doctoral thesis - tesis doctoral contribution to...
TRANSCRIPT
UNIVERSIDAD POLITÉCNICA DE MADRID
ESCUELA TÉCNICA SUPERIORDE INGENIEROS DE TELECOMUNICACIÓN
DOCTORAL THESIS - TESIS DOCTORAL
CONTRIBUTION TO ACTIVE ARRAY
ANTENNAS AT MICROWAVE BANDS
CONTRIBUCIÓN A LOS ARRAYS DEANTENAS ACTIVOS EN MICROONDAS
Gonzalo Expósito DomínguezIngeniero de Telecomunicación
2013
UNIVERSIDAD POLITÉCNICA DE MADRID
ESCUELA TÉCNICA SUPERIOR DE INGENIEROS DE TELECOMUNICACIÓN
DEPARTAMENTO DE SEÑALES, SISTEMAS Y RADIOCOMUNICACIONES
GRUPO DE RADIACIÓN
DOCTORAL THESIS - TESIS DOCTORAL
CONTRIBUTION TO ACTIVE ARRAY
ANTENNAS AT MICROWAVE BANDS
CONTRIBUCIÓN A LOS ARRAYS DEANTENAS ACTIVOS EN MICROONDAS
Autor :
Gonzalo Expósito Domínguez
Ingeniero de Telecomunicación
Directores:
Manuel Sierra-Castañer
Doctor Ingeniero de Telecomunicación
Profesor Titular de Universidad
José-Manuel Fernández-González
Doctor Ingeniero de Telecomunicación
Contratado doctor
Madrid, 2013
TESIS DOCTORAL: Contribution to active array antennas at microwave bands.
Contribución a los arrays de antenas activos en microondas.
AUTOR: Gonzalo Expósito Domínguez
Ingeniero de Telecomunicación
DIRECTORES: Manuel Sierra-Castañer
Doctor Ingeniero de Telecomunicación
Profesor Titular de Universidad
José-Manuel Fernández-González
Doctor Ingeniero de Telecomunicación
Contratado doctor
DEPARTAMENTO: Señales, Sistemas y Radiocomunicaciones
Univerisdad Politécnica de Madrid
El Tribunal de Calicación, compuesto por:
PRESIDENTE:
Prof. Dr.
VOCALES:
Prof. Dr.
Prof. Dr.
Prof. Dr.
VOCAL SECRETARIO:
Prof. Dr.
VOCALES SUPLENTES:
Prof. Dr.
Prof. Dr.
Realizado el acto de defensa y lectura de la Tesis, acuerda otorgarle la CALIFI-
CACIÓN de:
Madrid, a de de 2013.
A mi sobrino Martín
"Serenidad para aceptar las cosas que no puedo cambiar,
valor para cambiar las cosas que puedo
y sabiduría para poder diferenciarlas"
REINHOLD NIEBUHR
Resumen
Esta tesis que tiene por título Contribución a los arrays de antenas activos en
banda X, ha sido desarrollada por el estudiante de doctorado Gonzalo Expósito
Domínguez, ingeniero de telecomunicación en el Grupo de Radiación del Departa-
mento de Señales, Sistemas y Radiocomunicaciones de la ETSI de Telecomunicación
de la Universidad Politécnica de Madrid bajo la dirección de los doctores Manuel
Sierra Castañer y José Manuel Fernández González.
Esta tesis contiene un profundo estudio del arte en materia de antenas activas en
el campo de apuntamiento electrónico. Este estudio comprende desde los fundamen-
tos de este tipo de antenas, problemas de operación y limitaciones hasta los sistemas
actuales más avanzados. En ella se identican las partes críticas en el diseño y pos-
teriormente se llevan a la práctica con el diseño, simulación y construcción de un
subarray de una antena integrada en el fuselaje de un avión para comunicaciones
multimedia por satélite que funciona en banda X. El prototipo consta de una red
de distribución multihaz de banda ancha y una antena planar.
El objetivo de esta tesis es el de aplicar nuevas técnicas al diseño de antenas
de apuntamiento electrónico. Es por eso que las contribuciones originales son la
aplicación de barreras electromagnéticas entre elementos radiantes para reducir los
acoplamientos mutuos en arrays de exploración electrónica y el diseño de redes des-
fasadoras sencillas en las que no son necesarios complejos desfasadores para antenas
multihaz.
Hasta la fecha, las barreras electromagnéticas, Electronic Band Gap (EBG),
se construyen en sustratos de permitividad alta con el n de aumentar el espacio
disponible entre elementos radiantes y reducir el tamaño de estas estructuras. Sin
embargo, la utilización de sustratos de alta permitividad aumenta la propagación
por ondas de supercie y con ellas el acoplo mutuo. Utilizando sustratos multicapa
y colocando la via de las estructuras en su borde, en vez de en su centro, se consigue
reducir el tamaño sin necesidad de usar sustratos de alta permitividad, reducir
i
ii
la eciencia de radiación de la antena o aumentar la propagación por ondas de
supercie.
La última parte de la tesis se dedica a las redes conmutadoras y desfasadoras
para antenas multihaz. El diseño de las redes de distribución para antenas son una
parte crítica ya que se comportan como un atenuador a la entrada de la cadena
receptora, modicando en gran medida la gura de ruido del sistema. Las pérdidas
de un desfasador digital varían con el desfase introducido, por ese motivo es necesario
caracterizar y calibrar los dispositivos correctamente. Los trabajos presentados en
este manuscrito constan de un desfasador reectivo con un conmutador doble serie
paralelo para igualar las pérdidas de inserción en los dos estados y también un
conmutador de una entrada y dos salidas cuyos puertos están adaptados en todo
momento independientemente del camino del conmutador para evitar las reexiones
y fugas entre redes o elementos radiantes.
El tomo naliza con un resumen de las publicaciones en revistas cientícas y
ponencias en congresos, nacionales e internacionales, el marco de trabajo en el que se
ha desarrollado, las colaboraciones que se han realizado y las líneas de investigación
futuras.
Abstract
This thesis was carried out in the Radiation Group of the Signals, Systems and
Radiocomunications department of ETSI de Telecomunicación from Technical Uni-
versity of Madrid. Its title is Contribution to active array antennas at X band and
it is developed by Gonzalo Expósito Domínguez, Electrical Engineer MsC. under the
supervision of Prof. Dr. Manuel Sierra Castañer and Dr. José Manuel Fernández
González.
This thesis is focused on active antennas, specically multibeam and electronic
steering antenas. In the rst part of the thesis a thorough description of the state
of the art is presented. This study compiles the fundamentals of this antennas,
operation problems and limits, up to the breakthrough applications. The critical
design problems are described to use them eventually in the design, simulation and
prototyping of an airborne steering array antenna for satellite communication at X
band.
The main objective of this thesis is to apply new techniques to the design of
electronically steering antennas. Therefore the new original contributions are the
application of Electromagnetic Band Gap materials (EBG) between radiating ele-
ments to reduce the mutual coupling when phase shift between elements exist and
phase shifting networks where special characteristics are required.
So far, the EBG structures have been constructed with high permitivity substra-
tes in order to increase the available space between radiating elements and reduce
the size of the structures. However, the surface wave propagation modes are enhan-
ced and therefore the mutual coupling increases when high permitivity substrates
are used. By using multilayered substrates and edge location via, the size is redu-
ced meanwhile low permitivity substrates are used without reducing the radiation
eciency or enhancing the surface propagation modes.
The last part of the thesis is focused on the phase shifting distribution networks
for multibeam antennas. This is a critical part in the antenna design because the
iii
iv
insertion loss in the distribution network behaves as an attenuator located in the rst
place in a receiver chain. The insertion loss will aect directly to the receiver noise
gure and the insertion loss in a phase shifter vary with the phase shift. Therefore
the devices must be well characterized and calibrated in order to obtain a properly
operation. The work developed in this thesis are a reective phase shifter with a
series-shunt switch in order to make symmetrical the insertion loss for the two states
and a complex Single Pole Double Through (SPDT) with matched ports in order to
reduce the reections and leakage between feeding networks and radiating elements.
The end of this Ph D. dissertation concludes with a summary of the publications
in national and international conferences and scientic journals, the collaborations
carried out along the thesis and the future research lines.
Contents
Resumen i
Abstract iii
List of Figures x
List of Tables xi
1. Introduction 1
1.1. Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2. Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.3. Outline of the thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2. State of the art on multibeam and recongurable steering array
antennas 23
2.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.2. Multibeam passive antennas . . . . . . . . . . . . . . . . . . . . . . . 33
2.2.1. Wide band Butler matrix network at X band . . . . . . . . . . 35
2.3. Scanning active antennas . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.3.1. Airborne steering antenna . . . . . . . . . . . . . . . . . . . . 39
2.4. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
3. Mutual coupling reduction using EBG in steering antennas 57
3.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
3.2. EBG theory fundamentals . . . . . . . . . . . . . . . . . . . . . . . . 60
3.3. Surface wave supression . . . . . . . . . . . . . . . . . . . . . . . . . 63
3.4. Mutual coupling reduction . . . . . . . . . . . . . . . . . . . . . . . . 66
3.5. Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
v
vi CONTENTS
3.6. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
4. Phase shifting distribution networks for switchable beam antennas 85
4.1. Fundamentals of phase shifters . . . . . . . . . . . . . . . . . . . . . . 85
4.2. Topologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 88
4.3. Switchable beam antenna . . . . . . . . . . . . . . . . . . . . . . . . 92
4.3.1. Symmetric reective phase shifter . . . . . . . . . . . . . . . . 95
4.3.2. Matched SPDT switch . . . . . . . . . . . . . . . . . . . . . . 98
4.4. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
5. Conclusions 113
5.1. Original contributions . . . . . . . . . . . . . . . . . . . . . . . . . . 114
5.2. Future research lines . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
5.3. Publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
5.3.1. Journal publications . . . . . . . . . . . . . . . . . . . . . . . 115
5.3.2. Conference contributions . . . . . . . . . . . . . . . . . . . . . 116
List of Figures
1.1. Astronomic interferometry clusters [1.6]. . . . . . . . . . . . . . . . . 3
1.2. Classication of metamaterials [1.17]. . . . . . . . . . . . . . . . . . . 6
1.3. Schematic view of a Field Eect Transtistor. . . . . . . . . . . . . . . 9
1.4. Micro ElectroMechanical Systems schematic view. . . . . . . . . . . . 10
1.5. PIN diode and equivalent circuit. . . . . . . . . . . . . . . . . . . . . 11
1.6. Ferrite core phase shifter. . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.7. Varactor diode outline. . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.1. Array splitting problem. . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.2. Wavefront combination to get electronic steering. . . . . . . . . . . . 25
2.3. Visible margin of an array in ψ and θ. . . . . . . . . . . . . . . . . . 26
2.4. HPBW vs scanning angle. . . . . . . . . . . . . . . . . . . . . . . . . 30
2.5. Directivity reduction vs scanning angle. . . . . . . . . . . . . . . . . . 31
2.6. Crossing point losses vs number of beams. . . . . . . . . . . . . . . . 31
2.7. Quantization noise oor vs number of bits. . . . . . . . . . . . . . . . 32
2.8. Appearance of grating lobes due to failures in the phase feeding law. . 33
2.9. Wide band Butler Matrix network . . . . . . . . . . . . . . . . . . . . 37
2.10. Phase of the scattering parameters of a wideband Butler matrix net-
work. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
2.11. Airborne steering antenna for satellite communications [2.40]. . . . . 39
2.12. Double stacked patch with 3dB/90o hybrid coupler. . . . . . . . . . . 41
2.13. Block diagram system. . . . . . . . . . . . . . . . . . . . . . . . . . . 41
2.14. Radiation pattern of 16x1 array. . . . . . . . . . . . . . . . . . . . . . 42
2.15. Amplitude and phase of the active network elements. . . . . . . . . . 43
2.16. 1 to 4 unbalanced network divider. . . . . . . . . . . . . . . . . . . . 43
2.17. S parameters of 1 to 4 unbalance network divider. . . . . . . . . . . . 44
2.18. 3 dB/90o Hybrid coupler miniaturization. . . . . . . . . . . . . . . . . 45
2.19. Hybrid coupler S parameters. . . . . . . . . . . . . . . . . . . . . . . 46
vii
viii LIST OF FIGURES
2.20. Radiation pattern of azimuth plane. . . . . . . . . . . . . . . . . . . . 46
2.21. 4x4 array array prototype [2.40]. . . . . . . . . . . . . . . . . . . . . . 47
2.22. S parameters measurements of 4x4 array prototype (8 ports). . . . . . 48
2.23. Steering radiation pattern of 4x4 subarray. . . . . . . . . . . . . . . . 48
3.1. High impedance surface and its model with parallel resonant LC cir-
cuit. The substrate is transparent in order to get better visualization
of metallic vias. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
3.2. LC model for the mushroom like EBG structure. . . . . . . . . . . . . 61
3.3. Layer and top views of traditional EBG structures (left hand side)
and multilayered F structure (right hand side). . . . . . . . . . . . . . 63
3.4. Unit cell scheme for eigenmode solutions. Dimensions in mm. . . . 64
3.5. Brillouin diagrams of the EBG unit cell. . . . . . . . . . . . . . . . . 65
3.6. Simulation scheme for transmission parameters S21 analysis. . . . . . 66
3.7. Parametric study of isolation characteristics (S21 for original shape
mushrooms) when size and number of elements are swept. . . . . . . 67
3.8. Parametric study of isolation characteristics (S21 for original shape
mushrooms) when substrate thickness, via diameter and gap size are
sweept. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
3.9. Parametric study of width W and length L for F shape mushrooms.
Transmission S parameters (S21). . . . . . . . . . . . . . . . . . . . . 69
3.10. Samples of single and multilayered EBG mushrooms with dierent
shapes and number of elements. . . . . . . . . . . . . . . . . . . . . . 69
3.11. Comparison between measurements and simulations of transmission
parameters S21 for dierent types of EBG mushrooms. . . . . . . . . 70
3.12. Multilayered mushroom with rectangular shape, 4 elements and edge-
located via (F-shape). . . . . . . . . . . . . . . . . . . . . . . . . . . 71
3.13. |E| eld simulation of two round patches with dual circular polarization. 72
3.14. 2x1 test array of circular patch antennas with and without EBG F
shape mushrooms. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
3.15. Specic dimensions of 2x1 test array. . . . . . . . . . . . . . . . . . . 73
LIST OF FIGURES ix
3.16. Comparison of measurements and simulated S parameters for 2x1 array. 73
3.17. Radiation patterns of 2x1 test patch antenna array with and without
EBG structures. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
3.18. Layer view of the 4x4 array with EBG structures construction. . . . . 76
3.19. S parameters measurements of 4x4 array with EBG structures. . . . . 77
3.20. 4x4 array and Butler matrix network connection for LHCP congu-
ration. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
3.21. 4x4 steering array radiation pattern for RHCP, at the center frequency
(7.825GHz). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
3.22. 4x4 steering array axial ratio for frequencies 7.25 GHz and 8.4 GHz,
RH and LH circular polarizations over the scanning angles. . . . . . . 79
3.23. 4x4 steering array axial ratio for RH and LH circular polarizations
and dierent pointing directions over the working frequency. . . . . . 79
4.1. Single switches. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
4.2. Compound switches. . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
4.3. Tuned λ/4 switches. . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
4.4. Quadrature matched hybrid switches. . . . . . . . . . . . . . . . . . . 91
4.5. Switched line phase shifter. . . . . . . . . . . . . . . . . . . . . . . . . 92
4.6. Uniform radiation pattern beam shape. . . . . . . . . . . . . . . . . . 93
4.7. Multibeam radiation pattern unbalance description. . . . . . . . . . . 94
4.8. Multibeam radiation pattern isolation description. . . . . . . . . . . . 94
4.9. Reective phase shifter construction. . . . . . . . . . . . . . . . . . . 95
4.10. Reective phase shifter measurements. . . . . . . . . . . . . . . . . . 97
4.11. Insertion loss variation due to the PIN diode parameters model. . . . 98
4.12. Reective congurations. . . . . . . . . . . . . . . . . . . . . . . . . . 98
4.13. Series-shunt reective phase shifter results. . . . . . . . . . . . . . . . 99
4.14. Single and Double PIN diode based SPDT switch. . . . . . . . . . . . 99
4.15. Graphic equivalent of the double SPDT operation. . . . . . . . . . . . 100
4.16. Comparison of the features of the single and double SPDT. . . . . . . 101
4.17. Double PIN diode SPDT switch Photograph. . . . . . . . . . . . . . . 101
x LIST OF FIGURES
4.18. Double PIN diode SPDT switch measurements. . . . . . . . . . . . . 102
4.19. SPDT full measurements at 24 GHz. . . . . . . . . . . . . . . . . . . 102
4.20. SPDT switch measurements over the temperature. . . . . . . . . . . . 104
4.21. Coupling between networks due to the open circuit of the isolated
path of the conventional SPDT switch. . . . . . . . . . . . . . . . . . 104
4.22. Complex SPDT switch with matched output ports. . . . . . . . . . . 105
4.23. Insertion loss, isolation and return loss of the complex SPDT switch
with matched output ports. . . . . . . . . . . . . . . . . . . . . . . . 107
List of Tables
1.1. Characteristics summary of the digital switches. . . . . . . . . . . . . 12
1.2. Characteristics summary of the analog phase shifters [1.31]. . . . . . . 14
2.1. Gain reduction and angle widening for scanning angles. . . . . . . . . 30
2.2. Characteristics summary of the passive electronic steerable techniques. 36
2.3. Amplitude coecients of the passive network. . . . . . . . . . . . . . 44
3.1. Comparison of PEC, PMC, and EBG ground planes for low prole
antenna designs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3.2. Antenna specications. . . . . . . . . . . . . . . . . . . . . . . . . . 75
4.1. Summary of formulas for SPST Switches. . . . . . . . . . . . . . . . . 90
4.2. Reective phase shifter [S] matrix. . . . . . . . . . . . . . . . . . . . . 96
4.3. Complex SPDT [S] matrix. . . . . . . . . . . . . . . . . . . . . . . . . 106
xi
Chapter 1
Introduction
Since the discovery of electromagnetic wave propagation in 1888 by Hertz, an-
tennas have been used to transmit information from one point to another through
the free space. The IEEE Standards Denitions of Terms for Antennas [1.1] de-
nes the antenna or aerial as a means for radiating or receiving radio waves. In
other words an antenna is a transducer which adapts the impedance of a guiding
device to the free space impedance and allows the transmission or reception of elec-
tromagnetic energy. During the last decades, there has been a growing interest in
developing more complex antenna systems. Nowadays, these antenna systems can
be used as sensors such as vigilance, detectors or RADAR, in personal communica-
tions (xed or mobile), instrumentation equipment, measurement purposes, satellite
communications or space exploration.
The work carried out since September 2009 in the Radiation Group, of the Tech-
nical University of Madrid (UPM) has been focused on the research of the application
of new technologies to antenna arrays and radio frequency circuits. This doctoral
thesis parts fundamentally from the necessity to nd more exible solutions that
could be apply in the analysis, design and construction of passive and active anten-
na systems.
1.1. Motivation
Since the early beginnings of telecommunications, some of the research eort in
the antenna eld has focused on getting higher directivities in order to reach longer
distances or transmit higher amount of information. The old days where big anten-
nas which work at low frequencies using earth surface wave or ionosferic propagation
1
2 INTRODUCTION
have nished due to the large dimension and low exibility of the antennas. Mo-
reover, by using these frequencies the operation bandwidth is very narrow and high
data rate systems can not be used. If higher frequency bands are used, the physical
dimensions of the antenna decrease and the available bandwidth increase. However,
the free space losses increase too, and the distances that they can reach are shorter.
Therefore the quest to achieve higher directivities keeps on. The directivity of an
antenna is its ability to gather the energy in a specic region of the space, to say
so, if the energy is focused in a smaller beam, this electromagnetic waves will reach
longer distances. Nevertheless, not only for reaching higher distances the directi-
vity can be used, but also for avoid external interference sources. In example in
radioastronomy, the antennas have very large directivities in order to obtain narrow
beams for deep space exploration where to avoid another heavenly bodies is very
important. As long as de directivity is directly proportional to the efective area of
the antenna (eq. 1.1, directivity for parabolic antennas) [1.2], the antennas tended
to follow the rule of thumb "the bigger, the better".
D =
(πD
λ
)2
εap (1.1)
Where D is the antenna diameter, λ is the wavelength and εap is the aperture
eciency.
For 29 years the Eelsberg Radio Telescope [1.3] was the largest fully steerable
radio telescope on Earth. In 2000 it was surpassed by the Green Bank Telescope
[1.4], placed in West Virginia with an elliptical 100 by 110-metre aperture. The
biggest antenna in the world is placed in Arecibo, Puerto Rico with 305 m [1.5],
nevertheless, this antenna does not have movement capabilities.
The construction, maintenance and operation aspects of these huge antennas
make them very complex and expensive, consequently other solutions were explo-
red. An astronomical interferometer achieves high-resolution observations using the
technique of aperture synthesis, mixing signals from a cluster of comparatively small
telescopes rather than a single very expensive monolithic telescope.
