University of Central Florida

STARS

Retrospective Theses and Dissertations

1986

A Colpitts Oscillator Design Technique Using S-Parameters

Timothy K. Johnson
University of Central Florida

Part of the Engineering Commons

Find similar works at: https://stars.library.ucf.edu/rtd
University of Central Florida Libraries http://library.ucf.edu

This Masters Thesis (Open Access) is brought to you for free and open access by STARS. It has been accepted for
inclusion in Retrospective Theses and Dissertations by an authorized administrator of STARS. For more information,
please contact [email protected].

STARS Citation
Johnson, Timothy K., "A Colpitts Oscillator Design Technique Using S-Parameters" (1986). Retrospective
Theses and Dissertations. 4961.
https://stars.library.ucf.edu/rtd/4961
A COLPITTS OSCILLATOR DESIGN TECHNIQUE USING S-PARAMETERS

BY
TIMOTHY K. JOHNSON
B.E.E., Auburn University, 1983

RESEARCH REPORT

Submitted in partial fulfillment of the requirements

for the degree of Master of Science in Engineering
in the Graduate Studies Program of the
College of Engineering
University of Central Florida
Orlando, Florida

Fall Term
1986
 ABSTRACT

This research report describes a method for designing

a Colpitts oscillator using S-parameters. The oscillator
components are grouped into three functional blocks: 1) an
unstable active network (which includes the transistor,
feedback capacitor, and input resistor); 2) an output
matching network ( which includes the inductor, tuning
capacitor, and load); and, 3) an input matching network ·
(which consists of the remaining tank capacitor). Thi~

configuration not only satisfies the standard Colpitts

oscillator topology, but allows the use of three simple
criteria (based on the network S-parameters) to predict
oscillation. A computer program was developed to calculate
specific tank component values based on these criteria. An
example oscillator (at 100 MHz) was built to ·· verify the
procedure.
 TABLE OF CONTENTS

LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . iv

Chapter
I. INTRODUCTION. . . . . • . . . . . . . . . . . . . 1
Description of Colpitts Oscillator Topology. 1
General Criteria for Oscillation • . . . 2
II. DESIGN PROCEDURE. . . . . . . . . . . . . . . . . 6
Determination of Active Network S-Parameters 8
Determination of Output Network. . . . . 12
Determination of Input Network . . . . . 15
Description of Computer Program. . . . . 16
III. DESIGN EXAMPLE. . . . . • . . . • . • . . . . 19
Design of Biasing Circuit. . . . . • . . 20
Design of Tank Circuit . . . • • . . 21
Final Circuit Implementation . • • . 25
Matching Network • . . • . . . . • . 25
Circuit Testing. . . • . . . . . . . 28
IV. RESULTS AND CONCLUSIONS 32

Appendices
A. NETWORK PARAMETER TRANSFORMATIONS . 34

B. LISTING OF DESIGN PROGRAM . . . 37

c. EXAMPLE RUN OF DESIGN PROGRAM . 43

References • . . • . . . • • . . . 46

iii
 LIST OF FIGURES

1. Colpitts Oscillator. . . 3

2. General Oscillator Block Diagram . 5

3. Colpitts Oscillator Block Diagram. 6

4. Unstable Active Network. . . • . . 8

5. Active Network (h-parameter representation). 9

6. Active Network (y-parameter representation). . 10

7. Active Network (parallel network representation) 11

8. Output Network • . . . • . . • . • . . . • 13

9. Input Network. • 15

10. Design Program Flow Chart. . . 18

11. DC Equivalent Circuit. • • 20

12. Series and Parallel Models for a Real Inductor 23

13. Impedance of a Real Inductor at 100 MHz • • . • • • 24

14. 100 MHz Colpitts Oscillator. • . . . . • • . . 26

15. Matching Network 27

16. Circuit Testing Arrangement. . • . • • . • • . 29

17. Time Domain output . • • . . . . . • . 30

18. Frequency Domain Output (20 MHz per division). 30

19. Frequency Domain Output (200 KHz per division) 31

iv
 CHAPTER I

INTRODUCTION

Of the various types of feedback oscillators used at

radio frequencies, perhaps the most popular is the Colpitts
oscillator. It has certain advantages over the Hartley
oscillator (Sentz and Bartkowiak, 1968), in that with two
capacitors and a single inductor in its tank circuit (as
opposed to one capacitor and two inductors for the Hartley)
mutual inductance within the tank is not a problem;
furthermore, in some cases circuit size may be smaller for
the Colpitts oscillator (an important consideration at high
frequencies).

