wanted signal.

Here it describes mainly the reflective microwave filters constructed from purely
reactive elements. The practical filters have a small non-zero attenuation in the passband and
small-signal output in the attenuation or stopband due to the presence of resistive losses in
reactive elements and propagating medium.

A microwave filter is designed to operate between resistive source and load

impedances of 50 ohms in most microwave systems. Mostly, reactive elements L and C are
used, and the filter is assumed to be lossless with its nominal characteristic resistive
impedance R0. Applications can be found in virtually any type of RF or microwave
communication, radar, or test and measurement system.

The image parameter method of filter design was developed in the late 1930s and was
useful for low-frequency filters in radio and telephony. Filters designed using the image
parameter method consist of a cascade of simpler two-port filter sections to provide the
desired cutoff frequencies and attenuation characteristics but do not allow the specification of
a frequency response over the complete operating range.

Thus, although the procedure is relatively simple, the design of filters by the image
parameter method often must be iterated many times to achieve the desired results. A more
modern procedure, called the insertion loss method, uses network synthesis techniques to
design filters with a completely specified frequency response. The design is simplified by
beginning with low-pass filter prototypes that are normalized in terms of impedance and
frequency.

Transformations are then applied to convert the prototype designs to the desired
frequency range and impedance level. Both the image parameter and insertion loss methods of
filter design lead to circuits using lumped elements (capacitors and inductors). For microwave
5.22 Microwave Design Principles

applications, such designs usually must be modified to employ distributed elements consisting
of transmission line sections.

5.4.1 Filter Parameters

In designing a filter, the following important parameters are generally considered.

1. Pass-band width

2. Stop-band attenuation and frequencies

3. Input and output impedances

4. Return loss

5. Insertion loss

6. Group delay

The most important parameter among the above is the amplitude response given in
terms of the insertion loss vs frequency characteristics. Fig.5.18 shows a basic block diagram
of a filter fed by a generator and terminated by a load.

Fig 5.18 A filter between a generator and a load

Let Pi be the incident power at the filter input, Pr is the reflected power, P L is the
power passed on to the load. The insertion loss of the filter is then defined by, under load
match condition (RL = R0),

(5.46)

where PL = Pi – Pr if the filter is lossless and G is the voltage reflection coefficient given by
|Γ |2 = Pr /Pi.

The return loss of the filter is defined by

(5.47)

which quantifies the amount of impedance matching at the input port.

 Antennas and Microwave Engineering 5.23

The group delay is important for the multi-frequency or pulsed signals to determine
the frequency dispersion or deviation from constant group delay over a given frequency band
and is defined by

Where t is the transmission phase.

The most used microwave filters are the low-pass, high-pass, band-pass, and band-stop
filters. The electrical equivalent circuits of these filters along with typical insertion loss
response characteristics are shown in Fig.5.19.

Fig 5.19 Electrical equivalent circuits of filters and their responses: (a) low-pass

(b) high-pass (c) bandpass (d) band-stop

5.4.2 Filter design

Two-filter synthesis techniques are popular. These are the image parameter method
and the insertion loss method. Out of these two methods, only the insertion loss method gives
complete specifications of a physically realizable frequency characteristic over the entire pass
and the stop bands from which the microwave filters are synthesized or designed most
preferably.

Prototype Low-pass Design by Insertion Loss Method

The basic design of microwave filters of types low-pass, band-pass, and band-stop,
operating at arbitrary frequency bands and between arbitrary resistive loads, are from a
prototype low-pass design through some frequency transformation, element normalization and
the simulation of these elements using sections of the microwave transmission line. In this
5.24 Microwave Design Principles

method, a physically realizable network is synthesized that will give the desired insertion loss
vs frequency characteristics.

This method consists of the following steps:

1. Design of a prototype low-pass filter with the desired passband characteristics.

2. Transformation of this prototype network to the required type (low-pass, high-pass,

band-pass, or band-stop) filter with the specified center and band-edge frequencies.

3. Realization of the network in microwave form by using sections of microwave

transmission lines whose reactance correspond to those of distributed circuit elements.

Fig.5.20 shows a prototype low-pass circuit where the element values are designed
from standard low-pass response approximations such as Butterworth (or Maximally fl at or
binomial) pass band response and Chebyshev (or equal ripple) pass-band response. In the
basic prototype filter derived from these approximations, the angular cut-off frequency ωc*
and termination resistance r is both normalized to unity.

Fig 5.20 Prototype Low pass Filter

The Butterworth and Chebyshev response characteristics are shown in Fig.5.21 and are
described in the following paragraphs.

Butterworth Response

Insertion loss approximation for a low-pass Butterworth filter prototype is expressed

by

(5.48)

where the passband ranges from ω = 0 to ω = ωc. Butterworth approximation exhibits a flat
response in the passband and a monotonically increasing attenuation in the stopband. The
 Antennas and Microwave Engineering 5.25

maximum insertion loss in the passband is 3 dB at ωc so that am2 = 1. The rate of increase of
the insertion loss for ω > ωc depends on the exponent 2n, which in turn is related to the
number of filter sections used in a filter network of n reactive elements, Lx represents an
insertion loss at a given frequency ωx* = ωx/ωc in the stopband.

Fig 5.21 IL responses of filters: (a) Butterworth (b) Chebyshev Response

The approximation for a low-pass Chebyshev prototype is expressed by

IL = 1 + am2 Tn2 (ω2); ω’ = ω/ωc (5.49)

where n denotes the degree of approximation (i.e., number of reactive elements), and
am is the ripple factor.

Tn (ω’) is the Chebyshev polynomial of degree n given by

T1(x) = x,

T2(x) = 2x2 – 1,

T3(x) = 4x3 – 3x,

T4(x) = 8x4 – 8x2 + 1

Tn(x) = 2xTn–1(x) – Tn–2(x)= cos (n cos–1x) for x< = 1 ports and

= cosh (n cosh–1x) for x >= 1 stop band

where x = ω’. The insertion loss oscillates between 1 and 1 + am2 in the pass-band,
becomes 1 + am2 at the cut-off frequency ωc and increases monotonically beyond cut-off
(stopband ω’ > 1) at a much faster rate as compared with the Butterworth filter. Lx represents
5.26 Microwave Design Principles

insertion loss at a frequency ω’x = ωx/ωc in the stopband. The element values of low-pass
ladder network derived from Butterworth and Chebyshev responses for the normalized values
of cut-off angular frequency ω’c = 1, and load resistance r = unity, can be calculated from the
following equations:

Butterworth prototype element values

go = 1 = gn + 1 for all n

gk = 2 sin [(2k – 1) π/2n], k = 1, 2, ..., n (5.50)

Chebyshev prototype element values

go = 1 for all n

g1 = 2p1/sinh (β/2n)

gk =4pk-i pk / qk-I gk-j; k = 2, 3, 4, ...., n

gn + 1 = 1 for n odd

= coth2 (β/4) for even n (5.51)

where, pk =sin[(2k-1)π) / 2n , k = 1, 2, 3, ..., n

qk = sinh2(β/2n) + sin2(kπ/n), k = 1, 2, ..., n

β = ln [coth (Am/17.37)]; Am in dB (5.52)

Here, Am is the ripple height = am2. The values of gk can be obtained from tables of gk
as given below.

Table 5.1 Values of gk for Butterworth filter (IL = 3 dB at ωc = 1)

Antennas and Microwave Engineering 5.27
5.28 Microwave Design Principles

Table 5.2 Values of gk for Chebyshev filter 0.1 dB ripple

Table 5.3 0.2 dB ripple

Table 5.4 0.5 dB ripple

5.4.3 Filter Transformations from Prototype

To design actual low-pass, high-pass, band-pass and band-stop filters, the

transformations of the lowpass prototype filters with normalized cut-off frequency ω’c = 1 and
having the source and load resistance of 1 ohm are made into the desired type with required
source and load impedances using frequency and impedance transformations. These
procedures are described below.
 Antennas and Microwave Engineering 5.29

Low-pass Filters

For the design of a loss-pass filter from the low-pass prototype response parameters,
frequency scaling is required to change the normalized cut-off frequency to absolute cut-off
frequency ωc by dividing all inductances and capacitances by ωc without altering the
resistances. Impedance scaling is done by changing the source and load resistances from 1 to
RL (with RL = RG), multiplying all resistances and inductances by RL, and dividing all
capacitances by RL. The above two scaling are done simultaneously. Thus, the actual low-pass
filter element values are

farad (5.53)

henry (5.54)

High-pass Filters

The low-pass prototype network is transformed into a high-pass filter by transforming

series inductances into series capacitances and shunt capacitances into shunt inductances
using the frequency transformation

(5.55)

where ωc and ω are the band edge and variable angular frequencies of the high pass
filter. The element values are

(5.56)

(5.57)

Band-pass Filter

To map the low-pass prototype to a band-pass filter, the following frequency

transformation is used:

√ (5.58)

where, ω0, ω1, and ω2 – ω1 are center angular frequency, variable angular frequency,
and angular frequency bandwidth, respectively, and ω1 and ω2 are the angular frequency band
limits. Applying the frequency transformation to series inductances and shunt capacitances of
the low-pass prototype give
5.30 Microwave Design Principles

Series-tuned Series Elements

(5.59)

(5.60)

Shunt-tuned Shunt Elements

( ) (5.61)

(5.62)

Band-stop Filters
The transformation from low-pass prototype to band-stop is given by

(5.63)

where all the quantities are used in Eq. 5.63 are defined as in a band-pass filter. Here
series inductance is mapped into a shunt-tuned circuit in series with the generator and load
with element values

(5.64)

and shunt capacitance into a series-tuned circuit with element values

(5.65)

The above results are summarized in Table 5.5.

Table 5.5 Filter Element values

 Antennas and Microwave Engineering 5.31

5.4.4 Microwave Low Pass Filters

Low-pass filter elements can be designed from the low-pass prototype by using the
scaled Eqns 5.53 and 5.54. Low-pass filters are conveniently constructed using TEM
structures such as coaxial lines, strip lines, and microstrip lines. Microstrip filters are used in
satellite, airborne communication, and EW systems for their small size, lightweight, and low
cost. Coaxial line, stripline, and microstrip filters can be realized using such line sections
having appropriate impedances to simulate reactive elements.

The design is a good approximation of an idealized lumped-element circuit. In coaxial-

line and micro strip-line low-pass filters, the shunt capacitors can be realized by very short
sections (<< λg/4) of relatively low characteristic impedance line, and the series inductors by
short sections (< λg/4) of relatively high characteristic impedance. Fig.5.22 shows realization
of some filter elements. From the basic transmission-line theory, the length of the
predominantly inductive line and the predominantly capacitive line are obtained from the
input reactance and susceptance formulae of the filter section, respectively.

Fig 5.22 Microwave low-pass filters: (a) Coaxial π-section (b) Coaxial T-section
(c) Microstrip π-section (d) Microstrip T-section
5.32 Microwave Design Principles

Inductive Length

For inductive reactance XL of short length lL of a lossless line with short-circuit

termination,

ωcL = ZOLtanβIL , IL<λg/4 (5.66)

or (5.67)

Capacitive Length

For capacitive reactance Xc of a short length l c of lossless line with open-circuit

termination,

1/ωcC = ZOCCotβlc (5.68)

Or (5.69)

where lg represents the corresponding guide wavelength, L and C are the inductance
and capacitance required, and Z0 represents the characteristic impedance of the corresponding
section.

In practice, however, fringe capacitance Cf at the ends of the inductive line must be
considered and must be added to C where Cf is given by

( ) (5.70)

Waveguide form of low-pass filters cannot be realized since waveguides are high-pass
lines. A five-section low-pass microstrip filter configuration is shown in Fig.5.23

Fig 5.23 Five section Low pass microwave filter

 Antennas and Microwave Engineering 5.33

5.4.5 Microwave High Pass Filters

High-pass microwave filters can be designed from the low-pass prototype by using the
transformation (5.56) – (5.57) and the response characteristics of fig.5.24. In coaxial and
microstrip forms, short (< λg/4) lengths of relatively high characteristic impedance tee-
connected to the mainline, approximates the shunt inductors.

The series capacitors are obtained by very small (<< λg/4) gaps in the line as shown in
Fig.5.24. In waveguide form, the change in the broad wall dimension may be utilized to
simulate the lump equivalent circuit. The shunt inductor can be realized by an inductive iris or
by an H-plane T-junction.