Examples of Astronomic interferometry clusters are the Very Large Array (VLA),
Motivation 3
(a) Very Large Array (VLA) New Mexico. (b) Atacama Large Millimeter Array (AL-
MA).
Figure 1.1: Astronomic interferometry clusters [1.6].
in New Mexico [1.7] or Atacama Large Millimeter Array (ALMA) [1.8]. The VLA
consists of twenty-seven 25-m-diameter antennas arranged in a Y-shaped array. Each
arm of the Y is approximately 21 km long and the antennas can be moved to various
positions on the arms by a rail-mounted transporter and it gets the smallest angular
resolution of 0.05 arcseconds at a wavelength of 7 mm. On the other hand ALMA is
compound by 66 12-meter and 7-metter diameter radio telescopes observing at milli-
meter and sub-millimeter wavelengths placed at 5000 meters altitude over distance
of 16 km. With a spatial resolution of 10 milliarcseconds is 10 times better than the
Very Large Array (VLA) and 5 times better than the Hubble Space Telescope. The
Square Kilometer Array (SKA) [1.9] is the biggest project of interferometry teles-
cope under developement, it will be deployed in Australia and South Africa and it
will have a total collecting area of approximately one square kilometer. The SKA
will combine the signals received from thousands of small antennas (15 m diameter
dish) spread over a distance of more than 3000 km. These antennas cover from 70
MHz to 30 GHz and simulate a single giant radio telescope capable of extremely
high sensitivity and angular resolution.
The main advantage of the antenna arrays over the parabolic dishes is the mul-
tibeam or beam steering capability. If the signals are properly combined with a
determined amplitude distribution or phase dierence, the group of antennas can
transmit or receive dierent levels of power from dierent points of the space wit-
4 INTRODUCTION
hout moving physically the antenna. This is very useful for high interference envi-
ronments, adaptive systems, etc. Nowadays this type of antennas, which work at
higher frequencies to achieve size reduction, are integrated in mobile devices. Howe-
ver the design and the integration of this type of systems is really complex because
neither the technology nor the theoretical foundations can handle the problems that
appear in the radiation pattern composition of the antenna elements.
The purpose of the work carried out in the rst part of the thesis is triple: in
the rst place, to identify and describe the problems related to active antennas
focusing on steering systems. Secondly, to study the solutions already proposed
exploring several dierent paths and techniques and nally to contribute with new
solutions. Nowadays broadband communication systems are required everywhere,
therefore there is a growing interes on COTM (Communications On The Move)
systems. They are installed in cars, ships and more recently in aircrafts. To achieve
large data rates in such a high frequency, it is necesary high signal to noise ratio.
In order to get more radiated or received power, the gain of the antenna is increase
but at the same time the beam is reduced, therefore it is necessary to accurately
steer the antenna. The antennas can be mechanically or electronically oriented.
By means of a thorough study of the art in steering systems, simulations and the
identication of the main problems and constraints, an airborne sub array antenna
for satellite communications is built. This prototype will be used to illustrate the
capabilities of a broad band, dual circular polarized 4 x 4 array. This antenna
includes miniaturized hybrid couplers and it is electronically steered by means of a
wideband butler network.
The second part of the thesis is related to the use of metamaterials in acti-
ve antennas. Since the discovery of metamaterials, the interest on it has grown
explosively. More than 10 years have took place since the emergence of the rst
metamaterials and since then they have proved to be exceptionally promising for
both research and applications. The investigation of metamaterials is currently one
of the most active topics in engineering and physics. A number of detailed review
articles and books have been published recently [1.101.15].
These complex materials allow us to achieve extraordinary electromagnetic pro-
Motivation 5
perties which are sometimes even not available in natural materials. The name
for articial complex materials that agree with this characteristic of unconventio-
nal properties can be called metamaterials. Metamaterials are dicult to dene
and to classify. A unique denition for metamaterial does not exist within the re-
search community and in the literature. One formal denition proposed in [1.16]
is macroscopic composites having man-made, three-dimensional, periodic cellular
architecture designed to produce an optimized combination, not available in nature,
of two or more responses to specic excitation. Another denition in [1.13]: a
metamaterial is an articial structure of material that, in a certain frequency ran-
ge presents unusual electromagnetic properties (as propagation of backward waves,
negative refraction, presence of forbidden zones, etc.), which gains its properties
from its structure rather than directly form its composition. Therefore, in order to
clarify the situation, a global denition for metamaterials that might satisfy most
researchers is artitial engineered complex electromagnetic functional structured
materials, by placing them in a periodic manner, have a superior electromagnetic
properties that can not be observed in the constituent materials used to manufacture
them.
The classication of these materials (Fig. 1.2) is also a hard task since nowadays
combination of fabrication techniques, properties and applications are combined in
order to nd new materials. However the concept of metamaterials is treated quite
general among the researchers with topics such as frequency selective surfaces (FSS),
electromagnetic/photonic bandgap (EBG/PBG) structures, left-handed (LH) ma-
terials, articial magnetic conductors (AMC) or high-impedance surfaces (HIS) or
hard/soft surfaces, articial dielectrics (AD), and plasmonic medias. A raw descrip-
tion of these materials is the following:
Frequency selective surfaces (FSS) are dielectric layers of very large ex-
tent, which contain planar conductive elements on its true side and which
backside is free. If the true side of the selective surface is illuminated by
harmonic waves of various frequencies, some waves are transmitted with a mi-
nimum attenuation, some waves are totally reected back to the half-space of
6 INTRODUCTION
Figure 1.2: Classication of metamaterials [1.17].
the source, and some waves are partially transmitted and partially reected.
So, a FSS can be viewed as a lter for plane waves at any angle of inciden-
ce [1.18].
Articial magnetic conductors (AMC) (or high impedance surfaces (HIS))
have new emerged properties as a magnetic response, even if the component
materials are non-magnetic. We can associate the hard and soft surfaces in
the class of the AMC because their behaviour can be used as AMC and perfect
electric conductors (PEC).
Left-handed (LH) materials have been called by many names as negative
refractive index (NRI) materia or backward wave media. These properties have
simultaneously negative permittivity and permeability. Composite right/left-
handed (CRLH) concept is an articial transmission line (TL) approach that
describe the behavior (RH or LH) of these medium depending on the frequency
range of working.
Articial Dielectrics (AD) consist of a large number of subwavelength con-
ducting obstacles embedded in a homogeneous host medium [1.19]. Calling
Motivation 7
such dielectrics as metamaterials make sense because the conducting proper-
ties of metals are being changed to a dielectric-type behavior in the macros-
copic properties. One of the kind of AD can be the magnetic materials (mu
negative (MNG) media).
Plasmonic medias (or epsilon negative (ENG) media) is the name given to
a discipline seeking to benet from the resonant interaction obtained under
certain conditions between an electromagnetic radiation and the free electrons
with the interface between a metal and a dielectric material. This interaction
generates density waves of electrons, behaving like waves and called plasmons
or surface plasmons at optical frequencies [1.20].
Electromagnetic/photonic bandgap (EBG/PBG) materials are also
known as photonic crystals. They are periodic structures that can be made
by metallic, dielectric or metallodielectric elements. These structures are used
to control and manipulate the propagation of electromagnetic waves. The
EBG structures have two important attributes that are to create a bandgap
operation and to localize frequency windows in the bandgap by breaking the
periodicity of the structure. The rst property is useful in using EBG as a
spatial and frequency lter, while the second property is useful in propagating
the EM wave in a desired frequency and direction. There exist 1D, 2D or
3D periodic structures in which the propagation of electromagnetic waves is
inhibited in some frequency bands (called bandgaps or stopbands) or directions
that are determined by the periodicity of the materials and their dielectric
constants.
Therefore, metamaterials can contribute to enhance the performance of active
antennas. The surface wave supression properties of the last type of metamaterials
described, EBG materials are used in this work to reduce the mutual coupling bet-
ween radiating elements. The reason for the integration of EBG between radiating
elements as electromagnetic barriers is to replace the cavities which are heavier and
more expensive.
8 INTRODUCTION
The last part of this thesis is dedicated to the distribution networks that feed the
radiating elements in active antennas. The behaviour of these distribution networks
can be modied by placing components which change their capacitance, resistance,
permeability, etc. by means of a control signal. They can split, divert, block or
allow the current and therefore they change the array antenna features. This work is
oriented to explore the technologies, topologies and concepts regarding phase shifting
in the distribution networks. There are several ways to classify this components,
but from the point of view of steering antennas there are two main groups: digital,
which actuates as switches between dierent paths, and analog, which can modify
the characteristic impedance of the transmission line, thus the propagation speed is
change and therefore a phase shift is obtained.
Digital:There are two main switching techniques based on semiconductor bia-
sed: CMOS and PIN diodes. These devices change their conductor or isola-
tion behavior as a function of its chemistry composition. On the other hand,
MEMS recently attracts more interest, due to its mechanical operation and its
excellent features.
-MOSFET transistors use an electric eld to control the shape and hence
the conductivity of a channel of one type of charge carrier in a semiconductor
material. These devices has three terminals, gate, drain and source, in such
a way that a negative gate-to-source voltage causes a depletion region on the
channel for a n type. If the depletion region expands to completely close the
channel, the resistance of the channel from source to drain becomes large and
the FET is eectively turned o like a switch. Likewise a positive gate-to-
source voltage increases the channel size and allows electrons to ow easily
(Fig. 1.3).
CMOS switches Complementary metal-oxide-semiconductor is a technology
for constructing integrated circuits. The most important characteristics of
CMOS devices are high noise immunity and low static power consumption.
Since one transistor of the pair is always o, the series combination draws
signicant power only momentarily during switching between on and o states.
Motivation 9
Figure 1.3: Schematic view of a Field Eect Transtistor.
Consequently, CMOS devices do not produce waste heat and therefore the
power consumption is lower. However, the main drawback is that nowadays
the higher frequencies where they can work do not reach eectively 20 GHz.
There are several materials which the transistors are made from, the switches
are commonly designed using III-V semiconductor based such as AlP or more
recently GaAs, which are used in high power and high throughput applications
[1.21]. New dierent materials are currently tested, such as Silicon-on-Sapphire
(SOS) [1.22], which obtains higher frequency operation ranges but can not
handle high power.
The main advantage of the FET is its high input resistance. Thus, it is a
voltage-controlled device, and shows a high degree of isolation between input
and output. It has low noise level, it exhibits no oset voltage at zero drain
current and it has good thermal stability. Its main disadvantage is the inte-
gration of device in the systems, in order to reduce the insertion losses and
the noise gure, the introduction of these devices has to be made during the
photolitography processes, the only way to make protable the systems is for
mass production. However this problems are being currently addressed [1.23].
-The Micro ElectroMechanical Systems (MEMS) are micro mechanized
surfaces which use a mechanical movement to obtain a short or open circuit
in a radio frequency transmission line, with very low losses. The movement is
caused by an electro static force between to electrodes which are fed with a
large diference voltage. The MEMS can be used from µ-wave frequencies up
10 INTRODUCTION
to mm-wave frequencies (0.1-100 GHz) and they have a great performance in
high frequencies up to 30 - 40 GHz, however, in higher frequencies due to the
fabrication process the package capacitances yields higher insertion losses.
RF MEMS switches oer higher performance than pin diodes or FET diode
switches and those can work up to 120 GHz. However, important issues such as
long reliability, packaging techniques and production costs, are currently being
addressed. The switches are either fabricated using an attached membrane
xed in two points or a oating cantilever, and are modeled as mechanical
springs with an equivalent spring constant k [N/m]. The actuation mechanism
is achieved using an electrostatic force between the top and bottom electrodes.
This electrostatic force depends on the geometry of the switch, but mainly on
the actuation voltage [1.24]. Depending on the topology there are two main
types of MEMS: cantilever (Fig. 1.4(a)), where the application requires a
series switch and membrane or bridge (Fig. 1.4(b)) for shunt schemes.
(a) Cantilever [1.25]. (b) Capacitor bridge [1.24].
Figure 1.4: Micro ElectroMechanical Systems schematic view.
Due to the physical constrains the advantages of this systems are the near-zero
power consumption (only 10 - 100 nJ per switching cycle), very high isolation
thanks to its low o-state capacitances (2-4 fF) and very low insertion loss due
to its low series resistance up to 30 GHz. However, MEMS have low switching
speed, the power that the element can handle is not too large, the actuation
voltage is very high and the long term reliability is only up to 10 billion cycles.
Therefore this components can full the requirements in special applications
systems such as defense and ground station satellite communications, but are
expensive for terminal users.
The main advantages of this devices are the low insertion losses, linearity,
Motivation 11
null power compsumtion, and high isolation. On the other hand there are
some drawbacks such as low operation speed, high voltage operation, power
handling, limited life operation time or packaging.
-The P-I-N diode is a current controlled resistor at radio and microwave
frequencies. A unique feature of these diodes is their ability to control large
amounts of RF power with much lower levels of DC. It is a silicon semiconduc-
tor diode in which a high resistivity intrinsic I-region is between a P-type and
a N-type region. When the PIN diode is forward biased holes and electrons are
injected into the I-region. These charges do not immediately annihilate each
other. These charges stay alive (τ , carrier lifetime) and lowers the eective
resistance of the I-region to a value RS. When the PIN diode is at zero or
reverse bias there is no stored charge in the I-region and the diode appears as
a capacitor, CT , shunted by a parallel resistance RP [1.26].
By varying the I-region (Fig. 1.5(a)) width or area it is possible to construct
PIN diodes of dierent shapes but similar RS and CT characteristics. However,
the thicker I-region diode would have a higher bulk or RF breakdown voltage
and better distortion properties. On the other hand the thinner device would
have faster switching speed [1.27] - [1.28].
(a) PIN diode outline. (b) Forward and reverse bias.
Figure 1.5: PIN diode and equivalent circuit.
Another important issue for these components is the biasing. The higher IF ,
the lower RS and therefore the lower the insertion losses, when the series
topology is used. For this same series topology, when the reverse biased is
applied, the higher -VR, the lower the CT , and the higher isolation is obtained
12 INTRODUCTION
Table 1.1: Characteristics summary of the digital switches.
Parameter GaAs pHEMT GaAs PIN Si PIN RF-MEMS
Insertion loss [dB] 0.35-0.65 0.3-0.6 0.4-1.2 0.1-0.5
Isolation [dB] >30 12-25 34-55 20-30
Actuation Voltage [V] 5 5 2-5 30-80
Switching speed [µs] 0.02-0.06 0.02 0.02 2-10
Power consumption [µW] <100 <100 <1 0
Power handling [dBm] 36 33 30 23-27
Linearity IP3 [dBm] 55-72 65 70 65
[1.29]. However due to system restrictions, not always high IF and -VR are
possible.
Finally, in order to have an overview, in Table 1.1 a comparison between the three
technologies is shown. Due to the requirements of our design in terms of operation
frequency, power handle, insertion losses and isolation, the technology chose is PIN
diodes.
Analog:The phase shifters based on analog devices such as varactor diodes
or ferrite core have the main advantage of not to have quantization errors and
very low insertion loss. These devices follow a continuous signal (voltage or
current) which yields a behavior change in the conducting material properties
without steps. However the control signals are less robust against noise.
-The operation of FERRITE cores phase shifters is relatively easy. By
changing the magnetic eld inside a waveguide, the propagation constant can
be modied, and a phase dierence obtained. The current that ows along
a coil with a ferrite core is electronically adjusted (Fig. 1.6), obtaining with
that 360o phase variation [1.30].
-VARACTOR diodes are P-N diodes that changes its capacitance and the
Motivation 13
Figure 1.6: Ferrite core phase shifter.
series resistance as the reverse bias applied to the diode is varied. No current
ows but since the thickness of the depletion zone varies with the applied bias
voltage, the capacitance of the diode can be made to vary. Generally, the
depletion region thickness is proportional to the square root of the applied
voltage, capacitance is inversely proportional to the depletion region thickness
(Fig. 1.7). Thus, the capacitance is inversely proportional to the square root
of the applied voltage. This property of capacitance variation is utilized to
achieve a change of phase response or line impedance. Therefore, they can be
used as a phase propagation in phased array antenna applications.
Figure 1.7: Varactor diode outline.
Table 1.2 shows a summary of the analog phase shifters devices. In orther to
establish a comparison with digital devices the same parameters have been shown.
The quantitative gures in these type of devices are determined by the construction
process, therefore only qualitative values are presented.
By presenting this historical evolution of the active antennas focusing on the
14 INTRODUCTION
Table 1.2: Characteristics summary of the analog phase shifters [1.31].
Parameter Varactor Ferrite
Loss High Medium
Actuation Voltage [V] 0-30 100
Switching speed Nano seconds Milli seconds
Power consumption Negligible Negligible
Mounting complexity Low Medium/high
Cost Medium High
steering systems and the problems so far that are currently being addressed the
following objectives are pursued in this thesis.
1.2. Objectives
The eld of active antennas is very wide in surveillance radar technology, ATC
(Air Trac Control), SAR (Synthetic Aperture Radar), etc. Applications of active
antennas are already well identied what is still under investigation are the techno-
logies, materials, switching networks, etc. Considering that a lot of experiences in
the area of steering antennas are achieved, the objective of this doctoral thesis is
trying to contribute to an important series of aspects in the scope of the potential
application of these concepts in the design, analysis and prototyping in the eld of
active antennas, where the Radiation Group has wide experience. The electronic
steerable antennas can be divided in three dierent parts.
First, the selection of group topology, secondly the radiating elements itself, and
nally, the phase shifters placed in the distribution networks.
This thesis is divided in three main parts as well, where the following objetives
are pursued:
Objetive 1: this thesis analyzes the array antennas and their problems re-
Objectives 15
garding size, materials, construction or topology for steerable antennas. A
thorough study is carried out and trade o between angular resolution, direc-
tivity reduction in squinted angles, cross beam losses and number of bits of the
phase shifters are decided for a real satellite communications steering antenna
in X band (7.25-8.4 GHz). After a large number of simulations, the best op-
tion is identied and a 4x4 sub array prototype is constructed and measured
to demonstrate its operation.
Objetive 2: this thesis explores the use of metamaterials for mutual coupling
reduction purposes. By using double layer and via edge location, mushroom
EBG materials are reduced by 30% and placed between radiating elements
in low permittivity substrates. In this way the surface waves that appears in
planar antennas are suppressed in the same way that cavities work but this
solution is lighter and it can be integrated in a photolitography process. These
electromagnetic barriers are used in the previous sub array in order to enhance
its radiation properties.
Objetive 3: this thesis proposes the combination of the well known topolo-
gies and technologies regarding switches and phase shifters to get important
features for phased arrays antennas such as match or unmatch ports, maxi-
mum isolation between ports, balanced or unbalanced power distribution and
of course phase dierence accuracy.
With these objetives, the general purpose of this work is to be a practical appli-
cation of the proposed methods to real antenna systems. Therefore, this doctoral
thesis allows to extend the knowledge of the analysis, design and operation of active
antennas focusing on steerable ones, and proposes possible solutions that help to
improve the steerable antenna performances.
As observed from the description of the main objetives of the thesis, it can
be noticed that a similar methodology is followed in the three areas in order to
accomplish each of the objetives. First of all, a thorough study of the state of art in
the topic of interest is given, in order to know the situation of research in the area
of active antennas. Afterwards, theoretical studies are done in order to propose a
16 INTRODUCTION
solution, to construct and measure the prototype and identify the advantages and
drawbacks that the proposed solution may have. These last two parts, prototyping
and measuring, are the key aspects in this thesis which is not focused in theoretical
formulations but practical solutions. Finally, the analysis of the results is done, and
the main conclusions of each of the three areas of study are given.
1.3. Outline of the thesis
The main document is organized in ve chapters. Chapter 1, already exposed,
presents the motivation which leads this work. Furthermore, the main objetives are
highlighted and an outline of the thesis is detailed.
In Chapter 2, a complete revision of the array antenna basics is illustrated.
Issues such as the amplitude and phase feeding laws, and antenna feeding networks
are shown along this chapter. Through the design of a real airborne steerable anten-
na for satellite communications in X band, the most common problems regarding the
number of beams, bits selected, widening of the main beam and consequently direc-
tivity losses are addressed. At the end of this chapter measurements of a subarray
prototype are shown.
Chapter 3 presents the fundamental theory of Electromagnetic Band Gap (EBG)
Materials and their application for mutual coupling reduction. In this chapter, an
analytical design method is extended to double layer structures for size reduction.
These new smaller structures are used in steering antennas to cancel surface wave
propagation modes and reduce mutual coupling in planar antennas.
Chapter 4 illustrates a thorough revision of the topologies that are used in the
design of switches and phase shifters so far. In order to overcome the demanding
requirements of switchable beam antennas used for automovile RADAR a combina-
tion of switch topologies to obtain equal insertion losses in a reective phase shifter
is presented. Besides, a matched ports SPDT based on the previous reective phase
shifter is proposed.
Finally, Chapter 5 gives the conclusions in the form of the original contributions
of this Ph.D. dissertation and the future research lines that it has given rise to.