Description of Col2itts Oscillator Topology

The standard Colpitts oscillator implementation,
including a parallel tuning capacitor, is shown in Figure 1.
The tuned tank, or resonating portion of the circuit,
consists of two tapped capacitors in parallel with an
inductor. These elements are dominant in determining the
frequency of oscillation. Two capacitors are used (instead
 2

of one) so that a voltage divider can be used to feed a

portion of the output signal back into the input. This
positive feedback provides the drive necessary to sustain
oscillations. Analysis of this circuit is almost always
accomplished using a small-signal transistor model, such· as
hybrid-pi (or a modification thereof), and tuned-tank
network equations (Hayward, 1982). This technique relies on
estimates of transistor input and output impedances, since
they may have an appreciable effect on the frequency of
oscillation. This analysis can be complicated and, at best,
very approximate (Krauss, et al., 1980).

General Criteria for Oscillation

Many different sets of general criteria for oscillation
have been developed over the years, each set having its own
particular usefulness. The use of s-parameters to describe
circuit behavior at UHF and --microwave frequencies is
practically essential, but it can also be most helpful in
the VHF range. These parameters describe circuit response
in terms of incident, transmitted and reflected waves. Using
S-parameters, art oscillating system can be described very
simply as an unstable active network with input and output
matching networks (see Figure 2). The specific criteria for
oscillation corresponding to the diagram in · Figure 2 are
given by Vendelin as:
 3

v out

cl

R2
CA

I RFC c2

I
RE

Figure 1. Colpitts Oscillator.

 4

1) k < 1 ( 1)

2) s 11 •rG = 1 ( 2)

3) s 22 •rL = 1 (3)

where k (the stability factor) is given by:

2 2 2
1 - 18 111 - 18 221 + 18 11 8 22 - 8 12 8 211
k = (4)
2 IS12l IS21I
and

8 12 8 21rL
8
11 ' = 8 11 + (5)
1 - s22rL

8 12 8 21rG
8 ' = 8 22 + (6)
22
1 - s11rG

It can also be shown (Vendelin, 1982) that conditions 2) and

3) are redundant in that if one is satisfied, the other is
automatically satisfied.
 INPUT UNSTABLE OUTPUT

MATCHING ACTIVE MATCHING

NETWORK NETWORK NETWORK

Figure 2. General Oscillator Block Diagram.

 CHAPTER II

DESIGN PROCEDURE

In order to use the basic S-parameter criteria for

oscillation, the Colpitts os~illator topology shown in Fig.

l must be made to conform to the configuration illustrated

in Figure 2. Such a layout of the AC equivalent circuit is

shown in Figure 3.

L ct

Figure 3. Colpitts Oscillator Block Diagram.

6
 7

Thus, the combination of the transistor, feedback

capacitor and input resistor (R e ) ' becomes the
unstable active network. The parallel circuit containing
the inductor (L), tuning capacitor (Ct)' and load resistor
(RL) is considered to be the output matching circuit.
Finally, the second tank capacitor (C ) will be thought of
2
as the input matching element. Using this arrangement, the
requirements for oscillation will be satisfied in the
following way:

1) The overall S-parameters of the unstable active

network will be determined from the Common Emitter
transistor S-parameters and the values of c and Re.
1

2) For a given load and nominal value of ct, an

inductor will be selected such that the input match can be
accomplished with a single capacitor. (This condition is
necessary to satisfy the Colpitts topology.)

3) The appropriate resonating capacitor c2 will be

chosen.
 8

Determination of Active Network S-Parameters

Transistor S-parameters are normally supplied by the
manufacturer in the CE configuration. In order ·to combine
with these the effects of the input resistor and feedback
capacitor to obtain overall S-parameters for the active
network shown in Figure 4, several parameter conve·rsions
must be performed.

Figure 4. Unstable Active Network.

 9

The CE S-parameters must be changed to CE h-parameters

and then to CB h-parameters. These standard . transformations

(as well as the others mentioned in this paper) can be found
in Appendix A. The circuit can now be depicted as in Figure

5:

+ +

Figure 5. Active Network (h-parameter representation).

If Re is combined with h
and the result is
11
transformed to they-parameters (Van Valkenburg, 1974), the
arrangement of Figure 6 is obtained, where:

(7)

Y12 = -h12/(h11 · + Re) (8)

Y21 = h21/(hll + Re ) (9)

(hll + Re) (h22) - h12h21

Y22 = (10)
(hll + Re)
 10

+ +

Figure 6. Active Network (y-parameter representation).

Next, the real and imaginary parts of y 11 and y 22 are

separated:

Y11 = gll + j bll (11)

Y22 = g22 + j b22 (12)

This facilitates the regrouping of the elements of Figure 6

into two parallel networks as shown in Figure 7. The
y-parameters of the parallel networks can be combined to
give overall y-parameters for the active network (Van
Valkenburg, 1974).
Figure 7. Active Network (parallel network representation).
 12

The y-parameters of the Pi (lower) network are given by:

Y11 = j (bll + wc 1 ) (13)

Y22 = j (b22 + wc 1 ) (14)

Y12 = Y21 = -jwc (15)

1

Thus, the y-parameters of the active network are:

Y11t = g11 + j (b11 + wC 1 ) (16)

Y12t = Y12 - j wc (17)

1
Y21t = Y21 - j wc (18)
1
Y22t = g22 + j (b22 + wC 1 ) (19)

These y-parameters may now be converted directly to the

S-parameters of the active network (see Appendix A).

