0% found this document useful (0 votes)
143 views

Lecture 9: Quarter-Wave-Transformer Matching.: R JZ L Z Z Z JR L

Uploaded by

Ruth Enorme
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
143 views

Lecture 9: Quarter-Wave-Transformer Matching.: R JZ L Z Z Z JR L

Uploaded by

Ruth Enorme
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 13

Whites, EE 481 Lecture 9 Page 1 of 5

Lecture 9: Quarter-Wave-Transformer
Matching.

For a TL in the sinusoidal steady state with an arbitrary resistive


load (Fig. 2.16)

the input impedance of the right-hand TL is given as


R  jZ1 tan 1l
Z in  Z1 L (2.61),(1)
Z1  jRL tan 1l
Now imagine that we have a special length l  1 / 4 of TL, as
indicated in the figure above. At this frequency and physical
length, the electrical length of the TL is
2 1 
1l   rad (2)
1 4 2
Consequently, for a  / 4 -length TL, tan 1l   . Using this
result in (1) gives
Z12
Z in  (2.62),(3)
RL
This result is an interesting characteristic of TLs that are exactly
/4 long. We can harness this characteristic to design a matching
network using a /4-length section of TL.

© 2013 Keith W. Whites


Whites, EE 481 Lecture 9 Page 2 of 5

Note that we can adjust Z1 in (3) so that Z in  Z 0 . In particular,


from (3) with Z in  Z 0 we find
Z1  Z 0 RL (2.63),(4)
In other words, a /4 section of TL with this particular
characteristic impedance will present a perfect match (   0 ) to
the feedline (the left-hand TL) in the figure above.

This type of matching network is called a quarter-wave


transformer (QWT). Through the impedance transforming
properties of TLs, the QWT presents a matched impedance at its
input by appropriately transforming the load impedance.

This is accomplished only because we have used a very special


characteristic impedance Z1 , as specified in (4).

Three disadvantages of QWTs are that:


1. A TL must be placed between the load and the feedline.
2. A special characteristic impedance for the QWT is
required, which depends both on the load resistance and
the characteristic impedance of the feedline.
3. QWTs work perfectly only for one load at one frequency.
(Actually, it produces some bandwidth of “acceptable”
VSWR on the TL, as do all real-life matching networks.)
Whites, EE 481 Lecture 9 Page 3 of 5

Real Loads for QWTs

Ideally, a matching network should not consume (much) power.


In (4) we can deduce that if instead of RL we had a complex
load, then the QWT would need to be a lossy TL in order to
provide a match. So, QWTs work better with resistive loads.

However, if the load were complex, we could insert a section of


TL to transform this impedance to a real quantity (is this
possible?), and then attach the QWT. But, again, this would
work perfectly for only one load at one frequency.

Adjusting TL Characteristic Impedance

We see in (4) that the QWT requires a very specific


characteristic impedance in order to provide a match
Z1  Z 0 RL
With coaxial cable, twin lead, and other similar TLs this is often
not a practical solution for a matching problem.

However, for stripline and microstrip adjusting the characteristic


impedance is as simple as varying the width of the trace.
Consequently, QWTs find wide use in these applications.
Whites, EE 481 Lecture 9 Page 4 of 5

As we’ll see in Lecture 12, the characteristic impedance of a


microstrip

d W r

as a function of W/d is

Z0 
175
150
125
100

70.7 75
50.0 50
25
W d
1.3 2 2.4 4 6 8 10

To construct this curve, it was assumed that  r  3.38 , which is


the quoted specification for Rogers Corporation RO4003C
laminate that we’ll be using in the lab.

Example N9.1: Design a microstrip QWT to match a load of


100  to a 50- line on Rogers RO4003C laminate. Estimate
the fractional bandwidth under the constraint that no more than
1% of the incident power is reflected.
Whites, EE 481 Lecture 9 Page 5 of 5
High Frequency Design From April 2006 High Frequency Electronics
Copyright © 2006 Summit Technical Media
MATCHING NETWORKS

The Yin-Yang of Matching:


Part 2—Practical Matching
Techniques
By Randy Rhea
Consultant to Agilent Technologies

The Standard Quarter- where


The conclusion of this Wavelength Transmis-
ZL′ − 1
article covers transmission sion Line Transformer ρload = (43)
line matching networks, ZL′ + 1

