Wireless Transmission of Power for Sensors in

Context Aware Spaces

by
Jorge Ulises Martinez Araiza
B.S. Telecommunications Engineering
National Autonomous University of Mexico (2000)
Submitted to the Program in Media Arts and Sciences, School of
Architecture and Planning
in partial fulfillment of the requirements for the degree of
Master of Sciences in Media Arts and Sciences
at the
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
June 2002
( Massachusetts Institute of Technology 2002. All rights reserved.

Author ...........................................
Program in Media Arts and Sciences, School of Architecture and
Planning
May 10, 2002
Certified
by..........................
Edwin J. Selker
Associate Professor ofMedia Arts and Sciences
Thesis Supervisor

Accepted by ..........................
Andrew B. ippman
Chairperson
MASSACHUSETTS
INSTITUTE
OFTECHNOLOGY Department Committee on Graduate Students
JUN 2 7 2002
ARCHIVIES
LIBRARIES
 Wireless Transmission of Power for Sensors in Context
Aware Spaces
by
Jorge Ulises Martinez Araiza

Submitted to the Program in Media Arts and Sciences, School of Architecture and
Planning
on May 10, 2002, in partial fulfillment of the
requirements for the degree of
Master of Sciences in Media Arts and Sciences

Abstract
In the present thesis I create and use wireless power as an alternative to replace
wiring and batteries in certain new scenarios and environments. Two specific scenar-
ios will be highlighted and discussed that motivated this research: The Interactive
Electromechanical Necklace and the Wireless-Batteryless Electronic Sensors.
The objective is to wirelesslygather energy from one RF source and convert it into
usable DC power that is further applied to a set of low-power-demanding electronic
circuits. This idea improves the accomplishments of Radio Frequency Identification
(RFID) tags systems. The RF-to-DC conversion objective is accomplished by design-
ing and characterizing an element commonly known as a Rectenna, which consists
of an antenna and an associated rectification circuitry. The rectenna is fully char-
acterized in this dissertation and it is used for powering electronic lights, sounds,
transmitters, and different types of sensors as well.
The wireless power transmission system is presented in the first place with the
development of a special set of wearable beads for an interactive necklace. These
beads allow physical interaction between the necklace and electronic elements placed
in the environment. This scenario demonstrates that passive electronics without
batteries are possible. Next I also design and implement low-power sensors that will
use the energy delivered from the rectennas to perform active tasks. The switching
sensor provides visual/audio feedback to the user when there's a change in the state
of the sensed object (i.e. LEDs lit when a stapler runs out of staples); the humidity
sensor permits monitoring the humidity in the soil of a flower pot. The sensor actively
transmits the information of any of two possible stages (dry soil/humid soil) to its
environment. This scenario extends the capabilities of common RFID tags, where not
only they transmit information but also can react to their environment in an active
fashion.

Thesis Supervisor: Edwin J. Selker

Title: Associate Professor of Media Arts and Sciences

2
 The following people served as readers of this dissertation:

Thesis Reader .. ................... ........ .....................

Joseph Paradiso
Principal Research Scientist
MIT Media Laboratory
Technology Director, Things That Think Consortium

Thesis
Reader
................... ..................... ......
Yumi Mori
Visiting Scientist
MIT Media Laboratory

3
 Acknowledgments
To Ted Selker, for his invaluable guide throughout my graduate studies. It was
certainly a short period but it was a significant one. Thank you Ted for sharing your
experiences, your thoughts and thanks for those always interesting talks.
To my readers Joe Paradiso and Yumi Mori. Thank you for the suggestions
and invaluable comments. Thanks Joe for the constructive critique. Thanks Yumi
for the short but interesting talks and advice.
Thanks to Matt Reynolds, for having trusted in me, for your invaluable guide
and comments in the always-interesting field of RF. Gracias Mr. Reynolds!
To Barbara Dormady, Susan Knightly, and Linda Peterson, infinite thanks
for your infinite patience and help. You girls rock!
To thie Media Lab for being such a unique place, for I realized here that life
has a wide variety of interesting colors. To MIT for being the top best institute in
the world to study Engineering Sciences, also thanks to its professors that with their
lectures I have gained a little bit more of knowledge about what to be an engineer
means. Special thanks to professor Jin Au Kong,
To my beloved UNAM, that prepared me for being here. Thanks to Telmex,
especially to Carlos Slim, Javier Elguea and many more that made this possible,
so we could show that Mexicans also know how to play with Science.
To my teacher, mentor, roommate and friend Ante Salcedo. Thanks for having
me started into this beautiful world known as Electromagnetic Theory... Maxwell
rules!
To the Mexican Mafia, also known as the mischievous Pepetl band, for being
there in the good ones, the bitter ones and in the other ones: Ray the little Dic-
tator, all the Jorges and all the Juanes, Mr. nicknames Marquinho, Georgina
Pellizcona, Willy and his always accurate comments, Chema with the characteristic
laugh and Ernestin with his extreme adventures.
To Andrea Lockerd, Edison Thomaz, Surj Patel, Win Burleson, Teresa
Hyun-Yeul Lee and to those graduate students that slip from my memory, thanks.

4
 To the Holly Muffin Band, because no matter what the distance was they always
trusted in me and supported me with their emails and talks. Gustavin, Hunter,
Adriana. Edita, Saulinho (Saul), Marcela, Waldasco, Enriqueta... long live
the Pumas!
Thank you so much, holly muffin. To my beloved Keisha, thanks also.
To my brother, Omar, for I have realized here, at the distance, that there is no
other single best friend than he. My dear brother, you are the best, I am proud to be
your brother, your blood. There are still many experiences to have, and now I know
that it doesn't matter how bitter or sweet the future is, because it feels good to know
that I can always count on you no matter what.
To my father, Jose, for showing me the way, for his precise comments and because
I will never get tired of saying that you are the best father anyone could have had.
Some times, a man has to face huge changes in his life, but with guidance such as
yours, the path is easier to face. I am proud to be your son, Papi.
To my mother, Teresa, because no words are enough to describe what I feel, and
how grateful I am for all you have given me. To my mother because her tears and
laughs always gave me strength to keep going, and because since I was a baby she
prepared me for this moment. And there is no slight doubt in my mind that you are
the best mom I could have ever asked for, therefore muchas gracias Ma'.
To Ana, thank you. I will never get tired to say that I am really glad to have met
you. I am also glad that we have grown up at the same time, sometimes together and
sometimes not. I can't stop marvelling about the things you do and who you are,
and I can't stop marvelling myself on how much I have learned to love you. My dear
little best friend, gracias Enana.
To Rocio. Because I may or may not know exactly when, or where, or why, or in
which moment I started to love you this much, but the point is that I do. Thank you
for everything, for being there and for sometimes not being there, for having given me
the hugs, the virtual slaps on the face, for all the advice you gave me, for your voice
and all the rest you have shared with me. Thank you for being unique in the world,
for being so special, thank you for having accepted me for what I am. Thank you

5
Rocio, because without you I wouldn't have had the strength to survive some days to
accomplish this errand. And maybe I can do steal to the future a piece of a dream,
and then maybe inside that little piece these own dreams can fit in. And because the
best part of loving you is to know that you have been there, always, gracias Rocio.