Determination of Output Network

Once the S-parameters have been determined it can be
confirmed that k < 1. {This should not be a problem in that
c1 is providing positive feedback within the network.) The
next step is to determine the inductor value that will allow
the resonating of the input with a single capacitor (C ).
2
In other words, an inductor must be chosen such that:
 13

s 11 ' = 1.00 {__::

18

where o0 < e < 180°

This restriction is necessary since for a single capacitor:

rG = 1.00 L'!_
where -1ao 0 < ~ < o0

A tuning capacitor is usually placed in parallel with

the load and the inductor; its effects must also be

considered (see Figure 8).

r
Figure 8. Output Network.
 14

Thus, given the active network S-parameters, and nominal

values for Ct and RL, an acceptable value for L must be
found such that 1s 11 •1 = 1. Since a computer will be used,
the most straight-forward method of accomplishing this is by
trial-and-error.
The total network impedance is:

ZL = RL + j xeff (20)
-1 (21}
where Xeff =(wet - (l/wL)]

For the purposes of iteration, this parallel reactance can

be treated as the reactance of a single inductor, Leff. The
computer begins with a low value for Leff (1 nH} and
determines 1s 11 •1, using Equation (5). It continues to
perform this calculation (each time increasing the value of
Leff by 1 nH} until a value for Leff is found that results
in 1s 11 •1 = 1. For each value of Leff tried, a
corresponding value for the actual inductor, L, is
calculated (for the given nominal value of Ct). For certain
combinations of Re' c1, RL' and ct with a particular
transistor, it is possible. that no reasonable value for L
exists that will result in 1s 11 •1 = 1. For this reason a
maximum allowable value of L should be input to prevent
unnecessary iterations •.
 15

Determination of Input Network

All that remains . to complete the design is the proper

selection of c . The necessary value of rG has been fixed

2
in the previous determination of s • since:
11

rG = (S
11
')-l (2)

Therefore, c is determined from:

2

( 2 2)

where {23)

1
Figure 9. Input Network •
 16

Although an inductor was selected to give 1s • I = 1, a

11
further restriction is required to ensure that the input
match can be achieved with a capacitor (rather than an
inductor). As stated previously, for a capacitor:

rG = 1.00 L_!

where -180° < ¢ < o0

Thus, if the angle of rG is:

0
o < ¢ ' < 180°

the input cannot resonate with a single capacitor and the

Colpitts topology cannot be satisfied.

Description ·· of Computer Program

A program to perform the calculations described in this
chapter was written in FORTRAN on the IBM-PC. A listing is
given in Appendix B. The program requires the following
inputs:
 17

Transistor S-Parameters (CE)

RL
R
e

cl
ct
Frequency
Maximum allowable value for L
The program first determines the overall s-parameters of the
active network (transistor, c 1 , and Re) and calculates k.
At that point, Re and c 1 can be modified or the program can
proceed to determine values for L and c 2 • If no acceptable
values exist, the output so states. A flow chart for the
program is given in Figure 10.
 18

Input S-parameters,
f,RL,R
e
, c 1 ,c t ,L
max

Compute overall
S-parameters of
Active Network

Display
S-parameters &
k

Yes

Calculate IS
1
il
No

Increment
Leff

"Cannot "No inductor

resonate inpu found within
with a single specified
capacitor." range."

Calculate _ c2

Display c2

Figure 10. Design Program Flow Chart.

 CHAPTER 3

DESIGN EXAMPLE

To confirm the validity of the design technique

described in Chapter II, a 100 MHz Colpitts Oscillator was
built and tested. The transistor used was the HP HXTR-3101.
The CE S-Parameters supplied by the manufacturer (with
biasing conditions Ic = 10 mA and Vce = 10 V) are:

8
11 =
0.708 L-500
0.015 60°
8
12 = L
24.547 L149°
21 =
8

0.912 L-200
8
22 =
To confirm these values, the S-Parameters were measured in
the laboratory using the same biasing conditions. The
results were:
0.724 L-140
8
11 =
0.0002 L 75°
8
12 =
21.380 L154°
8
21 =
0.933 L-100
8
22 =

19
 20

which agreed well with those supplied by the manufacturer.

Design of Biasing Circuit

The DC equivalent circuit is shown in Figure 11.
+ vcc

Figure 11. DC Equivalent Circuit.