A
plus a discussion of how well-known dis-
(Im [ρ ])
2
characteristics of the load tributed matching
+ Re [ρload ]
load
affects matching band- network is the ρmax = (44)
Re [ρload ]
width and the choice of quarter-wavelength long
network topologies transmission line trans-
former. I will refer to this and the electrical length of the line is given by
network as a type 11. The characteristic (see text)
impedance of this line is given by

Z0 = 4 a2 + b2 − b
RS RL (41) θ12 = tan −1 + 90° (45)
2a
For example, a 100 ohm load is matched to
a 50 ohm source using a 90° line with charac- where
teristic impedance 70.71 ohms. The matchable
space of the quarter-wavelength transformer a = Z12 X L (46)
is small, essentially only the real axis on the
Smith chart. Nevertheless, it enjoys
widespread use. A quarter-wavelength line is
also used in filter design as an impedance
inverter to convert series resonant circuits to
parallel resonance, and vice versa [4].

The General Transmission Line


Transformer
Perhaps less well-known is that a single
series transmission line can match
impedances not on the axis of reals. The
matchable space of this type 12 network is
plotted in Figure 9. The characteristic
impedance of the series line is given by

Figure 9 · By allowing line lengths other than


1 + ρmax
Z12 = Z0 (42) 90°, the matchable impedance space for a
1 − ρmax single, series transmission line extends
beyond the real axis.

28 High Frequency Electronics


High Frequency Design
MATCHING NETWORKS

Figure 10 · Shortened double-section quarter-wave-


length transformer for matching real impedances.

b = X L2 + RL2 − Z12
2
(47)

If the real part of the load impedance is less than the


reference impedance and the load is capacitive, then 90°
must be added to the length computed by Eq. 45. If the
real part of the load impedance is less than the reference
impedance and the load is inductive, then 90° must be
subtracted from the computed length.

The Shortened Quarter-Wavelength Transformer Figure 11 · Smith chart plot for a one section L-network
Another less well-known but useful adaptation of the (red) to match 300 ohms to 50 ohms and a three sec-
quarter-wavelength transformer is the shortened, double- tion L-network in blue. Also plotted are corresponding
section transformer depicted in Figure 10. I will refer to constant Q curves (dashed).
this as type 13. Like the standard transformer, it is used
to match real impedances. But the required length is
shorter and it uses lines with characteristic impedance matches at a single frequency. Obtaining a good match
equal to the impedances being matched. These are both over an extended frequency range may require many ele-
practical features in many applications. ments and finding values is very challenging. Before I
Notice that the transmission line with characteristic cover this subject, let me introduce another fundamental
impedance equal to the load is adjacent to the source. concept.
Both transmission lines have the same length. The maxi-
mum line length is 30°, and it decreases as the load Q of the Load
impedance is much higher or lower than the source The term Q is used for several properties. Mastery of
impedance. each is critical to understanding oscillators, filters,
With normalized load resistance R′L, the line length is matching networks and other circuits [5]. One definition
of loaded Q is the center frequency divided by the 3 dB
1 bandwidth of a resonant circuit response. It is a finite
θ13 = tan −1 (48) value even if the circuit is built using components with
R′ + RL′ + 1
L
2
infinite Q. Component Q, or unloaded Q, is a measure of
RL′ component quality; the ratio of stored energy to dissipat-
ed energy in the component. It is as high as 200 for excel-
For example, a 100 ohm resistive load is matched to 50 lent inductors. But unloaded Q increases with physical
ohms using a single, 90° long line with a characteristic size, so modern miniature inductors have much lower Q.
impedance of 70.7 ohms. With a cascade of 50 and 100 Q of the load described in this section is yet a third defi-
ohm lines, each is 28.13° long for a total length of 56.25°. nition of Q. I often feel engineers would be less confused
if these properties were labeled Q, R and S. However,
The Challenge their definitions have similar roots.
Since all complex loads are matchable by two element Q of the load is a property of a complex termination.
networks and sometimes a single transmission line, why For series impedance it is given simply by
is matching sometimes difficult? For loads with a large
reflection coefficient, the element values may be difficult XL
to realize. This is particularly true for distributed circuits. Qof load = (49)
RL
But more often the problem is bandwidth. A simple circuit

