development of an automatic system to monitor the
TRANSCRIPT
Univers
ityof
Cape T
own
DEVELOPMENT OF AN AUTOMATIC SYSTEM TO MONITOR THE PERFORMANCE OF A
DENSE MEDIUM (MINERAL) SEPARATION PROCESS
PREPARED FOR: The department of Electrical and Electronics Engineering at the University of Cape Town.
PREPARED BY: Hr P.A. Alberts
B.Sc Eng (Elec) UCT
Submitted to the University the requirements for the Engineering.
of Cape Town in partial fulfilment of degree of Master of Science in
September 1989
i
i Town has been g\ven. The University of Cape this thesis In whole the right to reproduhceis held by the author. or In part. copyrlg t
I
PREPARED FOR: of
Engineering at the University Town.
PREPARED BY: Hr P.A. Alberts
B.Se Eng ( ) UCT
Submitted to the University of Cape Town partial fulfilment
for the of Kaster of
Engineering.
1989
i
The copyright of this thesis vests in the author. No quotation from it or information derived from it is to be published without full acknowledgement of the source. The thesis is to be used for private study or non-commercial research purposes only.
Published by the University of Cape Town (UCT) in terms of the non-exclusive license granted to UCT by the author.
Univers
ity of
Cap
e Tow
n
Univers
ityo
ape Tow
n
DECLARATION
I declare that this thesis is my own, unaided work. It is being
submitted for the degree of Master of Science in Engineering in
the University of Cape Town. It has not been submitted for any
degree or examination in any other University.
/Ukct; 29th day of September 1989
DECLARATION
I declare that this thesis is my own, unaided work. It is being
submitted for the degree of Haster of Science in Engineering in
the University of Cape Town. It has not been submitted for any
degree or examination in any other University.
I~ 29 th day of September 1989
Univers
ityo
ape Tow
n
ACKNOWLEDGEMENTS
The author is indebted to Prof B.J .. Downing for his supervision,
assistance and encouragement during this project, and to Dr. o. Salter (De Beers Diamond Research Laboratory, Johannesburg) for his enlightened technical input.
I would also like to thank Hr N. Wright, Mr P. Daniels and Hr A.
Vinnecombe, for their help in the construction of some of the
components used in this thesis.
ii
The author indebted to Prof B. J .. ass and Salter (De Beers Diamond Research his enlightened input.
his supervision, , and to Dr. D. Johannesburg)
I would like to thank HI N. Wright, HI P. Daniels and HI A. Vinnecombe, for their help the construction of some of the components
Univers
ityo
ape Tow
n
ABSTRACT
Dense Medium Separation (DMS) is a process used extensively in the minerals
processing industry to separate dense from less dense material in a dense fluid.
It may be considered to be a simple "sink-float" separation process. DMS is used
on a large scale in South African coal, iron ore and diamond operations.
There are, however, no commercially available systems that can determine the
separation efficiency of a DMS process on-line. This presents severe problems
to those operating DMS processes. The present study attempts to provide a
measurement technique for on-line application.
The technique is based on the use of density tracers. These are colour coded,
uniformly sized plastic beads of known density which have been used to monitor
DMS processes for many years. Their distribution between the "sink" and "float"
products, upon being discharged from the DMS process provides a measu_re of
its separation efficiency. Unfortunately, they can only be used when a DMS
process is not treating any material and, as such, they only provide an
approximate measure of the separation efficiency.
Density tracers are fed into, and collected from, DMS processes manually. An
automatic means of feeding and detecting these tracers on-line would therefore
satisfy the needs of OMS operators. (The feeding mechanism was not considered
as part of this study since it is a relatively simple piece of engineering.)
Active, or powered, tracers were discounted as being impractical. Metal doped
tracers with an associated metal detection system were ruled out. because of false
signals from the ubiquitous tramp metal in OMS feed streams. Passive microwave
corner reflectors proved to be equally unreliable because of similiar-strength
reflections produced by the OMS products.
Ill
Dense Medium Separation (OMS) is a process used extens
processing industry to separate dense from less dense
In the minerals
in a dense fluid.
It may be considered to be a simple "sink-float" separation process. OMS is used
on a scale in South African coal, iron ore and diamond operations.
There however, no commercially available systems that can determine the
separation efficiency of a DMS process on-line. This presents severe problems
to those operating DMS processes. The present study attempts to provide a
measurement technique for on-l application.
technique is based on the use of density tracers. These are colour coded,
uniformly sized plastic beads of known density which have been used to monitor
. DMS processes for many years. Their distribution between the "sink" and "float"
products, upon bei discharged from the OMS process provides a measure of
separation ciency. Unfortunately, they can only be when a OMS
process is not any material and, as such, only provide an
approximate measure of separation efficiency.
Density tracers are fed into, and collected from, OMS manually.
automatic means of feeding and these tracers on-line would refore
satisfy the n of OMS operators. (The feedi anism was not considered
as part of this study since it is a re ly of )
Active, or powered, tracers were discounted as impractical. Metal doped
with an associated metal detection system were ru out, because false
signals from the ubiquitous tramp metal in OMS streams.
corner reflectors proved to be equally unreliable because similiar-strength
reflections produ by the OMS products.
III
Univers
ityo
ape Tow
n
The solution to the tracer problem was for the tracers to receive
an applied signal and to re-radiate passively a modified signal.
Reception of the modified signal therefore indicates the presence of the tracer.
Frequency mixing and multiplication techniques were considered for the tracer application. Both rely on the fact that a mixer
diode will produce harmonics of an incident signal, if the transmitted signal is strong enough. Frequency mixing was
discounted, however, because of the need for complex, expensive, high frequency crystal locked oscillators.
The system chosen for development uses frequency multiplication. Each tracer contains a dual-frequency cross-dipole antenna with a
suitable Schottky mixer diode and RF choke. The applied frequency
is received by one dipole antenna causing the mixer diode to
produce harmonics of this fundamental frequency. The tracer•s
second dipole antenna is matched, however, to radiate only the
second harmonic.
Initial tests were conducted using a 500 MHz-1 GHz system. Following the success of these tests, a 3 GHz-6 GHz system was developed. The development, design and evaluation of this system
is described in detail.
iv
an to the tracer problem was
signal and to re-radiate of the signal
tracers to receive a
indicates signal.
of tracer.
mixing and multiplication tracer Both
of an signal strong
techniques were on the fact a mixer
the was
discounted, however, because of the complex, II
high frequency locked
chosen development uses frequency multiplication. tracer contains a dual-frequency antenna with a
and RF by one antenna to
harmonics dipole antenna harmonic.
tests were the success
matched,
conducted using ,
a a 3
frequency. tracer's only the , to
500 MHz-l GHz system. GHz-6 GHz was
. The development, design and evaluation system in
iv
Univers
ityo
ape Tow
n
LIST OF TABLES
1. Characteristics of transmit and receive dipoles
2. Characteristics of cross dipole antenna
3. Dimensions of 1/2 GHz cross dipole antenna
4. Loss characteristics for 1/2 GHz cross dipole
5. Characteristics of 3 GHz open circuit stub
6. Characteristics of 3 GHz notch filter
7. Predicted and actual characteristics of
6 GHz filter
8. Characteristics of 6 GHz power splitter
9. Characteristics of a simple three port device
27
28
35
37
67
72
79
85
91
10. Characteristics for three port with 70.70 lines 92
11. Characteristics of designed splitter 96
12. Coupling and reflection coefficients for coupler 110
13. Measured coupling coefficients for coupler
14. Comparing actual with predicted coupling values
15. Effects of parasitic etching on coupling
coefficients
16. Characteristics of varactor tuned oscillators
17. Characteristics of 3 GHz amplifier
v
.>
111
111
113
115
121
1. of . dipoles
2 .. Characteristics of cross dipole antenna
3. 1/2 GHz cross dipole antenna
4. Loss 1/2 GHz cross
5. Characteristics of 3 GHz open circuit stub
6. Characteristics of 3 GHz notch filter
7. Predicted and
6 GHz filter
8. Characteristics of 6 GHz power spl
9. of a simple
10. 70.7S! lines
11. Characteristics designed splitter
12. Coupling and
27 28
67
72 79
85
91
92 96
110
13. for coupler 111
14. Comparing actual with predicted coupling values III 15. Ef on coupl 113
16. Characteristics of varactor tuned oscillators
17. of 3 GHz
v
.)
115
121
Univers
ityo
ape Tow
n
LIST OF FIGURES .
1. Cutaway view of a cyclone 1 2(a). Dense medium cyclone 2 2(b). Dismantled cyclone (indicating over and underflow 3
orifices) 3. Schematic of the complete DHS system 4
4. Photograph of an underflow shaking screen 5 5. A typical Tromp curve 6 6. Metal detection test rig 11 7. Coil positions under shaking screen 12
8. Microwave corner reflector 13 9. Maximum area of reflection from reflector 14 10. Test rig measuring reflections from corner 15
reflector 11. Mixer diode mounting 18
12. The spiral antenna 18 13. Frequency mixing/density tracer counter system 20 14. Electrical length of 500 MHz dipole 22 15. The cross dipole antenna 24 16(a). Operation of fundamental mode element 25 16(b). Operation of second harmonic mode element 25 17(a). Return loss for transmit dipole 26 17(b). Return loss for receive dipole 26 18. Return loss for 500 MHz/1 GHz cross dipole antenna 27 19. Choke mounting on cross dipole antenna 28 20. Return loss of cross dipole with 12 turn choke 29 21. 500 MHz coaxial cavity resonator 30 22. Return and insertion losses for the 30
cavity resonator 23. Insertion loss of resonator 24. Frequency multiplication test setup 25. 1/2 GHz cross dipole antenna 26. Test rig to measure antenna match
vi
31 32 36 36
1. Cutaway view of a 2(a). Dense medium cyclone 2(b). Dismantled over
) 3. Schematic of the complete DMS system 4. of an underflow screen 5. A typical Tromp curve 6. detection test rig 1. screen
8. corner 9. Maximum area of reflection from reflector 10. Test corner
reflector
underflow
1
2 3
4
5
6 11
14 15
11. mounting 18 12. The spiral antenna 18 13. Frequency mixing/density tracer counter 20 14. 500 MHz 22 15. The cross dipole antenna 24
(a). Operation of fundamental mode 16(b). of second harmonic mode element 25 11(a). Return loss for transmit dipole 26 11(b). Return loss receive dipole 26 18. Return loss for 500 MHz/I GHz cross dipole antenna 21
. Choke mounting on cross antenna 28 20. Return loss of cross dipole ~ith 12 turn choke 29 21. 500 MHz coaxial resonator 30 22. Return losses the
resonator 23. of resonator 24. Frequency multiplication test setup
. 1/2 GHz cross antenna 26. Test rig to measure antenna match
30
31 32 36 36
Univers
ityo
ape Tow
n
27. Return loss for the 1/2 GHz cross dipole
28. 1-4 GHz anechoic chamber setup 29. Cross dipole antenna showing coax feed cable
30. Cross dipole antenna with reflective plate 31. Positive print of 3/6 GHz cross dipole antenna 32(a). Return loss of cross dipole antenna centered
on 3 GHz 32(b). Return loss of cross dipole antenna centered
on 6 GHz
37
39
39
40 43 43
43
33. Choke reactance test bed 45
34(a). Return loss of 5 turn choke over 100 MHz-1 GHz 45
band
34(b). Return loss of 5 turn choke over 3 GHz-6 GHz band 46 35. Section of horn antenna with wave front 49 36. Phase error of antenna along the slant length 37. Dimensions of 3 GHz circular flange
38. Side plate of 3 GHz pyramid horn
39. 3 GHz waveguide/coaxial transformer
40. Return loss for 3 GHz horn antenna 41. Dimensions of 6 GHz circular fiange 42. Side plates for the 6 GHz pyramidal horn
43. 6 GHz horn & waveguide-coaxial transformer 44(a)&(b). Return loss for 6 GHz horn & waveguide
to coaxial transformer 45. Block diagram of frequency multiplication system 46. Full frequency multiplication system
'
47\ /density tracer counter unit
Sweep and lock feedback loop I
481. Block diagram of transmitter unit 4~. Block diagram of the receiver unit
50. Open circuit stub at 3 GHz 51. Open circuit stub characteristics 52. Simulated 3 GHz notch filter characteristics 53. 3 GHz notch filter 54. Positive print of 3 GHz notch filter
vii
49 50
50
51
53 54
55 57 57
59 61
62 63 64 67 68 70 69 72
27. Return loss for the 1/2 GHz cross dipole 28. 1-4 GHz anechoic chamber setup 29. Cross antenna showing coax 30. Cross antenna with plate 31. Positive print of 3/6 GHz cross dipole antenna 32(a). Return of cross antenna
on 3 GHz (b). Return loss of cross dipole antenna
on 6 GHz
37 39 39 40 43 43
43
33. Choke reactance test 45 34(a). Return loss of 5 turn choke over 100 MHz-1 GHz 45
band 34(b). Return 5 turn choke over 3 GHz 46 35. Section of horn antenna with wave 49 36. error antenna slant length 49 37. Dimensions of 3 GHz circular flange 50 38. 3 GHz pyramid horn 50 39. 3 GHz waveguide/coaxial transformer 51 40. Return 3 GHz antenna 53 41. Dimensions of 6 GHz circular flange 42. 6 horn 43. 6 GHz & transformer 44(a)&(b}. Return loss for 6 GHzhorn & waveguide
to coaxial transformer 45. Block frequency 46. frequency multiplication system
\
\ / densi ty tracer counter 47J Sweep and lock feedback loop
I 48/• Block unit 4~. Block diagram of the receiver unit 50. Open at 3 GHz 51. Open circuit stub characteristics
3 GHz notch 53. 3 GHz notch filter 54. print of 3 GHz notch
57 57
59 61
62 63 64 67 68 70 69 72
Univers
ityo
ape Tow
n
SS. (a) .. (f). Practical characteristics of 3 GHz filter 73 S6. Simulated 6 GHz notch filter characteristics 74 S7. Simulated 6 GHz notch filter characteristics 77
(variable short circuit resistance) S8. 6 GHz notch filter 76 59. Positive print of 6 GHz notch filter 78
60 (a)&(b). Return and insertion loss of 6 GHz notch 79 filter
61. Single section Wilkinson power splitter 62. Circuit bisected for even mode analysis 63. Circuit bisected for odd mode analysis 64. Admittance circuit: even mode
65. Admittance circuit: odd mode 66. Simulated 6 GHz power splitter characteristics
81 82 82
83
83 86
67. VSWR of the power splitter 88 68. Simple three port circuit 89 69. Three port with 70.7Q impedance transformer lines 91
70. Simulated 6 GHz power splitter characteristics (including various microstrip approximations)
71. Positive print of power splitter
72(a) .. (f). Actual characteristics of 6 GHz power splitter
73(a). Microstrip coupled lines 73(b). Elec~ric fields due to even and odd mode
excitation 74(a). Coupled transmission lines 74(b). Equivalent circuit 7S. Various modes on a coupled transmission line 76. Simulated coupler characteristics 77. Positive print of 6 GHz directional coupler 78(a) .. (d). Actual characteristics of directional
coupler 79. Parasitic etching of copper below template 80. Schematic of VT0-8100 series oscillators 81. Oscillator power supply and ground rail
viii
97
98 99
101 102
102 103 104 109 111 112
113 116 117
55. ( a) •• ( f) •
56. 57. Simulated
Practical characteristics of 3 GHz 6 GHz notch filter characteristics 6 notch
(variable short circuit )
58. 6 GHz notch 59. Positive 6 notch 60 (a)&(b). Return and insertion loss of 6 GHz notch
61. Single section Wilkinson 62. Circuit even mode analys 63. odd moae 64. Admittance circuit: even mode 65. Admittance circuit: odd mode 66. 6 GHz 67. VSWR of the power splitter 68. port 69. Three 70.7Q impedance transformer lines 70. Simulated 6 GHz power characteristics
(
71. Positive print
72(a) .• (f). power splitter
characteristics of 6 GHz
73(a). 73(b) .
excitation ( a). Coupled
coupled lines due to even
74(b). Equivalent circuit 75. on a coupled 76. Simulated coupler characteristics 77. 6 GHz
odd mode
)
78(a) .. (d). Actual directional
79. Parasitic etching of copper below template 80. of 100 81. Oscillator power supply and ground
73 74 77
76 78 79
81 82 82 83
83
86 88
89 91 97
98 99
101 102
103 104 109 111 112
113 116 117
Univers
ityo
ape Tow
n
82. Po?itive print of oscillator power supply 117 83. Test rig to measure oscillators output power 118
and tuning voltage
84. Tuning unit for the 3 GHz oscillator 120 85. Single ended mixers in the frequency multiplication 123
system 86. V-I characteristic of diode 125 87. Time dependent reflection coefficient 125
88(a). Equivalent circu~t of a basic mixer (ref 24) 128 88(b). Basic mixer circuit 128 89. Microstrip single ended mixer biasing circuit 129
90. Biasing circuit
91. Positive print of single ended mixer 92. Test rig to measure the match of the mixer 93. Match of single ended mixer 94. Test rig to measure transmitter unit power
95. IF power measurement test rig 96. Balanced mixer position in receive unit
97. Counter unit position in the receiver unit
ix
130
131 133 133 136 137
140 141
82. 83. Test
of supply to measure oscillators output power .
117 118
84. Tuning unit for the 3 GHz oscillator 120 reou'encv multiplication 123
system 86. V-I 125 87. Time dependent reflection coefficient 88(a). Equivalent of a basic ( 24) 88Cb). circuit 89. Microstrip single ended mixer biasing circui~ 90. 91. Positive print of single ended mixer 92. Test to measure the the 93. Match of single ended mixer 94. Test rig to measure
. IF power measurement test rig 96. mixer position in 97. Counter position in the unit
125 128 128 129 130
1
133 1
136 137 140 141
Univers
ityo
ape Tow
n
GLOSSARY
Corner reflector
Cyclone
DAST
density tracer
diode
dipole
directional coupler
micros trip
microwave freq.
- right angled pyramids made of metal,
designed to reflect signals independent of orientation.
device for separating dense from less
dense material.
- Division of Aeronautics & Systems
Technology (CSIR).
pieces of plastic with exact specific
gravities used in cyclones to measure
their separation efficiencies.
- semiconductor device used to rectify RF signals. It can also be used to mix two
independent signals.
device used to launch and receive
electromagnetic signals.
- microwave component used to insert a signal from one transmission line, to an
adjacent one, by inductive and capacitive coupling.
guiding substrate for microwave signals.
signals with a wavelength from 1 m to 1
mm.
x
Corner reflector
DAST
tracer
diode
dipole
microwave freq.
- right angled pyramids of metal, designed to signals independent of
from dense material.
&
Technology (CSIR).
of plastic with exact specific gravities used in cyclones to measure
- semiconductor used to RF signals. It can also be used to mix two
- device used to launch and receive
to a
signal from one transmission line, to an one, and
coupling.
guiding substrate microwave signals.
signals with a wavelength from 1 m to 1 Mm.
x
Univers
ityo
ape Tow
n
notch filters
power splitter
spiral antennas
- device used to attenuate a certain
selected frequency band.
- device used to split a signal into two
or more paths.
- overmoded dipole antenna. The elements are
wrapped around the feed point with the
wavelength of radiation related to the
diameter of the spiral.
xi
notch
power splitter
spiral antennas
used to attenuate a
band. - device used to split a signal
or more paths.
. two
- overmoded dipole antenna. The elements are
wrapped around the feed point with the
of to
diameter of spiral.
Univers
ityo
ape Tow
n
TABLE OF CONTENTS
Title page
Acknowledgements
Abstract
List of Tables
List of Figures
Glossary
I. INTRODUCTION
I.I The operation of the heavy medium cyclone
I.2 The Tromp curve-measuring the efficiency of
the cyclone
i
ii
iii
v
vi
x
I
I
5
I.3 The density tracer 7
I.4 Problem statement 7
I.5 Objective of thesis 9
2. DETECTION OF THE DENSITY TRACER IO
2.I Passive & Active detection IO
2.2 Examination of various passive detection techniques II
2.2.I Metal detection II
2.2.I.l Theory of metal detection II
2.2.I.2 Conclusions I2
2.2.2 Investigating Microwave corner reflectors as I2
density tracers
xii
TABLE OF CONTENTS
. Title page
Abstract
List of Tables
Glossary
1. INTRODUCTION
1.1 The
1.2 The Tromp curve-measuring the efficiency of
cyclone
i
iii
v
x
1
1
5
1.3 The tracer 7
1.4 statement 7
1.5 Objective of 9
2. DETECTION OF THE TRACER 10
2.1 & Active detection 10
2.2 Examination of various passive detection techniques 11
2.2.1 11
2.2.1.1 Theory of metal detection 11
2.2.1.2 Conclusions
2.2.2 corner as
density tracers
xii
Univers
ityo
ape Tow
n
2.2.2.1 Theory of corner reflector operation
2.2.2.2 Measuring reflected power from corner
reflector
2.2.2.3 Results
2.2.2.4 Conclusions
3. COUNTING DENSITY TRACERS BY FREQUENCY MIXING OR
MULTI~LICATION
3.1 The technique of frequency mixing and
multiplication
12
14
15-
16
17
17
3.2 Analysing the frequency mixing technique 17
3.2.1 The spiral antenna 18
3.2.2 The frequency mixer/counter system 20
3.2.3 Conclusions 21
4. MICROWAVE FREQUENCY MULTIPLICATION 22
4.1 A description of a prototype multiplication system 22
4.2 The cross dipole 24
4.3 Constructing the antennas for the prototype 26
multiplication system
4.3.1 Choosing the correct choke for the cross dipole 28
antenna
4.4 Filtering the RF plug in's second harmonic
4.4.1 Designing the quarter wavelength cavity
resonator
4.5 Testing the complete frequency multiplication
xiii
29
29
31
2.2.2.1 of corner
2.2.2.2 Measuring reflected
reflector
2.2.2.3 Results
from corner
12
14
2.2.2.4 Conclusions 16
3. COUNTING DENSITY TRACERS BY FREQUENCY MIXING OR 11
3.1 The technique of frequency mixing and 11
3.2 Analysing the frequency mixing
3.2.1 antenna 18
3.2.2 The frequency mixer/counter system 20
3.2.3 21
4. MICROWAVE FREQUENCY MULTIPLICATION 22
4.1 A a 22
4.2 The cross dipole 24
4.3 antennas 26
multiplication
4.3.1 Choosing the correct choke for the cross dipole 28
antenna
4.4 Filtering the RF plug 's second harmonic
4.4.1 the
resonator
4.5 Testing the complete frequency multiplication 31
xiii
Univers
ityo
ape Tow
n
system
4.5.1 Results
4.5.2 Conclusions
4.5.3 Recommendations
4.6 Testing a cross dipole antenna
4.6.1 Results of tests on 1/2 GHz cross dipole
antenna
4.6.1.1 Measuring the return loss
4.6.1.2 Conclusions
4.6.1.3 Measuring the radiation pattern
4.6.2 Conclusions
5. DESIGNING THE ANTENNAS FOR THE FREQUENCY
MULTIPLICATION SYSTEM
5.1 Designing the cross dipole antenna
5.2 Designing a suitable DC biasing choke
5.2.1 Conclusions
5.3 Choosing a suitable mixer diode for the cross
dipole antenna
33
33
33
35
36
36
38
38
40
42
42
44
46
46
5.4 Designing the 3&6 GHz transmit & receive antennas 47
5.4.1 The 3 GHz horn antenna & waveguide-coaxial 47
transformer
5.4.2 Testing the 3 GHz horn antenna & waveguide-coax 52
transformer
5.4.3 Conclusions 53
5.4.4 Designing the 6 GHz horn antenna & waveguide 53
xiv
system
4.5.1
4.5.2 Conclusions
4.5.3
4.6 Testing a cross dipole antenna
4.6.1 of tests on 1/2 GHz cross dipole
antenna
4.6.1.1 the return loss
4.6.1.2 Conclusions
4.6.1.3
4.6.2 Conclusions
5. DESIGNING THE ANTENNAS FOR THE FREQUENCY
SYSTEM
5.1 DeSigning cross dipole antenna
5.2 a DC biasing
5.2.1 Conclusions
5.3 a the cross
dipole antenna
33
33
35
36
36
38
38
40
42
42
44
46
46
5.4 Designing the 3&6 GHz transmit & receive antennas 47
5.4.1 3 GHz horn antenna & waveguide-coaxial 47
5.4.2 Testing the 3 GHz horn antenna & waveguide-coax 52
5.4.3 ions
5.4.4 Designing the 6 GHz horn antenna & waveguide 53
Univers
ityo
ape Tow
n
-coaxial transformer
5.4.5 Testing the 6 GHz horn antenna & transformer 57
5.4.6 Conclusions 58
6. CONFIGURATION OF 3/6 GHz FREQUENCY MULTIPLICATION 59
SYSTEM
6.1 Problems with conventional receiver units
6.2 The receiver local oscillator feedback loop
6.3 The fundamental frequency transmitter and
antenna
6.4 The receiver unit
7. DESIGN AND CONSTRUCTION OF THE NOTCH FILTERS
7.1 Analysing the stub notch filter
7.2 The design, construction and testing of the 3 GHz
notch filters
7.2.1 Analysing the operation of the filter
7.2.2 Testing the 3 GHz notch filter
7.2.3 Conclusions
7.3 Design, construction and testing of the 6 GHz
notch filter
7.3.l Analysing the operation of the 6 GHz notch
filter
7.3.2 Testing the 6 GHz notch filter
7.3.3 Conclusions
xv
59
60
63
64
66
66
69
69
72
74
76
78
79
'--
;-
-coaxial transformer
5.4.5 the 6 GHz
5.4.6 Conclusions
antenna & 57
58
6. CONFIGURATION OF 3/6 GHz FREQUENCY MULTIPLICATION 59
SYSTEM
6.1 with
6.2 The receiver local oscillator feedback loop
6.3
antenna
6.4 The
and
unit
7. DESIGN AND CONSTRUCTION OF THE NOTCH FILTERS
7.1 Analysing stub notch
7.2 design, construction and testing of
notch filters
7.2.1 Analysing the of
7.2.2 3 GHz notch
7.2.3 Conclusions
3 GHz
7.3 Design, construction and the 6 GHz
notch
7.3.1 Analysing the operation of the 6 GHz notch
7.3.2
7.3.3
the 6 GHz
xv
59
60
63
64
66
69
69
72
74
76
78
79
Univers
ityo
ape Tow
n
8. DESIGN AND CONSTRUCTION OF 6 GHz POWER SPLITTERS/ Bl
COMBINERS
8.1 Theory of power splitter operation 81
8.1.1 Determining the isolation and transmission 81
coefficients of the splitter
8.1.2 Simulating the 6 GHz power splitter 84
8.1.3 Simulating the effects of the quarter 89
wavelengths and isolation resistor
8.2 Designing the 6 GHz power splitter/combiner 93
8.2.1 Physical constraints on power splitter operation 93
8.2.2 The schematic of the power splitter 94
8.2.3 Testing the power splitter 98
8. 2. 4. Conclusions 100
9. THE 6 GHz DIRECTIONAL-COUPLER 101
9.1 The theory of directional coupler operation 101
9.1.1 Determining the coupled mode formulas 103
9.2 Designing the 6 GHz directional coupler 106
9.2.1 Simulating the 6 GHz directional coupler 108
9.2.2 Construction and testing of the directional 110
coupler
9.2.3 Conclusions
10. DEVELOPING THE 3 GHz AND 6 GHz OSCILLATOR UNITS
10.1 Designing the 3 GHz oscillator power supply
xvi
114
115
116
9. DESIGN AND CONSTRUCTION OF 6 GHz POWER SPLITTERS/ 91
COMBINERS
9.1 Theory of 81
8.1.1 81
coefficients of the splitter
8.1.2 the 6 GHz power 84
8.1.3 Simulating the of the quarter 89
8.2 Designing the 6 GHz power splitter/combiner 93
8.2.1 on power operation 93
8.2.2 94
9.2.3 Testing the power splitter 98
8.2.4 100
9. THE 6 GHz DIRECTIONAL ,COUPLER 101
9.1 The theory of directional coupler operation 101
9.1.1 Determining the coupled mode formulas 103
9.2 6 GHz 106
9.2.1 Simulating the 6 GHz directional coupler 108
9.2.2 and testing of the directional 110
coupler
9.2.3 Conclusions
10. THE 3 GHz AND 6 OSCILLATOR
10.1 Designing the 3 GHz oscillator power supply
114
1
116
Univers
ityo
ape Tow
n
10.2 Measuring the output power and tuning voltage
of the oscillator
10.3 Designing the tuning unit of the oscillator
10.4 The 3 GHz power amplifier
11. THE 6 GHz SINGLE ENDED HIXER
11.1 Theory of single ended mixer operation
11.1.1 Operation of the mixer diode
11.2 The basic mixer circuit
11.3 Designing the 6 GHz single ended mixer
11.4 Testing the mixer
11.5 Results
11.6 Conclusions
11.7 Sweep and lock circuit
12. FINAL INTEGRATION AND TESTING OF FREQUENCY
MULTIPLICATION SYSTEM
12.1 Determining the transmitter unit power
12.2 Measuring the receiver unit IF output power
12.3 Measuring the IF power of the integrated
frequency multiplication system
12.4 Improving system performance
12.5 Counting the density tracer
12.6 Conclusions
13. CONCLUSIONS AND FUTURE WORK
xvii
118
119
121
123
124
124
127
131
132
133
135
135
136
136
137
138
139
141
141
141
10.2 Measuring the output power and tuning voltage 1
of oscillator
10.3 the of the
10.4 The 3 GHz power amplifier
11. THE 6 GHz SINGLE ENDED HIXER
11.1
11.1.1 Operation of the mixer diode
·11.2
11.3 Designing the 6 GHz single ended _~.~~,~
11.4 the
11.5
11.6 Conclusions
11.1 Sweep and lock circuit
12. FINAL INTEGRATION AND TESTING OF FREQUENCY
MID..TIPLICATION
12.1 Determining the unit
12.2 Measuring the unit IF output power
.3 IF power
fr'eauencv multiplication system
12.4 Improving system
.5 tracer
.6 Conclusions
13. CONCLUSIONS AND FUTURE WORK
1
121
123
124
124
1
1
132
133
1
136
136
131
138
139
141
141
141
Univers
ityo
ape Tow
n
13.1 Conclusions
13.2 Future work
References
Separator
Appendices
A. Line widths and lengths for various frequencies
B. E and H plane plots for cross dipole antenna
C. Characteristics of DMK 5068 Schottky mixer diode
D. Flange dimensions for horn antennas
E. 3 GHz notch filter simulation programs
F. A notch filter analysis technique using PIN diodes
G. 6 GHz notch filter simulation programs
H. Three port and power splitter simulation programs
I. Mode amplitude formulas and coupler simulation
programs
J. Avantek VT0-8100 oscillator characteristics
K. Line impedances for single ended mixer
L. Schematic of Counter circuit
and ANZAC MD 162 balanced mixer characteristics
xviii
142
144
145
149
156
161
166
169
173
180
183
190
196
199
203
.1 Conclusions
13.2 Future
Separator
A. and
B. E and H plane plots for cross dipole antenna
C.
D. dimensions
DHK 5068
horn antennas
142
144
145
149
156
1
E. 3 GHz notch programs 169
F. A notch technique using PIN diodes 173
G. 6 GHz
H. port and power simulation programs
I. Hode
programs
J. VTO-8100 characteristics
K. Line impedances for single ended mixer
L. of
and ANZAC HD 1 balanced mixer characteristics
180
190
199
203
Univers
ityo
ape Tow
n
CHAPTER 1
INTRODUCTION
1.1 Dense Medium Separation
Dense mediuin separation (OHS) is a process used by the minerals processing industry to separate dense from less dense material in
a dense fluid. The technique relies on the principle that two
particles of different density can be separated completely if they are immersed in a liquid with a density intermediate to these two c 1 1. This is provided that there is sufficient settling
time.
A variety of OMS equipment types exists. One of the most connnonly used, especially for treating material smaller than 32 nnn, is the OMS cyclone. Figure 1 shows a typical dense medium cyclone.
The cyclone consists of a cylindrical section connected to a
conical cyclone
section. through
Feed contained a tangentially
cylindrical section.
Figure 1. Cutaway view of a cyclone
1
in a fluid medium enters the mounted inlet pipe in the
~di um/ore inle~
over~lo~
orifice
INTRODUCTION
1.1
Dense medium separation (OKS) a used by the minerals to in
a dense fluid: The technique on the principle two partic can be separated completely
are a a to these two (1]. This is provided that there is sufficient settling
time.
A variety of OMS equipment types
OMS 1 shows a
• One the most commonly than 32 mm, is the
medium The cyclone of a cylindrical section connected to a
a fluid medium cyclone through a tangentially mounted inlet pipe in
cylindrical section.
1.
1
medium/ore inJe-t
overflow or-iric:e
Univers
ityo
ape Tow
n
High density particles are forced, under action from the
centrifugal force, to move through the medium to the walls of the
cyclone. They then spiral down to the underflow orifice.
Lower density particles are caught up in the air core inside the
cylone, forced through the vortex pipe, and expelled through the
overflow orifice at the top of the cylindrical section.
Figure 2(a) is a photograph of a typical dense medium cyclone
Figure 2(b) shows a dismantled version of the cyclone.
Figure 2(a). Dense medium cyclone
2
High density particles are forced, under action from the centrifugal force, to move through the .edium to the walls of the
cyclone. They then spiral down to the underflow orifice.
Lower density particles are caught up in the air core inside the cylone, forced through the vortex pipe, and expelled through the overflow orifice at the top of the cylindrical section.
Figure 2(a) is a photograph of a typical dense medium cyclone Figure 2(b) shows a dis~antled version of the cyclone.
Figure 2 (a). Qcnse medium cyclone
2
Univers
ityo
ape Tow
n
Figure 2(b). Dismantled. c yclone (indicating over and underflow
orifices)
The fluid medium used is typically composed
fluid so
of powdered
ferrosilicon dispersed i n typical density of 2700
water. kgm-3,
conditions in a cyclone kgm-3 is attained.
a pseudo
3
The although
density of
formed has a
under the dynamic approximately 3100
Figure 2{h). Dismantled cyclone (indicating Qver and underflow orifices)
The fluid medium used ferrosilicon dispersed in typical density of 2700
is typically composed of powdered water. The fluid so formed has a kgmoJ, although under the dynamic
conditions in a cyclone a pseudo density of approxbnately 3100 kgmoJ is attained.
Univers
ityo
ape Tow
n
Material that has been separated is then sent to shaking screens,
where it is washed to recover the ferro silicon. The washed
material falls off the edge of the screen and proceeds to futher
treatment. The ferro silicon has a high magnetic susceptibility
and is concentrated means of a magnetic separator for reuse. The
complete d i nse medium separation sy~tem is shown in fiqure 3.
I inlet pip~~--''--~
--~ C\dclone
underf!ol.J pipe
o v erfJ01,,1 product
L reco~ered medium __ '----__ . ________ ___.
Figure 3. Schematic of the complete OMS system
Figure 4 indicates a shaking screen (underflow product).
4
Material that has been separated is then sent to shaking screens,
where it is washed to recover the ferro silicon. The washed material falls off the edge of the screen and proceeds to futher
treatment. The ferro silicon has a high magnetic susceptibility and is concentrated means of a magnetic separator for reuse. The
complete d i nse medium separation system is shown in figure 3.
I inlei: pipe
-~ cyclone
underfloy pipe
overfJo,", product
~r~XD~~~ __ ~underfloy product
medium
Figure 3. Schematic of t he complete OMS system
Figure 4 indicates a shaking screen (underflow product).
4
Univers
ityo
ape Tow
n
Figure 4. Photograph of a n underflow shaking screen
1.2 The Tromp Curve-measur ing the efficiency of the cyclone It is important c 2 1 to determine the separating efficiency of a cyclone, since it gives an indication of the operable state of
the cyclone. Low separation efficiencies could be caused by a variety of circumstances . The result would be a change in the
operating specific gravi ty of the cyclone. This would cause
product destined for the underflow orifice to report to the overflow orifice, and vice-versa.
This is obviously not desirable from an economic point of view.
5
Figure 4. P~otograph of an underflow shaking sCreen
1.2 The TroJIIp Curve-measuring the efficiency of the cyclone
It is important ,2, to deternLine the separating effiCiency of a
cyclone, since it gives an indication of the operable state of
the cyclone. LOw separation effiCiencies could he caused by a
variety of circumstances . The result would be a change in the
operating specific gravity of the cyclone. This would cause
pcoduct destined foc the underflow orifice to report to the
overflow ocifice, and vice-vecsa.
This i.ll obviously not de.llicable from an econOlll.ic point of view.
5
Univers
ityo
ape Tow
n
It is thus important to be able to measure the separation efficiency of the cyclone, to timeously detect problems with the cyclone.
An indicator of the separation efficiency of a cyclone is the slope of a Tromp curve. This curve shows the percentage of feed material of a specific density collected at the underflow orifice, against specific gravity. Figure 5 indicates a typical Tromp curve.
% of particles
at underflo~
Figure 5. A typical Tromp curve
A
c
B
I
I
I I
ID
SG
The curve shows that below a certain cut-off specific gravity C, few particles report to the underflow orifice. Above this point, it is found that most of the particles report to the underflow orifice. If a particle of specific gravity D is required to be processed, then it must be ensured that the system cut-off specific gravity C, is sufficiently low to allow the particle to report with 100% certainty, to the underflow.
It is expected c 3 1 that the Tromp curve should have a pattern as shown in curve A in figure S. However, particles which have a specific gravity approaching the specific gravity of the medium, will not have sufficient time to reach either the overflow (or the underflow) orifice and will report to the other product. For this reason the Tromp curve appears as curve B in figure S. Particles which have the same specific gravity as the medium,
6
It to be able to measure separation of the cyclone, to timeous detect problems with the
cyclone.
An indicator of the separation efficiency a cyclone is the of a curve.
, specific curve.
" of iC'articl es at underflow
5.
curve collected at
c
The curve that below a certain cut-off
. I
I
I
I
I
I
I II)
of feed
a typical
gravity C,
to point, it is found that most of the particles report to the underflow
gravity D to be
processed, then· it must be that the system cut-off
It shown
low to the to to the underflow.
[3] that the curve should curve A 5.
as a
gravity approaching the specific gravity of medium, will not have sufficient time to either the overflow (or
underflow) and will to the product. For reason the Tromp curve appears as curve B in figure 5.
which have the same specific as the medium,
6
Univers
ityo
ape Tow
n
have an equal chance of being expelled through either the underflow or overflow pipes.
1.3 The density tracer
The material inside a cyclone is a complex mixture of ore and a
fluid medium mix. It is difficult to measure the density of this mixture or how efficiently the cyclone is separating the dense from less dense particles. One widely used technique c 4 1.cs1 is
to construct a Tromp curve with the aid of density tracers, and determine the separation efficiency from the slope of the curve.
Density tracers are colour coded, uniformly sized plastic beads
which have a range of known densities. The beads are usually weighted with heavy metal salts to obtain the density required. These density tracers are added to the fluid medium before it enters the cyclone. Only the medium is present, and ore from the
mine is not included in the feed.
The density tracers are collected by hand from the overflow and underflow screens. They are then sorted- into various specific gravity categories, according to their colour. From these results a Tromp curve can be drawn and the separation efficiency determined.
Density tracers can thus provide a quick and accurate estimation of system performance. Unfortunately, they can only be used when a OMS process is not treating any material and, as such, they only provide an approximate measure of the true separation efficiency.
1.4 Problem statement - A problem arises when attempting to measure the efficiency of the
cyclone on line while ore is being treated. It is not possible to count the tracers, as they are generally buried under a layer of ore particles and ferro silicon. They will thus be difficult to
7
have an equal chance or pipes.
fluid medium mix. It mixture or how
to construct a Tromp curve
through either the
a complex mixture of ore and a to measure
the cyclone of this
[4].[5] is
determine efficiency from the s tracers, and
of the curve.
Density tracers are colour which have a range of
coded, uniformly beads known . The beads are usually
These density tracers are enters cyclone. Only
not included in the
tracers are
to added to
medium the medium before it
, and ore from
by the underflow screens. They are then sorted into various specific
, a Tromp curve can be
to drawn and
'-V..LV' ......... From the separation efficiency
Density tracers can thus provide a quick and accurate estimation , they can only when
a DMS process is not treating any material and, as such, they only an measure the true ef
1.4 Problem statement - A to measure the of
cyclone on count
while ore is being treated. It is not possible to tracers, as they are a
ore particles and ferro silicon. They will thus be difficult to
7
Univers
ityo
ape Tow
n
collect as particles.
particles.
many could The others
be removed with some of the larger ore would be processed with the finer
This problem of identification prompted the following question: Could a system be devised which would automatically count density
tracers used in diamond and coal mineral processing?
The advantages of a . system which can count density tracers remotely, while ore is being treated, are as follows:
1) Since the efficiency of separation can be measured while the
ore is being treated, the cyclone does not have to be off-line to
the rest of the plant.
2) The density tracers are counted by computer, not by hand. This
means that the processing time of the information is increased,
and consequently the response time to the information.
3) The cyclone and the automatic counting system can be connected in a closed loop control system to ensure that the cyclone is
operating at its maximum efficiency. Control can be achieved by altering the density of the fluid medium used and determining its
affect on the separation efficiency.
A completely new type of density tracer is required to enable it to be detected by some remotely situated system. Analysis of ore treatment operations gave the following criteria for the tracers.
1) They must comparable in size to the ore being processed. In
the diamond mining industry, the maximum size would be 32 mm.
2) The cost per density tracer must be less than RI.SO.
3) There must be little upkeep. This would imply that the density
a
collect as many could be removed with some of the would
particles.
ore
finer
This problem of prompted following :
Could a system be devised which would automatically count density
tracers
The advantages of a
remotely, ore
which can count density tracers , are as follows:
1) the efficiency of separation can be measured while the
ore , not to to
rest of the plant.
2) by , not by
means that
ity tracers are
the processing of the information increased,
and consequently the response time to the information.
3) The cyclone and the automatic counting can connected
a loop to ensure
at its maximum efficiency. Control can be achieved by the density of the fluid medium determining
A of density tracer to it to be detected by some remotely situated system. Analysis of ore treatment the following for the tracers.
1) They must to the ore being processed. In
the diamond mining industry, the maximum size would be Mm.
2) The cost per dens tracer must be than R1.50.
3) There must be . This would imply that the
8
Univers
ityo
ape Tow
n
tracers contain no battery driven components. Thus, a passive means of detecting the density tracer must be employed.
4) The density tracer must be able to exist in the harsh environment of the dense medium cyclone.
1.5 Objective of thesis The aim of this thesis is to investigate various passive detection techniques for counting density tracers, and construct a prototype of the most suitable scheme.
The system must comply with the above specifications.
An analysis of possible detection techniques is presented. A suitable scheme is then recommended. The method is commented on, and experimental procedures and results are discussed. The design, construction and testing of a prototype is described.
Recommendations are given with regard to the final integration of the product in the dense medium separation system. Conclusions are drawn as to the suitability of the system.
9
tracers contain no battery driven components. Thus, a passive means of tracer must be
4) The density tracer must be able to exist the harsh of dense medium cyclone.
1.5
The aim of this thesis is to investigate various passive detection ity tracers, a prototype of most suitable scheme.
system must comply with the above specifications.
An analysis of possible detection techniques presented. A
on, and experimental procedures and results are discussed. The
, construction and a
are of the product in the dense medium separation system. Conc
are as to of
9
Univers
ityo
ape Tow
n
CHAPrER 2
DETECTION OF TIE DENSITY TRACER
2.1 Passive & Active detection
The terms passive and active detection are chosen to indicate the
targets status in the detection process. A passive detection technique can be defined as one ..,.here the target has no means of signalling its presence. It uses a physical phenomenon , such as reflection, to indicate its loca·tion. _An example of this would be
an object detected by radar. The object can be detected, and to
an extent identified, by its reflectivity at radar frequencies.
Identification can be estimated from the amplitude of the reflected signal or the radar cross section.
Active detection involves a target indicating its presence by signalling, in some manner, on request. A typical example of this
are transponder system on aircraft or spacecraft. The aircraft
will transmit a sequence of codes, identifying itself, on receipt of a specific signal from an airport radar.
In choosing which type of detection important to see the advantages
techniques . .
system will be used, it is and disadvantages of both
Active detection is a simpler technique. A transmitter could be mounted into a density tracer. The signal could· then be received, indicating the presence of a density tracer. The receiver would also not have to be as sensitive as that used in a passive detection system. This is
relies on a l/r2 rule as passive detection.
because an active detection system
opposed to the 1/r4 rule observed by
Active detection systems have the disadvantage that they require an energy source to power their transponder systems. The
10
DETECTION OF TiE DENSITY TRACER
2.1
The are chosen to indicate the status ion. A i ve
technique can be defined as one-There the target has no means of . It uses a phenomenon , such as
, to indicate its loca'tion. An example of this would be
an object detected . The can be detected, and to an extent , by at Identification can be from the amplitude of the reflected signal or the radar cross section.
Active detection involves a indicating presence by signalling, some manner, on . A typical example of are transponder system on aircraft or spacecraft. The aircraft
I a of, f, on of a specific signal from an airport
In , important to see the advantages and disadvantages of both
is a A
mounted into a density tracer. signal could then be received, a tracer. The would
also not have to be as sensitive as that used in a passive because an
rule as opposed to the rule observed by
Active an energy source to power their transponder The
10
Univers
ityo
ape Tow
n
practical result of this in the density tracer detection scheme would be an increase in the time spent servicing such a device. Active detection density tracers would be more expensive than those for a passive detection system, since the miniature transmitter circuits would be expensive to design-and build.
Passive density tracers are cheaper to upkeep than active density tracers.
build and require less However, they need more
powerful transmitters and more sensitive receivers.
Since cost and upkeep are of primary interest application, a passive detection system will be used.
2.2 Examination of various passive detection techniques 2.2.1 Metal Detection
2.2.1.1 Theory of metal detection
for this
Metal detection relies on the fact that an object with a high magnetic susceptibility , will cause an increase in the flux linkage between two neighbouring coils. See figure 6. The increase in the flux linkage between the coils will cause an increase in the induced voltage in coil B (figure 6).
voltage source
Figure 6. Metal detection test rig
oscilloscope
If these coils were now mounted below the underflow shaking screen, as shown in figure 7., a metal density tracer would indicate its presence by an increase in coil B's induced voltage.
11
practical result of this the tracer would an in the time spent such a device.
tracers would more those for a passive detection system, since the miniature transmitter circuits would to design> and build.
density tracers are cheaper to build and require less upkeep than tracers. However, more powerful transmitters and more sensitive receivers.
upkeep are of application, a passive detection will be used.
that an object with a high magnetic , will cause an in
two neighbouring coils. See 6. The increase in the flux linkage between the coils will cause an
the in B ( 6) •
oscilloscope
6.
If were now mounted below shaking screen, as shown in figure 7., a metal tracer would
presence by an coil B's induced voltage.
11
Univers
ityo
ape Tow
n
The density tracers could thus be counted by determining the number of times the voltage level rose above a certain threshold.
·--------· ---
underfloi.J from C:!::IC:l one
--------·----
shaking
Figure 7. Coil positions under shaking screen
A problem with this system, however, is that since it has a low resolution (difficult to focus magnetic fields) it is difficult to discriminate between different metal types or sizes. Accompanying the ore from the mine is a substantial amount of extraneous metallic material, hammer heads, pieces of girders, copper wire etc.
2.2.1.2 Conclusions It would be difficult to determine the difference between a metallic density tracer and another piece of metal. It has also been proven too expensive to remove the extraneous metal from the ore. Metal detection is thus not a viable technique to count density tracers.
2.2.2 Investigating Microwave corner reflectors as density tracers
2.2.2.1 Theory of corner reflector operation For the role of a density tracer it is desirable to have a small radar ~arget with a large radar cross section c 6 1. The larger the
12
The density tracers could thus be counted by the threshold. of level rose a
.7.
underflow from c!dclone
-.-----.----
shaking
A with , ht'1'loU"VIPT', since it has a low resolution (difficult to focus magnetic fields) it difficult to metal types or sizes. Accompanying the ore from the is a of
of girders, extraneous copper wire etc.
2.2.1.2
material, hammer heads,
It would be difficult to determine the difference between a tracer • It
been proven too expensive to remove the extraneous metal from the ore. density tracers.
2.2.2
For the radar
not a to count
a a large radar cross section [6]. The larger the
12
Univers
ityo
ape Tow
n
radar cross section, the greater the signal return. A flat metal plate could be
used, its radar cross section is given by:
where aP = radar cross section m m2
A = area of metal plate in m2
A. = wavelength in m
.. (2.1)
The reflection from a metal plate is very orientation dependent, the radar cross
section drops rapidly as the sheet is rotated. The return is strongest along the
normal.
A construction that provides better orientation dependance with good radar cross
section, is the microwave corner reflector.
Microwave corner reflectors are devices that reflect input signals back along
their incident paths. Figure 8 indicates a triangular corner reflector and the path
of a beam entering the device.
Triangular corner reflector
Beam
Figure 8. Microwave corner reflector
Microwave corner reflectors are simply three mutually perpendicular metallic
plates. As a beam enters the corner reflector, it is reflected three times and
returns along its incident path.
13
radar cross section, the the signal return. A flat metal plate could be
used, its radar cross section IS n by:
where a p ::::: radar cross section In m2
A ::::: area of metal plate in m2
,\ ::::: wavelength in m
.. ( 1)
The reflect from a metal plate 1S very orientation dependent, the cross
section drops rapidly as the sheet is rotated. The return is strongest alo the
normal.
construction that provides tter orientation dependance with good radar cross
section, is the microwave corner reflector.
Microwave corner reflectors are devices that reflect input signals back along
their incident paths. Figure 8 indicates a tria corner reflector and the path
of a beam e ng the device.
Triangular corner
Beam
Microwave corner reflectors are simply three mutually perpendicular metallic
plates. As a beam enters the corner reflector, it is reflected three times and
returns aI its incident path.
13
Univers
ityo
ape Tow
n
The cross sectional area of the triangular corner reflector can
be obtained by equating the effective area of the flat plate to
the maximum area of the triple reflections.
The maximum area of triple reflection is that afforded by the
corner when it is viewed along its axis of symmetry. This area is
a hexagon as shown in figure 9.
Figure 9. Maximum area of reflection from reflector
The maximum area of reflection of the triangular corner reflector
is then given by:
•. (2.2)
where a = edge length of a corner reflector in m.
subst. (2.2) into (2.1) yields for triangular reflectors:
.. (2.3)
2~2.2.2. Measuring reflected power from corner reflector
A microwave corner reflector was designed and constructed with an
edge length of 50 mm. The corresponding radar cross section ls
0.356 m2 9 from eq. (2.3))
14
The cross sectional area of the be obtained by equating the the area
corner can area of the flat plate to
area reflection that afforded by the corner when is viewed along a hexagon as shown in figure 9.
9.
maximum area of then given by:
=
where a edge length of a corner
subst. (2.2) into (2.1) yields
2~2.2.2.
edge of 50 mm. 0.356 m2 9 from eq. (2.3»
14
area
corner
.. (2.2)
in m.
triangular :
.. (2.3)
with an cross
Univers
ityo
ape Tow
n
A test setup consisting of a 35 GHz transmitter and horn antenna, a 35 GHz
receiver and horn antenna and spectrum analyser, is mounted 30 cm above a
table of rocks. The antenna gains were approximately 28 dB. The ,test rig is
shown in figure 10.
35 GHz horn 35 GHz osc.
spectrum analyser 35 GHz
/ "------------!
rock
oo
corner Q reflector
Figure 10. Test ng measuring reflections from corner reflector
2.2.2.3 Results
It was found that the reflected power from the rock mixture varied dramatically
with rock orientation. The return signal varied between -40 dB and -60 dB as
viewed on the spectrum analyser. The corner reflector was placed in the path of
the signal, to ensure maximum possible signal return. At best this was found to
be -40 dB.
Thus the best resolution available with a 100 mm diameter corner reflector at
35 GHz, is comparable to the return from certain rock orientations. The reason
for this is that at 35 GHz, there will not be an appreciable specular refl~ction
from these rocks. This means that seen along the normal a rock will appear to
be a relatively flat plate.
A large corner reflector ( side length 350 mm, radar cross section 855:-5o7m2)
was placed in the path of the 35 GHz signal. The assumption being that if there
was an appreciably better signal return with a 350 mm edge corner reflector at
35 GHz, then the original 50 mm edge length corner reflector would work well
at 350/50 *35 GHz = 245 GHz. However, the maximum signal was
15
A test setup of a Hz transmitter and horn antenna, a GHz
receiver and horn antenna and spectrum analyser, is mounted 30 em above a
table of rocks. antenna gains were approximately 28 dB. The ,test rig is
shown in figure 10.
35 GHz horn 35 GHz osc.
rock
spectrum 35 GHz
1
It was found that the reflected power from the rock mixture varied dramatically
with rock orientation. The return signal varied between ·40 dB and -60 dB as
viewed on the spectrum analyser. The corner reflector was placed in the path of
the aI, to ensure maximum possible return. At best this was found to
be -40 dB.