To satisfy the biasing conditions mentioned in the previous

section, the following standard values were chosen:

Rl = 4.7 Kn

R2 = 5.6 Kn

Re = 100 n
RE = 1 KQ
Vee = 21 v
 21

The value of Re' which is a rather insignificant part

of the DC circuit, plays a very important role in the AC
circuit, in that it is used to linearize the input
resistance of the active network and to limit the amount of
positive feedback (Sentz and Bartkowiak, 1968).

Design of Tank Circuit

In order to make the design of this example oscillator
as general as possible, a single resistor was used as the
load. It can represent the input impedance of an output
buffer stage, or it can simply be matched to whatever
network it is driving (Vendelin, 19_82). It should be noted
here that oscillation becomes less likely as the load
resistance decreases; a low parallel resistance tends to
drastically lower the Q of the tank (Krauss, et al., 1980).
The computer program was used to determine the tank
element values. In addition to the value for Re mentioned
above, the following were selected for input to the program:

c1 = 15 pF

RL = 4. 7KO to 600 n
Ct = 15 pF
 22

The results of the program were:

8 = 0.299 L-52.6°
11
8 = 0.541 L 57.4°
12
8 = 0.756 L 11.4°
21
8 22 = 0.847 L-33.o 0

k = 0.75

L = 83 to 100 nH

The program was run several times, varying the load

resistance value from 4. 7Kn down to 600 n. This was to
allow for the various resistances which are in fact parallel
to the load resistor, effectively reducing its value. The
primary parallel resistance comes from the equivalent
parallel resistance of the coil.
An inductor with a value in the range shown above was
wound and its impedance measured on the 8-Parameter test set
(see Figure 13). The inductor had a relatively low Q, given
by:

= X /R = 13.7 (24)
s s

where xs is the equivalent series reactance and Rs is the

equivalent series resistance. If the series equivalent
 23

circuit values are changed to parallel equivalent circuit

values, the parallel equivalent resistance is given by:

2
R
pe = (1 + Qs ) = 886 Q (25)

where Rpe is in parallel with the load resistor. (It should

be noted that for Qs > 10, the values for Ls and are
essentially the same.) Thus, the actual effective load
resistance of the output network (Krauss, et al., 1980) is:

(26)

Using this value as the load resistor, the program yields:

L = 93 nH
C2 = 82 pF

R
s
xs
Qs= R
s xpe R
pe
= (1 + Qs ) 2

xs

Figure 12. Series and Parallel Models for a Real Inductor.

 24

FULL SCALE 1.0000 MARKER 100 000 000.000Hz

PHASE REF 0.0deg Z RE(S11) 82.226E-3
REF POSN 0.0deg Z IM(S11) 1.1265

START 80 000 000.000Hz STOP 120 000 000.000Hz

AMPTD 15.0dBm

Figure 13. Impedance of a Real Inductor at 100 MHz.

 25

Final Circuit Implementation

The final circuit was built on two-sided copper-clad
board, one side being used as a ground plane. In addition
to the components previously mentioned, additional
components were used. The power supply was decoupled using
a parallel capacitor and series RF choke. The bypass
capacitors were realized by two parallel capacitors (to
lower effective series inductance of the bypass capacitors
and thus reduce the chance of spurious high frequency
oscillations). The size of the board was made as small as
possible (2 in. x 2 in.) to minimize stray capacitances;
leads were made as short as possible to lower inductances
(DeMaw, 1982). Variable capacitors were used in the tank
circuit to facilitate adjustments and fine tuning. A
complete circuit diagram is shown in Figure 14.

Matching Network
In order to measure the output at 100 MHz on standard
50 Ohm test equipment, a matching network was implemented.
The 4.7Kn output resistor was matched to a standard RF
characteristic impedance of 50 Ohms using a tunable

T-network as shown in Figure 15.

 26

RFC
v
cc

+ 21 v

L 90 nH
4. 7 KO

v out

cl 15 pF
R
R2 5.6 KO e
CA
-~ 4. 7 KO

I pF

c2

I3 pF

RE

Figure 14. 100 MHz Colpitts Oscillator.