30 High Frequency Electronics


High Frequency Design
MATCHING NETWORKS

Figure 12 · Amplitude transmission Figure 13 · Transmission amplitude Figure 14 · Smith chart plot for a
responses for 300 ohm to 50 ohm responses for 50 to 300 ohm single- one-section quarter-wavelength
matching L-networks. section (red), three-section (blue), transformer (red) and a three-sec-
five-section (green) and eleven- tion transformer (blue) to match 300
section transformers (magenta). to 50 ohms.

and for parallel admittance it is L-network. Plotted in red and blue are the transmission
and return loss responses of the three section L-network.
The 15 dB return loss has a bandwidth of about 17% for
GL
Qof load = (50) the one-section L-network and about 59% for the 3-sec-
BL
tion L-network. The ratio of the bandwidths is 59/17 =
3.5 and the ratio of the Q values is 3/0.9 = 3.3. The exact
The dashed red lines in Figure 11 are arcs of constant relationship between the bandwidth and Q arcs depend
Q of the load. These Q arcs pass through X′L = ±2.3 where on the return loss used to define the bandwidth.
R′L = 1 so Qof load = 2.3. The dashed blue arcs mark a Q of However, the relation is clear: matching networks with
0.9 since they pass through 0.9+j1.0 and 0.9–j1.0. To impedance arcs that remain closer to the real axis have
understand the significance of Q arcs, consider the solid better bandwidth.
arcs in Figure 11. Using impedance arcs is insightful when designing
both lumped and distributed matching networks. But this
Element Impedance Transforms Plotted on process is increasingly ineffective when attempting to
the Smith Chart match multiple frequencies over a wide bandwidth. The
The solid red arc that begins at 300 ohms (6 normal- blue arcs in Figure 11 were drawn for a single frequency,
ized) on the real axis, right of center, is the resulting 100 MHz. The solid blue return loss response in Figure 12
action at 100 MHz of a shunt 11.84 pF capacitor that reveals a near perfect match has been achieved at 100
transforms the 300 ohm load resistance to 50–j110 ohms. MHz as expected since the last blue arc in Figure 11 ends
The arc from this point to 50+j0 is the result of a series at 50 ohms. But further examination of the blue trace in
inductor of 178.4 nH, or +j110 ohms. These arcs are not Figure 12 reveals that the response is not centered on 100
responses plotted versus frequency but rather the length MHz, the design frequency. The dashed magenta and
of these arcs correspond to increasing values of the capac- brown plots in Figure 12 are the result of optimizing all 6
itor and inductor. These concepts are the basis of network element values to center the response on 100 MHz. When
design using the Smith chart [1, 2]. broad bandwidth is required, drawing arcs on a Smith
The red dashed Q arcs were drawn so they intersect chart, either with pencil and paper charts or by computer
the maximum extent of the solid L-network arcs, so the program, is cumbersome at best. Synthesis routines such
one-section L-network has a Q of 2.3. You can see that the as those in the Impedance Matching module of GENESYS
arcs of the 3-section L-network remain closer to the real are effective. Alternatively, optimization of an initial
axis. The Q of the 3-section L-network is only 0.9. Smith chart design via computer simulation is effective
Plotted in brown and green in Figure 12 are the for problems with well-behaved loads. This will be dis-
transmission and return loss responses of the one section cussed in more detail later.

32 High Frequency Electronics


A Simple Algorithm for Multiple-Section Transformer Loads with only real and no reactive component may be
The bandwidth of a quarter-wavelength transmission matched over infinite bandwidth if an infinite number of
line transformer decreases with loads higher or lower sections are employed. For loads with a reactive compo-
than the desired resistance. The bandwidth is about 22% nent the bandwidth is limited.
for a ratio of 6:1 as with a 300 ohm load in a 50 ohm sys-
tem. The bandwidth can be improved by cascading multi- Fano’s Limit
ple quarter wavelength transmission line sections. Given In a classic paper, Fano [6] offers an elegant formula
in this section is a simple but effective algorithm for com- for predicting the relationship between the bandwidth
puting the required characteristic impedance for each and the best achievable reflection coefficient using a loss-
section. less, infinitely complex matching network. His formula
As introduced in Eq. 41, the characteristic impedance involves Qof load introduced earlier.
of a single section is the square root of the product of the
source and load impedance. With the multiple-section − πQloaded
transformer, each line uses the same formula between Γ min = e
Qof load (56)
intermediate values of impedance, Zi(n). These intermedi-
ate values are computed to be a uniform geometric pro-
gression from the source to the load resistance. where Γmin is the best match that is achievable over a
bandwidth Fupper – Flower.
1
⎛R ⎞N
Step = ⎜ L ⎟ (51)
F0
⎝Z ⎠0 Qloaded = (57)
Fupper − Flower
where N is the number of sections. Then
An algebraic derivative of Fano’s equation is also helpful.
Zi (1) = Step × Z0 (52)