Thus the resolution available with a 100 mm diameter corner reflector at
GHz, is comparable to the return from certain rock ntations. The reason
for this is that at GHz, there will not be an appreciable specular refl~ction
from these rocks. This means that seen along the normal a rock will appear to
be a relatively flat plate.
corner reflector ( side length mm, radar cross section m'2)
was placed in the path the GHz signal. The assumption being that if there
was an appreciably better signal return with a 350 mm corner reflector at
z, then the original ngth corner reflector would work well
However, the maximum signal was
15
Univers
ityo
ape Tow
n
again measured to be -40 dB. This means that the corner reflector would not benefit from operating at a higher frequency.
The fact that increasing the radar cross section had no effect on the signal return can be explained as follows. The corner reflector is operating within the near field of the horn antennas. Thus the radar cross section is not valid here, and hence the reflected power is not linearly dependent on radar cross section or the inverse fourth power of distance. Further, for efficient operation the frequency would have to be increased to ensure specular reflections occurred from the rocks. At these high frequencies signal penetration through surface sludge or
contaminatio~ is negligible and the corner reflector would have to be kept perfectly clean.
2.2.2.4 Conclusions The reflections from rocks in certain orientations may be of the same magnitude as a corner reflector. For this reason corner reflectors are not suitable as density tracers.
16
to be -40 dB. means corner reflector would not benefit from operating at a higher
no on cross the signal return can be explained as lows. The corner
operating antennas. Thus the radar cross section not valid
hence
cross for ef
not
or the inverse fourth of distance. would to
to ensure specular reflections occurred from rocks. At
, and
high frequencies signal penetration through surface sludge or
and corner have
to clean.
2.2.2.4
The reflections from may
same as a corner For this reason corner are not suitable as density tracers.
16
Univers
ityo
ape Tow
n
,_ CHAPTER 3
COUNTING DENSITY TRACERS BY FREQUENCY MIXING OR MULTIPLICATION
3.1 The technique of freguency mixing and multiplication Another passive means of detection relies on frequency mixing or multiplication. A Schottky mixer diode has a non linear voltage current characteristic which produces harmonics of any sufficiently large signal incident on it. It can _also accept two signals and produces their mixed components.
These characteristics provide two ways of detecting density tracers. In both cases a mixer diode is inserted into a density tracer. The frequency mixing technique will require two transmitters operating with frequencies A and B. The diode will produce a mixed component of frequency A-B, this signal will be detected by a receiver and will thus indicate the presence of a
density tracer.
In the case of frequency multiplication, a single transmit frequency is used. The mixer diode receives the signal and produces harmonics. The second harmonic has the highest amplitude and will thus be the easiest to detect. The density tracers can thus be counted by noting the number of second harmonic pulses received.
3.2 Analysing the frequency mixing technique As stated previously the mixing technique would require two transmitters/antenna combinations. The mixer diode would be mounted across the feed points of an antenna, as shown in figure 11, inside a density tracer. This antenna would be efficient at receiving the · two transmit frequencies as well as transmitting their mixed component. A suitable antenna would be the spiral antenna.
17
CHAPTER 3
multiplication. A Schottky mixer diode has a non or
linear voltage current which of sufficiently large signal incident on it. It can also accept two
produces
These characteristics provide two ways of detecting density
tracers. In both cases a mixer diode is tracer. The transmitters operating with frequencies A
will B. The
a density two
diode I
a A-B, be detected by a receiver and will thus indicate the presence of a density tracer.
In the case of frequency multiplication, a transmit The the s
• The second harmonic has highest amplitude and will thus be the to detect. The dens tracers can
be counted by noting received.
3.2 As would two transmitters/antenna combinations. The diode would
of an antenna, as shown in
II, a density tracer. This antenna would be receiving the· two as well as their mixed component. A suitable antenna would be antenna.
17
at
Univers
ityo
ape Tow
n
mixer diode
~Di---___. antenna
Figure 11. Mixer diode mounting
3.2.1 The spiral antenna
densit!::I
tracer
This antenna has a very wide bandwidth and is thus able to operate at the three frequencies required by the mixing technique. This antenna has the useful property that it is circularly polarised. The density tracer falls off the screen in any orientation, thus the signal it receives from the transmitter will be in an unknown plane of polarisation. With a circularly polarised antenna, the plane of polarisation would not be an issue .. Generally the antenna efficiency would be 3 dB lower due to it being circularly polarised, but the benefits overshadow this slight loss in efficiency. The spiral antenna is shown in figure 12.
A
Figure 12. The spiral antenna
-opposite phase
The spiral antenna will radiate when currents of opposite phase are exactly one wavelength apart. This means that the antennas
18
mixer diode
densitw
tracer
Figure 11. Hixer diode mounting
3.2.1 This wide bandwidth and is thus able to
by the technique. This antenna has the property that circularly polarised. The tracer falls off the screen in any orientation, thus the Signal it receives from the
in an unknown polarisation. With a circularly polarised antenna,
."Generally the antenna to
of polarisation would not be an efficiency would be 3 dB lower due
overshadow this slight loss in efficiency. The spiral antenna shown in
12.
The antenna when currents of opposite phase are exactly one means antennas
18
Univers
ityo
ape Tow
n
bandwidth will be determined by its width. For lower frequency
signals the antenna will radiate near the edges (as in point A),
higher frequency signals will tend to radiate near the center of
the spiral (point B).
The relationship between the wavelength and the diameter of the
spiral is:
where
A. = rr*D
A = wavelength of signal
D = diameter of spiral
. . ( 3 . 1 )
Assuming that the maximum diameter of the density tracer is 32 mm, the wavelength of the lowest frequency is given by (3.1)
A = rr*D
= 3.1415 * 32
A.= 100.53 mm
The frequency of operation is given by:
f = c/A
= 3*10-0/100.53*10-3
= 2.98*10 9 Hz
.. (3.2)
Thus 3 GHz is the lowest frequency and will be the mixed
component in the system.
Assuming that one of the transmitters is operating at 4 GHz, the other would have to operate at 4+3 = 7 GHz.
19
the
mm, the wavelength of
the
lowest frequency
tracer
given by (3.l)
A.. = «*D
;: 3. 15 * 32
A. = 100.53 mm
operation is given by:
f == cIA •• (3.2)
== 10-0 /100. 10- 3
the and will be the mixed
component in
Assuming that one of the transmitters operating at
would to at ==
19
Univers
ityo
ape Tow
n
3.2.2 The frequency mixer/counter system
Figure 13 indicates a density tracer counting· system using frequency mixing.
,---e-1"· -=-f~4 f spiral signal , ~s antenna analyser I
_j
Figure 13. Fregyency mixing/density tracer counter system
To ensure that the receiver is detecting the mixed signal, i.e.
the mixed component, we will have to reference it to the transmitters. There are two ways of doing this.
1) Crystal locking the transmitter oscillators to form a stable reference for the mixed component.
2) Installing a sweep-and-lock circuit at the receiver. This circuit will sweep over a desired bandwidth and detect and lock
onto a sign~l received from the mixer diode.
The problem with solution 1 is that crystal locked oscillators are expensive, especially at 7 GHz. This would not be an economically viable solution. Solution 2 also has its problems, as shown below.
It is assumed, as in figure 13, that the density tracer is detected in free fall over the shaking screen. There is a maximum range at which the density tracer/ mixer diode can be detected. If this is assumed to be 30 cm {suitable distance from edge of
20
3.2.2 Figure
13.
mixing.
-z-~ signal analyser
counting' system using
To ensure that the detecting the mixed signal, i.e. we have to to
transmitters. There are two ways of doing
1)
for the mixed component.
2) a will sweep over a
onto a
problem are expensive,
as shown below.
1
especially at 7
to a
at the receiver. This bandwidth and detect and lock
diode.
GHz. This would not be an 2
It assumed, as in figure 13, that the density tracer
range at If
over the screen. is a maximum which the density tracer! mixer diode can be detected.
to be 30 cm {
20
Univers
ityo
ape Tow
n
screen to detection unit), then the time taken for the density tracer to fall through this distance is given by:
where
t v2*s*g = if2*30*10- 2 *9.W = 0.247 seconds
s = distance in m g = gravitational acceleration in m/s 2
t = time in s
.. (3.3)
Most sweep-and-lock circuits take longer than this to lock onto a received signal. The circuit used on the Plessey MRA-7 distance measuring survey instrument device locks after 0.5 s. Thus a sweep and lock circuit that could detect signals in time would need expensive and intricate circuitry.
To swmnarise then, the mixing technique has the disadvantage that
it's transmitter needs two oscillator/antenna systems.
Referencing the mixed component to the receiver, so as to ensure early detection, will require either expensive crystal locking oscillators or quick sweep and lock circuits.
3.2.3 Conclusions Considering the various disadvantages and advantages it appears that frequency mixing is not a satisfactory technique from the point of view of economics. Frequency multiplication will be analysed in the next chapter, to see if it is a more cost effective means of counting density tracers.
21
screen to detection unit), then taken for
given by: tracer to fall through distance
t
s
t
m
gravitational acceleration in m/s2
s
.• (3.3)
Most sweep-and-Iock circuits take longer than this to lock onto a on
measuring survey instrument device locks lock circuit that could
need ive and intricate circuitry.
KRA-7 after 0.5 s. signals in
·Thus a
would
To summarise then, the mixing technique has the disadvantage that
it's transmitter two the mixed component to , so as to ensure
early detection, or quick sweep and lock
3.2.3
Considering the various disadvantages and it
point analysed
mixing is not a factory technique from the
view of economics. Frequency multiplication next , to see it a more cost
means of counting density tracers.
21
Univers
ityo
ape Tow
n
CHAPTER 4
MICROWAVE FREQUENCY MULTIPLICATION
4.1 A description of a prototype multiplication system It has already been described how density tracers can be counted by detecting the second harmonic reflected from a Schottky mixer diode.
A frequency multiplication technique, since only one necessary. It was decided available components such as variable oscillator and a detection unit.
system has proven to be a simpler transmitter/antenna system is
to test this technique by using the HP 3595 RF plug in unit as a HP 8555 signal analyser as the
For ease of construction a half wave dipole antenna constructed from round brass rods would be used for the transmitter and receiver units. A fundamental chosen to test this principle, antenna dipole length of:
frequency c 7 1·t 0 1 . of 500 MHz was
as this gave a manageable transmit
1 = A.12 = c/(2*f) = 3*10°/2*500*10 6
= 300 nun
.. (4.1)
Figure 14 shows the dipole and its electrical length.
>../:2, 1--·-·········································--i I H
feedpoint-s
Figure 14. Electrical length of 500 MHz dipole
22
MICROWAVE FREOUENCY MULTIPLICATION
4.1 It by detecting the second
A
technique, since only one
necessary. It was decided to test
as HP
variable oscillator and a HP 8555
unit.
For ease a half wave
has
can be counted from a Schottky mixer
to be a simpler
technique by using
5RF plug in unit as a signal analyser as the
from round brass rods would be used for
antenna constructed
transmitter and
receiver [7],(0] 500 KHz was chosen to test this principle, as this gave a manageable transmit
antenna ::
I A/2 :::: c/(2*f) .. (4.1)
= 3*10 0 /2*500*10 6
14 the
f'eedpointlll
14.
Univers
ityo
ape Tow
n
Furtber, it is less critical to end effects loading and feed point impedance discontinuity effects than a spiral antenna.
The receiver unit will detect the second harmonic of the fundamental frequency reflected from the mixer diode. It will thus require a receiver operating at 1 GHz. For simplicity another dipole antenna will be used, and constructed from brass.
All that is required is to decide upon a suitable mixer diode and the antenna which will receive the fundamental frequency and transmit the multiplied second harmonic.
It was decided to use a HP 5802-2800 Schottky mixer diode as they are readily available and they operated adequately over the bandwidth of the system.
The choice of antenna upon which the mixer diode is important. It must be able to operate at the
mounted, is fundamental
frequency as well as the second harmonic. The antenna must also have a wide beamwidth. The reason for this is that the density tracer can be in any orientation after rolling off the edge of the screen. The antenna may then present a null in its radiation pattern towards the receiver, which might result in the density not being detected.
In the previous chapter, it was shown how a spiral antenna was used in a frequency mixing technique. The mixer diode was mounted between the feedpoints of the spiral antenna. The spiral antenna being wide band would receive the two fundamental frequencies and transmit the mixed component. This antenna has the following characteristics.
a) wide bandwidth b) circularly polarised
23
Furtber, it less critical to loading and
antenna.
The unit fundamental frequency
a antenna will
the second harmonic of the
It
operating at 1 GHz. For simplicity
brass.
All that is required to decide upon a mixer diode and antenna and
transmit the multiplied second harmonic.
It was to use a HP 5802-2800 as are readily avai·lable and they operated adequately over
of the
The choice of antenna upon which the mixer diode is mounted,
It must to at the
as as the second harmonic. The antenna must beamwidth. The reason for this is that the density
tracer can the screen.
of The antenna may then present a null in its radiation
not being detected.
In the previous chapter,
used in a frequency mixing between the feedpoints of
would
, which
it was shown how a . The
antenna was diode was mounted
the spiral. antenna. The spiral antenna
transmit the mixed component. antenna has the following
a) bandwidth b) circularly polarised
23
Univers
ityo
ape Tow
n
c) reasonable beamwidth
d) good radiation efficiency
This antenna would also be .suitable for the frequency multiplication technique. The mixer diode could be mounted across the feedpoints of the antenna, as shown in figure 11. A spiral antenna could not be obtained which operated between 500 MHz and
1 GHz (DAST supplied spiral antennas which operated from 2 GHz to
18 GHz).
It was decided that an antenna would be chosen that could operate at both transmit and receive frequencies with a reasonable
beamwidth . The antenna would have to be efficient at radiating at these frequencies (low return loss). The criterion of circular polarisation, would be dropped, as .the system was a prototype, and the density tracer would be detected statically (density tracer would not be detected as it was falling). An antenna which.
fulfilled this requirement was the cross dipole antenna.
It will be described in the next section.
4.2 The cross dipole antenna An c 9 1,c 10 1 antenna which satisfies the criterion of operating at
two frequencies, is the cross dipole antenna (shown in figure
15).
Figure 15. The cross dipole antenna
The elements of the dipoles are orthogonal to each other. This means that the electric fields of the two dipoles are at right
24
c) reasonable beamwidth
d) good radiation
antenna would also be . suitable for could multiplication technique. The
the feedpoints of the antenna, antenna could not be obtained
as shown in which
the frequency mounted across
11. A
500 MHz and
1 GHz (DAST
18 GHz).
antennas which~ operated from 2 GHz to
It was decided that an antenna would be chosen that could operate at both with a
antenna would have to be at radiating at these frequencies (low return loss). The criterion of circular
, would dropped, as the system was a prototype, and the tracer would detected (
tracer
fulfilled
as it was falling). An antenna which.
requirement was the cross dipole antenna.
1 be described the next section.
4.2 An [9],[10] antenna which satisfies the at
two , is the cross dipole antenna (shown in figure
) .
15.
dipoles are orthogonal to each other. This means that electric the two are at
24
Univers
ityo
ape Tow
n
angles to each other, minimizing any coupling between the dipoles.
The elements lengths are in the ratio 2:1 since the antenna has to operate at both the fundamental and second harmonic. This improves the isolation for the following reasons. The longer element will appear as a resistive impedance across the
feedpoints for the fundamental frequency (point A), but will be
an open circuit for the second harmonic (point B). The voltage waveforms indicate this in figure 16(a). Similarly the shorter elements appear as a resistive impedance at the feedpoints at the
second harmonic (point A figure 16(b)), but will be a high
impedance at (point B) for the fundamental, as can be seen in
figure 16 (b).
A/2 a1:
>../2 a1: Fo
Figure 16(a). Operation of fundamental mode element
>--/2 a1: Fo
Figure 16(b). Operation of second harmonic mode element
This means that the antennas are isolated from each other and the fundamental frequency will radiate more efficiently through longer element. Thus the second harmonic frequency radiate more efficiently through the shorter elements.
25
angles to each other, minimizing any coupling between
The are the ratio 2:1 s the antenna has to operate at both the and
for the following reasons. . This longer
element will as a fundamental the
impedance across the frequency (point A), but will be
an open ( B) •
waveforms indicate this in figure l6(a). shorter at
will be a high appear as a resistive
second harmonic (point A figure at ( B) the
figure 16 (b) .
Figure (a) .
)./2 ai: Fo
Figure 16 (b) .
at 16(b», but
, as can seen
This means that the antennas are isolated from each other and the more through
element. Thus the second harmonic frequency radiate more through the
Univers
ityo
ape Tow
n
4.3 Constructing the antennas for the prototype multiplication
system
The prototype multiplication system requires the following
antennas:
A 500 MHz transmit dipole
A 1 GHz receive dipole
A 500 MHz/ 1 GHz cross dipole
The dipoles would be constructed from brass rods (cross section 6
mm). Brass rods are easier to solder than the traditional antenna
material, aluminium.
Two dipoles were constructed, and their lengths reduced, until
the resonant frequencies were approximately 500 MHz and 1 GHz.
Figure 17 shows the return loss pl_ots for the two antennas. 0
+10
Return loss in dB
~2.o
+Jo
Figure 17(a).
+10
Return loss in dB +
4-30MHl 5&0 MHt
5oO 1000
Return loss f oJ;;: transmit dipole
:sco 1000
Figure 17(b). Return loss for receive dipole
Table 1 shows the characteristics more clearly.
26
14/o 1'1Hi-
-f'CM11i.) •$00
The prototype
antennas:
A 500 MHz
multiplication system
A 1 GHz receive dipole A 500 MHz/ 1 GHz cross dipole
the following
The dipoles would be constructed from brass rods (cross section 6 mm). Brass are easier to antenna
, aluminium.
Two dipoles were constructed, and their
resonant were 11 shows return loss plots
o
Return loss
11(a).
Return los~ in
Figure 11 (b).
Table 1 shows the more
26
lengths
500 MHz and 1 GHz. the two antennas.
Univers
ityo
ape Tow
n
Resonant frequency (MHz) bandwidth (MHz) Return loss (dB) 480 130 25
990 260 30
Table 1. Characteristics of transmit and receive dipoles
The 500 MHz/1 GHz cross dipole antenna was constructed using the
same 6 mm brass rod. Element lengths of 144 mm and 60 mm were
finally chosen. The dipole elements were soldered together and an
SMA-coax adapter was soldered to the feedpoint.
The return loss plot was obtained and can be seen in figure 18.
,,-..
·~•)_ ())"
0 ...J.
z
L -
-10
10
20
I
500 MHz COUPLED ANT
.. .• ..-----------'·..... (.... -........ , . ..··' ........ '•· ••••. I I.. .
......... '·., .. -·..-· \\ / ........
\ I
'· / './
4~ s ... ~r~·~w~;~1---+-~~1~+1~~•~-1-~~.1~-+-1 _____,, · .4-' ··0&>·div 1·.2
! F r ..i:- '«\ 1...1 e ri C_>" · i::' G H z ~ · ____,
Figure 18. Return loss for 500 HHz/1 GHz cross dipole antenna
Table 2 shows the characteristics of this antenna as obtained
from the network analyser.
27
1.
The 500 MHz/l GHz cross antenna was same 6 mm brass rod. Element lengths of finally chosen. The dipole were SMA-coax was soldered to feedpoint.
The return loss plot was can seen
(1 t1Hz COUPLED -10
,,-. . .
I~ 13 IV i '(r)
10 v'f 0 -J
z 213 a::: ::t I- 3t3 lJJ
~~+~.~~--+--+--~~~~~~~
.4--' l_
Figure 18.
using the
~-=~ were
18.
Table 2 shows the characteristics of this antenna as the
27
Univers
ityo
ape Tow
n
I
Resonant Frequency (MHz) Bandwidth (MHz) Return loss (dB) 500 150 15 1030 270 25
Table 2. Characteristics of cross dipole antenna
4.3.1 Choosing the correct choke for the cross dipole antenna The mixer diode placed across the feedpoint required a reference
bias voltage. Since we are using a schottky diode, the reference voltage is zero voltage. All that is needed then is to place a
choke across the diode, to provide a closed DC path for rectified current to flow. The mounting of the choke is shown in figure 19.
I Died~ J
Figure 19. Choke mounting on cross dipole antenna
It is important to choose a choke with a sufficient number of
turns. An unsuitable choke will effect the match of the antenna, as well as its resonant frequency.
Chokes were designed using found that the resonant
0.2 mm diameter copper wire. It was frequency decreased as the number of
turns of the choke was increased. When too few turns were used, the amplitude of the reflection coefficient was reduced. A choke with 12 turns was chosen. Figure 20 shows the return loss of the loaded cross dipole antenna.
28
Table 2.
4.3.1
feedpoint required
we are using a schottky diode, the reference voltage zero voltage. All to a choke across the diode, to provide a closed DC path for
to the 19.
. Choke mounting on cross dipole antenna
It to choose a choke with a suf number of
turns. An choke will fect the match of the antenna, as as resonant
were us 0.2 rom . It was found the resonant frequency decreased as the number turns of choke was turns were the amplitude of the reflection coefficient was reduced. choke
turns was 20 return loss loaded cross dipole antenna.
28
Univers
ityo
ape Tow
n
+10
Ret lJrn 1 oss
in dB
+lo
4'3.0
4-'=>4
soo lO~I
·soo 1000
Figure 20. Return loss of cross dipole with 12 turn choke
4.4 Filtering the RF plug in's second harmonic
An HP 3595 plug in was used as a transmitter. It was found that
the oscillator produced the second harmonic 20 dB below the
fundamental frequency. It is important to reduce the amplitude of
the second harmonic from the oscillator,because it could be
confused with the multiplied second harmonic from the mixer
diode. A suitable filter would be a quarter wavelength cavity
resonator (A= 500 MHz).
4.4.1 Designing the gyarter wavelength cavity resonator
A l 11 1 quarter wavelength cavity resonator is a device which
allows the fundamental and odd harmonics to propagate, while
filtering the even harmonics (including the second harmonic).
It was decided that a circular coaxial cavity resonator be used . •
To obtain the maximum match (maximum Q-factor), the ratio of the
radii r 1 and r 2 (as in figure 25) must be 3.6. Aluminium tubing
with an outside diameter of 101.85 mm and an inner diameter of 94
mm, was obtained. A solid brass rod of with a diameter of 15.07 mm would be used for the inner conductor. The ratio of the radii
is:
.. (4.2)
= 3.8
29
+10
R€1turn loss in dB
+20
Figure 20.
4.4
430 Soo
'soo
the oscillator produced the second . It is
the second harmonic
......... v ...... o;;;. A
resonator (X= 500 MHz).
a
.0101
'000
. It was found that harmonic 20 dB below the
the ,because it
4.4.1 Designing the quarter wavelength cavity resonator A (11) resonator which
propagate, while allows fundamental and odd harmonics to even .... Q ........ u:v .............. ) •
It was decided that a circular coaxial cavity resonator be To obtain the ), the
and (as in 25) must 3.6. Aluminium tubing
an of an inner of mm, was obtained. A solid brass rod of with a diameter mm would be used for the inner • The of is:
= (94.90/ .07)/2 3.8
29
.. (4.2)
Univers
ityo
ape Tow
n
The completed cavity resonator is shown in figure 21.
·nput coil 2 short
circuit11-~~~~~....---..~--4 r1
cen er conductor
........................................... L_. ---- .V4_ .~~ 1-\~t) ...
open circ:ui t side
Figure 21. 500 MHz coaxial cavity resonator
It can be noted that the signals are propagated by a loop on the short circuit side of the cavity. A metal plate shorts the aluminum outer conductor .to the brass inner conductor. Signals are received via the output probe at the open circuit side of the resonator.
Figure 22 shows the re~urn loss and insertion losses for the
cavity resonator.
- 1 5 e.~a.~ 1..oss . ;-4 . '
-5
5
' .1 s !
1 ·: .
IN ~e. . i i
. '· ! ,,,,------~,
j
i
J.
0
4
a
1'2
l •
1.6 . .004/div .S .46 .0~4/div l 35 l
~ 46 E.L~. ~.ld5 ri c '::'---~-~-Hz) .F:...r.~3. . .\,,l_~_nJ:;: Y. (GHz)_ Figure 22. Return and insertion losses for the cavity resonator
. ! • S' ,
I.
From the figures it can be noted that the cavity resonator has an insertion loss of 2 dB at 478 Hilz. From.figure 23 it can also·be
seen that the insertion loss at 956 MHz is 30 dB.
30
/
The completed cavity resonator
short ~ircuit~ ________ ~~ ____ ~
L-.
Figure 21.
It can noted that the signals
s aluminum outer conductor .to the are via the output probe resonator.
22 cavity resonator.
-15 1t~Q.N .... oss
1 .... ell!.
-5
5
·15
5 :
in 21.
circuit
are propagated by a loop on the A
brass inner conductor. Signals at the open circuit
1 e .'
6 .004/div .13'0 .. /' d i v . !
• S' , E.f_t_ '!~~ ~.:: ~ ___ ~:;"'-!.;:..=...'--- y (GHz) 22.
From the figures can be noted that the cavity resonator has an of 2 at 418 MHz. From 23 it can
seen that the insertion loss at 956 KHz is 30 dB.
30
Univers
ityo
ape Tow
n 'VJ--~~~~~
signal anal~ser
cross dipole antenn3
Figure 24. Frequency multiplication test setup
To obtain maximum benefit from the fact that the resonator
operated at 478 MHz, the RF plug in was tuned to this frequency.
·The fact that the frequency is 22 MHz short of the expected
fundamental frequency, is of little consequence. The antennas in
the system have wide enough bandwidths to accept both the 478 MHz fundamental and its 956 .MHz second harmonic.
The cavity resonator is attached to the RF plug in oscillator via
the short circuit input coil. The transmit antenna is attached to
the output probe of the cavity resonator. The receive antenna,
connected t~ the signal analyser, was placed at right angles to
the transmit antenna. This was to receive maximum isolation from
the planar travelling wave from the transmit antenna, and hence
any residual second harmonic component. The distance between the antennas was set at 20 cm.
An HP 5802-2800 schottky mixer diode and a 12 turn choke were
placed across the feedpoints of the cross dipole antenna. It was then placed between the transmit and receive antennas.
32
Figure 24. Frequency multiplication test setup
To operated at
'The fact that
froID. 478 MHz, RF plug
frequency is 22 was tuned to this MHz short
resonator
expected antennas
have wide enough bandwidths to accept both the 478 MHz and its 956 MHz second harmonic.
The cavity resonator to the RF plug in oscillator via input antenna to
the output probe of the cavity resonator. The receive antenna, connected to signal analyser, was placed at right angles to
antenna. was to the planar travelling wave from the transmit antenna, and hence
second the antennas was set at 20 cm.
An HP 5802-2800 mixer diode and a turn were of cross antenna. It was across the
then placed between transmit and receive antennas.
32
Univers
ityo
ape Tow
n
4.5.1 Results The sweep oscillator was set to a center frequency of 478 MHz, the power level of the RF plug in was set to 19.5 dBm. The signal
analyser indicated that the fundamental frequency (478 MHz) had
an amplitude of -30 dB.
It was found that the second harmonic rose from -50 dBm to -20 dBm as the cross dipole antenna was inserted between the transmit
and receive antennas. The -50 dBm signal was that of the second harmonic from the filtered sweep oscillator. Thus the signal returned from the schottky diode was 30 dBm higher.
In a further experiment the choke was removed and an annneter
connected across the feedpoints. The annneter acted as a low impedance hence satisfied the need for a DC path. It was found that the current rose from ~ to 1.5 mA when the cross dipole antenna was placed between the transmit and receive antennas. A
diode turns on hard when it draws more than about 1 mA of rectified current. This thus confirms that the schottky diode has
switched on, and is generating harmonic components.
4.5.2 Conclusions From the results it can be seen that the frequency multiplication system operates satisfactorily. The diode switched
provided a significantly large second harmonfc to be the receive system.
4.5.3 Recononendations
on hard and detected by
It has been proven that the 478 MHz/956 MHz frequency
multiplication system works as anticipated. However, to fit into the density tracer, the antenna would require a smaller diameter
of (32 mm), and hence a higher frequency than the 500 MHz/1 GHz system. Such a system could use spiral antennas. However, those obtained from DAST, which operated from 2 GHz to 18 GHz, were
33
was set to a center of 478 MHz, the power level of the RF plug in was set to 19.5 dBm. The signal analyser indicated fundamental (478 MHz) had an -30 dB.
It was found that dam as the cross
second harmonic rose from -50 dBm to antenna was
antennas. The -50 dam signal was that the second from the sweep oscillator. Thus the signal
the was
In a further connected across the
the choke was and an ammeter The ammeter as a low
impedance hence sat is the need for a DC path. It was found that current rose from Q to 1.5 rnA cross dipole antenna was the antennas. A
diode turns on hard when it draws more than about 1 mA of
4.5.2
current. on, and.
thus that the harmonic components.
From it can seen that the system operates satisfactorily. The
multiplication switched on hard
a to the receive system.
that the 478 MHz/956 MHz frequency multiplication works as anticipated. However, to into the tracer, antenna a diameter
~~~), hence a higher frequency than 'T~T~~. Such a could use antennas.
obtained from DAST, which from 2 GHz to 18 GHz, were
Univers
ityo
ape Tow
n
still too large (45 nun) to use in the system.
The alternative was to scale down the cross dipole antenna used in the 500 MHz/I GHz system, to a diameter of 32 nun.
A cross dipole antenna of 120 nun diameter was constructed, using brass rods~ as for the 500 HHz/1 GHz system. The antenna would not operate in its dual frequency mode and was highly inefficient. The reason for this was that the element lengths were much shorter and the ratio of diameter of rod to element length, was effecting the match of the antenna.
The solution was to load the load the antenna with a dielectric backing, so that it could operate at a lower frequency. Another advantage was that the axial ratio (ratio of diameter to length of the antenna element) could be reduced so that the effect on the antennas match was less. The material used was printed circuit board (PCB).
It was found that although the dielectric constant was not as accurately def lned for this material as RT-Duroid or other
microstrip boards, it could still be used at frequencies up to 6
GHz.
Instead of designing an antenna with a diameter of 32 nun (and hence higher frequency), it was decided to design an antenna at an intermediate frequency of approximately 1 GHz (with the multiplied harmonic at 2 GHz).
The antenna would be etched onto PCB and the match and radiation pattern measured. The reason for this would be to determine if the antenna operated as two distinct, isolated dipoles. If they operated correctly, then the antenna size could be scaled down even further to the limit required by the density tracer counter circuit.
34
still too (45 mm) use in the system.
was to scale down the cross antenna
in the ,. to a diameter 32 mm.
A cross dipole antenna
brass as for the 500 not
inefficient. The reason
were
length, was effecting the
was
, so that
advantage was
the antenna el.emlelllt) could
diameter was , KHz/! GHz system. The antenna would
was highly
was that the element of diameter
antenna. rod to element
load the antenna with a dielectric
at a
ratio (ratio of diameter to length
so on the antennas match was less. The material used was printed
c board ( PCB) •
It was found that although the constant was not as
as or other
microstrip boards, it could still be used at to 6
Instead of designing an antenna with a of 32 mm (and hence higher frequency), was decided to design an antenna at
an intermediate of approximately (with the multiplied harmonic at 2 GHz).
antenna would
pattern measured.
onto PCB
antenna operated correctly, even to
The reason
as two then the
this would be to determine ,
antenna by the
If could be down
tracer counter
Univers
ityo
ape Tow
n
4.6 Testing a cross dipole antenna
A cross dipole antenna was required to indicate its performance
as a dual frequency antenna. Frequencies of 1 and 2 GHz were
chosen. As shown previously, the antenna would be etched onto
PCB.
The simulation package "EESOF Linecalc" was lengths of each of the elements. The
characteristics.
Dielectric constant E= = 4.8 Height of dielectric H = 1.6 non Thickness of copper t = 0.017 non Ratio of copper loss RGO = 0.84
used to obtain the PCB had the following
Loss tangent
Characteristic impedance
= 9*10-4
= 50S2
As can be seen from Appendix A, the lengths and widths of the
lines are as summarised in Table 3.
freq (GHz) length (non) width (non) 1 ' 39.607 2.83
2 19.73 2.81
Table 3. Dimensions of 1/2 GHz cross dipole antenna
Figure 25 shows the positive printed version of this antenna
35
4.6
A cross antenna was to its as a dual frequency antenna. Frequencies of 1 and 2 GHz were
As shown antenna would onto PCB.
"EESOF Q was used to the lengths of each of the elements. The PCB had the following
Dielectric constant = 4.8 of H = 1.6 mm
Thickness of t = 0.017 mm Ratio of copper loss RGO = 0.84 Loss tangent = 9*
=
As can seen and of the lines are as summarised in 3.
3.
25 shows the printed version of this antenna
Univers
ityo
ape Tow
n Figure 25 1/2 GHz cross dipole antenna
The antenna was drawn four times oversize and then photo reduced
to improve accuracy.
4.6.1 Results of tests on 1/2 GHz cross dipole antenna 4.6.1.1 Measuring the return loss
A short length of copper coaxial-tq-SMA line was soldered to the
feedpoints of the antenna. The antenna was then connected up to the test arrangement, as shown in figure 26. The following equipment was used to test the match of the antenna:
HP 8350 B HP 8410 C HP 8746 B HP 3595 B
sweep oscilla.tor network analyser
S-parameter test set RF-plug in
I~ I~
Parame er s1
test cross dipole antenna
Figure 26. Test rig to measure antenna match
Figure 27 shows the return loss (Sll) 2 characteristic for this
36
Figure 25
antenna . was
to
4.6.1
A
feedpoints of the antenna. the test arrangement, as
was to test
HP 8350 B
HP 8410 C network HP 8746 B
HP 95 B RF-plug
I I
I
j
overs photo
to the The antenna was then connected up to
shown in figure 26. The following of the antenna:
analyser test set
in
te~t dipole:
Figure 26. Test rig to measure antenna match
Figure 27 shows the return loss (511)2 characteristic
36
Univers
ityo
ape Tow
n
antenna. As can be seen the antenna is not resonating at exactly 1 and 2 Ghz. The fundamental frequency is 1.43 GHz with the
second harmonic at 2.86 GHz. THe characteristics of this antenna can be seen in Table 4.
1. 43 GHz COUPLE At·n
.. -.. co u 4 ·-··
(1)
(;')
·=· _J
',
z 1 ·-:. .:..
0:::
=· I-.. LLJ 1 6 a:::
l 1
Figure 27. Return loss plot for the 1/2 GHz cross dipole
Center frequency (Ghz) HPBW (MHz) Sll(dB) 1.43 330 -17 2.86 790 -20
Table 4. Loss characteristics for 1/2 GHz cross dipole
There are many factors which influence the resonant frequency of an antenna. It has been shown that the end effects, produced by the ratio of the diameter of the element to its length, cause the antenna to resonate at a lower frequency. However, their would be an increase in the resonant frequency due to the field in the inunediate vicinity of the dipole elements being partly confined within the dielectric of the PCB. The wavelength then becomes:
. - .\ / A. 9 - C>/ Er .• (4.3)
37
antenna. As can seen the antenna not resonating at exactly 1 and 2 1.43 GHz with the second harmonic at 2.86 GHz. THe antenna can seen in
"-'. co "tl ',-,'
(I:' (;)
1=' i -l
I "
z Ct::
I =, l-
I ,IJ.I I Q;:
l Figure 27.
1.43 2.86
4.
4
.;.:. ,~.
12
1 6
4.
1 .43 GHz OUPLE ANT
1 .6 l·q.'3.'kl Fr
--------~~~~.~---------~--------~
Sll(dB) 330 -17 790 -20
There are many factors which influence the resonant frequency of an antenna. It has been shown that the , produced the ratio of the diameter of the length, cause antenna to resonate at a . However, would be an in resonant frequency due to field in the
the being within the of the PCB. The wavelength then becomes:
.• (4.3)
Univers
ityo
ape Tow
n
where A.g \<> E;
:r
= =
wavelength through microstrip
free space wavelength
effective dielectric constant of the PCB dielectric and air combination surrounding the dipoles.
This factor is, however, taken into account by the .simulation package LINECALC.
The final factor which influences the resonant frequency of the
antenna is the gap between the feedpoints. This was chosen at 0.5 nun, so as to acconunodate miniature chip mixer diodes. This gap
causes a certain amount of coupling between the two elements. The
result is a distortion of the electric field around the feedpoint (where the electric field has a predominantly reactive
component). This could explain the consequently higher resonant frequency.
4.6.1.2 Conclusions
The return loss plot shows that the antenna is operating at a higher fundamental frequency than initially designed. To ensure
the antenna operated at the desired frequencies, the element lengths were increased.
4.6.1.3 ~easuring the radiation pattern To determine if the two dipoles were acting independently of one
another, it was important to see if they produced satisfactory E and H radiation patterns at both the resonant frequencies. The antenna was sent to the Microwaves and Antennas Laboratory at the Division of Aeronautics and System Technology (DAST) of the
Council for Scientific and Industrial Research (CSIR). The patterns were obtained using DAST's 1-4 GHz anechoic chamber~ Figure 28 shows the anechoic chamber sef:.up.
38
where Ag = wavelength through microstrip Ao wavelength
constant of the PCB and combination surrounding the dipoles.
factor is, however, package LINECALC.
into account by
=~~n~ which resonant
antenna is the gap between the feedpoints. This was chosen at the
mm, so as to mixer This gap causes a certain amount of coupling result a distortion of the around the feedpoint
( has a component). This could explain the consequently higher resonant
4.6.1.2 The return higher fundamental
shows that the antenna operating at a . To ensure
the antenna operated at the desired frequencies, the element lengths were
To one another, it was important to see if they produced satisfactory E and H at both resonant . The antenna was sent to the Microwaves Antennas Laboratory at the Division of Aeronautics and Technology (DAST) of Council and Industrial (CSIR). The
were DAST's 28 shows the anechoic chamber se~up.
38
Univers
ityo
ape Tow
n
antenna <Rx) I-;;:-~ S.S. di p:>O 1 e
horn
Figure 28. 1-4 GHz anechoic chamber setup
The E field plots at 1.43 and 2.86 GHz, which Appendix B, both show the distinctive c 12 1
pattern. The 10 dB beamwidth appears to be 90~ at
difficult to determine the beamwidth at 1.43 be that the horizontal
can be found in "figure eight" 2.86 GHz. It is
GHz. A possible portion of the reason for this could
coaxial-to-SKA cable interferes with the electric field. This is the horizontal shown in figure 29, where
portion of the cable. position A indicates
The H planes at both 1.43 and 2.86 GHz, are omnidirectional. This was expected from a short dipole. The H-plane plots can also be found in Appendix B.
Coa><. cable
Figure 29. Cross dipole antenna showing coax feed cable
The slight oscillation in the H-plane plot at 2.86 GHz is caused by the vertical section of the coaxial cable. Position B in figure 29.
39
28.
at 1.43 2.86 GHz, can found eight" Appendix B, show the distinctive [12]
to at 2.86
to the beamwidth at 1.43 GHz.A poss
reason for this could be the horizontal portion of the the
shown in figure 29, where position A indicates the horizontal
of
The H planes at both 1.43 and 2.86 GHz, are omnidirectional. This was from a short dipole. plots can also be found in Appendix B.
29. Cross dipole antenna showing coax feed cable
2.86 GHz is caused
by vertical Pos B
figure 29.
Univers
ityo
ape Tow
n
A solution to obtaining more accurate H plane plots is to place a reflective
plate a distance l behind the antenna, covering the vertical portion of the coaxial
cable. For the cross dipole antenna, the distance I was chosen as:
= 1'/8 at 1.43 GHz
= >../4 at 2.86 GHz
Figure 30 shows the cross dipole antenna with its reflective plate.
Reflective plate
Coax. cable
L
Figure 30. Cross dipole antenna with reflective plate
It must be noted that a balun was not used to match the antenna to the
feedline. This was because the antenna was operating at two frequencies, and it
is not possible to match an antenna to two frequencies using quarter wavelength
sections. A quarter wavelength section at 1.43 GHz transforms to half wavelength
at 2.86 GHz.
4.6.2 Conclusions
The radiation plots show that both dipoles are acting independently of each
other. The E and H fields are exactly what
40
·'
A· sol to 0 nlng more aCCurate H plane plots to place a reflective
plate a distance I behind the ante cove the vertical portion of the coaxial
For the cross dipole antenna, the distance I was chosen as:
== "/8 at 1.43 GHz
== >-/4 at GHz
Figure 30 shows cross dipole antenna with
Ref
. cable
reflect plate.
It must be noted that a balun was not used to match the antenna to· the
fe line. This was because the antenna was rating at two frequencies, it
not possible to match an antenna to two using quarter wavelength
. A quarter wavelength section at 1.43 GHz transforms to half wavelength
at 2.86 GHz.
The radiation plots show both dipoles are acting independently of each
other. The and H are exactly what
40
Univers
ityo
ape Tow
n
would be expected from a traditional short wire dipole. It must be noted that the cross dipole used in practice will not have the
short section of coaxial cable to interfere with the radiation
patterns. In conclusion then~ the cross dipole antenna satisfies the criterion for the frequency multiplication method.
41
would be expected from a short be cross short section coaxial. cable to
dipole. It must will not have
with the radiation patterns. In conclusion then~ cross dipole antenna fies the the multiplication
41
Univers
ityo
ape Tow
n
CHAPTER 5
DESIGNING THE ANTENNAS FOR THE FREQUENCY
MULTIPLICATION SYSTEM
5.1 Design of the cross dipole antenna The frequencies of the system are ultimately determined by the
size of the density tracer. From section 4.6 it can be seen that a cross dipole antenna mounted on PCB, with a maximum diameter of 80 nun, will resonate at a fundamental frequency of 1.43 GHz. By
simply scaling the dipole down in size it is possible to produce
an antenna which has the correct diameter as specified. The antenna will now however be radiating at a higher fundamental and
second harmonic frequency.
It was found that by reducing the antenna by a factor of two will
produce a density tracer of diameter: 80/2 = 40 nun. The frequencies of operation should then be 2.86 GHz and 5.72 GHz.
If 3 GHz and 6 GHz are chosen as the frequencies of operation of
the antenna, the scaling down factor would be 2/1.43 = 2.09. The maximum diameter of the density tracer would be approximately 40 nun (includi~g feedpoint gap). This is larger than that specified (32 nun), which would have implied frequencies of 3.58 and 7.16 GHz for the antenna. After discussion with DRL it was decided that ~ and 6 GHz be chosen as the frequencies of the system, merely because of their ease of use components are available which operate specifically at these frequencies.
Working on these principles then, an antenna was designed by reducing the lengths by the factor 2.09. Figure 31 indicates the positive print of this antenna after it was photo reduced by
twenty five percent.
42
5.1 Design of the cross dipole antenna The are size of the density tracer. From section 4.6
by can be seen that
a cross dipole antenna on PCB, with a maximum of will resonate at a fundamental frequency of . By
simply scaling dipole down it possible to produce an antenna correct as antenna will now however be radiating at a higher fundamental and second harmonic
It was that by antenna by a two . a tracer of :
frequencies of operation should then be ~2~.~8~6~~~
If and of operation of the antenna,
are chosen as the frequencies scaling down factor would be -..... ......... ~"-----' ................ ~ The
tracer would approximately 40 mm (includi~g feedpoint gap). This larger than that specified (mm), would have of GHz for the antenna. After discussion with DRL
as the of the system, of their ease of use components are
specifically at
an antenna was by reducing the lengths by the 31 indicates
antenna it was photo reduced by twenty
42
Univers
ityo
ape Tow
n
Figure 31. Positive print of 3/6 GHz cross dipole antenna
Figure 32 (a) and (b) indicates
both bands of operation. 3 / 6 GHz ~)
the match of this antenna over
.. -.
.. _ ..
z Ct:
=· 1-UJ
' 0:::-.
COUPLED At·H
..------·---·---- ('! ·- ............ ----.::: •j ·· ..... ·--.. __ .··'
···· .. ,\ •... -·····J
... ...---J'·' 1,\ l
'• / \/
3 (1
I I I ~~t--K_'!.-t-1 -~l--11 .06.···div 3.2
F r;;o3qe_rio::::--_·_£_1~Hz ) __ _
l .J
. I
~igure 32(a). Return loss of cross dipole antenna centered on 3
.......
.. ..,,·
(),) (Jl 1::)
..J
iz Ct: ~ 1-IJ.J Ct:.
--;:-:--~ ··· .:1 ·' ,- H - C IJ- U P r --.E D A N T , r .'"" ........ "'-.· -
---------.... -....... l A.. ,._
. ,2.2._ qt.lg . . c '·-...: .... - .. .... ·~ ..... , .... .. ' .... , .... \ r
\- .,.--.. j \ / \ .... ./ \ i
30
40 I,• I/,.
. ! ., (OGM~·· 5 t.:.l -~- ;;'.-1 +-I --+-----,..¥-r-:_;..;;.,... ~t--'· ·-...~.-..·.__,__+----11-...,.-o.,
5 . 7· . ~/ d'i "' 6 . 3 _f r__e~~~n c: >'_(GHz)
Figure 32(b). Return loss of cross dipole antenna centered on 6
43
Figure 31.
Figure 32 (a) and (b) indicates the match of this antenna over both bands :3 6 z A t·l T
~j
..... , 00
(1 7J 1 .......
\"1..A.B (I)
20 (;').
'::0 .....J
Z :3 ~)
Ct:: .. ~ =, I- 40 w ~.
5 ~3 2 6 :3 Z
32(a) . ./·~··T·l
.,-.. ~ co "0 .... ,"
'2~ .....J
i Z 30 a:: :=. I- 4 (1 lI./ a::.
(b) •
43
Univers
ityo
ape Tow
n
As can be seen the antennas are well matched at 3 GHz and 6 GHz with return losses of 17 dB and 22 dB respectively. It can be
seen that the antennas have better return losses at~ and 6.16
GHz. It was found however that these troughs were susceptible to loading by objects placed in front of the antenna. Standing waves would be formed between the antenna and the object. The retu~n loss correspondingly rose from 20 to 40 dB. The resonant
frequencies at 3 GHz and 6 GHz remained more constant. It can thus be accepted that the antenna shown in figure 30 is a suitable cross dipole antenna at 3&6 GHz.
/
5.2 Designing a suitable DC biasing choke It has been shown earlier (sect. 4.3.1) that a choke is important to provide a DC path for the mixer diode. It is also used as a self biasing unit, giving the diode a reference bias at zero volts.
It must be noted that the chokes are self resonant at certain frequencies , and that their reactance increases with the frequency. The relationship between reactance and frequency is linear, as can be seen in eq. 5.1
where
Z1 = 2*n*f*L ... (5.1)
Z1 = reactance of choke in f = frequency in Hz L = inductance of choke in H
An ideal choke must be passive and have no affect on the match of the system, other than to provide a DC path. As we have seen however, this is not entirely possible, since reactance increases with frequency.
The test bed in figure 33 shows how chokes with various turns can
44
As can be seen
return
antennas are well matched at 3 GHz and of . It can be
seen the antennas return losses at . It was found that these troughs were susceptible to
by antenna. waves would be formed antenna and object. The return
correspondingly rose from 20 to 40 dB. The resonant
at more constant. It can thus accepted that the antenna shown 30 is a
suitable cross dipole antenna at
5.2
It to provide a DC path
f , volts.
It must be and
.......... ...., ...... .::;. It
a
choke is important also used as a
at zero
are resonant at certain
reactance increases with
frequency. The relationship between reactance and
linear, as
where
An ideal
. 5.1
... (5.1)
choke Hz
H
and have no
the system, other than to provide a DC on the match of
• As we have seen
however, this not , reactance with
test in 33 shows how chokes with various turns can
Univers
ityo
ape Tow
n
be tested to see their affect the on the match of a system. A SQQ line is etched onto microstrip, using the physical parameters from table S. A patch of copper is etched and shorted to the ground plate via a screw. Two SHA launchers are then soldered onto the board and the test bed is connected to a network analyser. The return and insertion losses over a_ range of frequencies, for each choke, is measured.
Figure 33. Choke reactance test bed
The principle is that as a choke becomes series resonant it will
be a low impedance anq reflect all of the energy incident upon on the SOQ line. There will thus be a decrease in the return loss and in the transmitted power (S12) 2 absorbed by the ~ixer.
The effect of a 5 turn choke on the return loss is shown in figure 34 (a) and (b). 32 SWG copper wire was used in this experiment.
0 W 1-n-t C HOK:E
lo
Return loss
in dB
100 1000
Figure 34(a). Return loss of Sturn choke over 100 .MHz-1 GHz band
45
to see on match a . A etched onto , us
5. A patch copper is etched and shorted to the ground plate onto the
a screw. Two SMA launchers are then soldered and the test bed to a
return and insertion over a range of , for choke, is
Figure 33.
The principle that as a choke becomes series resonant will
be a impedance anq of incident upon on
a the return and in the transmitted power (S12)2 absorbed by the ~ixer.
The of a 5 turn
figure 34 (a) and (b).
." Return lo'S'S
in dB
(a) •
100
on the return loss is shown in in this 32 SWG wire was
CHoc..E
100:;.
45
Univers
ityo
ape Tow
n
0
201====::::::
Return 1 oss
in dB
~----------------------------- -f (61-<~) 3
Figure 34(b). Return loss of 5 turn choke over 3 GHz - 6 GHz band
It can be seen that the choke is too low a reactive impedance and
is reflecting (X 1 = 2*rr*f*l) most of the energy over the band 100 Mhz-1 GHz, but is a much higher inductive reactance, thus
reflecting very little power over the range 3-6 GHz. Very little power is being reflected back to port 1.
5.2.1 Conclusions
A 5 turn choke using 32 SWG copper wire will be mounted across
the mixer diode, since it will have little affect on the match of the cross dipole antenna.
5.3 Choosing a suitable mixer diode for the cross dipole
antenna
The mixer diode used on the cross dipole has to conform to
certain specifications. It must be a schottky barrier diode. The
biasing voltage must be zero (no onboard power supplies). The
device should be packaged as a chip, i.e. without leads and plastic containment, since these will cause capacitive parasitic loading at high frequencies. The chip must be small enough to fit across the feedpoints of the cross dipole antenna. Chip diodes are cheaper than encapsulated diodes. However bonding facilities must be available in order to mount the chip device into the
density tracer.
The device chosen was an Alpha DHK 5068 Gallium Arsenide Schottky
46
o
10';;::=====
Rei:urn lo'Ss
in dB
(b) •
It seen choke too low a reactive and 2*«*f*1) most of the energy over the band 100
a much reactance, 1 power over the
is being reflected back to port 1.
5.2.1 Conclusions A 5 turn SWG will across
mixer diode, since have little affect on the match of the cross antenna.
5.3
The mixer diode used on the cross dipole has to conform to certain It must a biasing voltage must be zero (no onboard power The
as a , i.e. and plastic containment, since these will cause capacitive parasitic loading at high . The chip must be enough to fit across the feedpoints the cross dipole antenna. are than ......... v, .. 'C'o:>. However bonding must be available in order to mount the chip device into
tracer.
The chosen was an Alpha DHK 5068 Gallium Arsenide Schottky
46
Univers
ityo
ape Tow
n
barrier mixer diode. The characteristics of this diode can be
found in Appendix c. Although the diode operates best over the Ku band (12.4-18 GHz), it will still operate successfully at 6GHz.
For convenience the measurement reported here were performed using a packaged mixer diode, the package has dimensions of 1.4 nun by 1.27 nun. The gold contacts were 0.8 nun apart thus the unit will thus fit in perfectly between the f eedpoints (0.5 mm separation).
5.4 Designing the 3&6 GHz transmit & receive antennas As shown previously, the system will be operating with a 3 GHz
fundamental frequency and a 6 GHz second harmonic, thus 3 GHz
transmit and 6 GHz receive antennas are needed.
Metal fabricated dipoles at these frequencies would be very
short, especially at 6 GHz where the antenna length 1 would be
less than 3*10°/6*109/2 = 25 nun. A horn antenna would be a useful
substitute. They have wide bandwidths and high gains, as required by the system.
5.4.1 The 3 Ghz horn antenna & waveguide-coaxial transformer
The frequency multiplication system requires horn antennas which have a high gain and a wide beamwidth GO~.
It has been shown c 13 1 that there exist empirical formulas to
calculate the 10 dB width as a function of aperture, for the average horn antenna. These are:
1) For the electric field:
0E(l0) = 88*A/B degrees Bl.A < 2.5
2) For the magnetic field:
0u(l0) = 31 + 79*A/A degrees A/A <3
47
.. (5.2)
.. (5.3)
/
diode. The characteristics of this can found in Appendix C. ~though the diode best over the Ku band (12.4 GHz), will success at 6GHz. For convenience
a
the measurement reported
package mm by 1.27 Mm. The gold contacts were 0.8 mm
dimensions of 1.4
thus the will thus perfectly between the feedpoints (0.5 mm
5.4 As shown previously, a
fundamental frequency and a 6 harmonic, thus 3 GHz
6 GHz antennas are needed.
Metal fabricated at these would
short, where the antenna 1 would less than . A horn antenna would be a
substitute. They have wide bandwidths and high gains, as required
by
5.4.1
The frequency multiplication requires a high gain and a wide beamwidth 600 •
It has been shown [13] there exist empirical formulas to 10 dB as a of
average horn antenna. These are:
1) For the field:
B/A < 2.5 .. (5.2)
2) For the magnetic field:
A/>" <3 .• (5.3)
47
/
Univers
ityo
ape Tow
n
B,A are the apertures in the electric and magnetic fields.