 27

240 nH 15 pF
_,_...._.__..___"-""'r--..~~----~-+-------_. ~......,_~---o

46 nH
pF

Figure 15. Matching Network •

The nominal values were selected using the Smith Chart, and
the circuit was fine tuned using the s-parameter test set to
guarantee a good match at 100 MHz. The addition of this
parallel network to the load network results in yet another

change in Reff:

(27)

where Rmatch is the 50 n impedance as seen through the

matching network. With Rmatch = 4.7KO:
 28

Reff = 643 n
L = 97 nH

c2 = 55 pF

Circuit Testing

Th~ actual circuit values which resulted in oscillation

at the desired frequency are given below:

RL = 4. 7Kn
R
e
= 100 n
ct -- 17 pF

cl = 15 pF

c2 = 33 pF

L = 90 nH

The output was connected through the matching network

to the oscilloscope. The resulting time domain waveform can

be seen in Figure 17. The peak-to-peak voltage is 1.6 V,

corresponding to a power being delivered to a 50 n load of:

2
vrms
Pac = = 6.4 mW= 8.1 dBm (28}
Ro

Recalling the DC power required by the transistor at the

given biasing conditions

 29

= Vce I c = (10 V) (.01 A) = 100 mW (29)

it is noted that the oscillator efficiency is

p 6.4
Eff. =
-= ac
- = 6.4% ( 3 0)

It has been shown (Krauss, et al., 1980) that the maximum

theoretical efficiency for this oscillator configuration is

25%. Figure 16 illustrates the testing set-up.

100 MHz Matchi,ng 50 Ohm

Colpitts Network Test
Oscillator Equipment

Figure 16. Circuit Testing Arrangement.

 30

The frequency domain characteristics were measured on the

spectrum analyzer. Figure 18 illustrates the 6 dBm . signal
at 100 MHz and Figure 19, using 200 kHz per division,
illustrates the phase noise. Harmonics went virtually
undetected on the test equipment, since the matching network
also effectively serves as a bandpass filter (Bowick, 1982).
The maximum tuning range afforded by adjusting ct was about
1 MHz.

Figure 17. Time Domain Output.

 31

Figure 18 . Frequency Domain Output (20 MHz per division).

Figure 19. Frequency Domain Output (200 KHz per division).

 CHAPTER IV

RESULTS AND CONCLUSIONS

The performance of the oscillator described in the

previous chapter was very close to that predicted by the
design program. The most significant discrepancy arose in
the value of c 2 , which differed from the predicted value by
40%. Upon further investigation using the program, it was
found that a somewhat lower value for RL ( 580 n ) resulted
in the measured value of c
( 33 pF), although this also
2
raised the required value for L to 104 nH.
Several terms have been neglected in the analysis
presented in this paper; no one term by itself is very
significant, yet combined they may effect the results
substantially. Specifically, consideration should be given
to the resistance of the tank capacitors as well as the
reactance of the load resistor. The effects of these alone
could easily have caused the difference between the

predicted and actual values (Bewick, 1982).

32
 33

It should be noted in conclusion that a feedback system

is not necessarily required for all VHF oscillators. At
higher frequencies, feedback within the transistor itself
can make it unstable (particularly in the CB configuration)
without an external positive feedback path. The first part
of the design program can be used to assist in oscillator
design using topologies with no feedback simply by entering
R
e
= = o. The program then computes the standard CB
S-parameters and the stability factor, k. If k < 1, no
feedback is required, and a simple oscillator may be
designed using the CB S-parameters along with the general
criteria mentioned in Chapter I.
 APPENDIX A

NETWORK PARAMETER TRANSFORMATIONS

 35

I . 8-parameters (CE) to h-parameters (CE) (Vendelin, 1982)

(1 + 811> ~1 + 8 22> - 8 12 8 21
hll =
(1 - 811> (1 + 8 22> + 8 12 8 21

2812
h12 =
(1 - 811> (1 + 8 22> + 8 12 8 21

-2821
h21 =
(1 - 811> (1 + 8 22> + 8 12 8 21

(1 - s22> (1 - 8 11> - 8 12 8 21
h22 =
(1 - s11> (1 + 8 22> + 8 12 8 21

II. h-parameters (CE) to h-parameters (CB) (Casasent, 1973)

h11ce
hllcb =
(h21ce + 1)

h12cb = h11cbh22cb(h21ce + 1) - h 12ce

-h 21ce
h21cb =
(h21ce + 1)

h22ce
h22cb =
(h21.c e + 1)
 36

III. h-parameters to y-parameters (Van Valkenburg, 197 4.)

1
Y11 = -
hll

-h12
Y12 = -
hll

h21
Y21 =-
hll

hllh22 - h12h21
=
Y22
hll

IV. y-parameters to S-parameters (Vendelin, 1982)

(1 - Y11> {1 + Y22> + Y12Y21

8
11 = (1
+ Y11> {l + Y22> + Y12Y21

- 2 Y12
8
12 =
{l + Y11> {l + Y22> - Y12Y21

- 2 Y21
8
21 =
(1 + Y11> {l + Y22> - Y12Y21

(1 + y ·11> (1 - Y22> + Y12Y21

8 =
22 (1 + Y11> (1 + Y22> - Y12Y21 ·

NOTE: All two-port parameter transformations are normalized

to Z0 = 1.
 APPENDIX B

LISTING OF DESIGN PROGRAM

 38

COMPLEX s11.s12.s:1,s:2
COMPLEX HE11,HE12,HE21,HE22
COMPLEX HB11,HB12,HB21,HB22
COMPLEX YT11,YT12,YT21,YT22
COMPLEX Y011,Y012,Y021,Y022
COMPLEX so11.so12.so21,so22
COMPLEX DENH,DENS,DENY,DELTH
COMPLEX YC,ZL,ZO,GAML
COMPLEX XL,ZLOAD,SPRIM,SPRIG
COMPLEX GAMG,ZG
REAL MAG,K,LMAX,LEFF

