Lecture 9 - Lumped Element Matching Networks

Microwave Active Circuit Analysis and Design

Clive Poole and Izzat Darwazeh

Academic Press Inc.

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide1 of 55

 Intended Learning Outcomes

I Knowledge
I Understand the theory of L-section matching networks both analytically and graphically.
I Understand forbidden regions on the Smith Chart for various types of L-section.
I Understand the theory of three element matching networks (π-section and T-section).
I Understand bandwidth performance of lumped element matching networks.
I Skills
I To be able to select the appropriate L-section matching network (type 1 or type 2)
depending on the values of impedances to be matched.
I Be able to design an L-section matching network to match two arbitrary impedances
using either the analytical or Smith Chart (graphical) approach.
I Be able to design a T or π matching network to match two arbitrary impedances.

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide2 of 55

Table of Contents

The need for impedance matching

L-section matching networks

L-section matching network design using the Smith Chart

Three element matching networks

Bandwidth of lumped element matching networks

T to π transformation

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide3 of 55

 Impedance Matching

An impedance matching network is used between two dissimilar impedances in order

to ensure maximum power transfer between them. We typically want to match an
arbitrary load ZL to a transmission line Zo . The conditions for maximum power transfer
are therefore :

Zin∗ = Zo
Zo Matching Network ZL

and
∗
Zout = ZL
Zin Zout

The need for Impedance Matching :

I In many applications we require Maximum Power Transfer into the load. This is
achieved when the load to be matched to the line.
I An impedance matching network is required to present the optimum source
impedance to the input of a low noise amplifier, in order to achieve minimum noise
figure.

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide4 of 55

Table of Contents

The need for impedance matching

L-section matching networks

L-section matching network design using the Smith Chart

Three element matching networks

Bandwidth of lumped element matching networks

T to π transformation

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide5 of 55

 L-Section Matching Networks

I This is the simplest lumped component matching network, consisting of only two
lumped components : a shunt susceptance, jB, and a series reactance, jX.
I There are two basic configurations of L-Section matching network, depending on
the location of the shunt element. We shall refer to these as ’type 1’ and ’type 2’ :

jX jX

ZS jB ZL ZS jB ZL

Figure 1 : L-section : Type 1 Figure 2 : L-section : Type 2

Generally speaking, the shunt element, jB, is placed in parallel with the larger of ZS or
ZL .

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide6 of 55

 High pass vs Low pass L-Section networks

High Pass Configuration :

C
I

I When the series component is a

capacitor, it will block DC into the load. ZS L ZL
I The shunt inductor will act as a short
at low frequencies.

I Low Pass Configuration : L

I When the series component is an

inductor, it will allow DC into the load,
but will attenuate higher frequencies. ZS C ZL

I The shunt capacitor will act as a short

at high frequencies.

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide7 of 55

 L-section matching of a resistive source to a resistive load

Let us assume that we need to connect a resistive load RL to a resistive source RS , as

shown in figure 3(a), where, in the general case, RS 6= RL .

jX

RS RS

jB RL

RL

Matching

(a) (b)

Figure 3 : Matching arbitrary RL and RS

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide8 of 55

 L-section matching of a resistive source to a resistive load

We need to place a matching network between RS and RL , as in figure 3(b), to ensure

that the source is terminated by an impedance equal to itself, thereby ensuring
maximum power transfer.
Since we restrict ourselves to using only reactive elements in the matching network, the
matching of the two resistive elements proceeds as follows :
1. Firstly, we place a reactive element, represented by the susceptance jB, in parallel
with RL , such that the resistive part of the resulting combination is equal to RS .
2. We then cancel the reactive part of the combination (jB k RL ) by adding the equal
and opposite series reactive element jX .