And, as a few years ago, I won't thank Elvis, neither the Pope nor God.

6
And it all started here:

VxH(T, t) = ati(r, t) + J(T,t)

t- t)
VxE(, t) = 5B(~
V.D(T,t) = p(r, t)
V.B(, t) = 0

Thank you, Mr. Maxwell

7
Contents

Abstract
......
Acknowledgments
...........
........... ...........
...........
2

4
Table of Contents . . . . . . . . ........... ........... 8
List of Figures . . . . . . . . . ........... ........... 11

List of Tables . . . . . . . . . . . ........... ........... 14

1 Introduction 15
1.1 Motivation . 15

1.2 The history of power transmission by Radio Waves .......... 17

1.2.1 Early History. 17

1.2.2 The modern history . 19

1.3 Basic Principles .............................. 19

1.3.1 Free space transmission. 21

1.3.2 Choice of frequency ....................... 23
1.3.3 The rectenna. 23
1.4 RFID and WTPS ............................. 24
1.5 Thesis Objectives. 25

2 Theoretical Background 26
2.1 Classification of wireless systems ..................... 26
2.1.1 Wireless Local Area Networks .................. 28
2.1.2 Radio Frequency Identification (RFID) ............. 28
2.2 2450 MHz RFID Systems ..................... . 29
2.2.1 Operating Principle . 29

8
 2.2.2 Typical Tags. 30
2.2.3 Regulations ................ 31
2.3 The Radio Link and The Friis equation ........... 34
2.3.1 Effective Area. 35
2.3.2 Received power. 36
2.4 Radiated Power and Safety ................... 37

3 Rectenna Design 40
3.1 Microstrip radiators . 40
3.1.1 Advantages and limitations. 41
3.1.2 Microstrip patch antennas. 42
3.1.3 Design considerations for rectangular patch antennas. 43
3.2 Rectifier Design. 56
3.2.1 Small-signal Analysis of a Diode ............ 58
3.2.2 Diode Applications in Detector Circuits........ 61
3.3 Transmission-Line Impedance Matching . 62
3.3.1 Impedance Matching by Quarter-wave Transformer 63
3.3.2 Single-stub Matching. 64

4 Design and Experimental Results 66

4.1 Antenna Design .............. ........ . .66
4.1.1 Maxwell 2D Modeling ...... ........ . .69
4.1.2 Measured Antenna Performance ........ . .76
4.2 Rect'fier Design .............. . . . . . . . . . 76
4.2.1 Diode's Impedance Matching ........ . .82
4.2.2 Measured Rectification Data . . . ........ . .85
4.3 Radio Link calculation. ........ . .96
4.3.1 Measured Link Data ....... ........ ..100
4.3.2 Power Base Station ........ ........ ..102

9
5 Scenario Discussion 103
5.1 Interactive Necklace .............. ............. 103
5.2 Wireless Sensors ............ . . ................. 107

6 Conclusions 111
6.1 Fture Work ................................ 113

Bibliography 116

10
List of Figures

1-1 Transmission efficiency as a function of the parameter tau. 21

2-1 Basic Principle of microwave RFID Systems ........ . . . . 29

2-2 A block diagram for a general radio link .......... . . . . 34

2-3 IEEE C95.1-1992 Recommended safety density levels ... . 38

3-1 Microstrip Radiator . 40
3-2 Rectangular microstrip patch antenna ............ 42
:'-3 Examples of three typical feeding techniques ........ 45
.......

3-4 Radiating Slots Geometry .................. 48

3-5 Time and Frequency Domain Signals for a Rectifier . 57
3-6 Time and Frequency Domain Signals for a Detector . 57
3-7 Time and Frequency Domain Signals for a Mixer...... 58
3-8 Voltage-Current Curves ................... 59
3-9 Diode Equivalent Circuit Model ............... 60
3-10 Diode Detector Configuration ................ 61

3-11 Quarter-Wave transformer Coupling Sections ........ 63

3-12 Single Stub Matching Configuration ............. 64

4-1 Final Antenna Design ..................... 69

4-2 Simulation results for the edge-fed patch antenna ..... 70

4-3 Simulation results for the inset-fed patch antenna. 71

4-4 Simulation results for the modified inset-fed patch antenna 72

4-5 Predicted Radiation Pattern for the proposed patch antenna 73

11
 4-6 Measured values for the designed inset-fed patch antenna ...... . 74
4-7 Measured values for the tuned inset-fed patch antenna ........ 75
4-8 Manufactured Antenna .......................... 76
4-9 Proposed Rectification Circuit ...................... 77
4-10 Voltage Outputs for the two proposed diodes . ............. 78
4-11 Voltage-doubler Outputs for the two rectifiers ............. 79
4-12 Output power for different rectification efficiencies . .......... 80
4-13 Current Outputs for the proposed rectifier . .............. 81
4-14 Proposed Matching Sections ....................... 84
4-15 Simulation results for the HSMS8202Matching Circuit Section, RL =
2kQ .................................... 86
4-16 Simulation results for the HSMS2852 Matching Circuit Section, RL =
2kQ .................................... 87
4-17 Measured Input Return Loss for the HSMS2852 Matching Circuit Sec-
tion, RL = 2kQ ............................... 88
4-18 Measured VSWR for the HSMS2852Matching Circuit Section, RL = 2kQ 89
4-19 Measured Input Impedance for the HSMS2852 Matching Circuit Sec-
tion, RL = 2kQf .............................. 90
4-20 Rectification Circuit Board for the HSMS8202and Final Rectenna Sys-
tem ..................................... 91
4-21 Measured Input Return Loss for the HSMS8202 Matching Circuit Sec-
tion, RL = 2kQ2 .............................. 92
4-22 Measured VSWR for the HSMS8202Matching Circuit Section, RL = 2k 93
4-23 Measured Input Impedance for the HSMS8202 Matching Circuit Sec-
tion, RL = 2kQ. .............. ................ 94
4-24 Radiolink performance for Pt. = 30dBm, Gtx = 6dB .......... 99
4-25 Radiolink performance for Pt. = 30dBm, Gtx = 12dB ......... 100
4-26 Block Diagram for the Power Link Test ................. 101
4-27 Power Transmitter ............................ 102