Fig 5.24 High-pass filters: (a) π-section (b) T-sec on (c) Coaxial π-section
(d) Coaxial T-section (e) Microstrip π-sec on (f) Butterworth response of
prototype low-pass and high-pass
5.34 Microwave Design Principles

5.5. RF AND MICROWAVE AMPLIFIER DESIGN:

An amplifier is an electronic device that increases the voltage, current, or power of a

signal. Microwave amplifiers combine active and passive transmission line circuits to provide
functions critical to microwave systems and instruments. In earlier days, the design of
amplifiers relies on tubes such as Klystron tubes, Travelling wave tubes, and so on.

With the advent of solid-state devices, microwave amplifiers started using transistor
devices such as Si MOSFETs and GaAs MESFETs. Microwave transistor amplifiers can be
easily integrated with hybrid and monolithic integrated circuits because of its low cost,
reliability, and ruggedness. RF and microwave amplifiers can be used at higher frequencies.

RF and microwave amplifier design rely on the characteristics of transistors which can
be represented using S parameters. Transistor amplifiers can be used at frequencies more than
100 GHz in a wide range of applications requiring small size, low noise Figure, broad
bandwidth, and medium to high power capacity. To design the microwave amplifier, an
introduction regarding two port gain and stability concepts is required.

5.5.1 Two-port power gain

Consider a general two-port amplifier circuit connected to its source and load
impedances ZS and ZL as shown in fig.5.25. Let us develop the expressions for the three types
of power gain in terms of the scattering parameters of the two-port network and the reflection
S L of the source and load.

Fig 5.25 A two-port network with arbitrary source and load impedance

 Power Gain: It is the ratio of power dissipated in the load ZL to the power delivered
to the input of the two-port network.

G = PL / Pin

This gain is independent of ZS

 Antennas and Microwave Engineering 5.35

 Available Power Gain: It is the ratio of the power available from the two-port
network to the power available from the source.

GA = Pavn / Pavs

This gain depends on ZS but not on ZL

 Transducer Power Gain: It is the ratio of the power delivered to the load to the
power available from the source.

GT = PL / Pavs

This gain depends on both ZS and ZL

If the input and output are both conjugate matched to the two-port device, then the
gain is maximized and G = GA = GT.

The reflection coefficient seen looking towards load is

(5.71)

The reflection coefficient seen looking towards the source is

(5.72)

where Z0 is the characteristic impedance reference for the scattering parameters of the
two-port network.

The expressions for the refection coefficients looking towards the source and the load
and the power gains in terms of S parameters are as follows.

(5.73)

out
(5.74)

(5.75)
in in

(5.76)
5.36 Microwave Design Principles

(5.77)
avs in

Unilateral transducer power gain, GTU, where S12 = 0

(5.78)

Example 5.2

A silicon bipolar junction transistor has the following scattering parameters at

1.0 GHz, with a 50 Ω reference impedance: S11 = 0.38 ∟158◦, S12 = 0.11∟54◦, S21 =
3.50∟ 80◦, S22 = 0.40∟ −43◦. The source impedance is ZS = 25 Ω and the load
impedance is ZL = 40 Ω. Compute the power gain, the available power gain, and the
transducer power gain.

Solution:

The reflection coefficients at the source and load are

The reflection coefficient seen looking at the input and output of the terminated
network are

in

out

The power gain is,

The available power gain is,

out
 Antennas and Microwave Engineering 5.37

The transducer power gain is,

Single-stage microwave transistor amplifier

A single-stage microwave transistor amplifier can be modeled by the circuit shown

below.

Fig 5.26 A general transistor amplifier circuit

Separate effective gain factors for the input matching network and the output matching
networks are as follows

(5.79)

The overall transducer gain is then GT = GSG0GL. The effective gains GS and GL of the
matching networks may be greater than unity. This is because the unmatched transistor would
incur power loss due to reflections at the input and output of the transistor, and the matching
sections can reduce these losses.

If the transistor is unilateral, so that S12 = 0, the unilateral transducer gain reduces to
GTU = GSG0GL, where

(5.80)
5.38 Microwave Design Principles

5.6 STABILITY

The stability of a transistor amplifier depends on the source and load reflection
coefficients ΓS and ΓL respectively as Γin and Γout depends on source and load matching
networks.

Two types of Stability:

 Unconditional Stability : The network is unconditionally stable if |Γin| < 1 and|Γout | <
1 for all passive source and load impedances (i.e., |ΓS| < 1 and |ΓL| < 1).

 Conditional Stability (Potentially Unstable) : The network is conditionally stable if

|Γin| < 1 and |Γout | < 1 only for a certain range of passive source and load impedances.

Note that the stability condition of an amplifier circuit is usually frequency dependent
since the input and output matching networks generally depend on frequency. It is therefore
possible for an amplifier to be stable at its design frequency but unstable at other frequencies.

Stability Circles

For the amplifier to be unconditionally stable, the following conditions must be

satisfied by ΓS and ΓL

(5.81)

(5.82)
out

If the device is unilateral (S12 = 0), these conditions reduce to the simple results that
|S11| < 1 and |S22| < 1 are sufficient for unconditional stability. Finding this range for ΓS and ΓL
can be facilitated by using the Smith chart and plotting the input and output stability circles.

The stability circles are defined as the loci in the ΓL (or ΓS) plane for which |Γin| = 1
(or |Γout | = 1). The stability circles then define the boundaries between stable and potentially
unstable regions of ΓS and ΓL.

ΓS and ΓL must lie on the Smith chart (|ΓS | < 1, |ΓL | < 1 for passive matching
networks).

The determinant of the scattering matrix is represented as Δ

Δ = S11S22 – S12S21
 Antennas and Microwave Engineering 5.39

In the complex Γ plane, an equation of the form |Γ − C| = R represents a circle having

a centre at C (a complex number) and a radius R (a real number). The output stability circle
with a center CL and radius RL, where

(5.83)

The input stability circle by interchanging S11 and S22

(5.84)

Given the scattering parameters of the transistor, we can plot the input and output
stability circles to define where |Γin| = 1 and |Γout| = 1. On one side of the input stability
circle, we will have |Γout| < 1, while on the other side we will have |Γout| > 1. Similarly, we
will have |Γin| < 1 on one side of the output stability circle, and |Γin| > 1 on the other side. We
need to determine which areas on the Smith chart represent the stable region, for which |Γin| <
1 and |Γout| < 1.

Consider the output stability circles plotted in the ΓL plane for | S11| < 1 and | S11| > 1,
as shown in Fig.5.27. If we set ZL = Z0, then ΓL = 0, and fig.5.27(a) shows that |Γin| = | S11|.
Now if |S11| < 1, then |Γin| < 1, so ΓL = 0 must be in a stable region. This means that the center
of the Smith chart (ΓL = 0) is in the stable region, so all of the Smith chart (|ΓL | < 1) that is
exterior to the stability circle defines the stable range for ΓL. This region is shaded in Fig.5.27.

Alternatively, if we set ZL = Z0 but have | S11| > 1, then |Γin| > 1 for ΓL = 0, and the
center of the Smith chart must be in an unstable region. In this case, the stable region is the
inside region of the stability circle that intersects the Smith chart. Similar results apply to the
input stability circle. If the device is unconditionally stable, the stability circles must be
completely outside (or enclose) the Smith chart. It can be stated mathematically as

||CL| - RL| > 1 for |S11| < 1,

||CS| - RS| > 1 for |S22| < 1,

5.40 Microwave Design Principles

Fig 5.27 Output stability circles for a conditionally stable device

(a) |S11| < 1 (b) |S11| > 1

If | S11| > 1 or | S22| > 1, the amplifier cannot be unconditionally stable because we can
always have a source or load impedance of Z0 leading to ΓS = 0 or ΓL = 0, thus causing |Γin| >
1 or |Γout| S L must
be chosen in stable regions, and it is good practice to check stability at several frequencies
over the range where the device operates.

Tests for Unconditional Stability

K- Δ test:

A device will be unconditionally stable if Rollet’s condition, defined as

(5.85)

along with the auxiliary condition that

|Δ| = |S11S22 – S12S21| < 1

are simultaneously satisfied.

The K −Δ test is a mathematically rigorous condition for unconditional stability, it

cannot be used to compare the relative stability of two or more devices because it involves
constraints on two separate parameters. Recently, however, a new criterion has been proposed
that combines the scattering parameters in a test involving only a single parameter, μ, defined
as

(5.86)
 Antennas and Microwave Engineering 5.41

Thus, if μ > 1, the device is unconditionally stable. Also, it can be said that larger
values of μ imply greater stability.

Example 5.3

Determine the stability of this transistor by using the K −Δ test and the μ-test and
plot the stability circles on a Smith chart. The Triquint T1G6000528 GaN HEMT has
the following scattering parameters at 1.9 GHz (Z0 = 50Ω):

S11 = 0.869∟ −159◦,

S12 = 0.031∟ −9◦,

S21 = 4.250∟ 61◦,

S22 = 0.507∟ −117◦

Solution:

K and Δ can be computed as,

|Δ| = |S11S22 – S12S21| = 0.336

Thus, we have |Δ| < 1 but not K > 1, so the unconditional stability criteria are not
satisfied, and the device is potentially unstable. The stability of this device can also be
evaluated using the μ-test, which is found to be μ = 0.678, again indicating potential
instability.

The centres and radii of the stability circles are given by

5.42 Microwave Design Principles

L = 1.59∟132˚

RL = 0.915

CS = 1.09∟162˚

RS = 0.205

These data can be used to plot the input and output stability circles, as shown in
Fig.5.28. Since |S11| < 1 and |S22| < 1, the central part of the Smith chart represents the stable
operating region for ΓS and ΓL. The unstable regions are shaded.

Fig 5.28 Stability Circle

 Antennas and Microwave Engineering 5.43

5.7 SINGLE-STAGE TRANSISTOR AMPLIFIER DESIGN

Design for Maximum Gain (Conjugate Matching)

After the stability of the transistor has been determined and the stable regions for ΓS
and ΓL have been located on the Smith chart, the input and output matching sections can be
designed. Since G0 is fixed for a given transistor, the overall transducer gain of the amplifier
will be controlled by the gains, GS and GL, of the matching sections. The maximum gain will
be realized when these sections provide a conjugate match between the amplifier source or
load impedance and the transistor. Because most transistors exhibit a significant impedance
mismatch (large |S11| and |S22|), the resulting frequency response may be narrowband.

The Maximum power transfer from the input matching network to the transistor will
occur when

Γin = ΓS*

The maximum power transfer from the transistor to the output matching network will
occur when

Γout = ΓL*

With the assumption of lossless matching sections, these conditions will maximize the
overall transducer gain, this maximum gain will be given by,

(5.87)

Also, with conjugate matching and lossless matching sections, the input and output
ports of the amplifier will be matched to Z0. In the general case with a bilateral (S12 ≠ 0)
transistor, Γin is affected by Γout and vice versa, so the input and output sections must be
matched simultaneously.

Having the value of

|Δ| = |S11S22 – S12S21| < 1 (5.88)

Using the result that

Δ (S11*S22* – S12*S21*) = |Δ|2 (5.89)

5.44 Microwave Design Principles

S,

(S11 – ΔS22*)Γs2 + (|Δ|2 - |S11|2 + |S22|2 -1) Γs + (S11* - Δ*S22) = 0 (5.90)

The solution is,

L can be written as,

The variables B1,C1, B2,C2 are defined as

B1 = 1 + |S11|2 - |S22|2 - |Δ|2

B2 = 1 + |S22|2 - |S11|2 - |Δ|2

C1 = S11 – ΔS22*

C2 = S22 – ΔS11*

Solutions to ΓL are only possible if the quantity within the square root is positive, and
it can be shown that this is equivalent to requiring K > 1. Thus, unconditionally stable devices
can always be conjugately matched for maximum gain, and potentially unstable devices can
be conjugately matched if K > 1 and |Δ| < 1. The results are much simpler for the unilateral
case. When S12 = 0, it is shown that ΓS = S 11 and ΓL = S 22, and then maximum transducer
gain reduces to

(5.91)

The maximum transducer power gain occurs when the source and load are conjugately
matched to the transistor. If the transistor is unconditionally stable, so that K > 1, the
maximum transducer power gain can be simply rewritten as follows:

√ (5.92)

The maximum transducer power gain is also sometimes referred to as the matched
gain. The maximum gain does not provide a meaningful result if the device is only
conditionally stable since the simultaneous conjugate matching of the source and load is not
 Antennas and Microwave Engineering 5.45

possible if K < 1. In this case, a useful Figure of merit is the maximum stable gain, defined as
the maximum transducer power gain with K = 1. Thus,

Gmsg = |S21| / |S12| (5.93)

The maximum stable gain is easy to compute and offers a convenient way to compare
the gain of various devices under stable operating conditions.