Outline of the thesis 17
Furthermore, the framework in which this work has been carried out (research pro-
jects, nancial support, international collaborations...) are reported together with
the publications this Ph.D. dissertation has generated.
References
[1.1] IEEE standard denitions of terms for antennas. IEEE Std 145-1983, pages
01, 1983.
[1.2] A. W. Love. Reector Antennas. IEEE Press, 1978.
[1.3] O. Hachenberg, B.H. Grahl, and R. Wielebinski. The 100-meter radio telescope
at Eelsberg. Proceedings of the IEEE, 61(9):1288 1295, Sept. 1973.
[1.4] S. Srikanth, R. Norrod, L. King, and D. Parker. An overview of the green
bank telescope. In Antennas and Propagation Society International Symposium,
1999. IEEE, volume 3, pages 1548 1551, Aug. 1999.
[1.5] P.A. Castleberg and K.M. Xilouris. The Arecibo observatory. Potentials, IEEE,
16(3):33 35, Aug. 1997.
[1.6] National Radio Astronomy Observatory Nrao. http://www.nrao.edu/. Acces-
sed: 31/01/2013.
[1.7] P.J. Napier, A.R. Thompson, and R.D. Ekers. The very large array: Design
and performance of a modern synthesis radio telescope. Proceedings of the
IEEE, 71(11):1295 1320, Nov. 1983.
[1.8] T. de Graauw. The Atacama large millimeter/submillimeter array. In Infra-
red, Millimeter and Terahertz Waves (IRMMW-THz), 2011 36th International
Conference on, pages 1 4, Oct. 2011.
[1.9] A.J. Faulkner, P. Alexander, and J.G.B. de Vaate. System design for SKA ca-
pable aperture arrays. In Electromagnetics in Advanced Applications (ICEAA),
2012 International Conference on, pages 752 755, Sept. 2012.
[1.10] P.-S. Kildal, A.A. Kishk, and S. Maci. Special issue on articial magnetic
conductors, soft/hard surfaces, and other complex surfaces. Antennas and Pro-
pagation, IEEE Transactions on, 53(1):27, Jan. 2005.
19
20 REFERENCES
[1.11] C. Caloz and T. Itoh. Electromagnetic Metamaterials: Transmission line
theory and microwave applications. Wiley and IEEE Press, 2005.
[1.12] N. Engheta and R.W. Ziolkowski. Electromagnetic Metamaterials: Physics
and Engineering Exploration. Wiley and IEEE Press, 2006.
[1.13] R.W. Ziolkowski and N. Engheta. Metamaterial special issue introduction.
Antennas and Propagation, IEEE Transactions on, 51(10):25462549, Oct.
2003.
[1.14] G.V. Eleftheriades and K.G. Balmain. Negative-Refraction Metamaterials:
Fundamental Principles and applications. Wiley and IEEE Press, Jul. 2005.
[1.15] R. Marqués, F. Martín, and M. Sorolla. Metamaterials with negative para-
meters: theory, design and microwave applications. John Wiley and Sons, Feb.
2008.
[1.16] D.R. Askeland and Phulé. The Science and engineering of materials.
Brooks/Cole problishing/thomson learning, 2003.
[1.17] J.M. Fernández-González. Application of metamaterial structures in the de-
sign, analysis and prototyping of planar antennas. PhD thesis, ETSI de Tele-
comunicación, Technical University of Madrid (UPM), 2008.
[1.18] B.A. Munk. Frequency selective surfaces: theory and design. John Wiley and
sons IEEE press, 2000.
[1.19] R.E. Collin. Field theory of guided waves. New York: IEEE Press, 1991.
[1.20] H. Raether. Surface Plasmons on smooth and rough surfaces and on gratings.
Berlin: Springer-Verlag, 1988.
[1.21] P. Katzin, B.E. Bedard, M.B. Shifrin, and Y. Ayasli. High-speed, 100+w rf
switches using gaas mmics. Microwave Theory and Techniques, IEEE Transac-
tions on, 40(11):1989 1996, Nov. 1992.
REFERENCES 21
[1.22] D. Kelly, C. Brindle, C. Kemerling, and M. Stuber. The state-of-the-art of
silicon-on-sapphire cmos RF switches. In Compound Semiconductor Integrated
Circuit Symposium, 2005. CSIC '05. IEEE, Nov. 2005.
[1.23] Byung-Wook Min and G.M. Rebeiz. Ka-band low-loss and high-isolation
switch design in 0.13- µm cmos. Microwave Theory and Techniques, IEEE
Transactions on, 56(6):1364 1371, Jun. 2008.
[1.24] G.M. Rebeiz and J.B. Muldavin. Rf mems switches and switch circuits. Mi-
crowave Magazine, IEEE, 2(4):59 71, Dec. 2001.
[1.25] V.K Varadan, K.J. Vinoy, and K.A. Jose. RF MEMS and their applications.
John Wiley and sons IEEE press, 2003.
[1.26] Skyworks. Application Note APN1002, Design with PIN Diodes.
[1.27] J.V. Bellantoni, D.C. Bartle, D. Payne, G. McDermott, S. Bandla, R. Tayra-
ni, and L. Raaelli. Monolithic GaAs p-i-n diode switch circuits for high-power
millimeter-wave applications. Microwave Theory and Techniques, IEEE Tran-
sactions on, 37(12):2162 2165, Dec. 1989.
[1.28] M. Case, M. Matloubian, Hsiang-Chih Sun, D. Choudhury, and C. Ngo. High-
performance w-band GaAs PIN diode single-pole triple-throw switch CPW
MMIC. In Microwave Symposium Digest, 1997., IEEE MTT-S International,
volume 2, pages 1047 1051, jun 1997.
[1.29] K.W. Kobayashi, L. Tran, A.K. Oki, and D.C. Streit. A 50 mhz-30 GHz
broadband co-planar waveguide SPDT PIN diode switch with 45-db isolation.
Microwave and Guided Wave Letters, IEEE, 5(2):56 58, Feb. 1995.
[1.30] Shiban K. Koul and Bharathi Bhat. Microwave and Millimeter wave Phase
shifters Volume 1. Artech House Inc., 1991.
[1.31] R. Sorrentino, R.V. Gatti, and L. Marcaccioli. Recent advances on millimetre
wave recongurable reectarrays. In Antennas and Propagation, 2009. EuCAP
2009. 3rd European Conference on, pages 25272531, 2009.
Chapter 2
State of the art on multibeam and
recongurable steering array
antennas
Nowadays the communication systems are more and more demanding, the re-
quirements of these systems in terms of the data throughput, number and location
of users force the designers to explore new techniques to develop the systems. An-
tennas have a very important role in telecommunication systems because they can
focus the transmited power or radiate it in all the directions, to avoid another user
signals or to explore dierent angles. To say so, the antennas can vary their beha-
viour in order to adapt themselves to the environment and to oer the best possible
conguration. Those antennas are known as well as ADAPTIVE ANTENNAS.
2.1. Introduction
By applying the superposition principle due to the linearity of the Maxwell's
equations and considering that the current in each element is the same as that of
the isolated element (neglecting coupling), the total eld of the array is determined
by the vector addition of the elds radiated by the individual elements. This is
usually not the case and depends on the separation between the elements. To
provide very directive patterns, it is necessary that the elds from the elements of
the array interference constructively (add) in the desired regions and interference
destructively (cancel each other) in the other regions. If the radiating elements of
23
24 STATE OF THE ART ON MULTIBEAM AND . . .
the array are identical, there are ve parameters that can be used to shape the
overall radiation pattern of the antenna:
1. The geometrical conguration of the overall array (linear, circular, rectangular,
spherical, etc).
2. The relative displacement between the elements.
3. The excitation amplitude of the individual elements.
4. The excitation phase of the individual elements.
5. The relative pattern of the individual elements.
Usually, the elements are equal and they have the same spatial orientation, the-
refore they are completely equivalent by a simple translation (Fig. 2.1). In this
case, the array can be described by the relative position of the radiating elements,
the feeding currents and the radiation pattern of each element. The eld radiated
of each element can be described by means of the radiation pattern of the radiating
element placed in the origin axis fed by a reference current I0 and a phase term,
which takes into account the displacement respect to the origin axis (eq. 2.1).
En(r, θ, φ) = Ee(r, θ, φ)InI0ejk0rrn (2.1)
Figure 2.1: Array splitting problem.
Introduction 25
(a) In transmission. (b) In reception.
Figure 2.2: Wavefront combination to get electronic steering.
The introduction of a phase shift α between two consecutive elements in an
antenna array allows to adjust the steering direction of the main beam without
moving the antenna physically. Through the reciprocity theorem, this eect can be
explain in two dierent ways:
In transmission: The phase shifters placed in the distribution network create
a delay between the wavefront of each element. The combination of the wave-
fronts of all the radiating elements along the time shows the desired steering
direction (Fig. 2.2(a)).
In reception: The wavefront of the desired direction travels with a determined
velocity, however, this does not reach all the radiating elements at the same
time. Those radiating elements, which receive earlier the wavefront, must
delay their contribution to the summation in order to sum all the contributions
appropiately (Fig. 2.2(b)).
From the eq. 2.2, a phase shift α can be introduced, as a step forward respect
to the previous radiating element: An = anejnα
FA(θ, φ) =∑n
anejnαejnk0dcosθ =
∑n
anejn(k0dcosθ+α) =
∑n
anejnψ (2.2)
Applying the coordinate transformation shown in eq. 2.3, it is observed that the
array factor as function of ψ is the inverse DFT of the magnitude feeding law.
ψ = k0dcosθ + α (2.3)
26 STATE OF THE ART ON MULTIBEAM AND . . .
Similarly to eq. 2.3, where the periodicity of the function is 2π, in eq. 2.2 the
variable ψ has a period 2π. However, the radiation pattern has specic limits due
to the real values of the variable θ. The visible margin is the collection of the array
factor values which correspond directly with the radiation pattern of the antenna.
Its lenght is 2k0d and its center is ψ = α.
0 ≤ θ ≤ π → −k0d+ α < ψ < k0d+ α (2.4)
Figure 2.3: Visible margin of an array in ψ and θ.
In Fig. 2.3 the visible margin of a generic steering array is shown. From this
gure, three types of antennas can be named as function of the phase shift α:
Broadside array: the maximum of the radiation pattern is perpendicular to
the plane where the radiating elements are placed (θmax = π/2), in this case
Introduction 27
α = 0 and the visible margin is −k0d < ψ < k0d.
Exploration array: the maximum of the radiation pattern points to θmax
which is controlled by the progressive phase shift α (eq. 2.5).
θmax→ψ=0 = arccos
(−αk0d
)(2.5)
End re array: the maximum of the radiation pattern is located in the
surface where the radiating elements are placed (θmax = 0 or π). In this
case (θmax = 0) the value of α is equal to −k0d and the visible margin is
−2k0d < ψ < 0.
There are some important aspects which must be taken into account when an
exploration array is design. If the the distance between the elements d and/or the
phase shift α are higher than specic values, a replica of the main lobe could appear
in the radiation pattern, this phenomena is called grating lobes. This eect reduce
the separation between elements to 0.6 - 0.8λ for broadside arrays and 0.4 - 0.5 λ
for end-re arrays.
The recongurable antennas are dened as antennas that can radiate on demand
at several predetermined frequencies, or create rejection notches at various frequen-
cies, they can change their polarization or their radiation patterns on demand. The
motivation of this type of systems is to congure or to adjust the operation under
variable conditions, recover their functionality under any adverse or anormal condi-
tions by self-conguration (self-healing) such as degraded components that can be
recovered or damaged components that can be bypassed.
This is a very wide topic because it has a lot of application areas from personal
devices to high complexity facilities. It can be divided in its kind of application,
instant of change, used devices, parameters of change and exibility limits. These
systems are very wide spread because they are used in military phased arrays, mobile
base stations, broadcast facilities and industrial or domestic sensors.
Regarding instant of change the recongurable antennas can be divided in two
groups.
28 STATE OF THE ART ON MULTIBEAM AND . . .
The ones which change the property or propperties of the antenna dinamically,
that is to say, during their operation
and the ones which can be adjust before the operation.
An example for the rst ones are the phased arrays which recongures beam
pointing direction during operation to look for targets over wide area of sky. A
mobile base station antenna is mechanically tilted when they are set in order to
adjust for correct coverage before turn it on. The main advantages of the dynamic
systems are the speed and exibility, however they have some drawbacks such as
high cost, additional DC power and additive RF losses that not all the systems can
handle.
There are several ways to control the operation of a recongurable system. On
the one hand, the ones controlled mechanically (hand, motor, hydraulically) which
are slow but with a low cost and they do not add additional RF losses, and on
the other hand the ones controlled electronically such as switches based on MEMS,
PIN diodes or CMOS and analog or digital phase shifters which are faster but their
drawbacks are the high cost and the additional RF losses.
The exibility limit is an important parameter which directly aects on the
complexity of the system. For example, in a phased array, the fewer the pointing
beam steps, the simpler the system is. This issue limits the improvement of the
system regarding steering accuracy or gain in the desired direction.
Finally, there are many groups in which the recongurable antennas can be di-
vided in terms of the parameter change. They can change their radiation pattern,
beam direction (steering antennas) [2.1], beamwith or coverage angle (reector an-
tennas which are fed by several horn antennas in order to obtain dierent beam
shapes to earth coverage) [2.2], sidelobe level (in systems where the noise level can
change) [2.3], null steering (to avoid interferences) or single/multiple or split beam
(reector antennas) [2.4]. Another parameter of change is the polarization, it can
change from vertical to horizontal [2.52.7] or change the type of circular Left/Right
Handed Polarization (LHCP/RHCP) [2.8]. By feeding a radiating element with two
orthogonal modes which have a phase dierence of 90o the circular polarization is
Introduction 29
obtained, i.e. a patch antenna fed by a 90o 3dB branch line coupler ([2.9]). And
the last one is the operation frequency band [2.10], there are Ultra Wide Band
(UWB) antennas [2.11], there are systems which can broadering the band (large
number of military radios from HF to UHF) or systems that can use multiple bands
(the mobile phones use nowadays 5 or more frequency bands) [2.12]. The methods
in order to achieve these recongurable characteristics are not formal established,
they can change the antenna geometry, feeding points, add additional reactance,
recongurable matching circuits [2.13], etc.
There are no real design guidelines or theories for the operation of recongurable
antennas and some clear goals to advance the theoretical foundation of these anten-
nas are needed. In this way, many more recongurable antennas could be included
in complex and multifunctional systems.
The main target of this work is to study the steerable systems and to propose new
techniques to enhance their performance. In the next section, the main problems
and constructing strategies when using scanning arrays are described.
As it was previously mentioned by controlling the phase dierence between the
elements, the maximum radiation can be squinted in any desired direction to form
a scanning array. Since in the phased array technology the scanning must be conti-
nuous, the system should be capable of continuously varying the progressive phase
between the elements. In practice, this is accomplished electronically by the use of
analog or digital phase shifters. However, analog systems are less exible and more
complex than digital ones. Therefore, systems which are based on beam switching
are preferred. These systems are simpler but several drawbacks have to be taken
into account in the design process.
The Half Power BeamWidth (HPBW) of an array is dened as the angle where
the directivity in the main lobe is below the maximum 3 dB. In an array the direc-
tivity is directly related to the radiating element directivity and the array factor.
When electronic steering is applied, the visible margin (Fig. 2.3) distort the ψ to θ
transformation and the mean beam gets wider (2.6).
BW ∝ 1
cosθ(2.6)
30 STATE OF THE ART ON MULTIBEAM AND . . .
Table 2.1: Gain reduction and angle widening for scanning angles.
Angle 90o 70o 50o 30o 10o
Gain (dB) 30 29.72 28.84 26.98 22.39
Beamwidth (o) 14.32 15.24 18.69 28.64 82.48
Figure 2.4: HPBW vs scanning angle.
The eect of reduction in squinted angles is directly related to the previous one.
There is a directivity reduction for the farther steering angles which is a consequence
of the wider beamwidth. In Fig. 2.5 it is shown a directivity of 2 dB when the
steering angle is ±50o. Therefore, the farther the scanning angle, the higher are the
directivity losses. In table 2.1 the numerical values of gain and beamwidth for a 16
rows of elements array with scanning elevation are shown.
In a discrete scanning system, it is very important the number of beams that
this is going to use. As it can be seen in Fig. 2.6 there are two diferent systems with
±20o scanning angles. However, one system has four beams meanwhile the other
one has eight beams. For the four beams system the crossing point between beams
is under 15 dB, which means that the link budget could not be satised for certain
Introduction 31
Figure 2.5: Directivity reduction vs scanning angle.
scanning angles. Therefore, the system with eigth beams has better behaviour, since
the crossing point between beams is only 4 dB below the maximum.
Figure 2.6: Crossing point losses vs number of beams.
32 STATE OF THE ART ON MULTIBEAM AND . . .
In this case the dierent number of bits in the phase shifters is analyzed. When
discrete number of shift angles are used, each radiating element is not fed with
the necessary phase, but with one which is aproximated. In Fig 2.7, the radiation
patterns of three systems with ±30o but dierent number of bits are shown. It can
be noticed that the lower number of bits, the higher quantization noise. Eventhough
the same scanning angles are obtained, in systems where the SNR is important for
data rate transmission, high accuracy phase shifters are needed.
Figure 2.7: Quantization noise oor vs number of bits.
The antenna arrays can be divided in two groups depending on the way in which
the radiating elements are fed. On the one hand the passive arrays where the
radiating elements are fed by only one source, and the amplitude and phase feeding
are function of a propagation medium (i.e. the slot arrays antennas which are fed by
a waveguide, used for high power transmission). And on the other hand the active
arrays where each radiating element is fed by an independent source. This last group
is more exible because the feeding law can be controled in a easier way. But this is
not the only one advantage, these systems are more robust against failures since the
wrong operation of a radiating element do not aect signicantly the total radiation
pattern.
Multibeam passive antennas 33
(a) Steering angle of 70o (b) Steering angle of 50o
Figure 2.8: Appearance of grating lobes due to failures in the phase feeding law.
The last eect considered try to illustrate the problem when some phase shifters
do not work properly. There is a situation where the failure of certain parts could
be critical, in the case where a periodicity exists in the phase feeding law, grating
lobes can appear. In Fig. 2.8(a) and Fig. 2.8(b) the radiation patterns of three
systems with dierent failures are shown. It can be noticed that when the radiating
elements are fed in groups of two with the same phase, it is similar to place the
elements with double separation between them and therefore, when steer capability
is used, grating lobes appear. The same situation takes place when the elements are
fed in groups of three with the same phase.
In the next sections the dierent techniques to obtain the phase shift between
the radiating elements are described.
2.2. Multibeam passive antennas
There are several passive methods and devices in order to obtain the steering
angles of an array. After a thorough literature revision here are presented most of
the well known passive techniques.
34 STATE OF THE ART ON MULTIBEAM AND . . .
Butler Matrix Network
The Butler matrix is a passive network with 2n inputs, 2n outputs, 2n−1log22n
hybrid couplers, crossovers and phase shifters [2.14] [2.15]. The function of a Butler
matrix is to combine signals in phase going to or coming from an antenna array. It
produces 2n beams with constant angular separation. Each output signal at n port
(Sn) can be expressed as follows:
Sn =n∑
m=1
Amejαmn (2.7)
where Am is the input signal at port m and αmn is the phase dierence bet-
ween input ports. There are several dierent congurations in order to obtain the
broadside beam [2.16, 2.17], group a large number of radiating elements [2.18, 2.19]
or to get multiple beams in a hexagonal planar array with triangular distribution
[2.20]. In [2.21] a reective Butler network is proposed, its peculiarity resides on
the aditional phase shifts in order to get a symmetric structure, consequently, this
device can be used in transmission and reception at the same time.
Blass Matrix
The Blass matrix [2.22] consists in a determined number of interconnected feeding
lines which feed an antenna array. Two groups of interconnected lines by directional
couplers form the network. The applied signal in the input terminals travels along
the dierent paths which yields a phase shift. The amplitude can be modied by
the directional couplers design [2.23]. In [2.24, 2.25] a synthesis method for the
directional couplers is shown. In [2.26, 2.27] dierent congurations are presented
when applying to antenna arrays.
Nolen Matrix
The Nolen matrix combines the properties from both the Butler and the Blass
one and it was introduced in [2.28]. The Nolen matrix is actually lossless like the
Butler one and serial feeding network like the Blass one. This latter characteristic
is of great interest to avoid the crossovers. However being lossless, the Nolen matrix
Multibeam passive antennas 35
is limited like the Butler one to orthogonal set of output excitations while the Blass
matrix has no constrains on the output excitations. Therefore, a Nolen matrix can be
seen as a lossless Blass matrix [2.29]. As a counterpart, this lossless characteristic
imposes orthogonality on the output excitations that are not found on the Blass
matrices. In fact, the less orthogonal the output excitations of a Blass matrix the
higher the losses.
Rotman lens
The Rotman lens [2.30] are feeding networks which obtain in a easy way a
discrete main beam steering. The fundamental of these lens resides in the time
delay due to the path dierence, consequently the scanning angles are frequency
independent. In [2.31] an eight element array are fed by a Rotman lens.