12
5-1 Interactive Necklace System v2.0 ................... . 103
5-2 Transmitter and Receiver pair used in the Interactive Necklace Scenario. 104
5-3 Electronic Switch Schematic ................... ..... 105
5-4 Rectenna Implemented in the Interactive Necklace. .......... 106
5-5 CMOS Timer implemented in the Staples Sensor ............ 108
5-6 Piezo-buzzer circuit implemented .................... 109
5-7 Humidity Sensor. .............................. 110

13
List of Tables

2.1 Wireless System Frequencies ....................... 27

2.2 Frequency Limits for R.FID systems ................... 32
2.3 Limits for several frequency bands ................... . 33

3.1 Comparison between feeding techniques ................. 44

4.1 Design values for L and resulting resonant frequencies f ....... 67

4.2 SPICE parameters for the proposed diodes ............... 77
4.3 HSMS8202 Input impedances at different power input levels, RL = 100kQ 82
4.4 HSMS8202 Input impedances at different power input levels, RL = 2kQ 82
4.5 HSMS2852 Input impedances at different power input levels, RL = 100kQ 83
4.6 HSMS2852 Input impedances at different power input levels, RL = 2kQ 83
4.7 Measured Voltages for the HSMS2852 rectification circuit ....... 95
4.8 Measured Voltages for the HSMS8202 rectification circuit ....... 95
4.9 Measured and expected Powers for the experimental Radio Link . . . 101

14
 Chapter 1

Introduction

1.1 Motivation
The technological revolution has shown us a future of devices that can integrate
disparate Media (voice, data, images, etc.). Moreover, trends flood to the wireless
world, where huge amounts of data are transported by electromagnetic waves. Today
there is a plethora of such devices; if they don't yet fully cover all the expectations
they seem to be on the right track. PDA's, cellular telephones, laptop computers: all
these technologies try to integrate and to take advantage of various media.
At the same time, at the Media Lab we talk about and work on new intelligent
environments and context-aware spaces, which both try to make human lives easier.
These intelligent environments store information and use it afterwards; somewhere
they have stored a task user model, and with the help of an external sensor array,
they acquire data and create a representation of the ongoing events. This represen-
tation is compared to the model that is stored in memory, allowing the system to
react accordingly, making the system more context-aware. In other words, these are
computers that respond in an implicit way to the user needs, given the context in
which they're involved.
From the point of view that new technologies move towards portability, they
require the use of an onboard power supply, usually batteries. Batteries have allowed
us to build systems that are mobile, and more portability yields, among other things,

15
less hassle with cables. Running power lines back and forth is usually a demanding
task under certain circumstances. It is common these days to run power cords to
virtually any electronic system laying on a desk, namely computers, PDA's, etc.
Using this idea of wireless power transmission, it is possible to get rid of bulky cords,
and also remove the complications associated with the power interconnections and
the amount of cables needed.
Nevertheless wiring is the most confusing element of things; also batteries are often
the heaviest part and the most difficult to maintain (recharging them or replacing
them). We know how to make circuits that are less power consuming, which extends
battery life in a significant way. Accordingly, in this thesis the idea of battery-less
circuits will be explored, and also the notion of wireless power as an aide to existing
power sources will be presented and discussed.
Inside the broad radio electric spectrum, the ISM band (Industrial, Scientific
and Medical band located around the 2.4GHz frequency) is very popular given the
amount of equipment, services and standards that operate within it: wireless LAN
networks (IEEE 802.11), Bluetooth devices, wireless telephones, and even microwave
ovens! Today it is of common practice to have a wireless communication system
that works in this frequency range: examples range from such diverse environments
as universities, offices, and laboratories, even inside homes. Within 2.4GHz, the
availability of components is important and their efficiencies have been improved
over the years.
Therefore, from this rather strange combination of the need to provide energy to
small, low power-demanding devices such as certain types of sensors, and the need
to transmit radio frequencies through the air, the motivation for the present thesis
work arises: to use RF frequencies, in particular the ISM 2.4GHz band, to transmit
and provide power to electronic circuits (i.e. sensors) in a wireless fashion. Even
today, systems already use this wireless power transmission concept for RFID tag
purposes. An RFID tag is a device that is electronically programmed with unique
information, and a base station (reader) retrieves the information from the tag in a
wireless fashion. RFID Systems operating within the 2.4GHz band are categorized as

16
either active or passive. Active RFID tags are powered by an internal battery and are
typically read/write. Passive RFID tags operate without a separate external power
source and obtain operating power generated from the reader, by converting the RF
signal to DC voltage to supply the tag with energy.

1.2 The history of power transmission by Radio

Waves
Radio power transmission is defined as a three-step process in which:

* DC electrical power is converted into RF power

* The RF power is then transmitted through space to some distant point

* The power is collected and converted back into DC power at the receiving point.

Because the overall efficiency is necessarily the product of the individual efficiencies
associated with the three elements of the system, a premium is placed upon efficiency
conversion technology as well as upon aperture-to-aperture transfer efficiency. In
many respects the modern history of free space power transmission has been the
development of components for the transmitting and receiving ends of the system
that have sought to achieve the combined objectives of high efficiency, low cost, high
reliability, and low mass.

1.2.1 Early History

Power transmission by radio waves dates back to the early work of Heinrich Hertz [1].
Not only did he demonstrated electromagnetic wave propagation in free space by using
a complete system with a spark gap to generate high-frequency power and to detect
it at the receiving end, but he also used parabolic reflectors at both he transmitting
and receiving ends of the system.
Nikola Tesla carried out his experiments on power transmission by radio waves at
the turn of the century [2] [3]. He became interested in the broad concept of resonance

17
and sought to apply the principle to the transmission of electrical power from one
point to another without wires. By means of the alternating surges of current running
up and down a mast. Tesla hoped to set up oscillations of electrical energy over large
areas of the surface of the Earth, thus generate standing waves into which he would
immerse his receiving antennas at the optimum points.
Tesla carried out his first attempt to transmit power without wires at Colorado
Springs, Colorado, in 1899. He built a gigantic coil in a large square building over
which rose a 200-ft mast with a 3-ft diameter copper ball positioned at the top. The
Tesla coil was resonated at a frequency of 150,000 Hz and fed with 300 kW of low-
frequency power obtained from the Colorado Springs Electric Company. When the
RF output of the Tesla coil was fed to the mast an RF potential was produced on the
sphere that approached 100,000,000 V, according to Tesla. Such a high potential with
respect to the earth's surface resulted in very long and very visible discharges from
the sphere. There is, however, no clear record of how much of this power was radiated
into space and whether any significant amount of it was collected at a distant point.
Not until the early 1930's was another attempt made to transmit power without
wires, and this time experiments were cautiously made within the confines of a labo-
ratory. This experiment, performed in the Westinghouse Laboratory by H.V. Noble,
consisted of identical transmitting and receiving 100 MHz dipoles located about 25
ft from each other. No attempts were made to focus the energy, but several hundred
watts of power were transferred between the dipoles.
The major reason for the lack of serious interest in wireless power transmission
during the first fifty years of this century was that knowledgeable people realized
that efficient point-to-point transmission of power depended upon concentrating the
electromagnetic energy into a narrow beam. The only practical manner in which this
could be done would be to use electromagnetic energy of very short wavelengths and
collect energy over large diameters via optical reflectors or lenses. For the first thirty
five years of this century, devices did not exist to provide even a few milliwatts of
energy at these wavelengths. Sufficient power was not even available for experimental
work in communications and radar systems until the magnetron was developed during

18
 World War II [4].

1.2.2 The modern history

The modern history of free-space power transmission as it relates to microwaves in-
cludes not only the development of the technology of microwave power transmission
but also several proposed applications, ranging from RFID tags [5] to unmanned
vehicles [6].