Example 5.4

Design an amplifier for maximum gain at 4 GHz using single-stub matching

sections. Calculate and plot the input return loss and the gain from 3 to 5 GHz. The
transistor is a GaAs MESFET with the following scattering parameters (Z0 =50 Ω):

f(GHz) S11 S12 S21 S22

3.0 0.80∟-89˚ 0.03∟56˚ 2.86∟99˚ 0.76∟-41˚

4.0 0.72∟-116˚ 0.03∟57˚ 2.60∟76˚ 0.73∟-54˚

5.0 0.66∟-142˚ 0.03∟62˚ 2.39∟54˚ 0.72∟-68˚

Solution:

Calculate K and Δ from the scattering parameters at each frequency in the above table
gives the following results:

f (GHz) K Δ

3.0 0.77 0.592

4.0 1.19 0.487

5.0 1.53 0.418

It is found that K > 1 and |Δ| < 1 at 4 and 5 GHz, so the transistor is unconditionally
stable at these frequencies, but it is only conditionally stable at 3 GHz. We can proceed with
the design at 4 GHz but should check stability at 3 GHz after we find the matching networks
(which determine ΓS and ΓL ). For maximum gain, we should design the matching sections for
a conjugate match to the transistor. Thus ΓS = Γin* and ΓL = Γout*
5.46 Microwave Design Principles

Source and load reflection coefficients can be determined as,

B1 = 1 + |S11|2 - |S22|2 - |Δ|2

B2 = 1 + |S22|2 - |S11|2 - |Δ|2

C1 = S11 – ΔS22*

C2 = S22 – ΔS11*

Overall transducer gain is, GTmax = 6.20 + 8.30 + 2.22 = 16.7 dB

The matching networks can easily be determined using the Smith chart. For the input
S, as shown in Fig.5.29. The impedance, ZS, represented by this
reflection coefficient is the impedance seen looking into the matching section toward the
source impedance, Z0. Thus, the matching section must transform Z0 to the impedance ZS.
There are several ways of doing this, but we will use an open-circuited shunt stub followed by
a length of line. We convert to the normalized admittance ys and work backward (toward the
load on the Smith chart) to find that a line of length 0.120λ will bring us to the 1 + jb circle.
Then we see that the required stub admittance is +j3.5, for an open-circuited stub length of
0.206λ. A similar procedure gives a line length of 0.206λ and a stub length of 0.206λ for the
output matching circuit.
 Antennas and Microwave Engineering 5.47

Fig 5.29 Smith Chart for the design of the input matching network

The final amplifier circuit is shown in Fig.5.30. This circuit only shows the RF
components; the amplifier will also require bias circuitry.

Fig 5.30 RF Circuit

In Fig.5.31, it is shown that the expected gain of 16.7 dB at 4 GHz, with a very good
return loss. The bandwidth where the gain drops by 1 dB is about 2.5%.
5.48 Microwave Design Principles

Fig 5.31 Frequency Response

Note that the matching sections are frequency-dependent, so the impedances and
reflection coefficients are different at 3 GHz than their design values at 4 GHz.

Constant-Gain Circles and Design for Specified Gain

In many cases it is preferable to design for less than the maximum obtainable gain, to
improve bandwidth, or to obtain a specific value of amplifier gain. This can be done by
designing the input and output matching sections to have less than maximum gains; in other
words, mismatches are purposely introduced to reduce the overall gain. The design procedure
is facilitated by plotting constant-gain circles on the Smith chart to represent loci of ΓS and ΓL
that give fixed values of gain (GS and GL ). To simplify our discussion, we will only treat the
case of a unilateral device; the more general case of a bilateral device must sometimes be
considered in practice.

For many transistors |S12| is small enough to be ignored, and the device can be
assumed to be unilateral. This greatly simplifies the design procedure. The error in the
transducer gain caused by approximating |S12| as zero is given by the ratio GT/GTU. It can be
shown that this ratio is bounded by

1/(1+U)2 < GT / GTU < 1/(1-U)2 (5.94)

where U is defined as the unilateral Figure of merit,

U = |S12||S21||S11||S22| / (1- |S11|2) (1-|S22|2) (5.95)

Usually, an error of a few tenths of a dB or less justifies the unilateral assumption.

 Antennas and Microwave Engineering 5.49

The expression for GS and GL for the unilateral case is given by

Gs = 1-|Γs|2 /|1-S11Γs|2 (5.96)

GL = 1-|ΓL|2 /|1-S22ΓL|2 (5.97)

S =S 11 L = S 22, resulting in the maximum

values given by

GSmax = 1/1-|S11|2 (5.98)

GLmax = 1/1-|S22|2 (5.99)

Define normalized gain factors gS and gL as

(5.100)
ant

Then we have that 0 ≤gS ≤ 1 and 0 ≤gL ≤ 1.

For fixed values of gS and gL S L plane. To show this,

the equation can be expanded as

(5.101)

(5.102)

(5.103)

Simplifying gives

√
(5.104)

which is the equation of a circle with its center and radius given by

(5.105)
√
5.50 Microwave Design Principles

The results for the constant gain circles of the output section is given as

(5.106)
√

The centres of each family of circles lie along straight lines given by the angle of S 11
or S 22. Note that when gS (or gL ) = 1 (maximum gain), the radius RS (or RL ) = 0, and the
centre reduces to S 11 (or S 22), as expected. Also, it can be shown that the 0 dB gain circles
(GS = 1 or GL = 1) will always pass through the center of the Smith chart. These results can be
used to plot a family of circles of constant gain for the input and output sections. Then ΓS and
ΓL can be chosen along these circles to provide the desired gains. The choices for ΓS and ΓL
are not unique, but it makes sense to choose points close to the centre of the Smith chart to
minimize mismatch, and thus maximize bandwidth.

Example 5.5

Design an amplifier to have a gain of 11 dB at 4.0 GHz. Plot constant-gain circles

for GS = 2 and 3 dB, and GL = 0 and 1 dB. Calculate and plot the input return loss and
overall amplifier gain from 3 to 5 GHz. The transistor has the following scattering
parameters (Z0 = 50 Ω):

f(GHz) S11 S12 S21 S22

3 0.80∟-90˚ 0 2.8∟100˚ 0.66∟-50˚

4 0.75∟-120˚ 0 2.5∟80˚ 0.60∟-70˚

5 0.71∟-140˚ 0 2.3∟60˚ 0.58∟-85˚

Solution:

Since S12 = 0 and |S11| < 1 and |S22| < 1, the transistor is unilateral and unconditionally
stable at each frequency in the above table.

Maximum matching section gains can be calculated as

GSmax = 1/1-|S11|2 = 2.29 = 3.6 dB

GLmax = 1/1-|S22|2 = 1.56 = 1.9 dB

 Antennas and Microwave Engineering 5.51

The gain of the mismatched transistor is

Go = |S21|2 = 6.25 = 8.0 dB

The maximum unilateral transducer gain is,

GTUmax = 3.6 + 1.9 + 8.0 = 13.5

We, therefore, have 2.5 dB more available gain than is required by the specifications.

Gs = 1-|Γs|2 /|1-S11Γs|2

GL = 1-|ΓL|2 /|1-S22ΓL|2

ant

Using the above relations, the following data’s are calculated for constant - gain
circles.

Gs = 3 dB gs = 0.875 Cs = 0.706∟120˚ Rs = 0.166

Gs = 2 dB gs = 0.691 Cs = 0.627∟120˚ Rs = 0.294

GL = 1 dB gL = 0.806 CL = 0.520∟70˚ RL = 0.303

GL = 0 dB gL = 0.640 CL = 0.440∟70˚ RL = 0.440

5.52 Microwave Design Principles

The constant-gain circles are shown in Fig.5.32.

Fig. 5.32 Constant gain circles

We choose GS = 2 dB and GL = 1 dB, for an overall amplifier gain of 11 dB. Then we

select ΓS and ΓL along with these circles as shown, to minimize the distance from the center of
the chart (these places ΓS and ΓL along the radial lines at 120◦ and 70◦, respectively). Thus, ΓS
= 0.33∟120◦ and ΓL = 0.22∟ 70◦, and the matching networks can be designed using shunt
stubs.

The final amplifier circuit is shown in Fig.5.33.

Fig 5.33 RF Circuit

 Antennas and Microwave Engineering 5.53

In Fig.5.34, it is seen the desired gain of 11 dB is achieved at 4.0 GHz. The bandwidth
over which the gain varies by ±1 dB or less is about 25%, which is considerably better than
the bandwidth of the maximum gain design.

Fig.5.34 Transducer gain and return loss

The return loss, however, is not very good, being only about 5 dB at the design
frequency. This is due to the deliberate mismatch introduced into the matching sections to
achieve the specified gain.

5.8 LOW NOISE AMPLIFIER DESIGN

One of the design considerations of a microwave amplifier is its noise Figure.

The noise Figure helps determine the efficiency of a particular Low Noise Amplifier. LNA
suitability for an application is typically based on its noise Figure. In general, a low noise
Figure results in better signal reception. The first stage of a receiver front end has the
dominant effect on the noise performance of the overall system. Usually there exist a trade-off
between gain and noise and hence certain compromise must be done. This can be done by
using constant-gain circles and circles of constant noise Figure.

The noise Figure of a two-port amplifier can be expressed as

F = Fmin + |Ys – Yopt|2 (5.107)

where the following definitions apply:

YS = GS + j BS = source admittance presented to transistor.

Yopt = optimum source admittance that results in minimum noise Figure

Fmin = minimum noise Figure of transistor, attained when YS = Yopt.

RN = equivalent noise resistance of the transistor.

5.54 Microwave Design Principles

GS = real part of source admittance.

Instead of the admittance YS and Yopt, we can use the reflection coefficients ΓS and
Γopt, where

Ys = 1/Z0 (1-Γs / 1+Γs) (5.108)

Yopt = 1/Z0 (1-Γopt / 1+Γopt) (5.109)

ΓS is the source reflection coefficient. The quantities Fmin, Γopt, and RN are
characteristics of the transistor being used and are called the noise parameters of the device;
they may be given by the manufacturer or measured. We can express the quantity |YS − Yopt|2
in terms of ΓS and Γopt:

opt
opt (5.110)
opt

Also

(5.111)

The noise Figure is,

opt
(5.112)
opt

S
plane. First define the noise Figure parameter, N, as

opt
opt (5.113)

which is a constant for a given noise Figure and set of noise parameters. Then rewrite

(Γs - Γopt) (Γs* - Γopt*) = N(1-|Γs|2)

ΓsΓs* - (ΓsΓopt* + Γs*Γopt) + Γopt Γopt* = N – N |Γs|2 (5.114)

opt opt opt

(5.115)

Add |Γopt|2/ (N + 1)2 to both sides to complete the square to obtain

√ opt
opt
(5.116)
 Antennas and Microwave Engineering 5.55

This result defines circles of constant noise Fig. with centres at

CF = Γopt / N+1

and radii of

√ opt
(5.117)

Example 5.6

A GaAs MESFET is biased for minimum noise Figure with the following
scattering parameters and noise parameters at 4 GHz (Z0 = 50 Ω):

S11 = 0.6∟-60◦, S12 = 0.05∟ 26◦, S21 = 1.9∟ 81◦, S22 = 0.5∟ −60◦, Fmin = 1.6 dB,
Γopt = 0.62∟100◦, and RN = 20 Ω. For design purposes, assume the device is unilateral,
and calculate the maximum error in GT resulting from this assumption. Then design an
amplifier having a 2.0 dB noise Fig. with the maximum gain that is compatible with this
noise Figure.

Solution:

Δ = S11S22 – S12S21

K = 1-|S11|2 - |S22|2 + |Δ|2 / 2|S12S21|

Using the above formulas, K and Δ are calculated as,

K = 2.78

Δ = 0.37

Since K >1, Δ <1, the device is unconditionally stable even without the approximation
of a unilateral device.

Unilateral Figure of merit is,

U = |S12S21S11S22| / (1-|S11|2 ) ( 1-|S22|2) = 0.059

The ration GT/GTU is bounded as,

1/(1+U)2 < GT / GTU < 1/(1-U)2

Or

0.891 < GT / GTU < 1.130

5.56 Microwave Design Principles

In dB ,

-0.50 < GT - GTU < 0.53 dB

Thus, we should expect less than about ±0.5 dB error in gain. The center and radius of
the 2 dB noise Figure circle is calculated as,

opt
opt

N = 0.0986

CF

CF = 0.56∟100˚

√ opt

RF = 0.24

This noise Figure circle is plotted in Fig.5.35

Fig 5.35 Constant – gain and Constant – noise Figure circles

 Antennas and Microwave Engineering 5.57

We see that the GS = 1.7 dB gain circle just intersects the F = 2 dB noise Figure circle
and that any higher gain will result in a worse noise Figure. From the Smith chart the
S = 0.53∟ 75◦, yielding GS = 1.7 dB and F = 2.0 dB.