The system requires a 60<> in the electric field, thus from eq(s2
), the horn antenna.will require an aperture of:
Now the free space wavelength at 3 GHz is:
A<>(3 GHz) = elf = 3*10e/3*l09
= 100 mm
Therefore the aperture is:
B = 88*100/60 = 147 mm
•• (5.4)
•. (5.5)
.. (5.6)
To simplify construction an aperture of 150 mm was used. It was decided that square pyramidal horns would be used, thus the
beamwidth in the magnetic field would be:
Ou(lO) = 31 + J9*A/A = 31 + 79*100/150 = 84<>
•• (5. 7)
It has been noticed antenna is related
in a previous work that the beamwidth of an to the aperture of the antenna, and as
expected as the aperture increases in size, so the beamwidth narrows. Looking at this another way, as the · horn length of an
antenna with a given flare angle is increased, the aperture increases. This would cause the beam to narrow. It was found however that the effect of increasing the aperture is overshadowed by the phase error effects of the slant length of the horn. This phase error had to be determined and taken into
48
B,A are
, thus from (52 The system requires a 600
), the horn antenna will an aperture of:
Now the
Therefore
at 3 GHz
(3 GHz) elf == 3*10 0 /3*10 9
aperture
B = 88*100/60
,. .
.. (5.4)
::
.. (5.5)
.. (5.6)
To simplify construction an aperture of 150 mm was used. It was decided square horns would thus the
field would be:
9H (10) 31 + 79*A/A
It has antenna is
= 31 + 79*100/150 =
noticed in a to
expected as the aperture
.. (5.7)
work that the beamwidth of an the antenna, and as
in size, so the beamwidth narrows. Looking at another way, as the' horn of an antenna with a given flare angle is increased, the aperture
however that by
the horn. This
would cause to narrow. It was found effect of
phase error error
48
to be
ing of the
the is slant length of
Univers
ityo
ape Tow
n
account for the design of the horn antenna.
Figure 35 shows a section of a horn antenna. d represents the
maximum departure of the wavefront r 2 from the aperture plane.
Figure 35 . Section of horn antenna with wave front
The phase difference between the center of the aperture and the
edge, is given by 2*rr*5/A 9 • From figure 35 it can be seen that:
r 2 *cos(00 ) + 5 = r 2 .• (5.9)
therefore 5 r 2 (l-cos(00 ))
and
When the flare angle is small, o/ .>... 9 is small and the wavefront
approximates a plane wave. The field is thus more evenly spread across the aperture of the horn antenna. It has been shown that a suitable upper limit for 5/ A 9 is l..L!!.
A conservative estimate of 5/ A 9 = .lL2 for the phase error was assumed. This phase error ·is shown in figure 36.
/
/2
'--· Figure 36. Phase error of antenna along the slant length
Now .A 0 /6 = 100/6 - 16. 67 mm
and A/2 = 150/2 = 75 mm From Pythagoras then:
(x +A0 /6) 2 = x 2 +75 2 .. (5.10)
49
account design the horn
Figure 35 shows a section a antenna. d represents maximum departure of the wavefront r 2
Figure 35 *
The phase difference the center of the
edge, given by 2*rr*5/Ag* From figure it can
*cos(Oc::» + 5 therefore 5 ( I-cos ( 0 C» )
and 5 I >- 9 r21 )..g*(I-cos(Oc.>))
plane.
and the seen that:
•• (5.9)
When the angle is small, 51 A 9 small and the a plane wave. more
across the aperture of the horn antenna. It has been shown that a
limit 51 9 lLR.
A of 51 A 9 :: error was assumed. This phase error 'is shown in figure 36.
/
/2
Figure 36. Phase error of antenna along the slant length
Now).. = 100/6 = and A/2 = From Pythagoras then:
(x 6)2 x2 +7 .• (5.10)
49
Univers
ityo
ape Tow
n
x 2 + 33.34*x +277.89 = x 2 + 75 2
x 160.38 mm
The square pyramid will be constructed using four brass plates
brazed onto a circular flange. The dimensions of this flange can
be found in Appendix D. The flange is designed to work with a
system operating between 2.60-3.95 GHz. Figure 37 shows the flange. H3
- -0- ....... G 12.14 mm 'Q
lil'i20.6S l1W'll
' \ 72.14 nm
G I !6140 mm
I
/J
Figure 37. Dimensions of 3 GHz circular flange
As can be seen, the base of the pyramid has a width of 7.214 cm.
The template from which the pyramid is made, is shown in figure 38. . ! ._ .................... =1:§R-'-~~--~----·-·····-···--f T I
I
__ ,
. :iso mm
' 1 Figure 38. Side plate of 3 GHz pyramid horn
The length C is given by similar triangles:
75/160.38 36.07/C
C = 77.132 mm
From this we obtain:
so
.34*x 77.89 == X2 + 75 2
x
pyramid will be four brass onto a circular flange. The dimensions of can
found Appendix D. igned to work with a
system operating flange.
between H3
_ -0- ""-12.14 111m
37.
As can of the
38.
c
7 160.38 36.07/C
.65 111m . \ 12.14 111m
Q I !!S140 mm
a made,
I
.. .
C = 77.132 mm
From we obtain:
50
of shown
shows
Univers
ityo
ape Tow
n
D = 160.38-77.132
= 83.25 mm
Four plates of these dimensions are soldered to the flange in
figure 38. A system of launching waves from the coaxial feed cable to the 3 GHz horn was required. A waveguide-coaxial
transformer was constructed. From Appendix D, it was found that
the transformer would have dimensions as indicated in figure 39.
72.14
34.04 mm! >---·--·-·- .................................. . i __I
Figure 39. 3 GHz waveguide/coaxial transformer
Appendix D, it is found that the cut off frequency is:
fc: = 2.08 GHz
, and that the transformer waveguide
dimensions:
72.14 mm * 34.04 mm Free space wavelength is from (5.5)
A. 0 = c/f
= 3*10°/3*10 9
= 100 mm The wavelength of the cut-off frequency is:
,\ c: c/fc
3*10°/2.08*10 9
= 144.231 mm
The waveguide wavelength is given by:
51
has the following
.. (5.11)
D = 160. 7.132
Four plates of these dimensions are soldered to the flange in 38.· A launching waves the coaxial feed
cable to the 3 GHz horn was required. A waveguide-coaxial ~o.rmer was . From Appendix D, was
the transformer would have dimensions as indicated in 39.
72.14
. 1 L'--______ .Y
34.04 I
J
Appendix D, it is found that the cut off
, and
= 2.08 GHz
that the . .
transformer waveguide
The waveguide wavelength is given by:
51
is:
has the following
..(5.11)
Univers
ityo
ape Tow
n
l/A. 2 g >.. 2 g
Ag
The point of the 'Ag/4
The probe, shown >.. /4 <>
= =
l/A.<>2 - l/)\,c2
19257.047
138.77 mm
.. (5.12)
probe from the rear of the waveguide ts: = 34.693 mm
in figure 38, protrudes a length:
= 25 mm
into the transformer.
A transformer of these dimensions was constructed and soldered to
a circular flange, as in figure 36. The transformer was bolted to
the horn antenna, and then tested in the same manner as that in
section 4.6.1.1
5.4.2 Testing the 3 GHz horn antenna & waveguide-coax transformer The 3 GHz horn antenna was connected to port 1 of the network
analyser. The return loss plot, shown in figure 40, indicates the
match of the antenna at 3 GHz. As can be seen, the antenna has a
loss of 11.3 dB at 3 GHz.
3 GHz HOrU·J 6
.. -..
. ........ --·-
10 / ,_ ............ -·
I 2 .·· 11·1 OJI /,.,f,,.-···T-·---·~-·-
1 4 -·--··--.. --··-----· I
l 6 -+---+--+---r---1-c.-+~-'----+-·--+---i 2.95 01/div 3.05
I
I
--------'-F~r_e ·:t '-l e_n c ::: .. __ LG H~z~) ____ _
Figure 40. Return loss for 3 GHz horn antenna
The low efficiency of the antenna could be due to:
52
1 .. (5.
A = 192 .047
"'9 The point of the probe from the rear of the waveguide is:
\
The probe, shown
AcJ4 =
into the transformer.
a length:
)
A transl:ormer of these dimensions was constructed and soldered to
a , as 36. trans was to
the horn antenna, and then tested in the same manner as in
4.6.1.1
5.4.2 3 GHz horn antenna was connected to port 1 of the network
The return loss plot, shown in 40,
of antenna at 3 GHz. As can be seen, the antenna has a
loss of at 3 GHz.
-----_._- --------~--------------,
:3 GHz HOF.~t·~ 6
.'-',
m 8 7J
((, 1 ~~1 ((I
1=' \\·1 ~B ...J
z 1 ..... ,,;.
~ =, ...... 1 4 w ~
:2 , 95 .L]l .. ··oji, ... ' _______ ----'---'-.,e ."" tl 0;;0 n ': ),·_ .. ,"-'''''-'.L''''-'' ____ ----J
Figure 40. Return loss for 3 GHz horn antenna
The low the antenna could be due to:
52
Univers
ityo
ape Tow
n
a) Probe not in maximum position to intercept maximum electric
field.
b) Plate added to waveguide wall causes discontinuity in electric
field.
5.4.3 Conclusions
The return loss plot indicates that the 3 GHz horn antenna will
operate adequately in the system.
5.4.4 Designing the 6 GHz horn antenna & waveguide-coaxial
·transformer
The 6 GHz receive system requires a 6 GHz horn with a waveguide
coaxial transformer, similar to the 3 GHz unit. The dimensions
for the antenna and its transformer are calculated below.
The beamwidth of the 6 GHz horn antenna was chosen to be 60° ,
the same as the 3 GHz horn antenna. The aperture of the antenna
was determined as follows:
The free space wavelength at 6 GHz is obtained from eq (5.5)
A. 0 = c/f
= 3*108 /6*10 9
= 50 mm
The aperture is then determined using eq (5.6).
B = 88*A/Oi;:(l0)
= 88*50/60
= 73.33 mm
For convenience an aperture of 75 mm was chosen.
A phase error of A0 /6 along the slant length of the pyramid, is
assumed, as shown in figure 36.
Now A0 /6 = 50/6
53
a) not in maximum position to maximum
b) added to waveguide wall causes in
5.4.3
plot indicates that the 3 GHz horn antenna
5.4.4
6 a 6 GHz a
coaxial transformer, similar to the 3 GHz unit. The dimensions
for antenna are calculated below.
The beamwidth of the 6 GHz horn antenna was chosen to 60<=> , antenna same as 3 GHz antenna.
was as follows:
The at 6 GHz
= c/f
= 108 /6*10 9
=
The is then determined using eq (5.6).
B = 88*A/O:s:(10) = 88*50/60 =
For convenience an aperture of 75 mm was chosen.
A error A '0/6 of
assumed, as shown 36.
Now Ac:>/6 = 50/6
of
(5.5)
Dvramid, is
Univers
ityo
ape Tow
n
= 8.333 mm and A is chosen to be 75 mm, this implies that A/2 = 37.5 mm
x can now be determined using Pythagoras as in eq (5.11)
( x + A C> / 6 ) 2 = x 2 + (A/ 2 ) 2 •• ( 5 • 13 )
(x + 8.333) 2 = x 2 + 37.5 2
x 2 + 16.67*x +69.44 = x 2 + 1406.25
x = 80.192 mm
It was then decided to build a rectangular pyramidal horn using
the dimensions to fit the base of the 6 GHz. This would entail
constructing two types of side plates for the horn.
From Appendix D, the dimensions of the waveguide are found to be:
40.39 mm * 20.19 mm The circular flanges to be used on the 6 GHz horn and
transformer, are shown in figure 41.
~ - 8- -
' / ..... _-@_,,,.,,,.
Figure 41. Dimensions of 6 GHz circular flange
The flange is designed to operate over the band 4.90-7.05 GHz.
From figure 41 it can be seen that two sides of the pyramidal base have dimensions of 40.39 mm while the other two sides have
widths of 20.19 mm. The two side plates are shown in figure 42. The dimensions of the plate are calculated thereafter.
54
A is chosen to be implies A/2 x can now be determined using as in eq (5.11)
(x + A c>/6) 2 X2 + {A/2)2 . . ( 5 . )
(x + 8.333)2 = x 2 + 37.52
X2 + 16.67*x +69.44 = x 2 + 1406. x = 80.192 mm
It was to build a us the dimensions to the base of the 6 GHz. This would entail
two the
From Appendix D, the of the waveguide are found to be:
The circular flanges to be used on the 6 GHz horn and trans
Figure 41.
The
From base
are
designed to operate over the band
41 it can seen that two dimensions of 40.39 mm while the other two sides have
widths of 20.19 mm. The two are shown 42. of are calculated
54
Univers
ityo
ape Tow
n
t-·-·-·············-~-9.!.~-~-? ... ~---····-------·-·--I
----- r
Plate 1
I
i
~0.39 mrn i
-------I ._ .... G.=~-+-----------------JL ________________ ,
t------·-·····------~-9.!.~-~-? ... ~-------------------1 T '
-·'
' ~7S 11W1
1
i ........ t::: ..... _... ..•......... '. ..... f..: .............. --1
Plate 2
Figure 42. Side plates for the 6 GHz pyramidal horn
From the diagrams: 37.5/80.192 =20.195/C
thus C = 43.186 mm
and D 80.192-43.186
= 37.006 mm
This gives plate 1 a slant length of:
11 = v(37.5-20.195) 2 + 31.006 2 '
= 40.852 mm
Plate 2 will also have to have this slant length, to ensure that
the side plates fit flush against each other. The length of this plate can then be calculated using this criterion.
55
1
2
42.
From the
thus
.5/80.192
C = D
5/c
37.006 mm
~I
This gives plate 1 a slant length of:
11, =
2 have to slant ,. to the side plates flush against each other. The length of this
can
Univers
ityo
ape Tow
n
F = v(40.852) 2
30.296 mm (37.5 10. 095) 2 '
The horn antenna was then constructed using these plates and
soldered to the circular flange shown in figure 41.
A unit to launch waves into the horn ~as required. A suitable
launcher is the quarter wavelength waveguide-coaxial transformer.
The dimensions of the launcher are determined as follows.
From Appendix D we find that the flange dimensions are:
40.39 mm * 20.19 mm The waveguide is designed to operate over the band 4.90-7.05 GHz,
with a cut-off frequency of 3.71 GHz.
Now the free space wavelength is: A~ c/f = 3*10e/6*l09 .. (5.14)
= 50 mm The cut off frequency wavelength is: Ac= c/fc = 3*10°/3.71*109
80.86 mm From this the waveguide wavelength can be calculated.
1/ ,\ 92 = l/A~2 - l/Ac 2 •• ( 5 .15)
A 92 = 4047.665
>.. 9 = 63.62 mm
The coordinates of the.probe position are given by: A9 /4 = 63.62/4
= 15.905 mm
>-~t4 = 50/4 = 12.5 mm
The completed transformer is shown in figure 43.
56
F
horn antenna was then constructed using these plates and soldered to the shown in figure 41.
A unit to launch waves into horn was A launcher is the quarter wavelength waveguide-coaxial trans
of
From Appendix D we find
a cut-off
Now the free space wavelength is: A~
cut .. .
are:
over band
elf 3*10 0 /6*10 9 •• (5.14)
100 /3.71*1
From the waveguide wavelength can be calculated. 11 A == 11).. •• (5. )
A == 4047.665 A9' = 63.62 mm
The of ,probe are by:
A9'/4 = 63.62/4
The completed trans 43.
Univers
ityo
ape Tow
n
.... -::f. .
4
40 • 3'9 mm ./::_j__ ___ O·_'.prob&---·-----·---/ f ~ ,·1
I ,.•----·•·-------·-·•··········· 20.1'9 mm; j_/
Figure 43. 6 GHz horn & waveguide-coaxial transformer
5.4.5 Testing the 6 GHz horn antenna & transformer The 6 GHz horn antenna and transformer are attached to port 1 of
the HP 8410 B network analyser. The return loss for the antenna was determined and then plotted. The results of the test can be seen in figure 44.
1; c; H :z H Ct P t·4 1 1 ~· ..
:u t:'
'] 1 1 -I
-
···.
(!) 1. 1 -~
((I I
c: I ll· ts .A.I! •.. J
-~ 1 1 . 9 . ..:....
I .·, ·· ,,: 11.. c:LB I-·- 1 :-:'. . l
LLJ c~:
C, C.M'I ·--+·········-+·-···-+····-···-·+-·--··-t-· ····-··+-···---+··-··--·· + ·- ... · · ; ... ···--··-I
"'; . 9 '5 . (1 1 · (:! i. '..I ;~ ~:1 ":;
___ F_· ~t·~e_·'-l. IA enc::·· ( GHz )
Figure 44(a) & (b). Return loss for 6 GHz horn & waveguide to coaxial transformer
Figure 44(a) shows the cut-off frequency at 3.71 GHz and that the antenna is operating well within the band of 4.90-7.05 GHz.
57
43.
5.4.5 Testing the 6 GHz horn antenna & transformer 6 horn antenna trans are attached to 1
HP 8410 B . The return loss for antenna was determined and plotted. The of the test can be seen figure 44.
6 G H z H I) F: tl 1 1 .~~I
0 'J 1 c:: 1 -' -
(I) 1 1
.~
((I ... 4
C) ... B •• J
z 1 1 9 (}:' =,
1;2 1-" IJJ
1 ~:~ 1 Ct:::
C.C.M'l 1 ~~ '7 •.
.:: .::- :3 "" (J 1 d 1. './ r1 "" .. • .. ' • z Fr.;;.
. G / d i',,. (GH~:!
44(a) & (b).
44{a) shows frequency at 3.71 GHz and antenna operating within the of 4.90-7.05 GHz.
57
Univers
ityo
ape Tow
n
Figure 44(b) indicates that the return loss is approximately 11.85 dB at 6 GHz.
5.4.6 Conclusions Figures 44(a)&(b) indicate that the antenna is operating adequately over the band, and satisfactory for the operation of the frequency multiplication system.
58
44(b) at 6 GHz.
5.4.6
44(a)&(b) indicate band, and
return
that the antenna factory the operation of
58
Univers
ityo
ape Tow
n
CHAPTER 6
CONFIGURATION OF 3/6 GHz FREQUENCY MULTIPLICATION SYSTEM
6.1 Problems with conventional receiver units
The simplified block diagram for the 3/6 GHz frequency multiplication system is shown in Figure 45.
__ :._~ 3 GHz hornJ
~ '-.../
3 GHz cl"'oss osc. di iool e
Figure 45. Block diagram of frequency multiplication system
The problem with this system lies primarily with the receiver
unit. Receivers are generally wide band, and since we expect the 3 GHz signal to drift in frequency due to various physical factors, the second harmonic will not have a constant frequency
either. This complicates detection. If the second harmonic had a stable position in the frequency domain, then a narrow filter could re placed in front of the receiver. This would narrow the bandwidth of the receiver, easing detection.
A solution might be to place a sweep and lock circuit in series with the receiver. This device sweeps a band frequencies and locks onto signals above a certain threshold. The system then locks onto this signal. It is then put through a narrow bandpass filter after which it is sent to a receiver.
Two problems exist with this system:
1) Amplitude of second harmonic
59
CHAPTER 6
6.1
The simplified block diagram for the 3/6
multiplication system is shown in Figure 45.
T6 GiHz horn
l(o EO GHz recei.vef'1
I
Figure 45.
The problem with this system primarily with the
unit. Receivers are generally wide band, and since we expect the
3 GHz to to
factors, harmonic will not have a constant frequency
either. This harmonic had a
stable position
could re placed in
frequency domain, then a narrow filter
would narrow front of the
of the , detection.
A to series
with the receiver. This sweeps a band frequencies and
above a certain threshold. system then
locks onto this signal. It then put through a narrow
which it is sent to a
Two with system:
1) Amplitude of second harmonic
59
Univers
ityo
ape Tow
n
2) Time taken for circuit to sweep
1) The frequency multiplication system obeys a fourth order power law, similar to radar systems. It can thus be expected that the multiplied second harmonic will have a low amplitude. If the
second harmonic level entering is lower in amplitude than the
background noise of the receiver, the circuit will not lock.
finite amount of time to 2) The sweep and rock circuit takes a lock onto a receiver tone. As stated
0.5 seconds. After this period, the earlier, a typical period is density tracer would have
fallen out of the detection range.
One way to produce a stable source of second harmonic at the receiver, is to lock the 3 GHz oscillator. Crystal locking is
expensive, however.
A better solution is to lock the receivers local oscillator to
the second harmonic produced by the transmit oscillator.
The advantages are:
1) The transmitter can be free 2) The sweep and lock circuit is continually locked onto
running on the receivers local oscillator the transmit oscillators second
harmonic. The density tracer, with cross dipole antenna, does not have to be present.
6.2 The receiver local oscillator feedback loop The full frequency multiplication system/density tracer counter unit is shown in figure 46.
60
2) Time taken for circuit to
1) obeys a fourth order power law, similar multiplied
to radar systems. It can be the will have a low amplitude. If the
in amplitude than background , the will not
2) The sweep and circuit takes a amount of to onto a tone. As , a
0.5 seconds. After this period, ity tracer would have out detection
One way to produce a stable source second at , is to lock the 3 GHz Crystal locking
A better solution to lock the local oscillator to second produced the
The advantages are: 1) can be 2) The sweep
continually harmonic. The
circuit on the receivers local oscillator onto the transmit
tracer, with cross dipole antenna, does have to present.
6.2 The receiver local oscillator The tracer counter unit shown in figure 46.
60
Univers
ityo
ape Tow
n (j) p
3 GHz oscillator
6 GHz osci 11 at or
alQD
~
s\Je ep & l ock circuit
GHz
6 GHz directional coup 1 er
otch filter
poYer splitter 6 GHz
,single ended mi.xer
t'i_l ter
POYer corrt>iner 6 GHz
3 GHz horn·
filter
single ended mixer C6 oGHz >_
IF amp. detection unit
//' /
6 GHz horn
3/6 GH:?: cro~~ dipole antenna
& lock t
splitter
e ended
filter
)-.
unit
3/6 GHz c:ro~'!I dipole antenna
Univers
ityo
ape Tow
n
The 3 GHz oscillator in the transmit unit generates a second
harmonic at a lower amplitude. This second harmonic is coupled to
the receivers local oscillator/ sweep and lock unit. The sweep
and lock unit ensures that the local oscillator is locked onto
the second harmonic. It also adds 10.7 MHz (FM intermediate
frequency) to the second harmonic. This signal is power spl1t
into two paths. One closes the loop with the coupler and the
sweep and lock circuit.
This feedback loop ensures that the sweep and lock circuit
receives 10.7 MHz at its input. The signals are mixed by the
single ended mixer to form this FM-IF tone. Figure 47 shows the
sweep and lock loop.
1---_,To transmitter
'. t t+If
po\.!er splitter 1sing 1 e ended mixer
At center frequency Ft:6 GHz Ft/2:3 GHz If:!0.7 HHz
Figure 47. Sweep and lock feedback loop
From figure 46 it can be seen that fed by the local oscillator, will
second harmonic available. This
frequency as that multiplied by
the receiver unit, which is
have a continuous source of
signal will have the same
the mixer diode within the
density tracer. It must be noted that the second harmonic does not have to sit at exactly 6Ghz, it is free to move within the bandwidth of the system. For this reason, the second harmonic has
62
The 3 GHz oscillator the transmit unit generates a second at a second harmonic coupled to
the receivers local oscillator! sweep and lock The and lock unit ensures local oscillator is locked onto the frequency) to the
(FH intermediate second harmonic.
into two paths. One closes the loop with the coupler and the
This feedback loop ensures that sweep and lock circuit 10.7 MHz at The signals are mixed by
to form FH-IF tone. 47 sweep and lock loop.
-~--f::>--r-l--~ To transmi t ter
47.
From 46 it can seen that , by the local , will have a continuous source of
same frequency as that multiplied by diode within the
tracer. It must be not have to sit at exactly 6Ghz, it free to move within the
of. For reason,
62
Univers
ityo
ape Tow
n
been given the notation ft. The need for the 10.7 MHz FM-IF will be explained later.
The 3 GH~ notch filter removes the high power fundamental frequency from the feedback loop. This filter allows the second harmonic through with minor attenuation. The description of this
filter is given in chapter 7 of this thesis.
6.3 The fundamental frequency transmitter and antenna Figure 48 shows the transmitter unit consisting of oscillator, RF amplifier, directional coupler, 6 GHz notch filter and 3 GHz horn
antenna/waveguide-coaxial transformer.
3 GHz oscillator
FtJ Ft/2 J
-Ft:"'2
---Ft
Ft+Hf
t
Figure 48. Block diagram of transmitter unit
3 GHz horn
' /
The oscillator is chosen to produce a fundamental frequency centered at 3 GHz, with a lower amplitude 6 GHz component. A 3 GHz RF amplifier (the specifications of which will be given in chapter 10) is used to boost the power of the fundamental to
produce 1 watt.
The directional coupler ensures signal flow to the sweep and lock circuit as well as to the transmit antenna. The signal from the
63
given the notation . The need the .7 MHz FM-IF will be explained
3 notch removes the high frequency from the feedback loop. This filter allows the second harmonic through with of filter given in chapter 7 of this thes
6.3 The fundamental frequency transmitter and antenna 48 shows I' RF
amplifier, directional coupler, 6 GHz notch filter and 3 GHz horn
tranS1:0rmer
:3 GHz osci 11 ater :3 GHz horn
Ftf Ft/2!
t
48.
The oscillator is chosen to produce a fundamental frequency centered at 3 GHz, with a 6 GHz . A 3 GHz RF amplifier (the specifications of which will be given in
chapter 10) is to the to
produce 1 watt.
The directional coupler ensures flow to the sweep and lock as as to transmit antenna. from the
63
Univers
ityo
ape Tow
n
receiver local oscillator is also coupled to the seep and lock circuit at this point.
The 6 GHz notch filter attenuates the second harmonic while allowing most of the fundamental frequency to be transmitted. The
second harmonic is attenuated to ensure that it is not erroneously detected by the receiver, as a density tracer.
The 3 GHz horn antenna/waveguide-coax transformer, transmits the
fundamental frequency to the density tracer.
6.4 The receiver unit
Figure 49 shows the receiver unit
single ended mixer, 3 GHz notch horn antenna.
(From receive; local osc:.
' ~~~~~--· -Ft+H
consisting of power combiner,
filter, IF amplifier and 6 GHz
~--_.'/I ..------.~
Ft F"t
I I
ended I
IF amp & detection unit
Figure 49. Block diagram of the receiver unit
The figure indicates how the tone (ft + IF) from the receivers
local oscillator enters the power combiner from the left. The 6 GHz horn antenna receives the second harmonic (ft) from the mixer diode in the density tracer. There is a small amount of coupling between the two horn antennas so some 3 GHz component is also received. The 3 GHz notch filter attenuates most of this signal,
64
local oscillator is also at point.
The 6 GHz notch attenuates allowing most of the fundamental frequency to
harmonic attenuated to ensure , as a
to
that it ity tracer.
not
3 GHz horn antenna/waveguide-coax transformer, transmits the to tracer.
6.4
Figure 49 shows .the unit consisting of power combiner, , 3 GHz
horn antenna.
pO\"lElr Fr. om recei~er . J ocal oSC:. I
. __ ...
, IF
Ft
ion unit
Figure Block diagram of the receiver unit
figure how the tone ( + IF) from the local oscillator enters the power combiner from the GHz horn antenna ( )
and 6 GHz
6
diode the density tracer. There is a small amount of coupling the two antennas so some 3 GHz component also
3 GHz notch filter attenuates most this signal,
Univers
ityo
ape Tow
n
while allowing most of the multiplied second harmonic to the
power combiner.
The power combiner then sends both of the RF(ft) and LO(ft+IF) signals to a single ended mixer where the difference component
(IF) is removed and sent to an IF amplifier and bandpass filter.
The reason for using this complicated form of local oscillator/receiver system now becomes clear. Detection is made
at IF frequency. In this case FH-IF (10.7 MHz). A very narrow
bandpass filter· (300 kHz) can be made thus bandwidth and increasing the receiver
reducing the noise
sensitivity. The
intermediate frequency is then sent to a FM demodulator, for detection.
What becomes immediately obvious is that the second harmonic used
in the receivers local oscillator and that multiplied by the
mixer diode are the same. They will only vary in amplitude. These
signals will drift with frequency in the same manner.
In sunonary then, on presentation of a density_ tracer containing a
cross dipole antenna and mixer diode, the receiver will detect a pulse at 10.7 MHz. This pulse will last until the density tracer falls out of the detection range of the receiver.
This solution is very elegant since detection is done at a very low frequency. None of the oscillators need to be crystal locked, and there is no delay in detecting a density tracer while the
I
circuit sweeps the frequency band.
The following chapters will discuss the design, construction and testing of each of the components of figure 46.
65
allowing most multipl second harmonic to the combiner.
The sends both RF(f t } and LO(ft+IF) to a single mixer
(IF) removed and sent to an IF amp 1 and bandpass
reason this
now becomes . Detection at frequency. In case FH-IF (10.7 MHz). A
filter {300 can be reducing bandwidth and the
frequency then sent to a FH
obvious is
oscillator and that multiplied diode are the same. They will in amplitude.
s drift frequency same manner.
In then, on presentation of a . tracer
cross dipole antenna mixer diode, at .7 MHz. will
out the detection range of the
narrow
noise
11
the
These
a
a
tracer
This low and
is very
. None of no delay
since is done at a oscillators need to be crystal I'
detecting a tracer sweeps the band.
The the I'
each of the of
65
\
Univers
ityo
ape Tow
n
CHAPTER 7
DESIGN AND CONSTRUCTION OF THE NOTCH FILTERS
7.1 Analysing the stub notch filter As shown in figure 46 there is a need for a 3 GHz notch filter which allows 6 GHz signals through with minimum attenuation, as well as a 6 GHz notch filter with minimum 3 GHz attenuation. For each filter, the bandwidths of the notch, must be sufficiently wide to accept drift of the oscillators. The nulls must be deep enough to provide adequate attenuation of the high amplitude fundamental frequency. Another criterion is that, in its pass
band, the filter remain as flat as possible and have a transmission loss of as close to zero dB as possible.
A filter which satisfies the requirement is the quarter wavelength transmission line stub filter. The · operation of a single stub will be reviewed and it will be shown how the notch filters was constructed using stubs.
From Appendix A we find that for RT duroid with the following characteristics,
E:r = 2.2 H = 0.254 mm t = 0.01778 mm
RHO = 0.84 RGH = 0
the width and length of a quarter wavelength of transmission line at 3 GHz is:
w = 0.76 mm 1 = 18.29 mm
The effect of an open circuit quarter wavelength stub at 3 GHz, on a transmission line, is now simulated by the package "EESOF Touchstone". Figure 50 shows the transmission line and stub.
66
. ,
7.1
As shown in figure 46 there is a need for a 3 GHz notch filter which allows 6 GHz through , as well as a 6 GHz notch filter with minimum 3 GHz attenuation. For
each filter, bandwidths the notch, must be suf wide to the must be enough to provide adequate attenuation the high amplitude
, in
transmission loss of as close to zero dB as possible.
A which requirement
pass
a
wavelength transmission The· operation of a
and will how filters was constructed using stubs.
From A we find characteristics,
at 3 GHz
~
t RHO
RGH
.. ..
= 2.2
= 0.254 mm = 0.01778 mm
0.84 :::: 0
of a
w = 0.76 mm
The effect of an open circuit on a ion
for RT duroid
of
wavelength stub at by
3 GHz, ..
Touchstone". Figure 50 shows the line and stub .
. I
Univers
ityo
ape Tow
n
T
1·V4 <3 GHz) 1 2 3 ------------ ------------!-········-+········-!
¥4 3 GHz V4 3 GHz/SO Ohms
-b-
Figure 50. Open circuit stub at 3 GHz
The program (OPENIM.CKT) analyses this circuits operation, and
can be found in Appendix E. The return loss and transmission loss for the circuit over a band of frequencies are shown in figure 51. Table 5 sununarises the characteristics of the filter.
freq( GHz) Sll(dB) Sl2(dB) 3 -0.157 -41. 027
6 -49.853 -0.139
Table 5. Characteristics of 3 GHz open circuit stub
From the table it can be seen that the circuit reflects most of
the power at 3 GHz: % reflection = lOO*ALOG(-0.157/10)
= 96% .. (7.1)
There is a small amount of reflection at 6 GHz (from eq. (7.1) we obtain 0.001%). Seen differently, the circuit allows almost all of the power (97%) at 6 GHz to be transmitted through it. The attenuation at 3 GHz is very high as only 0.008% is transmitted.
The problem that arises with this filter is that since a single stub is used the bandwidth is narrow (quarter wavelength). Techniques for increasing the bandwidth of the 3 and 6 GHz notch filters will be discussed in the next section.
67
Ohms
Figure 50.
The program (OPENIK.CKT) analyses this circuits operation, and can be found in Appendix E. The return and transmission loss
over a band 51. Table 5 summarises the
5.
From can power at 3 GHz:
seen
% reflection = 100*ALOG( 96%
There is a amount of
are shown in of the
c reflects most of
.157/10) .. (7.1)
at 6 GHz (from obtain 0.001%). Seen ferently, the circuit allows almost all
) at 6 GHz to it. attenuation at 3 GHz is very high as only 0.008% is transmitted.
The problem that the
with this is that since a single
narrow ( ) . Techniques for increasing the bandwidth of the 3 and 6 GHz notch
be next
67
Univers
ityo
ape Tow
n
+\ ~ ~ :
I I I I
....
.....
-
- f
\I '
\ \ \ \,
6 ~
I J
\ ,/ +
I '
I
0 0 . 0 ~
~
\_ '->\.
~
; /
I I
,..., .. ,
....__ -~k
_ __.../' --- -· v---
65
--- -
i---
0 0
ci CT)
I
0 0 0
cd
N :::c C.!J I
CJ LU a:
0 0 0
Lrl
0 0 0 . "
0-0 . 0
" I
• • [0
-
I- I II
\ "
b
-o
:~--~~~~~---1-----t-----1-----1----~----~~ o· o
6
Univers
ityo
ape Tow
n
7.2 The design, construction and testing of the 3 GHz notch
filters As has been shown in the previous section, a quarter wavelength
open circuit stub, makes a useful notch filter at the fundamental
frequency and allows minor attenuation of the second harmonic. The problem lay with the depth and the narrow.bandwidth of the null. One way to increase the bandwidth of the notch and to
deepen it, is to increase the number of open circuit stubs. The
spacing between stubs is set at a quarter wavelength at 3GHz.
A four stub filter was designed and a program to test the filter was written for the simulation package "Touchstone", which can be
found in Appendix E. A plot of the return and transmission loss is shown in figure 52. Comparing figures 51 and 52, it can be seen that the depth of the notch has been dramatically increased from -40 dB to greater than -100 dB. The width of the null has also been increased from approximately 625 MHz (at 10 dB) to 1400
MHz.
This would appear to be a good solution to the problem of providing a wide band notch filter.
7.2.1 Analysing the operation of the filter Figure 53 shows the 3 GHz notch filter
open c:irc:ui t -»'4 J:GH2 ...... !-·---·.., ,,. .. "fL~ n n ~I
j_ _ __J
Figure 53. 3 GHz notch filter
This circuit can be analysed by using wave dynamics, to prove how the circuit rejects the 3 GHz components but allows transmission of any 6 GHz signal.
69
7.2 filters
As a quarter wavelength circuit stub, a at
and allows minor second harmonic. The lay the narrow, null. One way to increase bandwidth the notch and to
it,. is to of open . The
set at a at
A stub was was for the simulation
found in Appendix E. A plot of
a program package
the return and 51
to test It, which can
transmission loss and , can
seen that the depth of the notch has been dramatically increased from to of null also been increased from approximately 625 KHz (at 10 ) to 1400
This would appear to be a good solution to problem of providing a band notch
7.2.1
Figure 53.
wave , to how the circuit the 3 components but allows transmission of any 6 GHz signal.
69
Univers
ity of
Cape Tow
n--J 0
30.00
-40.00
-100.0
EEsof - Touchstone - Sun Feb 19 01:12:50 1989 - 3GNOTCH
D DB[ S11] + DB[ S21] FIL FIL
- - -
>-\\ .... - -\~
V' I I --- r-r I -··- -- -- ··- . T ----·--·-·+ -- .. --··-r-.
\ /- j:l- r---.. I ~ r , . I V
\~ \ \ f \ I I \J \ I J
\ I v \ I \ ~ L\ I v ~ I
I
\ /
\ I I
' - ----\
.• I
I
\ I
\ I
·-\
I
' ~ \
\
\ \ --
i FtGURE. 5 2
t '
J
~ nnn " nnn i:-1.::u=n-~i-11 R nnn
30.
~
-40.
-100.0 ~
- Touchstone - Sun Feb 19 01:1 50 1989 - 3GNOTCH
DBl S11] FIL
r:-- . ......".
~y~ .....
\ \ \ \
\
\ \ \ \ ,
\ , I
\
\ \
t nnn
I
j
DB( S21)
.......
)gr v'
r\ I v \¢ I
I I
V
! I / ,
.' I
I
.' I
I
,
.-
-•
,-=:.. -.,... .. 'T _ ..• -. ~-- ._--_._ .... -t- i·· ...... -;--.
r-~ B ~ I
\ r \ / \ ( \~! -\ 1 \ ~ Ii / T !
I
I
I -- -I
I I
I
-I -1
I ~ nnn I=r::U:n-~1 R nnn I
FI
Univers
ity of
Cape Tow
n
Analysing the stubs at 3 GHz we find that since the length of the
stub are a quarter wavelength, the input impedance is given by:
Z.:1_n.(3GHz) = Z02 /Z 1
= 502 /00
--- OQ
•. (7.2)
The open circuit A/4 stubs thus present a zero impedance across
the transmission line at 3 GHz. However, at 6 GHz it can be found
that the input impedance is given by:
Z.:1_n.(6GHz) = ZC>*il1 *cos(2*rr*l/A.) + j*Z<>*sin(2*rr*l/A)] [Z 0 *cos(2*rr*l/A) + j*Z~*sin(2*rr*l/~)]
. . ( 7 . 3 )
Letting 1 = A/2, since A(6) = A(3)/2, gives:
Z.:1_n.( 6GHz) ZC>*~1 *cos(rr) + j*Z<>*sin(rr)] [ZC>*cos(rr) + j*Z 1 *sin(rr)]
= ZC>*[-Z 1 + OJ [-Z0 + 0]
= .Z.i
The load is thus reflected back to the input of the stub, since
Z~oo (open circuit), therefore Z.:1_n.(6GHz)--oo
The stub will thus present an open circuit across the
transmission line at 6 GHz and will flow through the transmission
line unaffected.
From Figure 53 it can be seen that the elements areA(3)/4 apart. At 6 GHz the gap between the stubs is A. ( 6) /2, thus the 3 GHz
signal has a maximum at the input to the stub. Since the 6 GHz signal is phase shifted by 180°, it will be transformed from the
end of the filter to the beginning. Thus the filter will tend to
minimize attenuation of the 6 GHz signal.
71
Analysing the stubs at 3 we find that the
stub are a quarter wavelength, the input impedance is given by:
(3GHz) == 11..1 •. (7.2)
== 1m
A/4 a zero across
the transmission at 3 GHz. However, at 6 GHz it can be
that input impedance given by:
1..:t.n.(6GHZ) =1..o*il1*cos(2*n:*l/A) + j*1..o*sin(2*n:*l/A)] [1..~*cos(2*n:*l/A) + j*1..1*sin(2*n:*l/~)]
1
•• (7.3)
A 12, A(6) = A(3)/2, :
(6GHz) = 1..O*il1*COS(n:) + j*1..o*sin(n:)] [1..o*cos(n:} + j*1..:J,..*sin(n:)]
to input
(open Circuit), therefore 1..:t.n( )_at
The an open across the
transmission
line unaf
at 6 and will flow through the transmission
From 53 can be seen that the are A (3) 14 At 6 GHz the gap between the stubs is A (6) 12, thus the 3 GHz
signal is
of
a at input to 6 GHz
phase shifted by 180~, it will be transformed from the
to to
minimize attenuation of the 6 GHz signal.
71
Univers
ity of
Cape Tow
n
'
Appendix F provides another interesting way of analysing the 3
GHz open circuit stub, notch filter. The stubs are compared to PIN diodes which are used to protect mixer diodes on transmission lines.
7.2.2 Testing the 3 GHz notch filter
Two filters were constructed using RT-Duroid 5880 and given the
dimensions used in figure 51. The completed version is shown in
figure 54 (both filters were dimensionally identical, hence only one is produced here).
Figure 54. Positive print of 3 GHz notch filter
The filters were attached to the same test rig as in section 4.6.1.1. The return loss and transmission loss for both of them
was obtained and can be found in figures 55 (a) .. (f), (can be found on the following page). Comparing figure 55 to figure 52, it can be seen that the filters operate almost as well in practice as predicted by theory. Table 6 compares the theoretical and practical results.
Freg(Ghz) S12(dB) notch bandwidth (Hhz) Predicted 3 >100 1375 Actual 3 75 1350 predicted 6 +/-0 -actual 6 0.8 -
Table 6. Characteristics of 3 GHz notch filter
72
Appendix F provides another
GHz stub, notch PIN diodes which are used to lines.
7.2.2
Two were constructed
way of 3
The stubs are compared to on
5880 and given the dimensions used in
figure 54 ( 51. The completed .
were dimensionally is in
, hence only one produced here).
Figure 54.
The 4.6.1.1.
were attached to the same test rig as in section return both of
was obtained and can be found in figures 55 (a) .. (f), (can be on ) . to ,
can be seen that the filters operate almost as well in
6
and practical results. (dB) (MhZ)
1375 1350
predicted 6 +/-0
actual 6 0.8
Table 6.
72
Univers
ity of
Cape Tow
n
:::: GHz FILTEF: -·---- --------------
._J
f-
i::t UJ
-'-
·--·----
.. .J
z:
1-C(
Lu tfl
::: 0 ·-··--··+-·-··-··f·· --+--· ·:-·---·-+---+---·+---+---+-··- ··' !·~ 2 6/div ~
~-----·---"-F-'-t-_•::...? _,·=t__;;l-c:.._j ..... =-:: t Ii:: / -~~_H_:::_· _.:i ____ _
_J
.Z
··-, 1--· LLJ
·::·
6
1 (1
1 4
1 ·=· . ,_,
.:, c; H :,: F I L f E P
I I I
I i'
I I ,,, \
. 6 /di ..... ________ F,_t- e ·=t '-' 2.t.1 c :;.-- i:: c; H z :.1
~'j· 55(c)
-------------
.. -..
(J")
((1
0 ..J
z 1::i .....
35
45
e-·.J .')
65
75
3GNOTCH FILTER
\,
\\ \ ...... .
..... , ·--------.------
.. _..
l,J)
(J"J •=> ..J
z 0
/-I ;-
Cr.: UJ (.{)
z
3 GHz FILTEF.: .:.-~
. --------. --
::: (1 ----+-+-·--J-- -·+----+··--··+- ·---t-· ·-·--+. - --L ···-·--···! . :~: 6 / .-j i .... ;:; i
F r e ·=t IA '"' n c ::--- i:: 1-:; H ;:: .1 _J
:z GHz FILTEF: 2 .... ·l
·'1 ' r-··-.
I
,1 I
I I ,. ii I r---I 1, I '1
I I ,1 I I I I ,1
'1 ·' ,' 1, I I I I
I I I I '1
I I I ~- ... I I 'i I
1 I ',J 11 \I I ii
I
·I
l 6 +-~-1----1---+- ~·--+-·-·· +-·----1 .-, .::. . 6 /di .....
Ft·e·=t1...lenc>· <GHz.I
· \:",~ 55 (cl)
3GNOTCH FILTER -.'5
. '5
1 .5 .,,..~ .. ..,---.. . ...,,.._._ _____ ......
-~ - -... /- ---.. . ..--
2.5
3.5
2.8 . 0 6 ... ·di'•.•' Fre<1.uen·:~.-· (GHz>
3.4 5.8 . 04.-· .. d iv Fre11.1...1enc:,... t'.GHz)
F'j. 55 (f') 73
z
1-'LtJ
........
.3 GHz FILTEF: 1 :' I':; H z F I L T E F: c.
f--ctC: LU ,; (1
.. _+.... ·-I··_· .. t-- +-- ... +-+---i--........ I .. ' .] i '~:
::: 0 +--_.+--+"-/--'+_.'+----+_._.-4-.. _ .. -1-
::: C;~L: FILTEF: -2 -1
·1
1 .-, ,;:.
.-,
.:.: .6',:li\1
55 . \="':'155 (cO
-_ .... _ ... _--3GNOTCH FILTER 3GNOTCH FILTER
35 ...... -.5 co 7J
45 ' .... .5 if)
ii) e'" • .;J .:;) 0 1.5
..J
:z: 65 0 2.5 ....
.I-~ W 75 3.5 (.I")
:z: -3.4
:3 5 ......... --+--+--+---+--+---I--t--+---+--I 2.8 .06/div
___ F re,\uent:}" (GHz)
4.5+--+--~~--~~--I--~~--~--! 5.8 .04/div 6,2
F re,\l.JerICY (C;Hz)
Cf) 73
Univers
ity of
Cape Tow
n
From the table it can be seen that the notch is 75 dB compared to
the 100 dB predicted by Touchstone. In the program no algorithms were included that corrected for the microstrip propagation and
radiation losses, this could be the reason for the errors in the
prediction. The insertion loss .at 6 GHz is approximately 0.8 dB
which is satisfactory (83% of the power is transmitted through
the filter).
7.2.3 Conclusions The 3 GHz notch filter conforms adequately to what was predicted
by theory. The notch is deep enough to ensure adequate attenuation of the fundamental frequency. A lW signal at 3 GHz
will be reduced to 3.16*10-0 wafter passing through the filter.
7.3 Design, construction and testing of the 6 GHz notch filter
The 6 GHz notch filter is similar to the 3 GHz notch filter in
appearance and operation. A major difference however is that the 6GHz notch filter uses short circuit stubs. The stubs are
separated from each other byA(6GHz)/4 sections. The lengths of
the stubs areA(3GHz)/4 as before.
"Touchstone" was used to analyse the notch filter. Appendix G contains a copy of the program entitled "6GNOTCH.CKT". Figure 56
indicates the return and transmissions loss of the filter. Comparing figure 56 with figure 42, it can be seen that they are almost 'identical but symmetric around the 5 GHz line. Touchstone predicts that the notch at 6 GHz has an amplitude of >~100 dB and that the bandwidth is approximately 1400 MHz (at 10 dB points). The transmission loss at 3 GHz is approximately zero. In practice open circuit stubs are easier to fabricate than short circuit
stubs in microstrip.
74
From the table can be seen that the notch 75 dB compared to by • In the no
were included corrected for the propagation and radiation losses, this could be the reason for the errors the
,at 6 GHz which is satisfactory of the power is transmitted through the ).
7.2.3 The 3 to what was by theory. The notch deep enough to ensure attenuation of the fundamental frequency. A signal at 3 GHz
will through
7.3
appearance and 6GHz notch
in operation. A major difference however is that the
uses stubs. The stubs are from each other bYA(6GHz)/4 sections. The lengths of
the stubs are A ( 3GHz) 14 as before.
"Touchstone" was used to analyse the notch filter. Appendix G contains a copy "6GNOTCH.CKT". 56
return and transmissions Comparing 56 with figure 42, it can
but around the 5 predicts that the notch at 6 GHz has an
loss of the be seen that they are GHz . . Touchstone
of the bandwidth is approximately ~~~~== 10 dB points).
zero. In The transmission loss at 3 GHz is stubs are eas to fabricate than short circuit
stubs
74
University of Cape Town
:c
I u
t-
0 ~ lC
I C
") a
J
C'J
I '
I I I
I ~
C'J (T
l
c:c ...... ...... 0
I I I
C")
...... .c
Q
J LL..
c: ::l en
...... _.. N
Q
J en
c: -.... __J
0 co~
..._, 0
LL.. U
J .s=
+
u :::l 0
I I t-
.--...... ......
~
en 0
-..... _J
UJ
co~
UJ
Cl LL..
LU
D
... ~ ~
I L
---!
v " \
0eL { ~-~--I
...__ -. ""-·
' ~-.
I "'
I -
I
I 1 I .1 I
I '
I I
>
I J ~~+-
---· -
---
... ---~ ~--
:r--
/\ .... ~-
-~
.. \
... .
.. '/
"V 75
!'a
v
~
r--_ t--
~ i--
0 0 ~ I
l I
I '-1-I
-·-
I
.-.-
---~ -·--
I
I
R._
I
;
I 1 I
I I -I
I .
I ,_
I 1_
-· I I I
I I 0 0 0 aJ
N
~
j d, I
LU
a: -1
LL.. I
0 d
0 ...... I
0 0 0 Lri
0 0 0 . \,
- Touchstone - Sun Feb 19 01:16:39 1989 - 6GNOTCH
S11] DB[) L
30. --
-- .-
• r-1 r-.
K I 'I" _~_,
)X ...... L-I ..
) I- .","-t ,-' , / r-a _._.:.- I - - r 1 \ ----- -..,..
V \ l\ 7 \ I
" \ ,.
\ ¢ I \
I
\ i I
\ .J --40.00 ~, [ I~ \ Il ~I \
\ I '. --I
I
v ._--f-1-' ,
It' \ I , I
- \ I -I
\ I j
I I , i
\ ,
I I \-- ----
i
1. I -, -I I
-100.0 , I ----"- '-"--
J.l nnn I ~ nnn nnn I=CIJ:n-CU7
Univers
ity of
Cape Tow
n
A program entitled "6GSNOTCH.CKT" was written, the objective was to determine the effect of the resistance of the short circuit
stubs, which were fabricated using brass screws; on the
characteristics of the filter. The program can be found in
Appendix G. The microwave resistance of the screw was assumed to be lQ . The return loss and transmission loss for a four stub filter can be found in figure 57.
Comparing figures 56 and 57 it appears that the null is wider
(1800 MHz as compared to ~1_4~0~0~-HH~_z) and possibly deeper, if the screw resistance is taken into account. The transmission loss at
3 GHz appears to be deeper, it can thus be expected that slightly
less 3 GHz signal will propagate.
7.3.1 Analysing the operation of the 6 GHz notch filter Figure 58 shows dimensions of the filter.
.. --1 .>-.-(6)/4
Figure 58. 6 GHz notch filter
111hor-t c:ir-c:ui t sc:re~
_J At 6 GHz the short circuit stubs provide an input impedance of:
Zi~(6GHz) = Z~*l.Z.i*cos(2*n*l/A) + j*Z~*sin(2*n*l/A)] [Z~*cos(2*n*l/A) + j*Zi*sin(2*n*l/A)]
.. (7.4)
letting 1 = A/2, this yields:
Zin(6GHz) = Z~*l.Z.i*cos(n) + j*Z~*sin(n)] [Z 0 *cos(n) + j*Zi*sin(n)]
= Z~*(-Zil (-Zo)
= ~.1.
76
A program entitled "6GSNOTCH.CKT" was , the was
to determine of the of the short
were screws; on the
characteristics of the program can be found in microwave of
return loss and transmission
can be found in figure 57.
screw
loss
Comparing
(1800 MHz
screw res
56 and 57 appears that the
as compared to ) and poss
taken into account.
3 GHz to be deeper, it can thus be
less 3 GHz signal will
7.3.1
circuit
58.
was to
a four stub
wider
at
ightly
At 6 GHz an:
Z~n(6GHz) = Z~*La1*cos(2*R*1/A) + j*Zo*sin(2*R*1/A)] [Z~*cos(2*R*1/A) + j*Z1*sin(2*R*1/A)]
.• (7.4)
letting 1 ::::: A , :
( ) ==
==
76
University of Cape Town
:c
CJ ~
~ (.!) tO
I
O"J a
J
O"J ....... ..-1
(T
)
•• C
D
(T)
. . 0 0
.
O"J ..-1
.c
QJ
lJ.. c: ::::l en
QJ
c: 0
-6-' U
J .c
u ::::l 0 ~
""-0 U
J f±j - ...... C
\J en ~_J
CD
t-t
ClL
L
...... I
- ..-1
...... en ._
_ _
J
CD
t-t
ClL
L
D
I 0 0 d
(T)
I ' i I/
+
I
rr i. y C> ?1\ ~
·--.
~
h ] ~ v--·---
-
r..
J\ ~
i ... I('
\ ""· .. ,_ ' ... \ t
l..-C::r
'( 77
[::a
----h
~
/
~ c
!~ -~+
--
-
t>
r---_
r t:> 0 0 ~ I
.
--
---
I --.-
-
I I I
·--. -
-.__
I
---
-··~
-
-. ...:...._
.-0 d
0 ~ 0 0 0 . C
D
N
ra I C
J lJ.J a: LL.
0 0 0 . tn
0 0 0 . II.
30.
~
-40.
-100.0 ~
BEsof - Touchstone - Sun Feb 19 00:38:31 1989 - 6GSNOTCH
DB[ Sll]
X-t - I
't\ \~ V "
nnn
- .
DB[ S2l] FIL
.--.
/ v\ (" i\!