Another disciplinary consideration are those factors that have confined the fre-
quency used in the development of the technology, and thus far its applications to
the 2.4-2.5 GHz band, reserved for industrial, scientific, medical (ISM) application.
Although this frequency was originally used because of the designation of the ISM
band for experimental purposes and the availability of components there, it is also a
frequency where the efficiencies of microwave comp,.lunts are very high. If the Earth's
atmosphere is involved, it constitutes an excellent compromise between greater prop-
agation attenuation in heavy rainfall at higher frequencies, and the larger aperture
dimensions necessary for efficient transmission at lower frequencies.

1.3 Basic Principles

Besides, modern times present certain facts that motivated the present research. The
evolving needs for ubiquitous computing and sensorial spaces could provide a good
platform for such types of wireless power transmission systems. A few important
driving facts are:

* High frequency electronics are now very accessible, cheap, low power, small and
more reliable. In the past, access to this kind of technology was restricted to
military industry and high-end companies. Every day, more electronic compo-
nents are capable of handling radio frequency signals, and high frequencies are
now more accessible because the packaged RF front-ends are becoming more
transparent to the end users that take advantage of them in the most diverse

19
 applications, from remote-controlled electronic doors up to advanced telecom-
munication systems.

* The use of low power electronics employing the latest CMOS technology is
in evidence almost everywhere. The trend now is to design new Integrated
Circuits (IC's) that are extremely low power consuming and, at the same time,
powerful and capable of doing complex tasks. Power saving issues arise almost
everywhere and the new electronic systems that run with batteries tend to last
longer thanks to the low power drain these circuits have.

* From the design point of view, thanks to modern computers and processing
capabilities it is relatively easy to design high quality high frequency electronics
and, even more, to custom-make antennas that can solve specific problems. In
the past such design problems were delegated to research environments and
academic institutions, but now there is software that significatively simplifies
the design process, where with just a few parameters, complex electronic circuits
such as filters and antennas can be designed with a high level of confidence and
reliability.

As means of transferring energy from one point to another, beamed microwave

power transmission has these features:

o No mass, either in the form of wires or ferrying vehicles, is required between

the source of energy and the point of consumption.

· Energy can be transferred at the velocity of light

· The direction of energy transfer can be rapidly changed

* Little energy is lost in the Earth's atmosphere at longer microwave wavelengths

1CMOS stands for Complementary Metal-Oxide Semiconductor. Today, CMOS technology is

the dominant semiconductor technology for microprocessors, memories and application specific in-
tegrated circuits (ASICs). The main advantage of CMOS over NMOS and bipolar technology is the
much smaller power dissipation. Unlike NMOS or bipolar circuits, a CMOS circuit has almost no
static power dissipation. Power is only dissipated in case the circuit actually switches. This allows
to integrate many more CMOS gates on an IC than in NMOS or bipolar technology, resulting in
much better performance.

20
1.3.1 Free space transmission.
The means for transporting the energy through free space by microwaves are the
transmitting and receiving apertures. The size and expense of these apertures has a
direct relationship to the wavelength that is being used, the distance over which energy
is being sent, and the desired efficiencyof transmission. Goubau and others [7,8] have
derived the followingrelationship between the aperture to aperture efficiency and a
parameter given in figure 1-1.

= V;AA-/AD (1.1)

'1

0
0 0.5' 1.0 1.5 2.0 2.5 3.0
Ait-
Zke

Figure 1-1: Transmission efficiency as a function of.the parameter tau.

where

* At is the transmitting aperture area

* Ar is the receiving aperture area

* A is the wavelength of the microwave power being transmitted

* D is the separation distance between the two apertures

21
 From (1.1), a simple expression for the transmitter and receiver aperture areas
can be derived with the assumption that the aperture sizes are equal. Under these
conditions:

At = Ar = TAD (1.2)

This is a revealing expression because it shows that the aperture area, rather than
its diameter, varies with wavelength, and the advantages of going to higher frequency
are diminished if the aperture areas are approximately equal, as they tend to be for
total overall economy.
However, there are applications where the reception area may be limited and
where a particular intensity of the incident microwave illumination is desired. Under
those circumstances we may use:

Pd = AtPt/A 2 D2 (1.3)

where

* Pd is the power density at the center of the receiving location

* Pt is the total radiated power from the transmitter

* At is the total area of the transmitting antenna

* A is the wavelength

* D is the separation between the apertures

With this situation it is seen that to achieve a desired value of Pd at the re-
ceiver site, while being constrained by a fixed transmitted power Pt, the transmitting
aperture area varies as the inverse square of the wavelength of the radiation. For
some applications, where the area available for a transmitter is limited, the short
wavelengths are very attractive.

22
1.3.2 Choice of frequency.
If there were complete freedom to select the best frequency for power transmission,
the items that would have to be considered are:

1. The size of the aperture as given by expression (1.3).

2. The dependency of overall system efficiency,including the components at the

two ends of the system, upon frequency

3. The heat radiation losses associated with the inefficiency of components

4. The losses due to blocking and fading elements in the air

5. The existing state of the art of available components and

6. The impact of the use of the selected frequency upon other users of the electro-
magnetical spectrum

1.3.3 The rectenna

The rectenna is a unique device that was conceived and developed for beamed mi-
crowavepower transmission [9]. It is spread out over the receiving aperture area and,
as its nature suggests, combines the functions of an antenna and a rectifier. In its
simple form, the rectenna consists of a collection of rectenna elements, each composed
of a receiving element (usually a half wave dipole) that feeds a low pass filter circuit
terminated in a rectifying diode.
The rectenna has many desirable characteristics. They include:

1. In its pure form, a relatively non-directional aperture analogous to that of a

single dipole, regardless of the size of the aperture; in this form the aperture col-
lection efficiencyis independent of the illumination density distribution across
the aperture

2. An overall efficiency from incident microwave power to DC power output that

has been measured at over 85% [61

23
 3. A low specific mass of from 1 to 2 kg per kilowatt of DC power output

4. Relative insensitivity of the overall efficiency to changes in the level of power

input or load impedance

5. Small amounts of the critical GaAs material required, less than 1/100 000 of
that required for a solar photovoltaic array of the same area.