For the output section, L =S 22 = 0.5∟60◦ for a maximum GL of

GL = 1 / 1-|S22|2 = 1.33 = 1.25 dB

The transistor gain is

GO = |S21|2 = 3.61 = 5.58 dB

The Overall transducer gain is,

GTU = Gs + Go +GL = 8.53 dB

A complete AC circuit for the amplifier, using open-circuited shunt stubs in the
matching sections is shown in Fig.5.36. A computer analysis of the circuit gives a gain of 8.36
dB.

Fig 5.36 RF Circuit

Low-Noise MOSFET Amplifier

MOSFETs have a relatively low AC input resistance, making them difficult to

impedance match. External series resistance can be added to the gate, but this approach
increases noise power and degrades efficiency. By using a series inductor at the source of a
MOSFET however, it is possible to create a resistive input impedance without adding noisy
resistors.

This technique is called inductive source degeneration; similar methods can be used
with MESFETs and other transistors. The conceptual circuit is shown in Fig. 5.37(a), where
the inductor Ls is placed in series with the source of the device. The equivalent circuit of the
amplifier is shown in Fig.5.37(b), where we have simplified the model by assuming the
transistor is unilateral, and that Ri , Rds , and Cds can be ignored.
5.58 Microwave Design Principles

(a) (b)

Fig 5.37 Low-noise MOSFET amplifier. (a) Basic AC circuit.

(b) An equivalent circuit using a simplified unilateral FET model

For an input current I at the gate of the transistor, the capacitor voltage is
Vc = I/jωCgs . The gate voltage, relative to the ground, is then

(5.118)

The input impedance at the gate is

Z= = +j( ) (5.119)

showing that the circuit has produced an input resistance of gm Ls/Cgs. The series
inductor, Ls, can be chosen to match the input resistance of the amplifier to a source
impedance, Z0. The inductor at the gate, Lg, can then be chosen to cancel the residual input
reactance, which is usually capacitive. The combination of the series matching inductor, the
gate capacitance, and the effective input resistance forms a series RLC resonator. The Q of
this resonator is

Q= (5.120)

The bandwidth of this circuit may be relatively narrow if this Q is high. LNAs are used
in communication receivers such as in cellular telephones, GPS receivers, wireless LANs, and
Satellite communications.
 Antennas and Microwave Engineering 5.59

Example 5.7

An Infineon BF1005 n-channel MOSFET transistor having Cgs = 2.1 pF and gm

= 24 mS is used in a 900 MHz low-noise amplifier with inductive source degeneration, as
shown in Fig.5.37. Determine the source and gate inductors and estimate the bandwidth
of the amplifier. Assume a source impedance of Z0 = 50 Ω.

Solution:

Matching the input resistance to Z0 determines the source inductor as,

( )
Ls = = = 4.37 nH

The net reactance at the input is jX = j(ωLs – 1/ωCgs) = -j59.5 Ω, So the required
series inductance for matching is,

Lg = = = 10.5 nH

Q can be calculated as, Q= = 1.2

So, the bandwidth of the amplifier could be as high as 80%.

5.9. POWER AMPLIFIER DESIGN:

Power amplifiers are used in the final stages of radar and radio transmitters to increase
the radiated power level. Important considerations for RF and microwave power amplifiers
are efficiency, gain, intermodulation distortion, and thermal effects. Single transistors can
provide output powers of 10–100 W at UHF frequencies, while devices at higher frequencies
are generally limited to output powers less than 10 W. Various power-combining techniques
can be used in conjunction with multiple transistors if higher output powers are required. For
high input powers, transistors do not behave linearly.

In this case, the impedance seen at the input and output of the transistor will depend on
the input power level, and this greatly complicates the design of power amplifiers. The power
amplifier is usually the primary consumer of DC power in most hand-held wireless devices, so
amplifier efficiency is an important consideration. One measure of amplifier efficiency is the
ratio of RF output power to DC input power:

η = Pout/ PDC where η is the drain efficiency (or collector efficiency).

One drawback of this definition is that it does not account for the RF power delivered
at the input to the amplifier. Since most power amplifiers have relatively low gains, this tends
5.60 Microwave Design Principles

to overrate the actual efficiency. A better measure that includes the effect of input power is the
power added efficiency, defined as

ηPAE = PAE = =( ) =( )η (5.121)

where G is the power gain of the amplifier.

Another useful parameter for power amplifiers is the compressed gain, G1. It is defined
as the gain of the amplifier at the 1 dB compression point. Thus, if G0 is the small-signal
(linear) power gain, we have

G1(dB) = G0(dB) – 1 (5.122)

Nonlinearities can lead to the generation of spurious frequencies and intermodulation

distortion. This can be a serious issue in wireless transmitters, especially in a multicarrier
system, where spurious signals may appear in adjacent channels. Linearity is also critical for
non-constant envelope modulations, such as amplitude shift keying and higher-order
quadrature amplitude modulation methods.

5.9.1. Class A amplifiers:

They are inherently linear circuits, where the transistor is biased to conduct over the
entire range of the input signal cycle. Because of this, class A amplifiers have a theoretical
maximum efficiency of 50%. Most small-signal and low-noise amplifiers operate as class A
circuits.

5.9.2. Class B amplifiers:

The transistor in a Class B amplifier is biased to conduct only during one-half of the
input signal cycle. Usually, two complementary transistors are operated in a class B push-pull
amplifier to provide amplification over the entire cycle. The theoretical efficiency of a class B
amplifier is 78%.

5.9.3. Class C amplifiers:

They are operated with the transistor near cut off for more than half of the input signal
cycle, and generally use a resonant circuit in the output stage to recover the fundamental.
Class C amplifiers can achieve efficiencies near 100% but can only be used with constant
envelope modulations.

Higher classes, such as class D, E, F, and S, use the transistor as a switch to pump a
highly resonant tank circuit and may achieve very high efficiencies.
 Antennas and Microwave Engineering 5.61

Many communication transmitters operating at UHF frequencies or above rely on

class A, AB, or B power amplifiers because of the need for low distortion products. A
transistor behaves linearly for signal powers well below the 1 dB compression point (IP1 dB),
and so the small-signal scattering parameters should not depend on either the input power
level or the output termination impedance. However, for power levels comparable to or
greater than IP1 dB, where the nonlinearity of the transistor becomes apparent, the measured
scattering parameters will depend on input power level, output termination impedance,
frequency, bias conditions, and temperature. Thus large-signal scattering parameters are not
uniquely defined and do not satisfy linearity and cannot be used in place of small-signal
parameters.

A more useful way to characterize transistors under large-signal operating conditions

is to measure the gain and output power as a function of source and load impedances. One
way of doing this is to determine the large- SP
and ΓLP, that maximize power gain for particular output power.

Another way of characterizing the large-signal behaviour of a transistor is to plot

contours of constant power output on a Smith chart as a function of the load reflection
coefficient, ΓLP, with the transistor conjugately matched at its input. These are called load-pull
contours and they can be obtained using an automated measurement set-up with computer-
controlled electromechanical stub tuners.

5.9.4. Design of Class A Power Amplifiers

Large-signal parameters are used for the design of class A amplifiers. Since class A
amplifiers are ideally linear, it is sometimes possible to use small signal scattering parameters
for design, but better results are usually obtained if large signal parameters are available. As
with small-signal amplifier design, the first step is to check the stability of the device. Since
instabilities begin at low signal levels, small-signal scattering parameters can be used for this
purpose. Stability is especially important for power amplifiers, as high-power oscillations can
easily damage active devices and related circuitry.

The transistor should be chosen based on frequency range and power output, ideally
with about 20% more power capacity than is required by the design. Silicon bipolar transistors
have higher power outputs than GaAs FETs at frequencies up to a few GHz and are generally
cheaper; GaN HBTs are becoming very popular for high-power applications at RF and low
microwave frequencies. Good thermal contact of the transistor package to a heat sink is
essential for any amplifier with more than a few tenths of a watt power output.

Input matching networks may be designed for maximum power transfer (conjugate
matching), while output matching networks are designed for maximum output power (as
5.62 Microwave Design Principles

derived from ΓLP). The optimum values of source and load reflection coefficients are different
from those obtained from small-signal scattering parameters.

Low-loss matching elements are important for good efficiency, particularly in the
output stage, where currents are highest. Internally matched chip transistors are sometimes
available and have the advantage of reducing the effect of parasitic package reactance’s, thus
improving efficiency and bandwidth.

Example 5.8

Design a power amplifier at 2.3 GHz using a Nitronex NPT25100 GaN HEMT
transistor, with an output power of 10W. The scattering parameters of the transistor for
VDS = 28 V and ID = 600 mA are as follows: S11 = 0.593∟ 178◦, S12 = 0.009∟ −127◦, S21
= 1.77∟ −106◦, and S22 = 0.958∟ 175◦, and the optimum large signal source and load
impedances are ZSP = 10 − j3 Ω and ZL P = 2.5 − j2.3 Ω . For an output power of 10 W,
the power gain is 16.4 dB and the drain efficiency is 26%. Design input and output
impedance matching sections for the transistor, and find the required input power, the
required DC drain current, and the power-added efficiency.

Solution:

First, establish the stability of the device. Using the small-signal scattering parameters.

|Δ| = |S11S22 – S12S21| = 0.579 < 1

K = 1-|S11|2 - |S22|2 + |Δ|2 / 2|S12S21| = 2.08 > 1

Hence the device is unconditionally stable. Converting the large-signal source and
load impedances to reflection coefficients gives

ΓSP = 0.668 ∟187˚

ΓLP = 0.905 ∟-175˚

The source and load reflection coefficients for conjugate matching gives,

B1 = 1 + |S11|2 - |S22|2 - |Δ|2

 Antennas and Microwave Engineering 5.63

B2 = 1 + |S22|2 - |S11|2 - |Δ|2

C1 = S11 – ΔS22*

C2 = S22 – ΔS11*

Using the above relations, it is found that

Γ S = 0.508 ∟166˚

Γ L = 0.954 ∟-176˚

Note that these values are approximately equal to the large- SP LP,
but not exactly, since the scattering parameters used to calculate ΓS and ΓL do not apply for
large power levels. We should use the large-signal reflection coefficients to design the input
and output matching networks.

The AC amplifier circuit is shown in Fig.5.38

Fig 5.38 RF circuit for the amplifier

For an output power of 10 W, the required input drive power is

Pin = Pout(dBm) − G(dB) = 10 log(10,000) − 16.4 = 23.6 dBm = 229mW

The DC input power can be found from the drain efficiency as

PDC = Pout/η = 38.5 W,

so, the DC drain current is

ID = PDC/VDS = 1.37 A

The power added efficiency of the amplifier can be found to be

ηPAE = Pout – Pin / PDC

= 10.0 − 0.229 / 38.5

ηPAE = 25%.
5.64 Microwave Design Principles

5.10. MICROWAVE OSCILLATORS:

Microwave oscillators are the oscillators that operate at microwave frequencies using
negative resistance diodes or transistors. Microwave oscillators have their application in all
modern wireless communications, radar, and remote sensing systems as the signal source to
provide frequency conversion and carrier generation.

Microwave oscillators use an active nonlinear device such as a diode or a transistor

and a passive circuit to convert DC to a sinusoidal steady-state RF signal. At high frequencies,
the nonlinear devices diodes or transistors are biased to a negative resistance operating point
and used with the cavity, transmission line, or dielectric resonators that produces the
fundamental frequency of oscillations up to 100 GHz.

The following are the important considerations for the oscillators used in RF and
Microwave systems:

 Tuning range

 Frequency stability

 AM and FM noise

 Harmonics

The fig.5.39 shows the RF circuit of a one-port negative resistance oscillator.