V V
'+---"'-- ..... '~ ~ - -
/ r-" \
\
~ \
\
,
\
I,
-, \
s:; nnn
--,
'-' ~~ ~ >\r Et +-
,~
" I V \ , }J
I
,I 1
I
I. -" I
-I
l -I
I \
I
I
!-t"Dt"n, .C&' ~ nnn
"
Univers
ity of
Cape Tow
n
Now Zi~(6GHz)---O since Zi--0 (short circuit stub)
Therefore the stub presents a low impedance for signals at 6 GHz. At 3 GHz, however, A(3) 2*A(6) therefore \(6)/2 = \(3)/4 i.e. at 3 GHz, the stub is a quarter wavelength long, therefore:
z<> 2 /Zi 50 2 /0
therefore, Zi~(3GHz)--m , and hence the 3 GHz signal will tend not to propagate into the stub. The expectation thus is that
6 GHz frequency components will tend to be reflected, while the 3
GHz component is propagated with very little attenuation.
7.3.2 Testing the 6 GHz notch filter A 6 GHz notch filter was constructed using RT-Duroid 5880 and the
dimensions laid down in Appendix G. A positive photo print of the filter is shown in figlire 59.
Figure 59. Positive print of 6 GHz notch filter
The return loss and insertion loss for the 6 GHz filter was measured using an HP 8410 B network analyser. Figure 60(a)&(b) show the characteristics of this filter.
78
Now (6GHz)-0 ( )
Therefore the stub presents a low impedance for signals at 6 GHz. At 3 GHz, however, A(3) = 2*A(6)
)..{6)/2 :::: \(3)/4 i.e. at 3 GHz, the stub a wavelength long, therefore:
(3GHZ) = /z~
50 2 /0 , ( 3 GHz 1
tend not to propagate into stub. The expectation thus is that 6 GHz 1 , 3 GHz component propagated with very attenuation.
using RT-Duroid 5880 and the dimensions in G. A photo of
is shown in figUre 59.
The return loss and insertion loss for the 6 GHz was us an HP 8410 B 60{a)&{b)
show the characteristics of this filter.
78
Univers
ity of
Cape Tow
n
-·- 4 0
L:Li .-7) -·- .:~ 0
I (t 0 I t)
--I ' --_J
.-, (1 .:: ..:..
l.t ::t 1--LLl
4 0 i:i::
6 0
6 GHz FILTEF.:
I-er.: lJJ
-······--+-·--+--·-+---+--+--~-+---+----+---l ·-~
? 6/div 8 F t .::· 1 u e n r: ::-- i:: G H z :1
6 G H z F I L T E F: -- 2 0
0 -------···--·---·----· .. -
20
40 .
::: 0 ·---·+-+----+---+·--+--+--·+---+·-----1··-·-····-· 2 .6/div ~
F r· e ·'t '.A e n i: :: •:. 1::; H z •
Figure 60(a)&(b). Return and insertion loss of 6 GHz notch filter
Comparing figure 60 and figure 56, it can be noted how accurately
the filter conforms to theoretical prediction. Table 7 compares
the practical results with theoretical prediction.
Freq(GHz) S12(dB) notch bandwidth (MHz) predicted 6 >-100 1400
actual 6 -70 1800 predicted 3 +/-0 -actual 3 -0.4 -
Table 7. Predicted and actual characteristics of 6 GHz filter
The transmission is approximately ~ at 3 GHz, as expected. The depth of the null is less than what is expected, but this could be due to losses in the microwave substrates which were not taken
into account by the program.
7.3.3 Conclusions The 6 GHz notch filter compares favourably with that designed on
Touchstone. Table 7 indicates that it will work satisfactorily as
79
4 (1
L:::I .. '1'J
... ~. 8
-(I)
(1 tft
! =1 ...J
(1 ..;.
(.I:: ~'::I
1--w 4 0 fi:
IS 0
Hz: FILTEF: t, GH:: FILTEF.:
:. IS ,/ d i',,' f:t",·lu,:;.nc:" (GHz
60(a)&(b}.
-·2(1
. IS d i \.'
Comparing 60 and figure , it can be how conforms to theoretical Table 7
the practical results with theoretical prediction.
Freq(GHz) S12(dB) notch bandwidth predicted 6 >-100 1400 actual 6 -70 1800 predic 3 +/-0 -actual 3 -0.4 -
7.
(MHz)
The at 3 GHz, as . The depth of the is less than what is but this could be due to losses in the microwave substrates which were not taken into account by the program.
7.3.3 The 6 GHz notch filter compares favourably with designed on Touchstone. 7 indicates it 1 work satisfactorily as
79
Univers
ity of
Cape Tow
n
a notch filter at 6 GHz, and that there will be negligible
attenuation of the 3 GHz signal:
% attenuation 100-lOO*ALOG(-0.4/10) (at 3GHz)
= 9%
80
a notch at 6 GHz, and that there will negligible attenuation of 3 GHz :
% attenuation lOO-lOO*ALOG(-O.4/10) (at )
Univers
ity of
Cape Tow
n
CHAPTER 8
DESIGN AND CONSTRUCTION OF 6 GHz POWER SPLITTERS/COMBINERS
8.1 Theor:y of power splitter operation
8.1.1 Determining the isolation and transmission coefficients
of the splitter
From figure 46 in chapter 6, it can be seen that the frequency
multiplication system requires a power splitter and a power combiner. Both devices must operate at a center frequency of 6
GHz and have a wide enough bandwidth to ensure frequency drift of
the oscillators. The coupling between the input port and the
output ports must be as close as possible to 3 dB to ensure maximum power transfer. Thus there must be adequate return loss
at the input port and suitable isolation between the two output
ports.
A three port hybrid which works well as both a power splitter and
power combiner is that described by Wilkinson. This has been
described by Cohn ci4 1 Figure 61 shows a Wilkinson power
splitter with quarter wavelength transformers and lQQQ isolation
resistor.[ ......---··~~----·-----~ J
I 21=70.7\:~ /l : i Zo=SO Ohms ~ •• -------~""- /100 Ohms
Zl=70.7~~~---------r - ~ I
Figure 61. Sin_g_l_e_s-ection Wilkiiison-power splitter
As power enters port 1 it is split equally between the two output
ports 2 and 3. Theoretically, there is zero phase difference between the ports, with the isolation resistor across the output ports providing output matching and isolation.
81
8.1
8.1.1
From can seen multiplication a power splitter and a power
. Both must at a center 6 GHz and have a wide enough bandwidth to ensure frequency of
the The coupling between the input port and the output must as maximum power transfer. Thus at input and
A hybrid which works
combiner is that
by [14]
as to to ensure there must be adequate return loss
two
as both a power splitter and by Wilkinson. This has been
61 a splitter with quarter wavelength transformers and 100Q isolation
~~--~~-~
I 100 Ohms
61.
As enters port 1 it is split equally between two output 2 and 3. Theoretically, there zero
ports, with the isolation across output providing output
81
Univers
ity of
Cape Tow
n
The role of the resistor can be analysed by the method of even
and odd mode excitation of ports 2 and 3 with a load Z0 connected to port 1.
With even mode excitation, waves
phase difference are applied of equal
to ports 2 amplitude and zero
and 3. The voltage difference between the ports is zero, and no power is dissipated in the resistor. The power output at port 1 is the total power from ports 2 and 3 minus the reflected power.
Since there is no current flow through the resistor, figure 61
can be redrawn as in figure 62. The left hand load is replaced by 2*Z0 because of the bisection of the circuit.
Figure 62. Circuit bisected for even mode analysis
Odd mode excited waves are now applied to ports 2 and 3. These are waves of equal amplitude but opposite phase. The application of these signals to the output ports creates a substantial voltage difference across the resistor.
Because of port 1 are follows in
symmetry,the midpoint of· the resistor and the node at at ground potential, figure 61 can thus be redrawn as
figure 63. 1
l Z=O
Figure 63. Circuit bisected for odd mode analysis
82
the res can be analysed by the method of even
and odd mode
to. port 1.
of ports 2 and 3 a load
With even mode excitation, waves of amplitude and zero
are to 2 and 3. voltage
difference between the ports is zero, and no power
the output at port 1 the total power
from ports 2 and 3 minus the power.
no current through
can be redrawn as in 62. The hand load
2 of bisection of c
62.
Odd waves are now
of equal amplitude
to ports 2 and 3. are waves
to output
opposite phase.
ports creates voltage difference across the resistor.
a
61
by
Because of symmetry, the midpoint of the and the node at
1 are at , 61 can as follows in figure 63.
Figure 63. Circuit bisected for odd mode analysis
82
Univers
ity of
Cape Tow
n
If the situation is reversed and incident waves approach from the
left (port 1), then figures 62 & 63 can be redrawn as in figures
64 & 65. Yo=! o~----~Y~i~•t-.
~ Yin,e- ~Gl :Q,S
o~--~•--=:s-r--.. .j_-1_
........ _, ¢
Figure 64. Admittance circuit: even mode
·----- ----i
Yo=~----------~
... ----'l'in.o Gl= oo
.-.Jl:S .. ..; _ t-t-
- I<() J -------~~o ___ _
Figure 65. Admittance circuit: odd mode
Admittance representation is simpler to analyse. It has been shown , that the following substitutions can be made.
Y1 = l/Z1 , G1 = l/R1 G1 = l/(2*Z~) = 0.5
,
j>.,. and _p ~ are the even and odd mode reflection coefficients,
and 9 1 , 9 2 ,and 9 3 are the voltage reflection coefficients at ports 1,2 and 3. t 12 , t 13 and t 23 are the voltage transmission coefficients between the ports.
It has been shown that the following results hold for the
synunetrical three port case:
l.P .,.1 = l.P ~ t12 = t13 .. (8.1)
/t12I = lt13I = 0. S*v ( 1 -y .,.)2' .. (8.2)
83
If the situation and incident waves approach from the
( 1), 62 & 63 can be ~~~~--~~~--.
as in 64 &
Yo:::! o---_.u..... .......
::::0.13
Figure 64. Admittance circuit: even mode
Figure
Admittance representation to analyse. It has shown , that the following substitutions can be made.
.Ye and
= lIz:\. == 1/(2
"" are
,
and 9 l' 92,and ports 1,2 and 3.
)
G1 :::: 1
0.5
even
between the
It
, == 1 == 1
and odd mode reflection coefficients, voltage at
and 3 are the voltage
the hold symmetrical three port case:
11' t12
It
:::: 12 ~ = t 13 = !t13! o.
.. (8.1) •• (8.2)
Univers
ity of
Cape Tow
n
Y2 = Y3 = O.S*(J e + fo}
t2 3 = 0 • 5 * (J' e - 9 o}
.. (8.3)
•• (8.4)
The isolation provided by the resistor can be understood in a different way.
Firstly it is assumed that the source impedance is son . A match
at the input is achieved by transforming the son (characteristic
impedance) to lOOn This is done by using a quarter wavelength transformer, with a line impedance of 70.7Q The resistor R
(from figure 61) has a value of lOOn. Each output port
effectively has two loads of lOOQ in parallel with it (which
equals SOn). The input port is thus matched to the resistor.
Any signal incident at port 2 can reach port 3 via two routes.
One is through the resistor, which is assumed to have zero length. The other path is through the two quarter wavelength
sections. The two signals thus arrive at port 3 with equal
amplitude but 180° out of phase (odd mode). The two waves thus
destructively interfere and cancel each other. Therefore the
resistor isolates port 2 from port 3.
The magnitude of the isolation is determined by the difference in path length.between ports 2 and 3. A major source of error is the resistor length, which is theoretically zero. This is not possible and the consequence is that the two signals will not arrive exactly 180° of phase. This will decrease the isolation between the ports.
8.1.2 Simulating the 6 GHz power splitter The power splitter/ combiner used in the frequency multiplication system required a reasonably wide bandwidth to accommodate shifts in frequency of the oscillators. The return loss from the input and output ports had to be low, hence the VSWR for the ports had to be as close as possible to unity. The isolation between the
84
Y2 = .93 0.5*(J E!! + ) .. (8.3)
:3 = 0.5* ) .. (8.4) E!!
by can be understood in a different way.
assumed the source 500 . A at input is achieved by transforming the 500 (characteristic ~ .. v~~u.u~.~) to 1000 done by using a wavelength transformer, with a line impedance of 70.70 The R
(from figure 61) has a value of 1000. Each output port two
SOO). The
1000
thus (which
to the
Any signal 2 can port 3 via two routes. One through the resistor, which is assumed to have zero
The two at 3 with equal
amplitude but l80~ out phase (odd mode). The two waves thus each the
isolates port 2 port 3.
path length between ports 2 and 3. A major source the not ,
possible and the consequence l80~
between the ports.
zero. that the two signals will not
8.1.2 Simulating the 6 GHz power splitter The power splitterl
required a reasonably wide bandwidth to accommodate shifts return from the
and output ports had to be low, hence the VSWR for the ports had to as as to between the
84
Univers
ity of
Cape Tow
n
output ports had to
output ports had to
power transfer.
be high and the coupling between input and
be as tight as possible to ensure maxim\J.m
The standard single
before had a usable
isolation > 20 dB.
section Wilkinson power splitter
bandwidth of 1.44:1 for VSWR
The bandwidth could be increased
described
<l. 22 and
by adding
more resistive elements, but a bandwidth ratio of 1.44 :1 was
satisfactory for the frequency multiplication system. It was thus
decided that a single section Wilkinson power splitter be used
for the system.
A program entitled 6GSPLIT.CKT was written for the package Touchstone, to determine how closely it conformed to theoretical
prediction. The program can be found in Appendix H. It describes
the return loss from port 1, the isolation between ports 2 and 3,
and the transmission loss from port 1 to ports 2 and 3. It was
designed using widths and lengths for a 6 GHz transmission line
(obtained from Appendix A). A plot of the power splitters
characteristics can be found in figure 66. These characteristics
are summarised in table 8.
IF F;E·J-Co,-iZ
I
I ::..coooo 2. :1()()()i)
:; . oocoo I 3.SCOOU
4.00000 4. 5;j)00 :i. ocrx.o
6. COOCH) 6. Si::C(Jt::t 7.ocooo
·:· C·! ":'; _l I ;..._ .J.. I
-·1 :.: •. :::r;t7 - ~ 7. s,:~c;
-21.1~0
-15.~42
CE1[~>i2J ~;r'LIT
-3.545 -::! . 502
·-; .1 c:: ,-, -._:1 • 4._IL
-3.400 -3.351 -3.309 -~:. 278 - 3. 2t,:3 -3.264 - 3. '2.~;4 -3 . .321 -.~) • . 373 -.3. 4.37
-,..-1c1-r-·:· ·.·-·-:.-1--r;-.9·--L., --~-·;-· :=--. J·;-ij WU1..,_1 l-...J..; ,_,. _,..__
~; PL l T '3 t=-· L l I /
-3. 545 -0. ~ ti'?I -::5. 5t):2 -9.65i. -.3. 4:)2 - i i . 1 se) - ~~. 4()\) -i2.8 1:)6! -.3 .. 3~1 l -14.777 -.3 •. 3Cr? -17.314 -J.~78 -·20. CiSt1 - 3. 2,-=,.2 -:i:t). 72~1:3 -:-:::. 264 -55.99'3
-27.0Sb -:3 . . 321 -2tJ.9i6
-. --J~ -,
- .,:1 •. .J ( .:J -1:._:;·7'1 -1.:.;.. ~·23
Table 8. Characteristics of 6 GHz power splitter
85
output ports had to be high and the coupling between input and to as tight as possible to ensure
power transfer.
before had a usable bandwidth of 1.44:1 for VSWR <1.22 and > 20 bandwidth by
more resistive elements, but a bandwidth ratio of 1.44:1 was satisfactory for the frequency multiplication system. It was
a single section Wilkinson power splitter be used
the system.
A .CRT was written for the package Touchstone, to determine how closely to
The program can be found in Appendix H. It describes the return from port 1, the isolation between 2 and 3,
the
designed using (obtained
loss from widths and lengths for
Appendix A). A found in figure
are summarised in
Table 8.
1 to
a 6 of
66.
ports 2 and 3. It was
GHz transmission line
These characteristics
Univers
ity of
CapeTow
n()
J ()
)
10.0
0
,.;.20.
00
-50.
00 r ,..
EEso
f -
Touc
hsto
ne -
Sun
Feb
19 0
0:42
:24
1989
-6G
SPLI
T '
0 OB
[ S11
] +
OB
[ Si 2
] 0
08( S
13]
>< DB
[ S23
] SP
LIT
SPLI
T SP
LIT
SPLI
T -
··-·
-·
A.
.... A
. ....
A..
....
A
.... A
A
.... y
... y
-.r
... ...
'I'
.... v
~
--
6 t:
-2f-
-~
-'"'1
:::1--~
""!:::!
--·
-,
~ ~ -~
~ ~
) ~ \
fl ! \ \ p I
-l ·~
..
nnn
s; nn
n l=
lll=
n-~~
7
.... A
..
y T
~
~
-
~IGUQ.E.
IQ6
R n
nn ,
~~
\ \ · , CD
..;~
I- \. ......... -J
B5 to to
I 1"""""11 (T)
OJ ('\J I-m en ......... C'l ......... -J ~ c:c 0.. aen "IIif' ('\J
N "IIif' .. 0 ..-0 (T)
~I-
~t>
~ ! N
~
~ a5 I
E;-. .. c:I <;;. .~ UJ
~. a:
~ u...
~~ ~
~~ / . OJ en ........ ~ ......... -J
.CIJ ~B5 Ql I
0 0
~~ 0 · an
u... 0
C :::J en
0' I ~I-
Ql en ....... C m~ 0 ..... aen en .c u
~~ ff ,Do ¢
:::J 0 I-
1"""""11 '0 frJJ
~ ~I-
'3- en ....... 0 ......... -J en c:c 0..-
I±I oen ~,.
! ! ]
0 0 0
· ,"'
Univers
ity of
Cape Tow
n
It can be seen that the return loss from port 1 is lowest (-43.7 dB) at 6 GHz, the isolation between ports 2 and 3 is greatest (= 55.99 dB) at 6 GHz.1The coupling is tightest at 6 GHz with a transmission loss of -3.264 dB.
It can also be seen that the return loss and isolation is below=
20 dB for the 2:1 bandwidth, and the transmission loss is almost linear varying from -3.263 dB to -3.373 dB.
These values indicate that the simulation results conform to those predicted by theory.
The VSWR for each of the ports was determined using the formulas: I
VSWR(portl) = ·1 + ALOG (Sll/20) 1 - ALOG (Sll/20)
VSWR{ports 2&3) = 1 + ALOG (523/20) 1 - ALOG (S23/20)
.. (8.5)'
•• (8.6)
Figure 67 shows the VSWR for the splitter for the input and
output ports.
87
It can seen return loss from 1 lowest ) at 6 GHz, the isolation 3
at 6 GHz./The coupling at 6 GHz with a
It can seen return and isolation is be10w= the 2:1 bandwidth, and the 1055 is
varying from to -3.373 dB.
These values indicate that the simulation conform to
those by
The VSWR for each the, ports was determined using the formulas:
VSWR(port1)
V5WR(ports 2&3) = 1 + ALOG (523/20) 1 - ALOG (S23/20)
Figure 67 shows the V5WR
output
87
•• (8.5)"
.• (8.6)
splitter for input and
Univers
ity of
Cape Tow
n
VSWR OF INPUT AND OUTPUT PORTS OF 6 GHz POWER SPLITTER V3WR
2.5.---~~~~~~~~~~~~~~~~~~~~~~~~~~
2.0
1.5
to'--~~~~~---J.~~~~~~-L.~~~~~~....,::~~~~~___J
0
-2 4
Frequency (GHz) 6
VSWR PORT1 (MATCH) -+- VSWR PORTS 2&3 (ISO)
Figure 67. VSWR of the power splitter
8
From figure 67 it can be seen that the VSWR is approximately <1.22 for the bandwidth ratio of 1.44:1. The VSWR approaches unity at the center frequency of the power splitter. The characteristics of this splitter indicate that it will operate effectively over the bandwidth of the frequency multiplication system.
88
VSWR OF INPUT AND OUTPUT PORTS OF6 GHz POWER SPLITTER V8WR
2,5r-------------------------------------------~
2.0
1.5
1.0 L....-____ --1.. _______ ...L-____ -::111!~------1
o
-
61.
2 4
Frequency (GHz) 6
VSWR PORn (MATCH) -+- VSWR PORTS 213 (ISO)
8
From 61 it can be seen that the VSWR is approximately < 1. the bandwidth ratio of 1.44: 1. VSWR ;>7">, ............... "" ....
at the center of power splitter. The of that it
over the bandwidth of the frequency multiplication
aa
Univers
ity of
Cape Tow
n
8.1.3 Simulating .the effects of the quarter wavelengths
and isolation resistor
The simulation package Touchstone is useful in analysing the effects of resistance, line widths and lengths, and other factors
on the losses of the power splitter.
The first step in analysing the effect of each component of the power splitter on its operation was to determine the return loss and isolation for a simple three port circuit. The impedances of
the two output ports would then be increased from son to 70.7Q ,
and the effect of this change on the reflection coef f i~ients and
isolation, would be noted. Finally, a lOOQ resistor would be placed across the terminals of the output ports. This is the
standard power splitter/combiner.
A program entitled 3PORT1.CKT · was written for the simulation
package Touchstone, and can be found in Appendix H. The aim of the program was to determine the characteristics of a simple
three port circuit as in figure 68.
~so ,.. . -="'
Ohms
50 Ohms
1so Ohms
--~-··----_ / Figure 68. Simple three port circuit
89
B.1.3
The simulation package Touchstone is useful in analysing the
on the
The
power splitter on
and
the two output
, would
, line
power
and lengths, and other factors
operation was to determine of
return
of would then be increased from SOg to 70.7g ,
change on the lection coef and
, placed across the terminals of
standard power splitter/combiner.
a lOOg res
output
would
. This
A program entitled 3PORT1.CKT was for the simulation
Touchstone, can H. of
the program was to determine the characteristics of a simple
as in 6B.
Ohms
1r---__ ~----------r
SO OhlW5
Ohms
6B. Simple three port circuit
89
Univers
ity of
Cape Tow
n
This circuit can be analysed using simple transmission line formulas:
Impedance looking in at a : = SOQ//50Q
= 25Q
Reflection coefficient from a back to source
Power reflected:
Therefore Sll
Power transmitted:
= Pi.n./9
lO*LOG(Pi~/pi~/9)
9.54 dB
P.Ln. - P:re:t:(a)
pi~(l - 1/9) = pi~(8/9)
Power transmitted into each element: = PL~(4/9) Therefore transmission l·oss = -10*LOG(PL~/PL~(4/9))
= -3.52 dB Table 9 shows the return loss and reflections from the output ports for the program 3PORT1.CKT.
90
can analysed using simple transmission line formulas:
looking at a: :::::: SOn//50n
coef
Power . .
Power transmitted:
:::::
a
(a) =
Sll :::::
:::::
=
to source
191 2 * P'J...n./9
10*LOG(
(a)
(1 1/9)
8/9)
Power transmitted each element: ::::: P'J... .... (4/9) :::::: o *LOG (
9 the return ports for the program 3PORT1.CKT.
90
)
(4/9 »
output
Univers
ity of
Cape Tow
n
I .f' RE Q--:= GHZ I
2. 00000 2. 50000 3. 00000 3. 50000 4. 00000 4. 50000 5. 00000 5. 50000 6. 00000 6. 50000
)I 1. 00000 11 7. 50000 ~ 8. 00000
08£ S111 08£ S121 08£ S131 SPLIT SPLIT SPLIT
-9. 687 -3. 570 -3. 570 -9. 674 -3. 573 -3. 573 -9. 657 -3. 576 -3. 576 -9. 640 -3. 579 -3. 579 -9. 626 -3. 584 -3. 584 -9. 619 -3. 589 -3. 589 -9. 619 -3. 595 -3. 595 -9. 625 -3. 601 -3. 601 -9. 635 -3. 607 -3. 607 -9. 647 -3. 613 -3. 613 -9. 660 -3. 617 -3. 617 -9. 669 -3. 620 -3. 620
_-_9_ :~7_ 6_=~-,.;}. ~ 2 2 -3. 622
08CS231 SPLIT
-3. 554 -3.562, -3. 570 I -3. 576 t -3. 579 i
-3. 580 -3.579 -3. 576 -3. 573 -3. 571 ~ -3. 570~ -3. 573! -3. 577
Table 9. Characteristics of a simple three port device
It can be seen that at 6 GHz the return loss is -9.635 dB close
to the predicted value of -9.54 dB. Similarly, it was predicted
that the transmission loss would be -3.52 dB. From table 9 a loss
of -3.607 dB was obtained. This program thus confirms the results
obtained from theory.
~ A second program 3PORT2.CKT was written and can be found in
Appendix H. The circuit model contained quarter wavelength
sections with 50*v2 = 70.7Q characteristic impedances. Figure 69
shows the circuit and Table 10 contains the characteristics of ·---------
the circuit.
70,7 Ohms SO Ohms
Figure 69. Three port with 70.7Q impedance transformer lines
91
2. 00000 -9. 687 -3.570 2.50000 -9.674 -3.573 3. 00000 -9.657 -3.576 -3.576 -3.570 3.50000 -9.640 -3.579 -3.579 -3.576 4. 00000 -9.626 -3.584 3.584 -3.579 4. 50000 9.619 3.589 -3.589 -3.580 5.00000 -9. 619 -3.595 595 -3.579 5. 50000 -9.625 -3.601 -3.601 -3.576 6. 00000 -9.635 -3.607 -3.607 -3.573 6. 50000 -9.647 -3.613 -3.613 -3.571
-9.660 -3.617 -3.617 3.570 -9.669 -3.620 3.620 -3.573
6 6 622 -3.622 -3.577
9.
It can at 6 GHz the return loss is to it was
that the transmission loss would be 9 a loss the
A 3PORT2.CKT was and can be found in
Appendix H. circuit model contained quarter wavelength
50""2 == ~..:...L'=
shows the circuit and Table 10 contains the ----.-----
.7 Ohms so
69.
91
69 of
Univers
ity of
Cape Tow
n
l'FRE<i-GHZ
I 2. 00000 2. 50000 3. 00000 3. 50000
I 4. 00000 4. 50000 5. 00000 5. 50000 6. 00000 6. 50000 7. 00000 7. 50000 8. 00000
DBCS11l SPLIT
-10. 807 -11. 489 -12. 406 -13. 635 -15. 294 -17. 600 -21. 035 -27. 149 -46. 007 -26. 468 -20. 807 -17. 558 -__15. 349
DBC S12J SPLIT
-3. 449 -3. 392 -3. 331 -3. 269 -3. 213 -3. '164 -3. 127 -3. 104 -3. 099 -3. 112 -3. 142 -3. 189 -3. 247
DBCS13l SPLIT
-3. 449 -3. 392 -3. 331 -3. 269 -3. 213 -3. 164 -3. 127 -3. 104 -3. 099 -3. 112 -3. 142 -3. 189 -:-3. 247
DBCS23J,1
SPLIT I
-4.'4311
-4. 7841 -5. 123 -5. 4271 -5. 683 -5. 883
1
-6. 023 -6. 104 -6. 127 -6. 095
-6. 0.101 -5. 872 -5. 680
Table 10. Characteristics for three port with 70.7Q lines
Comparing table 10 and table 9, it can be seen that the return loss from port 1 has improved from -9.635 dB to -46 dB. The coupling slightly
between the from -3.607
input dB to
and output ports has tightened -3.099 dB, and the isolation has
increased from -3.573 dB to -6.127 dB.
It can be .seen that increasing the impedance of the two quarter wavelength transformers, has dramatically improved the return loss from the input port, with minor improvement to the isolation and coupling.
To complete the analysis of the splitter, a lOOQ resistor was added to the output stage. This is thus the standard single section power splitter. The results for the splitter have been shown in table 8. Comparing tables 8 and 10, it can be seen that the primary effect of the resistor is to increase the isolation from -6.127 dB to -55.993 dB. The return loss from port 1 has decreased slightly from -46.009 dB to -43.727 dB, as well as the
92
2.00000 10.807 -3,449 -3.449 -4,431 2.50000 -11.489 -3.392 -3.392 4. 784 3.00000 -12.406 -3.331 -3.331 -5.123 3.50000 -13.635 -3.269 -3.269 5, 427 4. 00000 15.294 -3,213 -3,213 -5.683 4.50000 -17.600 -3, '164 -3.164 -5.883 5. 00000 -21. 035 -3.127 -3.127 -6. 023 5.50000 -27.149 -3.104 3. 104 6. 104 6. 00000 46.007 3. 099 - 3. 099 -6.127 6.50000 -26.468 -3.112 -3.112 -6.095 7.00000 20.807 3. 142 -3.142 - 6. 0.101 7.50000 -17.558 -3.189 -3.189 5. 872
Table 10.
10 table 9, can seen ~hat the return loss from port 1 has improved from to ~~== coupling the input and output ports has tightened
increased from ~~~~~
It can be ·seen that increasing the impedance the two quarter wavelength trans, improved return
from the input port, with minor improvement to the isolation and coupling.
To complete the added to the output section
of the splitter, a 1000 is thus
. The results for shown in table 8. Comparing tables 8 and 10, it can be
was single
have been seen that
the
92
to return loss from port 1 has to , as well as the
Univers
ity of
Cape Tow
n
transmission loss which has decreased from -3.099 dB to -3.264 dB. These values are still satisfactory for the operation of the power splitter.
8.2 Designing the 6 GHz power splitter/combiner In designing the power splitter it is important to note the effects of the following physical parameters on the splitters performance.
8.2.1 Physical constraints on power splitter operation a) Front end T-junction b) Coupling between conductors c) Non zero length isolating resistors d) Curved conductors
a) front end T-junction It has been shown c 1 s1,c 16 1 that the first discontinuity that a
signal sees on entering a power splitter is the t-end junction-. The signal splits into two even sections at this point. It is also at this point that the transmission line width changes. The front end t-junction effect was modelled on Touchstone. It was found that its main effect was to alter the VSWR of the input port.
b) Coupling between conductors After separation from the T-end junction, the two quarter wavelength sections must be separated by the length of an resistor. If these lines were parallel to each other, there would be coupling between them. The center frequency at which the' splitter operates will also be the frequency at which coupling is greatest. The effect of this coupling was to reduce the input VSWR.
The coupling between the lines can be reduced by making the
93
transmission loss which has decreased -3.099 dB to
power
performance.
8.2.1
a)
are satis of the
to note
b) conductors c) Non zero d) Curved conductors
a)
It that the first discontinuity that a s sees on a t-end junction.
signal splits into two even sections at this point. It is also at point that
end t-junction found that main
b)
After separation wavelength
transmission line width changes. The was modelled on was to alter
It was VSWR of the input
T-end junction, the two quarter length an
resistor. were parallel to other, would be coupling between them. The center at
greatest. VSWR.
operates will also be the frequency at which coupling of was to the
The coupling between the lines can by making
93
is
Univers
ity of
Cape Tow
n
spacing between adjacent conductors as great as possible, and to ensure that conductors are not parallel to each other.
d) Non zero length isolating resistors
It has already been indicated that non zero length resistors change the path length difference between ports 2 and 3. Signals travelling from port 2 to 3 via the quarter wavelength sections
should be 180° out of phase with the signals flowing through the
resistor. However, if the resistor length is non zero, the phase
difference will not be 180° and the isolation between the ports will decrease.
It has been shown that the input return loss and the isolation nulls occur at higher frequencies than the center frequency of operation of the power splitter.
A solution is thus to use resistors with the shortest possible
lengths.
d) Curved conductors It is known ci 7 1 that parallel transmission lines will couple to
each other, and a solution is to curve the conductors. Slow curves of fer less radiation loss at high frequencies than the sharp bends. It has.been suggested that a minimum radius of 2 or
3 ti.mes the line width would be used. A radius of approximately 5 times the line width was used in the final design.
Touchstone did not have an algorithm to model thus no analysis of the effects of curved made.
8.2.2 The schematic of the power splitter
curved conductors, conductors could be
Line widths and lengths were obtained from Appendix A, for the 6 GHz power splitter. These dimensions are given below for two impedances.
94
spacing between adjacent conductors as great as possible, and to ensure are not to other.
d)
It non zero the
ling should be 18
difference between ports 2 and 3. Signals 2 to 3 the
of with the signals flowing through the Hr~~~v,~r, if the length non zero, the phase
not 1800
will decrease.
It has shown that the input return nulls occur at higher operation of the power
frequencies than the center frequency of
A
lengths. to use
d) It each
leI transmission lines will couple to a is to curve
curves less radiation loss at high frequencies than the a 2 or
3 times the line width would be . A radius of approximately 5 was the final
Touchstone did not have an algorithm to model no analysis of the of
made.
8.2.2 The schematic of the power splitter were
curved conductors, conductors could
for the 6 GHz power splitter. These dimensions are given below two
94
Univers
ity of
Cape Tow
n
z<> (6 GHz) w (mm) L(mm) so 0.76 36.57 70.7 0.43 37.24
The lengths of the quarter wavelength sections were chosen to be (S/ 4)A instead of the customary A'/ 4 lengths. This was done to ease practical problems with the positioning of isolation resistors. A
quarter wavelength at 6 GHz was 9.31 mm. The resistor had a length of 1 mm. This was a substantial part of a quarter wavelength. It was decided that if the transformer lengths could be increased, then there would be less constraint on the positioning of the resistor. Thus more gentler curves could be used. This would also have the advantage that the conductors could be placed further apart, thus reducing coupling.
The circuit was tested using Touchstone. The program called
6GSPLIT2.CKT can be found in Appendix H. Various algorithms were used to determine the operation of the splitter more accurately. Losses due to the front end T-junction, the change in line width at the resistor node, and the effects of the resistors length,width and resistivity,. could be analysed. The results of this program can be seen in figure 70.
Comparing figures 70 and 66, it can be seen that the depths of the nulls for the isolation and return loss have been reduced. The results can be seen more clearly in Table 11. -
95
The lengths of the (5/4)A
practical problems with quarter wavelength at
of
wavelength ).: /4
were chosen to be was to ease
the positioning of isolation resistors. A 6 GHz was had a was a a
wavelength. It was decided that if the transformer lengths could be increased, then would constraint on the
used. could be
would also have
The circuit was tested 6GSPLIT2.CKT can be found
to the Losses due to the end at , length, width and this program can seen
Thus more the advantage ,
that curves be the conductors
using Touchstone. The program called Appendix H. Various algorithms were
of the spl more T-junction, the change in line width and the
could be analysed. The 70.
Comparing nul
70 and 66, can be seen that depths of isolation and return loss have been
can be seen more 11.
Univers
ity of
Cape Tow
n
11 FREQ-GHZ ·- -- --·
DBf S2~ DBC S11J DBC S12l DBCS13J SPLIT SPLIT SPLIT SPLIT I
-8. 930 I I,
--11. 057 -3. 584 -3. 604 \; 2. 00000 I 2. 50000 -9. 700 -3. 771 -3. 756 -4. 148 I i
3. 00000 -12. 018 -3. 597 -3. 574 -10.1001: 3. 50000 -21. 975 -3. 339 -3. 335 -15. 867/ 4. 00000 -17. 560 -3. 402 -3. 418 -14. 9651 4. 50000 -10. 995 -3. 697 -3. 750 -8. 4101 5. 00000 -9. 799 -3. 887 -3. 840 -4.888:-5. 50000 -12. 502 -3. 666 -3. 633 -10. 7071; 6. 00000 -24. 183 -3. 421 -3. 420 -16. 436 ! ' 6. 50000 -16. 514 -3. 505 -3. 538 -14. 368 7. 00000 -10. 893 -3. 780 -3. 873 -7. 830. 7. 50000 -10. 025 -3. 964 :..3_ 887 -5.727 1
8. 00000 -12_. 918 __ -:3- 719 -3. 683 -11: 1 9_1_ L--
Table 11. Characteristics of designed s12litter
A notch at approximately 3.6 GHz has been created by the (5/4.)A
length sections. It was found that the reduction in the input match and isolation was predominantly caused by the isolation resistors length and width. The effect of the increased transformer lengths was to decrease the bandwidth of the power splitter, at 6 GHz, and to allow operation at a lower frequency. The bandwidth of the power splitter was still adequate for the needs of the frequency multiplication system.
Thus the reduction overshadowed- by the
in the practical
usual transformer sections.
bandwidth of the system, is advantages of using larger than
96
DB! S11J DB! S12] DS!S131 DB[ S SPLI T SPLIT SPLIT SP
2. 00000 11.057 3. 584 -3.604 -8. 2.50000 9. 700 -3.771 3.756 4. 3. 00000 -12. 018 -3.597 -3.574 -10.100 3.50000 -21.975 -3.339 -3.335 -15.867 4. 00000 -17.560 -3.402 3. 418 14.965 4.50000 -10.995 -3.697 -3.750 8. 410 5. 00000 -9.799 -3.887 -3.840 -4.888, 5.50000 -12.502 3.666 -3.633 10.7071 6. 00000 24.183 -3.421 -3.420 -16.436! 6.50000 16.514 3.505 -3.538 -14.368 7. 00000 -10.893 -3.780 -3.873 -7.830, 7. 50000 -10. 025 -3.964 :... 3. 887 -5.727 1
-3·719 1
11.
A notch at approximately 3.6 GHz has been created by the (5/4j)..
It was found the input match and isolation was predominantly caused by the isolation
was to bandwidth transformer lengths splitter, at 6 GHz,
bandwidth of and to allow operation at a lower
was needs of the frequency multiplication system.
Thus the reduction in the bandwidth of the
usual transformer sections.
96
the
system, than
Univers
ity of
Cape Tow
n
, '~ I
I I
,' -•K [)
'" . ·, \ "~ :\.,
-~ ''\.
I ''\ .. ~
I J ~-.... ··1 ~
/ v--« ~ ...
I ( ·~/I /
-~~ [~ '\
~ '~
....... .......... 1---.c:i
-~ ) ...... -
.;!). lY .i:,. / l¢
/ /
·~ l '
""'·"\.. \
0 0 0 0 . . 0 ~ ~
I
97
I I I
I I
-- ---
0 0 0 a:)
N ::c t!J
16 LLJ a: LL..
0 0 0
• Ln
0 0 0 . \,
0 0 . 0 Ln
I
OJ I-....... -l C-oo to to I
C1 al O'l ....... tn tn . . - to OJ •• ....... .......
.......
........
CIJ c::: 0 ......, en .c::: u :::::J 0 I-
~ 0 en ttl
,...... (T) OJ I-00 .......... CD C'l
' ,
,...... (T) .......1-00 .......
CD a: CJ 00
<>
....... 00 ....... ........ ....J
2595
,...... ....... ........ 1-00 .......
-l
95
I
o o d .......
I
,,~ ,I< ~ 0
//- 0 0
.. K [~
I\"~\' \ ~ ~ ,'\,
'~ I •
N .'')' ~---- ZE 16
,;0. ~ UJ / V a::
~ LL.
~
<>~ V I I
/ 0
<~ ~ 0 0
• tn '\
~ ~
"'-. -..........
I---s-0- ) --- --
~
~i» lY~
"/ ~ -'
/ i~ W
~ \ . ~'
7
Univers
ity of
Cape Tow
n
8.2.3 Testing the power splitter
A power splitter was designed and constructed using ethic
described before. A positive print of the device can be found in
figure 71.
Figure 71. Positive print of power splitter
The characteristics of the power splitter were obtained from
analysis with a network analyser. Figures 72(a) .. (f) show the
return loss, isolation between the output ports and coupling
between input and output ports.
From figure 72(d) it can be· seen that there is a null at 6.5 GHz
(return loss -24 dB). The return loss at 6 GHz was approximately
~- The return loss is best at 6.5 GHz, because of the length and width of the isolation resistor. Even though the return loss is low at 6 GHz, the power splitter will be driven from a good
source (6 GHz oscillator) which has an output impedance of 50Q .
The isolation at 6 GHz was found to be 25 dB, with a 1.4 GHz
bandwidth null. This value is close to what is defin~d in the
literature as typical power splitter characteristics.
The coupling between the input port and output port 3, is shown
in figure 72(f). The coupling at 6 GHz- was found to be approximately 3 dB. It was noticed, however, that the coupling tended to vary between 2.6 dB and 4.5 dB over the band from 5 to
6 GHz.
98
8.2.3 A power splitter was designed described before. A positive print of the
using can found
11.
11.
The characteristics of the analysis with a network return ,
input and output
can
power splitter were obtained . Figures 12(a) .. (f)
output
'seen that there is a null at From (return -=~=-). The return at 6 GHz was
the
~. The return loss best at 6.S GHz, because of the length width
is low at 6 source (6 GHz
isolation
. Even though return GHz, the power splitter will be driven from a good
at 6 ) which has an output
GHz was found to be of SOQ.
bandwidth null. This value is close to with a defin~d in the
as
and 3, in figure 12(f). The coupling at 6 GHz' was found to be
""'---=~ • I twas , hnwpvg,r, that the
tended to vary between 2.6 dB and ~ __ =-6 GHz.
98
over the band from S to
Univers
ity of
Cape Tow
n
6 GHz splitter (match) - 1 5 .. -..
6 GHz SPiitter (isolation - 1 0
.. _ ..
r.1) (J)
0 ....J
z C1 ..... let:: LLJ (1)
z .....
. ~. ((I
'rJ .. _ ..
t~(I
((1
1::i ....J
z 1::t:: =I. t-LLI 0:::
1.5
'=' "' '- . ·-·
3.5
4.5
5.5
t:' - _, ··~-··
1~(1
(,.(1
"' ._1 .------. _1::.J --· ,)··"'
'......... ·~ ........ , ··-
1 e ._1
2 5
,t''
...,\I ,,...--·---·-....... 1 ,/...... \1, I/ \ ,/' j
I I \ / v \ • ...' 1.1
z 1::0 ..... let: UJ IJ)
:z: 3 c:' ._1 ·+--,-+---l,__-+---+-+---+--1---1---1 ...... .-,
.::: .6/div ::: F t· e ·=i. u e n c >" ( G_H z ) ___ _
-r. ~· 12.. lG\.)
6 GHz SPiitter (couPling)
....... CCI T.I ".J°
(j')
1'J")
0
............ ,.,~11./~tll I
,• 1°1 I '• 111 I
,..... ....... ! • ••• •• .. I I •._.••, , ...... " \ l1 I
...J '•' \ /1/·\ ...'\ ~_... I~' 1'
1 .. .1' •• -.J
·-..,...-·1 I - \~, l~ 1,J ~ \./ ::i
1-1.JJ et::
6 . 5 .l--.._--1,_....--+--+-_--l-.J-.-+--+--+--+--9_...,, 2 .f:::i . .."t".:Jl\11
·.e"ll.H?t'lO::•' <GHz) ~-------~
'•.,.J'
(J)
(J)
1::i ....J
:z 1::i ..... let:: LLJ (I)
z ....
6 GHz SPiitter (isolation -15
-5. ·~·
5
15
1_r, (J)
__...,.- C• ... ----- ;.J
··~,., _,...-f_ ....... ,, , z i \ ,,..- ' 0 ' \ ,// I \ I ~ ~/ ··....... . l
et: ......_ ____ ·--_. __ . ....___ ~····"···· 2 5 ---------· UJ
(.t')
z 3 s +--f----1--+--+--+--+-·-1~-+--+--.,."""s
5 . 2 ./di 1•.•
1 I
Fre·;.1 .. H.•nc>" t'.GHz)
95
--.. ---. '\, ....... -.. ........... "'· ..... ,
\, /"'' \ I
'• ... -- .. ,.... •, I ........ _..- 11 l
1'1•
1
\I ... / \ /' ........ . '-. .. "· I, I
.................... /
1 0'
20
4 0 +---1---+--+----+--+-·-+--+---l-+---i . 6 ..... di 11/
F r e ., 1J e n c >" ( G H z ) +1'3 7Z.(b)
6 GHz SPiitter (match) 4
8
1 2
16
/.: .. ·····----·--... !\.,.,., /_,,...
,./... ,G· µ ~ \... l / '•. I
/ \ I/ I \ •' \ l / I 11
I ' f
'\ // 20
24+---4~-1--~-1--.......+---+---l---+l\ .......... /+1~-1----1 5 .2/div 7
Fre·:i.1.~er1c:: .. · <GHz)
6 GHz sPlitter (couPlin~)
1
2
'1 .. ·""1 ···,, ... - .. \
3 \ / - ' -.... L '· /"·--·-.... ., \ r 1~ ·.,-. 11 (\ ti ·-.\ \ / / .... .)I "\
'~'· •,.) I
4
5
6.l---i---+--+----+--+---+---+--+---+---1 .2 . ...-div 7 5
Fre·:i.1.~enc~ .. · <GHz)
-r.~ 72(f)
r.l) (.I' o ...J
:z: o .... I-0:: IJJ (I)
:z:
-15
.~.
QJ
SJ -5 ',-,'
_:0 5 (.i~1
C' ..J
1 ;:::-
:z: '_' Ct:: :)
I- 25 LU 0:::
6 GHz sPlitter (ma1ch) .,-.. QJ
-0
'J') ( .. I)
-113
~3
,~ 113 ..J
z ,;:, 2121 .... I-0:: IJJ 3121 'J' z
7C' \ • .* --' 4 13 +-.......... --+--+----1----+--+1--4----10+-+____�
2 6./div 2 .6 di",!' .-, ,::;
6 6 GHz splitter (match) 1.5 4
.......
co -0 8 'J C' '- , .... , ......
'GI-\~ (f)
12 if) 3,5 0 ...J
4.5 16
5.5 2121
6 . 5 +-.......... _--+_+-----4I----+-+-----4I---I--+_...--; 2 8 .6/ d i' ... ' 7
'\ 1...1 e ~_'_"''_''_--''-~~-''--__ -~.
~~--------------- ------------------------6 GHz splitter (isola~ion 6 GHz
-15 1 "" co -0
-5 .~. 2 if)
oj
5 0 ...J
:z: 0 1 5 4 .... .... I-0:: IJJ 25 5 (f)
:z: ......
7 - C" .:,. ,_I
5 5 , 2 "".,j i v
-tlj 72 TlCf)
9
Univers
ity of
Cape Tow
n
This means that the power coupled from the input to the output port varies from between 36 and 55%. The power reaching the single ended mixer might be too low to switch the diode on. The oscillator is not expected to wander in frequency from 5 to 6 GHz, however. A more realistic spread of frequencies would be 5.8
to 6.4 GHz where the power split varies from 3 to 3.5 dB (45 to 50%). It is expected that the single ended mixer could cope with a 45% reduction in oscillator power, as explained in chapter 10.
Further simulation on Touchstone placement was very critical, and return loss. A second version of
indicated that resistor predominantly effected the the power splitter was
constructed, and the position of the resistor optimized to produce minimum return loss. It was found that a match of -14 dB could be attained with negligible change in the isolation and transmission losses. ' 8.2.4 Conclusions The power splitter will operate effectively over the band from 5.8 to 6.4 GHz. The coupling varying from 3 to 3.5 dB, the isolation from 10 dB to 25 dB and the return loss from 8 to 18 dB over this band. It has been indicated that the poor return loss at the center frequency is improved with correct resistor placement.
100
This means power coupled from input to the output and The power between
might is not expected
be too low to the diode on. The to wander
GHz, however. A more in from 5 to 6
frequencies would be to where the power split 3 to 3.5 dB (45 to
). It is that the single ended a 45% , as explained in chapter 10.
Further simulation was
on Touchstone critical, and
indicated res predominantly effected
return constructed,
A
and
be transmission losses.
8.2.4
5.8
power splitter will to 6.4 GHz. The
over this from 10 dB to
. It
position of . It was negligible
the that a match
change the ,
was to
operate effectively over the band from coupling varying from 3 to 3.5 , the
25 dB and the return loss from 8 to 18 dB
at the center placement.
frequency improved the poor return
with correct resistor
100
Univers
ity of
Cape Tow
n
CHAPTER 9
THE 6 GHz DIRECTIONAL COUPLER
From figure 46 it can be seen that
centered at 6 GHz, is reqtiired for the
at 6 GHz, will be reversed coupled to
while the main component of 3 GHz signal
to the 3 GHz horn antenna. The next
a directional coupler,
directional
techniqtie.
coupler and discusses a
transmitter unit. Power,
the single ended mixer,
is transmitted directly
section
computer analyses the
aided design
9.1 The theory of directional coupler operation
A directional coupler is a four port device consisting of two parallel transmission lines separated by a short distance. The
coupling length is usually set at a qtiarter wavelength
corresponding to the center freqtiency of operation.
It is known c 1 a1 that the lowest mode of propaga~ion along
parallel lines in a homogeneous media is a TEM mode. Due to the inhomogeneities in microstrip, TEM modes cannot be propagated
because of different phase velocities in the different media.
In microstrip the inhomogeneity is caused by the dielectric which
has a relative permittivity E= and a relative permeability µ~. The relative permeability is generally close to unity, since the
dielectric is non magnetic. Figure 73(a) shows a microstrip coupler, while figure 73(b) indicates the electric field distribution for even and odd mode excitation of the coupled iines.
ground
_J~~~~~~~~~~~~~~======~plane Figure 73(a). =Mi=·=c=r~o~s~t=r=1=·p,.._==...=.o;=-"-=-~l=1="n=e=s=
101
From 46 it can be seen coupler, at 6 GHz, for the unit. Power,
at 6 GHz, 1 coupled to the single ended , while the main component is transmitted directly
to 3 horn antenna. next section analyses the directional coupler and technique.
a
A
parallel coupler is a port device consisting two
by a short distance. The coupl length set at a corresponding to the center frequency of operation.
It known [3.8] the along a homogeneous media is a TEM mode. Due to
inhomogeneities in microstrip, TEK modes cannot in
In microstrip the inhomogeneity is caused by which and a
The permeability generally close to unity, since the is non 13(a) a
coupler, while figure 13(b) indicates the electric field even and
iines.
9rot.Ar'ld
73(a). .~~~~~Y~~S~~~~~~~~~~======-Plane
101
Univers
ity of
Cape Tow
n
. I
• even mode odd mode 1 __ /
Figure 73(b). Electric fields due to even and odd mode excitation
Even mode propagation results
amplitude are transmitted down
propagation results when waves
when signals of equal phase and the coupled lines. Odd mode
of equal amplitude but opposite phase are propagated down the coupled line.
Characteristic impedances and phase velocities in, even and odd modes may be found from the even and odd mode inductances and
capacitances. The inductance is computed from a medium with dielectric constant of one. coupling coefficient K1 is not dependent
permittivity but on the relative permeability.
the capacitance of Thus the inductive
on the relative
The coupled transmission lines can be modelled with capacitances and inductances, as in fi~r~_!4(a)&(bJ
I
~ Z:<O>
12 I 1 : ~ E2 Z2<1>
I t z:<o> 1 El
Ii : } ZHl>
-- -- Z=l Figure 74(a). Counled transmission lines
102
even modI!' odd mode
73(b).
Even propagation results amplitude are down
propagation results when waves of equal ampl phase are propagated down coupled line.
equal phase and Odd mode
opposite
Characteristic impedances and phase velocities in, even and odd may found from even and odd mode
a medium coupling
inductance computed from with dielectric constant of one.
coefficient K1 is not dependent but on relative permeability.
the capacitance of Tnus the inductive
on the
The coupled transmission lines can be modelled with and
I Z2(1) I
Univers
ity of
Cape Tow
n
------.
1C2 L2 !C2 L2
_r.""V"'.
Cm L1 ;=L~J =Cm ~
IC! ' IC!
Z=Z! Z=Z!+dZ
Figure 74(b). Equivalent circuit
Figure 74 indicates the lossless case of two coupled lines where
L~,C~ (j=l,2) are the self inductance and self capacitance per
unit length of line j. r._ and c_ are the mutual inductance and
mutual capacitance per unit length of line j.
It is known c 19 1, c 2 .0 1 that couplers using microstrip generally
have unequal inductive and capacitive coupling coefficients. The
exact values of these coefficients are important in the
calculation of the coupling coefficients as well as the
directivity of the coupler. In the next section the formulas
governing these coefficients will be discussed.
9.1.1 Determining the coupled mode formulas
It has been ~hown c 21 1 that empirical formulas exist for the
coupling coefficients, based on the spacing (S) between the
lines, the width (W) of the lines and the height (H) of the substrate. These formulas are indicated below.
Where:
A1 (E=) = 1 + (1/4)*Ln(E=+l)/2 A2 (µ=) 1 + (l/4)*Ln(µ=+l)/2
..(9.1)
. . ( 9 . 2 )
103
L2
2=Z1 Z=Zl+dZ
Figure 74 (b) .
Figure 74 case two where self inductance and Lj,C j (j==1,2) are the
unit length of line j. and em are the mutual inductance and mutual j.
It is known [19],[2.0] that couplers using microstrip
exact values of these coefficients are important in the
governing
9.1.1 It coupling
of coupling as as of coupler. In the next section
will be discussed.
coefficients, based on width (W) of the lines
the and
substrate. These formulas are
spacing (5) between the the height (H) of the
below.
::::: o. *exp [-(A1*S/H + B1*w/H) ] .. (9.1)
::: O. *exp [-( *S/H + *W/H) ] ..(9.2)
. . A1 ( ) == 1 + (1/4)*Ln( )12
(Jlr) ::::: 1 + ( 1 )*Ln( 1)
103
Univers
ity of
Cape Tow
n
Bi (Er) = ( 1I10 ) *v' Er+ I' B2(µr) = (l/lO)*v'µr+l'
These formulas indicate that the capacitive coupling coefficient K~ depends on the relative permittivity Er, while the inductive coefficient K1 varies with the relative permeability µr.