The term "rectenna" is now used generically for the receiving aperture of any
beamed power transmission system that combines the function of capture and rec-
tification, even though in some formats there has been a departure from the "one
to one" relationship between dipoles and diodes in the original "pure" form of the
device. This departure results in directional sensitivity of the rectenna, which may
be tolerated for some applications.

1.4 RFID and WTPS

The Wireless Transmission of Power for Sensors (WTPS) proposed in the present dis-
sertation improves common RFID systems, now they can have sensors, computation,
and effectors. It is thus worth establishing the similarities and differences.
WTPS allows the existence of active elements (sensors) thanks to higher power
densities in comparison to the classic RFID systems. The raly of information can be
made in a different frequency band from that where the RF power is located. RFID
systems generally respond by modulating the reflection coefficient of their antenna.
Also, WTPS is centered around the far-field transmission principle, whereas RFID is
classically located in the near field region.
The proposed WTPS system and an RFID system are similar because both sys-
tems collect energy from the environment. Also they can provide a base station with
useful information (identification number) about the user. Finally both use Rectennas
as the transducer element between RF energy and DC power.

24
1.5 Thesis Objectives
This thesis has developed both the system that gathers the wireless power and the
sensors that will use this power to monitor certain environment variables, these devices
will not be limited to just store and identity operations, as with the RFID tags.
So for starters, the design of an antenna system that gathers the RF energy from
the environment and transforms it to DC power will be designed, constructed, and
extensively analyzed. The metrics to be analyzed are the amount of power delivered
to a given load and the conversion efficiency (defined as the amount of RF power
present at the receiver over DC power measured in the rectenna terminals). Chapter
2 will give a general theoretical background that is needed to understand the proposed
system in more detail. Chapter 3 will present in full detail the design-ruling physics
(equations) used for the development of the wireless power system, from the relations
utilized in the design of the microstrip antenna to basic concepts of signal rectification
using diodes. Chapter 4 presents all the experimental results obtained from the
implementation and realization of the rectenna system. Finally, Chapter 5 presents
the user scenarios with which the system was deployed and tested and discusses
sensors used within these scenarios. Chapter 6 gives the conclusions and future work
proposals for the Wireless Transmission of Power concepts.

25
Chapter 2

Theoretical Background

2.1 Classification of wireless systems

In the broadest sense, a wireless system allows the communication of information be-
tween two points without the use of a wired connection. This may be accomplished
using sonic, infrared, optical, or radio frequency energy. While early television remote
controllers used ultrasonic signals, very low data rates and poor immunity to inter-
ference make such systems a poor choice for modern applications. Infrared signals
can provide moderate data rates, but the fact that infrared radiation is easily blocked
by even small obstructions limits its use to short range indoor applications such as
remote controllers and local area data links. Similarly, optical signals propagating in
an unobstructed environment can provide moderate to high data rates, but require
a line-of-sight path, and cannot be used where dust, foliage, or fog can block the
signal. For these reasons, most modern wireless systems rely on RF and microwave
signals, usually in the UHF (100MHz) to millimeter wave (30GHz) frequency range.
Because of spectrum crowding and the need for higher bandwidth, the trend is to use
higher frequencies, so that the majority of wireless systems today operate at frequen-
cies ranging from about 800MHz to a few gigahertz. RF and microwave signals offer
wide bandwidths, and have the added advantage of being able to penetrate fog, dust,
foliage, and even buildings and vehicles to some extent.
One way to categorize wireless systems is according to the nature and placement

26
 Wireless System Operating Frequency
Advanced Mobile Phone Service (AMPS) T: 824-849MHz
R: 869-894MHz
Global System Mobile (European GSM) T: 880-915MHz
R: 925-960MHz
Personal Communication Services (PCS) T: 1710-1785MHz
R: 1805-1880MHz
US Paging 931-932MHz
Global Positioning Satellite (GPS) L1: 1575.42MHz
L2: 1227.60MHz
Direct Broadcast Satellite (DBS) 11.7-12.5GHz
Wireless Local Area Networks (WLANs) 902-928MHz
2.400-2.484GHz
5.725-5.850GHz
Local Multipoint Distribution Service (LMDS) 28GHz
US Industrial, Medical, and Scientific bands (ISM) 902-928MHz
2.400-2.484GHz
5.725-5.850GHz

Table 2.1: Wireless System Frequencies

of the users. In a point-to-point radio system, a single transmitter communicates with

a single receiver. Such systems generally use high-gain antennas in fixed positions to
maximize received power and minimize interference with other radios that may be
operating nearby in the same frequency range. Point-to-point radios are generally
used for dedicated data communications by utility companies and for connection of
cellular phone sites to a central switching office. Point-to-multipoint systems connect
a central station to a large number of possible receivers. The most common examples
are commercial AM and FM broadcasting radio and broadcast television. Multipoint-
to-multipoint systems allow simultaneous communication between individual users
(who may not be in fixed locations).
Wireless systems can be grouped according to their operating frequency. Table
2.1 lists the operating frequencies of some of the most common wireless systems.

27
2.1.1 Wireless Local Area Networks
Wireless local area networks (WLANs) provide connections between computers over
short distances. Typical indoor applications may be in hospitals, officebuildings, and
factories, where coverage distances are usually less than a few hundred feet. Outdoors,
in the absence of obstructions and with the use of high gain antennas, ranges up
to a few miles can be obtained. Wireless networks are especially useful when it
is impossibly or prohibitively expensive to place wiring in or between buildings, or
when only temporary access is needed between computers. Mobile computer users,
of course, can only be connected to a computer network by a wireless link.
Currently most commercial WLAN products in the United States operate in the
Industrial, Scientific, and Medical (ISM) frequency bands, and use either frequency-
hopping or direct-sequence spread spectrum techniques in accordance with IEEE
standard 802.11. Maximum bit rates range from 1 to 2 Mbps, which are much slower
than the data rates that can be achieved with wired Ethernet lines.

2.1.2 Radio Frequency Identification (RFID)

Radio Frequency Identification (RFID) systems are used for inventory tracking, ship-
ping, toll collection, personal security access, and other functions. As an example,
available now in several cities, automatic toll collection (ATC) uses a small transpon-
der in an automobile that can be interrogated by an RF system mounted at the
entrance of a highway or bridge. The transponder provides the vehicle's account
number, which is then debited, and a monthly bill sent to the driver. RFID systems
are much more specialized than cellular or WLAN systems, and use a wide range of
modulation methods, operating frequencies, and duplexing schemes.