Zi ( n) = Step × Zi ( n − 1) N
(53) −Qof load ln ( Γ min ) (58)
n= 2 Qloaded =
π
Zi ( N ) = RL (54)
An octave bandwidth is Qloaded = 1.5. Using Eq. 2 we
The characteristic impedance of each section is find a return loss of 15 dB is a reflection coefficient of
0.178. To achieve a 15 dB return loss would require
Z ( n) = Zi ( n − 1) Zi ( n) N
n =1 (55) Qof load less than 2.73. For another example, with a load of
300+j300, Qof load = 1.0. The resulting Qloaded = 0.549.
where Zi(0) is the source impedance. Normalized to a center frequency of 1 MHz, a 15 dB
The transmission amplitude responses for single, return loss could be achieved from 0.089 to 1.911 MHz. I
three, five and eleven section transformers are given in obtained a 15 dB return loss from 0.115 to 1.885, a load-
Figure 13. The response of this simple algorithm is some- ed Q of 0.565, with a 26-element L-C network designed
what Legendre in shape. The ripple is not equal across using a direct synthesis routine that is described later. A
the passband and it increases with increasing N and 26-element network is hardly practical, but it illustrates
increasing load reflection coefficient. The ripple band- Fano’s formula provides an absolute limit that is
width of the eleven-section transformer extends from approached with significant effort. Fano’s limit is used to
about 0.16 to 1.8 MHz, more than a decade. Notice how- discover if a solution is possible, thus avoiding effort on
ever a trend of diminishing return for an increasing num- an unsolvable problem. This introduction to Fano’s limit
ber of sections. An Agilent GENESYS workspace for this prepares us to consider broadband matching with reac-
N-section transformer also may be downloaded from the tive loads.
post “Matching Tutorial published in High Frequency
Electronics magazine” at the Founder’s Forum at Reactance Absorption using Filters:
www.eagleware.com. Equal Resistance
In Figure 14, impedance arcs are plotted on a Smith I am often asked “How do you design a lowpass filter
chart for single and three section transformers. These with unequal terminations?” Consider a lowpass filter
plots suggest that an infinite number of sections would between a 50 ohm source and a 300 ohm load. The defini-
result in no departure from the real axis, the Q would be tion of lowpass is a response from DC to an upper fre-
zero and the bandwidth infinite. In fact, this is the case. quency limit. But, at low frequency the shunt capacitors

April 2006 33
High Frequency Design
MATCHING NETWORKS

Figure 15 · Schematic and response of a bandpass filter used to match a 50 ohm load with 20 pF series capacitance.