The directional coupler can be modelled in terms of wave dynamics, as s own in fiqu~e 75.
Z>(Q) } Z t l-ine 2
J_ · zico}
r0i ~ Z:O
Figure L 75. Various
z:1
modes-- on a couplea --transmission line
The arrows in figure 75 indicate the direction of flow of the various waves:
a_._(z) and b_._(z) are the complex amplitudes of the forward travelling waves on line 1 and 2 respectively. a_(z) and b_(z) are the complex amplitudes of the backward travelling waves. The power flowing at any point z along line 1 is given by a-+-(z) 2 /2
(power in forward direction). The power flow in the backward direction is given by a_(z) 2 /2. The average power on line 1 at any point z is then given by [ a_._(z) 2 - a_(z) 2 ]/2. Similar relationships hold for line 2 and the b_.__ modes.
It has been shown c 22 1 that the coupling coefficients and directivity are defined as follows:
104
( ) = (1/10)
( ) (1/10)
indicate that the coupling depends on permittivity, while the
coefficient varies with permeability ~r'
coupler can be modelled in terms wave
The arrows in figure 75 indicate the direction of flow the waves:
a~(z) and b+(z) are waves on
complex amplitudes of 1 2 a_(z}
forward (z)
are the complex amplitudes of the backward travelling waves. The power flowing at any point z along line 1 is given by a+(z) 2/2
( forward direction). The power in the backward direction is given by a_(z) 2/2. The on 1 at any z is given by [ a~(z) 2 - a_(z) 2]/2. Similar relationships hold 2 and
It been shown [22] that the coupling coefficients and are as
c
104
. .
Univers
ity of
Cape Tow
n
C(dB) = -20*Log C
D(dB) = -20*Log D
where _pi(z) is the voltage reflection coefficient of line i at position z.
These are the definitive cases for the coupling coefficients and
directivity, however it has been shown that there are approximations for various special cases. A case of particular interest is that of where Ki ¥ Kc , ~1 = ~2 (The case for microstrip). This case holds for a medium where the inductive and capacitive coupling coefficients are not equal, but the phase constants for the lines are equal.
The following definitions are used:
01 = v (I + Ki)*(l Kc)' •• (9.3)
02 = v (I - K1 )*(1 + Kc)' •• (9.4)
91 ~1 *01*1 •• (9.5)
92 = ~2 *02*1 • • ( 9 • 6 )
6 = Ki - Kc • • ( 9 • 7 )
1 =A vi - Ki 2' -<>--
4 *v Ex-E> ~ f · vi - Ki*~ •• (9.8)
and ~<> = ~:re-££ ' v1 Ki*K~ c vi - K 2' i
105
C(dB) -20*Log C
o
D(dB) O*Log 0
where ~(z) is the voltage reflection z.
of i at
These are the definitive cases for the coupling coeffic and directivity, however it has
of where
are·equal.
The following definitions are used:
c
lOS
been shown that there are cases. A case
are not , but the phase
•• (9.3)
•• (9.4)
•• (9.S)
.. (9.6) •• (9.7)
•. (9.8)
Univers
ity of
Cape Tow
n
where ~1 ,~2 = line phase constants
Generally, ~0*1 =- rr/4
coupler. for the transmission line directional
Using the formulas above, the mode amplitudes can be obtained. The mode amplitudes formulas can be found in Appendix I.
For this case, it is assumed that the coupler is matched at all
ports. When K1 # Kc, coupling is introduced between b+ and the a+
and b_ modes. Similarly, the a_ mode couples to b+ and b_.
9.2 Designing the 6 GHz directional coupler In the design of the directional coupler, the spacing between the
\
lines is important because it determines the tightness of the
coupling. This value is important since the amplitude of the
signal coupled from the transmitter will be used as an RF signal
for a single ended mixer. This signal may at worst be 30 dB below
the magnitude of the local oscillator (in this case 4mW).
It must also be noted that the minimum spacing between the lines
that could be etched, was 0.4 mm. This was due to constraints of the etching and cartographic equipment.
The first step in designing the coupler was to determine the
inductive and capacitive coupling coefficients. A program called
COUPLER.TRU (which can be found in Appendix I) was written. The
aim of the program was to calculate inductive and capacitive coefficients for various microstrip characteristics. The
following values were chosen to describe the microstrip and the parallel coupling lines.
E~ = 9.6 µr = 1
106
constants
transmission
Using the formulas above, mode ampl can be obtained. formulas can be found Appendix I.
For this case, it matched
. When K1 ~ , coupling introduced a+ and modes. Similarly, the a_ mode couples to b+ and
9.2 directional coupler, the spacing between the ,
it determines the of lines is important coupling. This value important since amplitude of the
signal 1 be as an RF for a single ended . This signal may at worst be 30 dB below
the magnitude of the local oscillator (in this case 4mW).
It must noted that the minimum spacing between the lines that be , was 0.4 Mm. was due to constraints of
etching and cartographic equipment.
The step designing the coupler was to the inductive and coupling ients. A program called COUPLER.TRU (which can found in Appendix I) was
of the program was to calculate inductive and capacitive
following were chosen to describe the micros trip and the
9.6
~r = 1
106
Univers
ity of
Cape Tow
n
height of dielectric h = 0.635 nun
characteristic impedance = 50Q
length of line 1 = 4.8 nun
spacing between lines s = 0.4 nun
width of line = 0.56 nun
center frequency = 6*109 Hz
The program produced the following results.
Kc = 0.1691 and Ki = 0.2586
Inserting these values in equations (9.3) .. (9.7), the following
results were obtained.
01 = v(I + K1 )*(1 Kc)' = 1.0226
02 = "< 1 - K1 )*(1 + Kc)'
= 0.9310
91 = ~1*01*1 = 1.6063
92 ~2*02*1
1. 4624 D. = Ki - Kc
= 0.0825
These values are then inserted into eq's (9.9) .. (9.13), giving:
I b---L.Q..l! .. (9.18) 1a ... ( o >I = 0.2136
lb ... .il.ll lb_( o )I = 0.3216 .. (9.19)
l_g_..(JU! la---(O)I = 0.0174 .. (9.20)
l_g ___ fil .. (9.21) ja ___ ( o )I = 0.9769
107
height of dielectric h :::: 0.6 ram impedance :::::: 50Q
1 == 4.8 ram
5 :::::: 0.4 ram width :::: o. ram center 6*10 9 Hz
The program produced the following
and K1 = 0.2586
(9.3) .• (9.7), the
into eq's (9.9) .. (9.13), giving:
.. (9.18)
.. (9.19)
.. (9.20)
.. (9.21)
107
Univers
ity of
Cape Tow
n
lh_._il.l! Ja_._(0)1 = 0.0689
C(dB) = S 1 2 = -20 * log (0.2136) = 13.4076 dB
D(dB) = -20*log (0.3216) = 9.8526 dB
S11 = -20*log (0.0174) 35.21 dB
S13 = -20*log (0.0689) = 23.24 dB
S14 = -20*log (0.9769) = 0.2027 dB
.. (9.22)
.. (9.23)
.. (9.24)
.. (9.25)
.. (9.26)
~-(9.27)
Inserting eq's (9.20) into (9.16), the following VSWR was obtained.
VSWR = 1 + 0.0174 = 1.035 1 - 0.0174
9.2.1 Simulating the 6 GHz directional coupler It was decided to model the coupler on Touchstone to determine if the coupling coefficients determined earlier were correct. The simulation would also give an indication of the couplers performance over a wide band from 5 to 7 GHz. A program called 6GCOUP2.CKT was written and can be found in Appendix I. The program analyses the directional coupler described in the previous section. A plot of its characteristics can be seen in figure 76, and a swmnary of the coefficients can be found in Table 12.
108
C(dB) = S12 = -20 * log (0.2136)
D( )::::: -20*10g (0.3216)
=
5 11 = (0.0174)
S13 -20*10g (0.0689)
; -20*10g (0.9769)
0.2027 dB
's (9.20) obtained.
VSWR =
9.2.1 It was
(9.16),
the coupling ficients determined would an
• • ( 9 . )
.• (9.23)
· . ( 9 . )
• • ( 9 • 25 )
• • ( 9 . )
~.(9.27)
VSWR was
to if were correct. The
performance over a wide band from 5 to 7 GHz. A program .CKT was can I. The
program the directional coupler in the A
figure 76, and a summary of 12.
its
108
can be seen in ficients can be found in
Univers
ity of
CapeTow
n- 0 (0
10.0
0
-20.
00
-50.
00 ~
EEso
f -
Touc
hsto
ne -
Sun
Feb
19 0
1:04
:19
1989
-6G
COUP
2
0 DB
[ S11
] +
DB
[ S21
] 9
DB[ S
31]
>::
DB[ S
41]
COUP
CO
UP
, · C
OUP
COUP
' '
, '.,
'.
, '
, .
' .
, ,
' ,
' ,,
,
' ,
' ,
' ,
' '
, /'
>.
,,
I I
' I
----
---+
----
-+
I I
I I
I
------
x-:
I I
I r-i
-,
I ,
~
~
L--
--+
___.
~
.---,
~
-._
.. -
·~
---
+
--
T
L-J
ri
--
A
v
L---
E ~
~
y
~ ~
~
~
~
~
v ~
!="1
GU
RE
76
nnn
s:;;
nnn
C'O
C'n
-r-U
7
I D
n
nn
v
, I-
"
N
~ 0 U (!) to i/
OJ
~ l0-
...... OJ ...... , ,I'
" .:c:t 0 ••
?
OJ ...... , ?
.c " QJ
lJ... 0; C :::J ~ en
........... I C\1 ';v
en .Ii' ........, CD C
?
"
, ?
f"
,
\ 09
~ p I
+[
I \
Tff I
n"
-I- rn
[~ t r [] ~ - [ ~
r [~
0
- (
\
~
\1 i\
8 @ I
~
\ \ \
~
\
\
0 0 0
cd
,
N
D5 I
CI
~ lJ...
0 0 0
0 0 0
~
0' o g
I
Univers
ity of
Cape Tow
n
i FREQ-G_H_Z~~D_B_C_S-111 1 COUP
' 2. 00000 -35. 558
II 2. 50000 -33. 932
3. 00000 -32. 736
I' 3. 50000 -31. 862
4. 00000 -31. 246 I· 4. 50000 -30. 849 / 5. 00000 -30. 646 / 5. 50000 -30. 620
I 6. 00000 -30. 764
I 6. 50000 -31. 068 ii: 7 . 0 0 0 0 0 - 3 1 . 5 2 4
7. 50000 -32. 103 8 . 0 0 0 ()__9__-:: 3 2 . 7 4 4
DBCS21J DBCS31J DBCS41J COUP COUP COUP
-19. 844 -18. 146 -16. 859 -15. 873
-30. 498 -0. 061 i -28. 580 -0. 0871' -27. 029 -0. 115 -25. 731 -0. 143
-15. 123 -24. 612 -0. 170 -14. 571 -23. 625 -0. 194 -14. 190 -22. 734 -0. 214 -13. 966 -21. 913 -0. 229 -13. 889 -21. 141 -0. 239 -13. 958 -20.-404 -0. 243 -14. 178 -19. 692 -0. 242 -14. 556 -15. 113
-18. 998 -0. 236'
-1 C!: } __ ~ 0 - ------= 0. 2 2il
Table 12. Coupling and reflection coefficients for coupler
Comparing the results of Table 12 with those of eqs (9.23) (9.27), it is found that they agree within a few percent of each other.
9.2.2 Construction and testing of the directional coupler A 6 GHz directional coupler was designed using the dimensions from the previous section. The active coupling lengths were set at 4.8 mm. The positive print of the coupler can be seen in figure 77.
u n
Figure 77·. Positive print of 6 GHz directional coupler
110
FRE COUP
I 2.00000 35.558 19.844 2.50000 -33.932 -18.146 -28.580 -0.087 3.00000 -32.736 16.859 -27.029 -0.115
I 3.50000 31.862 -15.873 -25,731 -0,143 I 4. 00000 -31.246 -15.123 -24.612 -0,170
Ii 4.50000 -30.849 -14.571 -23.625 -0.194
! 5.00000 -30.646 -14.190 -22.734 -0.214 , 5.50000 -30.620 -13.966 -21.913 O. 229
I, 6. 00000 -30.764 -13.889 21.141 -0.239 6.50000 31.068 -13.958 -20. '404 -0.243!
!, 7. 00000 -31.524 -14.178 -19.692 -0.242
'I 7.50000 -32.103 14.556 18.998 -0.236 -32.744, -15.113 -18 20 -0 2
12.
12 (5'-23) ..
(9.27), found that they agree within a few percent each
9.2.2 A 6 from previous section. The active coupling lengths were set at 4.8 mm. figure 77.
Figure 77.
u n
1
can be seen
Univers
ity of
Cape Tow
n
The coupler was attached to a network analyser and the coupling and reflection losses determined. The results of the test are
shown in figures 78(a) .. (d). These results are summarised in table 13.
Sii(dB) S12(dB) S13(dB) S14(dB) 16 16.3 16.7 0.6
Table 13 .. Measured coupling coefficients ~or coupler
The results from table 13 can be combined with those from table 12 and those from eq's (9.23) (9.27), as below in table 14.
Description S u. ( dB ) S12(dB) S13(dB) S14 (dB)
theory(T'stone) ..,..30.76 -13. 889· -21.141 -0.239
theory(eq's J -35.21 -13.408 -23.24 -0.203
actual -16 -16.3 -16.7 -0.6
Table 14. Comparing actual with predicted coupling values
As can be seen from table 14, the values obtained from the actual measurements of the coupler differed from what was expected. Only the coupling from ports 1 to 2 (S12) and 514 were close to their
predicted values.
It was discovered, however, that during the etching of the directional coupler, the line widths were reduced to below their
initially designed values. This was due to the effect of parasitic etching under the template. Figure 79 indicates this
phenomenon.
111
was attached to a The and shown table
losses determined. The results of the test are 18(a) .. (d). results are summarised in
Table
The 13 can
and 's (9.23) •. (9.21), as below in table 14.
-13.889-.21 .408
-16.3 -0.6
Table 14.
As can be s~en from table 14, the values obtained from measurements what was . Only
couplinq from ports 1 to 2 (512) and 514 were to
predicted
It was discovered,
initially
coupler, designed
etchinq phenomenon.
, that etchinq of the the line widths were reduced to below
values. was due to of
the template. Fiqure 19 indicates this
1
Univers
ity of
Cape Tow
n
.. -.. CJ:i ";'.) .. _ ..
.. -..
·._ ..
1 1
1 . ..., ·~'
1 5
6 GHz couPler (m~tch)
...... .,.-·... ,.•''11"1 ··· ... _ ......... ·· ........ ;
' ~~ I ·· ... __ ... ···., ..... / ··., .................... M,/ '\ /
·-·~··, ...
.. -.. 14.5
.. _ .. 1 5 . 5
, .. -1:1:. /.. (.(I
/ C1 / _J
16.5
" .. --........ ··' z
rCt:: w 1 ::i. 5 ((t
z
6 GHz couPler ($12)
1?. -+---·-1---4----·+--·--...._-·+-·-~---+-- .... 1 9 . 5 ___ __.._~
1 .., ·-·
1 5
1 7
1 9
21
5 .:2 . ...-di\11 7 5
o'
F t· e ·'f. '.A e n c ::-·· ( G H z )
.;-,~ 7iCo..)
6 GHz coupler <813) .. -.. co ·;:i
··-"
- 1)) ..... -··------· - .. ---....... __ ..... (J')
-------- -- _ .... ...........
, .... /-•. /
.....
.... -·-·""·"' .. •···
·' ..... -~ l=I
...J
z l=I
....... I-Ci w (1:1
z ·~
. 1
. 2 /di"··'.
6 GHz couPler 0::814)
2 3 +---11-----+--+-·-------------·-+-----l .,,. . 2 ..... ,J i \.' c._; 7 ·-' . ;2 ..... di 11.t'
F r· e ·:u.~ e n c >' ( G Ho.....;;zw)!.._ ___ -----:-- ------'--F-'t-'· "°':::..: _,·'l......::t...::?....::..1e..!....n!...:c"-"-::-"_.c_( ..:=:G!....!.f..!....~ ~z-'::_• --
112
7
z coupler (match) z co I..~ P 1 e r ( S 1 2 ) ...,. f ..... ,. 1 4 5
.'-.. co CD "0 9
"tI "./ 1 5 5
,._ .. ((I
(.(. 1 1 (I)
1=.
':'0 CI 1 6 5 -I
_J Z
Z 1 :3 ~=. 1 7 5
...... IL
=' l- i r.:"
W ,-' 0:::
r-0:::
5 w 1 Po -' ((t
Z
1 ? .....
1 ~3 c- , I I -+ ............... -,,.I co <, .. ' , :~: "" d i ',.' 7 5 2 ..... d i tfl .., ,
.. _-_ ..
f\~
6 GHz coupler (813) .. -... 1 3 ,' ..... 1
1 5
1 7
19
21
;2/d i v 7 5 .2/div 7 ---'--'--''''--''-''''-'''-'-'--''''' y (G ,-'-"''-''-___ --- ----~'---!...-=-~~~...!.---"-...:::..,!~-"----..
1
Univers
ity of
Cape Tow
n
/ template
, parasitic r-1 ---,....r:~~J.!l"A&d@momt~--...,1 et·=hing Tt
hr1
:1 f ~~ dielectric _
1
s 0ersr rtrie ;;;i--- substrate I i L_ ' ground
plane ·--- __J
Figure 79. Parasitic etching of copper below template
This parasitic etching reduces the width of the coupling lines. The reduction in width is proportional to the thickness of the copper. The problem in couplers is compounded by the fact that as the widths of the lines is reduced, so the gap between them increases (by twice the thickness of the copper strip). A program called 6GCOUP3.CKT {in Appendix I) was written which would simulate the effect of this reduction corresponding ~ncrease in gap spacing. program can be found in table 15.
in line width and The results of this
FREQ-GHZ DBC S111 DBC S211 DBC S31l DBC S41J COUP COUP COUP COUP
2. 00000 -26. 362 -20. 554 -29. 438 -0. 065i 2. 50000 -24. 724 -18. 870 -27. 641 -0. 091
3. 00000 -23. 508 -17. 599 -26. 232 -0. 120 3. 50000 -22. 603 -16. 630 4. 00000 -21. 945 -15. 898 4. 50000 -21. 494 -15. 364 5. 00000 -21. 226 -15. 001 5. 50000 -21. 126 -14. 796
-25. 094 -24. 152 -23. 353 -22. 659 -22. 034
-0. 148, -0. 1751 -0. 197! -0. 21.5:_ -o. 220'.
6. 00000 -21. 192 -14. 739 -21. 453 -0. 234 6. 50000 -21. 426 -14. 831 7. 00000 -21. 845 -15. 077 7. 50000 -22. 473 -15. 489
__J!_,__Q_90.QQ__ __ -=---~3. 3 5 5 ______ -1_6_. O!!~
-20. 890 -20. 328 -19. 751 -19. 152
-0. 2351'--o. 229. -0. 2191; -0. 206
Table 15. Effects of parasitic etching on coupling coefficients
113
r I
Figure 19.
This coupling The reduction width copper. The problem in couplers
the lines (by~twice the
to compounded by the
so the gap of copper
called simulate
6GCOUP 3 • CKT ( in the fect of
Appendix I) was written
this reduction in line
program can gap
in table 15.
2. 00000 2. 50000 3. 00000 3.50000 4. 00000 4.50000 5.00000 5.50000 6. 00000 6.50000 7. 00000 7.50000
26.362 -24.724 -23.508 -22.603 -21.9'+5 -21. '+94 -21.226 -21.126 -21.192 -21. '+26
21. 845 -22.473
.35
-20.554 18.870
-17.599 -16.630 -15.898 -15.364
. 001 -14.796 -14.739 -14. 831 -15.077 15.489
-16.08
-29.438 -27.641 -26.232 -25.09'+ -24.152 -23.353 -22.659 -22.034 -21.453 -20.890 -20.328 -19.751 -19.152
that as
). A program which would
width and
-0.17 -0. 19 -0. 21.5 -0.228: -0.23'+ 0.235
1'
-0.229. -0.219.
1
.
-0.206
Table 15. Effects of parasitic etching on coupling coefficients
113
Univers
ity of
Cape Tow
n
Comparing table 15 to table 14 it can be seen that the return loss now more accurately reflects what was actually measured (S11
reduced from -30.76 dB to -21.19 dB). The coupling coefficients
(S 12 and S 13 ) in table 14 are now more closely simulated. S12 has been reduced from -13.889 dB to -14.739 dB closer to the -16.3 dB actually measured. S13 and S 14 remained almost unaffected.
9.2.3 Conclusions
It has been shown that the directional coupler would couple to a backward travelling wave with a coefficient of 16.3 dB. This
reduced value of coupling was shown to be due to parasitic etching of the parallel transmission lines. Parasitic etching
reduces the coupling line widths, and consequently increases the gap between lines. In the next section it will be shown that this value of coupling is sufficient to allow 6 GHz second harmonic
signals to couple from the transmitter to the single ended mixer
shown in figure 46.
114
Comparing table 15 to table 14 it can be seen that the return now more was
reduced from -30.76 dB to 1.19 dB). The coupling coefficients ( and 5 13 ) 14 are now more closely simulated. 5 12 has
reduced -13.889 dB to -14.739 to .3 actually . 5 13 and 5 14 remained almost unaffected.
9.2.3 It has been shown that the directional coupler would couple to a
travelling wave with a of 16.3 dB. value of coupling was shown to be due to parasitic
etching the parallel transmission lines. etching reduces the coupling , gap lines. In the next section it will be shown that value of coupling to allow 6 GHz harmonic
to from the transmitter to the single ended mixer shown in figure 46.
114
Univers
ity of
Cape Tow
n
CHAPTER 10
DEVELOPING THE 3 GHz AND 6 GHz OSCILLATOR UNITS
In chapter 6 it was shown that the frequency would require a 3
GHz transmit oscillator and a 6 GHz local oscillator. The specifications for the units were as follows:
a) At least 10 mW output at center b) Varactor tuned over a band
bandwidth of the power splitters)
frequency
of at least 500 Mffz (half the
c) Fast tuning over this band, to ensure synchronisation of 6 GHz
local oscillator with the 3 GHz transmit oscillator d) For the 3 GHz oscillator, a second harmonic available at approximately 20 dB below the fundamental
Suitable oscillators in both the 3 GHz and 6 GHz frequency range,
were obtained. Both belonged to the Avantek VT0-8100 series of
varactor tuned oscillators. A description of the oscillators can be found in Appendix J. A summary of the oscillators' characteristics can be found in table 16.
Model .freq. range power output voltage harmonics no. (GHz) (dBm) (VDC) (dBc)
8240 2.4-3.7 +10 +15 -18 8520 5.2-6.1 +10 +15 -25
Table 16. Characteristics of varactor tuned oscillators
The oscillators produce exactly the power output required. They operate over bandwidths in excess of 500 MHz, and according to the specification sheet, these oscillators can sweep across their frequency band in less than one microsecond. They thus satisfy the specifications shown previously. The 3 GHz oscillator has a second harmonic available at -18 dBc from the fundamental.
115
.'#
In chapter 6 GHz transmit
was the frequency would a 3 oscillator and a 6 oscillator. The
units were as follows:
a) At least 10 mW output at center b) Varactor tuned over a band of at least 500 Mnz (half the bandwidth of the )
c) Fast tuning over this band, to ensure synchronisation 6 GHz oscillator with the 3 GHz transmit oscillator
d) For the 3 GHz oscillator, a approximately 20 dB below fundamental
Suitable oscillators in both the 3 were obtained. Both belonged to the
and 6 GHz AvantekVTO-8100 series
at
,
varactor . A ae!sc:r be found in Appendix J. A summary
16. of the
can oscillators'
no. 8240 8520
Table 16.
(GHz) 2. .1
5.2-6.1
found
output (dBm)
+10 +10
The oscillators produce exactly
harmonics (VDC) ( ) +15 -18
operate over bandwidths in excess of 500 KHz, and according to
/I' the specification frequency band in less the specifications shown
than one can across
microsecond. They thus satisfy . The 3 a
harmonic available at -18 dBc from the fundamental.
115
Univers
ity of
Cape Tow
n
In the next section, power supply designs will be described.
However, since both units require the same DC voltages, only the 3 GHz oscillator power supply will be described.
10.1 Designing the 3 GHz oscillator power supply
Figure 80 shows two views of the oscillator package. The 3 GHz
and 6 GHz oscillator have the same package dimensions .
tuning 11oltage
+DC volta.ge
.. --- ------,
gr-ound []::
Ohimo
Figure 80. Schematic of VT0-8100 series oscillators
From this figure it can be seen that the oscillator requires a 15
voe power supply, a ground rail, a tuning voltage unit and a
section of sog transmission line, which was connected to the
output pin.
The oscillator had to be mounted onto microstrip so as to provide a sog line for the output pin. It was decided that the power
supply (without transformer) would be mounted onto the micr~strip board so that the power supply and oscillator would be effected
by the same thermal transients.
Figure 81 shows the power supply which would provide a stable +15 voe supply (within+/- 5%).
116
In the next , power supply designs will be described. However, since both units same , only 3 GHz oscillator power supply will be described.
10.1
3 GHz and 6 GHz have the same package dimensions.
Figure 80.
From this figure VOC power supply,
can be seen that the oscillator requires a 15 a ground rail, a tuning voltage unit and a
50g was to the pin.
had to onto so as to provide a 50g line for the output pin. It was decided that the power supply (without ) would mounted onto board so that the power supply and oscillator would be effected by the same thermal transients.
Figure 81 shows the power supply which would provide a +15 VOC (within +/- 5%).
116
Univers
ity of
Cape Tow
n
LH 317
~--.-----rN OUT --------1-----<0Vout
1220 YAC Input
R 1 220 Ohms
30 VAC + Ci 1 uF
Ohms
Ohms
Figure 81. Oscillator power supply and. ground rail
The transformer and bridge rectifier were connected together on a
separate section of PCB. The voltage regulator and the assorted
resistors and capacitors were attached to the microstrip board
shown in figure 82.
Figure 82. Positive print of oscillator power supply
117
I 220 VAC. Input
81.
The
30 VAC
LH 317
r-----~N OIUT~-1~ __________________ -OVOyt
BR 1
lN41 Ohms
Ohms
uF
~------------------~------------------~~------~nd
and bridge rectifier were connected together on a
and capacitors were attached to the microstrip board
shown in figure 82.
82.
117
Univers
ity of
Cape Tow
n
A DC blocking capacitor was added to the SQQ output transmission line. A value of 220 pF was chosen for the chip capacitor. The reactance of the capacitor was found to be:
Xe = 1/w*c = l/(3*109*2*R*220*1Q-12)
= 0.24
Thus at 3 GHz the capacitor has a small reactance, and will thus have a very small effect on the fundamental frequency.
The phase lag produced by the capacitor is correspondingly small.
phase lag= tan-1 (0.24/50)
=~
The oscillator and power supply unit were connected together. A variable power supply was attached to the tuning line. In the next section the oscillators' output power will be measured and the tuning voltage will be determined.
10.2 Measuring the output power and tuning voltage of the
oscillator Figure 83 shows the test rig used to measure the output power of the oscillator and its tuning voltage.
po1.11rr supply
Vi-----. ~ s'°'eep r::.:::J generator
Figure 83. Test rig to measure oscillators output power and tuning voltage
118
A blocking was to . A value of 220 pF was chosen
was found to reactance of the
Xc ::::: l/w*c
= 1/(3*109 *«*220*10- 12 )
=
at 3 GHz has a
a very small on the
lag by the
phase lag tan 1 (0.24/50) :::::
the . ..
output transmission
chip . The
reactance I' and·
frequency. thus
correspondingly small.
The oscillator supply were CO~D.IILeC:1: .A
next
power
the the tuning voltage
10.2
was attached to
• output determined.
tuning .. In the
will be measured and
test rig to measure the of
and tuning voltage.
83.
r I
118
Univers
ity of
Cape Tow
n
The first step was to check that the oscillators power supply was
set to +15 V. The oscillator was then connected to the spectrum
analyser via SMA - coaxial connectors. By alternately attaching
the oscillator- from the spectrum analyser to the freq. counter,
it was determined that the oscillator produced a maximum
amplitude 3 GHz signal with the tuning voltage set at 10.03 volts.
A sweep generator was then attached to the spectrum analyser. It
was set to a center frequency of 3 GHz. The power of this signal
was varied until it had the same amplitude on the spectrum
analyser at the 3 GHz oscillator signal. The sweep generator was
then attached to the power meter, where the output power was
found to be 17.5 mW at 3.002 GHz.
Using this same technique the second harmonic was found to have
an amplitude of 0.8 mW.
Thus the second harmonic is:
-lO*Log (0.8/17.5) = -13.4 dBc below the fundamental frequency. This value is better than
anticipated by the oscillator specifications.
10.3 Designing the tuning unit of the oscillator It was shown in the previous section that the oscillator required
the tuning voltage to be set at 10.03 volts to produce the 3 GHz signal at maximum power. The tuning unit was thus deigned to tune
between 9.5 - 10.5 volts. A schematic of the tuning unit is shown
below in figure 84.
119
The
set to
was to check that was lator was then connected to the
connectors. By alternately attaching
to the freq. counter,
v. The
SMA -
the oscillator- from
was determined
amplitude 3 GHz that the oscillator produced a
with the tuning voltage set at 10.03
A sweep generator was then attached to the spectrum analyser. It was set to a center of 3 GHz. The power signal was until had the same
3 was
to the power where the output power was
found to be 17.5 mW at 3.002 GHz.
Us same the second harmonic was found to have
an amplitude of 0.8 mW.
. .
-10*Log (0.8/17.5) = below the than
oscillator specifications.
10.3 It was shown in the previous section that the
to be set at 10.03 volts to produce the 3 GHz
signal at maximum power. The tuning unit was thus to tune 9.5 - 10.5 . A of tuning shown
below figure 84.
119
Univers
ity of
Cape Tow
n
The tuning unit was constructed on PCB and mounted on top of the
microstrip board used by the power supply. This was done to ensure thermal uniformity with the oscillator and the power
supply. The input pins of the tuning unit were attached to the output of the voltage regulator used by the power supply, while the output of the tuning unit was attached to the tuning pin of the oscillator power supply.
The complete 3 GHz oscillator unit was tested and it was found to produce a 3 GHz signal at 17.5 mW with a signal at 6 GHz of 0.8 mW. The 6 GHz oscillator was constructed using the' same techniques as the 3 GHz oscillator. It was found to produce a
fundamental frequency signal (at 6 GHz) with an amplitude of 15 mW.
10.4 The 3 GHz power amplifier From figure 46 it can be
system requires a 3 GHz to boost the 3 GHz signal
seen that the frequency multiplication
power amplifier. This amplifier is used from 10 mW to 1 W. Reception of the
reflected 6 GHz harmonic from the density tracer, is improved if the transmit power is increased.
A power amplifier suitable for this application of increasing the transmit power from 10 mW to lW, is the minicircuits ZHL-42 broadband linear amplifier. Details of the amplifier are given below in table 17, with a full description in Appendix J.
Freq (GHz) Gain (dB) Maximum power (dBm)
min output input 0.7 - 4.2 30 +/- 1.0 +29 +10
Table 17. Characteristics of 3 GHz amplifier
As can be seen, this amplifier requires a maximum input power.of +10 dBm, and will produce a maximum output power of 1 w.
121
The tuning was constructed on PCB and mounted on of by power supply. This was to
ensure
supply~
output of
with oscillator and the power input pins the were attached to the
the voltage regulator used by the the
oscillator the tuning unit was attached to the tuning pin
supply.
The complete 3 oscillator was and was found to produce a 3 signal at 17.5 mW with a signal at 6 GHz of 0.8 mW. The 6 GHz was using the' same techniques as the 3 GHz oscillator. It was found to produce a
fundamental frequency (at 6 GHz) with an amplitude of 15 mW.
10.4 From figure 46 it can be seen that the frequency multiplication
a 3 GHz power . This used
to reflected 6
GHz mW to 1 w. of GHz harmonic from the density tracer, is improved
transmit power
A power amplifier suitable for this application increasing the from 10 mW to lW, the ZHL-42
broadband linear amplifier. Details of the are below in 17, with a full description in Appendix J.
Table 17.
As can +10 dBm,
be seen, this amplifier a maximum will produce a maximum output of 1 W.
1
power
Univers
ity of
Cape Tow
n
In the next chapter, the single ended mixer and sweep and lock
circuit will be discussed.
122
In the next chapter, single ended mixer and and lock
122
Univers
ity of
Cape Tow
n
CHAPTER 11
THE 6 GHz SINGLE ENDED MIXER
Figure 85 indicates the two single ended mixers required for the frequency multiplication system.
3 GHz to densit~ tracer
~r 2
6 GHz rro• densit~ tracer
Figure 85. Single ended mixers in the fregyency multiplication
system
The single ended mixer/power combiner unit in the receiver (mixer
2) section could have been replaced by a balanced mixer. This device would have ensured local oscillator AM noise reduction as
well as cancellation of the third order intermodulation products.
However, the double beam lead diodes needed for this mixer were
not available and were considered too expensive and delivery time
too long (12 weeks), to warrant the~r purchase. However, since
local oscillator noise is not a problem with this system' there
was no need to use balanced mixers. Thus single ended mixers were
used, as single mixer diodes were available.
The other advantage was that both the single ended mixer and the
power combiner were duplicates of units used in the local
oscillator feedback loop (mixer 1). This meant that only one design had to be made for each unit. A theoretical description of
a single ended mixers operation, will be given in the following section.
123
Figure indicates the two single ended for the
85. system
The in 2) could have been replaced by a balanced
would have ensured local oscillator AM noise reduction as of
However, the double beaa lead diodes not were too delivery too long ( weeks) , to warrant purchase. However, local oscillator noise is not a problem. this system, was no to use • Thus were used, as single mixer diodes were available.
other advantage was that both the single ended and the were used the
oscillator feedback loop (mixer 1). This meant that only one design had to made each unit. A theoretical description of a will in section.
123
Univers
ity of
Cape Tow
n
11.1 Theory of single ended mixer operation
A c231 single ended mixer is a single input, single output
device. It is designed to convert radio frequencies (RF) to an intermediate frequency (IF) as efficiently as possible. The reason for this is that amplication and detection is simplified at IF frequencies. The frequency conversion is produced by a mixer diode with a fast response and high cutoff frequency. The diode is turned on and off by the local oscillator signal.
Both the RF and LO signals enter the mixer via the input port. As
can be seen from figure 85, the 6 GHz second harmonic from the transmit oscillator is coupled to the RF local oscillator signal
(nominally 6 GHz+ 10.7 MHz IF). These two signals then enter the single ended mixer where they are mixed and the difference component (IF), is sent via the output (IF) port to the sweep and lock circuit.
The second single ended mixer produces the difference between the local oscillator ( 6 GHz+ 10.7 MHz) with the received signal (nominally 6 GHz). The two signals are initially combined with a power combiner ,described in .chapter 8, and then fed into the
mixers input port.
11.1.1 The operation of the mixer diode
Central to the operation of the single ended mixer is the mixer diode, which is a nonlinear device (there is a nonlinear relationship between the current induced in the diode and the
-voltage across it). This is indicated in figure 86.
124
11.1 A (23]
reason
at IF
to convert frequency (IF) as
this is
frequencies.
input, output frequencies (RF) to an
as and detection
conversion is produced by a diode with a
amplication
frequency
response cutoff • The
Both
can
lock
The
turned on and by the
RF and 1.0 enter the via the input • As
seen from , the 6 GHz harmonic the oscillator coupled to the RF local oscillator
6 GHz + 10.7 MHz IF). These two signals then enter the
ended mixer they are and the di
(IF), is sent the output «
single ended _~._~.~ produces
+ 10.7 MHz) ( 6
to
difference the received s
(nominally 6 GHz). The two signals are combined then
a
power combiner , chapter 8,
port.
11.1.1 The operation of the mixer diode
Central to the operation the single
diode, which is a nonlinear device between current
This indicated
124
(
mixer is the mixer is a nonlinear
diode
Univers
ity of
Cape Tow
n
v
Figure 86. V-I characteristic of diode
When an alternating voltage is applied to the diode, the current
takes the form as indicated in figure 86 along the positive
current (I) axis. It can be seen that there is less current
flowing during the negative half cycles than during the positive half cycles. There is thus a net positive current whicb has a magnitude related to that of ~he applied voltage. The current
through the diode can be expressed in terms of voltage via a
Taylor series (because of its nonlinear nature). Generally the current is related accurately by the square of the applied
voltage.
Analysed in terms of transmission line theory, the local
oscillator drives the diode into heavy forward conduction for nearly half a cycle and into reverse bias for the other half cycle. This causes the reflection coefficient on the line to vary periodically, as a function of time, as shown in figure 87.
I +1 -···-························-·········-····-···-····-· I rlF'l-----
f(J.) -1 -·· ··-·· ·······-··············-····-·········-···········
i..~-"4
Figure 87. Time dependent reflection coefficient
125
v
86.
When an takes
voltage applied to the diode, the current form as indicated in 86 the positive
current (I) It can be seen that there less current flowing during the negative half cycles than during the positive
through Taylor current voltage.
related to the diode
thus a that of can be
of
in terms of oscillator drives the nearly half a cycle and into
. This causes periodically, as a function of
87.
net current which has a applied current
expressed in terms voltage via a nonlinear nature). Generally the
by the of
theory, the local into heavy
reverse bias
125
, as shown in I
conduction the other
to figure 87.
Univers
ity of
Cape Tow
n
With the forward and reverse reflection coefficients given by:
rF = -1 + 2*(RB + Rs)/Zo •• (11.1)
•. (11.2)
and the conduction angle e given by:
9= 2*arc cos(VF/ (8*Z 0 *PL)
where RB = Barrier resistance
Rs = Epitaxy layer resistance
Cj = Junction capacitance
Zo = line impedance
VF = Forward voltage
PL = local oscillator power
The time periodic reflection coefficient is a square and can be expressed as a square wave, however if the RF signal voltage is applied, the voltage of the reflected signal can be calculated ~nd will be given by:
VR(t) = r(t)*V3 *cos(w*t) = ro*Vs*COS(w*t) + (l/2)*r1*Vs*(COS(WL - Ws)t + COS{WL +
Ws}t) + •• (11.3)
The important term difference frequency frequency to the efficiency.
1 = Pri/Ps
is that involving WL w3 , this is the (IF). The ratio of reflected power at this incident power at w3 is the conversion
= ( 0 • 5 * rJ *Vs ) 2 /Vs 2
= r1 214 •• (11.4)
For the ideal case the conversion angle is approx.iinately 180°,
126
With the forward and reverse
::: + 2*(RB + )/Zo
rR = 1 - 2*Zo*w2*C j 2
and e
6= 2*arc cos(Vp/ (S*Zo*PL)
::::::
RS Epitaxy resistance Cj = Junction Zo =
= Forward voltage PL = local oscillator power
reflection coefficient expressed as a square wave, however if the applied, the of the ~nd will be given by:
(t) r(t)*vs*COS(w*t)
given by:
.. (11.1)
•. (11.2)
a square and can be
RF signal voltage is can
= ro*Vs*COS(w*t) + (i/2)*r1*Vs*[COS(WL - ws)t + COS(WL + ws)t] + •. (11.3)
The important term is that involving WL Ws , this the difference frequency (IF). frequency to incident power at Ws is
at this the conversion
1 ::: Pli/ps
= (0. rl *v s ) 2/v s 2
= r1 2 /4 .. (11.4) For the ideal case the angle approxiinately 1800 ,
Univers
ity of
Cape Tow
n
and the conversion loss becomes 3.9 dB. This value is usually worse since most mixer diodes do angles. They vary between 120°
not have such large conduction and 170°. The smaller the
conduction angle, the worse the conversion loss.
Since the maximum gain of a diode is unity, and it is a square law device, as the signal strength decreases so its gain decreases. This gain factor is important in determining how efficiently a mixer diode is operating and is hence related to the conversion efficiency which is the ratio of the power of the IF signal to that of the local oscillator.
Generally a problem exists whereby not only will a,RF signal of frequency fx.o + f xF produce an intermediate frequency of fxp , but also that with a frequency of fx.o - fxp• This is known as the image rejection problem. For the frequency multiplication system, where the local oscillator and the transmit oscillator are free running, the system would lose.information if one were rejected. Thus image rejection filtering suitable in this context.
11.2 The basic mixer circuit
of the signals would not be
Figure 88(a) shows c 24 1,c 2 s1 in ·equivalent circuit of a basic
mixer system.
127
(
and the conversion loss becomes 3.9 dB. This value is usually worse most diodes not have such large conduction angles. They vary between 1200 1700 •
conduction angle, the worse the conversion loss.
the maximum of a unity, and it a law device, as the signal strength decreases so its gain
This gain important in determining how a diode hence to
the conversion efficiency which is the ratio of the power of the IF signal to of
Generallya problem exists whereby not only will a,RF signal frequency + produce an intermediate frequency , but with a is known as the image rejection problem. For the frequency multiplication system, where the local are
were rejected.
11.2
would if one of the signals Thus image rejection filtering would not be
context.
Figure 88(a) shows (24],(25] a basic mixer system.
127
/ I
Univers
ity of
Cape Tow
n
I j9.c.eJ.. .. 9s.rn;J.aL9: :x.9Hlll-. .t.BaliSfil.t.Tl'•
l : :
!il~- .. f;l.9f;~ ! CRYSTAL ! ! ...................................... :
I
. . . . .
RF' CHOKE RF BYPASS IF choke IF' LOAD adllli. ttance Yb
. . . . ---· ....................... -... -............. --------- .. --- ...... --....... --............................. -··-
Figure 88(a). Eg.u.ivalent circuit of a basic mixer (ref 23)
Each unit described in figure 88(a) can be compared to a similar
unit in the frequency multiplication system. The signal
transmitter is · comparable to the transmitter/coupler unit in
figure 46. The local oscillator in figure 88(a) has an equivalent
unit in the receiver feedback loop. The RF choke, RF bypass and
the crystal are all contained in the single ended mixer. Figure
88(b) shows the frequency multiplication equivalent of the basic
mixer circuit.
Tran-smittttr antenna
---- ---·---
Figure 88(b). Basic mixer circuit
128
receiver
.1.,..9.C.eL •• 9.s.C.t.I,;J,.J.;I.L9J~. .s1.9.f!4B.I.... •• T.Mt.!I.s.~.U.tj::R ~ .. . . , : ; . ·
RF BYPASS IF choke Vb
; yy !
: · RF CHOKE:
i · ! .............. .: : ............. "'''' .. * .......... e .. ++ .... ,. ........ '''~~ .. ..,..:
---------------- -_ .... _------------------------------------
Figure 88(a). Egpivalent circuit of a basic mixer (ref 23)
in 88(a) can to a unit the frequency multiplication system. The signal transmitter is . comparable to transmitter/coupler unit in
46. 88( a) an unit in the receiver feedback loop. The RF choke, RF and the crystal are all contained the single ended mixer. Figure 88(b) the equivalent of
1,1onol ~n.
88(b) •
Univers
ity of
Cape Tow
n
From figure 88(a) it can be seen that the signal generator is represented by a current generator i~ and an internal admittance Y~ while the local oscillator by a current generator i_ and an
internal admittance Y_ These two current generators are connected in parallel to the mixer diode, and hence to the output terminals of the IF output port which is connected to the IF amplifier.
The RF and IF chokes on either side of the diode provide a low resistance path to the rectified current. There. is thus no appreciable current flowing through the diode. It is thus not biased by the local oscillator. Hence Schottky mixer diodes (such
as the DMK 5068) would be suitable in this context. These diodes have been shown earlier to operate adequately at this frequency.
The microstrip biasing circuit for the single ended mixer is shown in figure 89.
I. ... ••••••• 1::-.~/..oj _____ ,
HIXER DIODE f p..s/4 >. s=RF si gna 1
F' Output
Figure 89. Hicrostrip single ended mixer biasing circuit
Comparing figures 88(a) and 89 it can be seen that the short circuited quarter wavelength in figure 89 is equivalent to the RF choke. Both units provide a low resistance for the RF rectified current.
The open circuited patch ensures that the RF and LO signals are not applied across the terminals of the IF output. This has the
129
From 88(a) can seen that the signal generator is by a current 'and an
while the local oscillator by a current generator and an
y- two current generators are connected parallel to to terminals of the IF output port which is connected to the IF
The RF and IF chokes on either side of the diode provide a low to current. no
appreciable current flowing through the diode. It thus not biased by the local oscillator. Hence Schottky mixer diodes (such as the DMK 5068) would in context. diodes
been shown earlier to operate adequately at this frequency.
biasing circuit for the single ended mixer shown in figure 89.
MIXER DI
RF'&LO C==== RF'
Figure 89.
ZOi
T i rs / 4
· · · 1-.-___ -...11
liignal
compar.ing figures 88(a) and 89 can be seen the circuited quarter wavelength in figure 89 is equivalent to the RF choke. Both units provide a low RF current.
The open circuited patch ensures that the RF and LO signals are not applied across IF output.
129 .
Univers
ity of
Cape Tow
n
same operation as the RF bypass capacitor, which ensures that the
RF voltaqes of both the siqnal and local oscillators are applied
across the diode. There is thus a hiqh impedance to RF and LO
siqnals but a lower resistance for IF siqnals.
In Appendix K it has been proved that the short circuit line acts
as an RF choke and that the low impedance open circuited stub is
a,hiqh impedance at RF/LO frequencies.
Fiqure 90 summarises the operation of the biasinq circuit .
......... .:Y..~---····-t l
HIXER DIODE .
L F Output
RF
Fiqure 90. Biasing circuit
At 6 GHz point A is a very hiqh impedance while point B is at
qround potential. This means that the RF bypass and RF choke
ensure that RF&LO siqnals are applied across the diode.
At 10.1 Miiz point A is at qround potential while point B is a
very ~iqh potential. Thus the diode is reverse biased at IF
frequencies, ensurinq that the IF siqnals flow throuqh the IF output'S..iine.
It miqht seem that the diode should be matched to the.
transmission line for all the harmonics of the fundamental
frequency, but this is not necessary. The fundamental frequency will have the hiqhest amplitude, thus the diode need only be
130
same operation as the RF bypass capacitor, which ensures that the RF of both and local are applied across the diode. There thus a high impedance to RF LO signals but a lower resistance for IF signals.
In Appendix K has been proved that the short circuit line acts as an RF and low open stub a,high impedance at RF/LO frequencies.
Figure 90
H1:XER DIODE
RF&LO ~----------7
L 90.
.......... .:Y..."J ....... ....
ZOl
1 · · 1"'/4 · P.-___ ..lIdi
At 6 GHz point A ground potential.
a high B at This means that the RF bypass and RF choke
ensure RF&LO are applied across the
At 10.1 MHz very ~igh
output~iine.
It might
A at ground potential. Thus the
that IF diode
B a is reverse biased at IF
flow through the IF
seem that the diode should be matched to the.
frequency, but this is not necessary. The fundamental frequency
130
Univers
ity of
Cape Tow
n
matched to the transmission line using a quarter wavelength stub,
the RF stub. Similarly the diode IF impedance should be matched
to the IF amplifier. However, the impedance at IF frequencies is
of the order of 100-200ll, this is close enough to the characteristic impedance of son to not warrant additional I.F.
matching.
11.3 Designing the 6 GHz single ended mixer'
A DMK 5068 mixer diode was used in the single ended mixer. From appendix c, it can be seen that the diode .has a high cutoff frequency of 500 GHz. It could thus be expected that the mixer
would have a low conversion loss.
It was assumed that the single ended mixer would be operating
with an RF signal of approximately 6 GHz, and a local oscillator at 6 GHz+ 10.7 MHz. From Appendix A the lengths and widths for
the mixer were obtained as follows:
w = 0.76 mm 1 = 36.57 mm
The single ended mixer was designed according to figure 89. A positive print of the completed mixer can be seen in figure 91.
Figure 91. Positive print of single ended mixer
131
matched to using a quarter wavelength stub, the RF stub. Similarly the diode IF to IF amplifier. However, the impedance at IF of 100-200g, this is close enough
is to the
I.F. characteristic impedance of 50g to not warrant matching_
11.3
A DMK 5068 diode was used in the ended .. From appendix C, it can be seen that the diode.has a high cutoff
500 GHz. It mixer would loss.
would operating 6 GHz, a
It was with an RF at 6 GHz + 10.7 MHz. From Appendix A the lengths and widths for the mixer were obtained as follows:
w = 0.76 mm 1= 36.57 mm
The was to .. A
positive print of the completed mixer can seen in figure 91.
Figure 91. Positive print of single ended mixer
131
Univers
ity of
Cape Tow
n
/
11.4 Testing the mixer It is possible to determine how efficiently a mixer diode is matched to the transmission line by measuring the reflected signal from the diode at a particular RF frequency, generally the local oscillator~ frequency. The lower the amplitude of the
reflected signal, the greater the match at that particular frequency. It is assumed that the LO signal strength is much greater than that of the RF signal.
It would seem that measuring the match of the mixer on an ordinary automatic network analyser (such as the HP 8410 B) would· be the solution. However, network analysers such as the 8410
produce very low power (in order of microwatts) test port frequencies. This is far too low. The match of the mixer must be
determined while the diode is hard on, and is rectifying. The minimum local oscillator power necessary to turn the diode is generally about 1 mW. For this reason a scalar network analyser
was used to measure the match. Scalar network analysers use an external source of power, in the experiment a sweep oscillator was used. Figure 92 shows a test rig to measure the reflected power from the single ended mixer.
132
/
11.4 It possible to how a diode matched to the transmission line by measuring the reflected signal from the at a generally the local oscillator-' frequency. The the amplitude
signal, the greater the match at that particular . It LO
greater than that of the RF signal.
It would seem that match on an ordinary automatic network analyser (such as the HP 8410 B) would
such as 8410 (in order of ) test low
frequencies. This determined while
far too low. The match of the mixer must be
on, minimum local oscillator power necessary to turn the diode is generally about 1 mW. For this reason a scalar network analyser was to measure external source of power,
used. Figure 92 shows power from ended
use an in the experiment a sweep oscillator
a test rig to measure
132
Univers
ity of
Cape Tow
n
uept a....,litude nalyser P 8350 B
c:h A R f
ueep ascillatar i..-~~~~~~~~~~~--L:p 875S C
blank Cneg)
inpu
ef •
"'eep aut
directional c:aupler
inpu
Figure 92. Test rig to measure the match of the mixer
11.5 Results
a1..1er ter
435 A
The mixer was attached at the test port of the directional coupler and the reflected power over a frequency band from 5.2 GHz to 6.5 GHz was determined and can be seen in figure 93.
1u:-ructt.J t...o~S
"" ( J.@,) 0
+S
+10
+I~
Figure 93. Match of single ended mixer
.133
c:::h A R f
Figure 92.
11.5 Results
direction<'iI1 coupler
inpl.!
was at test of the
ower tel"
435 A
coupler and the reflected power over a frequency band 5.2 GHz to 6.5 GHz was and can seen 93.
1t.E:'TVIlN L..OioS , .... lJ.@,)· 0
+S
+\0
boO
93.
,133
Univers
ity of
Cape Tow
n
..
From figure 93 it can be seen that the match is best at 5.3 GHz
(-18 dB), while the match at 6 GHz is found to be -10 dB. It was found that the local oscillator had a power of 6.8 mW at 6 GHz.
Even though the mixer is best matched at 5.3 GHz, the mixer can still operate adequately with a match of 10 dB.
The isolation of the mixer to local oscillator signals was determined as follows. A sweep oscillator was connected to the
mixer LO port and set to 10 dBm at 6 GHz. The amount of LO power
leaking through the IF port was then measured. It was found that
the mixer had an isolation of 31 dB, which was more than adequate.
The conversion loss was measured by coupling an RF signal at 6.0107 GHz of -10 dBm, onto the transmission line. The IF power was measured in the same way as for the isolation test. The IF
power was found to be -29 dBm. The conversion loss is thus -29
dBm + 10 dBm = -19 dBm. This is quite low as conversion losses of
-8 dB are typical. It has been proved that the diode is matched correctly to both the RF and IF ports, as indicated by the return losses and the isolation. The poor efficiency was caused by the
softening of the mixer diodes V-I characteristics. This was most likely caused by electrostatic· damage to the diode, during
h!llndling.
The second single ended mixer was constructed and was also found to have as poor a conversion efficiency. The results for the tests were:
Return loss = -9.5 dB Isolation = 30 dB
Conversion loss = -19.5 dB
As only two diodes were available, this version of the single ended mixer would have to be used for the system. Futher, the system should still perform adequately but with a reduced range
134
From figure 93 it can be seen that the match is best at 5.3 GHz
(-18 dB), while the match at 6 GHz is found to be -10 dB. It was found that the local oscillator had a power of 6.8 mW at 6 GHz.
Even though the mixer is best matched at 5.3 GHz, the mixer can still operate adequately with a match of 10 dB.
The isolation of the mixer to local oscillator signals was
determined as follows. A sweep oscillator was connected to the mixer LO port and set to 10 dBm at 6GHz. The amount of LO power
leaking through the IF port was then measured. It was found that
the mixer had an isolation of 31 dB, which was more than adequate.