Fig 5.39 Circuit of a One port negative resistance oscillator

In the above circuit, the input impedance of the active device is given by

Zin=Rin+jXin. (5.123)

Since the impedance is depended on voltage or current and frequency the above
equation is expressed as

Zin (I, jɷ) = Rin (I, jɷ) + j Xin (I, jɷ) (5.124)
 Antennas and Microwave Engineering 5.65

the passive load impedance ZL=RL + j XL (5.125)

Applying Kirchhoff’s voltage law gives

(ZL + Zin) I = 0 (5.126)

To occur oscillations, the RF current must be non-zero. Hence the following

conditions are to be satisfied:

RL + Rin = 0 (5.127)

and XL + Xin = 0 (5.128)

Since the load is passive RL > 0, equation (5.127) implies that Rin < 0. This means that
the positive resistance denotes energy dissipation and a negative resistance denotes energy
source.

Equation (5.128) determines and controls the frequency of oscillation.

From equation (5.126), considering ZL = -Zin and during steady-state oscillation, the
reflection coefficients ΓL and Γin are,
in in
(5.129)
in in in

The process of oscillation is dependant on the behavior of Zin as follows.

1. Initially, the circuit is unstable at a certain frequency i.e.) Rin (I, jɷ) + RL <0.

2. Any transient excitation or noise induces an oscillation at a frequency ɷ.

3. As I increase, Rin (I, jɷ) must become less negative until the current I0 is reached such
that Rin (I0, j ɷ0) + RL =0 and Xin (I0, jɷ0) + XL (jɷ0) = 0.

4. Now the oscillator runs in a stable state.

5. The final frequency ɷ0 differs from the frequency Xin because Xin is a current
dependant and hence Xin (I, jɷ) ≠ Xin (I0, jɷ).

The equations (5.127) and (5.128) does not guarantee a stable state of oscillations. The
stability of the circuit requires a concern in current or frequency to allow the oscillator to
return to its original state. The above condition is enumerated by considering the effect of the
ɷ.

Considering ZT (I, s) = Zin (I, s) + ZL(s) and using Taylor series about the stable
operating point (I0, ɷ0), ZT(I, s) becomes
5.66 Microwave Design Principles

(5.130)

ZT(I, s) must be equated to zero to allow oscillations to occur.

Use the fact that

Therefore,

| |

| (5.131)

If the transient ɷ

{ } Or substitute ZT=RT+jXT

(5.132)

The equation eqn.(5.132) is sometimes known as Kurokawa’s condition. For a passive

load,

By substituting ZT=Zin + ZL, the stability equation reduces to

(5.133)

Where Zin = Rin + j Xin

ZL =RL + jXL

Equation (5.133) implies that a high Q circuit will produce maximum oscillator
stability. Cavity and dielectric resonators are used for this often.

Effective oscillator design requires consideration of other issues such as:

 Selection of an operating point for stable operation

 Maximum power output

 Antennas and Microwave Engineering 5.67

 Frequency pulling

 Large signal effects

 Noise characteristics

5.10.1. Transistor oscillators:

In a Transistor Oscillator, a potentially unstable transistor is designed with an

impedance to drive the device in an unstable region is terminated to create a negative
resistance one-port network.

Fig 5.40 Circuit model of a Transistor Oscillator

The Fig.5.40 shows the circuit model of a Transistor oscillator. In the circuit, the RF
output port acts as a part of the load network on the output side of the transistor. The
terminating network used to the left of the transistor can also act as output port.

For designing an oscillator, a device with a high degree of instability is required.

Hence common source or common gate FET’s or common emitter or common base BJT’s are
used with positive feedback to enhance the instability of the device. Once the transistor
configuration is selected, the output stability circle is drawn in the ΓL plane. Then the
terminating impedance ZS = RS + j XS is chosen to match Zin.

RS is chosen to be RS+ Rin <0 because Rin becomes less negative as the oscillator
power builds up. If not, oscillation may cease as increasing RF power increases Rin to the
point RS + Rin >0.

Practically the value of RS will be

RS = (5.134)

To resonate the circuit the reactive part of ZS is chosen as

XS = -Xin (5.135)
5.68 Microwave Design Principles

When oscillation occurs between the termination network and the transistor,
simultaneous oscillations occur at the output port.
At the input port for steady-state oscillations ΓS Γin =1. From the definition of
scattering parameters,

1 = S11 1 + S12 2 (5.136)

2 = S21 1 + S22 2 (5.137)

Since 2 = ΓL 2 the above equations become,

1 = S11 1 + S12 ΓL 2 (5.138)

2 = S21 1 + S22ΓL 2 (5.139)

Eliminating 2 and solving for Γin we get

2(1- S22ΓL) = S21 1 (5.140)

2 = (5.141)

1 = S11 1 + (S12 ΓL S21 1 / (1- S22ΓL)) (5.142)

1 = (S11 + (S12 ΓL S21 / (1- S22ΓL))) 1 (5.143)

Γin = 1/ 1 (5.144)

= S11 + (S12 ΓL S21 / (1- S22ΓL)) (5.145)

= (S11 – Δ ΓL) / (1- S22 ΓL) (5.146)
Where Δ = S11S22 – S12S21
Solving for ΓL gives,
ΓL = (1-S11 ΓS) / (S22 - Δ ΓS) (5.147)
The reflection coefficient ZS is,

out (5.148)

Taking LCM,
Γout = (S22 - Δ ΓS)/ (1-S11 ΓS) (5.149)

Multiplying ΓL Γout, we get 1, and hence the condition for oscillation ZL = -Zout is
satisfied at the load network.
 Antennas and Microwave Engineering 5.69

EXAMPLE 5.9

A one-port oscillator uses a negative resistance diode having Zin = 1.25∟ 40◦ (Z0
= 50 Ω) at its desired operating point, for f = 6 GHz. Design a load matching network for
a 50 Ω load impedance.

Solution:

By direct calculation, the input impedance of the diode is,

in
in
in

Then, the load impedance is,

ZL = - Zin = 44 – j 123 Ω

A shunt stub and series section of line can be used to convert 50 Ω to ZL, as shown in
the Fig.5.41

Fig 5.41 Load matching circuit for the one - port oscillator
5.70 Microwave Design Principles

EXAMPLE 5.10

Design a transistor oscillator at 4 GHz using a GaAs MESFET in a common gate

configuration, with a 5 nH inductor in series with the gate to increase the instability.
Choose a load network to match to a 50 Ω load, and an appropriate terminating network
at the input to the transistor. The scattering parameters of the transistor in a common
gate configuration are (Z0 = 50 Ω) S’11 = 2.18∟ −35◦,S’12= 1.26∟ 18◦,S’21= 2.75∟
96◦,S’22= 0.52∟ 155◦.

Solution:

L plane) parameters gives

( )
| |

CL = 1.08 ∟33˚

RL = 0.665

Since S’11= 2.18 > 1, the stable region is inside this circle, as shown in the Smith
chart in Fig.5.42(b) .There is a great L , but one
in L located on the opposite
L = 0.59∟ −104◦. Then we can design a
single-stub matching network to convert a 50 Ω load to ZL =20 − j35 Ω, as shown in Fig.
(5.42(a)).

Fig 5.42 (a) Oscillator circuit

 Antennas and Microwave Engineering 5.71

L in as

or Zin = −84 − j1.9 Ω. Then, we find ZS as

Using Rin/3 should ensure enough instability for the start-up of oscillation. The easiest
way to implement the impedance ZS is to use a 90 Ω load with a short length of the line, as
shown in Fig.5.42(b). It is likely that the steady-state oscillation frequency will differ from 4
GHz because of the nonlinearity of the transistor parameters.

Fig 5.42(b) Smith chart for determi L

5.72 Microwave Design Principles

5.10.2. Dielectric Resonator Oscillators:

In Microwave oscillators, the oscillator stability is enhanced using high Q- tuning

networks. The resonant network uses lumped elements or microstrip lines and stubs provide Q
to a few hundred.

The waveguide cavity resonators can provide Q of 104 or more but have a
disadvantage that it is not suited for integration in miniature microwave integrated circuitry.

Usage of metal cavities has a disadvantage of significant frequency drift caused by

dimensional expansion due to the variations in temperature.

The dielectric cavity resonator overcomes the above disadvantages as it produces

 Q in several thousand,

 compact and can be easily integrated with planar circuitry, and

 made from ceramic materials that have excellent temperature stability.

Hence transistor dielectric resonator oscillators (DROs) are commonly used over the
microwave and lower millimeter-wave frequency range.

Fig 5.43 (a) Geometry of a dielectric resonator

(b) Equivalent circuit coupled to a microstrip line

The fig.5.43(a) shows a dielectric resonator coupled to an oscillator circuit coupled to

a microstrip line. The resonator operates in the TE01 mode and couples to the firing magnetic
field of the microstrip line. The distance d between the resonator and microstrip line
determines the strength of coupling. As the coupling is made via the magnetic field, the
resonator acts as a series load on the microstrip line. As the equivalent circuit shows the
resonator is modelled as a parallel RLC circuit. The coupling to the feedline is modelled by
the turn’s ratio of the transformer N.
 Antennas and Microwave Engineering 5.73

The impedance of a parallel RLC resonator circuit is given by,

(5.150)

Where Q0 = R/ɷ0L is the unloaded resonator Q

ɷ0 = 1/√LC is the resonant frequency
Δɷ = ɷ - ɷ0

The coupling factor between the resonator and the feedline g is the ratio of the
unloaded Q (Q0) to external Q(Qe).

(5.151)

Where RL = 2Z0 is the load resistance for a feedline with source and termination
resistance Z0.
In case if the feedline is terminated with an open circuit λ/4 from the resonator, the
magnetic field is maximized and RL becomes Z0 and the coupling factor becomes twice the
value of equation (5.151).

The reflection coefficient at the terminated microstrip line looking towards the
resonator is given by,

(5.152)

Where g = N2 R/2Z0
The coupling coefficient g= Γ/ (1- Γ) is found by measuring Γ at resonance and the
resonant frequency and Q can also be found by measurement.

To provide frequency stability, a dielectric resonator is incorporated into the circuit

either as parallel feedback or a series feedback arrangement as shown in Fig.5.44.

Fig 5.44 (a) Dielectric resonator using (b) Dielectric resonator using
Parallel feedback series feedback
5.74 Microwave Design Principles

The parallel feedback arrangement uses a resonator coupled with two microstrip lines
which acts as a high Q pass bandpass filter and couples a portion of the transistor output back
to its input. The spacing between the resonator and the microstrip lines controls the amount of
coupling and the length of the lines controls the phase. Designing an oscillator using parallel
feedback is done using CAD software.

The series feedback arrangement uses only one microstrip feedline and does not have a
tuning range as a parallel feedback circuit. Designing an oscillator using series feedback uses
the same procedure as a design of two-port oscillators.

EXAMPLE 5.11

A wireless local area network application requires a local oscillator operating at

2.4 GHz. Design a dielectric resonator oscillator using the series feedback circuit with a
bipolar transistor having the following scattering parameters (Z0 = 50 Ω): S 11 = 1.8∟
130◦, S12 = 0.4∟45◦, S21 = 3.8∟ 36◦, and S22 = 0.7∟ −63◦. Determine the required
coupling coefficient for the dielectric resonator, and the required microstrip matching
0 for small variations in
frequency about the design value, assuming an unloaded resonator Q of 1000.

Solution:

The DRO circuit is shown in Fig.5.45(a) .

Fig 5.45 (a) Circuit for the Dielectric resonator

The dielectric resonator is placed λ/4 from the open end of the microstrip line; the line
length lr can be adjusted to match the phase of the required value of ΓS. In contrast to the
oscillator of the previous example, the output load impedance for this circuit is part of the
terminating network.
 Antennas and Microwave Engineering 5.75

The stability circles for the transistor can be plotted if desired but are not necessary for
the design since we can begin by choosing ΓS to provide a large value of |Γout|.

out

Which indicates that we can maximize Γout by making 1 − S11ΓS close to zero. Thus,
we choose ΓS = 0.6∟−130◦, which gives Γout = 10.7∟132◦. This corresponds

to an impedance

out
out
out

Applying the analogous start-up condition for the output side gives the required
termination impedance as

ZL = (−Rout /3)− j Xout = 5.5 − j6.1 Ω.

The matching network can now be designed using a Smith chart. The shortest
transmission line length for matching ZL to the load impedance Z0 is lt = 0.481λ, and the
required open-circuit stub length is ls = 0.307λ.

Next, we match ΓS to the resonator network. We know that the equivalent impedance
of the resonator seen by the microstrip line is real at the resonant frequency, so the phase
’
S must be either zero or 180◦. For an under
coupled parallel RLC resonator, R < Z0, so the proper phase will be 180◦, which can be
achieved by transformation through the line length lr. The magnitude of the reflection
coefficient is unchanged, so we have the relation

which gives lr = 0.431λ. The equivalent impedance of the resonator at resonance is

then

The coupling coefficient can be found, with a factor of two to account for the λ/4 stub
termination, as
5.76 Microwave Design Principles

5.11. MICROWAVE MIXERS:

A mixer is a three-port device that performs the function of frequency conversion.