The weight of dielectric lens does not allow the use of these lens at µwave fre-
quencies, however these are nowadays important for EHF and mmwave frequencies.
The parallel metallic lens solve the weight problem [2.32], where the spacing between
parallel plates is substituted by a guided medium and the medium permittivity is
directly related to the distance between parallel plates.
Summary of passive electronic steerable techniques
In table 2.2 the main characteristics of the passive techniques in order to get
scanning angles are shown. In this way, it is easy to compare the advantages and
the disadvantages of these systems. It can be concluded for all the systems that the
higher number of beams, the bigger the lattice. Therefore the space constraints in
the system will limit its exibility.
2.2.1. Wide band Butler matrix network at X band
In this work, a wideband Butler matrix [2.33, 2.34] with four inputs and four
outputs has been designed. The output phase dierences are ±45 and ±135, and
crossovers are implemented with two hybrid couplers in cascade. The hybrid coupler
is the key design part since the rest of the elements are based on it. By applying
36 STATE OF THE ART ON MULTIBEAM AND . . .
Table 2.2: Characteristics summary of the passive electronic steerable techniques.
Type Scanning
range
Aperture
size
SLL Bandwidth Eciency
Rotman ±45 10λ -20dB 4:1 >63%
Hybrid lens ±10 130λ -16dB - >60%
Gradual index lens ±360 20λ -13dB 9:1 -
Hybrid reector ±30 230λ -22dB - >76%
Blass matrix ±60 15λ -13dB <1% 75%
Butler matrix ±60 16λ -13dB >2:1 40%
Barlett's [2.35, 2.36] theorem to two stages hybrid coupler resolution the following
equations are drawn:
(Z02
Z0
)2
=2Z01Z03
Z20 + Z2
01(Z01
Z0
c0 − 1
)= 1− c20 (2.8)
where c20 is the power ratio between ports 2 and 3. In this case, the design
condition is c20 = 12. There are three missing values (Z01, Z02 and Z03) and only two
equations. Therefore a degree of freedom exists. In this case, the periodic solution
(Z02 = Z0) is selected.
The prototype of the Butler network can be seen in Fig. 2.9(a) [2.37]. As it
can be seen in Fig. 2.9(b), a plain response over the frequency is obtained for
transmission parameters (Sji=-6.5±2). Besides, the reection coecient and the
isolation are adequate (Sii and Sjj < −14dB). The phase dierence between the
outputs is ±45 for port 1 (Fig.2.10(a)) and ±135 for port 2 (Fig.2.10(b)) along
the whole bandwidth.
Multibeam passive antennas 37
(a) Construction (b) Amplitude Scattering parameters
Figure 2.9: Wide band Butler Matrix network
(a) Port 1 (b) Port2
Figure 2.10: Phase of the scattering parameters of a wideband Butler matrix net-
work.
38 STATE OF THE ART ON MULTIBEAM AND . . .
2.3. Scanning active antennas
The complexity of the scanning active antennas which obtain a continuous beam
steering (instead the discrete group of beams obtained with a multibeam passive
antenna) resides on its feeding distribution network. In it every element (for 2D
scanning arrays), row (for elevation scanning) or column (for azimuth scanning)
is feed with a dierent phase. This phase is provided by an element called phase
shifter.
A phase shifter is a device used in automation, conversion technology, and mea-
suring technology to change the phase of electromagnetic oscillations. The design
of a phase shifter depends on the range of operating frequencies, the limits of the
phase change, and the accuracy of the equipment. They can be analog or digital.
Electrically controlled analog phase shifters can be realized with varactor diodes
that change capacitance with voltage or ferrite phase shifters where the current
that ows along a coil with a ferrite core changes the speed propagation inside a
waveguide.
Most phase shifters are of the digital variety, as they are more immune to noise on
their voltage control lines. Digital phase shifters provide a discrete set of phase states
that are controlled by two-state "phase bits."The highest order bit is 180 degrees,
the next highest is 90 degrees, then 45 degrees, etc., as 360 degrees is divided into
smaller and smaller binary steps. A three bit phase shifter would have a 45 degree
least signicant bit (LSB), while a six bit phase shifter would have a 5.625 least
signicant bit. These set of phases are produced in two ways, one posibility is to
obtained dierent values of L and C with discrete components such as capacitors
and inductors or by switching lines of dierent lengths which are commuted with
CMOS switches, MEMS or PIN diodes, this last group is know as well as true time
delays because they do not change the phase dierence between the outputs with
the frequency variance.
Scanning active antennas 39
2.3.1. Airborne steering antenna
The purpose of this work is to build a broadband airborne satellite communi-
cation system for low observability airplanes. The system will be able to transmit
and receive at X band MPEG-2 video format with IP-v.6 encapsulation under the
DVB-S2 standar.
The antenna developed for this system must full the recommendations ITU-R
S580-6 [2.38] and ITU-R S.465-5 [2.39]. The antenna consists of double stacked
printed elements grouped in an array, this terminal works in a frequency band from
7.25 up to 8.4 GHz (15% of bandwidth), where both bands transmission (7.9-8.4
GHz, RHCP) and reception (7.25-7.75 GHz, LHCP) are included simultaneously
(full duplex system). The antenna has a gain of 31 dBi to ensure 99.9% system
operation (the gain limit of the visible region is 28 dBi for θ=40o), and it has a
radiation pattern with a beamwidth smaler than 10o. The antenna operates with
dual circular polarization and it has the capability to steer in elevation from 90o to
40o electronically and 360o in azimuth with a motorized junction. A diagram of the
antenna can be seen in Fig. 2.11.
Figure 2.11: Airborne steering antenna for satellite communications [2.40].
A rectangular structure is selected formed by an array of 16x24 elements. These
elements are treated in two dierent ways:
Rows: First grouped in rows, separated a distance between them of 0.85λ|fminin order to increase the eective area and consequently the gain. In case of
40 STATE OF THE ART ON MULTIBEAM AND . . .
using an uniform excitation the side lobe levels exceed the limit of the recom-
mendations mentioned before. Due to that restriction, a progresive reduction
of the feeding amplitude from the centre to the edges of each row wil be im-
plemented with a passive unbalanced feeding network.
Columns: Thus, 16 rows of 24 elements are grouped vertically and separated
0.5λ|fmin. The steering direction of the main beam is achieved due to the
phase shifting feeding between each row. The terminal will be integrated in
the fuselage and because of that, it will be necessary to steer close to endre
locations. Therefore to avoid the closer angles to endre, where the directivity
is reduced, the planar antenna is raised. With the installation of a wedge of
30o the nal steering directions will be from 90o to 40o. The antenna nal
dimensions are 33 x 85 x 20 cm.
Radiating element
Each radiating element is composed of two stacked patches (as shown in Fig.
2.3.1): the upper one is fed by electromagnetic coupling (thickness substrate 1
=0.254 mm), meanwhile the bottom one (thickness substrate 2=1.143 mm) is fed
by two via holes (diameter=1.1 mm) to get the circular polarization. These vias are
connected with a two-stages miniaturized hybrid coupler [2.41] which enhances the
bandwidth of the circuit, permits two circular polarizations (RHCP and LHCP) at
the same time, and increases the isolation between both channels (TX and RX). The
two patches are separated by one foam layer (thickness=4 mm) in order to get higher
bandwidth (Fig. 2.3.1). In the bottom part (thickness substrate 3=0.254 mm), the
feeding distribution network and connectors are placed. It is necessary to remark
that specications and space constraints make the network design a challenging one,
since the hybrid couplers are especially shaped according to the available space. In
order to introduce the distribution lines, and 1 to 4 dividers, the double stage hy-
brid couplers were miniaturized. This miniaturization reduces the dimensions from
10.90 mm x 14.41 mm to 8.07 mm x 10.67 mm (26 %). Also the connection with the
radiating array has to be carefully dened (in this case by means of via holes). The
Scanning active antennas 41
(a) Radiating element 3D view. (b) Radiating element layer view.
Figure 2.12: Double stacked patch with 3dB/90o hybrid coupler.
(a) Transmiter. (b) Receiver.
Figure 2.13: Block diagram system.
substrate permittivity is 2.17 in order to get good radiation of the antenna. Despite
the reduction of the Q factor, the substrates and foam thickness are high in order
to enhance the bandwidth (∼ 15 %).
Active feeding network
The active feeding network is presented in Fig. 2.13(a) and Fig. 2.13(b). Each
subsystem will be the responsible for adapting the signal in amplitude, noise or
phase. In both cases, transmitter and receiver, the subsystem in charge for the
beam steering will be the phase shifters, these, can be digital or analogical, but in
both cases TTD (true-time-delay) in order to cover the whole frequency operation
band.
In the receiver system, another lter is added after the antenna in order to
minimize the power transmitted, coupled in the receiver system. The necessary
steering range in elevation goes from 10o to 60o above the horizon. However, the
plane in which the antenna is supported is tilted, thus, the nal steering range is from
40o to 90o. Electromagnetic simulations (in dashed line) and the analytic simulations
(continuous line) for the dierent steering directions are presented in Fig. 2.14. It
42 STATE OF THE ART ON MULTIBEAM AND . . .
can be seen that there is a directivity reduction between the perpendicular steering
direction (90o, or broadside) and the worst case (40o).
Figure 2.14: Radiation pattern of 16x1 array.
The feeding amplitude and phase coecients were calculated in Fig. 2.15 for
reducing the side lobe levels, and obtain the desired steering angles.
The most important decision in the active chain is the election of the phase
shifter. For this high performance communication system the noise gure it is very
important because it is desirable to have high data rates. In order to reduce the
noise oor it is necessary to have a phase shifter with the higher number of bits as
possible because those are directly related to the noise oor. Therefore a MAPS-
010166 of MA-COM with 6 bits and least signicant bit =5.625o, able to operate
from 8 to 12 GHz is choose. With this devide the continuous mean beam exploration
is ensured (there are no cross beam losses) and periodic phase feeding failures are
avoid since each row of the antenna is fed by a dierent phase shifter.
Passive feeding network
The antenna has two dierent distribution networks, the passive one which feed
each all the elements (or columns) inside a row and an active one which feeds every
row. The mask that the radiation pattern must full states that the side lobe level
must be 42 dB below the main beam directivity at ±48o for the azimuth plane.
Scanning active antennas 43
Figure 2.15: Amplitude and phase of the active network elements.
Therefore, enourmous tapering is design for the feeding distribution network. In
order to make feasible the construction of the transmision lines, the characteristic
impedance of the radiating elements is reduce to Z0 =25Ω.
In Fig. 2.16 the 1 to 4 element distribution network is presented. The unbalanced
power distribution between the output ports in order to obtain the rigorous azimuth
radiation pattern can be seen in the width of lines.
Figure 2.16: 1 to 4 unbalanced network divider.
The results of this circuit are presented in Fig. 2.3.1. The dierent output power
for each of the output ports is shown in Fig. 2.17(a). However, all the dividers
and transmission lines that reach each radiating element must have the exact same
lenght, so all the radiating elements are feed with the same phase Fig. 2.17(b).
In the same way the unbalanced 1 to 3 power divider was design to gather three
of the unbalanced 1 to 4 power divider. This group of power dividers will feed 12
elements, and by means of a balanced 1 to 2 power divider the rest of the elements
44 STATE OF THE ART ON MULTIBEAM AND . . .
are feed with the exact same network but with a transversal symmetry. Finally the
following coecients are obtained 2.3.
(a) Module. (b) Phase.
Figure 2.17: S parameters of 1 to 4 unbalance network divider.
Table 2.3: Amplitude coecients of the passive network.
Element Amplitude Element Amplitude
1 &24 0.299 7 & 18 0.682
2 &23 0.363 8 & 17 0.745
3 &22 0.427 9 & 16 0.809
4 &21 0.491 10 & 15 0.873
5 &20 0.555 11 & 14 0.936
6 &19 0.618 12 & 13 1
Scanning active antennas 45
(a) Quarter-wavelength transmis-
sion line T-model equivalent.
(b) Construction of a miniaturized
hybrid coupler.
Figure 2.18: 3 dB/90o Hybrid coupler miniaturization.
The available space in the passive feeding network is one problem for this kind of
antenna, because the vertical separation between the rows 0.5λ|7,25GHz is 20.68 mm
and to t the distribution network and the hybrid coupler is a hard issue. For this
reason, a miniaturized hybrid coupler was desing. Basically, the reduction consists
in the substitution of each line in the model of a branch line for its equivalent
quarter-wavelength transmission line of T-model [2.41].
A hybrid coupler was designed, simulated and constructed following these gui-
delines. Final dimensions are smaller than 8 mm x 10 mm (Fig.2.18(b))
Electromagnetic optimizations were carried out to improve the hybrid operation.
These simulations and measurements can be seen in Fig.2.19(a) (Module of S pa-
rameters) and Fig.2.19(b) (phase of S parameters). In Fig. 2.19(b) it can be seen
that the dierence between the phases S21 and S31 is ≈90o±5o.
The nal simulation results of the distribution network for the 1 x 24 elements
array is presented in Fig 2.20. The side lobe level slightly surpasses the mask,
however the average energy surpassed is low enough following the ITU-R S.732
[2.42].
46 STATE OF THE ART ON MULTIBEAM AND . . .
(a) Module. (b) Phase.
Figure 2.19: Hybrid coupler S parameters.
Figure 2.20: Radiation pattern of azimuth plane.
4x4 sub array construction
In this section a 4x4 element array for satellite communications in X band is
designed, constructed and measured to prove the feasibility of the system above
describe. This antenna is a subarray of a larger system which is compact, narrow
beamwidth and reaches a gain of 16 dBi. It has the capability to steer in elevation
to 45o, 75o, 105o and 135o electronically by means of a butler matrix (section 2.2).
Scanning active antennas 47
Fig.2.21(b) shows the bottom part of the construction of the 16 elements array. In
this gure, it can be seen the uniform feeding distribution network, the miniaturized
hybrid couplers and the connectors, which will be connected to the steering network.
If the Butler network is connected to left side ports, the antenna radiation will be
LHCP and vice versa. The rest of the ports will be loaded with 50Ω loads.
(a) Top view (b) Bottom view
Figure 2.21: 4x4 array array prototype [2.40].
Fig. 2.22 presents eight input ports of the sixteen elements array. The continuous
line show the left side ports, the dashed line represents the right side ports and the
marked line corresponds to the isolation between two ports together. The isolation
is proper enough with a value under 15 dB for the whole band (SLR < −15 dB) and
the reection coecient for all the ports is under -10 dB (Sii < −10 dB).
In Fig. 2.23(a) and Fig. 2.23(b) the radiation patterns at the center frequency of
each band (7.5 GHz and 8.15 GHz) for every steering angle are shown. The steering
angles −45, −15, +15 and +45 correspond to phase dierence feed α = 135,
α = 45, α = −45 and α = −135 respectively. The continuous line presents
the measured data, while the dashed one shows simulated data, which are in good
agreement.
It can be seen in both gures (Fig. 2.23(a) and Fig. 2.23(b)) a gain reduction
in the main beam far from the broadside direction as it was previously shown in
section. This is reasonable due to the main beam widening. The level dierences
between measurements (continuous line) and simulations (dashed line) are because
of the fact that loss tangent in the dielectric substrate, cables and connectors were
48 STATE OF THE ART ON MULTIBEAM AND . . .
Figure 2.22: S parameters measurements of 4x4 array prototype (8 ports).
(a) LHCP 7.5 GHz. (b) RHCP 8.15 GHz.
Figure 2.23: Steering radiation pattern of 4x4 subarray.
not accurately taken into account.
The integration of all the necessary elements, passive networks and active devices
to full the system requirements in such a small space is a hard task and trade o
must be done. After a large amount of simulations and analytical calculations, this
proposed design full all the requirements and it yields good performance.
2.4. Conclusions
This chapter gives a theoretical and practical overview of the factors that have to
be taken into account in the array exploration design process. This in an important
Conclusions 49
starting point to understand thoroughly the basics of this antennas, to look for the
principal contributions that have been already made, and what it is more important
to identify the elds in which some new contributions can be made. A design of an
steering antenna for a satellite communications in X band is carried out. In this
system the link budget requirements and antenna specications are very demanding
and state of the art technologies and techniques are requided in order to full the
high performance in terms of radiation eciency and reduced available space.
One of the detected problems in printed antennas is the mutual coupling bet-
ween the radiating elements when a phase shift (α=0o) is applied to get the steering
direction. It has been observed that the surface wave modes along the substrate are
enhanced when the radiating elements are not fed in phase. To avoid this problem
the use of cavity patch antennas is commonly utilized, however, this technique in-
creases the weight of the system, the mounting process becomes more dicult and
therefore the cost rises. To solve this problem without increasing the complexity of
the antenna fabrication the use of Electromagnitic Band Gap (EBG) metamaterials
is explored. By placing these structures in between the radiating elements the sur-
face wave modes along the substrate are suppressed, obtaining therefore a similar
eect to cavity back radiating elements.
Another problem of the phased array antennas resides in their distribution fee-
ding networks. There are several kind of phase shifters, and the election of the
devices and the circuit topologies is directly related to the return and insertion
losses and the isolation between radiating elements. Therefore, several designs com-
bining topologies and technologies are proposed and developed in order to obtain
dierent properties in the feeding distribution network. Features such as match or
unmatch ports, maximum isolation between ports, balanced or unbalanced power
distribution and of course phase dierence.
References
[2.1] H. Iizuka, P.S. Hall, S. Sugiura, and K. Sato. Varactor-loaded H-shaped an-
tenna with radiation pattern control. Antennas and Propagation, IEEE Tran-
sactions on, 56(9):2833 2840, Sep. 2008.
[2.2] V.B. Khaikin, S.V. Kaplaya, M-H. Ghung, V.N. Radzikhovsky, and S.E. Kuz-
min. Simulated characteristics of trao 14-m mm telescope with multi-beam
FPA at 85-115 GHz. Gravitation and Cosmology, (1-2):155160, 2005.
[2.3] U.R. Kraft. Gain and G/T of recongurable active receive antennas-a parame-
tric approach. In Antennas and Propagation Society International Symposium,
1999. IEEE, volume 4, pages 2250 2253, Aug. 1999.
[2.4] F. Kira and T. Hori. Beam forming network design for cluster feeding of
scanning antenna. In Phased Array Systems and Technology, 2000. Proceedings.
2000 IEEE International Conference on, pages 133 136, 2000.
[2.5] Xue-Xia Yang, Bing-Cheng Shao, Fan Yang, A.Z. Elsherbeni, and Bo Gong. A
polarization recongurable patch antenna with loop slots on the ground plane.
Antennas and Wireless Propagation Letters, IEEE, 11:69 72, 2012.
[2.6] M.S. Nishamol, V.P. Sarin, D. Tony, C.K. Aanandan, P. Mohanan, and K. Va-
sudevan. An electronically recongurable microstrip antenna with switchable
slots for polarization diversity. Antennas and Propagation, IEEE Transactions
on, 59(9):3424 3427, Sep. 2011.
[2.7] Wen-Jiao Liao, Sheng-Jie You, and Hsi-Tseng Chou. Polarization recongu-
rable patch array antenna. In Wireless Information Technology and Systems
(ICWITS), 2010 IEEE International Conference on, pages 1 4, 28 2010-sept.
3 2010.
[2.8] J.A. Ruiz-Cruz, M.M. Fahmi, S.A. Fouladi, and R.R. Mansour. Waveguide
antenna feeders with integrated recongurable dual circular polarization. Mi-
51
52 REFERENCES
crowave Theory and Techniques, IEEE Transactions on, 59(12):3365 3374,
Dec. 2011.
[2.9] G. Expósito-Domínguez, J. M. Fernández-González, P. Padilla, and M. Sierra
Castañer. Dual circular polarized steering antenna for satellite communications
in x band. Progress In Electromagnetic Research, 122:6176, 2012.
[2.10] C. Kalialakis and P.S. Hall. Analysis and experiment on harmonic radia-
tion and frequency tuning of varactor-loaded microstrip antennas. Microwaves,
Antennas Propagation, IET, 1(2):527 535, Apr. 2007.
[2.11] A. Mirkamali and P.S. Hall. Log periodic printed dipole array for wideband
frequency reconguration. In Wideband, Multiband Antennas and Arrays for
Defence or Civil Applications, 2008 Institution of Engineering and Technology
Seminar on, pages 95 110, Mar. 2008.
[2.12] F. Ghanem, P.S. Hall, and J.R. Kelly. Two port frequency recongurable
antenna for cognitive radios. Electronics Letters, 45(11):534 536, 21 2009.
[2.13] Z. H. Hu, J. Kelly, C. T. P. Song, P. S. Hall, and P. Gardner. Novel wide tu-
nable dual-band recongurable chassis-antenna for future mobile terminals. In
Antennas and Propagation (EuCAP), 2010 Proceedings of the Fourth European
Conference on, pages 1 5, Apr. 2010.
[2.14] J. Butler and R. Lowe. Beam-forming matrix simplies design of electronically
scanned antennas. Electronic Design, 9:170173, 1961.