The conversion loss was measured by coupling an RF signal at
6.0101 GHz of -10 dBm, onto the transmission line. The IF power was measured in the same way as for the isolation test. The IF
power was found to be -29 dOm. The conversion loss is thus -29 . dBm + 10 dBm = -19 dBm. This is quite low as conversion losses of
-8 dB are typical. It has been proved that the diode is matched correctly to both the RF and IF ports, as indicated by the return losses and the isolation. The poor efficiency was caused by the
softening of the mixer diodes V-I characteristics. This was most likely caused by electrostatic' damage to the diode, during
h~ndling.
The second single ended mixer was constructed and was also found to have as poor a conversion efficiency. The results for the tests were:
Return loss = -9.5 dB Isolation = 30 dB
Conversion loss = -19.5 dB
As only two diodes were available, this version of the single ended mixer would have to be used for the system. Futher, the system should still perform adequately but with a reduced range
134
Univers
ity of
Cape Tow
n
perfonnance with these mixers.
11.6 Conclusions
The single ended
frequency of 6 GHz.
mixer operates effectively at the
It has a satisfactory match of 10 dB, center
and an isolation of 31 dB. The mixers conversion loss is poor, between-
19 dB and -19.5 dB, but due to lack of mixer diodes, these will be used for the system.
11.7 Sweep and lock circuit It has been shown in Chapter 6 how the local oscillator will track the second harmonic of the transmit oscillator using a
sweep and lock circuit. The sweep and lock circuit used was from
a Plessey S.A. MRA-7 distance measuring device.
A circuit diagram for the sweep and lock circuit can be found elsewhere [ 25 1. A detailed description of the circuits operation
will not be given here, save to indicate that it requires a 10.7 MHz tone at its input.Its output will drive the tuning voltage
pin of the local oscillator until the difference between its
output (nominally 6GHz + 10.7 MHz) and the second harmonic of the
transmit oscillator (nominally 6 GHz) is 10.7 MHz.
In the next section,
multiplication system be tested.
the final integration of the frequency will be performed, and the full system will
135
performance with these mixers.
11.6
The single ended mixer frequency of 6 GHz. It
of 31 dB.
operates effectively a satisfactory match
at the 10 dB, poor,
center and an
19 dB and -19.5 dB, but due to lack of mixer diodes, these will
6 how the track
sweep and
second harmonic of the transmit oscillator using a
lock circuit. The sweep and lock circuit used was from
a S.A. MRA-7 distance measuring
A circuit for and lock can found elsewhere [ 25 1. A detailed description of the circuits operation
not be given here, save to indicate that it requires a 10.7 MHz tone at .Its tuning pin of the local oscillator until the difference between its output (nominally 6GHz + 10.7 MHz) and the second harmonic of the
( 6) 10.7 MHz.
In the next section, multiplication be tested.
the final· integration of will be performed, and the
135
the frequency
Univers
ity of
Cape Tow
n
CHAPTER 12
FINAL INTEGRATION AND TESTING OF FREOQENCY KQLTIPLICATION SYSTEM
12.1 Determining the transmitter unit power It was decided that it was best to determine the operation of each sub unit before integrating the complete frequency multiplication system. The first unit to be tested was the transmitter unit including transmit oscillator, RF amplifier, directional coupler, 6 GHz notch filter and 3 GHz horn antenna. The test was to determine the power at point A in figure 94.
3 GHz osc. RF aq:o
direction.al coupler
6 GHz not ch f il 't er
t
Figure 94. Test rig to measure transmitter unit power
3 GHz horn
The transmitter unit was connected together as in figure 94. The transait oscillator was powered up and its output connected to a HP 53518 frequency counter. It was found to be producing a fundamental frequency of 3.03 GHz. The oscillator was then connected to a spectrum analyser and the amplitude of the signal noted. Another signal was connected via a sweep analyser and its amplitude adjusted until it had the same magnitude as the output of the oscillator. The sweep generator was then attached to a
136
12.1
the operation of unit before integrating the
The first unit to complete be tested was the
transmitter unit including , RF directional coupler, 6 GHz notch filter and 3 GHz horn antenna.
test was to
3 GHz osc.
94.
direction.al coupler
The unit was trans.Lt oscillator was powered UP counter. It fundamental frequency 3.03
at point A in figure 94.
6 GHz 3 GHz horn notch r U ier
t
as 94. The up and its output connected to a
was found to a GHz. The oscillator was then
and the noted. signal was connected via a sweep analyser and its
the same as the output of the oscillator. The sweep generator was then attached to a
136
Univers
ity of
Cape Tow
n
power meter, where its output power was found to be 11.3 dBm.
The power amplifier stipulates
+10 dBm, for this reason a 3
that the input should not exceed
dB attenuator was between the
amplifier and the transmit oscillator. The power at point A in
figure 94, was then measured. Using a similar measurement
technique, the power was found to be +28 dBm or 631 mW. This was
considered an adequate value.
This amplitude was lower than the 30 dBm expected, because of the losses in the coupler and the 6 GHz notch filter. The power out of the amplifier was measured to be 30.4 dBm at 3 GHz.
12.2 Measuring the receiver unit IF output power Figure 95 shows the receiver unit with power splitter and power
combiner, single ended mixer and 3 GHz notch filter, and receive signal and local oscillator signal provided by sweep generators.
--------------------
local
pa1.1er splii:ter
ta directional coupler
><Rr
-, I
6 GHz horn I I I
f'il ter
I _______________________________ _:::_:____:_:c_:: ________ _,
Figure 95. IF power measurement test rig
The output of the single ended mixer was attached to a low frequency spectrum analyser ( HP 8590A). The sweep generator was attached to a frequency counter, and its frequency altered by
137
where power was found to be
The amplifier input should not exceed +10 dBm, for this reason a 3
that
dB attenuator was the
of
and the 94, was
oscillator. The
measured. a power at point A in
measurement , the power was found to
an adequate value.
similar
or was
amplitude was
in the coupler amplifier was
single local
local
the 30 and the 6 GHz notch filter. The
to be at 3 GHz.
and 3
signal notch filter,
by
to rec::t.tonal cOUP ell"'
the out
and power receive
not c:h ,. Hi er
• IF power measurement test rig
of the ended mixer was attached to a low spectrum ( HP 8590A). The sweep was
attached to a frequency counter, and frequency altered by
131
Univers
ity of
Cape Tow
n
The system was powered up and the output of the receivers single ended mixer was connected to an HP 8590A spectrum analyser. An IF signal at 10.7 MHz was recovered with an amplitude of
between .=.6.1 and -63 dBm.. The cross dipole antenna was positioned 30 cm from the transmit and receive antennas.
This is adequate for detection of the cross dipole antenna.
12.4 Improving system performance It has been shown that the single ended mixer designed previously has a low conversion efficiency. The effect of this is to reduce the magnitude of the IF signal in the receiver unit. After
investigation, it was found that the mixer diodes V-I characteristic had softened, and that this was the cause of the poor conversion efficiency. It was decided that a commercial balanced mixer should replace the power combiner and single ended mixer in the receiver unit. The balanced mixer had an improved conversion loss and Would reduce local oscillator noise and third order intermodulation products.
An ANZAC MD-162 balanced mixer was chosen. It operated over a band from 3 to 10 GHz, with a conversion efficiency of better than 7.5 dB over this range·(-6 dB at 6 GHz). Details of the balanced mixer can be found in Appendix L.
The balanced mixer was placed in the receiver unit as shown in figure 96.
139
The system. was powered up and the output of receivers single ended was to an HP 8590A analyser. An IF at 10.7 MHz was an between and -fiJ dBm. The cross dipole antenna was positioned
from and antennas.
This is adequate detection of the cross dipole antenna.
12.4 It has been shown that the single ended mixer designed previously
a to reduce the IF signal
investigation, it was found mixer diodes V-I characteristic softened, and cause the poor conversion efficiency. It was decided that a commercial balanced should replace the power combiner and single ended
in the had an conversion loss and Would reduce local oscillator noise and third order intermodulation products.
An ANZAC MD-162 balanced mixer was chosen. It operated over a band from 3 to 10 GHz, with a conversion efficiency of better
7.5 dB over -(-6 dB at 6 GHz). of the balanced mixer can be found
The balanced mixer was placed figure 96.
Appendix L.
the receiver unit as shown in
139
Univers
ity of
Cape Tow
n
from local osc.
balanced mixer
F a11P
6 GHz horn
3 GHz f' il-ter
i i
IF OYtput _ _J Figure 96. Balanced mixer position in receive unit
The IF output power was measured and found to be -48 dBm with the
density tracer placed 30 cm from the receive antenna. The system
performance has thus improved by 61-48 = 13 dBm. The system could
thus detect the density tracer to a greater distance, as shown below:
Pr = PR/r4 since the received power has increased by 13 dBm or 19.95 .
The increase in range is thus given by:
(19.65)0.25*30 cm= 63 cm
This is in excess of the original 20 cm specification. As can be seen the increase in system performance is 13 dBm, which is the same as the increase in conversion efficiency:
19 dBm - 6 dBm = 13 dBm.
The increase in received power at IF, is thus due to the increased conversion efficiency of the balanced mix~r.
140
from loc:al Otic:.
96.
balanced rn.ixer 6 GHz horn
;3 GHz f"il1:er
! i
~
IF was found to -.:l!;..:!:1:-~= with density tracer the antenna. The performance has thus improved by ~6~1~-~4~8~=~1~3,-d=~Bm=. The system could thus the density tracer to a distance, as shown below:
s
The
This is As can
which
The
power
in range thus given by:
excess of the original =2=0 __ = seen
dam or
same as increase in conversion efficiency:
19 dam 6 dBm ::: 13 dam.
received power
increased conversion at IF, is thus due the balanced mix~r.
140
dBm,
to the
Univers
ity of
Cape Tow
n
12.5 Counting the density tracers
The counter unit follows the IF amplifier in the receiver as shown in figure 97.
local P<>Yer
directional COUPier-
osc. !!.pl i tter-
counter unit
6 GHz antenna
filter
Figure 97. Counter unit position in the receiver unit
The output of the IF amplifier is passed through a FM demodulator
where the IF is detected and a pulse of DC is produced. The IF amplifier and FM demodulator are included with the sweep and lock circuit but will not be discussed here c25 1. The DC pulse is then
sent to a hysteretic comparator which changes the slowly changing
pulse into a square pulse. This pulse is then sent to MC 14553 B
three digit counter which counts the pulses and sends the number
to a seven segment display. The counter circuitry used here can
be found in Appendix L.
12.6 Conclusions It has been shown that the frequency multiplication system can detect a 32 mm density tracer with adequate sensitivity at the
specif~ed distance of 20 cm. Techniques for improving the sensit:.ivity of detection have been discussed. A circuit to count the IP,i.;., pulses and hence the number of density tracers, has been
described.
141
12.5 The counter shown 97.
97.
IF
dlrec:1:1onal coupler
in as
The output of the IF amplifier is through a FH demodulator IF and a pulse of The IF
amplifier and PH demodulator are included with the sweep and lock circuit but will not be [26]. The DC pulse sent to a comparator which changes the slowly changing pulse into a square pulse. This pulse then sent to He 14553 B three digit counter which counts pulses to a seven display. counter used be found in Appendix L •
• 6 It has been shown that the frequency multiplication
a 32 .. specif~ed distance
tracer of em.
sensitivity at for improving
can
can
the u...II..o ...... u:to.c;u. A to count
and hence the number of density tracers, has
141
Univers
ity of
Cape Tow
n
CHAPTER 13
CONCLUSIONS AND FUTURE WORK
13.1 Conclusions The thesis begins by introducing the user to dense medium separation and the techniques used to measure the separation efficiency of cyclones. It was shown that the separation efficiency is the efficiency whereby the cyclone separates desired product from undesirable reject material from the mine.
The problems associated with discussed and the indicator described. The disadvantage of
measuring this separation were used, the density tracer, was standard means of density tracer
counting was discussed, and techniques using active and passive means of detection was discussed. Passive detection techniques were considered suitable. Metal detection, microwave corner reflectors and frequency mixing and multiplication were attempted.
It was found that there were discrimination problems with both metal detection and microwave corner reflectors. Frequency multiplication and frequency· mixing were then attempted. Frequency mixing was found to be expensive, needing two transmitters and horn antennas. Fast detection was necassary hence the frequency mixing system required either crystal locked oscillators or fast sweep and lock circuits. Both these solutions are caaplex and expensive.
For these reasons frequency multiplication was investigated. This technique utilised a mixer diode characteristic of producing a second harmonic of an input signal if driven hard enough. The diode was mounted inside a density tracer on a dual frequency antenna. The density tracer received a transmitted signal and transmitted the second harmonic. This signal was detected by a
142
CONCLUSIONS AND FUTURE WORK
13.1
The thesis separation
ficiency efficiency
begins by introducing the techniques used to
of cyclones .. It was shown is the efficiency whereby
The problems associated with measuring and the indicator used,
user to dense medium measure the separation that the
the cyclone separates mine.
this separation were density tracer, was
tracer counting was discussed, and techniques using active and passive means of was were considered suitable.
and Metal detection, microwave corner mixing and multiplication were
attempted.
It was found metal detection multiplication
and and
was
were microwave frequency·
found to mixing transmitters and horn antennas. hence the
corner mixing
Fast
with
were then attempted. , two
detection was necassary
oscillators or fast sweep and lock circuits. Both these solutions are and
For reasons technique utilised second harmonic of
was a mixer diode characteristic an input signal driven
diode was mounted inside a density tracer on
of producing a hard enough. The a dual
antenna" The tracer a transmitted transmitted the second harmonic. This signal was detected by a
Univers
ity of
Cape Tow
n
receiver and indicated the passage of a density tracer.
A cross dipole antenna was designed. This antenna was mounted
inside the density tracer with the diode across its feedpoints.
A prototype system was designed and constructed. It utilized a
500 MHz transmitter unit and a 1 GHz receiver. The test proved succesful and for this reason a higher frequency system was constructed. The density tracer was reduced in size to that of
the specified 32 mm. This resulted in a system with a 3 GHz
transmitter and 6 GHz receiver.
A unique
utilized a which kept
frequency multiplication system was devised which
sweep and lock circuit in the receiver feedback loop
the receivers local oscillator locked onto the transmit oscillators second harmonic.
This technique made it possible to use free running oscillators
in both the transmitter and the reciver, a considerable saving had crystal locked oscillators been used. It also made it possible to reduce the frequency of detection to 10.7 MHz (FM
IF).
The bulk of this thesis was concerned with the construction of the compone~ts of the frequency multiplication system. Each of these components, shown below, were constructed and after testing proved adequate for their task.
a) 3/6 GHz cross dipole antenna
b) 3 and 6 GHz horn antennas
c) 3 and 6 GHz notch filters
d) 6 GHz power splitter/combiner
e) 6 GHz directional coupler
f) 3 and 6 GHz oscillator power supplies
g) 6 GHz single ended mixers
143
A cross dipole antenna was designed.
tracer with the A prototype system was designed and 500 KHz transmitter unit and a 1 GHz
a tracer.
This antenna was mounted
across constructed. It a
receiver. The test proved reason a was
constructed. The density tracer was reduced in size to that of the specified 32 mm. This resulted a with a 3 GHz
and 6 GHz
A unique a in the feedback loop
which kept the receivers local oscillator locked onto the
This technique made to use running
both the had crystal locked
, a considerable saving oscillators been used. It also made it
possible to
IF).
of to 10.7 KHz (FH
The bulk was with of the compone~ts of the frequency multiplication system. Each of
shown were and proved adequate for their task.
a) 3/6 cross dipole antenna
b) 3 and 6 GHz horn antennas
c) 3 and 6 GHz notch
d) 6 GHz power splitter/combiner
e) 6 GHz
f) 3 and 6 GHz oscillator power
g) 6 GHz single
143
Univers
ity of
Cape Tow
n
The diodes for the 6 GHz single ended mixers were found 'to have
softened V-I characteristics, possibly due to electrostatic
discharge during handling. However, it was decided to use them in
any event, since no other diodes were available. The system was
intergrated and it was found to operate adequately. The IF output
power was found to be -61 dBm with the single ended mixers. The
power combiner and single ended mixer in the receiver output
stage were replaced with a balanced mixer, and the IF output
amplitude rose to -48 dBm. This increase was found to be related
to the improvement in conversion efficiency.
1~.2 Recommend.ations The cross dipole antenna is polarisation sensitive, it would thus
be advisable to use circularly polarised antennas for the
transmitter and receiver. A future project could consist of the
design and construction of these antennas, as well as arrays of
these antennas, to ensure the detection of the density tracer in
any orientation. There is a 3 dB loss in antenna gain when using
circular polarisation, but the advantage of being able to detect the density tracer irrespective of orientation overshadows the
minor loss in IF signal strength.
Another field of research could be into the needed to protect the diode/cross dipole antenna.
need to be found whlch is durable to the rigours
plastic radomes
A material will
of the mining
.environment and also reasonably pervious to microwaves.
The use of this system in other industrial applications could be
investigated.
144
;_
The diodes the 6 single ended mixers were found 'to have
V-I , possibly to electrostatic
during handling. However, was to use in any event, since no other diodes were available. The system was
. The IF output it was found to
power was found to be IdBm power combiner and single ended
with a balanced
mixer in the receiver output
were mixer, and IF output
rose to -48 dOm. was found to
to the improvement conversion efficiency.
.2 cross antenna is , would
be advisable to use circularly polarised antennas for the
of the transmitter and • A and construction
these antennas, to ensure
these antennas, as well as arrays of
detection of the density tracer in
any orientation. is a J dB
the
in antenna gain when
of able to
the density tracer irrespective of orientation overshadows the
minor in IF signal strength.
field of research could be into the plastic radomes
needed to protect the diode/cross dipole antenna. A
need to found whi,ch durable to rigours of the mining
.environment and reasonably pervious to microwaves.
The use of this system in industrial applications could
144
Univers
ity of
Cape Tow
n
REFERENCES
1. Allan, F.F. 1985. Heavy media Separators. SHE mineral
processing handbook. Society of mining engineers. New York. pp
4.3-4.19.
2. Davis, J.J. 1986. Density tracers give option to floating sink
tests. Coal Age, Vol 91 no. 5, pp 60-62.
3. Wills, B.A. Heavy medium separation. Mineral processing
technology. Pergammon press. New York, Vol 29, pp 256-275.
4. Napier-Munn, T.J. March 1985. Use of density tracers for
determination of the Tromp curve for gravity separation
processes. Trans. Inst. Min. Metal!. pp c47-c53.
5. Scott, I.A. and Lyman, G.J. 1987. Metallurgical evaluation of
iron ore drum separators using density tracers. Bull. Proc.
Australas. Inst. Min. Metal!., Vol 202 no.1, pp 49-56.
6. Hall, J.S. 1947. Radar aids to navigation. MIT Radiation
laboratory series. McGraw-Hill, New York, 389p. ~
7. Dubost, G. 1981. Flat radiating dipoles and application to
arrays. Research studies press, John Wiley & sons, London, 103p.
145
REFERENCES
1. Allan, F.F. 1985. Heavy media Separators. 5MB mineral
New York .. pp
4.3-4.19.
2. Davis, J.J. 1986. Density tracers give option to floating sink
tests. ~~~~, Vol 91 no. 5, pp 60-62.
3. Wills, B.A. medium
~~~~~~. Pergammon press. New York, Vol 29, pp 256-275.
4. Napier-Munn, T.J. March 1985. Use of density tracers
determination of the Tromp curve
processes. • pp c47-c53.
5. Scott, I.A. and Lyman, G.J. 1987. Metallurgical evaluation of
iron ore drum separators using density tracers.
Australas. Inst. Min. Metall., Vol 202 nO.1, pp 49-56.
6. Ball, J.S. 1947.
laboratory series. McGraW-Hill, New York, 389p • • y
7. Dubost, G. 1981. Flat radiating dipoles and application to
, John Wiley & sons, London, 103p.
145
Univers
ity of
Cape Tow
n
8. James, J.R. et al. 1981. Microstrip antenna theory. IEE
electromagnetic waves series 12. Peter Peregrinus Ltd, London,
290p.
9. Compton, R.T. 1981. On the performance of a polarization
Sensitive Adaptive Array. IEEE-AP, Vol AP-29, no.5, pp 718-725
IO. Alberts, P.A. 1988. Development of an automatic system to
monitor the performance of a Dense Medium (mineral) Separation
process. Second joint symposium. on Antennas & Propagation and
Microwave Theory & Techniques, pp 13.1-13.5
II. Gandhi, O.P. 1984. Microwave Engineering & aQPlications.
Pergammon press, New York, 260p._
12. Kraus, J.D. 1985. Electromagnetics. Third edition. Singapore,
McGraw-Hill, 775p.
13. Silver, s. 1949. Microwave antenna theory and design. MIT
Radiation laboratory series. New York, McGraw-Hill, 623p.
-· 14. Cohn, S.B. 1968. A class of broadband three port TEM Mode
hybrids• IEEE-M'l".l', Vol MTT-16, no.2, pp 110-116.
15. Easter, B. 1975. The equivalent circuit of some microstrip
discontinuities. IEEE-M'.l'T, Vol MTT-23, no.8, pp 655-660.
146
8. James, J. et 81. Microstrip antenna theory. lEE
electromagnetic waves series 12. Peter Ltd, London,
290p.
9. Compton, R.T. 1981. On the performance of a polarization
Sensitive Adaptive Array. Vol , no.5, pp 118-125
10. AlJbe]~ts, P.A. 1988. of an to
monitor the performance of a Dense Medium (mineral) Separation
process.
.1-13.5
11. Gandhi, o.P. 1984. Microwave Engineering' applications.
Pergammon press, Hew York, 260p_.
12. Kraus, J.D. 1985. Electrgmagnetics. Third edition. Singapore,
, 11
13. Silver, S. 1949.
Radiation laboratory
-. 14. Cohn, S.B. 1968. A
• Hew York, McGraw-Hill, 623p.
of broadband
hybridS .. IEEE-IftT, vol KTT-16, no.2, pp 110-116.
15. B. equivalent some
discontinuities. IEBE-IftT, Vol KTT-23, no.8, pp 655-660.
146
TEM Mode
Univers
ity of
Cape Tow
n
16. Rachman, D.M. 1988. Computer aided design of broad-band
microstrip power dividers. Second joint symposium on Antennas &
Propagation and Microwave Theor;y & Techniques, pp 8.1-8.15.
17. Rachman, D.M. 1988. Computer aided design of broad-band
microstrip power dividers. Second joint symposium on Antennas &
Propagation and MicrowaYe Theor;y & Techniques, pp 8.1-8.15.
18. Krage, M.K. and Haddad, G.I. 1970. Characteristics of coupled
microstrip transmission. IEEE-MTT, Vol MTT-18, no.4, pp 217-228.
19. Krage, M.K. and Haddad, G.I. 1970. Characteristics of coupled
microstrip transmission. IEEE-MTT, Vol .M'l'T-18, no.4, pp 217-228.
20. Akhtarzad, s. et al. 1975. -The design of Coupled Microstrip
lines. IEEE-.MTT, Vol .M'l'T-23, no.6, pp 486-492.
21. Kal, s. et al. 1981. Empirical relations for capacitive and
inductive coupling coefficients of coupled microstrip lines.
IEEE-MTf, Vol M'l'T-29, no.4, pp 386-388.
22. Krage, M.K. and Haddad, G.I. 1970. Characteristics of coupled
microstrip transmission. IEEE-MTT, Vol .M'l'T-18, no.4, pp 217-228.
23. Pound, R.V. 1948. Microwave mixers. MIT Radiation laboratory
147
16. D.H. 1988.
microstrip power dividers. Second joint sxmPOsium on Antennas i
~QR~~~~~~~~~~~~~~~~~~~, pp 8.1-8.15.
17. Rachman, D.H. 1988. Computer aided design of broad-band
microstrip power dividers.
18. Krage, H.K. and Haddad, G.I. 1970. Characteristics of coupled
microstrip transmission. Vol HTT-18, no.4, pp 217-228.
I. 1970. coupled 19. Krage, H.K. and Haddad,
microstrip transmission. Vol HTT-18, no.4, pp 217-228.
20. Akhtarzad, s. et ale 1975. The design of Coupled Hicrostrip
Vol , no.6, pp 486-492.
21. Kal, S. et ale 1981. Empirical relations for capacitive and
inductive coupling coupled
IEEE-MfT, Vol ~-29, no.4, pp 386-388.
22. Krage, H.K. and Haddad, G.I. 1970. coupled
micros trip transmission. IEEE-HTt, Vol HTT-18, no.4, pp 217-228.
23. Pound, R.V. 1948. Microwaye mixers. HIT Radiation
Univers
ity of
Cape Tow
n
series. New York, McGraw-Hill, pp 19-28.
24. Pound, R.V. 1948. Microwave mixers. MIT Radiation labor~tory
series. New York, McGraw-Hill, pp 120-122.
25. Katoh, M~ and Akaiwa, Y. 4-GHz Interqrated circuit mixer.
IEEE-MTT, Vol .MTT-19, no.7, pp 634-637.
26. MRA-7 Operators manual, Plessey SA.
148
• New York, , 19-28.
24. Pound, R.V. 1948. Microwave mixers. MIT Radiation 1abor~tory
• New, , pp
25. Katoh, M. and Akaiwa, Y.
Vol , no.7,
26. MRA-7 manual,
4-GHz Intergrated circuit mixer.
634-637.
SA.
148
Univers
ity of
Cape Tow
n
APPENDIX A
LINE WIDTHS AND LENGTHS FOR VARIOUS FREQUENCIES
149 149
Univers
ity of
Cape Tow
n
LineCalc <TM) - Configuration< 400 1200 100 16517 1968 1000 1 l -Synthesis Tue Sep 26 01: 06: 14 1989
MSUB ER=4. 8 H=1. 6 T=O. 017 RHO=O. 84 RGH=O
TAND TAND=9e-4
FREQ F=1
DIM FREQ gHz LNG mm ANG deg
MLI N . W=2. 83 L=158. 43
Z0=50 Keff=3. 58 A,dB=O. 08
E< eff) =360
I so
Li neCalc (no Configuration( 400 1200 100 16517 1968 1000 1 ) , Synthesis Tue Sep 26 01: 06: 14 1989
HSUB ER=4.8 H= 6 T=0.017 RHO=0.84 RGH=O
TAND TAND=ge-4
FREQ F=1
DIM FREQ gH2 LNG mm ANG deg
HLIN . 1-1"'2.83 L=158.43
ZO=50 Keff=3.58 A,dB=0.08
E( err) =360
15
Univers
ity of
Cape Tow
n
LineCalc ( TMJ - Configuration< 400 1200 100 16517 1968 1000 1 l Synthesis Tue Sep 26 01: 06: 51 1989
MSUB ER=4. 8 H=1. 6 T=O. 017 RHO=O. 84 RGH=O
TAND TAND=9e-4
FREQ F=2
DIM FREQ gHz LNG mm ANG deg
MLI N . 11=2. 81 L=78. 92
Z0=50 Keff=3. 61 A, dB=O. 06
ECeffl=360
15~
LineCalc (TM) - Configuration( 400 1200 100 16517 1968 1000 1 ) Synthesis Tue Sap 26 01: 06: 51 1989
MSUB ER 4.8 H=1.6 T=O.017 RHO::0.84 RGH=O
TAND TANO=ge 4
FREQ F=2
DIM FREQ gHz LNG mm ANG deg
MLI N . 101=2.81 L 78.92
20=50 Keff=3.61 A, dB O. 06
E( err) 360
15:1-
Univers
ity of
Cape Tow
n
LineCalc (TM> - Configuration< 400 1200 100 16517 1968 1000 1 > Synthesis Tue Sep 26 01: 02: 35 1989
MSUB ER=10 H=. 635 T=O. 01778 RHO=O. 84 RGH=O
TAND TAND=9e-4
FREQ F=6
DIM FREQ gHz LNG mm ANG deg
MLI N . W=O. 57 L=4. 83
Z0=50 Keff=6. 70 A, dB=O. 02
E< eff> =90. 00
\52
LineCalc (TM) - Configuration( 400 1200 100 16517 1968 1000 1 ) S sis Tue Sep 26 01; 02: 35 1989
HSUB ER=10 H=.635 T=0.01778 RHO=O.84 RGH=O
TAND TAND=ge-4
FREQ F=6
DIM FREQ LNG mm ANG deg
HUN. W=O. 57 L=4. 83
ZO=50 Keff::::6.70 A, dB:O. 02
EC err) =90. 00
2
Univers
ity of
Cape Tow
n
LineCalc (TM> - Configuration< 400 1200 100 16517 1968 1000 1 ) Synthesis Tue Sep 26 01: 04: 07 1989
MSUB ER=2. 2 H=Q. 2548 T=O. 01778 RHO=O. 84 RGH=O
TAND TAND=9e-4
FREQ F=3
DIM FREQ gHz LNG mm ANG deg
MLI N . W=O. 76 L=18. 29
Z0=50 Keff=1. 87 A, dB=O. 04
EC effl =90. 00
153
LineCalc (TM) - Configuration( 400 1200 100 10517 1908 1000 1 ) Synthesis Tue Sep 20 01: 04: 07 1989
MSUB ER 2.2 H 0.2548 T=0.01778 RHO=0.84 RGH=D
TAND TAND=ge-4
FREQ F=3
DIM FREQ gH2 LNG mm ANG deg
MLl N . W=0.70 L=18.29
20=50 Kerf=1.87 A, dB 0.04
B( erf)· 90.00
\5
Univers
ity of
Cape Tow
n
LineCalc (TM> - Configuration( 400 1200 100 16517 1968 1000 1 l Synthesis Tue Sep 26 01: 04: 45 1989
MSUB ER=2. 2 H=O. 2548 T=O. 01778 RHO=O. 84 RGH=O
TAND TAND=9e-4
FREQ F=6
DIM FREQ gHz LNG mm ANG deg
MLIN .. W=O. 76 L=9. 14
Z0=50 Keff=1. 87 A, dB=O. 03
EC effl =90. 00
154-
LineCalc (TM) - Configuration( 400 1200 100 16517 1968 1000 1 ) thesis Tue Sep 26 01: 04: 45 1989
MSUB ER=2 2 H=0.2548 T=0.01778 RHO=0.84 RGH=O
TAND TAND=ge-4
FREQ F 6
DIM FREQ LNG mm ANG deg
MLI N . W=0.76 L=9.14
ZO=50 Keff=1.87 A, dB=O. 03
E( erf) =90.00
I 54-
Univers
ity of
Cape Tow
n
LineCalc (TM) - Configuration( 400 1200 100 16517 1968 1000 1 l Synthesis Tue Sep 26 01: 05: 21 1989
MSUB ER=2. 2 H=O. 2548 T=O. 01778 RHO=O. 84 RGH=O
TAND TAND=9e-4
FREQ F=6
DIM FREQ gHz LNG mm ANG deg
MLI N . W=O. 43 L=9. 31
Z0=70. 7 Keff=1. 80 A, dB=O. 04
E< effl =90. 00
155
LineCalc (no - Configuration( 400 1200 100 16517 1968 1000 1 1 Synthesis Tue Sep 26 01: 05: 21 1989
MSUB ER=2.2 H=0.2548 T 0.01778 RHO=0.84 RGH=O
TAND TAND=ge-4
FREQ F:::6
DIM FREQ gHz LNG mm ANG deg
MLIN .. 101=0.43 L=9.31
ZO=70.7 Keff:1.80 A, dB=O. 04
. E( eff) =90.00
Univers
ity of
Cape Tow
n
I
APPENDIX B
E AND H PLANE PLOTS FOR CROSS DIPOLE ANTENNA
156
/
Univers
ity of
Cape Tow
n(Jl
-..J
v
DATE:
I ~. ·;;. ~"'t
TESTED BY:
TRANSMITTING ANTENNA:
" ~
- z 0
-180 -150
GAcstr 74HL2134'1173
FREQUENCY·
Fl-(' c~ == 2.Sf. Cj H~ ANTENNA UNDER TEST:
(, r; SS-'?" D;ft'les
PLANE OF CUT:
'E I
~
~
v
-120 -90 -60 \
COMMENTS:
z 2! - 5> r-- en ~ -I
~
~
-v
.
--30 p 30 60 90 120 150 180
v
DATE:
~. ~(l
TESTED BY:
TRANSMITTING ANTENNA:
z "
GAcs" 74HL2134 "1173
FREQUENCY" COMMENTS:
2:b Cf H~ ANTENNA UNDER TEST:
o
PlANE OF CUT: I --11--1
Univers
ity of
Cape Tow
nUl ())
DATE:
IS.:s.8'9
TESTED BY:
TRANSMITTING ANTENNA:
~
L
I-z I I I II I 11 0
-180 -150
GAcsir 74HL2134'1173
-
I I II
-120
FREQUENCY: COMMENTS:
1-~GlH~ ANTENNA UNDER TEST:
(-ri:.·-~~ b ~r:il~
PLANE OF CUT:
t H'
,
,.._
I I 11 I I I I I 11 I I I I
-90 -60 -30
'
z z - ;;: -
~ ~ f -I ~
,- ~~ ...
~
lo o!)
- 7l
/
,___ I I ~I !.I 111 l 11 l I I 111 I 111 I I II II I 111 I
0 /
30 60 90 120 150 180
,
'
DATE: I FREOUENCV: I COMMENTS:
is.;,, 9 )Sc, lH?c
TESTED BY: ANTENNA UNDER TEST:
blr:01t-s
TRANSMITTING ANTENNA: PLANE OF CUT:
(fI
ACSi,74Hl2134'1113
10 t?E>
/
Univers
ity of
Cape Tow
n()) c.o
DATE:
TESTED BY:
TRANSMITTING ANTENNA:
'
'-----z fT TfT 1 Tl 0
-180 -150
. ..
GAcsor 74HL2134' 1173
~
11
FREQUENCY:
\. L( s s H l.
ANTENNA UNDER TEST:
C.-c--~ I:>l.\P~
PLANE OF CUT:
I E'
IJ
'
f T 11 T T 'I I 1 I I I
-120 -~O
I -60
fl
COMMENTS:
~
·1 II I I I 1 I II
-30 0
~
0 c: (J
'
0
--~
~
11
30 I lT
A ~ ~"' (L f}v-, (l ( ~._:/ ( I ,. .·._,-, )
P3', !. ;; .~·
fj
~
l TT T T TfT
60 90
./'
() 0
1,
1 TTI 1 1
120
0 (i ,)
11 l I
150
z -r en -t
/f /Oc
1
v
I 1111
180
2
~ ~
~ / 1 " l.-v\. t?,
~I p
-
Ii
oJ Jfr
(]I <P
DATE:
TESTEO BY:
TRANSI\IlITTING ANTENNA:
AcSif 74HL2134'l173
FREQUENCY:
• L[ H:c. ANTENNA UNDER TEST:
D\'~
PlANE OF CUT:
p
j c. ()
COMMENTS:
-
A ~v ( \ '-"'--'j )
Univers
ity of
Cape Tow
n& 0
DATE:
TESTED BY:
TRANSMITTING ANTENNA:
-z 0
-180 -150 -120
GAcsir 74HL2134" 1173
-·-~ --- -- -~- -
FREQUENCY: COMMENTS:
l, Lf 3 9 r{c ANTENNA UNDER TEST:
(\.,.~ Dtr~
PLANE OF CUT:
\ ~l l
-90 -60 -30 0
. '2-·
'
" . DD
,0
-~ ' "'lo z z ' /I - -r l>
CJ) CJ) .~!'I , , I -I -I - , lJ' C)
()
0 " / tl ()
~
.
-
'
I
-30 60 90 120 150 180
A )--'---, ~- . ~ll ----=::::>
I ~ r_ '"· ·( , I ' \
DAtE:
TESTED BY:
TRANSMITTING ANTENNA:
FREQUENCY:
l,4 ANTENNA UNDER TESt:
(.~D~
PLANE OF CUT:
\
COMMENTS:
(
A
Rl ~ ,
" ..
"
I , I
Univers
ity of
Cape Tow
n
'APPENDIX C
CHARACTERISTICS OF DHK 5068 SCHOTTKY MIXER DIODE
161 161
Univers
ity of
Cape Tow
n
~ :~
-~ .. ,,}. /c:
1-.--.
Features
• Low Noise Figure
• Excellent Cutoff
• Ideal for Image Enhancement Mixers
• Passivated Planar Construction for Reliability
Description
Alpha's series of gallium arsenide Schottky barrier diodes are available in beam lead. chip and packaged forms for mixer applications through 90 GHz. They are designed for low junction capacitance as well as low series resistance and exhibit calculated cutoff frequencies in excess of 900 GHz.
.-·~"!packaged diodes are hermetically sealed and may be used in waveguide, stripline or coaxial configurations.
Beam lead diodes are particularly well suited for MIC work. The beam lead design eliminates the problems associated with bonding to the junction, as is the case with a chip diode. A line of chip diodes is available for those who prefer to use chip and wire techniques for their MIC work. Capacitance ranges and series resistances on the beam lead and chip diodes are comparable to those of their packaged counterparts.
Beam lead and chip diodes may be mounted on a variety of standard or special substrates; if desired, Alpha will also bond them directly into a customer circuit.
ffJ.Alpha
!62
Schottky Diodes REFERE:\CE -:<L\ fBER: 10 h\~
Gallium. Arsenide Schottky Barrier Mixer Diodes
Outline Drawings
. I ,\.- -, __i_
I
i I
!(~,'.i . ""' _.l .iie _...._
Sar on Graph;c Symbol Denotes Catnoae
/·
t '""1"l;""i'"I' J' I )o.i· 1: I Jo•· 1 ·:;
J~;;~~~; ! :----- •}10\~61
Note: Millimeters in parentheses.
l i4-001
Zi0-805
Pnnted in U.S.A. Speciiicarions subjecr ro change wirhour norice.
.,
Features
• Low Noise
• Excellent Cutoff
• Ideal for Enhancement Mixers
• Passivated Planar Construction for Reliability
ription
series of m arsenide Schottky barrier diodes are available in beam lead. chip and packaged forms for mixer through 90 GHz. are
for low junction capacitance as well as low series resistance and exhibit calculated cutoff frequencies in excess of 900 GHz.
,'~,,! packaged diodes are sealed and may be used in waveguide, stripline or coaxial configurations.
Beam lead diodes are particularly well suited for MIC work. The beam lead eliminates the associated with bonding to the junction, as is the case with a chip diode. A line of chip diodes is available for those who prefer to use chip and wire techniques for their MIC work, ranges and series resistances on the beam lead and chip diodes are to those of their
Beam lead and chip diodes may be mounted on a variety of standard or jf will also bond them directly into a customer circuit.
ffJAlpha
2
Schottky Diodes 1
; 1
Gallium. Arsenide Schottky Barrier Mixer Diodes
Outline Drawings
-I Sar on Graphic Symbol Denotes Cathoae
r"
/ .. ~
• t -+-
Note: Millimeters in parentheses.
1 i4·001
liO·80S
Primed in U.s.f\. without notke.
Univers
ity of
Cape Tow
n
• ..... - • -- - J> ~ - - •• -- f-r- . . . . . . ·' , • - . ' . .... ~ . - - . ., .· .. .. ~ - .
These diodes are categorized by noise figure for mixer applications in three frequency ranges: X, Ku, and Kabands. Chips are available for use up to 90 GHz. Gallium arsenide diodes are particularly well suited for image enhancement mixer circuits due to their high cutoff frequency. Conversion loss for these diodes appr.oaches the theoretical minimum of 3.0 dB (single sideband) in X-band and is significantly lower than silicon Schottky diodes at frequencies above 12 GHz.
Matched pairs of mixer diodes are used in conjunction with a hybrid or magic-tee primarily for suppressing noise originating in the local oscillator. They are also used to isolate the local oscillator arm from the signal arm. thus minimizing radiation <ind absorption of signal power. Other uses are for specific reflection of signals through the hybrid and for balanced modulators and discriminators.
The matching criteria for packaged mixer diodes are as follows:
al Conversion loss (within 0.3 dB of each other) b) IF impedance (within 25 ohms of each other) . c) ihe VSWR of individual diodes, when not otherwise
restricted (such as 1.3 on premium units), is limited to 1.6 max.
These specifications allow the noise figure of the receiver to deteriorate no greater than 0.1 dB due to local oscillator noise. The VSWR limit allows a maximum of 5% leakage; in practice, this leakage is generally less than 2%.
A typical V1 vs 11 curve is plotted in Figure 1. Figure 2 shows a typical plot of capacitance vs bias voltage.
Noise Figure and IF Impedance as a function of Local Oscillator drive level with DC bias is shown in Figure 3.
Diodes may also be especially tailored to meet your particular electrical specifications or package configuration needs.
See Catalog Section VIII for Application Notes: 80800 Mixer and Detector Diodes
· 80000 Bonding Methods 80850 Handling Precautions for Schottky Barrier
and Point Contact Mixer and Detector Diodes
163
(' .. 0
"~
: ~'.c-l
11, : : .,: I ' ~' ',
I ~'·,'Ci
i '------....Jl__J_
.::111
I - ~.1;, 0 J ':~) i::, ... :_f3,
c~oo noa· NO~ \ 1 78/ i ~ S1
0 os: 0 ,:'it>(
OCI':' f'llC..W
Note: Millimeters in parentheses.
347-001
247-001
207-001
These diodes are categorized by noise figure for mixer in three frequency ranges: X, Ku, and Ka
bands. Chips are available for use up to 90 GHz. Gallium arsenide diodes are particularly well suited for image enhancement mixer circuits due to their high cutoff frequency. Conversion loss for these diodes the theoretical minimum of 3.0 dB X-band and is significantly lower than silicon diodes at above 12 GHz.
Matched of mixer diodes are used in conjunction with a hybrid or magic-tee pdmarily for suppressing noise in the local oscillator. are also used to isolate the local oscillator arm from the arm. thus minimizing radiation and of
Other uses are for reflection of signals through the hybrid and for balanced modulators and discriminators.
The matching criteria for packaged mixer diodes are as follows;
al Conversion loss (within 0.3 dB of each other) b)IF (within 25 ohms of each other) C) The VSWR of individual diodes, when not otherwise
as 1.3 on premium units), is limited to 1.6 max.
These allow the noise of the receiver to deteriorate no than 0.' dB due to local oscillator noise. The VSWR limit allows a maximum of 5% leakage; in practice, this leakage is generally less than 2%.
A typical V, vs If curve is plotted in Figure 1. Figure 2 shows a typical plot of capacitance vs bias voltage.
Noise Figure and IF as a function of Local Oscillator drive level with DC bias is shown in Figure 3.
Diodes may also be especially tailored to meet your particular electrical or configuration needs.
See Catalog Section VIII for Application Notes: 80800 Mixer and Detector Diodes
. 80000 Bonding Methods 80850 Handling Precautions for Schottky Barrier
and Point Contact Mixer and Detector Diodes
163
347·001
t ' i~ "
207·001
Millimeters in parentheses.
Univers
ity of
Cape Tow
n
Typical Ku-Band N{ixer Diodes
I
I 1.0f---------
l
i . 0.001~-------·--h"'--o~l--------l
I 0 ·0001 '--0--1 o'-0_2_00--Joo---.i-oo--soo"""'_s_oo--1-00--aoo-'
Voltage !Millivolts)
Figure 1. Forward DC Characteristic Curve Range-Voltage vs Current
.,,
164-
Figure 2. Junction Capacitance Range vs Voltage
4.0 ; o.oi 0.02 o.o5 0.1 0.2 o.5 i.o 2.0 s.o 10.0
Local Oscillator Drive (milliwansi
Figure 3. RF Parameters vs Local Oscillator Drive-Level
kal Ku-Band ~I
Voltage (Millivolts)
1. Curve
...
164-
er Diodes
~ O.16;-;~-----------------------
-2 -3 VOllage IVI
2. Junction Range vs Voltage
local Oscillator Dri •• (milliwartsl
3. RF Drive-Level
vs Oscillator
Univers
ity of
Cape Tow
n
- - ~. _,..,._~~. -- ... --~ "· ~ ... ·-·-:..- .... ~--' - - - ~ .. . .. :.-•_. . . ...._. _,. ..... - .. ·- ~·- . ·---- - . -~-
Type Number
Frequency Polarity
Band Reversible
Packaged:
I x DMK6600A x DMK6601A
x I DMK6600 x I DMK6601
Ku DMK6602A Ku DMK5068A
Ku DMK6602
v Ku DMK5068
Ka DMK6603A Ka DMK4058A
Ka DMK6603 ! Ka DMK4058
Beam Lead Singles:
I x I
DMK6604A x DMK6604
i Ku ! DMK6605A I Ku DMK6605 I
I Ka j DMK6606A .. Ka j DMK6606 I i DMK4791 i mm
I mm I DMK478415'
Beam Lead Pairs:
I Ku DMK6591
Beam Lead Quads:
I Ku DMK6592
Chips:
' x CMK7703A I x CMK7703
Ku CMK7704A Ku CMK7704
Ka CMK7705A Ka CMK7705 Ka CMK7701
mm CMK7702
All types: Forward .voltage. V, = 600-800 mV@ 1 mA Breakdown voliage. V6 = 4.0V Min.@: 10 µA
Notes: Maximum operating temperature= 150° C Note 1. Single sideband noise figure measured with
L.O. = 7 mW and including N,, = 1.0 dB.
Electrical Characteristics
NF<1l Fco<21 dB (GHz)
Max. Min. Min.
4.5 750 0.10 4.5 750 0.10
5.0 500 0.10 5.0 ' 500 0.10
4.8 I 750 0.05 4.8 750 0.05
5.3 500 0.05 53 500 0.05
5.5 600 -5.5 600 -6.0 350 -6.0 I 350 -
5.0i3: 500 0.10 5_5!31 350 0.10
5_3i31 500 0.05 5.8'3' 350 0.05
6.013 ' 350 -5_5131 300 -- 900 -- i 1000 -
500 0.05
500 0.05
4_5131 750 0.10 5.0131 500 0.10
4.813) 750 0.05 5_313\ 500 0.05
5.o<31 I 600 I -5.5(3\
I 350 (
I -
5.5'31 600 -
5_913' 1200 -
Note 2. F00 = 2
R1
C where Rs= Rr (@ 10 mA)- R8 ; R8 = 0208 (for 10 mA)
r s ~ . 1
Note 3. Noise figure is determined by lot sampling. Note 4. Electrical characteristics are specified for each diode in a pair or quad configuration. Note 5. To be supplied bonded by Alpha on customer substrate
)65
Cj@OV pf Package I Max. Outline
0.20 207-001 0.20 247-001
0.20 207-001 I 0.20 I 247-001
I 0.15 207-001 0.15 247-001
0.15 I 207-001 0.15 247.001
0.08 207-001 0.08 247-001 I
I 0.08 i 207-001 I I 0.08 I 247-001 !
0.20
I 174-001
I 0.20 174-001
I 0.15 I 174-001 I
I 0.15 174-001 !
I i
I 0.10 i 174-001 i 0.10 I 174-001 i 0.07 I 174-001
; I :
0.04 I 366-001 i
0 15 378-012
0.15 294-003
0.20 270-801 I I
0.20 270-801 ! 0.15 270-801
I 0.15 270-801
0.08 270-801
I 0.08 270-801 0.06 i 347-00: :
I 0.04 I
347-001 !
I I
Frequency Table
Band Freq.(GHz) I x 8.2-12.4
I Ku 12.4-18.0 Ka 26.5-40.0 I mm 40.0-100 0 I
;
Packaged:
X DMK6600A X DMK6601A
X I DMK6600 X I DMK6601
Ku DMK6602A Ku DMK5068A
Ku DMK6602
V Ku DMK5068
Ka DMK6603A Ka· DMK405BA
Ka DMK6603 ,
Ka DMK4058
DMK6604A DMK6604
Ku DMK6605A DMK6605
Beam Lead Pairs:
Chips'
I it. CMK7703A X
Ku Ku
Ka Ka Ka
mm
All
Noles:
, I
CMK7703
CMK7704A CMK7704
CMK7705A CMK7705 CMK7701
CMK7702
600-800 mV@l rnA VB::: 4.0V [email protected]
Maximum operating temperature:: 150"C Nole 1. Single sideband noise figure measured with
l.O 7 rnW and including N" = 1.0 dS.
45 4.5
5.0 5.0
4.8 4.8
5.3 5.3
5.5 5.5
6.0 6.0
5.0(3: 5.513\
5.3(3\ 5.8,3\
4.5(3) 5.0(3!
4.8!31 5.3!3\
5.013\ 5.5(3, 5.5!31
6.913\
Note 2. where RT (@ 10 mA) :::
is determined by lot
750 0.10 750 0.10
500 0.10 , 500 0.10
750 0.05 750 0.05
I 500 0.05 500 0.05
600 -600 -350 -350 1 -
500 010 350 0.10
500 0.05 350 005
750 0.10 500 0.10
750 0.05 500 0.05
600 -I 350
I
600 I -1200 -
Note 4. characteristics are tor each diad"" in a pair or quad configuration. Nole 5. To be supplied bonded by Alpha on customer substrate
}65
0.20 207·001 0.20 247·001
0.20 207·001 0.20 247·001
0.15 015 247.00 1
0.15 I 0.15 247·001
0.08 207· 008 247-001
[ 0.08 i 207·001
1 008 I 247-001 I
174·001 174·001
174·001
0.20 270·801 ! 0.20 270·801 !
0.15 270·801 i 0.15 270-801
0.08 270·801 0.08 270·801 0.06 347·001
0.04 347 ·001 I
Univers
ity of
Cape Tow
n
APPENDIX D
FLANGE DIMENSIONS FOR HORN ANTENNAS
166 166
Univers
ity of
Cape Tow
n
-··- ·-----·- -·--~--1
GENERAL INFORMATION u·-~ · Waveguide [hp] 93
HP flange dimensions •
:ircular flanges
S Band G Band
3.250 Diameter-\ -~--
~ e ---~, 0 \ !
Hole Diameter
I ' I \ I .
~--m·· --~ \ I
\ '. ! \ \ I ' I \ .: \ ' . /
0 i \, 0 '~ ©>--::
J Band
I
Band Frequency I Material Cover Hole Designations I
I Range I B=Brass Flange Dia. HP EIA WR- !GHzJ ! A=Alum. UG- (In.)
;
B 53 s ! 284 2.60. 3.95 0.281 A 584
D1amete· ; G 187 3.95. 5.85 B 149A 0.219 i A 407
13i 8.20 ~ 344 0.219 ::. .. ~::· . r. 441 "'· '-----·---------·--
l67
GENERAL INFORMATION . Waveguide 93
HP flange dimensions
:ircular flanges
S Band G Band
J Band
137 5.3:: ·8.20 , ----~ -~-------.--
Univers
ity of
Cape Tow
n
EEJ GENERAL INFORMATION Waveguide Standard data chart
--··----------- -- ----------,--------
'
i I !
I
: !
I r I I
! i j
! I
I i
I i I
I !
I ' !
I
T~eoretical fheoiet:C3t !Al+' Attenuation !tter.uation ih!cretical'
~p
3and ?;e~1.m1cy ~an1e
IT , \lode SHzl
3.rna :cs;.~~at;a!'!s i:'lant!
; 1 ! Other ~~~~!~~! i Oes11n~tia1 1n!ide O:.otside l.J# to ~1gh . :~;w t::i Hi(h ~W P;•er
;:tsi~I~) IEC ! tlA - British iAN·'. Common' A·41um. ·Choke I Canr -W-idttl--i -H-e-i1_1rt_l_T_ol-,i_W_i_dt_h.,1_H_e-ig-ht"'"'i_T_o_I -i, Nam. Wall i Cutoff : a%~~~~~~'/ i a%~~~i~~~/ , lf~~~e~:
CAI 1 cm : .• m I cm I c11 I _m i Thickness : freouencyj Sil'ter I Silver · ~~ia,.iltts R- , WR- j 'NG· RS- i Vsace S-Silm . UG- , UG- ;inJ (in.l i :mils) 1in.l j inJ i :mils} i mm UnJ [ ~GHzJ ~ dBi 100 ft '.dB, 100 Meters; Kilowatts)
.. :2 · i.~O :~ . ;5() I 16 5i0 5.50>
11 .,551
·2ii 1 s=i·· 8661.',1=12.7! 2~3 QS08 11')~12·0272 lJjJ.·1.394~1 ·,_-251 11, =_~,-· ', ·b~.~_6',0 • ' • · ·3llr ·-::Si 0.JdGI lJ.,69-D.li3 J.383-G.5&4, i l.3. i7.~
: !5 2.20 : d ~ 5 ! 0 j l . I
l.70 - 2.SO i 22 1 430 I J I
2 20 - ],JO i 16 340 I I
2.60' 3-95 I 3212114 I 10 ' I
JJO • 4_<lQ 40 229 J ; IA ' I
H5·585 1~J1a1i 12 l i
, ;:a I i i;g i," 1 ·-
13
I 5 85 8 20 r ·o J 137 I 14 I' I I i
H I 7.05. 10.00 j84j112 15 I i i I
1.00. 11 00 J _I i02 I I -I I I
x : !.20' 12.40 i 100 ! 90 I 16
I l '
M I 0.00 · 15.00 I 120 I 15 17 I i
p i 12.40 · 11.00 140 I 62 11
I
N i 15.00 . 22.00 1180
51 19
I K i 18.00 . 26.50 220 42 20
I 22.00 . 33.00 260 34 21
R 2S.50 - 40.00 320 21 22
33-00. 50.00 400 22 23
46.00 -60.00 500 !9 24
v 50.00. 75.00 620 15 25
60.00 -90.00 140 12 26
75.00. 110.00 900 10 27
AMlrnimon,
;o4 i05 I
43 75
-I -i
cS.R
49 11 ~.H
95 !