Mixers translate the input frequency into a different frequency. The mixer uses the nonlinear
or time-varying element to achieve the frequency conversion. The output of the mixer consists
of the sum and difference frequencies of the two input signals. As the nonlinear components
such as diode or transistor generate a wide variety of harmonics of input frequencies, filters
must be used to select the desired frequencies. The operation of the mixer depends on the
nonlinearity provided by the nonlinear component used.

The important characteristics of mixers are:

 Image frequency

 Conversion loss

 Noise effects

 Intermodulation effects

5.11.1. Mixer Characteristics:

The output of the mixer is proportional to the product of the two input signals. Mixers
are used to perform both frequency upconversion and downconversion. The Fig.5.46
illustrates the frequency conversion.

Fig 5.46 Illustration of frequency conversion

Two of a mixer’s three ports serve as inputs, while the other port serves as an output
port. An ideal mixer produces an output that consists of the sum and difference frequencies of
its two input signals. In other words:

fout = fin1 ± fin2 (5.153)

 Antennas and Microwave Engineering 5.77

The three ports of a mixer are known as the intermediate-frequency (IF),

radiofrequency (RF), and local-oscillator (LO) ports. The LO port is usually an input port. The
RF and IF ports can be used interchangeably, depending on whether the mixer is being used to
perform upconversion or downconversion. The LO signal is typically the strongest signal
injected into a mixer. The required LO drive level is dependent on several factors, including
the mixer’s configuration and device technology.

Frequency Upconversion:

When a mixer is used to perform upconversion, an input signal enters the IF port and
an LO signal enters the LO port. These two input signals produce an output signal at the RF
port. The frequency of this output signal is equal to the sum of the IF input signal’s frequency
and the LO signal’s frequency. Upconversion is normally part of a transmitter, while
downconversion is typically used in a receiver.

The local oscillator signal is represented as,

VLO(t) = cos 2ΠfLOt (5.154)

A lower frequency baseband or intermediate frequency (IF) signal is expressed as,

VIF(t) = cos 2ΠfIFt (5.155)

The output of the mixer is the product of the LO and IF signals which is expressed as,

VRF(t) = K VLO(t) VIF(t) (5.156)

substituting eqn (5.154) and eqn (5.155) in eqn (5.156), we get

VRF(t) = K cos 2ΠfLOt cos 2ΠfIFt

= K/2[ cos (2Π (fLO - fIF) t + cos (2Π (fLO + fIF)] (5.157)

Where K is a voltage conversion loss of a mixer which is a constant.

Frequency Downconversion:

When a mixer is used to perform downconversion, an input signal enters the RF port
and an LO signal enters the LO port. These two input signals produce an output signal at the
IF port. The frequency of this output signal is equal to the difference of the RF input signal’s
frequency and the LO signal’s frequency.
5.78 Microwave Design Principles

The RF input signal is expressed as,

VRF(t) = cos 2ΠfRFt and

the LO oscillator signal is expressed as

VLO(t) = cos 2ΠfLOt

Now the output of the mixer is,

VIF(t) = K VRF(t) VLO (t)

= K cos 2ΠfRFt cos 2ΠfLOt

= K/2[ cos (2Π (fRF - fLO) t + cos (2Π (fRF + fLO)] (5.158)

Since RF and LO frequencies are very close to each other the sum frequency is almost
equal to twice the RF frequency and the difference frequency is much smaller than RF
frequency. The desired output of the mixer fIF is the difference frequency fRF - fLO which is
obtained using a low pass filter.

fIF = fRF - fLO (5.159)

Image Frequency:

In a receiver, the antenna delivers the RF input frequency at frequency fRF from a
wide band of frequencies. For a receiver with fLO local oscillator frequency and fIF IF
frequency gives the RF input frequency down-converted to the IF frequency as,

fRF = fLO + fIF (from eqn. (5.159)) (5.160)

Consider the RF input frequency as,

fIM = fLO – fIF (5.161)

The frequency in eqn. (5.161) is identical to the frequency in eqn. (5.160) as the
Fourier spectrum of any real signal is symmetric about zero frequency and hence contains
positive as well as negative frequencies. The RF frequency is expressed in the eqn. (5.161) is
called the image response.

The image frequency plays an important role in receiver design as it is difficult to

distinguish at the IF stage from the desired RF signal of frequency. Hence steps are to be
taken in the RF stages of the receiver to preselect signals within the desired RF frequency
band. The identification of the desired RF frequency and the image frequency is arbitrary and
depends on whether the LO frequency is above or below the desired RF frequency. It is clear
 Antennas and Microwave Engineering 5.79

from eqn. (5.160) and eqn. (5.161) that the desired and the image frequencies are separated by
2fIF.

From eqn. (5.159) it is found that fIF may be negative and there are two LO
frequencies used for the RF and IF frequency,

fLO = fRF± fIF (5.162)

In practice, most receivers use a local oscillator fLO = fRF + fIF because a smaller LO
tuning ratio is enough to select RF signals over a band of frequency.

Conversion Loss:

In passive mixers, conversion loss is defined as the difference in signal level between
the amplitude of the input signal and the amplitude of the desired output signal. In a mixer
used for downconversion, the conversion loss is the difference between the RF input signal’s
amplitude and the IF output signal’s amplitude.

Lc = 10 log (available RF input power/ available IF output power) ≥ 0dB.

In a mixer used for upconversion, the conversion loss is the difference between the IF input
signal’s amplitude and the RF output signal’s amplitude. Conversion loss is expressed as a
positive number in decibels. Conversion loss occurs due to the resistive losses in a mixer and
frequency conversion process from RF to IF ports. Minimum conversion loss is more critical
at the receiver side because RF stages at the receivers operate at such low power levels
compared to the transmitters. Minimizing conversion loss in the RF stages maximizes the
receiver noise Figure.

Typical values of conversion loss can range from approximately 4.5 to 9 dB.
Conversion loss values of 6 to 8 dB are common in standard double-balanced mixers, while
triple-balanced mixers generally have a higher conversion loss than double-balanced mixers.
It is also possible to achieve conversion gain in active mixers. One of the factors that affect
conversion loss is the LO power level and minimum conversion loss occurs for LO powers
between 0 and 10dBm.

Noise Figure:

Noises are generated in mixers by the nonlinear devices such as diodes or transistors
and by thermal sources due to resistive losses. The practical range of noise Figure is from 1 to
5 dB. In general, diode mixers achieve lower noise Figure compared to transistor mixers.
Noise Figure of a mixer depends on its input signal either single side band or a double side
band signal. The mixer down converts the noise at both side bands hence the power of SSB
signal is one half of that of the DSB signal.
5.80 Microwave Design Principles

Consider a DSB signal of the form,

VDSB(t) = A [ cos (ɷLO - ɷIF)t + cos (ɷLO - ɷIF)t] (5.163)

After mixing with a LO oscillator cos ɷLOt and passing through a low pass filter the IF
signal is,

VIF(t) = cos (ɷIFt) + cos(-ɷIFt) = AK cos ɷIFt (5.164)

Where K is a constant for conversion loss for each sideband. The average power for
the DSB input signal is expressed in eqn. (5.163) is given by,

Si =A2 (5.165)

The average power of the output IF signal is,

So (5.166)

The input noise power for the noise Figure is defined as N1=kT0B, where To = 290k
and B is the IF bandwidth. The output noise power is equal to the input noise plus the noise
added by the mixer Nadded divided by the conversion loss.

N0 = (KT0B +Nadded)/ Lc (5.167)

Now the DSB noise Figure of the mixer is given by,

FDSB= = (1+ ) (5.168)

Let’s consider for SSB case,

The SSB input signal is expressed as,

VSSB(t) = A cos (ɷLO - ɷIF)t (5.169)

Mixing with the LO signal and passing through low pass filter, the down converted IF
signal is,

VIF(t) = cos (ɷIFt) (5.170)

The average power of the SSB input signal is,

Si = (5.171)
 Antennas and Microwave Engineering 5.81

and the average power of the output IF signal is,

So (5.172)

The input and output noise powers of SSB are same as the DSB case and hence the
noise Figure for an SSB input signal is

FSSB= = (1+ ) (5.173)

Comparing eqn. (5.168) and eqn. (5.173) the noise Figure of the SSB case is twice that
of the DSB case.

FSSB = 2FDSB (5.174)

Isolation:

Isolation is a measurement of the amount of power that leaks from one port to another.
Isolation is defined as the difference in signal level between the amplitude of an input signal
and the amplitude of the leaked power from that input signal to another port. When isolation is
high, the amount of power leaking from one port to a different port is small.

Three types of isolation in microwave mixers are:

 LO-RF isolation,

 LO-IF isolation, and

 RF-IF isolation.

As an example, if a 5-GHz signal with an amplitude of +15 dBm is injected into the
LO port, a portion of this signal will leak into the RF port. If this LO input signal causes a 5-
GHz signal with an amplitude of -20 dBm to be measured at the RF port, the LO-RF isolation
is 35dB. LO-IF isolation and RF-IF isolation are calculated in the same manner.

LO-RF isolation is critical in frequency down-converting systems because LO power

can leak into the RF circuitry. If the LO-RF isolation is poor, the LO power can contaminate
the RF line.

Poor LO-RF isolation can also cause problems in frequency upconverting systems
when the LO frequency is very close to the RF output frequency. Because the LO frequency
and the RF output frequency are so close, no amount of filtering can remove the LO leakage.
As a result, that leakage interferes with the RF output, potentially degrading the RF output
circuitry.
5.82 Microwave Design Principles

1-dB Compression Point

A mixer’s conversion loss remains constant when the mixer is in linear operation. As
the amplitude of the input signal increases, the amplitude of the output signal rises by the
same amount. However, once the input signal’s amplitude reaches a certain level, the
amplitude of the output signal ceases to exactly follow the input signal.

The mixer deviates from linear behavior and its conversion loss begins to increase.
When a mixer’s conversion loss increases by 1 dB, the 1-dB compression point has been
reached. The 1-dB compression point of a mixer is defined as the amplitude of the input signal
required to increase the conversion loss by 1-dB. A mixer’s 1-dB compression point
determines the upper limit of its dynamic range.

Fig 5.47 1-dB Compression Point

A mixer’s 1-dB compression point is usually related to the LO drive level. Mixers
with higher LO drive level requirements have a higher 1-dB compression point. Yet higher
LO power also must have to be delivered to such mixers. In general, the 1-dB compression
point is anywhere from 4 to 7 dB below the mixer’s minimum recommended LO drive level.

Intermodulation Distortion:

Two-tone third-order intermodulation distortion (IMD) occurs when two signals

simultaneously enter the mixers IF or RF input port. In practice, this could happen in a multi-
carrier signal environment. These two signals interact with each other and with the LO signal,
which creates distortion. In a receiver, two-tone third-order IMD is a serious problem because
it can generate third-order distortion products that fall within the IF bandwidth.

If fRF1 and fRF2 represent two separate RF input signals and fLO represents the LO
signal, the third-order distortion products generated at the mixer’s IF port is:
 Antennas and Microwave Engineering 5.83

Interferer1 = 2fRF1 - fRF2 - fLO

Interferer2 = 2fRF2 - fRF1 - fLO

These third-order distortion products are extremely close to the desired IF output
frequency. No amount of filtering can remove these unwanted distortion products. Thus, the
signal-to-noise ratio of the received signal is degraded, highlighting the need to suppress these
distortion products.

The third-order input intercept point (TOI or IP3) is a widely accepted Figure of merit
used to describe a mixer’s capability to suppress third-order distortion products. TOI is used
in predicting the nonlinear behaviour of a mixer as the amplitude of its input signal increases,
which causes the third-order products to increase by a 3:1 ratio. For any 1-dB increase in the
input signal’s amplitude, the third-order products increase by 3 db.

Fig 5.48 Graphical representation of how TOI is derived.

5.12. Mixer Types:

5.12.1. Single-ended diode mixer:

Fig.5.49 uses only one diode element hence it is called as a single ended mixer. The
RF and LO are the two inputs given to the diplexer which superimposes these two input
voltages to drive the diode.
5.84 Microwave Design Principles

Fig 5.49 Schematic circuit diagram of One diode Mixer

The diode is biased at DC voltage (VDO), decoupled from RF and LO signal paths
through a DC blocking capacitor (CB). RF choke blocks the RF/LO signals entering the bias
source. High-frequency components produced due to diode nonlinearity are filtered by IF
filter, allowing only IF component to appear at the output. The above describes the application
of a down converter but the same circuit can be used as an upconverter as each port can be
used interchangeably as an input port or output port.