F'I=3. 1415926
WRITE<*,*) ' ENTER MAGCNON-DB> AND PHASE(DEG> OF CE S-PARAMS'
WRITE<*~* )' INPUT 511 MAG AND PHASE'
READ(*,*lS11M,S11D
WRITE<*,*)' INPUT 512 NAG AND PHASE'
READ(*,*lS12M,S12D
WRITE<*,*) I INPUT S21 MAG AND PHASE'
READC*,*>S21M,S21D
WRITE<*,*) ' INPUT 522 MAG AND PHASE'
READC*,*)S22M,S22D
WRITE<*,*)' INPUT FHEO (IN MHz)'
READC*,*>FREQ
WPJTE(*,*) 'INPUT MAXIMUM ALLOWABLE 'JALUE FOR L <IN nH)'
READ<*,*}LMAX
WHITE«~,*)' INPUT NOMINAL VALUE FOR TUNING CAPACITOR <IN pF) '
HEAD<-~,*}CT
WRITE(*,*)' INPUT LOAD RESISTANCE (IN Ohms)'
READ<*,*}HL
2 WRITE<*,*)' INPUT F:e <IN Ohms)'
READ<*,*iRE
WRITE<*,*>' INPUT Cl <IN pF>'
HEAD<*,*>C1

S11P=S11D*PI/180.
S12P=S12D*PI/180.
S21P=S21D*PI/180.
S22P=S22D*PI/180.

C CALCULATE CE S-PARAMETERS IN STANDARD FORM <TRANSISTOR ONLY>

S11=S11M*CMPLX(CQS(S11P) ,SIN<SllP))
S12=S12M*CMPLX(C05(512P) ,SINCS12F'))
 39

S..'.'.1=S21f'ltCMPLX !COS(S21P) ,Srn(S21Pl)

S22=S::2M+.CMPLX (C0S(S22P> ,SIN(S22f'))

L TRANSFORM TO CE H-PARAMETERS (TRANSISTOR ONLY)

DENS=(1.0-S11)*ll.O+S22)+(512*S21)
HE11=((1.0+S11)*(1.0+S22J-(512*S211 I/DENS
HE12=(2.0*S12)/DENS
HE21=<-2.0*S211/DENS
HE22= ( ( 1. 0-S l 1) * ( 1. O-S22)-(Sl2*S21)) /DENS

C TRANSFORM TO CB H-PARAMETERS (TRANSISTOR ONLYJ

DENH=HE21+1.0
HB11=HE11/DEMH
HB21=-HE21/DENH
HB22=HE22/DENH
HB12=HB11*HB22*DENH-HE12
C TRANSFORM TO CB Y-PARAMETERS <TRANSISTOR AND RE)
RN=RE / 50.0
DELTH=(HB11+RN)*(HB22)-(HB12*HB21)
YT1t=l./(HB11+RN)
YT12=-HB121<HB11+RN>
YT21=HB21/(H811+RN)
YT22=DELTH/CHB11+RN)

C CALCULATE Y-PARAMETERS ASSOCIATED WITH Cl & RE AND COMBINE

C WITH Y-PARAMETERS OF CB TRANSISTOR

G11=REAL<YT11>
B11=AIMAGCYT11)
G22=REAL<YT22)
B22=A I MAG ( YT22)
Y=2.0*PI*FREQtC1*(50.0E-06)
YC=CMPLX<0.0,Y)

Y011=CMPLXCG11,B11+Y)
Y012=Yi12-YC
Y021=YT21-YC
Y022=CMPLX<G22,B22+Y)

C CONVERT OVERALL Y-PARAMETERS TO OVERALL CB S-PARAMETERS

DENY=Cl.O+Y011)*(1.0+Y022l-(Y012•Y021)
S011=CC1.0-Y011)f(1.0+Y022)+(Y012*Y021J)/((1.0+Y011>*<1.0+Y022)+
:t <Y012*Y021))
S012=(-2.0*Y012)/DENY
 40

0 2 1 = 1• - 2 • 0 * YCl '::'. 1 I i DE N Y
'..3
*
so ·2 2= ( ~ 1. O+Y 011.) !. 1. (l-'{1]22) + (YQ12*Y021)) /DENY

l~RITE ( *.~· ), '

!1JRITE(*,*i '
WF:ITE(*,*) ' F:e= ', F:E, ' Ohms '
WPITEU·,*) .