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide9 of 55

 L-section matching of a resistive source to a resistive load

We can analyse the circuit of figure 3(b) as follows : Since we know that, to satisfy the
matching condition, the total impedance of the parallel combination (jB k RL ) should be
the complex conjugate of RS + jX , we can write :

1 RL + jBRL2
RS + jX = = (1)
(1/RL ) − jB 1 + B2 RL2
From (1) we get RS and X in terms of RL and B as follows :

RL
RS = (2)
1 + B2 RL2
and

BRL2
!
X= (3)
1 + B2 RL2

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide10 of 55

 L-section matching of a resistive source to a resistive load

Equation (??) implies that a real value of Q is only obtained if RL /RS > 1. If this is not
the case then we need to reverse the position of X and B in figure 3(b), in other words
B is placed in parallel with the source, not the load. We can apply exactly the same
design procedure, only treating the source as if it were the load and vice versa. We can
therefore write (??) in a form that covers any values of RS and RL as :
s
Rhigh

Q= −1 (4)
Rlow
Where Rhigh is the higher value of RS and RL and Rlow is the lower value. Another way
of intuitively understanding where the parallel arm should be placed is to consider that,
if RL > RS then RL needs to be reduced by adding a parallel resistance. On the other
hand, if RL < RS then it needs to be increased by adding a series resistance.

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide11 of 55

 L-section matching of a resistive source to a resistive load

We can now set out the basic design procedure for an L-section to match resistive
loads as follows :
1. Calculate the Q for a given RS and RL using (??) (Note the orientation of the
parallel arm based on whether (RL /RS ) is greater or less than unity).
2. Calculate B from :
Q
B=± (5)
RL
3. Calculate X from (??).
Note that the sign of B in step 2 above may be chosen arbitrarily, since the load is
purely resistive and we are free to choose B to be either capacitive or inductive. The
difference being that the type of reactance chosen for B will determine whether the
L-section has a high or low pass frequency characteristic away from the centre
frequency. If a negative value of B is chosen (i.e. parallel capacitance) then the
L-section will have a low pass characteristic. If a positive value of B is chosen (i.e.
parallel inductance) then the L-section will have a high pass characteristic.

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide12 of 55

 L-section matching of a complex source and load

For simplicity, we will restrict our analysis to the most common situation, namely where
we need to match the complex load ZL to the system characteristic impedance, Zo . As
before, the choice of whether to use a type 1 or type 2 matching network will depend on
the relationship between the resistive part of the load, RL , and Zo . As was shown for
the case of purely resistive loads, the parallel element, jB, should be placed in parallel
with whichever is larger of RL or Zo , in other words :
If RL > Zo : use type 1 L-section (shunt element is next to the load).
If RL < Zo : use type 2 L-section (shunt element is next to the source).

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide13 of 55

 L-section matching of a complex source and load

In general, the load and source will be complex. We can generalise the above
technique to cover a complex ZL and ZS as shown in figure 4 by considering only the
resistive parts of ZL and ZS first and then absorbing the reactive parts into the resulting
matching components X and B.

jX jX

ZS jB ZL ZS jB ZL

(a) L-section : Type 1 (b) L-section : Type 2

Figure 4 : Generalised L-section matching networks

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide14 of 55

 Analytical design of Type 1 L-section

Consider a generalised type 1 L-section used to connect an arbitrary load, ZL to a

transmission line of characteristic impedance Zo :

jX

Zo jB ZL

Zin Matching network

From the figure above we have three equations with 2 unknowns :

ZL = RL + jXL (known)
1
Zin = jX + jB+(RL +jXL )−1
(B, X : unknown) (6)
ZS = Zo (known)

Matching is defined as Zin = ZS . This implies Zin = Zo .

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide15 of 55

 Analytical design of Type 1 L-section

From the previous slide :

1
Zo = jX + (7)
jB + (RL + jXL )−1
The above can be rearranged and separated into real and imaginary parts to yield the following pair
of equations :
B(XRL + jXL Zo ) = RL − Zo (8)
X (1 − BXL ) = BZo RL − XL (9)
Which can be solved to yield :
q q
RL
XL ± Zo RL2 + XL2 − Zo RL
B= (10)
RL2 + XL2

Once B is determined, X can be found using equation (9):

BZo RL − XL
X = (11)
1 − BXL

Note that the term inside the square root in equation (10) can be negative. Therefore type 1 is only
used in the case when Rs > RL .

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide16 of 55

 Analytical design of Type 2 L-section

Consider a generalised type 2 L-section used to connect an arbitrary load, ZL to a

transmission line of characteristic impedance Zo :

jX

Zo jB ZL

Zin Matching network

From the figure above we have three equations with 2 unknowns :

ZL = RL + jXL (known)
1
Yin = jB + RL +j(X +XL )
(B, X : unknown) (12)
ZS = Zo (known)

Matching is defined as Zin = ZS . This implies Zin = Zo .