[2.15] J. Shelton and K. Kelleher. Multiple beams from linear arrays. Antennas and
Propagation, IRE Transactions on, 9(2):154161, Mar. 1961.
[2.16] W.-P. Delaney. An RF multiple beam-forming technique. Military Electronics,
IRE Transactions on, MIL-6(2):179 186, Apr. 1962.
[2.17] M. Koubeissi, C. Decroze, T. Monediere, and B. Jecko. Switched-beam anten-
na based on novel design of Butler matrices with broadside beam. Electronics
Letters, 41(20):1097 1098, Sep. 2005.
REFERENCES 53
[2.18] W. White. Pattern limitations in multiple-beam antennas. Antennas and
Propagation, IRE Transactions on, 10(4):430 436, Jul. 1962.
[2.19] D.-C. Chang and S.-H. Jou. The study of Butler matrix BFN for four beams
antenna system. In Antennas and Propagation Society International Sympo-
sium, 2003. IEEE, volume 4, pages 176 179, Jun. 2003.
[2.20] J.P. Shelton. Multibeam planar arrays. Proceedings of the IEEE, 56(11):1818
1821, Nov. 1968.
[2.21] J. Shelton and J. Hsiao. Reective Butler matrices. Antennas and Propaga-
tion, IEEE Transactions on, 27(5):651 659, Sep. 1979.
[2.22] J. Blass. Multidirectional antenna - a new approach to stacked beams. In
IRE International Convention Record, volume 8, pages 4850, Mar. 1960.
[2.23] P.S. Hall and S.J. Vetterlein. Review of radio frequency beamforming tech-
niques for scanned and multiple beam antennas. Microwaves, Antennas and
Propagation, IEE Proceedings H, 137(5):293303, Oct. 1990.
[2.24] M. Fassett, L. Kaplan, and J. Pozgay. Optimal synthesis of ladder network
array antenna feed systems. In Antennas and Propagation Society International
Symposium, 1976, volume 14, pages 58 61, Oct. 1976.
[2.25] S. Mosca, F. Bilotti, A. Toscano, and L. Vegni. A novel design method
for Blass matrix beam-forming networks. Antennas and Propagation, IEEE
Transactions on, 50(2):225 232, Feb. 2002.
[2.26] S.J. Vetterlein and P.S. Hall. Novel multiple beam microstrip patch array
with integrated beamformer. Electronics Letters, 25(17):11491150, 1989.
[2.27] P. J. Wood. An ecient matrix feed for an array generating overlapped beams.
pages 371374, Mar. 1987.
[2.28] J Nolen. Synthesis of multiple beam networks for arbitrary illuminations. PhD
thesis, Baltimore MD., Apr. 1965.
54 REFERENCES
[2.29] N.J.G. Fonseca. Printed s-band 4 × 4 Nolen matrix for multiple beam antenna
applications. Antennas and Propagation, IEEE Transactions on, 57(6):1673
1678, june 2009.
[2.30] J. Herd and D. Pozar. Design of a microstrip antenna array fed by a Rotman
lens. In Antennas and Propagation Society International Symposium, 1984,
volume 22, pages 729 732, Jun. 1984.
[2.31] W. Rotman and R. Turner. Wide-angle microwave lens for line source ap-
plications. Antennas and Propagation, IEEE Transactions on, 11(6):623632,
Nov. 1963.
[2.32] J. Ruze. Wide-angle metal-plate optics. Proceedings of the IRE, 38(1):5359,
Jan. 1950.
[2.33] T.A. Denidni and T.E. Libar. Wide band four-port Butler matrix for switched
multibeam antenna arrays. In Personal, Indoor and Mobile Radio Communi-
cations, 2003. PIMRC 2003. 14th IEEE Proceedings on, volume 3, pages 2461
2464, Sep. 2003.
[2.34] S.J. Foti and T. MacNamara. Design of wideband Butler matrices using
Schiman lines. InMultiple Beam Antennas and Beamformers, IEE Colloquium
on, pages 5/1 5/8, Nov. 1989.
[2.35] D. M. Pozar and D. H. Schaubert. Microstrip Antennas: The analysis and
design of microstrip antennas and arrays. IEEE press, 1995.
[2.36] G. Matthaei and L. Young. Microwave Filters, Impedance-Matching Net-
works, and Coupling Structures. Artech House Publishers, 2000.
[2.37] G. Expósito-Domínguez, J. M. Fernández-González, P. Padilla-Torre, and
M. Sierra-Castañer. Matriz de Butler de banda ancha en banda x para antenas
recongurables. XXV URSI Spain, Sep. 2011.
REFERENCES 55
[2.38] ITU-R. Radiation diagrams for use as design objetives for antennas of earth
station operating with geostacionary satellites. Recommendation 580-6, Inter-
national Telecommunication Union, 2004.
[2.39] ITU-R. Reference earth-station radiation pattern for use in coordination
and interference assessment in the frequency range from 2 to about 30 ghz.
Recommendation 465-5, International Telecommunication Union, 1993.
[2.40] G. Expósito-Domínguez, P. Padilla-Torre, J. M. Fernández-González, and
M. Sierra-Castañer. Electronic steering antenna onboard for satellite commu-
nications in x band. 5th European Conference on Antennas and Propagation
(EuCAP 2011), pages 21202123, Apr. 2011.
[2.41] C. W. Tang and M. G. Chen. Synthesizing microstrip branch-line couplers
with predetermined compact size and bandwidth. IEEE Trans on Microwave
Theory and Techniques, 55(9):19261934, 2007.
[2.42] ITU-R. Method for statistical processing of earth station antenna side-lobe
peaks to determine excess over antenna. Recommendation 732, International
Telecommunication Union, 2012.
.
Chapter 3
Mutual coupling reduction using
EBG in steering antennas
The behaviour of radiating elements varies when they are placed surrounded
by other elements. This eect is due to some amount of energy transfered between
them whether one and/or several are transmiting or receiving. The amount depends
basically on the radiation characteristics of each, relative separation between them
and the relative orientation of each. There are many dierent mechanisms that can
cause this interchange of energy. For example, even if both antennas are transmit-
ting, some of the energy radiated from each will be received by the other because of
the non-ideal directional characteristcs of practical antennas. Part of the incident
energy on one or both antennas may be re-scatered in dierent directions allowing
them to behave as secondary transmitters. This interchange of energy is known as
MUTUAL COUPLING, and in many cases it complicates the analysis and design
of an antenna.
3.1. Introduction
In microstrip antennas, not only the radiating elements contribute to the mutual
coupling, but also the feeding networks and feeding probes (if exist). These antennas
are printed in a thin layer of metal (18-35µm) on a substrate by chemical process or
laser etching. The relative permittivity εr and the thickness of the substrate are very
important because they change the size and the properties of the radiating elements.
It is desirable to use thick substrates to increase the operation bandwidth and high
εr to reduce the size of the printed antennas. However, by using this strategy, the
57
58 MUTUAL COUPLING REDUCTION USING EBG IN . . .
surface wave propagation modes are enhanced because of the thick substrate, and
the electromagnetic elds instead of being radiated, are conned in the substrate
due to the high εr which lead to a stronger interaction and therefore, higher mutual
coupling between radiating elements.
This undesired energy transfer characteristic causes several damages in the ra-
diation pattern of the antenna such as directivity drop, power losses and higher
side lobe level (SLL) and crosspolar (XP) component.The most common mutual
coupling reduction techniques are cavity structures [3.1], non uniform feeding distri-
bution [3.2], unequally space distribution [3.3] and Defected Ground Plane (DGP)
[3.4], but lately Electromagnetic Band Gap (EBG) structures are being used. The
introduction of several rows of EBG structures between two printed antennas has
been proved to increase the isolation [3.5].
EBG structures are usually realized by periodic arrangement of dielectric mate-
rials and metallic conductors. In general, they can be categorized into three groups
according to their geometric conguration: three-dimensional transmission lines,
two-dimensional planar surfaces and one-dimensional transmission lines. This work
is focused on the 2-D EBG surfaces, which have the advantages of low prole, light
weight, low fabrication cost, and are widely considered in antenna engineering.
Among others, EBG structures are used as ground planes emulating Articial
Magnetic Conductors (AMC) in a small frequency range and they are also used to
reduce the mutual coupling between elements.
EBG as Articial Magnetic Conductor (AMC) structure has one important
feature: the in-phase reection coecient for plane waves [3.6]. This property
can be used to design low-prole wire antennas [3.7]. The low-prole design
usually refers to the antenna structures whose height is less than one-tenth.
In chapter 12 of [3.8], a comparison between PEC, PMC, and EBG ground
planes is carried out. The conclusions are drawn in table 3.1.
In [3.9], EBG ground plane is used to reduce the eect of multipath by blocking
the propagation of surface waves in a low prole GALILEO antenna. Its
performance is similar to a classical choke ring antenna.
Introduction 59
Table 3.1: Comparison of PEC, PMC, and EBG ground planes for low prole an-
tenna designs.
Ground plane Reection phase Comments
PEC 180 Reverse image.
PMC 0 In-phase reection.
EBG from 180 to −180 with frequency Suitable frequency band.
as Filters, thanks to its frequency selective feature, this EBG structure can
be used as a lter. By introducing three elements as a ground plane for a
microstrip line, in [3.10] isolation higher than 55 dB in the rst frequency
band (2.1 GHz) and 40 dB in the second one (2.45 GHz) are achieved.
as Mutual Coupling barriers: Patch antennas are found to have very strong
mutual coupling due to the severe surface waves on thick and high permitti-
vity substrates. In the literature, it can be found a variety of works which
apply metamaterials for the reduction of this eect. In [3.5], four rows of EBG
mushrooms are inserted between the patch antennas in a εr=10.2 and thick-
ness (h=2 mm) substrate. With this conguration, 8 dB of mutual coupling
reduction is obtained. In [3.11], dierent substrates are combined: radia-
ting elements are suspended over a thick foam layer in order to increase the
bandwith, meanwhile EBG structures are printed in a thin high permitti-
vity substrate for size reduction and surface wave suppression. In [3.12], by
using edge-located vias, the size of mushroom-type EBG is reduced by 20 %.
Among other strategies, in [3.13], a fork shape is used. The area occupied by
the fork-like structure is less than 25% of the mushroom-like structure. Be-
sides, MicroElectroMechanical Systems (MEMS) are used, and recongurable
stop band is obtained. Another studied technique is the use of metal strips.
Basically, the idea is to combine the EBG concept with soft surfaces. A com-
parison of mushroom-type EBG surfaces and corrugated and strip-type soft
60 MUTUAL COUPLING REDUCTION USING EBG IN . . .
surfaces is shown in [3.14]. This stripe-type is used in [3.15] and in [3.16] to
reduce mutual coupling. Finally, dual band planar soft surfaces are developed
in [3.17]: two sizes of strips are mixed in order to get dual forbidden band.
In this chapter, EBG structures as electromagnetic barriers for mutual coupling
reduction are studied. In order to solve the problem for printed radiating elements
which are very close, a new type of EBG is developed. By combining several tech-
niques the size of these structures is reduced meanwhile the barrier characteristics
are preserved.
3.2. EBG theory fundamentals
EBG based on Frequency Selective Surfaces (FSS) [3.18] are one type of meta-
materials with particular electrical properties [3.19]. EBG technique appears as an
application of truncated frequency selective surface (FSS) [3.20]. These structures
consist of an array of metal protrusions on a at metal sheet and can be visualized
as mushrooms protruding from the surface as shown in Fig.3.1. When the period p
is small compared to the wavelength of interest (p << λ), it is possible to analyze
the material as an eective medium, with a surface impedance. These mushrooms
present very high impedance for vertical an horizontal modes at certain frequencies.
The behavior of this structure is similar to a LC circuit, eq. 3.1. Below the reso-
nance frequency, the surface is inductive and supports TM modes meanwhile above
resonance frequency, the surface is capacitive and TE surface waves are supported.
ω0 =1√LC
(3.1)
Nearby ω0, the surface impedance (Zs) is much higher than the impedance of free
space (Z0=120π), as eq. 3.2 depicts. Therefore, no vertical or horizontal propagation
modes are allowed. The high impedance surface also ensures that a plane wave will
be reected without the phase reversal that occurs on a perfect electric conductor
(PEC).
EBG theory fundamentals 61
Figure 3.1: High impedance surface and its model with parallel resonant LC circuit.
The substrate is transparent in order to get better visualization of metallic vias.
(a) EBG parameters. (b) LC model.
Figure 3.2: LC model for the mushroom like EBG structure.
Zs =jωL
1− ω2LC(3.2)
The value of the capacitor is given by the fringing capacitance between neigh-
boring coplanar metal plates. This can be derived using conformal mapping, this
technique establish two-dimensional electrostatic eld distributions. The derivation
starts with a pair of semi-innite plates separated by a gap and then truncates them
with a nite patch size. Finally, the edge capacitance for the the proximity of the
metal plates C, according to [3.20] is given in eq. 3.3:
C =
[wε0(εeff )
π
]cosh−1
(2w
g0
)(3.3)
The value of the inductor is derived from the current loop in Fig. 3.2, consisting
62 MUTUAL COUPLING REDUCTION USING EBG IN . . .
of the vias and metal sheets. For a solenoid current, the magnetic eld can be
calculated using Ampere's law. The equivalent inductor is then computed from the
stored magnetic eld energy and the excitation current. After a simple derivation,
the inductance L is expressed as eq. 3.4 which depends only on the thickness t of
the structure and the permeability of the free space µ0:
L = µ0t (3.4)
In [3.20], Yang et al. propose the analitical solution when the radius of the via
is not r << t eq. 3.5.
L = 2 ∗ 10−7t
[ln
(2t
r
)+ 0,5
(2r
t
)− 0,75
](3.5)
Substituting (3.3) and (3.4) or (3.5), into (3.1) and (3.2), the surface impedance
and resonant frequency can be computed. Other characteristic parameters, such
as the reection phase, can also be derived accordingly. This LC model is easy to
understand, but the static eld approximations limit its accuracy.
Analytically, it can be noticed that the band gap operation is proportional to√L/C. Thus, for a xed separation of the metal plates the value of the capacitance
is set, and the working bandwidth can be improved by increasing the length of the
via and therefore the thickness of the substrate, which is adequate for wideband
antennas. On the other hand, when the thickness of the substrate is xed and the
gap size can not be reduced, C can be increased by using multilayer circuit boards
with overlapping plates, in order to decrease the frequency. In Fig. 3.3, on the left
hand side, traditional mushrooms are shown, meanwhile on the right hand side a
multilayered mushroom structure is presented.
Finally, EGB structures can be seen as bend corrugations, where the length of
the corrugation is the sum of the via and the patch two times, as it is depicted
in Fig. 3.3. The band gap operation of the corrugations starts and nishes when
the length is similar to λ/4 and λ/2 respectively, which yields an octave regarding
operation frequency band [3.21].
Surface wave supression 63
Figure 3.3: Layer and top views of traditional EBG structures (left hand side) and
multilayered F structure (right hand side).
3.3. Surface wave supression
Thanks to the high impedance surface, horizontal or vertical modes are not
allowed in the mushroom structures at certain frequencies. In order to nd the
allowed frequencies for each wave vector, a single unit cell with periodic conditions
is simulated with CST and the Eigenmode solver. The simulation is carried out
by means of a variable sweep which goes across the unit cell (from the x point to
m and p) in the two dimensions x and y. The x-axis of the graphs represents the
dimensions inside the unit cell (x-m-p) in which the boundary conditions are fullled
at certain frequencies for the transmission modes of the structure. In Fig. 3.4(a)
and Fig. 3.4(b) the unit cell is symmetric respect to the via, thus, only three traces
have to be computed. However, for the F and H shape cases in Fig. 3.4(c) and Fig.
3.4(d) respectively, the EBG structure does not have the same dimensions in x and
y. Therefore, ve traces have to be analyzed to describe all the possible propagation
modes for these boundary conditions.
All the studied structures are developed in a low permittivity substrate εr=2.17
and thickness of 1.143 mm. For the traditional mushroom, the size of the original
patch in X band (7.25 - 8.4 GHz) is 3 mm x 3 mm, on the other hand for multilayer
mushrooms 2.8 mm x 2.8 mm size is used. For this last conguration and the rest of
multilayer congurations, two stacked substrates of 0.762 mm and 0.381 mm thick
64 MUTUAL COUPLING REDUCTION USING EBG IN . . .
(a) Original mushroom. (b) Double layer mushroom.
(c) Double layer and edge-location
via mushroom (F shape).
(d) Double layer interdigitated
mushroom (H shape).
Figure 3.4: Unit cell scheme for eigenmode solutions. Dimensions in mm.
are used.
The electric eld is described in terms of an eigenvalue equation, which is solved
numerically. In Fig. 3.5(a), Fig. 3.5(b), Fig. 3.5(c) and Fig. 3.5(d) the mode solu-
tions that satisfy the boundary conditions are shown. The abscissa value represents
the wave number which fulls the requirements at a certain frequency. The lowest
line is the TM mode, the second and third lines are TE modes. A frequency band
gap (dashed lines), in which the surface does not support surface wave propagation
of either polarization, horizontal nor vertical, extends from the top of the TM band
to the point where TE band crosses the light line. The comparison between original
shape and double layer mushrooms yields the result of 2 GHz bandwidth for the
original shape and 1.2 GHz for the multilayer solution.
Surface wave supression 65
(a) Original mushroom. (b) Double layer mushroom.
(c) Double layer and edge-location via mush-
room (F shape).
(d) Double layer interdigitated mushroom (H
shape).
Figure 3.5: Brillouin diagrams of the EBG unit cell.
In this work, the combination of multilayer structure and edge-location via [3.22]
for mushroom size reduction is discussed. In order to keep the frequency working
band (15%), and radiation eciency, the same substrate (εr=2.17 and thickness of
1.143 mm) is used. As long as the inductance depends on the substrate thickness,
which is xed already, the only available parameter is the capacitance C. In order
to increase this parameter, a multilayered structure with edge-location [3.23] via is
presented. The dimensions of the patches are 2.1 mm x 3.6 mm (this value means
30% size reduction). The Brillouin diagram for this conguration is presented in Fig.
3.5(c). In this case, the bandwidth remains its value but signicant size reduction is
noticed. Another dierent attempt to reduce the size of the mushroom is to increase
the capacitance by constructing H shape mushrooms. The main idea is to avoid the
66 MUTUAL COUPLING REDUCTION USING EBG IN . . .
multilayered congurations by using interdigitated elements. However, due to the
construction limits and frequency of operation, this solution is not feasible (Fig.
3.5(d)).
3.4. Mutual coupling reduction
The size of the mushroom patches and the necessary number of periods for mu-
tual coupling reduction is higher than the available space between radiating elements
for low permittivity substrates. To study this space constraint, a simulation scheme
for S21 analysis is proposed in Fig. 3.6. In this scheme, a 50 Ω transmission line is
placed over an EBG ground plane. In this simulation, dierent topologies (original,
double or F shape), size of the mushroom w, gap size g and number of elements n
are tested in order to obtain the surface wave suppression behavior.
Figure 3.6: Simulation scheme for transmission parameters S21 analysis.
In order to keep the study in the same frequency operation band, the thickness
of the substrate is the same that in the previous section (t=1.143 mm) and the
gap size is x to (g=0.2 mm) to reduce the space. In Fig. 3.7(a), a comparison
between the mushroom size w and the working operation band is shown. It can be
noticed that an increase of operation bandwidth is obtained for higher frequencies
Mutual coupling reduction 67
when L is xed and C is reduced because of w reduction, as it was forementioned
in section 3.2. In the same way, a study of the required number of elements n is
carried out and the results are drawn in Fig. 3.7(b). Now, with the mushroom size
x to w=3 mm, dierent number of periods are placed. In the original study seven
elements are used, when this number is reduced, it can be observed how the isolation
characteristic decrease.
(a) Isolation S21 for dierent EBG sizes (w). (b) Isolation S21 for dierent number of EBG
periods (n).
Figure 3.7: Parametric study of isolation characteristics (S21 for original shape mush-
rooms) when size and number of elements are swept.
Other simulations are done maintaining the xed values of n=4 and w=3 mm.
In Fig. 3.8(a) a swept of the substrate thickness is shown. In this case of study
the values for g and dvia are 0.2 mm and 0.4 mm respectively. It can be noticed
that the behavior of resonance frequency tends to move to a higher values when the
substrate is thinner. The reason for that behavior is the dependance of L with the
thickness of the substrate (as it is explained in eq. 3.4). With values of t=1.143
mm and g=0.2 mm a variation of the via diameter from 0.2 mm to 1.2 mm of the
mushroom structure is carried out in Fig. 3.8(b). With this parameter variation,
the eq. 3.5 and the dependence of the resonance frequency with the via diameter
of the structure can be proved. By examining this gure, it can be noticed that for
dvia=0.2 mm the bandwidth operation is roughly 1.5 GHz meanwhile for dvia=1.2
mm, the bandwidth operation is increased to ≈2 GHz. At the end of section 3.2 it
is stated that the bandwidth operation is proportional to√L/C. However, here,
68 MUTUAL COUPLING REDUCTION USING EBG IN . . .
through these simulations L is changed and this property regarding the bandwidth
is proved. Eventually, in Fig. 3.8(c), the variation response of S21 transmission
parameter is shown. Just like in the other gures, a frequency operation change can
be observed. This change is due to the value of C in the mushroom structure, which
is proportional to the gap between the parallel plates. Therefore, when the gap size
increases, C decreases, and the band gap appears at higher frequencies.