CALL STABILITYCS011,S012,S021,S022,K)
WRITE(lf-,*) ' Cl= ',Cl,' pF'
~JRITE<*,*)'
Wf;:ITE<*,*), K= ',K
WRITE(*,*)'

CALL CTMAP<S011,S011M,S0110)
CALL CTMAP\S012,S012M,S012D)
CALL CTMAP<S021,S021M.S021Dl
CALL CTMAP CS022,S022M,S022D)

l~R I TE ( *. *) . OVERALL S-PAPAMETERS (INCLUDING RE t-t: C 1) '

,
trnJTE(*~*) . ', 'MAG', , '
DEG'
l~ F: IT E ( * , * > ' Slt= ',S011M,S011D '
WRITE(*,*)' S12= ',S012M,S012D
WRITE(*,*) ' S21= ', S021M,S021D
l~HITE(*,*)' S22= ' ,S022M,S022D
WF-:ITE(*,*) ' TR.Y AN0 TH ER C1 0 R HE? ( YES= 1 , N0 = 2 ) '
READ<*,*)8
IF <B.EQ.1.0) GOTO 2

C DETERMINE A VALUE FOR L THAT WILL ALLOW A SINGLE CAPACITOR

C TO RESONATE THE INPUT

L=O.O
LEFF=l.O
Z0=(50.0,0.0J
~ XX=2.011-PI*FREQ*LEFF*<1.0E-03)
iL=CMPLX<O.O,XX)
ZLOAD=CMPLXCRL,0.0)

ZL=(ZLOAD*XL)/(ZLOAD+XL)

GAML=<ZL-ZO)/(ZL+ZO)

SPRIM=S011+(CS012*S021*GANL>l<1.0-<S022*GAML>>>
S=CABS<SPRIM)
IF (S.GE.1.0) GOTO 4
YEFF=!.O/XX
YCT=2.0*PI1FREQ*CT1<1.0E-06)
Y IND~YEFF+YCT
 41

ZIND=l.U i YIND
L=ZIND / (2.UH'I*FPEQ*(l.OE-03)}
IF (~.GT.LMAX) GOTO 44
LE F F =L E F F + 1 • i)
GOTO 3

44 WRITE(*,*) 'NO INDUCTOR FOUND WITHIN SPECIFIED RANGE'

GOTO 40
45 WRITE<*,*1 'CANNOT RESONATE INPUT ·WITH A CAPACITOR'
GOTO 40

4 WRITE<*,*>
l~ R I TE ( * , * }
~JRITE«~,*)
WRITE(*,*> lJ ITH RL=
1
, RL, Ohms '
WRITE\lf-,*) pc: , ,F:E, Ohms'
tff I TE < * , * ) Cl= 'c 1 ' pF'
WRITE<*,*) Ct= ,CT, pf'
~JRITEU·,*) v-
·.. -
I .'
' I'·.

CALL CTMAP(SPRIM.SPRM,SPRDl
l~ R IT E ( * , *) '
WR IT E o:, *) 'IN ORDER TO RESONATE INPUT AND OUTFlUT'
WRITEO·,*)'
WRITE<*,*)' L= ',L,' nH'
WRITE<*,*)'
WRITE<*,*i'
CALL CTMAP<GAML,GAMM,GAMD>
WRITE<*,*) 'GAMMA OF LOAD= ',GAMM,' AT ',GAMD,' DEG.'
WRITE(*,*> '
tiJRITE(*,*) '511 PRIME= , ,SPRM,' AT ',SPRD,' DEG. I

C DETERMINE A VALUE FOR C2 THAT WILL RESONATE THE INPUT

DEG=ATAN2<AIMAG\SPRIM> ,REAL<SPRIM) )*180.0/PI

GDEG=-DEG
IF <GDEG.GT.O.O> GOTO 45
GAMGP=GDEG*PI/180.0
GAt1G=CMPL X (COS ( GAMGP > •SIN< GAMGP))
ZG=ZO*<<l.O+GAMG)/(1.0-GAMG))
XC=CABS(ZG)
C2=1.0/C2.0*PI*FREQ*XC*<l.OE-06))
~~RITE<*,*}, C2= ',C2,' pF'
torn IT E ( ~ , * ) '
CALL CTMAP<GAMG,GAIM,GAID>
WRITE<*,*) 'GAMMA OF IMPUT= . ,GAIN,, AT I ,GAID, I DEG •.
~JH IT E < * , * > '
SPRIG=S022+((5012*S021*GAMG)/(1.0-<S011*GAMG>>>
CALL CTHAP<SPRIG,SPGM,SPGD)
WRITE<*,*)'522 PRIME= ',SPGM,' AT ',SPGD,' DEG.'
WRITE<*,*>'
 42