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide17 of 55

 Analytical design of Type 2 L-section
From the previous slide :
1
Yin = jB + (13)
RL + j(X + XL )
The above can be rearranged and separated into real and imaginary parts to yield the
following pair of equations :

BZo (X + XL ) = Zo − RL (14)

(X + XL ) = BZo RL (15)
Which can be solved to yield :
s
1 (Zo − RL )
B=± (16)
Zo RL

and
p
X = ± RL (Zo − RL ) − XL (17)

Once again, the requirement that RL < Zo ensures that the terms under the square
roots in the expressions for B and X and are real. Again, with two solution pairs for B
and X , there are two different matching network solutions.

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide18 of 55

 The effect of adding reactive elements

L-Section design is best performed using the

Admittance/Impedance Smith chart of figure ??.
I Adding series reactive loads will modify the

1
0.9
0.8
0.7
impedance by adding negative reactance

1.5
0.6
(series C), or positive reactance (series L)

5
0.

2
4
0.
I Adding shunt reactive loads will modify the -b

admittance by adding negative susceptance

0 .3 3
+x
Shunt L
(shunt C), or positive susceptance (shunt L) 0.2 Series L 4
A 5
Let’s say we start at point ’A’ : 0.1 Series C
Shunt C
I Adding a series component: 10
-x +b
20
0.1 0.2 0.3 0.4 0.5 1 1.5 2 3 4 5 10 20
I move along constant resistance circle 0

I Series L : move clockwise −20

−0.1
I Series C : move counter-clockwise −10

Constant Resistance Circle

−0
.2
I Adding a shunt component: −5

.3 −4
−0 Constant Conductance Circle

I move along constant conductance 4 −3

0.
circle −

Shunt L : move counter-clockwise

5 .
I

−0

−2
.6
Shunt C : move clockwise

−0
I

.5
−0.7

−1
−0.8

−0.9

−1
I Generally, the point is to arrive at the origin
(Zin = 50Ω)

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide19 of 55

Table of Contents

The need for impedance matching

L-section matching networks

L-section matching network design using the Smith Chart

Three element matching networks

Bandwidth of lumped element matching networks

T to π transformation

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide20 of 55

 The effect of adding reactive elements

From the previous slide we can draw some general conclusions :

1. only the unit resistance and unit conductance circles pass through the origin,
therefore :
1.1 Adding series L or C alone will only match those loads lying on the unit resistance circle.
1.2 Adding shunt L or C alone will only match those loads lying on the unit conductance
circle.
2. If the load is inductive (i.e. it lies in the upper half of the Smith Chart) then we
need to add series C or shunt C to match.
3. If the load is capacitive (i.e. it lies in the lower half of the Smith Chart) then we
need to add series L or shunt L to match.
Points 2 and 3 above are kind of obvious : we need to add opposite reactances to
cancel out the load reactance.

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide21 of 55

 Unit resistance / conductance circles

1
0.9
0.8
0.7

1.5
0.6
5
0.

2
4
0.

There are two important 0 .3 3

circles on the Smith chart 4

that we need to be aware
0.2
5

of when designing 0.1

L-sections : 10

20
I Unit resistance circle 0 0.1 0.2 0.3 0.4 0.5 1 1.5 2 3 4 5 10 20

: the locus of all −20

impedances of the −0.1

−10

form z = 1 ± jx Unit Conductance Circle Unit Resistance Circle

−0
.2

Unit conductance
−5
I
.3 −4
circle : the locus of −0

all admittances of the 4 −3

0.
−

form y = 1 ± jb
5 .
−0

−2
.6
−0

.5
−0.7

−1
−0.8

−0.9

Figure 5 : Unit conductance and resistance circles −1

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide22 of 55

 Demarcation regions on the Smith Chart
1
0.5 2
I The choice of matching network Region 3
topology depends on the location of the
load, zL , on the Smith chart.
0.2 5
I Richard Li [?] has proposed the division
of the Smith Chart into 4 distinct regions
as follows :
0.2 0.5 1 2 5
Region 1: Low resistance or High 0
conductance loads
Region 2: High resistance or Low Region 1 Region 2
conductance loads
Region 3: Low resistance and Low −5
−0.2
conductance loads
Region 4: Low resistance and Low
conductance loads Region 4