(a) Isolation S21 for dierent substrate thick-
ness (t).
(b) Isolation S21 for dierent via diameter
(dvia).
(c) Isolation S21 for dierent gap size (g).
Figure 3.8: Parametric study of isolation characteristics (S21 for original shape mush-
rooms) when substrate thickness, via diameter and gap size are sweept.
All the substrates have permittivity of εr=2.17. However four dierent thickness
values are used. The TLs impedance is 50 Ω and those TLs are printed in a 0.254
mm thick substrate. Mushrooms in single layer case are printed in a 1.143 mm thick
substrate, meanwhile double layer case are printed in 0.762 mm (bottom layer) and
Mutual coupling reduction 69
0.381 mm thick (upper layer) substrates. Therefore, total thickness value maintains
its value as it can be seen in Fig. 3.12(a).
F shape solution is deeply developed and a parametric study regarding the sizes
of the mushrooms is carried out. In Fig. 3.9 the deviation of the isolation response
is shown. When the size is very small, the band gap tends to appear in higher
frequencies, and it can be seen how the sum of both dimensions, W and L must
remain under certain values.
Figure 3.9: Parametric study of width W and length L for F shape mushrooms.
Transmission S parameters (S21).
Figure 3.10: Samples of single and multilayered EBG mushrooms with dierent
shapes and number of elements.
In order to validate the whole process, prototypes of four and seven rows are
70 MUTUAL COUPLING REDUCTION USING EBG IN . . .
built. Six prototype Transmission Lines (TL) with EBG ground plane are shown in
Fig. 3.10. On the left hand side, the two circuits are single layered, the two circuits
in the middle are double layered, and the last two circuits on the right hand side
combine double layer with edge-location via (F-shape).
(a) Original shape mushroom. (b) Double layer mushroom.
(c) F shape mushroom.
Figure 3.11: Comparison between measurements and simulations of transmission
parameters S21 for dierent types of EBG mushrooms.
In Fig. 3.11(a), Fig. 3.11(b) and Fig. 3.11(c), comparisons between the simula-
ted structures and the measured prototypes are shown. Ensembling diculties due
to the small size of the circuits (size circuit approximately 1 cm) add some dieren-
ces between simulations and measurements. The fabrication process in PTFE, such
as metal vias and stack up process is a hard task. However, the overall behavior
of a LC lter in a certain frequency band is observed. It can be noticed, compa-
ring the gures, that traditional mushrooms have larger frequency operation band
Mutual coupling reduction 71
than F shape mushrooms. However, F shape fulls the bandwidth and isolation
requirements.
Finally, the trade o solution between isolation and available space is carried
out, being the necessary number of periods for surface wave suppression: n=4. In
Fig. 3.12(a) and Fig. 3.12(b), the chosen topology is shown. With nal dimensions
of 2.1 x 3.6 mm, four elements, the double layered structure and edge-location via
fulls the requirements of available space (10.2 mm) between two printed antennas
separated 0.6λ0 in a εr=2.17 and 1.143 mm thick substrate. A bandwidth operation
in X band, from 7.1 GHz to 8.2 GHz (15%) and 10 dB of isolation are obtained.
(a) Schematic (b) Prototype
Figure 3.12: Multilayered mushroom with rectangular shape, 4 elements and edge-
located via (F-shape).
In order to prove the eectiveness of these electromagnetic barriers, four rows of
double layered edge-located via EBG mushrooms are introduced between two round
patches with double circular polarization. The radiating elements are integrated
in the same substrate and they are circular polarized. The elements are fed by a
90/3dB branch line coupler, in order to get the double circular polarization.
In Fig. 3.13(a), |E| eld simulation for the two patches is shown for Left Handed
Circular Polarization (LHCP). In Fig. 3.13(b), it can be seen graphically, how |E|
eld value decays quicker when using double layer edge-location via EBG structures.
72 MUTUAL COUPLING REDUCTION USING EBG IN . . .
(a) Without EBG structures (b) With EBG structures
Figure 3.13: |E| eld simulation of two round patches with dual circular polarization.
Thus, two circular patches fed by the 90/3dB branch line coupler with and
without four rows of EBG F-shape mushroom are built. These printed antennas, se-
parated 0.6λ0, are mounted with double stacked patch with permitivity of εr=2.17
and a foam layer between them of 4 mm in order to cover the whole bandwidth
(20%). The thicknesses of the substrates are 1.143 mm for the bottom patch and
0.254 mm for the upper patch and the feeding network. Fig. 3.14 shows the cons-
truction process of both test antennas.
Figure 3.14: 2x1 test array of circular patch antennas with and without EBG F
shape mushrooms.
The radius for the via fed and the coupled fed patches are 7.31 mm and 7.17 mm
Mutual coupling reduction 73
Figure 3.15: Specic dimensions of 2x1 test array.
respectively as it can be seen in Fig. 3.15. With these prototypes, a comparison
between simulations and measurements of S21 parameters for a 2x1 array antenna is
presented in Fig. 3.16(b). LHCP ports are excited and RHCP ports are loaded with
50 Ω loads. Similar behavior between simulations and measurements is achieved.
There is a mutual coupling reduction between the two patch antennas of approxi-
mately 5 dB and an improvement of reection coecient Sii of ∼3 dB in most of the
operation band. This coupling reduction remains for the dierent phase feedings of
the elements.
(a) Matching S11 parameters (b) Isolation S21 parameters
Figure 3.16: Comparison of measurements and simulated S parameters for 2x1 array.
In Fig. 3.17, the radiation patterns for φ cuts are presented. It can be noticed
that for the 90 pattern, the beamwidth is reduced for the EBG case. EBG structures
surrounding microstrip antennas tend to make narrower the radiation pattern of
74 MUTUAL COUPLING REDUCTION USING EBG IN . . .
the radiating element. This beamwidth reduction increases the directivity of the
antenna, since the eect of the EGB barriers is similar to the eect of cavities.
On the other hand, due to the structures placed on the edge of the antenna, a ∼5
dB of back lobe reduction is obtained. The radiation pattern for φ=90 presents
asymmetry due to the measurement setup in the anechoic chamber.
Figure 3.17: Radiation patterns of 2x1 test patch antenna array with and without
EBG structures.
The maximum separation between elements for avoiding grating lobes in steering
antennas is a double problem. First, by placing the radiating elements very close,
|E| eld interaction between radiating elements is stronger. In the second place, the
eective area of the antenna is reduced. For multimedia applications, broadband
capability is required, and thick substrate is used. Thus, the surface wave propaga-
tion modes are enhanced and the mutual coupling between elements grows. For that
reason, the steerable antenna prototype from [3.24] has been built to measure the
radiation patterns and to verify the features and performances when mushroom F
shape structures are introduced between elements in order to suppress surface wave
propagation and, consequently, to reduce mutual coupling.
Applications 75
3.5. Applications
In this particular electronic steering antenna, the separation between radiating
elements must not be higher than 0.6λ0 in order to avoid grating lobes in the visible
region when the beam is not pointed to broadside, as it was explained in Chapter
2. In this section, a module of a wideband planar array antenna with dual circular
polarization (LHCP and RHCP) and electrical elevation steering for satellite com-
munication systems is provided (Table 3.2). In order to reduce the mutual coupling
between radiating elements, EBG structures are introduced.
Table 3.2: Antenna specications.
Parameter Value Units
Frequency RX 7.25 − 7.75 GHz.
Frequency TX 7.9 − 8.4 GHz.
Polarization RX LHCP* * Interchangeable.
Polarization TX RHCP*
Gain ∼16 dBi.
Elevation steering ±10 and ±40 Degrees
Dimensions <0.2 m.
Antenna Eciency >60 %.
Axial Ratio <3 dB.
CP/XP >25 dB.
Matching Sii >13 dB.
Isolation Sij >15 dB.
For the antenna, it is important to have good radiation eciency, therefore, a
εr=2.17 substrate is used. The available space between radiating elements under
this circumstance is not enough to place original mushroom shape EBGs. Therefore,
76 MUTUAL COUPLING REDUCTION USING EBG IN . . .
multilayer and edge location via techniques are combined to reduce the size of the
mushrooms by 30% in the vertical direction where the elements get the steering
elevation.
In Fig. 3.18, the printed array layers are presented. In the bottom layer (0.254
mm thick) the feeding distribution network and 90/3dB hybrid couplers are printed.
In the second layer (0.762 mm thick) the rst mushrooms and the vias for the patches
are placed. On top of it, the third layer (0.381 mm thick) contains the via fed circular
patches and the second mushrooms, which are stacked with the rst mushrooms in
the previous layer. On the top layer (0.254mm thick) parasitic fed patches are
printed. Finally, between via feed patches and parasitic fed patches a 4 mm foam
layer is placed in order to enhance the bandwidth [3.25]. The distance between the
radiating elements is 0.6λ0 for vertical axis in order to avoid grating lobes when
electronic steering is used. However for horizontal axis 0.85λ0 is used in order to
reduce the number of elements and to get the required directivity.
Figure 3.18: Layer view of the 4x4 array with EBG structures construction.
In Fig. 3.19(a) and Fig. 3.19(b) the measurements for eight input ports of
the sixteen elements array are presented. The continuous lines show the left side
ports, meanwhile the dashed lines represent the right side ports. The dotted lines
Applications 77
correspond to the isolation between two closer ports. The measured isolation is
adequate enough with a value below 15 dB for the whole band and the measured
reection coecient for all the ports is below 13 dB, thanks to the EBG structures,
otherwise these values do not reach 10 dB. The rest of the ports (ports 3 and 4) are
not shown but they have similar results.
(a) Port 1: right and left inputs (b) Port 2: right and left inputs
Figure 3.19: S parameters measurements of 4x4 array with EBG structures.
Figure 3.20: 4x4 array and Butler matrix network connection for LHCP congura-
tion.
In order to get the beam steering in elevation, the passive Butler matrix network,
(Chapter 2) is connected to the 4x4 subarray with EBG barriers as it can be seen
in Fig. 3.20. If the Butler network is connected to left side ports the antenna
radiation will be LHCP and vice versa, therefore, this system obtains double circular
78 MUTUAL COUPLING REDUCTION USING EBG IN . . .
polarization. The rest of the ports are loaded with 50Ω loads. Butler Matrix network
yields -45, + 135, -135 and +45 phase shift between output ports when exciting
ports 1 to 4 respectively.
These phase shifts between elements (α= ±45 and ±135) are used to calculate
the pointing directions of the antenna. Those directions are calculated by applying
exploration beam equation. The steering directions for the antenna: +10, -40,
+40 and -10 are obtained.
Figure 3.21: 4x4 steering array radiation pattern for RHCP, at the center frequency
(7.825GHz).
In Fig. 3.21, the steering vertical plane radiation patterns for RHCP at the
center frequency (7.825 GHz) and dierent pointing directions are presented. Good
agreement between simulations (dashed lines) and measurements (continuous lines)
is obtained. The gain dierences between simulations and measurements are due
to the connectors and cables used to connect the Butler matrix network with the
antenna, which are not introduced in the simulations. CP/XP ratio is better than
15 dB. Due to the uniform distribution, Side Lobe Level (SLL) is under 12 dB for
±10 and under 7 dB for ±40 beams.
Finally, in Fig. 3.22(a) and Fig. 3.22(b), the axial ratio for the dierent beams
and dierent steering angles at the end and the beginning of the frequency band are
shown. In this case, dierent frequencies, polarizations and steering directions are
Applications 79
shown in order to prove the good features of the antenna.
(a) +10 LHCP and -40 RHCP beams (b) -10 LHCP and +40 RHCP beams
Figure 3.22: 4x4 steering array axial ratio for frequencies 7.25 GHz and 8.4 GHz,
RH and LH circular polarizations over the scanning angles.
As it was expected, the introduction of EBG structures between the radiating
elements has not signicant inuence in the circular polarization of the antenna,
and only slightly dierences are appreciated. Finally, in Fig. 3.23, the axial ratio
of the antenna is shown, related to the frequency. The purity of the polarization is
better than 3 dB over the whole frequency band [3.26].
Figure 3.23: 4x4 steering array axial ratio for RH and LH circular polarizations and
dierent pointing directions over the working frequency.
80 MUTUAL COUPLING REDUCTION USING EBG IN . . .
3.6. Conclusions
In the beginning of this chapter, the basic metamaterial properties are overvie-
wed. Nevertheless, the main objective is to summarize the EBG structures which
are one type of metamaterials based on FSS. By understanding its electromagne-
tic behavior, a design method is proposed for calculating the size of the structures
analytically. These structures can block the surface waves when they are introduced
in between of two printed antennas and therefore its properties as electromagnetic
barriers are thoroughly described. In the rst place, the allowed propagating modes
are presented by means of Brillouin diagrams. Secondly, a electromagnetic simula-
tion scheme is built and parameterized in CST. With this scenario, the inuence of
the dierent parameters such as, gap g, patch size w, via diameter dvia or substrate
thickness t is analyzed.
So far, the solution in order to introduce the EBG structures was to increase
the substrate permittivity which leads to a size reduction of the printed radiating
elements, however, this strategy reduces the radiation eciency and enhance the
surface wave propagation modes.
In order to reduce the mutual coupling between radiating elements of a steering
array, using low permittivity substrates and therefore maintaining good radiation
eciency a new smaller mushroom shape is proposed. This structure combines
multilayer substrate and via edge location reducing the eective size by 30%.
Finally the new EBG structures are inserted in a 4x4 steering array showing how
these electronic barriers reduce the mutual coupling between elements of dierent
rows and increase the performance of the antenna in terms of directivity and back
lobe reduction.
References
[3.1] R. Mailloux. On the use of metallized cavities in printed slot arrays with dielec-
tric substrates. Antennas and Propagation, IEEE Transactions on, 35(5):477
487, May 1987.
[3.2] C. A. Balanis. Antenna Theory, analysis and design. Wiley india, third edition
edition, 1997.
[3.3] B.P. Kumar and G.R. Branner. Design of unequally spaced arrays for per-
formance improvement. Antennas and Propagation, IEEE Transactions on,
47(3):511523, Mar. 1999.
[3.4] S. Mohsen, M. Alireza, T. Ahad, and H. Teimur. Mutual coupling reduction
of microstrip antennas using defected ground structure. In Communication
systems, 2006. ICCS 2006. 10th IEEE Singapore International Conference on,
pages 15, Oct. 2006.
[3.5] F. Yang and Y. Rahmat-Samii. Microstrip antennas integrated with electro-
magnetic band-gap (EBG) structures: a low mutual coupling design for array
applications. Antennas and Propagation, IEEE Transactions on, 51(10):2936
2946, Oct. 2003.
[3.6] J.M. Fernández-González. Application of metamaterial structures in the design,
analysis and prototyping of planar antennas. PhD thesis, ETSI de Telecomuni-
cación, Technical University of Madrid (UPM), 2008.
[3.7] F. Yang and Y. Rahmat-Samii. Reection phase characterizations of the EBG
ground plane for low prole wire antenna applications. Antennas and Propaga-
tion, IEEE Transactions on, 51(10):26912703, Oct. 2003.
[3.8] N. Engheta and R. Ziolkowski. Metamaterials, physics and engineering explo-
rations. John Wiley and Sons, 2006.
81
82 REFERENCES
[3.9] R. Baggen, M. Martínez-Vázquez, J. Leiss, S. Holzwarth, L.S. Drioli, and
P. de Maagt. Low prole galileo antenna using EBG technology. Antennas and
Propagation, IEEE Transactions on, 56(3):667674, Mar. 2008.
[3.10] L. Inclán-Sánchez, J.-L. Vázquez-Roy, and E. Rajo-Iglesias. High isolation
proximity coupled multilayer patch antenna for dual-frequency operation. An-
tennas and Propagation, IEEE Transactions on, 56(4):11801183, Apr. 2008.
[3.11] E. Rajo-Iglesias, O. Quevedo-Teruel, and L. Inclán-Sánchez. Mutual coupling
reduction in patch antenna arrays by using a planar EBG structure and a
multilayer dielectric substrate. IEEE Trans. on Antennas and Propagation,
56(6):16481655, Jun. 2008.
[3.12] E. Rajo-Iglesias, L. Inclán-Sanchez, J.-L. Vázquez-Roy, and E. García-Muñoz.
Size reduction of mushroom-type EBG surfaces by using edge-located vias.
Microwave and Wireless Components Letters, IEEE, 17(9):670672, Sept. 2007.
[3.13] L. Yang, F. Mingyan, C. Fanglu, S. Jingzhao, and F. Zhenghe. A novel
compact electromagnetic-bandgap (EBG) structure and its applications for mi-
crowave circuits. Microwave Theory and Techniques, IEEE Transactions on,
53(1):183190, Jan. 2005.
[3.14] E. Rajo-Iglesias, M. Caiazzo, L. Inclán-Sanchez, and P.-S. Kildal. Comparison
of bandgaps of mushroom-type EBG surface and corrugated and strip-type soft
surfaces. Microwaves, Antennas Propagation, IET, 1(1):184 189, Feb. 2007.
[3.15] E. Rajo-Iglesias, O. Quevedo-Teruel, and L. Inclán-Sánchez. Planar soft sur-
faces and their application to mutual coupling reduction. Antennas and Pro-
pagation, IEEE Transactions on, 57(12):38523859, Dec. 2009.
[3.16] O. Quevedo-Teruel, L. Inclán-Sánchez, and E. Rajo-Iglesias. Soft surfaces
for reducing mutual coupling between loaded PIFA antennas. Antennas and
Wireless Propagation Letters, IEEE, 9:9194, 2010.
REFERENCES 83
[3.17] E. Rajo-Iglesias, J.-L. Vázquez-Roy, O. Quevedo-Teruel, and L. Inclán-
Sánchez. Dual band planar soft surfaces. Microwaves, Antennas Propagation,
IET, 3(5):742748, Aug. 2009.
[3.18] D. Sievenpiper, L. Zhang, R. Broas, N. Alexopoulos, and E. Yablonovitch.
High-impedance electromagnetic surfaces with a forbidden frequency band.
IEEE Trans. On Microwave Theory and Techniques, 47:20592074, Nov. 1999.
[3.19] C. Caloz and T. Itoh. Electromagnetic metamaterials: transmission line
theory and microwave applications: the engineering approach. John Wiley and
Sons, 2006.
[3.20] F. Yang and Y. Rahmat-Samii. Electromagnetic Band Gap Structures in
Antenna Engineering. The Cambridge RF and Microwave Engineering Series,
2008.
[3.21] P.-S. Kildal. Denition of articially soft and hard surfaces for electromagnetic
waves. Electronics Letters, 24(3):168170, Feb. 1988.
[3.22] F. Yang and Y. Rahmat-Samii. Polarization-dependent electromagnetic band
gap (pdebg) structures: Designs and applications. Microwave and Optical Tech-
nology Letters, 41(6):439444, Jun. 2004.
[3.23] E. Rajo-Iglesias, L. Inclán-Sanchez, J.-L. Vázquez-Roy, and E. García-Muñoz.
Size reduction of mushroom-type EBG surfaces by using edge-located vias.
Microwave and Wireless Components Letters, IEEE, 17(9):670672, Sept. 2007.
[3.24] G. Expósito-Domínguez, J. M. Fernández-González, P. Padilla, and M. Sierra-
Castañer. Dual circular polarized steering antenna for satellite communications
in X band. Progress In Electromagnetic Research, 122:6176, 2012.
[3.25] R. Garg, I. Bahl, and A. Ittipiboon. Microstrip antennas design handbook.
Artech House, 2001.
84 REFERENCES
[3.26] G. Expósito-Domínguez, J. M. Fernández-González, P. Padilla, and M. Sierra-
Castañer. Mutual coupling reduction using EBG in steering antennas. IEEE
Antennas and Wireless Propagation Letters, 11:12651268, 2012.
Chapter 4
Phase shifting distribution networks
for switchable beam antennas
In electrical engineering, a switch is an electrical component that can break
an electrical circuit, interrupting the current or diverting it from one conductor to
another. A Single Pole Single Throw (SPST) can be used to enable or disable specic
parts of the circuit meanwhile a Single Pole Double Throw (SPDT) is used to select
one of the two posible paths. In active steering antennas, they can be used to switch
between to dierent distribution networks, feeding points of the radiating elements
to change their polarization or phase shifters.
4.1. Fundamentals of phase shifters
The phase shifter is a two-port device used to produce either a xed or con-
trollable phase shift to the input signal. Phase shifters are used in a variety of
communication and radar systems. They are essential components in antenna sys-
tems and phased arrays, where the radiating elements must be fed with appropiate
phases so as to create the required radiated beam. In general, xed phase shifters
can potentially become variable phase shifters by replacing some xed reactances
with electronically tunable ones.