become high impedance and they vanish, and the series but very slowly.
inductors become low impedance and they vanish. So if the Capacitors or inductors are also absorbed using cook-
components effectively vanish, how can the network book bandpass filter topologies. The simplest form also
match 50 ohms to 300 ohms and have a low-loss response requires equal termination resistance. Bandpass filters
at low frequency? It can’t! The solution is theoretically absorb a much more reactive load at the expense of reduced
unrealizable. A match and maximum power transfer can bandwidth. For example, a fractional bandwidth of 10%, or
be achieved over a limited bandwidth that does not extend bw = 0.1, allows absorption of ten times larger components
to DC. The response is pseudo lowpass. The return loss as seen in Eqs. 59 and 60. Bandpass filters also support
and insertion loss at low frequency are the same values series capacitors and shunt inductors at the load.
that would exist without the network with 50 ohms driv-
ing a 300 ohm termination as computed by Eqs. 2 and 4. bw
Fmax = (61)
The bandwidth degrades and the low-frequency insertion 2πgRCseries
loss increases with increasing source/load resistance ratio.
However, a lowpass filter can match a reactive load up
bw × R
to the cutoff frequency if the load resistance equals the Fmax = (62)
2πgLshunt
source resistance. Consider a complex load with 50 ohms
in parallel with 20 pF capacitance. A standard lowpass fil-
ter with shunt 20 pF or higher capacitor at the output can For example, suppose we wish to match a 50 ohm load
absorb the load capacitance. The shunt capacitor in the with series capacitance of 20 pF to 20 dB return loss from
lowpass is simply reduced by the value of the load capac- 80 to 120 MHz using a three-section Chebyshev bandpass
itance. Likewise, if the load has series inductance, a low- filter. What is the minimum value of series capacitance
pass filter with a series output inductor is used. A match that can be absorbed? Using Eqs. 2 and 4 we find that the
is achieved from DC to a maximum frequency given by required passband ripple is 0.0432 dB. The final g-value
for a three-section, 0.0432 dB ripple Chebyshev is 0.852.
g When using an L-C filter, the center frequency is found
Fmax = (59)
bw × 2πRCshunt geometrically. That is,

gR Fmax = 80 × 120 = 97.98 MHz (63)


Fmax = (60)
bw × 2πLseries
rather than 100 MHz. Therefore bw = 40/97.98 =
where g is the lowpass prototype g-value adjacent to 0.408. From Eq. 61 we see that Cseries = 15.6 pF. Therefore,
the load, C or L is the load capacitance or inductance, R is a series load capacitor of 15.6 pF or larger may be
the load resistance and bw = 1. Increased reactor values absorbed so the 20 pF series capacitance of the load may
may be absorbed by accepting higher passband ripple. g be absorbed. Given in Figure 15 are the schematic and
ranges from 1 for a 5th order Butterworth response to responses of this matching network. The 70 pF network
2.13 for a 1 dB ripple Chebyshev. For example, g = 1.15 for capacitor and the 20 pF load capacitors in series form the
a 0.1 dB ripple 5th order Chebyshev. 20 pF of shunt 15.6 pF capacitor required by the filter. As you can see,
capacitance is matched up to 183 MHz using this filter. the line of distinction between filters and matching net-
The g-value also increases with increasing filter order, works is not always clear.

34 High Frequency Electronics


High Frequency Design
MATCHING NETWORKS

X N −1
Xa = Xa = X
1− N N
X X
Xb = Xb =
N N
X 1− N
Xc = Xc = X
N ( N − 1) N2 Figure 17 · Shunt Norton transform applied to capaci-
tor C2 in Figure 15 (top) and after absorbing C2a into
Figure 16 · Series Norton transform (left) and shunt C1 and transferring the transformer to the load (bot-
Norton transform (right). tom). Notice the output load resistance has changed.

The Norton Transform Reactance Absorption using Filters:


A great source of yang is the Norton transform. This Unequal Resistance
underutilized transform is useful for a wide range of Once the load resistance is transformed, the tech-
matching and filter applications [7]. There are two forms niques described in the section Reactance Absorption
as shown in Figure 16, Norton’s first, or series transform, using Filters: Equal Resistance may be used to handle
and his second, or shunt transform. These transforms reactive loads. For example, in Figure 17 you can see that
convert a series or shunt reactor into three reactors and a series inductors less than 1017 nH or capacitors greater
transformer. The transformer turns ratio, N, is equal to than 2.6 pF can be absorbed. By using initial filters with
the square root of the impedance ratio and it may be a parallel last section, shunt inductors and capacitors are
either greater or less than 1. Notice from the formula that absorbed. I have therefore described procedures for the
one of the three reactors is always negative. Why would exact design of broadband matching networks for resis-
we be interested in a transform that converts one compo- tive or complex loads. Mastering the use of Norton trans-
nent into two positive valued components, one negative forms requires practice but the effort is worthwhile. The
component and a transformer? It seems like a very poor S/Filter module in GENESYS includes convenient rou-
trade. The yang is hidden in the details! tines for learning and using the Norton and many other
To design a matching network, I’ll start with the con- network transformations [3].
ventional cookbook three-section bandpass filter given in
Figure 15. First, swap inductor L2 and capacitor C2. This Determining the Load Type
has no affect on the response. Next, the Norton shunt Knowledge of the nature of the load is required to
transform is applied to C2 using N = 0.408. Ultimately, select a matching network that can absorb the load reac-
this steps 50 ohms up to 300 ohms. The resulting tance. Consider Figure 18 with the impedance of five dif-
schematic is given at the top in Figure 17. The bottom ferent loads plotted vs. frequency on a Smith chart. The
schematic is after C1 and the negative capacitor C2a are brown trace at the lower right is on the bottom half of the
combined and the transformer is shifted to the right and chart so it is capacitive. The reactance ranges from a nor-
is absorbed by increasing the load resistance. For wider malized 1.0 to 2.8, or –50 to –140 ohms on a 50 ohm chart
bandwidth using higher order filters, the Norton trans- while the resistance is constant. This load is a capacitor
form should be applied to as many nodes as possible. To in series with 120 ohms resistance. A matching network
step the impedance up on the right, the shunt Norton with a series capacitor smaller in value than the load
requires a series capacitor to the left of the shunt capaci- capacitance could be used to absorb the load reactance.
tor. In Figure 15 note that this occurs once in this three- The green arc lies on a circle whose left side is tangent to
section bandpass. In higher order filters it will occur more the left end of the real axis. This circle is constant con-
often. The intermediate impedances, and therefore the ductance on an admittance Smith chart. This arc is in the
transformer turn ratios, are defined by Eq. 53 where N is top half of the Smith chart so the load is inductive. It
the number of applied Norton transforms. The Norton intersects the real axis at 0.5 so the load is 25 ohms in
technique for matching is reasonably economic. In other parallel with an inductor.
words, good bandwidth is achieved with a minimum num- The blue and magenta arcs intersect the real axis. All
ber of elements. arcs for positive values of inductors, capacitors and trans-

36 High Frequency Electronics


High Frequency Design
MATCHING NETWORKS

mission line length rotate clockwise on the chart with


increasing frequency. The blue arc is on a circle of con-
stant resistance and it is capacitive at low frequency and
inductive at high frequency. It is a series L-C in series
with constant 25 ohms resistance. A matching network
for this load could end with a series L-C with L larger
than the load and C smaller than the load, or it could end
with the load being the final series resonator in a band-
pass filter with alternating series and parallel resonators.
The magenta arc is on a circle of constant conductance
and it is inductive at low frequency and capacitive at high
frequency. It is a parallel L-C in shunt with 100 ohms
resistance. A matching network and corresponding
responses for this load are given in Figure 19. This net-
work is a conventional five-section bandpass filter that
was designed so that the final resonator equaled the
load’s parallel L-C. Also, using a Norton transform, the
center shunt capacitor was transformed into two capaci-
tors, C3 and C4, to step the 50 ohm design impedance of Figure 18 · Various types of loads plotted vs. frequen-
the filter up to 100 ohms to equal the load resistance. cy on a Smith chart.

Ill-Behaved Loads
The red trace in Figure 18 is more interesting. It is optimization would be ineffective.
modeled data for a 9.24 meter tall, 0.3 meter diameter To match this antenna from 7 to 14.35 MHz I used the
monopole antenna tower mounted over ground. At low L-C Bandpass algorithm in the Impedance Match module
frequency it is capacitive. With increasing frequency it of GENESYS. This algorithm finds an appropriate topolo-
intersects the real axis at 35 ohms, it becomes inductive gy and computes element values based on user specified
and then intersects the real axis at 400 ohms. At high fre- frequency limits and the network order. This algorithm
quency it is again capacitive. It is generally parallel to cir- uses four general automated steps. First, a best-fit RLC
cles of constant resistance at the first resonance and par- model is determined for the load. The algorithm then
allel to circles of constant conductance at the higher fre- finds the poles and zeros of the matching network using a
quency resonance. As is typical with monopole antennas, Chebyshev approximation [8,9]. A network is then syn-
it has two resonances, a series resonance at 6.75 MHz fol- thesized using a continued-fraction expansion. The first
lowed by a parallel resonance at 14.25 MHz. The double three steps are very effective at finding the correct topol-
resonance, and the fact that the resistance ranges from ogy and in dealing with reactive loads. It is not particu-
34.4 ohms at series resonance to 405 ohms at parallel res- larly effective in transforming the resistive component of
onance makes this load difficult to match. In fact, the match. Therefore, as the fourth step, the user launch-
attempting to match this load manually using an auto- es an optimization that has been pre-defined by the algo-
mated Smith chart program or a modern simulator with rithm. This four-step process is extremely effective in

Figure 19 · A matching network for the parallel resonator load plotted in magenta in Figure 18 (left) and the cor-
responding responses (right).