−0.5 −2
−1

Region 1 Region 2 Region 3 Region 4

r<1 r>1 r<1 r<1
x<|0.5| −∞ < x < +∞ x>0 x<0
g>1 g<1 g<1 g<1
−∞ < b < +∞ b<|0.5| b<0 b>0

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide23 of 55

 Forbidden Regions for L-Sections

I Not every L-network topology can perform the Zo C ZL

required matching between arbitrary load and source

impedances.
I There is no guarantee that a solution exists for any
given L-section. Depending on the L-Section topology,
there will be Forbidden Areas on the Smith Chart
which cannot be matched.
I For example, assuming a 50Ω source, addition of a
Forbidden region
shunt capacitance will result in moving clockwise
away from the origin along the constant conductance
circle , as shown.
Zo
I This implies that all load impedances within the
shaded region opposite cannot be matched with this Shunt C
particular configuration.
Matchable region

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide24 of 55

 Forbidden Region: L-section type 1 (LC)

Type 1 L-Section has the series element next to the source and the parallel element next to the load.

C
L

ZS L ZL ZS C ZL

Type 1a can only match capacitive loads lying outside unit admittance Type 1b can only match inductive loads lying outside unit admittance
circle or inductive loads lying inside unit resistance circle circle or capacitive loads lying inside unit resistance circle

Inductive Loads Inductive Loads

Forbidden region

Forbidden region

Capacitive Loads Capacitive Loads

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide25 of 55

 Forbidden Region: L-section type 1 (LL/CC)

Type 1 L-Section has the series element next to the source and the parallel element next to the load.

C1
L1

ZS L2 ZL ZS C2 ZL

Type 1c can only match capacitive loads lying outside both unit circles Type 1d can only match inductive loads lying outside both unit circles

Inductive Loads Inductive Loads

Forbidden region

Forbidden region

Capacitive Loads Capacitive Loads

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide26 of 55

 Forbidden Regions : L-section type 2 (LC)

Type 2 L-Section has the series element next to the load and the parallel element next to the source.

C
L

ZS C ZL ZS L ZL

Type 2a can only match capacitive loads lying outside unit resistance Type 2b can only match inductive loads lying outside unit resistance
circle or inductive loads lying inside unit admittance circle circle or capacitive loads lying inside unit admittance circle

Inductive Loads Inductive Loads

Forbidden region

Forbidden region

Capacitive Loads Capacitive Loads

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide27 of 55

 Forbidden Region: L-section type 2 (LL/CC)

Type 1 L-Section has the series element next to the source and the parallel element next to the load.

C2
L2

ZS L1 ZL ZS C1 ZL

Type 2c can only match capacitive loads lying outside both unit circles Type 2d can only match inductive loads lying outside both unit circles

Inductive Loads Inductive Loads

Forbidden region

Forbidden region

Capacitive Loads Capacitive Loads

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide28 of 55

 4-Step Design procedure for L-Sections

Step 1: Normalize (zL =ZL /50) and locate zL on the Smith Chart
Step 2: Select the appropriate L-section topology, based on where ZL lies in relation to the various
forbidden regions.

For type 2 L-section

For type 1 L-section
Step 3: Move along the constant resistance
Step 3: Move along the constant conductance
circle until it intersects with the unit
circle until it intersects with the unit
conductance circle. Record the reactance
resistance circle. Record the susceptance
change and thus determine the value of
change and thus determine the value of
series L or C.
shunt L or C.
Step 4: Move along the unit conductance circle
Step 4: Move along the unit resistance circle to
to the origin, record the susceptance
the origin, record the reactance change and
change and thus determine the value of
thus determine the value of series L or C.
shunt L or C.

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide29 of 55

 L-Section design example

Match a load ZL =(25 + j75)Ω to 50Ω at 10GHz.