In the ideal case, a phase shifter can be dened as a two-port device perfectly
matched (S11=S22=0) and without loss (thus |S21|=|S12|), where the output signal
undergoes a prescribed phase shift ∆φ with respect to a given reference. The phase
shift of the output signal is usually evaluated with reference to the phase shift of
85
86 PHASE SHIFTING DISTRIBUTION NETWORKS FOR . . .
a line section of equal length as the device. Being the lenght of such a line section
and β its phase constant; the phase shift is thus evaluated as:
∆φ = θ − βl (4.1)
Depending on the frequency behavior of ∆φ(f), there are two fundamental types
of phase shifters:
The true− phase phase shifter, whose ∆φ(f) is constant with the frequency.
The true − time − delay (TTD) phase shifter, whose ∆φ(f) varies linearly
with frequency, so that the device behaves as a delay line.
In phased array antennas, it is very important to use TTD, because the elements
are placed physically at a certain distance, but the eective distance changes with
the frequency. The delay line behavior of TTD adjust the phase dierence in each
frequency which is very appropiate. The work presented in this thesis is focused on
TTD phase shifters.
Phase shifters can be broadly classied as mechanical or electronic devices, de-
pending whether the phase control is achieved through mechanical or electronic
tuning. Depending on the type of operation, they can be categorized as analog or
digital, having reciprocal or nonreciprocal characteristics. Other classications can
be in terms of the type of employed transmission structure to realize the phase shif-
ter (waveguide, planar transmission line, dielectric guide, etc.) and the technology
adopted for fabrication (semiconductors, monolithic or ferrites).
An important sector where this technology can be used is automotive radars.
These systems integrated in cars can be widely used with a relative low cost design.
There are two frequency bands reserved for automotive applications, at 24 GHz
a Butler matrix network is modied with CMOS chips in order to get continuous
scanning ±90o [4.1]. In [4.2] by changing the phase of 18 power ampliers the
steering beam is obtained with a directivity of 12 dB and 5o beamwidth. 60 GHz
is another working frequency band for automotive applications in [4.3] a 4 elements
array with ±25o steering angles is shown.
Fundamentals of phase shifters 87
The use of PIN diodes in steering antennas is widely known. Examples of the
use of these devices can be found in [4.4] where PIN diodes are use to change
the radiation pattern of a Yagui-Uda and a planar inverted F antenna. In [4.5]
the diodes are applied to a grid antenna which obtains a ±10o and ±20o steering
angles in elevation and azimuth respectively. This antenna has 2 dB insertion loss
at 3 GHz. Finally, in [4.6] the purpose of the work is focused on the design low
series resistivity PIN diodes, the reason for that is to reduce the insertion loss in a
reectarray compound of dipoles circularly placed which get circular polarization at
18.2 GHz.
The good linearity of the MEMS devices allows to build switches with an ope-
ration band from 0.1 to 120 GHz. In [4.7] two phase shifters, the rst one with
two bits and 0.6 dB of insertion loss and the second one with four bits and 1.2 dB
insertion loss at 10 GHz. Another example in [4.8] shows a transmission line with
the ability to change the propagation speed, by changing the line impedance with
the capacitance variation. These concepts have been applied to electronic steering
antennas. In [4.9] a Ka band antenna constructed with 6 layers LTCC technology is
shown. This antenna has a 8 dBi gain and ±15o steerable beam. Meanwhile in [4.10]
the antenna is able to steer the beam in both directions, (azimuth and elevation)
and it works in V band (40-75GHz). These devices can be used also in reectarrays
such as [4.11] where the device acts as a periodic structure, it obtains a phase shift
of 150o, and it has between 0.4-1.5 dB insertion losses at 2 GHz. In [4.12] a patch
antenna is coupled fed by an slot, by changing the slot dimensions with MEMS the
phase shift of the radiating elements can be changed. High performance antennas
for radar applications are developed with MEMS devices too [4.13]. This antenna
has a scan range of ±60o, eective area of 0.4 m2 and it works at X band.
In [4.14] a beam steering traveling wave antenna is designed. It is compound
by interconnected patch antennas with transmission lines where varactor diodes are
placed. By changing the capacitance of the diodes, the eective distance between
the radiating elements change and consequently the phase feeding law and the stee-
ring direction. Similarly, in [4.15] the eective lenght of a meander line antenna at
X band is electronically controlled with varactor diodes. The 60 GHz band is used
88 PHASE SHIFTING DISTRIBUTION NETWORKS FOR . . .
for wireless communication, many private networks and indoor multi-channel trans-
mission are being developed. For this purpose, a ±7o steering antenna composed of
24 radiating elements is proposed in [4.16]. This antenna is fed by phase shifters
based on transmission lines nished with varactor diodes.
The problem of ferrite based phase shifters is their poor calibration accuracy.
Hysteresis can cause changes of 30o in the phase shift prevailing at a given applied
magnetic eld. In some cases, thermal drifts may also alter the obtained phase shift.
This is an important issue that can be handle with a servo system [4.17]. Another
important issue is to build these devices at very high frequencies due to the space
constraints. However, in [4.18], ferrite based phase shifters are designed at 35 and
94 GHz. In order to avoid the space and weight problems, a low cost microstrip
line based ferrite phase shifter is proposed in [4.19]. The design is based on three
microstrip lines arranged and fed with phase dierences so as to produce circular
polarization in the ferrite region. This obtains a phase shift of 360o in less than an
eective wavelength yielding an important size and weight reduction.
4.2. Topologies
Depending on the circuit topology, the device can be considered as transmission
or reection phase shifters, depending on where the phase shift is produced. In
transmission, the signal phase shift is through the device, while the reection one
exploits the phase shift produced by the reection from a reactive load. Below, the
well known topologies with its advantages and drawbacks are thoroughly discussed.
Series-shunt
There are two fundamental methods of connecting switches to a transmission
line to provide a switching function: in series with the transmission line so that RF
power is conducted when the switch is forward biased (Fig. 4.1(a)) and reected
when reverse biased; or in shunt with the transmission line so that the RF power
is conducted when the switch is reverse biased and reected when forward biased
Topologies 89
(Fig. 4.1(b)). A simple reective SPST switch can be designed utilizing one or more
switches in either conguration as shown in Fig. 4.1. The combination of several
devices will increase the isolation regarding the diversion of the current, however it
will introduce higher insertion losses.
(a) Series switch. (b) Shunt switch.
Figure 4.1: Single switches.
Compound
A multi-throw microwave switch essentially consists of combination of SPST
switches connected to a common junction and biased so that each switch port can
be enabled individually. The common junction of the switch must be designed to
minimize the resistive and reactive loading presented by the OFF ports in order to
obtain low insertion loss and VSWR for the ON port. By using series mounted PIN
diodes connected to the common junction, a path is selected by forward biasing its
series diode and simultaneously reverse biasing all the other diodes. This provides
the desired low-loss path for the ON port with a minimum of loading from the OFF
ports. The two most common compound switch conguration are Series-Shunt or
TEE designs. The schematic diagrams for both switches are shown in Fig. 4.2.
(a) Series-Shunt SPST switch. (b) TEE SPST switch.
Figure 4.2: Compound switches.
90 PHASE SHIFTING DISTRIBUTION NETWORKS FOR . . .
Table 4.1: Summary of formulas for SPST Switches.
Type Isolation Insertion loss
Series 1 +(XC
2Z0
)2 (1 + RS
2Z0
)2Shunt
(1 + Z0
2RS
)21 +
(Z0
2XC
)2Series-Shunt
(1 + Z0
2RS
)2+(XC
2Z0
)2 (1 + Z0
RS
)2 (1 + RS
2Z0
)2+(Z0+RS
2XC
)2TEE
[1 +
(XC
Z0
)2] [(1 + Z0
2RS
)2+(XC
2RS
)2] [(1 + RS
Z0
)2] [1 +
(Z0+RS
XC
)2]
A summary of formulas used for calculating insertion loss and isolation for com-
pound and simple switches is given in table 4.1.
Tuned
This method utilizes shunt mounted PIN diodes located a quarter wavelength
from the junction. The diode(s) of the selected ON port is reverse biased while the
OFF ports are forward biased to create a short circuit across the transmission line.
As a result of the quarter wavelength spacing, the short circuits are transformed
to open circuits at the junction (Fig. 4.3). By proper choice of transmission line
impedances and minimization of stray reactance it is possible to construct a switch
of this type with low insertion loss and VSWR over a three to one bandwidth.
(a) Tuned series switch. (b) Tuned shunt switch.
Figure 4.3: Tuned λ/4 switches.
Topologies 91
Reective
It is often desirable to have a switch that presents a low VSWR in its OFF
position as well as in its ON state in order to maintain desired system performance.
Reective or also known as absorptive switch incorporate 50Ω terminations in each
of the output ports. Fig. 4.4 shows the two possible congurations. The operation of
this switch is as follows: The power incident on port A divides equally between ports
B and C, port D is isolated. The mismatch produced by the switch in series (Fig.
4.4(a)) or parallel (Fig. 4.4(b)) with the load resistance at ports B and C reects
the power. The reected power exits through port D isolating port A. Therefore, A
appears matched to the input signal.
Both of the above types of hybrid switches oer good features. The series con-
nected diode conguration is, however, recommended for attenuators used primarily
at high attenuation levels while the shunt mounted diode conguration is better for
low attenuation ranges.
(a) series connection. (b) shunt connection.
Figure 4.4: Quadrature matched hybrid switches.
Switched transmission lines
A basic example of a switched line phase shifter circuit is shown in Fig. 4.5.
In this design, two SPDT switches employing PIN diodes are used to change the
electrical length of transmission line by some length ∆φ. The phase shift obtained
from this circuit varies with frequency and is a direct function of this dierential
line length as ∆φ = 2π∆l/λ. The switched line phase shifter is inherently a broad
92 PHASE SHIFTING DISTRIBUTION NETWORKS FOR . . .
band circuit producing true time delay, with the actual phase shift dependent only
on ∆l.
Figure 4.5: Switched line phase shifter.
There are some other techniques based on the reactive part variation. By chan-
ging the biasing of the switch device (i.e PIN diode) the XC changes and therefore
there is phase shift produced along the transmision line depending on the biasing.
This topology is called loaded line phase shifter and it main advantages are high
power handling and continuous phase shift. However the biasing is very sensitive
and the insertion loss are higher. This topology is not further developed because its
phase shift is dependant with the frequency. It is not a TTD and therefore is not
adequate for steering array antennas.
4.3. Switchable beam antenna
So far the topic of the thesis has been focussed on steerable systems, the descrip-
tion of the system requirements and antenna features were shown and the problems
about the radiating elements and their mutual coupling has been addressed, howe-
ver, none of the phase distribution networks have been presented. In this section
the implementation of several feeding networks, based on phase shifters or switches,
able to achieve the adequate phase distribution are presented.
Standard users can not aord high accuracy scanning antennas in daily used
systems such as satellite television, WIFI connection or automotive RADAR and
therefore another dierent techniques not based in active phased arrays antennas
must be explored. By means of digital signal processing the SNR can be enhanced
Switchable beam antenna 93
when using several radiation patterns such as multibeam systems. These systems do
not require narrow step scanning angles, and consequently the distribution networks
are cheaper due to the lack of phase shifters with large number of bits. However,
the radiation pattern must be well dened, the insertion loss must be equal for all
the beams, and the isolation between dierent networks or signal paths must be as
large as possible to avoid cross talk interferences.
Switchable beam antennas can be used in RADAR systems to avoid ambiguity
when two objets are at the same distance but in two dierent azimuth or elevation
positions in the same temporal window [4.20, 4.21]. However, the following issues
must be taken into account carefully:
Crossing beam losses. When using an specic beam it is desirable to transmit
or receive the same amount power regardless θ or φ angles (Fig. 4.6(b)). The
cross beam losses could yield a position or velocity miscalculation (Fig. 4.6(a)).
(a) Received power as a function of the
incoming signal direction.
(b) Equal received power for two dierent
incoming signals.
Figure 4.6: Uniform radiation pattern beam shape.
Dierent insertion loss for each feeding network. The construction of the
distribution networks can contain dierent components in each of the signal
paths yielding dierent insertion loss for each beam and therefore a reduction
in its gain (Fig. 4.7).
94 PHASE SHIFTING DISTRIBUTION NETWORKS FOR . . .
(a) Desired Scenario. (b) Dierent beam insertion loss scenario.
Figure 4.7: Multibeam radiation pattern unbalance description.
Isolation between networks. Eventhought a beam is selected, the rest of the
beams are presented at the same time. The gain level of the undesired beams is
dened by the isolation between the networks, therefore the lower the leakage
is in the distribution network switches, the better performance the system has
(Fig. 4.8(a)).
(a) Gain level of the isolated beams. (b) Worst case scenario.
Figure 4.8: Multibeam radiation pattern isolation description.
In Fig. 4.8(b) the worst case is presented were undesired clutter appears in
the isolated beams and the desired signal is weakly received because of higher
insertion losses. To overcome this problem, the insertion losses in the distri-
bution network must be equal for every path and clutter reduction techniques
Switchable beam antenna 95
based on digital correlation signal must be applied.
4.3.1. Symmetric reective phase shifter
As a key component for a distribution network, a 120o reective phase shifter at
Ka band (24 GHz) is designed and constructed (Fig. 4.9). This device is construc-
ted with microstrip technology on a ROGERS 4350B substrate,ε=3.66, 0.25 mm
thickness and tested with two dierent PIN diodes MA912 and MA200, to nd the
best response under the same biasing conditions (IF=100 mA and VR=0V). Deco-
upling capacitors and lter structures are place to isolate RF and DC biasing and
the connectors used are rosenberger mini SMP able to operate up to 40 GHz.
Figure 4.9: Reective phase shifter construction.
The operation of this circuit is similar to a reective switch (section 4.2) but with
a transmission line which introduce a phase shift. The phase dierence is obtained
when the PIN diode is reverse or forward biased. The following equations show the
behaviour when the PIN diodes are reverse bias.
96 PHASE SHIFTING DISTRIBUTION NETWORKS FOR . . .
Table 4.2: Reective phase shifter [S] matrix.
ON OFF 0 ejπ/2
ejπ/2 0
0 ej(π/2+2φ)
ej(π/2+2φ) 0
[S]Refl =−1√
2
0 j 1 0
j 0 0 1
1 0 0 j
0 1 j 0
· ejφ · ejφ · −1√
2
0 j 1 0
j 0 0 1
1 0 0 j
0 1 j 0
(4.2)
S21 · ejφ · ejφ · S42 + S31 · ejφ · ejφ · S43 =
−1√2j · ejφ · ejφ−1√
21 +−1√
21 · ejφ · ejφ−1√
2j =
j
2· ej2φ +
j
2· ej2φ = j · ej2φ = ej(π/2+2φ) (4.3)
And the case for the PIN diodes forward biasing is:
[S]Refl =−1√
2
0 j 1 0
j 0 0 1
1 0 0 j
0 1 j 0
·
0 j 1 0
j 0 0 1
1 0 0 j
0 1 j 0
(4.4)
S21 · S42 + S31 · S43 =
−1√2j−1√
21 +−1√
21−1√
2j =
j
2+j
2= j = ejπ/2 (4.5)
Therefore, the [S] matrix of the reective phase shifter for the OFF and ON case
are:
The amplitude measurements (Fig. 4.10(a)) show a frequency shift (22.5 GHz
instead 24 GHz) in the crossing point between the OFF and the ON state. The
Switchable beam antenna 97
insertion loss are 1.5 ±0.5dB for the whole operation band (BW=1 GHz, 4%) and
the phase shift between the two states (Fig. 4.10(b)) is 120±10.
(a) Module. (b) Phase.
Figure 4.10: Reective phase shifter measurements.
This frequency shift is due to the inaccurate PIN diode model, which series
resistance and total capacitance for the ON and OFF states, depend on the bias
current (IF ) and voltage (VR), and physical dimensions. The non-linear model obey
the following equations:
Q = IF τ (4.6)
RS =W 2
(µp + µn)Q(4.7)
Where τ is the carrier lifetime,W is the Intrinsic region width, and µn and µp are
the electron and hole mobility respectively. In Fig. 4.11, the insertion loss variation
due to the PIN diode parameters model is presented. After a thorough simulation,
by comparing the simulation expected from Fig. 4.10 with the measurements, the
diode parameters for MA200 are found and accurate model is parametrized for the
next designs. CT=40pF, L=0.1nH, Rp=1000Ω and Rs=5Ω.
To overcome the problem that the dierent insertion loss yield on each beam
gain of the radiation pattern a series-shunt switch is introduced in the outputs (B
and C) of the reective phase shifter (Fig. 4.12(b)).
98 PHASE SHIFTING DISTRIBUTION NETWORKS FOR . . .
(a) ON state series resistance varia-
tion.
(b) OFF state total capacitance va-
riation.
Figure 4.11: Insertion loss variation due to the PIN diode parameters model.
(a) Reective switch. (b) Reective series-shunt switch.
Figure 4.12: Reective congurations.
With this circuitry topology, the overall insertion loss are slightly increased but
the dierence between ON and OFF states are signicantly reduced while maintai-
ning the desired frequency shift. To symmetrize both paths regarding insertion loss
is very important because this phase shifter is placed in the distribution network
several times and the current path to each radiating element must remain constant
at all times [4.22].
4.3.2. Matched SPDT switch
To feed an array antenna with several distribution networks is a well known
technique in order to obtain multiple radiation patterns. However, it is very impor-
tant to ensure an adequate isolation between the dierent distribution networks and
Switchable beam antenna 99
(a) Amplitude. (b) Phase.
Figure 4.13: Series-shunt reective phase shifter results.
reduce the leakage as much as possible. When using a Single Pole Double Through
switch for diverting the current from one input to one of the two possible outputs
it is very important to take good care of the technology utilized and the chosen
circuitry topology.
For a two beam array antenna, which is fed by means of two feeding networks a
SPDT based on PIN diodes is designed and built. The diodes utilized are MA 200,
which are well characterized for the biasing conditions at the operation frequency
(24 GHz). To choose the better topology for the system, the isolation, leakage and
insertion loss where study for the cases of single and double SPDT switch.
(a) Single. (b) Double.
Figure 4.14: Single and Double PIN diode based SPDT switch.
In Fig. 4.14, two schematics for single and double PIN diode based SPDT switch
are shown. The operation of both circuits is controlled biasing one of the output with
forward current (IF ) meanwhile the other output is reversed biased (VR). The rst
diode is placed λ/2 away from the junction, in this way the diode high impedance
produced in the reversed bias output is translated to the junction as a open circuit.
The transmission lines where the diodes are placed are Z0 = 50Ω. In contrast to a
100 PHASE SHIFTING DISTRIBUTION NETWORKS FOR . . .
wilkinson divider, at any time one of the branches is an open circuit and therefore
the output impedance do not change. This technique enhance the isolation between
the outputs and solve the possible leakage of placing diodes too close.
Due to the limitations of the biasing the equivalent series resistance (RS) and the
total capacitance (CT ) of the PIN diodes do not yield perfect open or short circuits
and the possibility of introducing other PIN diode in each output is considered. This
second diode must be placed λ/4 away from the rst diode, in this way the isolation
performance is increased by introducing a second pole in the transfer funtion (Fig.
4.15). In these set of gures the operation of the switch is shown. The PIN diodes
are shunt placed to reduce the insertion loss during the operation.
(a) Schematic equivalent. (b) Diodes in output 1 are reversed bia-
sed and forward biased in output 2.
(c) The second diode is an open circuit in
series with the output.
(d) The rst diode is shunt short circuit
with the output.
Figure 4.15: Graphic equivalent of the double SPDT operation.
Switchable beam antenna 101
Figure 4.16: Comparison of the features of the single and double SPDT.
In Fig. 4.16 the comparison of the single and double SPDT switches is shown.
At the operation frequency, the insertion loss for both cases are -1 dB, however it
can be see how the isolation and the leakage are ≈-12 dB for the single diode SPDT
meanwhile these values for the double diode SPDT are ≈-22 dB. The increase of
complexity in the design and cost is worth in comparison with the features obtained.
Figure 4.17: Double PIN diode SPDT switch Photograph.
The construction of one SPDT switch with two PIN diodes in each output is
shown in Fig. 4.17. The PIN diodes are shunt placed and are fed by high impedance
λ/2 lines. The forward current that fed the diodes is limited by polarization resistors,
in this way the series resistance of the diode is characterized accurately at all the
102 PHASE SHIFTING DISTRIBUTION NETWORKS FOR . . .
times. Several circuits were constructed with dierent λ/2 lengths in order to obtain
a precise short circuit at the operation frequency (Fig. 4.18).
(a) Isolation. (b) Transmission.
Figure 4.18: Double PIN diode SPDT switch measurements.
In Fig. 4.19 the matching, isolation and transmission parameters for the circuit
adjusted to 24 GHz are shown. The insertion loss are smaller than -2 dB when the
eect of the connectors is taken into account and the matching of the ports is better
than -10 dB over the whole operation band. The isolation obtained (<-40 dB) shows
that double SPDT circuit model presented in Fig. 4.15 works properly.