28
 Reader Antenna

I I ,I
lOO
P on.0~~ 1 IIwIIIVJ.
,,c~t~t~ IRF
-_SnaI

Tag Data

Figure 2-1: Basic Principle of microwave RFID Systems

2.2 2450 MHz RFID Systems

2.2.1 Operating Principle

Microwave RFID systems have been in widespread use for over 10 years in trans-
portation applications (rail car tracking, toll collection and vehicular access control).
Systems operating in the UHF and microwave regions are divided between "active
power" and "passive power" tags. Operational range and functionality can be ex-
tended with active power (a battery in the tag), and passive power provides extended
life and lower costs. In the past, microwave RFID tags have been relatively complex
and expensive because of the challenge of processing microwave signals with CMOS
integrated circuits. This barrier of tag construction simplification has been overcome,
and the majority of present microwave RFID systems for article tracking use a single
integrated circuit and passive power. This has positive implications on cost, life-
time, and environmental situation. The basic operating principle of 2450 MHz RFID
systems is energy and data transmission using propagating radio signals ("E" field
transmission). A basic diagram of a microwave RFID system is shown in 2-1. This is
exactly the same principle as used in long-range radio communication systems. An
antenna on the reader generates a propagating radio wave, which is received by the
antenna in the tag. A passive power tag converts the signal to DC voltage to supply
the tag with energy. Data transmission from the reader to the tag is done by changing
one parameter of the transmitting field (amplitude, frequency or phase).

29
 The return transmission from the tag is accomplished by changing the load of the
tag's antenna (amplitude and/or phase). In this context, the microwave, 13.56 MHz
and < 135 kHz systems use the same principle. For microwave RFID systems, this
method is called modulated backscatter. Alternatively, a signal of different frequency
can be generated, modulated, and transmitted to the reader. These later types of
systems are referred to as using "active RF transmitter" tags.
Differing from inductive RFID systems (13.56 MHz and also < 135 kHz), UHF
systems and microwave RFID systems operate in the "far field" of the reader trans-
mission antenna. Achievable operating distances 0.5 to 12 meters is possible for pas-
sive power tags and beyond 30 meters for active power tags depending on microwave
frequency, regional regulations, and antenna details (for single reader antenna con-
figurations). Since the systems operate with the tags in the far field of the reader
antennas, the field strength decreases with the first exponent of the distance, which
means that the required transmission power increases with the second exponent of
the distance. Limitations are set by regulations. As a reference value and depending
on the chip technology, power consumption and data rate used, a passive 2.45 GHz
tag at 4 W ERP (Effective Radiated Power) yield some 0.5-1 meter range and active
tags range approx. up to 15-20 meter with standard power readers.
UHF and microwave signals are attenuated and reflected by materials contain-
ing water or human tissue, and are reflected by metallic objects. Unlike inductive
RFID systems, it is possible to design tags that work flat on metallic objects. Line of
sight transmission is not required for operation. UHF and microwave signals easily
penetrate wood, paper, cardboard, clothing, paint, dry, dirt, and similar materials.
Additionally, because of the short wavelength of the radio signal and reflective prop-
erties of metallic objects, reader systems can be designed to have high reliability in
regions with high metallic object content.

2.2.2 Typical Tags

Today 2450 MHz tags are available in many different shapes and with different func-
tionality, influenced by the applications' requirements. Unlike inductive RFID tags,

30
 which require substantial surface area, many turns of wire, or magnetic core material
to collect the magnetic field, UHF and microwave tags can be very small requiring
length in only one dimension. Thus, in addition to having longer range over the in-
ductive systems, the UHF and microwave tags can be easier to package and come in
a wider variety of configurations. Tag lengths of 2 to 10 cm are typical. The tag's
thickness is limited only by the thickness of the chip as the antenna can be fabricated
on thin flexible materials. Since tags operating in the E field do not require anten-
nas with extremely low impedances, inexpensive flexible antennas able to withstand
considerable bending are achievable.

Shape

Two main classes for 2450 MHz tags can be found:

* Rigid industrial tags for logistical purposes

* Thin and flexible smart label RFID products (also disposable products)

The expectation is that all different types of tags will also be used in future. It is
one of the benefits of 2450 MHz that many different shapes and sizes can be realized.
An additional benefit for article tracking and supply chain applications is that UHF
and microwave tags are widely used in the transportation function and thus a single
reader system can read tags on rail cars, inter-modal containers, trucks, trailers and
automobiles as well as parcels, packages, garments, pallets, boxes and the like.

2.2.3 Regulations
Most express delivery services, for example, use handheld terminals than can scan bar
codes on packages and relay information to a central station. Compared to barcode,
the impact of regulations is much greater for RFID. There are two areas that have to
be considered:

31
 Frequency Electric Field Magnetic Field Power Density
125 kHz 614 V/ni 130 A/n __
13.56 MHz 136 V/m 1.2 A/m
mW/cm 2
915 MHz -3

2450 MHz 8mW/cin

able 2.2: Frequency Limits for RFID systems

Human safety

In the area of hllumansafety compliance to the most important safety standards (world-
wide) has been verified by leading manufacturers. This has been clone for UHF and
microwavesystems as well as proximity and vicinity systems operating at 13.56 MHz.
RFID systems of all frequencies can be used without any doubts or reservations and
provides a safe means of RFID application. In simple words, it has been demonstrated
that system performance is not linited significantly by today's radiation safety regu-
lations and technology capabilities.
There are many standards in place. In the US, the standard ANSI C95.1-1991
was reaffirmed in 1997 and is presently undergoing revision. The latest standard is
IEEE Standard C95.1, 1999 Edition that reaffirms and updates the ANSI standard
and is summarized here. The standards have many facets, but the basic premise is
that no adverse effects occur at a specific absorption rate (SAR) level of 4 W/kg. A
safety factor of 10 is provided for occupational environments' and 50 for the general
public. The SAR is the rate at which energy is absorbed by the human body from
electromagnetic waves. The standard covers all frequencies of interest for RFID.
The standards also list the strengths of the fields permitted at the various fre-
quencies, derived from the allowed SAR values. Table 2.2 lists the limits for several
frequencies of interest (using the occupational limits):
In EUROPE, the EU Commission has issued on July 12th a Council Recom-
mendation for the general public on the limitation on EM Fields and specified basic
restrictions and reference levels. Either the reference levels or the basic restrictions
must be met.
The basic restrictions for localized SAR between 100 kHz and 10 GHz is 2 W/kg

32
for 'head and trunk' and 4 W/kg for 'limbs'.