Step 1: Normalize zL = (25 + j75)/50 = 0.5 + j1.5 Transmission line : Zo C ZL =25+j75

and locate this point on the Smith Chart

Step 2: Based on the location of zL we select a
type 1 LC L-section.
Step 3: Move clockwise along the constant
conductance circle until it intersect the unit
resistance circle at 1-jx. Susceptance zL
1 1
change = −100 − 75 = −0.0230.
0.023
Therefore C = 2πf = 0.371pF
Zo
Step 4: Move clockwise along the unit resistance
circle to the origin. Reactance change
100
=100Ω. Therefore L = 2πf = 1.59nH .
(1-jx)

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide30 of 55

Table of Contents

The need for impedance matching

L-section matching networks

L-section matching network design using the Smith Chart

Three element matching networks

Bandwidth of lumped element matching networks

T to π transformation

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide31 of 55

 3 element matching networks

I L-sections provide no flexibility over the circuit Q, which is a function of the load
and source impedances.
I High ratios of source and load impedance will have high Q, resulting in a narrower
bandwidth than we would like.
I In order to have control over circuit Q we need to add more flexibility by going from
2-elements to 3-elements in the matching network.
I There are basically two types of 3-element matching network defined by the
topology : the ’T-network’ and the ’π-network’.

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide32 of 55

 Three element matching network
We sometimes need to specify a Q value for the matching network, in which case, we
need more degrees of freedom in the design and this requires more circuit elements.
The next step up in complexity from the two element L-sections just described is the
three element matching network.

There are two basic configurations of three element matching network, which are
referred to as the π-section and the T-section, according to their respective topologies,
as shown in figure 6.

jX jX1 jX2

ZS jB1 jB2 ZL ZS jB ZL

(a) π-matching network (b) T-matching network

Figure 6 : Three element matching networks

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide33 of 55

 π-section matching network
A π-section matching network consists of 3 elements arranged as follows:

jX

ZS jB1 jB2 ZL

Figure 7 : π-section matching network

The π-Section can be considered as consisting of a type 1 L-section and type 2

L-section in cascade as shown below, where X = X1 + X2 :

jX1 jX2

ZS jB1 jB2 ZL

Figure 8 : π-section matching network

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide34 of 55

 The π-section matching network

I The π-section can be considered as consisting of a type 2 L-section followed by a

type 1 L-section in cascade as shown in figure 9, where the central element, X ,
has been split into two reactive elements of the same type, i.e. X = X1 + X2 .
I We also presuppose a "invisible" load resistance, Rx , interposed between the two
L sub-networks in figure 9.
I The purpose of L-section 1, therefore, is to match the source to Rx . Similarly, the
purpose of L-section 2 is to match Rx to the load. The individual L-sections can be
designed according to the principles set out in the previous section, provided we
know the value of Rx , that is.
I The value of Rx can be chosen arbitrarily, but it should be smaller than both RS
and RL , since it is connected to the series arms of the two L-sections[1]. If we start
with a required value of Q, however, this will determine the choice of Rx .

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide35 of 55

 The π-section matching network
Consider the deconstructed π-section matching network shown in figure 9. Applying (4)
we get the loaded Q of L-network 1, which matches RS to Rx . We have already
stipulated that Rx must be smaller that RS so we have:
s 
RS
QL1 = −1 (18)
Rx
Applying the same logic, the loaded Q of L-network 2, which matches Rx to RL (which
is larger than Rx ), is given by :
s 
RL
QL2 = −1 (19)
Rx
Since the loaded Q of the overall circuit is determined by the branch of the circuit
having the highest loaded Q value, we can write the overall Q of the circuit in figure 9,
Qπ , as[1] :
s
Rhigh

Qπ = −1 (20)
Rx
Where Rhigh is the larger of RS and RL . By inspection of (18) to (20) we can see that
the overall Q of the π-section will be equal to the highest Q of the two constituent
L-sections.

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide36 of 55

 The π-section matching network design

With known values of QL1 , QL2 and Rx we proceed with the π-section design by
designing each constituent L-section, using the techniques set out in section ??, and
then combining the central elements.

The reader will note that we have so far only considered the resistive parts of the
source and load, RS and RL . This is because any reactive parts of the source and load
can be absorbed into the parallel branches of the π-section; B1 and B2 .

jX1 jX2

RS jB1 Rx jB2 RL

L-section 1 L-section 2

Figure 9 : π-section matching network as a cascade of two L networks

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide37 of 55

 π-network matching example
As an example, we will design a π-section matching network having a Q of 4, to match
an antenna having an impedance of 20 + j25Ω at 2GHz to a 50 Ω transmission line.
The matching network should be able to pass DC current.

Since the matching network should be able to pass DC current, we need to use the
topology shown in figure 10.