Figure 4.19: SPDT full measurements at 24 GHz.
As long as these devices must operate under a wide range of circumstances, a
Switchable beam antenna 103
temperature study is carried out. In order to study the behaviour of the device with
the temperature is necessary to remember that the thicker intrinsic region of the
PIN diode W , the lower the switching time, ts, however the higher power handling
capacity. The operation or ambient temperature, TA, is directly related with the
maximum allowable power dissipation, PD in Watts, which is determined by:
PD =Tj − TA
θ(4.8)
where Tj is the maximum allowable junction temperature (usually 175). Power
dissipation may be computed as the product of the RF current squared multiplied by
the diode resistance, RS. For CW applications the value of the thermal resistance,
θ, used is the average thermal resistance, θAV . However in most pulsed RF and
microwave applications where the duty factor, DF, is less than 10 % and the pulse
width, tp, is less than the thermal time constant of the diode, good approximation
of the eective value of θ in C/W is obtained (eq. 4.9).
θ = DF · θAV + θtp (4.9)
Where θtp is the thermal impedance of the diode for the time interval correspon-
ding to tp.
The double PIN SPDT switch is tested over temperature (Fig. 4.20), under the
same biasing conditions but with a temperature variation from -20C to 80C. An
insertion loss rise can be noticed when the temperature is raised (Fig. 4.20(a)).
This makes sense because the power that the diode can handle is reduced when the
ambient temperature raises and therefore its behaviour as a short circuit is worse.
As long as the diodes are placed as a shunt, more current goes through the isolated
path, hence the overall insertion loss to the desired output are increased.
When the outputs of a SPDT switch are connected to an antenna array the iso-
lated port is an open circuit for the distribution network (network 2), therefore all
the power is reected and re-radiated or coupled to the other network (network 1)
through the radiating element (Fig.4.21). This situation could yield strong interfe-
rences and modify the desired radiation pattern.
104 PHASE SHIFTING DISTRIBUTION NETWORKS FOR . . .
(a) Isolation. (b) Transmission.
Figure 4.20: SPDT switch measurements over the temperature.
Figure 4.21: Coupling between networks due to the open circuit of the isolated path
of the conventional SPDT switch.
To overcome this problem, a combination of several circuits that select one from
two possible outputs matching all the ports at any time is proposed [4.23]. A
Wilkinson divider splits the input power equally in two outputs paths. The rst
ouptut is connected to a 270 delay line meanwhile the other output of the Wilkinson
divider is connnected to a 180 reective phase shifter. Both paths are connected
to the input (A) and isolated (C) ports of a branch line 3dB/90 . The distribution
networks are connected to the outputs of the branch line (B, D). Fig. 4.22 shows a
circuit schematic of the complex SPDT with matched output ports.
Thus, the operation of the complex SPDT can be expressed in form of [S] para-
meters in the following form for the OFF state and the electrical length of the delay
lines is φ = 90o:
Switchable beam antenna 105
Figure 4.22: Complex SPDT switch with matched output ports.
[S]SPDT =−j√
2
0 1 1
1 0 0
1 0 0
·e
j3π/2 · −1√2
0 j 1 0
j 0 0 1
1 0 0 j
0 1 j 0
+ ej2π · −1√2
0 j 1 0
j 0 0 1
1 0 0 j
0 1 j 0
(4.10)
Output 1:
S21,Wil · e−j3π/2 · S21,Refl + S31,Wil · e−j2π · S24,Refl =
−j√2
1 · e−j3π/2 · −1√2j +−j√
21 · e−j2π−1√
21 =
j
2· e−j3π/2 − 1
2e−j2π = 0 (4.11)
Output 2:
S21,Wil · ej3π/2 · S31,Refl + S31,Wil · ej2π · S34,Refl =
−j√2
1 · e−j3π/2 · −1√2
1− −j√2
1 · e−j2π−1√2j =
1
2· e−jπ − 1
2e−j2π = 1 (4.12)
Therefore, the output power goes to the output port 2, the output port 1 is
isolated and the input and the outputs are matched at anytime. Similar development
can be done for the ON state:
106 PHASE SHIFTING DISTRIBUTION NETWORKS FOR . . .
Table 4.3: Complex SPDT [S] matrix.
ON OFF0 1 0
1 0 0
0 0 0
0 0 1
0 0 0
1 0 0
Output1:
S21,Wil · e−jπ/2 · S21,Refl + S31,Wil · e−j2π · S24,Refl =
−j√2
1 · e−jπ/2 · −1√2j +−j√
21 · e−j2π−1√
21 =
−1
2· e−jπ/2 +
1
2e−j3π/2 = 1 (4.13)
Output 2:
S21,Wil · ejπ/2 · S31,Refl + S31,Wil · ej2π · S34,Refl =
−j√2
1 · e−jπ/2 · −1√2
1− −j√2
1 · e−j2π−1√2j =
1
2· e−j0 − 1
2e−j2π = 0 (4.14)
Finally the [S] of the complex SPDT is shown in Table 4.3:
In Fig. 4.23, the simulation results for the complex SPDT switch with matched
output ports are shown. The insertion loss are 1.1±0.15 dB and the return loss of
all the ports, isolation and leakage between the outputs have values below -23 dB
for the whole operation frequency band.
4.4. Conclusions
In this chapter, the phase shifting distribution networks are thoroughly studied
in order to apply them to switchable beam antennas. A briey fundamentals des-
cription, which is essential to understand the operation of this networks is presented.
The technology overview (MOSFET, MEMS, PIN and Varactor diodes and ferri-
te) employed in switches and phase shifters is presented in the introduction. This
Conclusions 107
Figure 4.23: Insertion loss, isolation and return loss of the complex SPDT switch
with matched output ports.
chapter focuses on the development of distribution networks with the capability of
phase shifting between radiating elements to obtain a steering direction, therefore a
thorough description of switch and phase shifter topologies is shown.
The application of this circuits to automotive RADARs at the frequency band
of 24 GHz requires high performance in terms of isolation, insertion loss, leakage,
temperature behaviour, etc. In this work, a symmetric reective phase shifter with
series-shunt switch at the outputs gets a small dierence of 0.2 dB for the ON and
OFF states while maintaining the desired 120 degrees. The design task carried out
here can be exported to any frequency because the circuit topology is not frequency
dependant.
Diverting the input to one of the two dierent output networks with a single
SPDT can involve low features in terms of matching, isolation and leakage. In order
to solve this problem, a complex SPDT with matched ports regardless operation
constrains is proposed. This circuit combines a 180 reective phase shifter, a
Wilkinson divider, a 270 delay line and a 3 dB 90. The results show a 1.1 dB
insertion loss and return loss, isolation and leakage below -23 dB.
References
[4.1] T.-Y. Chin, S.-F. Chang, J.-C. Wu, and C.-C. Chang. A 25-GHz compact low-
power phased-array receiver with continuous beam steering in cmos technology.
Solid-State Circuits, IEEE Journal of, 45(11):2273 2282, Nov. 2010.
[4.2] J. Wernehag and H. Sjoland. A 24-GHz automotive radar transmitter with
digital beam steering in 130-nm cmos. In Research in Microelectronics and
Electronics 2006, Ph. D., pages 481484, 2006.
[4.3] J.-L. Kuo, Y.-F. Lu, T.-Y. Huang, Y.-L. Chang, Y.-K. Hsieh, P.-J. Peng, I-
C. Chang, T.-C. Tsai, K.-Y. Kao, W.-Y. Hsiung, J. Wang, Y.A. Hsu, K.-Y.
Lin, H.-C. Lu, Y.-C. Lin, L.-H. Lu, T.-W. Huang, R.-B. Wu, and H. Wang. 60-
GHz four-element phased-array transmit/receive system-in-package using phase
compensation techniques in 65-nm ip-chip CMOS process. Microwave Theory
and Techniques, IEEE Transactions on, 60(3):743 756, Mar. 2012.
[4.4] Hyung-Suk Ham, Young-Ro Kim, and Jong-Myung Woo. Active printed in-
verted f antenna for beam steering. In Antennas and Propagation Society In-
ternational Symposium, 2007 IEEE, pages 3348 3351, Jun. 2007.
[4.5] J. Mazotta, M. Delisio, and Jung-Chih Chiao. Quasi-optical discrete beam
steering grids. In Microwave Symposium Digest, 1999 IEEE MTT-S Interna-
tional, volume 4, pages 18251828, 1999.
[4.6] P.V. Rainey, M. Bain, R. Jackson, S.N. Mitchell, D. Linton, V.F. Fusco, and
H.S. Gamble. Design, manufacture and performance of monolithic pin diodes
for beam steered reectarray. In Antennas and Propagation, 2007. EuCAP
2007. The Second European Conference on, pages 1 5, nov. 2007.
[4.7] G.-L. Tan, R.E. Mihailovich, J.B. Hacker, J.F. DeNatale, and G.M. Rebeiz.
Low-loss 2- and 4-bit TTD MEMS phase shifters based on SP4T switches.
Microwave Theory and Techniques, IEEE Transactions on, 51(1):297304, Jan.
2003.
109
110 REFERENCES
[4.8] K. Topalli, O.A. Civi, S. Demir, S. Koc, and T. Akin. A monolithic phased
array using 3-bit distributed RF MEMS phase shifters. Microwave Theory and
Techniques, IEEE Transactions on, 56(2):270277, Feb. 2008.
[4.9] W. Gautier, V. Ziegler, A. Stehle, B. Schoenlinner, U. Prechtel, andW. Menzel.
RF-MEMS phased array antenna on low-loss LTCC substrate for Ka-band data
link. In Microwave Conference, 2009. EuMC 2009. European, pages 914 917,
Sep. 2009.
[4.10] C.-W. Baek, S. Song, J.-H. Park, S. Lee, J.-M. Kim, W. Choi, C. Cheon,
Y.-K. Kim, and Y. K.on. A V-band micromachined 2-d beam-steering antenna
driven by magnetic force with polymer-based hinges. Microwave Theory and
Techniques, IEEE Transactions on, 51(1):325331, Jan. 2003.
[4.11] H. Rajagopalan, Y. Rahmat-Samii, and W.A. Imbriale. RF MEMS actua-
ted recongurable reectarray patch-slot element. Antennas and Propagation,
IEEE Transactions on, 56(12):3689 3699, Dec. 2008.
[4.12] O. Bayraktar, K. Topalli, M. Unlu, O.A. Civi, S. Demir, and T. Akin. Beam
switching reectarray using RF MEMS technology. In Antennas and Propa-
gation, 2007. EuCAP 2007. The Second European Conference on, pages 1 6,
11-16 2007.
[4.13] J.J. Maciel, J.F. Slocum, J.K. Smith, and J. Turtle. MEMS electronically
steerable antennas for re control radars. Aerospace and Electronic Systems
Magazine, IEEE, 22(11):1720, Nov. 2007.
[4.14] E. Ojefors, Shi Cheng, K. From, I. Skarin, P. Hallbjorner, and A. Rydberg.
Electrically steerable single-layer microstrip traveling wave antenna with varac-
tor diode based phase shifters. Antennas and Propagation, IEEE Transactions
on, 55(9):2451 2460, Sep. 2007.
[4.15] N. Gokalp and O.A. Civi. Beam steerable traveling wave meander line antenna
using varactor diode for x-band applications. In Antennas and Propagation
REFERENCES 111
Society International Symposium, 2008. AP-S 2008. IEEE, pages 1 4, Jul.
2008.
[4.16] Hiroki Tanaka and Takashi Ohira. Beam-steerable planar array antennas
using varactor diodes for 60-ghz-band applications. In Microwave Conference,
2003. 33rd European, pages 10671070, Oct. 2003.
[4.17] D.D. King, C.M. Johnson, and C.M. Barrack. Precise control of ferrite phase
shifters. Microwave Theory and Techniques, IRE Transactions on, 7(2):229
233, Apr. 1959.
[4.18] R. Babbitt and R. Stern. Fabrication and performance of ferrite phase shifters
for millimeter frequencies. Magnetics, IEEE Transactions on, 15(6):17441746,
Nov. 1979.
[4.19] W.W.G. Hui, J.M. Bell, M.F. Iskander, and J.J. Lee. Low-cost microstrip-
line-based ferrite phase shifter design for phased array antenna applications.
Antennas and Wireless Propagation Letters, IEEE, 6:8689, 2007.
[4.20] I. Gresham, N. Jain, T. Budka, A. Alexanian, N. Kinayman, B. Ziegner,
S. Brown, and P. Staecker. A compact manufacturable 76-77-GHz radar module
for commercial acc applications. Microwave Theory and Techniques, IEEE
Transactions on, 49(1):44 58, Jan. 2001.
[4.21] J. Putnam, M. Barter, K. Wood, and J. LeBlanc. A monolithic GaAs PIN
switch network for a 77 GHz automotive collision warning radar. In Microwave
Symposium Digest, 1997., IEEE MTT-S International, volume 2, pages 753
756, Jun. 1997.
[4.22] G. Expósito-Domínguez, B. Sanadgol, and B. Zhou. Phase shifters for swit-
chable antenna arrays. XXVI URSI Spain, Sep. 2012.
[4.23] G. Expósito-Domínguez, J.-M. Fernández-González, M. Sierra-Castañer, and
M. Martínez-Vázquez. Phase shifters for switchable antenna arrays. COST-
VISTA IC 1102, Thessalonika, Greece, May 2013.
Chapter 5
Conclusions
This thesis has been carried out in the Radiation Group of the Signal, Systems
and Radiocommunication Department, in the Escuela Técnica Superior de Ingenie-
ros de Telecomunicación of the Technical University of Madrid from October 2009
to September 2013. The work presented in this thesis has been supervised by Pro-
fessor Dr. Manuel Sierra Castañer and co-directed by Dr. José Manuel Fernández
González.
Part of the work reported in this Ph. D. dissertation (specically, chapter 4) was
carried out within a research period (January-June 2012) at the following foreign
institution: Institute of Mobile and Satellite Communications Technology (IMST),
Kamp-Lintfort, Nordrhein-Westfalen, Germany.
A three month stay during 2013 was accomplished in Colorado University at
Boulder, United States of America, under the supervision of professor Dr. Zoya
Popovic. In this stay waveguide to microstrip transtitions at G band (140-325 GHz)
for power ampliers based on Double Heterojunction Bipolar Transistors (DHBT)
were desgined.
The research projects for which the work reported in this Ph. D. dissertation
has generated results are:
Project title: Communication systems for emergency environments (SICOMO-
RO) TEC2011-28789-C02-01. Financial institution: Ministerio de Educación
y Ciencia. Research institutions: Universidad Politécnica de Madrid.Duration:
2011 - 2014. Research in chief: Belén Galocha Iragüen.
Project title: Caracterización de Canales Radio, Optimización y Calibrado de
la Antena GEODA para Comunicaciones Espaciales (CROCANTE) TEC2008-
113
114 CONCLUSIONS
06736-C03-01. Financial institution: Ministerio de Educación y Ciencia. Re-
search institutions: Universidad Politécnica de Madrid Universidad de Vi-
go.Duration: 2009 - 2011. Research in chief: Leandro de Haro y Ariet.
Project title: STEALTH, plan Avanza I+D TSI-020100-2009-76. Financial
institution: Ministerio de Educación y Ciencia. Research institutions: Univer-
sidad Politécnica de Madrid TTI Santander.Duration: 2009 - 2010. Research
in chief at UPM: Manuel Sierra Castañer.
The main nancial support for the developement of this Ph. D. has come from
a grant of Technical University of Madrid CH/003/2011.
5.1. Original contributions
This thesis has yielded the following original contributions:
The design of a wideband steering antenna for satellite communication at X
band. In order to full the coverages and data rate throughput under the space
constraints and demanding antenna features several miniaturization and band-
width enhancement techniques are utilized. By combining these techniques,
new printed circuits and radiating elements are generated.
The size reduction of the mushrooms EBG structures to use them as mutual
coupling barriers. So far, the utilization of these structures is made in high
permitivity substrates. The use of high permitivity substrates enhances the
surface wave modes and lower the radiation eciency of the antenna, hence,
it is not very useful. With the EBG miniaturization originally presented in
this thesis the structures can be placed between the radiating elements in low
permitivity substrates.
The design of a symmetric 120 reective phase shifter, which includes a series-
shunt switch to obtain equal insertion loss for the both states (ON and OFF).
With this circuit the amplitude dierence between the switchable lobes is
minimized.
Future research lines 115
The design of a complex SPDT. The purpose of this circuit is to reduce the
leakage, increase the isolation and to match the three ports (input and outputs)
at any time regardles operation constrains.
5.2. Future research lines
Integration and interoperability of steerable antennas in X, Ku and Ka band
for airplanes. This task will be carried out in Airbus Military. The designs
must full the requirements in terms of EMC and EMI in order to operate
safely under any operation scheme.
Scale and development of the reective phase shifter and complex SPDT pre-
sented in Chapter 4 from 24 GHz to X band. In this way, future multibeam
antennas could be developed.
Design of an hybrid 3 dB/90 at 220 GHz in aWR4.3 based on the waveguide to
microstrip transition developed during the short stay in University of Colorado.
5.3. Publications
This Ph. D. dissertation has given rise to the following publications:
5.3.1. Journal publications
P. Padilla, J.M. Fernández-González, J.L. Padilla, G. Expósito-Domínguez, M.
Sierra-Castañer and B. Galocha-Iraguen. 'Comparison of dierent methods for
the experimental antenna phase center determination using a planar acquisi-
tion system ,' in Progress in Electromagnetics Research, Vol. 135. pp.331-346,
2013.
G. Expósito-Domínguez, J.M. Fernández-González, P. Padilla and M. Sierra-
Castañer. 'EBG size reduction for low permittivity substrates,' in Internatio-
nal Journal of Antennas and Propagation, Metamaterials special issue, 2012.
116 CONCLUSIONS
G. Expósito-Domínguez, J.M. Fernández-González, P. Padilla and M. Sierra-
Castañer. 'Mutual coupling reduction using EBG in steering antennas,' in
IEEE Antennas and Wireless propagation Letters, Vol.11. pp. 1265-1268,
2012.
J.M. Fernández-González, P. Padilla, G. Expósito-Domínguez and M. Sierra-
Castañer. 'Lightweight portable planar slot array antenna for satellite com-
munications in X band,' in Antennas and Wireless Propagation letters, Vol.10,
pp. 1409-1412, 2011.
G. Expósito-Domínguez, J.M. Fernández-González, P. Padilla and M. Sierra-
Castañer. 'Dual circular polarized steering antenna for satellite communica-
tions in X band,' in Progress In Electromagnetics Research, Vol. 122. pp.
61-76, 2012.
5.3.2. Conference contributions
5.3.2.1. International
G. Expósito-Domínguez, J.M. Fernández-González, M. Sierra-Castañer and
M. Martínez-Vázquez. 'Switches and phase shifters characterization for au-
tomotive switchable antennas,' in COST VISTA MC Meeting and Workshop.
Greece, Thessaloniki, May. 2013.
G. Expósito-Domínguez, J.M. Fernández-González, P. Padilla and M. Sierra-
Castañer. 'Revision of EBG metamaterials and active antennas,' in 6th Euro-
pean Conference on antennas and propagation - EuCAP 2012. Prague, Czech
Republic, Mar. 2012.
G. Expósito-Domínguez, J.M. Fernández-González, P. Padilla and M. Sierra-
Castañer. 'New EBG solutions for mutual coupling reduction,' in 6th European
Conference on antennas and propagation - EuCAP 2012. Prague, Czech Re-
public, Mar. 2012.
Publications 117
G. Expósito-Domínguez, J.M. Fernández-González, P. Padilla and M. Sierra-
Castañer. 'Mutual coupling reduction techniques in electronic steering an-
tennas in X band,' in 33rd ESA Antenna Workshop on Challenges for Space
Antenna Systems, Noordwijk, Oct. 2011.
G. Expósito-Domínguez, J.M. Fernández-González, P. Padilla and M. Sierra-
Castañer. 'Electronic steering antenna onboard for satellite communications
in X band,' in 5th European Conference on antennas and propagation - EuCAP
2011, Rome, Italy, pp. 2245 2248, Apr. 2011.
5.3.2.2. National
G. Expósito-Domínguez, B. Sanadgol, B. Zhou. 'Phase shifters for switchable
antenna arrays,' in XXVII Simposium Nacional de la Unión Cientíca Inter-
nacional de Radio - URSI 2012, Eche, Spain, Sep. 2012.
G. Expósito-Domínguez, J.M. Fernández-González, P. Padilla and M. Sierra-
Castañer. 'Matriz de Butler de banda ancha en banda X para antenas de haz
recongurable,' in XXVI Simposium Nacional de la Unión Cientíca Interna-
cional de Radio - URSI 2011, Leganes, Spain, Sep. 2011.
G. Expósito-Domínguez, A. Sánchez, N. Ortiz and M. Sierra-Castañer. 'An-
tena impresa con escaneo electrónico en sistemas embarcados para comuni-
caciones por satélite en banda X,' in XXV Simposium Nacional de la Unión
Cientíca Internacional de Radio - URSI 2010, Bilbao, Spain, Sep. 2010.