-! ! !
50 I XN C 6 ! 106 i ''
I 51 i X8.w 68 i
I r -I
I -l r 1
52 I ! 67 '
I-1-
! 91 I KU.Y.U
1-1 i 107
--
I
53 121 56
---
- Y,KA,U -91
- a 97
--
- M 98
- E 9<J
--
!EC-International Electrotechnical Commissron JAr~-Joint Army Navy El'-tlectronic Industry Association
r ! I
I I l I I
I
8 A
B A
B A
B A
A
B A
8 A
8 A
8 A
8 A s
8 A s
8 A s
B A s 8 A s 8 s 8 s B s
8 s
i 54 ! S85A,
.1358 !379
553A 554A
53 5114
i295-l\ ?.477 11.:.:l2.7 I !3.360 i5.i01 2j5! I =5 :5 25i
10 922 5.461 I= ;u 11.328 .4.31 '2.!5! =5 ,4 46)
8.6361 J 318 =12.; 9 J42 ·3.-lOl 1.701 :::5 I J.561
I
7.214 l.4a4 i :::12.7 7.620 !2.1141 il.341 :::5 •3.001
5867 -::!27 I -2.3!1 I i=S' !
I C'-!R22~' 5 817 l - i 2 29'. ,
2.9C8 I -::12.i 15.142 3233 II =!2.7 !.:.151 : =5 . 2.-!~S: 1 \!.r3l '..=SJ j
; 1~C i : io68 I r ,
149A , us5 ! 2_m 1 ·= 12. 1 I 5.080 j' 2.540 1 = 12.1 1 -«17 i rt 8721: ·o.arn =5 , :.2.001 11.001 ! '=51 I
i CMR159 ; J.039 i I - I ,l ,9\ i
1 34381 34-4 I
' 4408 ' 4-41 j :urn 1
n19 i=:o2J4364 2344:=:021 o.:;91 1 --=4 : u 1a1 ;o q231 , :::41 I 1580 '021H10' 1905 ·021 I=·.
1 ri.501 1-· :o:s221 I :::4 i0.7501' '=41 : I
1 528 I 51 , 2.850 I 1.262 :::10.2 3. 175 1.581 :::10.21 I IJ78 . 138 \ 11.1221 I ,o.4971 =4 11-251 (06251 !=41
I ! 2.m I 1 295 j ::: 1 6 I 2.345 I 549 =l 6 r I '.!02! i ·0.5101 =3 J rl.12! 0.510! ':::31
I 468 i 39 ' 2.286 1.016 ! = 1.s : 2_540 I u10 -:::7.6 ' ! 1168 I 135 ! 0 901 !0.411) = 3 ' 11.001 i •0.501 !-:::31 I
1.905 0.953 ::7.6 2.159 1 207 :::7.6 10.751 10.3751 =3 I0.8501 !0,4751 1::31
I 5~A 419 1 580 0.790 -=&.4 I 1.783 0.993 :::7.6 - 10.6221 10.311) ::2.5 10.7021 10.3911 1:::31
i- -1.295 0.648 :::6.4 1.499 0.851 :::7.6
10.5101 10.255) ::2.5 10.5901 10.335) 1:::31
596A ' 595 1425µ 1.087 0.432 ::5.1 1.270 1~:m11 ::7.6 598A 597 10.420) 10.1701 ::2 10.5001 1:::31 - -
i 0.864 0.432 ~5.1 I •J67 0.635 =7.6 '034()1 iO !IOI =2 ,04201 r0.2501 =:3l
:530
1599 {381µ 0.711 0.356 ::U 0.914 0.559 ::5.1 - (0.2801 (0.140) ::1.5 10.3601 10.2201 1=2l
600A -383 0 569 0.284 -:::2.5 0.772 0488 :::5.l - i0.2241 10.1121 ::1 10.3041 :0.1921 (=:2}
0.478 0.2388 :::2.5 0.681 0-442 :!:5.1 i0.1881 I0.0941 ::1 I026lll i0.1741 1::21
385 0.3759 0.1879 =:2.5 0.579 0.391 ::5.l - !0.1481 10.0741 ::1 10.2281 I0.!541 1::21
387 0.3099 0.1549 :::1.3 0.5130 0.3581 =5.1 - 10.1221 10.0611 :::0.5 I0.2021 i0.l4ll 1::21
0.2540 0.1270 ::1.3 0.4572 0.3302 =5.1 iO.IOOl 10.050! ::0.5 10.180! 10.!301 1::2l
'For more informition refer to U.S. Military 59ecrfrcation, Mll·W-85, Waveguide, Ri1id, Rectancular. 2 For more information refer to U.S. Military Spec1ficat1on, Mll-F-3922. '1anps. w .. ecuide Cover.
2.C3 0.0801
213 0 0801
2.03 0 080}
2.03 '0.0801
l.63 e.J5.t'
1.63 '0.0641
I.OJ 0.0641
1 63 ·o 0541
1.63 i0.0641
!
1.27 I 0.0501
1.27 i :0050) I
1.27 10.0501
1.02 i004l))
1 02 i0.040!
1.02 10.040!
I 02 0.04-01
1.02 10.0401
1.02 ,004()1
1.02 I0.0401
1.02 '.0.0401
1.02 !0.0401
1.02 :0.0401
\.16
l.3i5
l 735
2.080
2.59
3.16
3.71
4 29
5.26
6.50
6.56
7.81
9.49
11.6
14.1
11.3
21.1
26.35
31.4
39.9
48.4
59.0
i S33 . l 230 I ; s -: J 5 !.2'9 · 0 836 1
o.:s9. ~ 504 2 492. 1.655 s.2 . rs ~.!96 '0 329 i .526 ' i 080
l 030 '0.7\6 3.382. 2.352 0.673 -0 ~68 2.207. !.535
3.4-Vl
2.18. 3.1
I I.828 1.296 ; 002 '4 255 l 56. 2_;4 1
1 c ~4 - J846 3 9 l 7 · 2. 777
'. 2.$35 · 1.869 H49·6.134 : 1941 · 13171 1 uso. i.220 5.774. 4.il03
: 3G9l · 2.J24 I iO 15 · 7 630 i 1754 · 3831 i 20\3-\5;81 -;.622-~',80 i 1
3821 3018 1254 9907 I 554 .96) ! I . b '2:!96-1,971 i s.ie1. 6.465
: 5.355 ·4.161 l 17.58. 13.6& : 1355' 454) \ 3.497 · n11 I 11.47. 8.913 i
6.939· UGO I 22.78 _ :4 31 I -Z80. 424i 4 532. 2.848 I 1487 - 934
8.362 ' 5. 7841 27.45. !8.99 ! '.206 - 2931 5.461. 3.778' 17.91 . 12.39 ' 9.893 '6.909 32.48. 22.68 1166. 2291 6.461 . 4.51 z 21.19-14.10
12.48. 0 162 40.92. 30.08 (119· 15n rnu:m 1 26.71·19.63 20.22·14.87
17.02 . 12.33 55.88 . 40.49 179 · 106) 11.12. 8.054 36.47. 2S.42 8.418. 6.099 27.62. 20.01
26.6& . 19.58 87.51. 64.21 143 - 581 17.41·12.79 57.11. 41.95 13.18 · 9.6&4 I 43.25. 31.77 I 32.58. 22.ss J ;05.;i4. ;4 37 :34. 47l 2127.14.80 69)9' 48.54 16.1l.1120 52.36 - 36.76
44.2'9. 30.33 145.4. 99.57 I 123. 321
I 28.92. 19.81 94.81 · 64.91 21.9 · 15.00 71-85 . 49.21
30.84. 20 96 iOl.2 - 67-78 114 - 20!
38.79. 27 21 127-3 -89.25 <I0· 14)
5730-3915 188.0. 128.4 16. 91
78.33. 52.51 256 9. 172.3 14·61
100.5. 70.71 329.7. 231.9 13. 4)
1 Attenuation computatrons: Rectan1ular 1uide, TE,, mode: Resrst,vrties. Brass. !6S-35J 6.63 x 10 •fl/cm, Aluminum, 2.83 1 10-• n:cm: Silver, 162 • 10·• !l1cm. •CW Power computations, Breakdown strength of arr taken at 15.000 volts per centimeter. Sltety factor of approxrmately 2 at sea level rs assumed. 1 HP rnstrumentation flan1es mite with reclangular cover flanges :ioted. Fla':, adaotert are anilable to mate wrth UG-4251 U and UG-
381/U. Specify 11515A IK·81nd, 18.0-26.5 GHzl or 11516A :RBand.26.5- .OGHll.
166
! I I
I r
EIJ N RAL INFORMATION Waveguide Stand aid data chart
/I i 15,00,22,00 I I
K : IUlIl· 26.50 53 8 121 A 66 S
22.00 . 33,00 260 34 21 B :34,471 A s
16,50 . 48.110 3211 2& U UA.U 8 (23 32) A
9& S 6!1OA
3300,50,00 400 22 23 8 97 S 67.78 ,14 2m
4000 60.00 SIlO 19 24 8 S 38.79 2711 1273 89,25 <10,141
50,00 . 7500 620 15 25 III 8 98 S 5730 3915 188,0·128,4 16·91
60,00 90,00 74() 12 26 99 78,33 5251 256,9 • 172,3 14 ·61
75.00, llQ,OO 900 10 27 100,5, 70.71 329,/ . 231.9 13,4)
AlllltUhltillllS, IEC~;nt .. nalion.1 Electrolethnical CommISSIon JAK-k)101 Almy H.", E!"--"I.etro.ie 100000stry Assotilll .. I For more infarmllia. relol t. U,s. Military ~.clflc.tlOn, Mll,W,8S, W ... ,uid/l, Rllie!. Re<:t"" •• lar. I For mOl" i"IONna!i •• r.l", t. U,S. Military SpeCIfication, Mll,f·3922. ~.n .. s. WM(Ilido Co""" • Ant •• ati .. tampulallo,s, Rett.n(llIM iuido. TE .. modo: Res"t,,,ties, Br4'S$, 165-l!>l 6,63 • 10 • illem: 2,83 , 10" • C'N Po",.r compUtalioM: Breakdo ... ,tre.lII" 01 ." Iale •• t 15,000 ,,01" pe, ctntim.I"'~ S:Ii.ty factor approximately 2 ,I sea 'HP '.stl'1,lme.talion 111",11 mil. "tin (fCUnlUi., COYe' !I,ng .. ""'ed, Fla'it .t.n. ar. "".,Ioblo 10 m.ate .. 'th UG-41SI U .nd 18IIU~ Specify Jl51SA IK·S.oo, 18,0,26,5 GHzl or ll516A :RB,nd. 26,5, 4{),Q GIll).
10' fVcm,
.;i:
Univers
ity of
Cape Tow
n
APPENDIX E
3 GHz NOTCH FILTER SIMULATION PROGRAMS
169 169
Univers
ity of
Cape Tow
n
TJucnstone (TM) - Configuration~ 100 1600 102 165i7 1963 lOGO OPENIM.CKT Sat Dec 10 08:28:31 1988
1 :~PEDANCE OF AN OPEN CIRCUIT STUB ON RT-OUROIO S82J 'JPERATING RANGE 2-8 GHz 1 l 9 JANUARY 1989
I: j M u..1.I ·
::"REQ GHZ F:E3 OH I:'JO NH CAP PF L\JG MM TI r1E P:::
. OH ANG
MSUB ER=2.2 H=0.254 T=0.01778 MLIN 1 ~ W=0.76 L=l8.~9 MLOC 2 W=0.76 L=18.29 MLIN 2 3 W=0.76 L=l8.29 DEF2P 3 FIL
FREQ SWEEP 2 8 O.OS
OUT =IL FIL
GRID ,~ANGE
GR1
DBCSllJ GRl OBCS12J GRl
2 8 o.s -70 10 l(i
\10
RHO=O. 84 .RGH=O
1 0CSTUB AT 3 Gl-iz
TJucnstone (TM) - Configuratlon~ i00 1 i7
o I i~
OH I:'JO NH CAP L\JG MM T:
Ct-(T
z
OPENIM.CKT t Oec 10 :28:31 1988
AN OP CIRCUIT STUB ON RT- ROID 5820 RANGE 2-8 GHz
1
ER=2.2 , ,) ~
T' , .01778 .RGH=O
2 L= 8.2!7' L= ZI.29 L= 8 .
IOCSTUS AT 3 '-l .3 ....
FREQ P 2 8 0.05
OUT FIL [S11 J GRl FIL [S12J GRI
GRID RANGE 2 80.5 GRI 10 11.)
3
, 70
- ?4
Univers
ity of
Cape Tow
n
Touchstone <TM) - Configuration( 100 1600 102 16517 l~~~
.=KEO-GHZ
2.00000 2.soooo 3.00000 3.50000 .:.;. . 00000 .+.soooc S. OOOOr) s.scooo 6. c:.oooo 6.50000 7.00000 7. '.:10000 3.00000
OPENIM.OLlT Sat Dec 10 10:21:57 1988 ·
DBCSllJ FIL
-3.803 -1.2s1 -0. 157 -1.233 -3.008 -7.116
-11 . 2ti6 -:7.674 -49.~)33
-17.532 -11.:23 -7.110 - 3. 846
DBCS12J FIL
-2.547 -6.656
-41.027 1' ,. ,.... ,,
-:i.b:J'+
-1 . 100 -0.482
-0. 139 -0.223 -0.509 -.1. .160 -2.675
171
Touchstone <TM) - Configuration( 100 1600 102 16517 l~~~ OPENIM.OGT Sat Dec 10 10:21:57 1988·
~?EQ-GHZ 08[S11J 08[S12J
2.)0000 2.50000 3.00000 3.50000 4.00000 -+.50000 5.0000l) 5.50000 6.(,0000 6.50000 7.00000 7. :::,0000 3.00000
FIL FIL
-3.803 -1 .251 -0.157 -1 .2.33 -3.308 -7.116
- L 1 .2tJ6 -:7.674 -49.~)33
-17.532 -11.::23 -7.110 -:;.846
-2.547 -6.656
-41.027 l ",.....,
-:;, • b~"+ .-,
-.c..'
-1 . 100 -0.482
-(;. 139 -0.223 -0.509 - J. • 160 -2.6
171
Univers
ity of
Cape Tow
n
Touchstone CTMJ - Configuration( 100 1600 102 16517 1968 1000 1 3294 I 3GNOTCH. CKT Tue Sep 26 00: 51: 2o 1989
THIS PROGRAM IS USED TO DETERMINE THE TRANSMISSION AND REFLECTION COEFFICIENTS FOR A 3 GHz NOTCH FILTER I 6 GHz PASS FILTER 29 JULY 1988
DIM FREQ GHZ RES OH IND NH CAP PF LNG MM TI ME PS COND /OH ANG DEG
CKT
MSUB MLIN MLOC MLIN MLOC MLIN MLOC MLIN MLOC MLIN DEF2P
ER=2. 2 . H=O. 2548 T=O. 01778 RHO=O. 84 1 2 W=O. 76 L=72. 79 2 W=O. 76 L=18. 29 2 3 W=O. 76 L=18. 29 3 W=O. 76 L=18. 29 3 4 W=O. 76 L=18. 29 4 W=O. 76 L=18. 29 4 5 W=O. 76 L=18. 29 5 W=O. 76 L=18. 29 5 6 W=O. 76 L=72. 79 1 6 FIL
OUT
FIL DBC S11 l GR1 FIL DBC S211 GR1
FREQ SWEEP 2 8 .03
GRID RANGE 2 8 . 5 GR1 -100,30 10
\72
RGH=O ! RT 5880 ! INPUT SEC TI ON !1ST OC STUB ! QW AT 3 GHz ! 2ND OC STUB ! QW AT 3 GHz ! 3RD OC STUB ! QW AT 3 GHz !4TH OC STUB ! OUTPUT SEC TI ON
Touchstone (TM) - Configuration( 100 1600 102 16517 1968 1000 1 3294 3GNOTCH. CKT Tue Sap 26 00: 51: 20 1989
THIS PROGRAM IS USED TO DETERMINE THE TRANSMISSION AND REFLECTION COEFFICIENTS FOR A 3 GHz NOTCH FILTER I 6 GHz PASS FILTER 29 JULY 1988
01 M FREQ GHZ RES OH IND NH CAP PF LNG MM TI ME PS COND 10H ANG OEG
CKT
MSUB ER=2. 2 . H=O. 2548 T=0.01778 MLIN 1 2 W=0.76 L=72. 79 MLOC 2 H=O. 76 L=18. 29 MLIN 2 3 W=O. 76 L=18.29 MLOC 3 H=O. 76 L=18.29 MLIN 3 4 W=O. 76 L=1B.29 MLOC 4 W=O. 76 L=18. 29 MLIN 4 5 W=O. 76 L=18. 29 MLOC 5 W=O. 76 L=1B.29 MLIN 5 6 W=0.76 DEF2P 1 6 FIL
OUT
FIL DBf S11] GR1 FIL DBf S21] GR1
FREQ SHEEP 28.03
D RANGE 2 8 . 5 GR1 -100'30 10
L=72. 79
\7
RHO=O. 84 RGH=O ! RT 5880 ! INPUT SECTION ! 1 ST OC STUB ! QW AT 3 GHz ! 2ND OC STUB ! QW AT 3 GHz ! 3RO OC STUB ! QW AT 3 GHz !4TH OC STUB ! OUTPUT SECTI ON
Univers
ity of
Cape Tow
n
APPENDIX F
A NOTCH FILTER ANALYSIS TECHNIQUE USING PIN DIODES
173 173
Univers
ity of
Cape Tow
n
An important analysis technique in describing the stub filter is
to compare its operation to that of isolation PIN diodes used to protect mixer diodes.
PIN diodes are often used to protect mixer diodes, by absorbing
any excess energy which might damage the mixer diodes. The degree
of protection (i.e the amount of attenuation of the input signal)
is determined by the number of PIN diodes connected in parallel
with the mixer diode, as well as the spacing (in fractions of
wavelengths) between the PIN diodes.
A mathematical technique will be used to show how the protection
technique works and then to analyse the operation of filter.
To simplify
stub filter,
reason for
the explanation and make
only forward biased PIN
this is that PIN diodes
isolation in this bias position.
it more applicable· to the
diodes will be used. The
provide most of their
The assumption is made that the impedance of a forward biased PIN
diode is H2 , and that of the mixer diode is SOQ
Single PIN diode protection
A simple circuit realisation is:
The two resistances are taken in parallel and hence provide an
input impedance of:
Z(a) = 1//50
= 0.980
174
An is the stub
to compare its operation to that of isolation PIN diodes used to
PIN diodes are often used to protect diodes, by absorbing
any excess
(i.e the amount of the input s )
is determined by the number of
with the mixer diode, as
PIN diodes connected in parallel
as the spacing (in fractions of
) the PIN
A mathematical to show how protection
works then to of
To simplify the explanation and make it more applicable" to the
only PIN
reason for this that PIN diodes provide most of
isolation in this bias position.
The assumption made that the impedance of a forward biased PIN
10 ,
PIN diode
::
The two resistances are taken in
input
Z{a)
. . =1//50
= 0.980
174
50g
and provide an
Univers
ity of
Cape Tow
n
.The reflection coefficient at {a) is:
J) Z ( a)-Z<> Z(a)+Z<>
= 0.98-50 0.98+50
-0.9616
The amount of power reflected from {a) is:
Pr.,,,£{a) = \f I 2*Pin.
= 0.925 Pin.
The amount of power transmitted is:
p tr { a ) = pin. 0.925*Pin.
Of this ratio the amount of power absorbed by the PIN diode is:
The amount absorbed by the mixer diode:
The isolation provided by the forwarded biased PIN diode:
= lO*Log10* l.47*10- 3 *Pin. _pin.
175
. The reflection
amount
amount of
at (a)
(a)
. .
. .
(a)
(a) = - O.
:
Of this ratio the amount of power absorbed by
amount by
= o.
. .
= O.075*P~n.*1 51
::::::
PIN diode
The isolation provided biased PIN diode:
:::::: 10*Log10*
175
:
Univers
ity of
Cape Tow
n
= -28.3 dB ... (1)
Double PIN diode protection
A second PIN diode is connected in parallel with the first PIN
diode. It is assumed initially that the two PIN diodes are half a
wavelength apart, and that both are forward biased. The isolation
provided is calculated as follows:
As before, a parallel branch of models the protection scheme.
at (b)
a >-/2
o--__,a,...r-·-···-··-··~·t .... )
0
Z.t..n.(b) = 1//50 = 0.98Q
11 Ohm }z _J
Since the impedance at (b) is >- /2 away from (a), the resistances
are effectively on top of each other. Hence the impedance at (a) can be s ... """ ~"'-'.i"''-t......J..e
~~b calculated as follows:
Z(a) = 1//0.98 = 0.495Q
Reflection coefficient: y = Z (a)-Z<> Z(a)+Z<>
0.495-50 0.495+50
176
••• ( 1)
with the
diode. It assumed that the two PIN diodes are
wavelength II and that both are . The
provided is as follows:
As II a of the
Ohm'll
at (b) Z.Ln.(b) = 1//50
at (b) A /2 from (a) II
are
Hence
on each other.
impedance at (a) can
as follows:
Z(a) 1//0.98 = 0.495$1
: =
176
PIN
half a
Univers
ity of
Cap
e Town
-0.9804
Power reflected at (a): P=~f(a) = l..Pl 2 *Pin
0.9612*Pin
Power transmitted at (a): Pt=(a)
Power at (b):
Power absorbed by mixer dioqe:
pin - pin*IJl 2
0.0388*Pi:n.
0.0388*Pi:n.*1 1.98
= 0.0196*Pin.
= 0.01996*Pin.*1. 51
Isolation: = lO*Log10* 3.843*10-4 *Pin pin.
= -34.15 dB . . ( 2 )
This is 34.15 - 28.3 = 5.85 dB
If the second PIN diode is placed A/4 away from the first PIN
diode, the isolation is determined as follo~s:
: r-::1 Ohn ~50 Oh~ I .. _j
Both PIN's forward biased.
zin.(b) = 1/)50 = 0.98ll
177
:
-34415 dB
is
If PIN PIN
diode, the isolation is determined as follo~s:
Both PIN's
Z ... n{b) ::::: 1/150
=
177
Univers
ity of
Cap
e Town
Since the two diodes are A / 4 apart the impedance at ( b) can be
transformed to (a):
Z(a) = z 0 2/z(b) = 50 2 /0.98
255H2
The impedance at (a) is then: Zi:n
Reflection coefficient at (a):
1//2551
= 0.9996Jl
= 0.9996-50 0.9996+50
Power reflected at (a):
= -0.9608
P:r.,,.f (a) = j9j2 *P .t..:n
0.9608 2 *P.t..:n
= 0.923l*Pi:n.
Power transmitted Pt:r(a) = 0.0768*P.i..:n
Power transmitted to (b): Pt:r(b) = 0.0768*P.i.:n*-1_ 1+2551
= 3*10-5 *P.i..n.
Power absorbed by mixer: 3 *10-5 *P *l .i..n. --51
Therefore the isolation provided by this combination is:
lO*Log10* 5.88*10-7 *P.t..:n p .i..:n
178
two
transtormed to (a):
are >- /4
Zeal = Zo2/Z(b} /0.98
The impedance at (a) is then:
Reflection at (a):
1//2551
0.99961l
Power reflected at (a) : (a) . = !912 *p in<
O.9608 2 *Pi n<
= O.923l*Pin
Power transmitted (a) = O.0768*Pin
Power transmitted to (b): (b) = O.0768*Pi;n*
Power by :
the by
178
at (b) can be
1
:
Univers
ity of
Cap
e Town
= -62.3 dB . . . ( 3 )
Comparing (1),(2) and (3), it can be seen that two forward biased PIN diodes placed A/4 apart provide the greatest protection for a mixer diode. It can also be seen that placing the PIN diodes at
A/4 rather than A/2 spacings provides the mixer diode with almost
30 dB more protection.
Analysing the stub filter Looking at the operation of the stub filter, it can be seen that the stub spacing fulfills the same isolation effect as the spa_cing between the PIN diodes. Looking at the 3 GHz notch filter we find the following:
0'4 3GHz I-·:·---~
A/4 3GH•:_f .~ ~
It can be seen that at 3 GHz, the spacing is .A I 4 i.e. greatest isolation. This means that most of the power will be absorbed in the open circuit stub (enfused by the fact that the stub offers little impedance to a 3 GHz signal).
For any and as
6 GHz component, the spacing between the stubs is A/2, we have seen this configuration provides the least
isolation. Hence since theA(3)/4 open circuit stubs offer a high impedance to the 6 GHz signal, and there is little isolation between the stubs, there is little attenuation to a 6 GHz component in this filter.
The 6 GHz notch filter can be similarly described but the explanation would be superfluous and will not be attempted here.
179
= . • . ( 3 )
Comparing (1),(2) and (3), it can be seen that two forward biased PIN diodes placed A/4 apart provide greatest protection for a
It can seen the PIN diodes at A/4 rather than A/2 spacings provides the mixer diode with almost
30 dB more
Looking at the operation of the stub , it can seen that
the spacing fulfills same as between the PIN diodes. Looking at the 3 notch
we find the following:
"':>'4 3QHz
It can be seen at 3 GHz, the A /4 i.e. means that most the power 1 be absorbed in
the open circuit stub (enfused by the that the stub to a 3 GHz ).
For any 6 GHz component, the spacing A. /2, and as we seen isolation. Hence since theA(3)/4 open circuit stubs of a high
to the 6 GHz , and 1 between the stubs, there attenuation to a 6 GHz
in
6 GHz can but the explanation would be superfluous and will not be attempted
179
Univers
ity of
Cap
e Town
APPENDIX G
6 GHz NOTCH FILTER SIMULATION PROGRAMS
180 180
Univers
ity of
Cap
e Town
Touchstone <TM> - Configuration< 100 1600 102 16517 1968 10_00 1 3· 6GNOTCH. CKT Tue Sep 26 DO: 52: 55 1989
THIS PROGRAM DETERMINES THE REFLECTION AND TRANSMISSION COEFFICIENTS FOR A 6 GHz NOTCH FILTER I 3 ghZ PASS FILTER 13 JULY 1988
DIM FREQ GHZ RES OH IND NH CAP PF LNG MM TIME PS COND /OH ANG DEG
CKT
MSUB ER=2. 2 H=O. 2548 T=O. 01778 RHO=O. 84 RGH=O ! RT 5880 MLIN 1 2 W=O. 76 L=36.19 ! INPUT SECTION MLSC 2 W=O. 76 L=18.29 ! 1 ST SC STUB MLIN 2 3 W=O. 76 L=8. 3825 ! QW AT 3 GHz MLSC 3 W=O. 76 L=18. 29 ! 2ND SC STUB MLIN 3 4 W=O. 76 L=8. 3825 ! QW AT 3 GHz MLSC 4 W=O. 76 L=18. 29. ! 3RD SC STUB MLIN 4 5 W=O. 76 L=8. 3825 ! QW AT 3 GHz MLSC 5 W=O. 76 L=18. 29 !4TH SC STUB MLIN 5 6 W=O. 76 L=36. 19 ! OUTPUT SECTION DEF2P 1 6 FIL
OUT
FIL DBCS111 GR1 FIL DBCS211 GR1
FREQ SWEEP 2 8 . 03
GRID RANGE 2 8 .5 GR1 -100 30 10
181
Touchstone (TM) - Configuration( 100 1600 102 16517 1968 10,00 1 3' 6GNOTCH.CKT Tue Sep 26 00:52:55 1989
THIS PROGRAM DETERMINES THE REFLECTION AND TRANSMISSION COEFFICIENTS FOR A 6 GHz NOTCH FILTER I 3 PASS FILTER 13 JULY 1988
DIM FREQ GHZ RES OH IND NH CAP PF LNG MM T! ME PS COND 10H ANG DEG
CKT
MSUB MLIN MLSC MLIN MLSC MLIN MLSC MLIN MLSC MLIN DEF2P
OUT
ER=2.2 H=0.2548 T=0.01778 1 2 W=0.76 L=36. 19 2 W=0.76 L=18.29 2 3W=0.76 L=8.3825 3 W=0,76 L 18.29 3 4 W=O. 76 L=8. 3825 4 W=0.76 L=18.29. 4 5 W=O. 76 L=8. 3825 5 W=0.76 L=18.29 5 6 W=0.76 L=36.19 1 6 PI L
FIL DBC S111 GR1 FIL DBr S211 GR1
FREQ SWEEP 2 8 .03
GRID RANGE 2 8 . 5 GR1 -100 30 10
RHO 0. 84 RGH=O ! RT 5880 ! INPUT SECTION ! 1ST SC STUB ! QW AT 3 GHz !2ND SC STUB ! QW AT 3 GHz ! 3RD SC STUB ! QW AT 3 GHz ! 4TH SC STUB ! OUTPUT SECT! ON
Univers
ity of
Cap
e Town
Touchstone (TM> - Configuration< 100 1600 102 16517 1968 1000 1 3294 ) 6GSNOTCH. CKT Tue Sep 26 00: 54: 27 1989
EDIT THIS FILE TO CREATE NEW CIRCUIT FILES. DESIGN FOR 6 GHz NOTCH FILTER I 3 GHz PASS FILTER INCORPORATING THE SCREW RESISTANCE TO GROUND <SC SECTIONS> 13 JULY 19 8 8
DIM FREQ GHZ RES OH I ND NH CAP PF LNG MM TIME PS COND JOH ANG DEG
CKT
MSUB ER=2. 2 H=O. 2548 T=O. 01778 MLIN 1 . 2 W=O. 76 L=36. 19 MLIN 2 3 W=O. 76 L=18.29 RES 3 0 R=1 MLIN 2 4 W=O. 76 L=8. 3825 MLIN 4 5 W=O. 76 L=18.29 RES 5 0 R=1 MLIN 4 6 W=O. 76 L=8. 3825 MLIN 6 7 W=O. 76 L=18.29 RES 7 0 R=1 MLIN 6 8 W=O. 76 L=8. 3825 MLIN 8 9 W=O. 76 L=18.29 RES 9 0 R=1 MLIN 8 10 W=O. 76 L=8. 3825 MLIN 10 11 W=O. 76 L=18.29 RES 11 0 R=1 MLIN 10 12 W=O. 76 L=8. 3825 MLIN 12 13 W=O. 76 L=18.29 RES 13 0 R=.1 MLIN 12 14 W=O. 76 L=36. 19 DEF2P 1 14 FIL
OUT
FIL DBC S11J GR1 FIL DBCS21J GR1
FREQ SWEEP 2 8 . 03
GRID RANGE 2 8 .5 GR1 -100 30 10
RHO=O. 84 RGH=O ! RT 5880
182-
! INPUT SECT I ON ! 1 ST SC STUB ! RES TO GROUND ! QW AT 3 GHz ! 2ND SC STUB ! RES TO GROUND ! QW AT 3 GHz ! 3RD SC STUB ! RES TO GROUND ! QW AT 3 GHz ! 4TH SC STUB ! RES TO GROUND
! OUTPUT SECT! ON
Touchstone (nn - Configuration( 100 1600 102 16517 1968 1000 1. 3294 ) 6GSNOTCH.CKT Tue Sap 26 00: 54: 27 1989
EDIT THIS FILE TO CREATE NEW CIRCUIT FILES. DESIGN FOR 6 GHz NOTCH FILTER / 3 GHz PASS FILTER INCORPORATING THE SCREW RESISTANCE TO GROUND (SC SECTIONS) 13 J UL Y 1988
DIM FREQ GHZ RES OH IND NH CAP PF LNG MM TIME PS COND 10H ANG DEG
CKT
HSUB ER=2.2 H=0.2548 T=0.01778 HLIN 1 . 2 W=O. 76 HLIN 2 3 W=O. 76 RES 3 0 R=1 HLIN 2 4 W O. 76 HLIN 4 5 W O. 7.6 RES 5 0 R=1 HLIN 4 6 W=O. 76 MLIN 6 7 W=O. 76 RES 7 0 R=1 HLIN 6 8 W=O. 76 MLIN 8 9 W O. 76 RES 9 0 R=1 MLIN 8 10 W=0.76 MLIN 10 11 W=O. 76 RES 11 0 R=1 HLIN 10 12 W=0.76 HLIN 12 13 W=0.76 RES 13 0 R=.1 MLIN 12 14 W=O. 76 DEF2P 1 14 FIL
OUT
FIL DBl S11) GR1 FIL DBl S211 GR1
FREQ SWEEP 2 8 . 03
GRID RANGE 2 8 . 5 GR1 -100 30 10
L=36.19 L=18. 29
L=8. 3825 L 18.29
L=8.3825 L=18. 29
L=8. 3825 L=18.29
L=8.3825 L=18.29
L=8. 3825 L=18.29
L=36. 19
RHO=0.84 RGH=O !RT 5880 !INPUT SECTION ! 1 SC STUB ! RES TO GROUND ! QW AT 3 GHz !2ND SC STUB ! RES TO GROUND ! QW AT 3 GHz ! 3RD SC STUB ! RES TO GROUND ! QW AT 3 GHz ! 4TH SC STUB ! RES TO GROUND
! OUTPUT SECTION
Univers
ity of
Cap
e Town
APPENDIX H
THREE PORT AND POWER SPLITTER SIMULATION PRQGRAMS
183 183
Univers
ity of
Cap
e Town
Touchstone < TMl - Configuration< 100 1600 102 16517 1968 1000 1 3294 l 3PORT. CKT Tue Sep 26 00: 31: 03 1989
! CALCULATING THE ISOLATION BETWEEN PORTS FOR A SIMPLE SPLITTER ! OPERATING RANGE 2-8 GHz ! 23 JANUARY 1989
DIM FREQ GHZ RES OH IND NH CAP PF LNG MM· TI ME PS COND /OH ANG DEG
CKT MSUB ER=2. 2 H=O. 254 T=O. 01778 RHO=O. 84 RGH=O MLIN 1 2 W=O. 76 L=18. 29 MLIN 2 3 W=0 .. 76 L=18. 29 MLIN 2 4 W=O. 76 L=18. 29 DEF3P 1 3 4 SPLIT
FREQ SWEEP 2 8 0. 5
OUT SPLIT 08[ S11J GR1 SPLIT OBCS12l GR1 SPLIT DBC S131 GR1 SPLIT OBC S23l GR1
GRID RANGE 2 8 0. 5 GR1 -70 10 10
IB4-
Touchstone (TM) - Configuration( 100 1600 102 16517 1968 1000 1 3294 ) 3PORT. CRT Tue Sep 26 00: 31: 03 1989
! CALCULATING THE ISOLATION BETWEEN PORTS FOR A SIMPLE SPLITTER !OPERATING RANGE 2-8 GHz ! 23 JANUARY 1989
DI M FREQ GHZ
. RES OH INO NH CAP PF LNG MMTI ME PS COND JOH ANG DEG
CRT MSUB ER=2.2 H 0.254 T 0.01778 RHO=0.84 RGH=O MLIN 1 2 W 0.76 L=18.29 MLIN 2 3 "=0 .. 76 L=18.29 MLIN 2 4 "=0.76 L=18.29 OEF3P 1 3 4 SPLIT
FREQ SWEEP 2 8 O. 5
OUT SPLIT OB[S11] GR1 SPLIT DB[S12J GR1 SPLIT DB[ S13] GR1 SPLIT DB[ S23J GR1
GRID RANGE 2 8 0.5 GR1 70 10 10
Univers
ity of
Cap
e Town
Touchstone <TM> - Configuration< 100 1600 102 16517 1968" 1000 1 3294 3PORT.OUT Tue· Sep 26 00: 32: 29 1989
FREQ-GHZ DBC S11J DBC S12J DBCS13J DBC S23J S.PLI T SPLIT SPLIT SPLIT
·2. 00000 -9. 686 -3. 586 -3. 586 -3. 586 2. 50000 -9. 655 -3. 593 -3. 593 -3. 593 3. 00000 -9. 621 -3·. 6 00 -3.600 -3.600 3. 50000 -9. 593 -3. 607 -3. 607 -3. 607 4. 00000 -9. 577 -3. 612 -3. 612 -3. 612 4. 50000 -9. 578 -3. 618 -3. 618 -3. 618 5. 00000 -9. 594 .._ 3. 6 2 3, -3. 623 -3. 623 5. 50000 -9. 621 -3. 628 -3. 628 -3. 628 6. 00000 -9. 653 -3. 633 ·-3. 633. -3. 633
. 6. 50000 -9. 683 -3. 637 -3. 637 -3. 637 7. 00000 -9. 705 -3. 642 -3. 642 -3. 642 7.50000 -9. 714 -3. 646 -3. 646 -3. 646 8. 00000 -9. 710 -3. 650 -3. 650 -3. 650
155
Touchstone ( T H) - Configuration( 100 1600 102 16517 1968' 1000 1 3294 3 PORT. OUT Tue 26 00: 32: 29 1989
FREQ-GHZ DBr S11] D8[S121 DBr S131 ( S2 3] SPLIT SPLIT SPLIT SPLIT
'2.00000 -9.686 -3.586 -3.586 -3.586 2.50000 9 655 -3.593 -3.593 -3.593 3. 00000 -9.621 - 3'. 600 -3.600 -3.600 3.50000 -9 593 -3.607 3.607 -3.607 4. 00000 -9 577 -3.612 -3.612 -3.612 4.50000 -9.578 -3,618 3. 618 3. 618 5. 00000 9.594 '... 3. 6 2 3~ -3.623 -3.623 5.50000 -9.621 -3.628 -3.628 3 628 6. 00000 -9.653 -3.633 '-3.633 . -3.633
·6.50000 -9.683 -3.637 -3.637 -3.637 7.00000 9. 70S -3.642 -3.642 -3.642 7.50000 -9.714 -3.646 -3.646 3. 646 8. 00000 -9.710 -3.650 -3.650 -3.650
Univers
ity of
Cap
e Town
Touchstone ( TMl - Configuration( 100 1600 102 16517 1968 1000 1 3294 3PORT2. CKT Tue Sep 26 00: 33: 58 1989
! CALCULATING THE ISOLATION BETWEEN PORTS FOR A SIMPLE SPLITTER ! OPERATING RANGE 2-8 GHz ! 23 JANUARY 1989
DIM FREGl GHZ RES OH IND NH CAP H LNG MM TI ME PS COND /OH ANG DEG
CKT
MSUB ER=2. 2 H=O. 254 T=O. 01778 RHO=O. 84 RGH=O MLIN 1 2 H=O. 76 L=18. 29 MLIN 2 3 W=O. 42 L=9. 31 MLIN 2 4 H=O. 42 L=9. 31 DEF3P 1 3 4 SPLIT
FREQ SHEEP 2 8 0. 5
OUT SPLIT DBC S111 GR1 SPLIT DBC S12l GR1 SPLIT DB( S13l GR1 SPLIT DBCS231 GR1
GRID RANGE 2 8 0. 5 GR1 -70 10 10
187
Touchstone (TM) - Configuration( 100 1600 102 16517 1968 1000 1 3294
3PORT2. CKT Tue Sep 26 00: 33' 58 1989
! CALCULATING THE ISOLATION BETWEEN PORTS FOR A SIMPLE SPLITTER ! OPERATING RANGE 2-8 GHz ! 23 JANUARY 1989
DIM FRE(;) GHZ RES OH INO NH CAP PI." LNG MM TIME PS COND 10H ANG OEG
CKT MSUB ER=2.2 MLIN 1 2 MLI N 2 3 MLIN 2 4 DEF3P 1 3
FREQ SWEEP 2 8 0.5
OUT SPLIT DB! S11J GR1 SPLIT OB[S12J GR1 SPLIT DB( S13J GR1 SPLIT DB(S23J GR1
GRID RANGE 2 8 0.5 GR1 -70 10 10
H=0.254 T=0.01778 RHO=O. 84 RGH=O W 0.76 L:18.29 w 0.42 L 9. 31 W=0.42 L 9. 31
4 SPLIT
Univers
ity of
Cap
e Town
Touchstone ( TMl - Configuration< 100 1600 102 1b517 1968 1000 1 3294 6GSPLIT. CKT Tue Sep 26 00: 40: 37 1989
! EDIT THIS FILE TO CREATE NEW CIRCUIT FILES. ! 6GHz POWER SPLITTER
DIM FREQ GHZ RES OH IND ~lH
CAP PF LNG MM TI ME PS COND , OH ANG ·DEG
CKT MSUB ER=2, 2 H=O. 2548 T"" 0 .. 017 7 8 MLIN 1 2 W=O. 76 L=36. 57 MLI N 2 3 w = 0. 4 3 L=9. 3 'l MLI N 2 4 W=O. 1~3 L=9. 31 RES 3 4 R=100 MLI N 3 5 W=O. 7b L=36. 57 MLI N 4 6 W=O. 76 L-=3b. 57 DEF 3-P 1 5 6 SPLIT
OUT SPLIT DB( S11l GR1 SPLIT DBfS12l GR1 SPLIT DBCS13l GR1 SPLIT DB( S23l GR1
FR E(~ SWEEP 2 8 . 03
GRID RANGE 2 8 . 5 GR1 -so 10 10
\BB
RHO=O, 84 RGH=O. 00 ! 50 OHM INPUT ! 1 ST QW SPLIT ELEMENT ! 2 ND QW SPLIT ELEMENT ! RES. ACROSS PORTS 2 & ! 5 0 OHM OUTPUT SECTION ! 50 OHM OUTPUT SECTION
3
Touchstone (TM) - Configuration( 100 1600 102 16517 1968 1000 1 3294 6GSPLIT. CKT Tue Sep 26 00' 40: 37 1989
! EDIT THIS fILE TO CREATE NEW CIRCUIT FILES. ! 6GHz POWER SPLITTER
DIM F REfJ GHZ RES OH IND ~l H CAP PF LNG MM TIME PS COND . OH ANG ,DEG
CKT MSUB ER 2. 2 H=0.2548 T",O .. 01778 RHO=O.84 RGH=O. 00 MLIN MLIN MLIN RES MLIN MLIN DEF 3-P
OUT SPLIT SPLIT SPLIT SPLIT
FREGI SWEEP
GRID RANGE GR1
1 2 W=O.76 2 3 H 0 43 2 4 W O. li 3
3 4 R 100
3 5 W O. 7b 4 6 W=0.76
1 ') 6 SPLIT
DB[ S11] GR1 DB[ S12] GR1 DB[S13J GR1 DB[ S23J GR1
2 8 . 03
2 8 .5 -1)0 10 10
L=36.5 L=9.31 L=9.31
L 36. ') 7 L.::3b.57
B
! 50 OHM INPUT ! 1ST QW SPLIT ELEMENT ! 2ND OW SPLIT ELEMENT ! RES. ACROSS PORTS 2 g. 3 ! 50 OHM OUTPUT SECTION ! 50 HM OUTPUT SECTION
Univers
ity of
Cap
e Town
Touchstone ( TMi - Configuration< 100 1600 102 16517 1968 1000 1 3294 J
6GSPLIT2. CKT Tue Sep 26 00: 36: 42 1989
! EDIT THIS FILE TO CREATE NEW CIRCUIT FILES. ! 6GHz POWER SPLITTER
DIM FREQ GHZ RES OH IND NH CAP PF LNG MM TI ME PS COND I OH ANG DEG
CK T MSUB ER=2. 2 MLI N 1 2
H 00 0. 2548 T=O. 01778 RHO=O. 84 W=O. 7244 L=36. 57
RGH=O. 00 ! 50 OHM INPUT
MTEE 2 3 4 W1=0. 7244 W2=0. 3944 W3=0. 3944 ! APPROX. TEE JUNCTION MLIN MLIN TFR MTAPER MTAPER MLI N MLI N DEF3P
OUT SPLIT SPLIT SPLIT SPLIT
FREQ SWEEP
GRID RANGE GR1
3 5 W=O. 3944 L=45. 88 4 6 W=O. 3944 L=45. 88 5 b W=O. 8 L=O. 5 RS=100 F=O
5 7 W1=0. 3944 W2=0. 7244 L=O. 4 6 8 W1=0. 3944 W2=0. 7244 L=O. 4
7 9 W=O. 7244 L=36. 57 8 10 W=O. 7244 L=36. 57
1 9 10 SPLIT
DB[ S11l GR1 DBlS12l GR1 DBCS13l GR1 DB[ S23l GR1
2 8 . 03
2 8 . 5 -50 10 10
185
! 1ST QW SPLIT ELEMENT ! 2ND QW SPLIT ELEMENT
! RES. APPROX ACROSS PORTS 2&3 ! APPROX. OF TAPER EFFECT
! 50 OHM OUTPUT SECTION ! 50 OHM OUTPUT SECTION
Touchstone (TM) - Configut'ation( 100 1600 102 16517 1968 1000 1 3294 ) 6GSPLIT2. CKT Tue Sap 26 00: 36: 42 1989
! EDIT THIS FILE TO CREATE NEW CIRCUIT FILES. ! 6GHz pOWER SPLITTER
DI M FREQ GHZ RES OH IND NH CAP PF LNG MM TI ME PS COND 10H ANG DEG
CKT MSUB ER=2.2 MLIN 1 2
H,.,0.2548 T=0.01778 RHO=0.84 W:=O.7244 L=3b.57
RGH=O. 00 ! 50 OHM INPUT
HT 2 3 4 \011 =0.7244 W2=O.3944 W3=0.3944 ! APPROX. TEE JUNCTION HLIN 3 5 w=o 3944 L=45.88 ! 1 ST ,.;)w SPLIT ELEMENT MLIN It 6 W=0.3944 L=45.88 ! 2ND 101 SPLIT ELEMENT TFR 5 b Ii' 0.8 L=0.5 RS=100 F=O ! RES. APPROX ACROSS PORTS 2&3 MTAPER 5 7 w1=0.3944 W2=0.7244 L=0.4 ! APPROX. OF TAPER EF F ECT MTAPER 6 a W1=0 3944 W2=0.7244 L=O. 4 MLIN 7 9 W=0.7244 L=36.57 ! 50 OHM OUTPUT SECTION MLIN 8 10 101=0.7244 L=36.57 !50 OHM OUTPUT SECTI ON DEF3P 1 9 10 SPLIT
OUT SPLIT DB[ 311J GR1 SPLIT DB[5121 GR1 SPLIT DB[:::;131 GR1 SPLI T DB[S23J GR1
FREQ SwEEP 2 8 .03
GRID RANGE 2 8 .5 GR1 -50 10 10
\85
Univers
ity of
Cap
e Town
APPENDIX I
MODE AMPLITUDE FORMULAS AND COUPLER SIMULATION PROGRAMS
190 190
Univers
ity of
Cap
e Town
1 Coupler mode amplitude formulas i
Case 2 (kL¢kc, fh =f32 and matched terminations): In this case
----·--- -·------- --- ,-----·--
4-(0) (kL + kc){j[(</>2 - </>1) sin (81 + 82) + (<1>2 + <1>1) sin (81 - 82)] - ~[cos (81 - 82) - cos (81 + 82)]J ---= -----------------------::----~-----,-----a_(Q) [4(1 + ¢1¢2) - A2
] cos (81 + 82) - (4(1 - ¢1<1>2) - .1 2 ) cos (81 - 82) + j\ [4(¢1 - </>2)
+ 2.1(4>2 - ¢J J sin (81 + 82) - [4 (</>1 - <1>2) - 2.1(¢1 + 4>2) J sin (81 - 82)}
a .. (l) 2¢d2<1>1 cos 81 + j(2 + A) sin 81] + 2¢1[2¢2 cos 82 -f- j(2 - .1) si; 82f ----- -------- ------ = -----------------a.,.(O) DEX of (32)
b_(O) (k L + kc) { [2 cos (81 + 82) - 2 cos (81 - 82) J + j[ (4>1 + ¢2) sin (81 + 82) + (¢2 - ¢1) sin (81 ...:. 82)]} -= a .. (O) DEX of (32)
b+-(l) 2¢1 [2¢: cos 82 + j(2 - A) sin 82] - 21>2 [24>1 cos 81 + j(2 + A) i:in 81] -- = - ------------~(O) DEX of (32)
The coupling is given by
c = I b-(0) ! , a+(O)
and the directivity is given by
D=lb+(l)I b_(O)
2(<1>1<1>2)(cos82 - cos81) i + j(2(<1>1 sin 82 - <I>: sin 81)
1 - ~(4>1 sin 82 + <1>2 sin 81) J 2[cos (81 + 82) - cos (81 - 82)]
(kL +kc)
2 + j((<1>1 + </>2) sin (81 + 82)
+ (<1>1 - </>2) sin (81 - 82)]
The input VSWR and impedance may be calculated from •
I a_(O) I 1+ -.-
. t VSW~R a+(O) mpu =
I a_(O) I
1- --a+(O)
where a__(O)/ a+(O) is the input reflection coefficient.
I 5 I
I Coupler mode amplitude formulas i
Case 2 (kL~kc, (jl={j: and matched terminations): In this case
4-(0) (kL + kC)U[(¢2 - ¢1) sin (81 + 82) + (¢2 + ¢1) sin (81 - 82)] - ~[eos (81 - 82) - cos (81 + 82)H ---= -(L(O) [40 + ¢192) - t.V] cos (8 1 + 82) - [-1(1 - ¢I¢~) - tl~] cos (81 - 82) + i! [4(¢1 - ¢2)
+ 2tl(¢2 - ¢J] sin (81 + 82) - [4(4)1 - ¢2) - 2j.(4)1 + ¢:) J sin (81 - 82)}
a .... (l) 2¢J2¢1 COS 81 + j(2 + A) sin od + 2¢1[2¢2 cos 82 -+- j(2 - A) si~ 82] -- = --=---.-----.:...-------=--------(1.,.(0) DEX of (32)
1.._(0) (I; L + kc) { [2 cos (81 + 82) - 2 cos (81 - 82)] + j[ (¢1 + ¢~) sin (81 + (2) + (¢2 - ¢l) sin (81 ...:. O2) H -= d~(O) DEX of (32)
b ... (l) 2¢1[2¢: cos 82 + j(2 - A) sin 82] - 2<Pz[2¢1 cos 81 + j(2 + A) sin 8d - = - -- ------------a,.{O) DEX of (32)
The coupling is given by
and the directivity is given by
D = I b+(l) I b_(O)
2(¢1¢2)(COS 8: - cos 81) i + j[2(¢1 sin 82 - I/J: sin 81) I - j.(¢l sin 82 + ¢2 sin ( 1)]
2[cos (81 + 82) - cos (81 - 8%) J
(kL + kc)
2 + i[(l/Jl + I/J:) sin (91 + (2)
+ (I/Jl - 1/J2) sin (81 - 92)]
The input VSWR and impedance may be calculated from .
where 0-1...0)/0+(0) is the input reflection coefficient.