38 High Frequency Electronics


High Frequency Design
MATCHING NETWORKS

Figure 20 · Six element network designed to match the monopole antenna from 7 to 14.3 MHz (left) and the
resulting VSWR plotted on the right (blue). Also plotted are the original monopole VSWR (red) and the monopole
matched with a single L network (green).

dealing with ill-behaved loads. Simulation and HF Filter Design and Computer
On the left side of Figure 20 is a six-element matching Simulation and has taught seminars on oscillator and fil-
network designed by this algorithm for the monopole ter design.
antenna. Plotted in blue on the right is the resulting
VSWR maintained at roughly 2:1 from 7 to just over 14 References
MHz. The original antenna VSWR is plotted in red. The 1. P. Smith, Electronic Applications of the Smith Chart,
resistance at the series resonance is close to 50 ohms and 2nd edition, 1995, SciTech/Noble Publishing, Raleigh,
the raw VSWR is quite good. But the resistance at paral- North Carolina.
lel resonance is high and the VSWR is poor. Plotted in 2. G. Parker, Introduction to the Smith Chart, (CD-
green is the resulting VSWR with a simple L-network ROM tutorial), 2003, SciTech/Noble Publishing, Raleigh,
that successfully centered the frequency but it has poor North Carolina.
bandwidth and the VSWR at the desired band edges is 3. Agilent Technologies, EEsof EDA Division, Santa
over 5:1. Rosa, CA. www/agilent.com/find/eesof
4. G. Matthaei, L. Young, and E.M.T. Jones, Microwave
Summary Filters, Impedance Matching Networks and Coupling
I had two goals in this tutorial series on matching. Structures, Artech House, Dedham, MA, 1967/1980.
First, I introduced many important concepts in matching 5. R. Rhea, Practical Issues in RF Design (3 set CD-
to help you grasp the underlying objectives in matching ROM tutorial), 2003, SciTech/Noble Publishing, Raleigh,
network design. Second, I provided formula and tech- North Carolina.
niques for the practical design of many types of both L-C 6. R. M. Fano, “Theoretical Limitations on the
and distributed networks. Not included were techniques Broadband Matching of Arbitrary Impedances,” Journal
applicable to transmission line transformers. The later of the Franklin Institute, January 1950.
are particularly effective in dealing with resistive loads in 7. R. Rhea, Filter Techniques (3 set CD-ROM tutorial),
the HF and VHF frequency range. This topic is well cov- 2003, SciTech/ Noble Publishing, Raleigh, North Carolina.
ered by Sevick [10,11,12]. 8. R. Levy, “Explicit Formulas for Chebyshev
Impedance-Matching Networks,” Proc. IEEE, June 1964.
Author Information 9. T.R. Cuthbert, Jr., Circuit Design Using Personal
Randall Rhea is a consultant to Agilent Technologies. Computers, John Wiley, New York, 1983.
He received a BSEE from the University of Illinois and 10. J. Sevick, Transmission Line Transformers, 4th ed.,
and MSEE from Arizona State. He worked at the Boeing 2001, SciTech/ Noble Publishing, Raleigh, North Carolina.
Company, Goodyear Aerospace and Scientific-Atlanta. He 11. J. Sevick, “Design of Broadband Ununs with
is the founder of Eagleware Corporation which was Impedance Ratios Less Than 1:4,” High Frequency
acquired by Agilent Technologies in 2005 and Noble Electronics, November, 2004.
Publishing which was acquired by SciTech Publishing in 12. J. Sevick, “A Simplified Analysis of the Broadband
2006. He has authored numerous papers and tutorial Transmission Line Transformer,” High Frequency
CDs, the books Oscillator Design and Computer Electronics, February, 2004.

40 High Frequency Electronics

You might also like