Fig 5.50 AC equivalent Circuit of the mixer

The equivalent circuit of the mixer shown in Fig.5.50 has a RF and LO input voltages
represented as two series connected voltage sources. Let the RF input voltage be

VRF(t) =VRF cos ɷRF t (5.175)

and the LO input signal is,

VLO(t) = VLO cos ɷLOt (5.176)

 Antennas and Microwave Engineering 5.85

The total diode current using small-signal approximation is given by,

(5.177)

The first term in eqn.5.177 is the DC bias current which is blocked by the DC blocking
capacitors hence not available at the IF output. The second term is the replication of the RF
and LO input signals and is filtered out by the low pass filter. Now the eqn.5.177 becomes,

(5.178)

(5.179)

(5.180)

The several components present in the above equation are blocked by the blocking
capacitor and filtered out by the low pass filter. The resultant IF output will be

(5.181)

Where ɷIF = ɷRF - ɷLO is the IF frequency. The spectrum of the down converter single
diode mixer is like the spectrum of the ideal mixer.

5.12.2. Single-ended FET mixer:

FET has various nonlinearity parameters that can be used to mix, but the important
parameter used is transconductance gm. Often FET is operated in a common source
configuration with a negative bias.

The Fig.5.51 shows the variation of transconductance with the gate bias.
5.86 Microwave Design Principles

Fig 5.51 Variation of transconductance with gate bias

When FET is used as an amplifier and the gate bias voltage is chosen to be zero or
slightly positive, the transistor operates as a linear device and the transconductance will be
maximum. When the gate bias voltage reaches the pinch-off region, a small positive variation
in the gate voltage causes a large change in the transconductance that leads to a nonlinear
response.

Hence the LO voltage is applied as the input to the gate of the FET to pump the
transconductance between low and high states thus provides the operation of a mixer. The
circuit for a single-ended FET mixer is shown in Fig.5.52. RF/LO diplexer is used to combine
the RF and the LO signals which is fed as the input to the gate of the FET. An impedance
matching network is used to provide a very low input impedance. RF chokes are used to bias
the gate of the FET at a negative voltage near pinch-off and to provide positive bias to the
drain of the FET. A bypass capacitor used at the drain provides the return path for the LO
signal. Low pass filter provides the final IF output signal after filtering.

Fig 5.52 Circuit diagram of a single-ended FET Mixer

The equivalent circuit based on the unilateral equivalent circuit of a FET is shown in
Fig.5.53. The RF and LO input voltages are given by,

VRF(t) =VRF cos ɷRF t

 Antennas and Microwave Engineering 5.87

VLO(t) = VLO cos ɷLOt

Let Zg =Rg +jXg be the Thevenin source impedance for the RF input port and ZL =
RL+jXL be the Thevenin source impedance of the IF output port. These impedances are
complex to allow the complex conjugates match at the input and the output ports to get
maximum power transfer. Since LO signal is not concerned with maximum power transfer LO
has only real generator impedance of Z0.

Since the LO drives the FET transconductance, its time variation expressed as a
Fourier series in terms of harmonics of LO is given by,

(5.182)

Fig 5.53 Equivalent Circuit of FET

The conversion gain of the FET mixer is found by,

IF-avail
(5.183)
avail

Where VIFD = IF drain voltage, the impedances Zg and ZL are chosen to obtain
maximum power transfer at the RF and IF ports. The RF frequency component of the phasor
voltage across the gate to source capacitance in terms of the voltage divider between Zg, Ri,
and Cgs is,

(5.184)

Multiplying the transconductance eqn.5.182 by VRFc = VRFccosɷRFt and expanding we

get,

(5.185)

The down-converted IF frequency component is extracted from the second term using
trigonometric identity,
5.88 Microwave Design Principles

(5.186)

Where ɷIF =ɷRF- ɷLO. Then the IF component of the drain voltage in phasor form is,

(5.187)

Now conversion gain before conjugate matching becomes,

not (5.188)

To maximize the conversion gain RF and IF ports are matched conjugately.

Considering Rg=Ri, Xg=1/ɷRFCgs, RL= Rd and XL=0 the above equation reduces to,

(5.189)

The quantities gi, Rd, Ri, and Cgs are the parameters of the FET. Matching circuits are
used to transform the FET impedance to 50Ω for the RF, LO and IF ports.

5.12.3. Balanced Mixer:

A balanced mixer consists of two single-ended mixers combined with a hybrid

junction (900 or 1800). Balanced mixer with a 900 hybrid improves RF input matching and
with 1800 hybrid improves the RF-LO isolation over a wide range of frequencies. Both these
types of mixers reject all the even-order intermodulation products.

Fig 5.54 Balanced mixer (a) 900 Hybrid (b) 1800 Hybrid

Let us consider the RF and Lo voltages as

 Antennas and Microwave Engineering 5.89

VRF(t) =VRF cos ɷRF t (5.190)

VLO(t) = VLO cos ɷLOt (5.191)

The scattering matrix of a 900 hybrid is given by,

(5.192)

The total RF and Lo voltages applied to the two diodes are,

√
(5.193)
√

The diode current i(t) using the small-signal diode approximation is given by,

(5.194)

(5.195)

Where negative sign on i2 indicates the reversed diode polarity and K is a constant.

Adding the above two eqns., we get

(5.196)

Where is the IF frequency. Upon combining all the DC components

of the diode current gets cancels. After low pass filtering the IF output will be,

(5.197)

Assume that the diodes are matched and exhibit a voltage reflection coefficient at
the RF frequency then the phasor expression for the reflected RF voltages at the diodes will
be,

(5.198)
√
5.90 Microwave Design Principles

(5.199)
√

The outputs at the RF and LO ports are given by,

(5.200)
√ √

(5.201)
√ √

Thus, the phase characteristics of the 900 hybrid lead to the cancellation of reflections
at the RF port. The isolation between the RF and LO ports depends on the diode matching and
hence difficult to maintain over a reasonable frequency range.

5.12.4. Image Reject Mixer:

The concept of the mixer is an RF frequency mixed with a LO frequency generates an

IF frequency. The frequency ɷLO - ɷIF is called the Lower Sideband (LSB) and the frequency
ɷLO + ɷIF is called the upper sideband (MSB). The image rejects the mixer shown in Fig.5.55
is used to isolate these two frequencies into separate output signals. The circuit is used as an
up conversion which is called as a single sideband modulator. For this the IF input signal is
delivered to either the LSB or the USB port of the IF hybrid, and the output is produced at the
RF port of the mixer.

Fig 5.55 Circuit diagram of Image reject mixer

Let the RF signal is expressed as

(5.202)
 Antennas and Microwave Engineering 5.91

Where VU and VL, are the amplitudes of the upper sideband and lower sideband
respectively.

Using the Scattering matrix,

(5.203)

The RF voltages at the diode are represented as,

(5.204)
√

(5.205)
√

Mixing with the LO voltage, and passing through the low pass filter
the IF inputs to the IF hybrid is

(5.206)
√

(5.207)
√

Where K is the mixer constant for the squared term of the diode response.

The phasor representation of the IF signals is,

(5.208)
√

(5.209)
√

Combining these voltages in the IF hybrid gives,

(5.210)
√ √
5.92 Microwave Design Principles

(5.211)
√ √

In the time domain, the outputs are expressed as,

(5.212)

V1(t) and V2(t) has a 900 phase shift between them. The image rejection mixer does not
incur additional losses other than the conversion losses of the single rejection mixer. The
practical difficulties of the image rejection mixer are:

 Fabricating a good hybrid at low IF frequency

 Losses

 Noise Figure which is always greater than a simple mixer.

5.12.5. Differential FET Mixer:

The differential FET mixer has two FETs connected in a differential balanced
configuration like the differential amplifier. The LO input voltage and IF output voltage is
balanced signals. These signals are converted into single-ended signals using baluns. Fig.5.56
shows the circuit diagram of a differential FET mixer. The RF input is single ended and
applied to the bottom transistor.

Let the RF and LO voltages are expressed as

(5.213)

Fig 5.56 (a)Balanced FET Mixer (b) Equivalent Circuit

 Antennas and Microwave Engineering 5.93

(c) LO voltage waveform and idealized switching waveform of the top left FET

The circuit operates as a switch operated by the LO which turns the top two FETs on
and off with its alternate half cycles. The transistors are biased slightly above the pinch-off
voltage so each of the transistors will be conducting slightly above half of each LO cycle.

During the positive half cycle of the FET at the top left conducts with low
resistance and turns off during the negative half cycle. During the positive half cycle of
the FET at the top right turns on. One of the upper FETs is always conducting. The lower FET
is biased into saturation and operates as an RF amplifier. The RF current at the drain of the
bottom FET is The RF and LO ports require impedance matching. The IF
output circuit provides a return path to ground for the LO signal. A current source or inductive
degeneration is used as the source of the lower FET.

In the equivalent circuit, the upper FETs are replaced by the ideal switches. As per the
idealized conductance waveform shown in Fig.5.56(c), the effect of the switches is modelled
using a Fourier series. The first few terms of the Fourier series are given by,

(5.214)

The voltages at the IF terminals are expressed as,

(5.215)

(5.216)

The net output IF the voltage is,

(5.217)
5.94 Microwave Design Principles

The voltages at the RF and LO frequencies are cancelled without filtering and leaves
the terms at the frequencies at The IF output after passing through a low pass
filter is,

. (5.218)

5.12.6. Gilbert Cell Mixer:

The Gilbert cell mixer uses two single balanced FET mixers to form a double-balanced
mixer. It has fully balanced differential ports for LO, RF and IF signals. The RF and LO
signals are cancelled at the IF output port because of the circuit symmetry and its excitations.
The operation of the Gilbert cell mixer is the same as the single balanced FET mixer where
the four FETs operate as switches which are controlled by the LO voltage. The lower two
FETs operate as a balanced amplifier for the RF input signal. The circuit shown below uses a
current source at the sources of the amplifier FETs. This type of mixer is used in CMOS
RFICs for wireless applications.

Fig 5.57 Gilbert Cell Mixer

5.12.7. Double-Balanced Mixer:

The double-balanced mixer is shown in Fig.5.58 uses two hybrid junctions or

transformers which provides good isolation between the three ports and rejects all even
harmonics of the RF and LO signals. This provides very good conversion loss but less than
ideal input matching at RF port. This mixer also provides a higher third order intercept point
than single ended or balanced mixer.
 Antennas and Microwave Engineering 5.95

Fig 5.58 Double Balanced Mixer

5.12.8. Antiparallel Diode Mixer:

This type of mixer is used for sub harmonically pumped millimeter-wave frequency
conversion. The back to back diodes functions as a frequency doubler. The nonlinearity of the
diode operates as a resistive frequency multiplier to generate the second harmonic of the LO
to mix with the RF signal to produce the desired output frequency. This type of mixer has
symmetric I-V characteristics that suppress the mixing product of the RF and Lo signals
which leads to better conversion loss.

Fig 5.59 Sub harmonically pumped mixer using an antiparallel diode mixer

Table 5.6 summarizes the characteristics of the various mixers:

5.96 Microwave Design Principles

TWO MARKS QUESTIONS AND ANSWERS

1. What is the impedance transformation?

The ability to change impedance by adding a length of a transmission line is known as

impedance transformation. When operated at a frequency corresponding to a standing
wave of 1/4-wavelength along the transmission line,
the line's characteristic impedance necessary for impedance transformation must be
equal to the square root of the product of the source impedance and the
load's impedance.

2. What is a quarter-wave transformer?

A quarter-wave transformer is a simple impedance transformer that is

commonly used in impedance matching to minimize the energy which is reflected
when a transmission line is connected to a load. It presents at its input the dual of the
impedance with which it is terminated.

3. What is the lossless and lossy line?

A lossless line is
defined as a transmission line that has no line resistance and no dielectric loss. This
would imply that the conductors act like perfect conductors and the dielectric acts as a
A lossy transmission
line includes a term to represent the resistance of the signal flowing. down
the line and a conductance to represent the possibility of a leakage current between
the conductors through the insulator. (Here the attenuation coefficie

4. What is impedance matching?

Impedance matching is the practice of designing the input impedance of an electrical

load or the output impedance of its corresponding signal source to maximize the power
transfer or minimize signal reflection from the load. Here (ZL=ZO)

5. Why impedance matching or tuning is important?

Impedance matching or tuning is important for the following reasons:

1. Maximum power is delivered when the load is matched to the line (assuming the
generator is matched), and power loss in the feed line is minimized.