\ 5 \
Univers
ity of
Cap
e Town
THIS PROGRAM IS DESIGNED TO AID IN THE CALCULATION OF ODD AND EVEN IMPEDANCES FOR MICROSTRIP DIRECTIONAL COUPLED LINES
BY P. ALBERTS SEPTEMBER 1988
* Definition of variables* ZOE - EVEN MODE IMPEDANCE ZOO - ODD MODE IMPEDANCE ZO - CHARACTERISTIC LINE I~PEDANCE W - WIDTH OF LINE I IN mm) S - SPACING BETWEEN LINES IIN mm) L - LENGTH OF LINE <IN mm> ER - RELATIVE PERMEABILTIY OF DIELECTRIC H - HEIGHT OF DIELECTRIC IIN mm> F - FREE SPACE FREQUENCY I IN Hz> * INPUT THE MICROSTRIP VARIABLES
INPUT PROMPT "ER OF BOARD ". ER INPUT PROMPT "UR OF BOARD! NOMINALLY=1> ":UR INPUT PROMPT "EREF OF BOARD ". EREF INPUT PROMPT "HEIGHT OF DIELECTRIC I IN mm> ... H INPUT PROMPT "CHOOSE CHARACTER! STI C LINE IMPEDANCE ". zo INPUT PROMPT "ENTER LENGTH OF LINE I QUARTER WAVELENGTH> INPUT PROMPT "ENTER SPACING BETWEEN LINES I IN mm> ". s INPUT PROMPT "ENTER WIDTH OF LINE I IN mm> ... w INPUT PROMPT "ENTER CENTER FREQUENCY C I N Hzl ". F
! CALCULATING THE FREE SPACE WAVELENGTH
LET C=3E8 !--ET LO = C I F
! * CALCULATING THE CAPACITIVE AND INDUCTIVE COEFICIENTS
LET A1 1 + 0. 25*LOGI C ER+1) 12>
LET B1 I SQRI ER+1> > /10 LET KC 0. 55*EXPI -( A1*S/H + B1*WIH> >
LET A2 1 + 0. 25*LOGC C UR+1) /2) LET B2 I SQRC UR+1> > /10 LET KL 0. 55*EXPC -C A2*S/H + B2"'WIH> >
PRINT "KC= ";KC,"KL= ";KL
". L
! * CALCULATING THE PHASE CONSTANT AND THE LENGTH OF THE COUPLING SECTION
LET B I 2*PI*F/C) *C SQRC EREFl l *SQRC C 1-KL*KCl /C 1-KL*KL> l LET L CLO/I 4*SQRI EREF> l l *SQRI C 1-KL*KLl /( 1-KL*KCl > LET 0 = B*L
PRINT "LENGTH CIN mml ";L*1E3
1 * CALCULATING THE COUPLING, DIRECTIVITY AND VSWR CONSTANTS
LET C LET C
< < KL+KCl /2l *SIN< Ol -20*LOG10C Cl
LET D = I ( KL-KCl I< KL+KCl l *O/ SI NI Ol LET D -20*LOG10CDl
)92
THIS PROGRAM IS DESIGNED TO AID IN THE CALCULATION OF ODD AND EVEN IMPEDANCES FOR MICROSTRIP DIRECTIONAL COUPLED LINES
BY P. ALBERTS SEPTEMBER 1988
'" Definition of variables'" ZOE - EVEN MODE IMPEDANCE ZOO ODD MODE IMPEDANCE ZO CHARACTERISTIC LINE I~PEDANCE W - WIDTH OF LINE (IN mm) S - SPACING BETWEEN LINES (IN mm) L - LENGTH OF LINE (IN mm) ER - RELATIVE PERMEABILTIY OF DIELECTRIC H - HEIGHT OF DIELECTRIC (IN mm) F - FREE SPACE FREQUENCY (IN Hz) '" INPUT THE MICROSTRIP VARIABLES
I NPUT PROMPT .. ER OF BOARD to: ER I NP UT PROMPT "UR OF BOARD( NOMI NALLY:: 1) ": UR INPUT PROMPT "EREF OF BOARD n: EREF INPUT PROMPT "HEIGHT OF DIELECTRIC (IN mm) ": H INPUT PROMPT "CHOOSE CHARACTERISTIC LINE IMPEDANCE ": ZO INPUT PROMPT "ENTER LENGTH OF LINE (QUARTER WAVELENGTH) ". L I NP UT PROMPT .. ENTER SPACI NG BlnWEE N LI NES (I N mm) ": S I NPUT PROMPT" ENTER WI DTH OF LI NE (I N mm) ": W INPUT PROMPT "ENTER CENTER FREQUENCY (IN Hz) It: F
! CALCULATING THE FREE SPACE WAVELENGTH
LET C 3E8 p:T LO == C I F
! '" CALCULATING THE CAPACITIVE AND INDUCTIVE COEFICIENTS
LET A1 :::: 1 + O. 25"'LOG( (ER+1) 12)
LET B1 == (SQR( ER+11 I 110 LET KC == O. 55"'EXP( -( A1"'S/H + B1 "'WI HI I
LET A2 == 1 + O. 25"'LOG( (UR+1) 12) LET B2 ::: (SQR( UR+1)) 110 LET KL ::: O. 55"'EXP( -( A2"'S/H + B2"'W/H»
PRINT "KC:::: "; KC, "IL:::: "; KL
I '" CALCULATING THE PHASE CONSTANT AND THE LENGTH OF THE COUPLING SECTION
LET B :::: (2*PI"'F/CI "'( SQR( EREFI I "'SQR( (1-KL"'KC) I( 1-KL"'KL» LET L :::: (LOt( 4"'SQR( EREF») "'SQR( (1-KL"'KLl I( 1 KL"'KC» LET 0 B"'L
PRINT "LENGTH (IN mm) n;L"'1E3
! '" CALCULATING THE COUPLI DIRECTIVITY AND VSWR CONSTANTS
LET C :::: «KL+KCI/2)"'SIN(Q) LET C -20"'LOG10(C)
LET D :::: « KL KC) I( KL+KC» "'O/SINI OJ LET D -20"'LOG10(D)
Univers
ity of
Cap
e Town
LET A= 1+CKL+KCl*<<KL-KCl/4)*(CSIN<Ol) 2+0*CO+SINC2*0lll LET VSWR = C1+ABS<All/(1-ABS<All
PRINT "CC IN dBl = "; C, "DC IN dBl = "; D. "VSWR= "; VSWR
1 "' ODD AND EVEN MODE IMPEDANCE
END
[33
LET A == 1+( KL+KC) *( (KL-KCl/4l *( (SIN( 0» ~2+0"{ O+SIN( 2"'0») LET VSWR ::: (1+ABS(A)}f(1-ABS(A»
PRINT "C( IN dB);;:: n; C. "D( IN dB);;:: "; D. "VSWR= "; VSWR I " ODD AND EVEN MODE IMPEDANCE
END
183
Univers
ity of
Cap
e Town
Touchstone < TMJ _- Configuration< 100 1600 102 16517 1968 1000 1 3294 6GCOUP2. CKT Tue Sep 26 00: 45: 17 1989
THIS PROGRAM DETERMINES THE COUPLING CONSTANTS FOR A COUPLER WITH A QUARTER WAVELENGTH COUPLING SECTION. CENTRE FREQUENCY IS 6 GHz
DIM FREQ GHZ RES OH IND NH CAP PF LNG MM TIME PS COND I OH ANG DEG
CRT MSUB ER=9. 6 H=O. 635 T=O. 034 RHO=O. 84 RGH=O MCLIN 1 2 3 4 W=. 56 S=. 4 L=4. 86 ! SEPAR. =O. 4mm DEF4P 1 2 3 4 COUP
OUT COUP DBCS111 GR1 COUP DBC S21 l GR1 COUP DBCS.311 GR1 COUP DBCS41J GR1
FREQ SWEEP 2 8 0. 5
GRID RANGE 2 8 0. 5 GR1 -50 10 10
! 84-
Touchstone (TM) _- Configuration( 100 1600 102 16517 1968 1000 1 3294 6GCOUP2. CRT Tue Sep 26 00: 45: 17 1989
THIS PROGRAM DETERMINES THE COUPLING CONSTANTS FOR A COUPLER WITH A QUARTER WAVELENGTH COUPLING SECTION. CENTRE FREQUENCY IS 6 GHz
DI M FREQ GHZ RES OH INO NH CAP PF LNG MM TIME PS CONO ; OH ANG DEG
CRT
OUT
MSUB MCLIN DEF4P
ER=9 6 H=0.635 T=0.034 RHO=0.84 1234 W.56 3=.4 L=4.86 1 2 3 4 COUP
COUP DBC S11J GR1 COUP DBC321J GR1 COUP DB[ 1 J GR1 COUP OB[S411 GR1
FREQ SWEEP 2 8 O. 5
GRID RANGE 2 8 O. 5 GR1 50 10 10
RGH=O !SEPAR. =0. 4mm
Univers
ity of
Cap
e Town
Touchstone <TM> - Configuration< 100 1600 102 16517 1968 1000 1 3294 >
6GCOUP3. CKT Tue Sep 26 00: 48: 04 1989
! THIS PROGRAM DETERMINES THE COUPLING CONSTANTS FOR A COUPLER ! WI TH A QUARTER WAVELENGTH COUP LI NG SEC TI ON. CENTRE FREQUENCY IS 6 GHz
DIM FREQ GHZ RES OH I ND NH CAP PF LNG MM TIME PS COND /OH ANG DEG
CKT MSUB MCLIN DEF4P
ER=9. 6 H=O. 635 T=O. 034 1 2 3 4 W=. 492 S=. 468 1 2 3 4 COUP
OUT COUP DBCS11J GR1 COUP DBC S21J GR1 COUP DBCS311 GR1 COUP DBC S41J GR1
FREQ SWEEP 2 8 0. 5
GRID RANGE 2 8 0. 5 GR1 -50 10 10
RHO=O. 84 L=4. 86
\95
RGH=O ! SEPAR. =O. 4mm
TouchstonQ (TM) - Configuration( 100 1600 102 16517 1968 1000 1 3294 ) 6GCOUP3. CKT Tue Sep 26 00: 48: 04 1989
! THIS PROGRAM DETERMINES THE COUPLING CONSTANTS FOR A COUPLER ! WI TH A QUARTER WAVELENGTH COUPLI NG SECTION. CENTRE FREQUENCY IS 6 GHz
DIH FREQ GHZ RES OH IND NH CAP PF LNG HH TIME PS COND IOH ANG DEG
CKT HSUB HCLIN DEF4P
ER=9.6 H=0.635 T=0.034 1 2 3 4 W=. 492 S= 468 1 2 3 4 COUP
OUT COUP DB[S111 GR1 COUP DB[S21] GR1 COUP DB[S311 GR1 COUP DB[S411 GR1
FREQ SWEEP 2 8 O. 5
GRID RANGE 2 8 0.5 GR1 50 10 10
RHO=O. 84 L=4.86
RGH=O ! SEPAR. =0. 4mm
Univers
ity of
Cap
e Town
APPENDIX J
AVANTEK VT0-8100 OSCILLATOR CHARACTERISTCS
196 196
Univers
ity of
Cap
e Town
....-•• I •
LIMITED FREQUENCY RANGE VARACTOR-TUNED OSCILLATORS
COMMERCIAL VARACTOR TUNED OSCILLATORS Avantek VT0-8000 Series oscillators use a silicon transistor chip as a negative resistance oscillator. The oscillation frequency is determined by a varactor diode acting as a voltage-variable capacitor in a thinfilm microstripline resonator. This provides extremely fast tuning speed. limited primarily by the internal impedance of the user-supplied voltage driver.
VT0-8000 SERIES
A typical oscillator can be swept across its frequency band in less than one microsecond.
The VT0-8000 Series varactor-tuned oscillators are packaged in T0-8 transistor cans for simple installation in a conventional 50-0hm microstripline PC board. They are ideal for the most compact. lightweight commercial and military equipment designs.
Guaran:eed Specifications at 25' C Case Temperature (0 to 65° C Operating Temperature) PCS
Po wet Pow•t Tuning Volfa91' LJmitl lnputPoW9t' Output Output (at ~actt end of ( :1-. R09-) ...
Fre-ou~ncy •nto SO Variation specifle-d lreq. ra~e) 'current Harmonk:a Ra~• Ohms. Min. Mui mum •YOC@ •YOC@ Yoilage (mA} Typial c-
"'od•I {GHz) (dBm) (dB} Low Freq. High F..q. (YDC) Maximum !dBcl TY!>O
VT0-8030 O.HIAS -10 :1.5 5::4 50::10 -15 so -15 T0-8V VT0-8040 0.-J-0.6 -13 ::1 5 3:'.:1 40::'.:8 •15 so -15 TO-av VT0-8060 0.6-1 0 ·13 :1.5 3::1 40::8 •15 so -15 T0-8V VT0-8080 0.8-1.4 ·13 :::1.5 2::1.5 35±10 •15 so -15 T0-8V VT0-8090 0.9-1.6 ·13 :::1.5 2±1 48+8; -10 •15 so -15 T0-8V VT0-8100 1.0· 1.4 ·10 :!:1.5 2::1 48±8 •15 so -15 T0-8V VT0-8150 1.5·2.5 ·10 ::1.5 2.5±1 47::8 •15 so -18 To-av VT0-8200 2.0-3.0 -10 ::1.5 2•2. -1 30:'.:8 •15 50 -18 T0-8V VT0-8240 2.4-3.7 ·10 ::1 5 2•2; -1 30:::8 •15 so -18 T0-8V VT0·8300 3.0-3.5 ·10 :'.:1.5 3.5 min. 11 max. +15 50 -18 T0-8V VT0-8350 3.5-4.5 •10 ::1.5 5 min. 35 max. +15 so -20 T0-8V VT0-8360 3.6-4.3 ·10 :!:1.5 8±2 24±4 +15 50 -25 T0-8V VT0-8400 4.0-4.5 ·10 ::1.5 2 min. 14 max. ·15 so -25 T0-8V VT0-8420 4.2-5.0 -10 ::1.5 7.5c:2.5 25•2.5; -4 ·15 50 -25 T0-8V VT0-8430 4.3·5.8 ·10 ::1.5 5.5±2 24+3 •15 so -25 T0-8V VT0-8490 4.9-5.9 •10 ±1.5 5.5±2 24+3; -4 +15 so -25 T0-8V VT0-8520 5.2-6.1 ·10 ±1.5 5.5±2 24±3 +15 so -25 T0-8V VT0-8540 5.4-5.9 •10 :!:1.5 8 min. 28 max. +15 so -15 T0-8V VT0-8580 5.8-6.6 •7 ±1.5 5±2.5 24+3; -5 +15 so -25 T0-8V VT0-8650 6.5-8.6 •10 ::1.5 2:!:1 20±5 •15 100 -20 TO:.SV VT0-8790 7.9-10.1 •10 ::2 3±2 26±4 •15 150 -10 T0-8V VTo.a&10 8.1-9.1 •iO ±2 2 min. 16 max. +15 100 -15 TO-'BV VT0-8850 8.5-9.6 •1a ±1.5 5±2 13±5 +15 100 -25 T0-8V VT0-8950 9.5-10.5 '10 ±1.5 4±1 10 max. +15 100 -20 To-av VT0-81000 10.0-10.25 •10 :':1.5 0 min. 15 max. +15 100 -15 T0-8V
COMMERCIAL HYPERABRUPT VARA'cTOR TUNED OSCILLATORS This family of oscillators is similar to the standard commercial VT0-8000 Series except for the incorporation of a silicon hyperabrupt varactor tuning diode. This enables the oscillator to be tuned over the
specified range in less than 20 volts rather than 40-50 volts in conventional oscillators. They feature extremely fast tuning speed. limited primarily by the internal impedance of the user-supplied voltage driver.
VT0-9000 SERIES Guaranteed Specifications at 25°C Case Temperature (O" to 65°C Operating Temperature) PCll
p- Po- Tunln9 Voltage Umtta Input Power OUlput Output (at Heh end ot (~1 .. Aog.I All
F..quoncy Into 50 Yarl•tton ·~tied fNq. ,.,., Cul'Nftl H.....-R•ngo o_,...._ Maatmum +VOC@ •YDC@ Voltage (mAI ry....., c..o
M-1 (GHzl (dllmt (dB} Low Freq. . High F- (YDC} Maxlntu11t (-1 Typo
VT0-9032 0.32-0.64 ·10 ±2 O min. +20 max. •15 so -14 T0-8V VT0-9050 0.5-0.9 ·10 ±2 0 min. •20 max. +15 so -10 T0-8V VT0-9088 0.68-1.36 •10 ±2 0 min. +20 max. +15 50 -14 T0-8V VT0-9090 0.9-1.6 ·10 ±2 +2 min. +18 max. +15 so -14 T0-8V VT0-9120 1.2·2.0 +10 ±2 •2 min. +14 max. +15 so -14 TO-SV VT0-9130 1.3-2.3 •10 ±1.5 ,..2 min. +20 max. +15 so -15 T0-8V VT0-9140 1.4-2.1 •10 ±1.5 +4±2 +10±2 +15 50 -15 T0-8V
Anntek, Jnc. • 3175 BoWers Ave .. $.anla Clara. Ca 95054 • Phone. {408) 727-0700 • FAX: (408) 727-0539 • TWX. 310·371-8717 Of 310-371-8478 • TELEX 34--6337
VARACTOR Tl .A.vantek \HOtemper'!ture : VT0-8000 Se Jators. Trey ar guaranteed to of -54° :o ~s: higher reliab1
MT0-8000 SI Guaranteec Spec
MTQ-8040 MTQ-8060 MT0-8090 MT0-8240 MT0-8360 MTQ-8650 MT0-8950
CASE DRAWING CASE DRAWING
BUFFERED V The VTD Ser g ral bu lier arr in-line packa~ than 0.2 oz. fabricated or using advan< silicon transi~
Internal butte in load impec producing +1 VTD to be Uc wideband isc lightly-loadec
VTD SERIES Guaranteed Spe·
Model
VT0-600 VT0·2000 VTD-2800 VTD-3800 VT0-4900
IQ!antek. Inc. • :
• __ ...,,,...,.....,,, __ ...,.~ _ _,,.,w,.. ... 4€--. _.,.P_"l!t4P--WWW"""'~-•"'·-"'·IE"'' _ • .,,,, .. , _,..,...~·"'-._,_._ t''...,.""!Jlllill!lll!iilll""'-...-•
\ ff7 \
LIMITED fREQUENCY RANGE VARACTOR- TUNED OSCillATORS
COMMERCIAL VARAC10R TUNED OSCILLATORS Avantek VTO-BOOO Series oscillators use a silicon transistor chip as a resistance oscillator. The oscillation frequency determined by a varactor diode acting as a voltaga-variable capacitor in a thinfilm microstripline resonator. This provides extremely fast tuning speed, limited primarily by the internal
. impedance of the user-supplied voltage driver.
VTO-8000 SERIES
lITO-SOlO 0.3-0"5 -'0 =1.5 5,,::4
IITO-S040 0.-1-0.6 -13 =1 5 3'::' IITO·801iO 0.6·10 '13 =1.5 3::1 lITO-BOlIO 0.8-14 ~13 ::1.5 2:!:1.5 IITO-8O!lO 0.9-1.6 -13 ::1.5 2±1 IITO-81oo 1.0-1.4 ·10 :::1.5 2=1 IITo-al50 15-2.5 -10 ::1.5 2.5±1 IITO-S2oo 2.0-3.0 -10 -.::;1.5 2-2. IITO-8240 2.4-3.1 -10 =15 2.2: -1 IITO-8300 30-3.$ -10 ::1.$ 3.5 min. IITO-33S0 3.5~.$ -10 -.::;1.5 5 min. IITO-8360 3.6-4.3 -10 t1.5 8:2 VTO-S4oo 4.0~5 -10 =1.5 2 min. IITO-S420 4.2-5.0 -10 ::15 1.5,,::2.5 IITO-II430 4.3~5.8 -10 -.::;1.5 55±;2 VTO-11490 4.9-5.9 .10 :t1.5 5.S±2 VTO-8520 5.2-6.1 -10 ::1.5 5.5±2 IITO-8S40 5.4-5.9 -10 :';1.5 8 min. IITO-8511O 5.8-6.6 ·1 ::1.5 5±2.5 IITO-Ii650 6.S-8.6 -10 ::l.S 2::1 IITo-e790 7.9-10.1 -10 ::2 3±2 IITa-.10 B.1-9.1 -10 ±2 2 min. VTO-8850 85-9.6 .,. ±1.5 5±2 IITO-8950 9.5-10.5 dO ±15 4±1
10.0-10.25 '10 o min.
A typical oscillator can be swept across its frequency band in less than one microsecond.
The VTO-aooo Series varactor-tuned oscillators are in TO-8 transistor cans for simple installa
tion a conventional 50-Ohm microstripline PC board. They are ideal for the most compact. lightweight commercial and military equipment
50,,::10 -15 50 -15 TO-SV 40::8 .15 50 ·15 TQ-aV 40,,::8 ·15 50 -15 TO-SV 35:tl0 '15 50 -15 TO-8V
48+8; -10 -15 50 -15 TQ-aV 48±!I '15 50 -15 TQ-aV 47±8 '15 50 -H! TO-8V 30t8 '15 50 -18 TO-8V 30:::8 -15 50 ~la TQ-aV
11 mal(. +15 50 -18 TQ-a\l 35 max.. '15 50 -20 T0-8\1
24±;4 '15 50 -25 rQ-a\l 14 mal(. -15 50 -25 TO-8V
25+2.5: ~ -15 50 ·25 TO-8V 24+3 -15 50 -25 TO-S\!
24+3; -4 +15 50 -25 TQ-aV 24±3 +15 50 -25 T0-8\1
28 max.. +15 50 -15 T0-8\1 24+3;-5 +15 50 -25 TO-8V
2O±5 -IS 100 -20 r04l\l 2S±4 .15 150 -10 TO-8\1
16 max, '15 100 -15 T04l\l 13±5 +15 100 -25 T0-8\1
10 max. +15 100 -20 TO-8\1
COMMERCIAL HYPERAIRUPT VARA'cTOR TUNED OSCillATORS This family of oscillators is similar to the standard commercial VTO-aooo Series except for the incorporation of a silicon hyperabrupt varector tuning diode. This enables the oscillator to be tuned over the
VTO-9OOO SERIES
IITO-9032 0.32-0.64- -10 :';2 o min. IITO-90SO OS..().9 -10 ±2 o min. IITO-1iIOIY 0.68-1.36 '10 ::2 o min. IITO-9OIIO 09-1.S -10 ::2 '2 min. IITO-9120 1.2-2.0 ·10 :t2 '2 min, VTO-!l130 13-2.3 -10 !15 .. 2 min. IITO-!!I40 1.4-2.1 -10 ~4:t2
specified range in less than 20 volts rather than 40-50 \/olts in conventional oscillators. feature ex-tremely fast tuning speed. limited by the internal driver.
·20 max. ·15 +20 max. +15 +20 mu. +15 +18 max. .15 +14 max. +15 +20 mal(. '15
+10::2 -IS
50 50 50 50 50 50 50
-14 -10 -14 -14 -14 -15
TO-8\! T0-8V TO-8V T0-8V TO-8V TO-8V
..,-- .--=--,-----........."-.,.".,..---,~--,.",......, . ...,.--~.-...,.....;..,.....,,,..-~---.;"""""'-----......... -• , •
•
\
VARACTOR TI Allantek ~TOtemper'!lure : VTO-8000 S€:r Jators. Trey guaranteed to of _54° :0 ·85 higher reJiabl
MTO-SOOO 51 GU3r3Meec Spec
MT().8!)IO MTo-806O MT0-8090 MT0-8240 MTC>-8360 MTO-865O
CASE ORAWING CASE DRAWING
BUFFERED V The VTD Ser gral buffer arr in-line packa~ than 02 oz. fabricated or using advanc silicon transi~
Internal buffe in load impec producing +1 VTD to be u, wideband isc lightly-Ioadec
VTD SERIES Guarant~ Spe·
Univers
ity of
Cap
e Town
broadband linear
Power Amplifiers up to 1 Watt (+30 dBm)
700 MHz to 4.2 GHz
case style selection outline drawings see section 1
ZHL-42
FREQlENCY GHz
GAN, ~ I MAXM.M POWER, dBm I DYNAMC RANGE i
I I
VSWR I
I MOOO. Aotness. Oulput (qcll) ~ I t\f. dB
ntercept pt., dBm 3rd order
Typ. i h OUt I NO. ~- Max CorT'1)ression (no danc:Jge) Typ.
ZHL 42 ZHL-42 I 0.7-4 2 case~ 36 ZHL-4240 I O 7 -4 2
=, 0 1·
48
' j() =15
NOTES: 1 _ 0C€fatr,g terrperatu-e - 20' C to
-65'C Staage temperatu-e - -55'C to -;- 100°c
2 Witti ro load output. demte cnaxrrun rout ocwer (no dcmage) by10cfl
3. "nces md specifications Slbject to ctxr.ge without notice
FEATURES • up to1w, covers 700 MHz
to 4.2 GHz • ultra-linear Class A design
provides unconqitional stability
• any load impedance can be connected without. concern for damage or oscillation
• ideally suited for T ACAN, mobile rado. telemetry. anti-collision. and satellite CornrTU"'lications systems
•use to boost sigialjsweep generator output, achieve broadband isolation, and provide up to fOll test set ups with a fOll-way splitter
• exceptionally low price •one-year guarcr1tee • irrrnediate delivery
-·o
ZHL-42 FREQ.
~)
!CO 25 345.84 991 B5
1137 60 1283.50
!429 70 1575_50 !72090 1866.60 2013.10
2158 00 2304 20 2448.60 25%20 2741.40
2887.60 l:'m 40 3033 . .JO 3179.70 3324 .JO
:!470 . .JO 3616.70 3763 00 3909 00 J053 50 4199.90
c
ZHL-4240 700.00 875.00
1050.00 122500 1400.00
157500 1750 00 1925.00 2100.00 227500
2450.00 2625.00 2800.00 2975.00 3150.00
3325.00 350000 3675.00 3850.00 402500 4200.00
·o - oB 1 2 ~ 1 2 5. ~ l ! 2 ~ 1 2 s.1 I
G~. dB LMAAITY Forwcrd Reverse C~. Pout
3356 33 44 33 61 34 37 3454
3409 34 06 3465 35.05 35.42
35.21 34.72 34.42 34.55 3378
3403 34.25 3405 33 85 34.01
34.34 34.49 3465 32.36 33.32 34.60
45.20 45.38 44.65 4440 44.47
44.93 45.46 45.48 44.99 44.57
44.37 44.45 44.70 4488 45.64
46.25 45.92 45.19 44.85 45.18 46.71
47.42 47 61 47 66 47 97 47 96
48.08 48.00 48.03 48.15 48 .JO
4836 48.35 48.27 48.36 48.26
48.29 48.39 48.50 48.35 48.62
48.51 48.76 48.61 48.49 48.74 48.69
(dB) (dBm)
1 !8 '. 05 G 86 0.93 0.61
0.07 0. 16 J 80 ~ 41 175
1.40 1..JO 122 1.05 031
0.32 0.49 0.29 0.36 0.77
0.79 1.07 0.72 0.33 C.05 0.08
U8 1.05 0.86 0.93 0.61
0.07 0.80 175 1 .JO 1.22
105 0.31 0.32 0.49 0.36
0.77 0.79 107 0.33 0.05 0.08
28 91 28.82 2"1.29 2998 .J0.42
.J0.65 30.SJ 30.65 .J0.46 .JO 57
.JO.SJ 30.13 29.77 29.91 30.01
3035 .JO 38 .JO 50 .J0.27 .J0.04
30.19 29.91 29.73 2969 30 . .JO 30.04
2891 28.82 2929 29.98 .J0.42
3065 .J0.65 .J0.57 .J0.13 29.77
29.91 .J0.01 .J0.35 .J0.38 .J0.27
.J0.04
.J0.19 29.91 29.69 .JO.JO 30.04
DC POWER ' i PRICE $ I
I
Voll_ Cooent Ea. Qty.
·s
NOISE AGUlE (dB)
7 3 7 17 7 17 7 17 7 20
7 •7 7 ~Q 7 21 7 24 7 28
7 33 7 42 7 39 7 38 7.38
7 39 7 40 7 41 7 43 7.43
7 47 7 52 7 .sa 7 62 7 96
3.9 39 3.9 3.8 3.8
3.8 3.9 3.9 3.9 3.8
3.6 3.6 3.6 3.6 3.6
3.6 3,6 3.8 3.8 3.8 4.1
l 095 ( 1-9)
I 1395 ('-9)
VSWR n out
157 -198 154 1.67 I 51 1 42 '46 137 1.43 1.36
1.60 1.33 I 67 1.47 1.72 1.4J 1 62 1.32 157 1.20
1 47 1 21 148 1.16
-;~ ii~ 1.37 1.24
1.34 1.52 1.28 1.94 1.24 1.99 1.24 1.74 1.JO 1.56
'39 1.43 1.58 1.28 167 1.11 1.54 1.21 139 139 1.44 1.84
123 1.65 120 1.29 123 1.22 1 22 1.37 1.34 1.33
150 1.20 1 . .JO 1.20 1.26 1.22 123 122 1.40 118
1.49 1.20 1.64 1.41 1.59 1.76 1.52 1.90 1.32 1.98
128 1.73 137 1.32 1.41 1.24 1.55 1.48 1.4'1 1.55 1.66 1.96
c::JMini·Circuits Po. sox 166. aroo1<1vn. New v011< .,1235 (11a1934-4500
------,-----I 0 0.
r
cc
broadband linear
Power Amplifiers up to 1 Watt (+30 dBm)
700 MHz to 4.2 GHz
case style selection outline drawings see section 1
ZHL-42
FREQIBlCY GAJI4. cS MAXMJ\.1 POWER. dBm DYNM'IC RANGE VSWR <;Hz
htercept pt., d8m
MOOS.. Rotness 0uIput (qdB) ~ l\f.dB 3rd order i NO.j Min. Max C OfT'4')ression (no dcmoQe) Typ. Typ. h OUt
I 0./-42 Xl .=10 -21 ! -,a 1
2 5.' I ZHl-42 I -'0 '0 i 2,5 1 . -. " ! - - --ZHl-42 case U 36 ZHl-4240 I I cUJ -/..'.)1 L:).11 07-42
NOTES: ZHL-42 ,. Ooe;atr,g terrperatU'e - 20' C to FREQ, GA~. dB LMAA!TY
-65'C ~) Forward Reverse C~, Pout StOfoge terrperatU'e - '55'C to .;- 1OO'C (dB) (ciJm)
2 Witt', r{) lOad output. ooate cnaxlTun rcut ocwer (no damage) by1Cdl 100 25 3356 47.42 1 18 2891
3A5.84 3344 4761 'OS 28.82 3. "'rICes end specifications st.bject to 99185 3361 4766 086 21.29
chCT>ge with::lvt notice 113760 34 37 4797 0.93 2998 1283.50 3454 4796 0.61 30.42
1429 70 3409 aa.08 007 3065
FEATURES '575.50 3406 aa.oo 0.16 30.53 P20.90 3465 aa03 050 30.65 1866.60 .J.5.OS aa.15 ~ 41 30.<16
• Up to1w, covers 700 MHz 2013.10 3542 aa30 175 3057
2158 00 35.21 aa36 t.40 30.53
to 4.2 GHz 2304 20 34.72 aa . .J.5 130 30.13 2448.60 34.42 aa.27 122 29.77
• ultra-linear Class A design 2595.20 3455 aa.36 1.05 29.91 274140 3378 aa.26 0.31 30.01
provides unconc;litional 2887.60 34.03 aa.29 032 3035
stability 2999 40 34.25 aa.39 0.49 3038 3033.30 3405 48.50 0.29 3050 317970 1385 48.35 0.36 30.27
• any load impedance can 332430 34.01 aa.62 0.77 30.04
be connected without, Ja10.3O 34.34 48.51 0.79 30.19 361670 34.49 48.76 1.07 2991
concern for damage or 3763 00 3465 48.61 0.72 29.73 3909 00 32.36 aa.49 033 2969
oscillation cUJ5J 50 33.32 48.74 C.OS 30.30 4199.90 34.60 aa.69 0.08 30.04
• ideaRy suited for T ACAN. ZHL-4240 mobile radio. telemetry.
anti-colfision, and satellite 700.00 45.20 U8 2891 875.00 45.38 1.OS 2882
communications systems 1050.00 44.65 0.86 2929 122500 4440 0.93 29.98
• use to boost sigtal/sweep 1400.00 4447 0.61 30.42
157500 44.93 0.07 3065 generator outp..Jt, achieve 1750 00 45.46 0.50 30.65
1925.00 45.48 175 30.57 broadband isolation. and 2100.00 4499 130 30.13
227500 4457 1.22 2977 provide up to fOLr test
2450.00 44.37 105 29.91
set ups with a fOLr-way 2625.00 44.45 0.31 30.01 2800.00 44.70 0.32 30.35
splitter 297500 4488 0.49 30.38 3150.00 45.1>4 0.36 30.27
• exceptionaRy low price 3325.00 46.25 0.77 30.04 .J.5OO 00 45.92 0.79 30.19
• one-year guarO'"ltee 3675.00 45.19 107 29.91 3B50.00 44.85 0.33 2969
• immediate deHvery 4025.00 45.18 0.05 3030 4200.00 <16.71 a.os 30.04
48
DC POWER
Vail. ClJTent
'5 :' :)QA, ,- .
NOISE AGl.&lE (dB)
, 3 717 717 717 720
7 '7 7 ~9 721 724 728
733 742 739 738 7.38
739 740 741 743 7.43
747 752 758 762 796
3.9 39 3.9 J.8 3.8
3.8 3.9 3.9 3.9 3.8
3.6 3.6 3.6 3.6 3.6
3.6 3.6 3.8 3.8 J.8 4.1
r;;::1Mini .. Circuits PO BOX 166, Broo!<lyn. New YOI'k ., 1235 (718) 934-4500
~--.---' __ ....... __ --r .---..- .. - ~.---
--_ .. --I a 0.
r
! PRICE $
Ea. Qty.
i <395 ('·9) I , c;.:
VSWR n auf
~ 57 '198 154 1.67 151 142 '<16 137 1.43 136
1.60 U3 167 1.47 1.72 14.3 • 62 132 157 1.20
147 121 148 116
.1<19 1.15 1.51 1.18 1.37 1.2&
134 1.52 128 1.94 1.2& 1.99 1.24 !.7& 130 1.56
139 1.43 1.58 1.28 167 111 1.54 1.21 139 139 1.44 1.84
123 1.65 120 1.29 123 1.22 122 137 1.34 1.33
150 1.20 1.30 1.20 1.26 1.22 1.23 122 1.40 1.18
1.49 1.20 1.1>4 1.41 1.59 1.76 1.52 1.90 1.32 1.98
1.28 1.73 1.37 132 1.41 1.24 1.55 1.48 141 1.56 1.66 1.'16
Univers
ity of
Cap
e Town
APPENDIX K
LINE IMPEDANCES FOR SINGLE ENDED MIXER
199
APPENDIX K
LINE IMPEDANCES FOR SINGLE ENDED MIXER
199
Univers
ity of
Cap
e Town
The figure below indicates a microstrip single ended mixer with
RF choke and RF bypass. The RF choke is the shorted quarter
wavelength line. It is used to provide a path for the IF
rectified current, while being a high impedance to RF signals.
The RF bypass is the low impedance, open circuit patch on the IF
output. It ensures that the RF signals are applied across the
diode, and that the IF component is sent to the IF output port.
The following section proves these principles.
RF
HIXER DIODE A
1'4···--···?'/...~---··--t
T . Z01
1Y4 . _______ .t
F Output I I I !
'··~~~~~~~~~~~~~~~~~~~___J
The RF bypass
. . At a z 1-:n. = .Z.a *Zi + j*Z0 *tan ~l Za + j*Zi*tan JU
now : j3 = 2*lt T
For the case of the RF signals 1 = A/4
then : Z.t..:n. = £~i*cosn/2 + j*Z0 *sinn/2 Z0 *cosn/2 + j*Zi*sinn/2
200
below a single mixer with
RF and RF The RF the quarter
line. It is used to a path the IF
current, while being a impedance to RF
R.F bypass low impedance, circuit patch on the IF
• It ensures the RF are across the
........... 'IJU,'<# , and that IF component is sent to the IF port.
At a
now :
MIXER DIODE A
RF&lO "------7
F Output
For case of the RF signals 1 = A/4
then : Z1...n. =
200
Univers
ity of
Cap
e Town
The load is a short circuit, thus Z1 =0 and hence Zi~ = 502 /0--m
Therefore at RF frequencies, the RF bypass acts as a high impedance. The stub will now be analysed at IF frequencies.
The wavelength at 6 GHz is ~ 1 0.05 m (RF signal) at 10.7 MHz ~2 = 28.04 m (IF component)
As can be seen the at 10.7 MHz the short circuit stub is not /4
long, but rather A* 28.04 = A*560.8 long. 4*0.05 4
Now at 6 GHz A/4 = 0.05/4 = 0.0125 m
The length of the stub at 10.7 .MHz, in terms of wavelengths, is
thus: 0.0125*A = 4.458*10-4 *A 28.04
Thus the impedance of the stub at 10.7 MHz is:
Zi~ = ! 0 *Z 1 *cos(2*n*4.458*10-4 ) + j*Z0*sin(2.8*10~ 3 ) Z0 *cos(2.8*10-3 ) + j*Z1 *sin(2.8*10-3 }
= Zo*Zi*0.999 + j*Za*2.8*10- 3
Z0 *0.999 + j*Z1 *2.8*10- 3
letting Z0 = sog , Z1 = 0
Thus, at an IF frequency of 10.7 MHz, the stub has a low impedance of o.14g .
The stub thus provides a DC return for the rectified RF current.
The RF bypass Similiar to the previous analysis, the impedance of the patch to the RF and IF signals can be calculated and the operation as a RF
bypass confirmed.
The impedance looking in at b can be calculated as follows. The
201
a , thus Z.l.=O
Therefore at RF frequencies, the RF bypass acts
/0-00
as a high
. The at IF The wavelength at 6 GHz /-.;1 = 0.05 m (RF
at 10.7 MHz 28.04 m (IF cOlillDonerlt )
As can be seen the at 10.7 MHz the short circuit is not /4 long, but :::::
Now at 6 GHz A/4 = 0.05/4
The length the stub at 10.7 MHz, in terms wavelengths, is
., .,
letting
:::::
z * ""
= = 0
1= 0.14U
Thus, at an IF frequency impedance of 0.14U .
stub at 10.7 MHz is:
10.7 MHz, stub has a low
The stub thus provides a DC return the rectifiedRF current.
the , the the patch to
the RF signals can calculated and the operation as a RF
at b can be as
201
Univers
ity of
Cap
e Town
impedance at c is an open circuit, the characteristic impedance of the patch Z01 is low since the patch is wide. The impedance at b is given by:
Thus there is very little impedance to RF signals through this
patch. This patch thus also provides isolation of the RF signals from the IF output. An important quantity in single ended mixer
design. The higher the isolation, the better.
At 10.7 MHz the A(6 GHz)/4 section is 4.458*10-4 *A(l0.7 MHz). The impedance looking in at b is then given by:
Zi~(at b) = Z01*Z 1 *0.999 + j*Z01*2.8*10-3
Z01*0.999 + j*Z1*2.8*10- 3
letting X = 0.999 and Y = 2.8*10- 3 , and inserting into the
equation above, gives.
Zi~(at b) = koi*Zi*X + j*Z0 1*Y Z01*X + j*Z 1 *Y
Using L'hospitals rule:
Therefore
= lim Zoi*~i*X + j*Zoi*Y.) Z.i--m (Z01*X + j*Z 1 *Y)
2-= lim Z01 * 0Z1 ( Zi *X + j*Z01 *Y)
Zi-m _.£_ (Z01 *X + j*Zi*Y) 02.1
Zi~(at b) = lim koi*X Zr-m j*Y
= Z'° 1 *356.8 j
[zi~< at 10. 7 MHz >I = lzoi *356. a I This is still a reasonably high resistance, eventhough Z01 is a
low characteristic impedance. The IF signals will travel down the low impedance IF output line, rather than through the open
circuit quarter wavelength patch.
202
an open circuit, characteristic wide. The impedance at
b given by: /Z1--..... 0
there is very impedance to RF signals through this
IF output. An important
• The higher isolation,
At 10.7 MHz the A(6 GHz)/4 section looking at b is then
of RF signals enaea mixer
4.458*10-4*A(10.7 KHz). The by:
x = 0.999 Y = 2.8*10-3 , and inserting
,
L'·hospitals . .
== lim Zc:>1
= Ze>1*356.8 j
IZ1-n(at 10.7 MHz)1 =
a
characteristic impedance IF
high ~oeaance. The IF
line, quarter wavelength patch.
202
.81
, will
than
a down the
open
Univers
ity of
Cap
e Town
APPENDIX L
SCEMATIC OF COUNTER CIRCUIT
AND ANZAC MD 162 BALANCED MIXER CHARACTERISTICS
203 203
Univers
ity of
Cap
e Town
f\)
0 _p.·
12 Vdco-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~--.~~~~~~~~~~~~~~--11>--~~~~~-o
level shifter
R'S llk Ohms 4 70k Ohms
18k <>hms
1 uF
Di 1N414S
Hysteretic COl!1f>arator
R4 470K Ohms
100k Ohms
Schmitt trigger
O Vdc~~~~~~~~-J1--~~~~~~~~~~~~+-~__,l-~~~~~~~~~~~~~~~~
level shifter and co,,.,arator unit
•ocumo.nl: Number figure 1
Januar- L- 1990 eet of'
o ~.
.. III 0 12 VdcO>----------------------------------------------------------,-------------------------~------------
o '\IdcO ill
Illvel shiftvr
it
R'5 11k Ohms 470k Ohms
r
18k ()hms
1 I.IF R4 1)1 4701( Ohms 1 ... 4148
iii ill 0
level shifter and cOmParator unit
'oc::ume-nt-NUmbll!!r figure: 1
Januar r:;'-I99"O 1Sh-•• f: of'
Univers
ity of
Cap
e Town
·'
N 0 01
:-J
?
~i I i i I . Ii It-re---.~
HAN 3640
, I • I • I OVdd
counter unit and display
oc:ullll!'nt Number figure 2
Januar r.--..c9ao ISheet
1
o<
" N
:J
H~M 3640
counter unit and display
oc::ulOent Humber figure 2
JaMuer I. 198IHihe@'t of'
Univers
ity of
Cap
e Town
j
\i' J;..· MC14553B
OPERATING CHARACTERISTICS
The MC145538 three-digit counter, shown in Figure 3, consists of three neptiwe edge-trigger~ BCD counter1 which•• cuc1dltd in 1 synchronous f1sttion. A quid l1tch at the output of each of the thrtt BCD counters permits storage of 1ny given count. The three sets of BCD outputs (active high). after goine through the l1tches. are time division multiplexed, providing one BCD number or digit at 1 time. Otgit select outputs (.:tive low, ire provided for display control. All outputs •re TTL compatible.
An on ·chip oscillator provides the low frequency scanning clock which drives the multiplexer output selector. The frequency of the oscillator can be controlled externally by a capacitor betvween pins 3 and 4, or it un be overridden and driven with an external clock at pin 4. Multiple devices can be cascaded using the overflow output, which provides one putse for everv 1000 counts.
The Master Reset input, when taken high, initializes the three BCD counters and the mulliptexer scanning circuit. While Master Reset is high the digit scanner is set to digit one; but all three digit select outputs are disabled to, prolong display life, and the scan oscillator is inhibtted. The Disable input, when high, prevents the input clock from reaching the counters. while shll retaining the last count. A puhe shaping circuit at the clock input permits the counters to continue operating on input pul~ with very slow rise times. Information present in the counters when the l•tch input goes hiQh, will be stored in the latches and will be retained while the latch input is high, independent of other inputs. lnfo,mation c•n be recovefed from
the latches after the countms have been reset if latch Enable remains high during the entire reaet cycle.
FtOUlll! I - .... ANDl!D •LOCK DtAOllAll
Cloe:1t 12
13 .. ~ (Ae:tl-Hllf'J
Owtftoww
6-402
CLIDI Dl9't $elect CMIOI
CActl- t.ow)
~00
~QI
~02
~03
BCD Output1 (ActlHi9hl
5 .. ~
2 M ill I . .. a:
• ::>
~
'''
II I I
ii D 0: ..
~ .... u u
ci I~~ I c !!l:i lo -
~ :!
" ~Ii:::
ll
~ .. .. - ..
s~ 0
Ld_tl ~ ~ -.: "
c !?1:1
"
MC14553B
.. .. ~ . ;:; ll
0 0 >
.. ~·--' . u
!§ 0 ~ .. 0 ..
... ~ 0 Ct
!!r;r
ij ..,_
6-403
I I· I I I I I I I~~ : E ~ r ~ i I c ~ E
~-------~ • 0 . -- v >
I - -...,· I i i I I I I I I I Q =
0 ~
I ~ :! u ll
II
1-t" .',:
[\)
0 ()l
•
IIIIC14S538
OPt!RATING CHARACTERISTICS
The MC145538 three-digit counter, shown in figure 3. consists of three I"1etltiw. ed9!t-trigger6d BCD counters wt1ich If. cascaded in .. synchronous famion. A qUid tatch at the output of each of the thrft aCD Counters permits stOfage of Iny given Count. The three sen of BCD outputs (ICtive high), Ifter gOi"G through the Iltch"_ .re time diwj,jon multipieJII:td, providing one BCD number Of digit It I lime. Dlgit sefect outputs (active low) are provided for dis.piay conuol. All OUtputs ,ne TTL comparibte,
An on chip oscillator provides the low frequency scanning clock which drives the multiplexer output sefector. The frequency of the oscillator can be controlfed ex. ternally by a capacitor bet'N!en pins J and 4, or it can be ovt!'uu1rlen iJnd driven with an external clock at pin 4. Multiple de .... ice\ Coin be cascaded using the overflow output, which provides one puke for every 1000 counts.
The Mas1er Reset input, when taken high, inittalile5 the three 8CD COUnters and the muHipfe.-:er scanning circuit. While Master Reset is high the digit scanner is set to digit one; but all three digit select outputs are disabled to. prolong display life, and the scan oscillator is inhibited. The Disable input, when high, prevents the input clock from reaching the counters, while shU retaining the last count. A pulse shl~ng circuit at the clock input permits the counters to continue operating on input pulses with very slow rise times. Information present in the counters when the latch input goes hiQh. will be stored in the latches and will be 'etained while the latch input ill high, in.
dependent of o1her inputs. Infofmation can be fecovered from
the latches ahm the countms have been feset if latch Enable femein. high during the enti,e reset cycle.
"QURII J - EXPANDeD IILOCK DiAQIIIIAMI
ClOd,
12
O'Mb'. (Act·woe HiGh I
I I
I I
.. "A
(Aetl_Hipl
00
O.
a. ~ 10 03 Unit.
c 00
O.
02 R +- 10 03
ren,
C DO
o. a. A +10 03 Hund,edS
O .. ,fteo"""
6-402
h.IDI DIIIII Select eMIDI CAulwoet.owi
~ao
~o,
~02
~03
BC~
Outputs (Active Hiettl
5 .. • Q
! ill I .. ..
IE ,:I
" Ii:
, , ,
" I I
Ii ~ " ..
~ . .. u u 0
g
!;!pl
~ ! "
gl~ :I
5 ~
~
i"i .;
Ld.! . ..
IIIIC145538
'~b liS -
~~ " .. ai
.-J~L-!:::--'-~~I} ~ ~
~ f ~ i ~ . ~ E
~--=-----'I • e • v : . H a::
a ~
_ _ L
U U o I~ g
~I:I
.. ~I....I
!!
il u_
.. ::I : u :I
I~~ ra~ g}<a o,~
SI ..
~ u :II
alo I II I I I I
6-403
II
Univers
ity of
Cap
e Town
Guaranteed Specifications* (From -55:)C to +85:iC)
Frequency Range: RF, LO Ports IF Port
Conversion Loss:** 1.5·5.5 GHz 1-7 GHz
Isolation: LO to RF (1-3 GHz)
(3-7 GHz) LOtolF (1-3GHz)
(3-7 GHz) RF to IF (1-3 GHz)
(3-7 GHz)
1-7GHz 10.2000 MHz
7.5 dB Max 8.5 dB Max
15 dB Min 17dBMin 20dB Min 13 dB Min 17 dB Min 12dB Min
·' Operating Characteristics Impedance: Maximum Input:
Total Power
RF Input: 1 dB Compression 1 dB Desensitization
SSS Noise Figure:
50 Ohms Nominal
3SOmW Max@ 25°C Oerated 3.5 mWfC
+8 dBm Typical +6 dBm Typical
Within 1 dB of Conversion Loss Max
3rd Order Input Intercept: 2.0 GHz +16.5 dBm Typical 7.0 GHz +18.0 dBm Typical
3rd Order Intercept Degradation: 1.5 dB Typical@
IF Termination VSWR 3:1 • All specifications apply when operated at +13 dBm
available LO power with 50 ohm source and load impedance.
• • For IF frequencies of 10-500 MHz and an AF of -10dBm or lower.
t Independent of sum frequency match.
This product contains elements protected by United States · Patent Number 4,224,572.
TERMINATION· INSENSITJVE MJCROWAVE MIXER 1-7 GHz • Intermodulation Ratio Insensitive to IF
Port Mismatches·· • 6 dB Typical Midband Conversion Loss
1-,.. w (,J a: w
10
4
27
Typical Performance**
CONVERSION LOSS VS IF PORT TERMINATIONt
I ~--
~~ ~~ ... 3:~ VSWR ------- ~-- I SOfl ~
2 3 4 5
3rd ORDER INTERCEPT VS IF PORT TERMINATIONt
L0@-13dBm
8 1 8
~-22 a;~ ·~
50fl ..,,, - ----~:!! 17
a: 0 12 1
--. .. --2
- 3:1 VSWR --I L00+13d8m
3 4 5 7 • FREQUENCY IN GHz
Adams~ Russell ANZAC •.. the qualitative difference .\NZAC DI\ ISION
164
207
1 I .
Guaranteed Specifications* ( C to C)
LOtoRF (1-3GHz) (3-7
LO to IF (1-3 GHz)
RF to IF (1·3 GHz)
15 dB Min 17 dB Min 20 dB Min 13 dB Min 17 dB Min 12dBMin
.\
Operating Characteristics Impedance: Maximum Input:
Total Power
1 1 dB
sse Noise Figure:
50 Ohms Nominal
+8dBm +6dBm
Within 1 dB of
3rd Order Input Intercept: 2.0 GHz +16.5 dBm
1.5 dB Typical @
1 .. All specifications apply when operated at +13 dBm
available LO power with 50 ohm source and load impedance •
.. For IF frequencies of 10-500 MHz and an AF of -10 dSm or lower.
f Independent of sum frequency match.
This product contains elements protected by United States ' Patent Number 4.224.572.
TERMINATION· INS NSITIVE MICROWAV MIXER 1 GHz .. Intermodulation Ratio Insensitive to IF
Port Mismatches" .. 6 dB Typical Midband Conversion Loss
Typical Performance **
CONVERSION lOSS VS IF PORT TERMINATIONt
10~~~~~~~---.---.---r--,
3Td ORDER INTERCEPT VS If PORT TERMINATIONt
27~~~~=r~~~~--,----,----,
Adams~ Russell ANZAC ••• the qualitative difference ~NZ-\c DI\ ISION
164
1 t .
..
Univers
ity of
Cap
e Town
I
I
environmental ii unit has bee!l d~signed to meet or exceed the following
1'!t ·ronmental criteria: !!!!!--- Governing Document: MIL-STD·202E - Test Method Condition vr;u&l Inspection . Anza
0c ~ork0matn1 .ship0Manual Ji! hanical Inspection ev1ce u me wg.
tf;~metic Seal• · 112 D ShoCk 213 C Acceleration 212 A \''bration 204 D R~sistance to Soldering• 210 C 'fhrrmal Shock 107 A .Moisture Resistance 106 ~cific device te~t!ng to these and other environmental tests ii available at add1t1onal cost. • These tests apply to MD-162 only
= 35
~ 30
~ 25 ~ < 20 ... ~ 15
Typical Performance
5 10 50 100 FREQUENCY IN MHz
ISOLATION
2 3 4 5 6 FREQUENCY IN GHz
1.0
-1.0
500 1000 2000
'""
7 8
RF & Lo··1MPEDANCE_ .....
IF IMPEDANCE
I Schematic
Mechanical Data MDC-162
0.375 0. 190 (953/ 1 f (4.831 r ~~;~z~:CTOR .~CJ/PLACES
MOUNTING HOLE 0.116 DIA THRU (41 PLACES
,.
0.896 (22.761
UNLESS OTHERWISE NOTED •. XXX = ! 0.020 •t.X ., 0.51 WEIGHT (APPROX.I: 1.2 OUNCES 34.5 GRAMS 'DIMENSIONS IN ( I ARE IN MM FINISH: CHEMICAL CLEAR PER MIL.C.5541 8, CLASS 3.
MD-162
0.018 :!: 0.005 PIN DIA. ~(.461 (.131 JPLACES
0.06 RAD. TYP . .--------(1.521
0.740:: 0.020 1 ( 18.801 C.511 L L MO. 162 R
Lf x LO. 165 MIN T p (4.191 . y.
o 200 J 0'i.!~1 t- j IS.08IMAX~1. 100:: 0.020 0.095 ! 0.020 J__ 121.941 c.5.11 12.411 c.511
~~-·--E==i UNLESS OTHERWISE NOTED .. XXX •: 0.010 '(.X •: 0.031 WEIGHT !APPROX.I: 0.35 OUNCES 9 9 GRAMS "DIMElll~IONS IN 11 ARE IN MM
Fl~ISH· CASE AND LEADS -GOLD ELECTROPLATED PER MIL·G.45204 TYPE 3, CLASS 2
LEADS: WELDABLE ANO SOLDERABLE PER MIL.ST0-12768
Ordering Information
MODEL NO.
- MD-162 MDC-162
PART NO.
9149 9144
CONNECTORS
SMA
UNIT PRICE (1-9 UNITS)
$380 475
Delivery is from stock.
80 Carn bridge Street, Burlington, i\1ass. 01803 (617) 273-3333 lWX 710 • 332 • 0258
165
20B
Environmental . nit has been designed to meet or exceed the following
1/111 U I" 'ronmenta Criteria: fII<r1 IL-STD.202
112 213 212 204 210 107 106
C A D C A
these and other environmental tests is Iva at cost . • These tests apply to MD-162 only
Typical Performance PORT RESPONSE i 3nr-----r---r-------r---r------,--,r-~
~. 2~~r_+----r_+----~~ ... ILl ,. ; ~ oL---~~~~:=~~~~==~~~~_ a::
·1.0 RF '" LO·-IMPEDANtE~·"· IF IMPEDANCE
I Schematic
Mechanical Data MDC·162
" UNLESS OTHERWISE NOTED •• xxx· ! 0.020 '!.X., O.Sl WEIGHT tAPPROX.I: 1.2 OONCES 34.5 GRAMS 'OIMENSICNS IN I) ARE IN MM fiNISH: CHEMICAL CLEAR PER MIL·C·5541 B. CLASS 3.
0.140 = 118801
MD·162
:: 0.005 PIN DIA. L III 3 PLACES
TVP.
UNLESS OTHERWISE NOTED. ,XXX, 0010 'c.x·! 0.031 WeiGHT IA"PROX.L 0.35 OUNCES 9 9 GRAMS 'DIMEN~IONS IN II ARE IN MM FINISH
MODEL NO. PART NO.
- MD-162 MDC-162
9149 9144
Information
CONNECTORS
SMA
UNIT PRICE
$380 415
Delivery is flam Slock.
80 Can1bridge St Burlington, Mass. 01803 (617) lWX 710·
165