40 ;i.J F I TE ( * , * ) , TRY AN0 THE R RE , E L 0 R C 1 ? ( YE S =i , 1'~ O=2 ) '

FEAD(l,it)A
IF i.A. E!J.1.0) GOTO 1

STOP

EMO

C THIS SUBROUTINE CONVERTS A COMPLEX NUMBER INTO MAGNITUDE

L AND DEGREES

SUBROUTINE CTMAPCCOMP1MAG,DEG)

COMPLEX COMP
REAL MAG

PI=3.1415926
;~ XR=RE AL ( C0 MP )
YYI=AIMAG(COMP)
MAG=SQRT<XXR**2+YYI**2)
DEG=ATAN2<AIMAG<COMPJ , REAL<COMP))*180.0/PI

PETURN
END

C THIS SUBROUTINE CALCULATES THE STABILITY FACTOR, K

SUBROUTINE SfABILITYCS011,S012,S021,S022,K>

COMPLEX so11,so12,so21,so22
HEAL K

B=CABS<S011)
C=CABSCS022)
D=CABS<S011*S022-S012*S021>
K=,1.0-B**2-C**2+D**2)/(2.0*CABS CS012)
~ *CABS<S021)J

RETURN
END
 APPENDIX C

EXAMPLE RUN OF DESIGN PROGRAM

 44

A: \ .>
A: \ >
A: \ >
A:\ >tkj
ENTER MAG <NON-OBJ AND PHASE(DEG) OF CE S-PARAMS
INPUT Sll MAG AND PHASE
(J • 724,-14.
INPUT S12 i'lAG AND PHASE
0.0002,75.
INPUT S21 i1AG AND PHASE
21. .3 8, 154.
INPUT ""~..::..a::.
.-. ,-, MAG AND PHASE
0.933,-10.
INPUT FREQ ( IN MHz )
100.
INPUT MAXIMUM ALLOWABLE VALUE FOR L ( IN nH)
12 (i.
I NPUT NOM INAL VALUE FOR TUNING CAPACITOR CIN pF>
f ~
L ..J a

INPUT LOAD RESISTANCE <IN Ohms)

650.
INPUT Re i.W Ohms)
100 .
INPUT Cl (IN pF)
c::
! __, .
.1.

Re= 100.0000000 Ohms

[: 1 = 15.0000000 pF

K= 7.476355E-001

OVERALLS-PARAMETERS (INCLUDING RE & Cl)

MAG DEG
s 11 = 2.989655E-001 -52.6134100
S12= 5. 40970BE-001 57. 350170(1
:321 = 7.564967E-001 11.4056300
522= 8.468690E-001 -33.0495300
TRY ANOTHER Cl OR RE? CYES=1,N0=2)
 45

~1J1 TH RL= 65 o. 0000000 Ohms

F'. e= 100. 0000000 Ohms
Cl= 15. 0000000 pF
Ct= 15. 0000000 pF
f::"= I . 476355E-001

IN fJRDER TO ~'.ESONA IE INPUT ANO OUTPUT

I-
L...- nH

8.716151E-001 AT 38.5130300 DEG.

311 PP!ME= 1.0027850 AT 120.5401000 DEG.

.-.-:-
!-· ~- 55. 73-79500 pF

GA ri MA 0 F I NP U T = 1.0000000 AT -120.5401000 DEG.

S22 PHIME= 1. 1479440 AT -38.5225600 DEG.

TRY ANOTHER RE, RL OR Cl? <YES=1,N0=2)

 REFERENCES

Bewick, Chris. RF Circuit Design. Indianapolis: Howard W.

Sams and Co., 1982.
Casasent, David. Electronic Circuits. New York: Quantum
Publishers, Inc., 1973.
DeMaw, Doug. Practical RF Design Manual. Englewood Cliffs:
Prentice-Hall, 1982
Hayward, W. H. Introduction to Radio Freauencv Design.
Englewood Cliffs: Prentice-Hall, 1982.
Hewlett-Packard, Inc. Diode and Transistor Designers's
Catalog. Palo Alto, 1984.
HXTR-3101 "Low Cost General Purpose Transistor"
Krauss, Herbert L.; Bostian, Charles W.; and Raab, Fredrick
H. Solid State Radio Engineering. New York: John
Wiley and Sons, 1980.
Sentz, R. G.; Bartkowiak, Robert A. Feedback Amplifiers
and Oscillators. New York: Holt, Rinehart and
Winston, 1968.
Van Valkenburg, M. E. Network Analysis. Englewood Cliffs:
Prentice-Hall, 1974.
Vendelin, George D. Design of Amplifiers and Oscillators by
the S-Parameter Method. New York: John Wiley and
Sons, 1982.

46