2. Impedance matching sensitive receiver components (antenna, low-noise amplifier,

etc.) may improve the signal-to-noise ratio of the system.
 Antennas and Microwave Engineering 5.97

3. Impedance matching in a power distribution network (such as an antenna array

feed network) may reduce amplitude and phase errors.

6. What are the factors that may be important in the selection of a particular
matching network?

1. Complexity—As with most engineering solutions, the simplest design that

satisfies the required specifications is generally preferable. The simpler matching
network is usually cheaper, smaller, more reliable, and less lossy than a more
complex design.

2. Bandwidth—Any type of matching network can ideally give a perfect match (zero
reflection) at a single frequency. In many applications, however, it is desirable to
match a load over a band of frequencies. There are several ways of doing this,
with, of course, a corresponding increase in complexity.

3. Implementation—Depending on the type of transmission line or waveguide being

used, one type of matching network may be preferable to another. For example,
tuning stubs are much easier to implement in waveguide than are multisection
quarter-wave transformers.

4. Adjustability—In some applications, the matching network may require

adjustment to match a variable load impedance. Some types of matching
networks are more amenable than others in this regard.

7. What are the different methods of impedance matching?

The various methods are using:

1. L section matching network

2. Single stub matching

3. Double stub matching

4. Quarter wave transformer

8. What is the L section matching network?

The simplest type of matching network is the L-section, which uses two reactive
elements to match an arbitrary load impedance to a transmission line. A simple
electrical impedance-matching network requires one capacitor and one inductor.
5.98 Microwave Design Principles

9. What is single stub tuning?

Another popular matching technique uses a single open-circuited or short-circuited

length of the transmission line (a stub) connected either in parallel or in series with the
transmission feed line at a certain distance from the load. Such a single-stub tuning
circuit is often very convenient because the stub can be fabricated as part of the
transmission line media of the circuit, and lumped elements are avoided. Shunt stubs
are preferred for microstrip line or stripline, while series stubs are preferred for slot
line or coplanar waveguide.

10. What are the drawbacks of single stub matching?

The single-stub tuner can match any load impedance (having a positive real part) to a
transmission line but suffers from the disadvantage of requiring a variable length of
line between the load and the stub. This may not be a problem for a fixed matching
circuit but would probably pose some difficulty if an adjustable tuner was desired. In
this case, the double-stub tuner, which uses two tuning stubs in fixed positions, can be
used. Such tuners are often fabricated in coaxial lines with adjustable stubs connected
in shunt to the main coaxial line.

11. What is double stub matching?

A double-stub matching network matches a complex load impedance (Zload) to a

desired complex input impedance (Zin) using two shunt stubs and a connecting line.

12. What are the applications of the Smith chart?

1. It is used to calculate impedance and admittance on any load.

2. It is used to find Vmax, Vmin, SWR, and reflection coefficient K.

3. It is used to find the length and position of the stub.

13. What is a microwave filter?

Microwave filters are two-port, reciprocal, passive, linear devices that heavily
attenuate the unwanted signal frequencies while permitting transmission of wanted
frequencies.

14. List the important filter parameters.

In designing a filter, the following important parameters are generally considered.

1. Pass-band width
 Antennas and Microwave Engineering 5.99

2. Stop-band attenuation and frequencies

3. Input and output impedances

4. Return loss

5. Insertion loss

6. Group delay.

15. What is insertion loss?

The insertion loss is defined as the ratio of incident power to the load power. It is
given by IL(dB) = 10 log incident power(Pi)

Load power (PL)

16. Define return loss.

It is defined as the ratio of incident power to the reflected power, which tells about the
amount of impedance matching at the input port.

17. Define the group delay.

The group delay is important for the multi-frequency or pulsed signals to determine
the frequency dispersion or deviation from constant group delay over a given
frequency band.

18. What are the commonly used filters?

The commonly used filters are

1. Low pass filter

2. High pass filter

3. Bandpass filter

4. Bandstop filter

19. What are the types of filter design?

Two methods are normally used. They are

1. Image parameter method

5.100 Microwave Design Principles

2. Insertion loss method

20. What are the steps involved in the insertion loss method?

Insertion loss method consists of the following steps:

1. Design of a prototype low-pass filter with the desired passband characteristics.

2. Transformation of this prototype network to the required type (low-pass, high-

pass, band-pass, or band-stop) filter with the specified center and band-edge
frequencies.

3. Realization of the network in microwave form by using sections of microwave

transmission lines whose reactance are corresponded to those of distributed circuit
elements.

21. List the advantages of microstrip filters.

Microstrip filters are used in satellite, airborne communication, and EW systems for
their small size, lightweight, and low cost.

22. What is a microwave amplifier?

A microwave amplifier having a base element whose reactance can be varied

periodically by an ac voltage at a pumping frequency.

23. Define power gain.

Power gain=G=PL/Pin is the ratio of power dissipated in the load ZL to the power
delivered to the input of the two-port network. This gain is independent of ZS,
although the characteristics of some active devices may be dependent on ZS.

24. Define available power gain and transducer power gain.

Available power gain=GA=Pavn/Pavs is the ratio of the power available from the two-
port network to the power available from the source. This assumes conjugate matching
of both the source and the load, and depends on ZS, but not ZL. Transducer power
gain=GT=PL/Pavs is the ratio of the power delivered to the load to the power available
from the source. This depends on both ZS and ZL.

25. What are the two types of stability?

Unconditional stability: The network is unconditionally stable if |Γin|<1 and |Γout|<1 for
all passive source and load impedances (i.e., |ΓS|<1 and |ΓL|<1).
 Antennas and Microwave Engineering 5.101

Conditional stability: The network is conditionally stable if |Γin|<1 and |Γout|<1only for
a certain range of passive source and load impedances. This case is also referred to as
potentially unstable.

26. What is Rollet’s condition?

Rollet’s condition (K-∆) test is used for checking the unconditional stability.

and

27. What is a microwave power amplifier?

Power amplifiers are used in the final stages of radar and radio transmitters to increase
the radiated power level. Typical output powers may be on the order of 100–500 mW
for mobile voice or data communications systems, or in the range of 1–100 W for
radar or fixed-point radio systems. Important considerations for RF and microwave
power amplifiers are efficiency, gain, intermodulation distortion, and thermal effects.

28. What is drain efficiency?

It is defined as the ratio of RF output power to the DC input power. It is also called
as collector efficiency.

29. What is the compressed gain?

Another useful parameter of a power amplifier is compressed gain G1, defined as the
gain of the amplifier at the 1 dB compression point. Thus, if G0 is the small-signal
(linear) power gain, we have G1(dB)=G0(dB)−1.

30. What is a microwave oscillator?

RF and microwave oscillators are found in all modern wireless communications, radar,
and remote sensing systems to provide signal sources for frequency conversion and
carrier generation. A solid-state oscillator uses an active nonlinear device, such as a
diode or transistor, in conjunction with a passive circuit to convert DC to a sinusoidal
steady-state RF signal. Basic transistor oscillator circuits can generally be used at low
frequencies, often with crystal resonators to provide improved frequency stability and
low noise performance.

31. List the important consideration of oscillators in the microwave system.

5.102 Microwave Design Principles

Important considerations for oscillators used in RF and microwave systems include the
following:

1. Tuning range (specified in MHz/V for voltage-tuned oscillators)

2. Frequency stability (specified in PPM/◦C)

3. AM and FM noise (specified in dBc/Hz below the carrier, offset from the carrier)
Harmonics (specified in dBc below carrier)

32. Write the advantages of a transistor oscillator.

1. Transistor oscillators generally have a lower frequency and power capabilities

than diode sources (e.g., tunnel, Gunn, or IMPATT diodes), but offer several
advantages over diodes. First, oscillators using transistors are readily compatible
with monolithic integrated circuitry, allowing easy integration with transistor
amplifiers and mixers, while diode devices are often less compatible.

2. Also, a transistor oscillator circuit is much more flexible than a diode source. This
is because the negative resistance oscillation mechanism of a diode is determined
and limited by the physical characteristics of the device itself, while the operating
characteristics of a transistor can be adjusted to a greater degree by the bias point,
as well as the source or load impedances presented to the device. Transistor
oscillators usually allow more control of the frequency of oscillation, temperature
stability, and output noise than do diode sources.

3. Transistor oscillator circuits also lend themselves well to frequency tuning, phase
or injection locking, and various modulation requirements. Transistor sources are
relatively efficient but usually are not capable of very high-power outputs.

33. Draw the model of a transistor oscillator.

In a transistor oscillator, a negative resistance one-port network is effectively created

by terminating a potentially unstable transistor with an impedance designed to drive
the device in an unstable region. The circuit model of a transistor oscillator is shown
below. In this circuit, the RF output port is part of the load network on the output side
of the transistor, but it is also possible to use the terminating network to the left of the
transistor as the output port.
 Antennas and Microwave Engineering 5.103

34. What is a microwave mixer?

A mixer is a three-port device that uses a nonlinear or time-varying element to achieve

frequency conversion an ideal mixer produces an output consisting of the sum and
difference frequencies of its two input signals. The operation of practical RF and
microwave mixers is usually based on the nonlinearity provided by either a diode or a
transistor.

35. Define image frequency.

Image frequency: In a receiver, the RF input signal at frequency fRF is typically

delivered from the antenna, which may receive RF signals over a relatively wide band
of frequencies. For a receiver with an LO frequency fLO and IF frequency fIF, gives the
RF input frequency that will be down-converted to the IF frequency as fRF=fLO+fIF.
Now consider the RF input frequency given by fIM=fLO−fIF.

36. List the various types of mixers.

1. Single-ended diode mixer

2. Single-ended FET mixer

3. Balanced mixer

4. Image reject mixer

5. Differential FET Mixer and Gilbert cell mixer.

37. Define the noise Figure

Noise Figure of a two-port amplifier is given by

F= Fmin + RN / GS │YS –YOPT│2

where the following definitions apply:

YS=GS+jBS=source admittance presented to a transistor.

5.104 Microwave Design Principles

Yopt=optimum source admittance that results in minimum noise Figure

Fmin=minimum noise Figure of transistor, attained when YS=Yopt.

RN=equivalent noise resistance of the transistor.

GS=real part of source admittance.

38. What is a Gilbert cell mixer?

The Gilbert cell mixer uses two singly balanced FET mixers of the type to form a
double-balanced mixer. It has fully balanced (differential) ports for LO, RF, and IF
signals. Due to the symmetry of the circuit and its ex-citations, the RF and LO signals
are cancelled at the IF output port. Operation is the same as the singly balanced FET
mixer, with the four upper FETs operating as switches controlled by the LO voltage,
and the lower two FETs operating as a balanced amplifier for the RF input signal.

39. What are the two mixer characteristics?

1. Frequency upconversion

2. Frequency down-conversion.

40. List the applications of microwave mixers.

Mixer circuits can be used to shift the frequency of an input signal like as in a receiver.
They can also be used as a product detector, modulator, frequency multiplier, or phase
detector.

41. List the applications of an oscillator.

They provide a critical clocking function for high-speed digital systems, generate
electromagnetic energy for radiation, enable frequency up and down conversion when
used as local oscillators, and are used as a reference source for system
synchronization.
 Antennas and Microwave Engineering 5.105

REVIEW QUESTIONS

1. Explain the impedance of impedance transformation in transmission lines.

2. With one example, explain how the L section is used in impedance matching.

3. Explain the concept of single stub impedance matching with one example.

4. List the drawbacks of single stub matching. How double stub matching overcomes?
Explain.

5. Explain how QWT is used in impedance matching?

6. Explain the design of the microwave filter using the insertion loss method.

7. Derive the design equation of microwave low pass filter.

8. Derive the equation for unilateral transducer power gain.

9. Explain the concept of stability circles with neat diagrams.

10. Explain the design of a single-stage transistor amplifier design.

11. Explain the various types of broadband transistor amplifiers.

12. With a neat diagram, explain the operation of low noise amplifier.

13. Discuss the design of microwave oscillators using i) Transistors ii) Dielectric
resonator.

14. Explain the operation of the microwave power amplifier with neat sketches.

15. Explain the design of class A power amplifier with diagrams.

16. Explain the operation of the single-ended diode mixer.

17. Discuss the working principle of i) Single-ended FET mixer ii) Balanced mixer.

18. Explain Differential FET Mixer and Gilbert cell mixer.