XL (20Ω)

50 Ω line C1 C2

RL (25Ω)

Figure 10 : Π-section matching network example

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide38 of 55

 π-network matching example
Since we are using a π-section, it is more convenient for us to work with an equivalent
load admittance, as we need to absorb the reactive part of the load into the π-section
element jB2 . We therefore convert the load to an equivalent parallel configuration, so
that we can simply add the susceptances. We therefore calculate the load admittance
as :

1
YL = = 0.0195 − j0.0244S
20 + j25
We can now redraw figure 10 broken down into two L-sections in figure 11, which also
shows the virtual resistance, Rx .

L1 L2

50 Ω line C1 Rx C2 BL GL

L-section 1 L-section 2

Figure 11 : π-section matching network example

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide39 of 55

 π-network matching example

Assuming we absorb the equivalent load susceptance, BL , into C2 of the π-section, the
equivalent purely resistive load then becomes :

1 1
RL = = = 51.28Ω
GL 0.0195
We can now calculate the value of Rx using (??) :

Rhigh 51.28
Rx = 2 +1
= = 3Ω
Qπ 42 + 1
We now employ the procedures learned in section ?? to design the two constituent
L-sections in figure 11. Since RL > RS we know that the L-section comprising L2 and
C2 , has the highest Q of the two, i.e. :

QL2 = Qπ = 4

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide40 of 55

 π-network matching example

We now calculate the shunt susceptance, B2 , from (5), remembering to compensate for
the negative load susceptance, BL , by adding an equivalent positive susceptance to B2
:

4
B2 = + 0.0244 = 0.078 + 0.0244 = 0.10240
51.28
We can now calculate the capacitance C2 as follows :

B2 0.1024
C2 = = = 8.15pF
ω 2π × 2 × 109
We now apply equation (??) to determine the reactance of L2 :

QRL 4 × 51.28
X2 = 2
= = 12Ω
1 + Qπ 1 + 42
So, the inductance, L2 , is calculated as :

X2 12
L2 = = = 0.96nH
ω 2π × 2 × 109

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide41 of 55

 π-network matching example

We now repeat the above design procedure for the L-section comprising L1 and C1 . We
firstly need to calculate QL1 using (18) :
s
50

QL1 = − 1 = 3.96
3
QL1 3.95
B1 = = = 0.079 S
RS 50
B1 0.079
C1 = = = 6.30pF
ω 2π × 2 × 109
QL1 RS 3.96 × 50
X1 = 2
= = 11.87Ω
1 + QL1 1 + 3.962
So, the inductance, L1 , is calculated as :

X1 11.87
L1 = = = 0.945nH
ω 2π × 2 × 109

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide42 of 55

 π-network matching example

By combining L1 and L2 to give L = 1.90nH, and removing Rx , we arrive at the final

configuration shown in figure 12 :

1.90nH

50 Ω line 6.30pF 8.15pF ZL = 20 + j25Ω

Figure 12 : π-section matching network example - final configuration

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide43 of 55

 T-section matching network
A T-section matching network consists of 3 elements arranged as follows:

jX1 jX2

ZS jB ZL

The T-Section can be considered as consisting of a type 2 L-section and type 1

L-section in cascade as shown below, where B = B1 + B2 :

jX1 jX2

ZS jB1 jB2 ZL

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide44 of 55

 T-network matching example

As an example, we will design a T-section matching network having a Q of 4, to match

an antenna having an impedance of 20+j25 Ω at 2GHz to a 50 Ω transmission line.
This matching network should be able to pass DC current.

We now proceed to design the type 2 L-section comprising L2 and C2 , shown in

figure 13.

L1 L2

XL (20Ω)

ZS C1 Rx C2

RL (25Ω)

L-section 1 L-section 2

Figure 13 : T-section matching network example

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide45 of 55

 T-network matching example

Since RL < Zo in this case, the Q of the L-section C2 , L2 is equal to the overall Q of the
T-section, i.e. :

QL2 = QT = 4
We can now calculate the shunt susceptance, B2 , from (5), noting that the ’load’, in this
case, is Rx :

QL2 4
B2 = = = 0.0118S
Rx 340
The value of C2 can now be calculated as :

B2 0.0118
C2 = = = 0.94pF
ω 2π × 2 × 109

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide46 of 55

 T-network matching example

Since the matching network should be able to pass DC current, we need to use the
topology shown in figure 14.