Frequency (f) Electric Field (V/m) Magnetic Field (A/rn)

3 -- 150 kHz 87 5
0.15--- 1 MHz 87 0.73/f
/2
1 ---10 MHz 87/f' 0.73/f
10 - 400 MHz 28 0.073
400 - 2000 MHz 1.375f' /2 0)0037f1/2
2 -- 30() GHz 61 0.16

T'Ille 2.3: Limits for several frequllency an(ls

Radiation (to ensure proper use of the spectrum)

In general RFID technology does not yet face hlarnionized( iternational regulations
(as legislation is still a issue of the (lifferent local authorities). However. 2450MTNHz
is one of the most advanced frlequlenciesin terms of this allocation an(l harmonization
p)rocess, thlls offering global Ilse tlnder (lifferent environmental (e.g. noise level) and
regulatory conlditions. Operation is possible in the UHF and/or microwave regions
worldwi(le. Operation is possible unlder both unlicernse(l and licensed condlitions.
This is a short summary of the regullatory situation for radiation limits at 2450
MIHz:

United States: FCC 15.247

* Frequency: 902-928 MHz, 2400-2483.5 MHz. 5725-5850 MHz

* Transmitter power : 1 W

* Antenna gain : 6 dBi

* Resulting EIRP maximum: +36 dBm

* Communication protocol : spread spectrum, DSSS or FHSS

* Licensing: unlicensed

FCC Sub Part M, 90.350-90.357

33
 * Frequency: 902-904, 909.75-921.25 MHz

* Maximum EIRP: 30 W

* Licensing: licensed

Performance

Ranges of 0.5 to 5 meters from a single antenna are typical for passive power UHF and
microwave RFID systems. The ranges of the UHF (915 MHz) systems are generally
longer than the microwave systems (2450 MHz) because the wavelength is longer at
the lower frequency (equation 1.3), antennas collect more energy, and parasitic effects
are lower.
Compared to the low frequency inductive systems, the UHF and microwave sys-
tems can have longer range, higher data rates, smaller antennas, and more flexibility
in form factors and antenna designs. On the other hand tolerance to object penetra-
tion and reading range under no-line-of-site conditions can be better with inductive
systems. Date Rates with active UHF and Microwavesystems are higher than induc-
tive systems, whereas with passive systems, the difference might not be significant.

2.3 The Radio Link and The Friis equation

- R -

P G. Gr Pr
Figure 2-2: A block diagram for a general radio link

A general radio link is shown in figure 2-2, where the transmit power is Pt, the
transmit antenna gain is Gt, the receive antenna gain is G, and the received power
(delivered to a matched load) is Pr. The transmit and receive antennas are separated
by the distance R.

34
 By conservation of energy, the power density S,,v9 radiated by an isotropic antenna
(D = 1 = OdB) at a distance R is given by

2
Pt
Sav = 4 W/m 2 (2.1)

where Pt is the power radiated by the antenna. This result is deduced from the fact
that we must be able to recover all of the radiated power by integrating over a sphere
of radius R surrounding the antenna; since the power is distributed isotropically, and
the area of a sphere is 47rR2 , (2.1) follows. If the transmit antenna has a directivity
greater than 0 dB, we can find the radiated power density by multipliying by the
directivity, since directivity is defined as the ratio of the actual radiation intensity to
the equivalent isotropic radiation intensity. Also, if the transmit antenna has losses,
we can include the radiation efficiency factor, which has the effect of converting
directivity to gain. Thus, the general expression for the power density radiated by an
arbitrary transmit antenna is

Savg W/m2 (2.2)

2.3.1 Effective Area

For a receiving antenna, it is of interest to determine the received power for a given
incident plane wave field.
We expect that received power will be proportional to the power density, or Poynt-
ing vector, of the incident wave. Since the Poynting vector has dimensions of W/m 2 ,
and the received power Pr has dimensions of Watts, the proportionality constant must
have units of area. Thus we write:

Pr = AeSavg (2.3)

where Ae is defined as the effective aperture area of the receive antenna. The
effective aperture area has dimensions of m2 , and can be interpreted as the "capture
area" of a receive antenna, intercepting part of the incident power density radiated

35
toward the receive antenna. P in (2.3) is the power available at the terminal of the
receive antenna, as delivered to a matched load.
The maximum effective aperture area of an antenna can be shown to be related
to the directivity of the antenna as:

DA2
Ae = 4DA (2.4)

where A is the operating wavelength of the antenna. For electrical aperture an-
tennas' the effective aperture area is often close to the actual physical aperture area.
But for many other types of antennas, such as dipoles and loops, there is no sim-
ple relation between the physical cross-sectional area of the antenna and its effective
aperture area. The maximum effective aperture area as defined above does not in-
clude the effect of losses in the antenna, which can be accounted for by replacing D
in (2.4) with G, the gain, of the antenna.

2.3.2 Received power

If the power density obtained in (2.2) is incident on the receiver antenna, we can use
the concept of effectiveaperture area defined above to obtain the followingexpression
for the received power:

P = AeSavg= 47rR22 (2.5)

47rR

Where P units are Watts []. Next, (2.4) can be used to relate the effective area
to the directivity of the receive antenna. Again, the possibility of lossesin the receive
antenna can be accounted for by using the gain (rather than the directivity) of the
receive antenna. So the final result for the received power is

GtG A 2
P = (47rR)2
(4 ) Pt (2.6)

'Many types of antennas can be classified as aperture antennas, meaning that the antenna has a
well-defined aperture area from which radiation occurs. Examples include reflector antennas, horn
antennas, lens antennas, and array antennas.

36
 The units of Pr are Watts [W].
(2.6) is also known as the Friis equation, which addresses the fundamental question
of how much power is received by a radio antenna. In practice, the value given by (2.6)
should be interpreted as the maximum possible received power, as there are a number
of factors that can serve to reduce the received power in an actual radio system. These
include impedance mismatch at either antenna, polarization mismatch between them,
propagation effects leading to attenuation or depolarization, and multipath effects
that may cause partial cancellation of the received field.
Observe in (2.6) that the received power decreases as 1/R 2 as the separation
between transmitter anct receiver increases. As mentioned earlier this dependence is
a result of conservation of energy. While it may seem to be prohibitively large for
large distances, in fact the space decay of 1/R 2 is much better than the exponential
decrease in power due to losses in a wired communications link. This is because the
attenuation of power on a transmission line varies as e -2 z (where a is the voltage
attenuation constant of the line), and at large distances the exponential function
always decreases faster than an algebraic dependence like 1/R2. Thus for long distance
communications, radio links will perform better than wired links.