L1 L2

XL (20Ω)

C
50 Ω line
RL (25Ω)

Figure 14 : T-section matching network example

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide47 of 55

 T-network matching example

For clarity, we can redraw figure 14 broken down into two L-sections in figure 13, which
shows the virtual resistance, Rx , the value of which can be calculated using (??) :
 
Rx = 20 42 + 1 = 340Ω

We now apply equation (??) to determine the series reactance X2 , noting that we need
to subtract the inductive reactance of the load, XL :

QL2 Rx 4 × 340
X2 = 2
− XL = − 25 = 55Ω
1 + QL2 1 + 42
We can now calculate L2 as :

X2 55
L2 = = = 4.38nH
ω 2π × 2 × 109

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide48 of 55

 T-network matching example

We now repeat the above design procedure for the type 1 L-section comprising L1 and
C1 to give :
s
340

QL1 = − 1 = 2.41
50
QL1 2.41
B1 = = = 0.0071 S
Rx 340
B1 0.0071
C1 = = = 0.56pF
ω 2π × 2 × 109
QL1 Rx 2.41 × 340
X1 = = = 120.4Ω
1 + Q2 1 + 2.412
We can now calculate L1 as :

X1 120.4
L1 = = = 9.58nH
ω 2π × 2 × 109

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide49 of 55

 T-network matching example

We now complete the design by removing Rx and combining C1 and C2 to give

C = 1.5pF . The final design is shown in figure 15.

9.58nH 4.38nH

50 Ω line 1.5pF ZL = 20 +

Figure 15 : T-section matching network example - final configuration

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide50 of 55

Table of Contents

The need for impedance matching

L-section matching networks

L-section matching network design using the Smith Chart

Three element matching networks

Bandwidth of lumped element matching networks

T to π transformation

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide51 of 55

 Bandwidth of lumped element matching networks

The bandwidth of any lumped element matching network can be determined from the
loaded Q of the circuit, QL , by applying (??), i.e.:

fo
∆f = (21)
QL
In the case of L-sections, the Q and therefore the bandwidth, are solely a function of the
load and source resistances. The bandwidth of an L-section can therefore be
determined by applying (4) as follows:

fo
∆f = s (22)
Rhigh

−1
Rlow
The advantage of the three element networks, (T and π), by contrast, is that Q can be
chosen, to some extent, as an independent design parameter. This means that we have
some degree of choice of bandwidth, independent of the load and source resistances,
provided that the chosen Q is larger than that which is available with an L network. This
means that T or π-networks are only really suitable for narrow-band applications.

If wider bandwidth are required, a matching network based on cascaded L-sections

may be used.

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide52 of 55

Table of Contents

The need for impedance matching

L-section matching networks

L-section matching network design using the Smith Chart

Three element matching networks

Bandwidth of lumped element matching networks

T to π transformation

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide53 of 55

 T to π transformation
The well known Star-Delta transformation allows us to convert any given π-network of
generalised impedances into an equivalent T-network and vice versa. With reference to
figure 16 we can write the following:

T to π transformation : Zc

(Z1 Z2 + Z1 Z3 + Z2 Z3 )
Za = (23)
Z2
Za Zb
(Z1 Z2 + Z1 Z3 + Z2 Z3 )
Zb = (24)
Z1
(Z1 Z2 + Z1 Z3 + Z2 Z3 )
Zc = (25) (a) π-network
Z3
Z1 Z2

π to T transformation :

Za Zc
Z1 = (26) Z3
Za + Zb + Zc
Zb Zc
Z2 = (27)
Za + Zb + Zc
(b) T-network
Za Zb
Z3 = (28)
Za + Zb + Zc Figure 16 : π to T transformation

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide54 of 55

 References

C. Bowick.
RF Circuit Design.
Newnes Elsevier, Burlington, MA, USA, 2008.

R. Ludwig and G. Bogdanov.

RF Circuit Design.
Pearson Education Inc., Upper Saddle River, NJ, USA, 2 edition, 2009.

D M Pozar.
Microwave Engineering.
John Wiley and Sons Inc., New York, USA, 2 edition, 1998.

P H Smith.
Electronic Applications of the Smith Chart.
Noble Publishing Corporation, October 2000.

© Poole-Darwazeh 2015 Lecture 9 - Lumped Element Matching Networks Slide55 of 55