2.4 Radiated Power and Safety

Safety is a legitimate concern for users of wireless equipment, particularly in regard
to possible hazards caused by radiated electromagnetic fields. The body absorbs RF
and microwave energy and converts it to heat; as in the case of a microwave oven,
this heating occurs within the body, and may not be felt at low power levels. Such
heating is most dangerous in the brain, eyes, genitals, and stomach organs. Excessive
radiation can cause cataracts, cancer, or sterility. For this reason it is important to
define a safe radiation level standard, so that users of wireless equipment are not
exposed to harmful power levels.
ANSI/IEEE Standard C95.1-1992 [10]gives the U.S. safety standard for human ex-
posure to electromagnetic radiation. In the RF microwave frequency range of 100MHz

37
 L·.·L
If#

if)
,l
1:
..I
t: I I

,S

It

I0 1

I hIgl" e ; 4 al I

Figure 2-3: IEEE C95.1-1992 Recommended safety density levels

to 300GHz, exposure limits are set on the power density (in Watts/cm 2 ) as a function
of frequency, as shown in figure 2-3. The recommended safe power density limit is as
low as 0.2mW/cm2 at the lower end of this frequency range, because fields penetrate
the body more easily at low frequencies. At frequencies above 1.5GHz the power den-
sity limit rises to 10mW/cm2 , since most of the power absorption at these frequencies
occurs near the skin surface. By comparison, the sun radiates a power density as
high as 100mW/cm 2 on a clear day, but the effect of this radiation is much less severe
than a corresponding level of microwave frequency radiation because the sun heats
the outside of the body, with much of the generated heat being reabsorbed by the air,
while microwave power heats inside the body. At frequencies below 100MHz, elec-
tric and magnetic fields interact with the body differently than at higher frequencies,
hence separate limits are given for field components at these frequencies.
In addition to the above power density limits, the FCC sets limits on the total
radiated power of some specific wireless equipment. Vehicle-mounted cellular phones
(using an external antenna) are limited to a maximum radiated power of 3W. For
handheld cellular phones, the FCC has set an exclusionary power level of 0.76W,
below which phones are exempt from the ANSI standard on radiated power density.
Most cellular phones radiate a maximum of 0.6W, and newer PCS phones radiate
even less power. Although cellular and PCS base stations are limited to a total

38
effective radiated power of 500W, depending on antenna height and location, most
urban base stations radiate a maximum of 10W. Wireless products using the ISM
bands are limited to a maximum radiated power of 1W.
While other countries have different (sometimes lower) standards for RF and mi-
crowave exposure limits, most experts feel that the above limits represent safe levels
within a reasonable margin. Some researchers, however, are concerned that health
hazards may occur due to non-thermal effects of long-time exposure to even low levels
of microwave radiation [11].

39
Chapter 3

Rectenna Design

3.1 Microstrip radiators

Dilctr SubtMrte

Figure 3-1: Microstrip Radiator

As shown in figure 3-1, a microstrip antenna in its simplest configuration consists

of a radiating patch on one side of a dielectric substrate (er<10), which has a ground
plane on the other side. The Patch conductors, normally of copper or gold, can
assume virtually any shape, but regular shapes are used are generally used to simplify
the analysis and performance prediction. Ideally, the dielectric constant, o,, of the
substrate should be low (Er < 2.5), to enhance the fringe fields that account for the
radiation. However,other performance requirements may dictate the use of substrate

40
materials whose dielectric constants can have greater values.

3.1.1 Advantages and limitations.

Microstrip antennas have several advantages compared to conventional microwave
antennas, and therefore many applications cover the broad frequency range from
100MHz to 100GHz. Some of the principal advantages of microstrip antennas com-
pared to conventional microwave antennas are:

o Light weight, low volume, and thin profile configurations, which can be made
conformal

· Low fabrication cost; readily amenable to mass production

· Linear and circular polarizations are possible with simple a feed

* Dual frequency and dual-polarization antennas can be easily made

* No cavity back is required

· Can be easily integrated with microwave integrated circuits

* Feed lines and matching networks can be fabricated simultaneously with the
antenna structure.

However, microstrip antennas also have some limitations compared to conventional

microwave antennas:

* Narrow Bandwidth and associated tolerance problems

* Somewhat lower gain (6dB)

e Large ohmic loss in the feed structure of arrays

* Lower power handling capability (100W)

41
 * Microstrip antennas fabricated on a substrate with a high dielectric constant
are strongly preferred for easy integration with MMIC RF front-end circuitry.
However, use of high dielectric constant substrate leads to poor efficiency and
narrow bandwidth.

3.1.2 Microstrip patch antennas.

A microstrip patch antenna (MPA) consists of a conducting patch of any planar or
non- planar geometry on one side of a dielectric substrate with a ground plane on the
other side. Radiation characteristics have been calculated for a large number of patch
antennas [12]. Rectangular and circular patch antennas are widely used. Typically, a
patch antenna has a gain between 5 and 6 dB and exhibits a 3 dB beamwidth between
70" and 90° .
As shown in figure 3-2, the basic antenna element is a strip conductor of dimensions
L x W on a dielectric substrate of dielectric constant r and thickness h backed by a
ground plane. When the patch is excited by a feed, a charge distribution is established
on the underside of the patch metallization and the ground plane. At a particular
instant of time, the underside of the patch is positively charged and the ground plane
is negatively charged. The attractive forces between these sets of changes tend to hold
a large percentage of the changes between the two surfaces. However, the repulsive
force between positive charges on the patch pushes some of these charges toward the

Radiating Slots

I\'fr L i
--

7
-4 --h--
0 ////~~
I

Figure 3-2: Rectangular microstrip patch antenna

42
edges, resulting in large charge density at the edges. These charges are the source of
fringing fields and the associated radiation. Using a thicker substrate with a lower
value of dielectric constant can increase the fringing field and the radiated power.

3.1.3 Design considerations for rectangular patch antennas.

Usually, the overall goal of a design is to achieve specific performance characteristics
at a stipulated operating frequency. If a nlicrostrip antenna can achieve these overall
goals, then the first decision is to select a suitable antenna geometry. A rectangular
patch antenna can be designed using the followingconsiderations.

Feeding techniques.

Microstrip antennas have radiating elements on one side of a dielectric substrate, and
thus early microstrip antennas were fed either by a microstrip line or a coaxial probe
through the ground plane. Since then, a number of new feeding techniques have
been developed: prominent among these are coaxial feed, microstrip (coplanar) feed,
proximity-coupled microstrip feed, aperture-coupled microstrip feed, and coplanar
waveguide feed. These feed structures are summarized in table 3.1

Substrate Selection.

An important design step is to choose a suitable dielectric substrate of appropriate

thickness h and loss tangent. A thicker substrate, besides being mechanically strong,
will increase the radiated power, reduce conductor loss, and improve impedance band-
width. However, it will also increase the weight, dielectric loss, surface wave loss,
and extraneous radiations from the probe feed. A rectangular patch antenna stops
resonating for substrate thickness greater that 0.11A0o (Er = 2.55) due to inductive
reactance on the probe feed [13]. The substrate dielectric constant E, plays a role
similar to the substrate thickness. A low value of Erfor the substrate will increase the
fringing field at the patch periphery, and thus the radiated power. An increase in the
substrate thickness has a similar effect on antenna characteristics as a decrease in the

43
 r...,..mum r

P,
0
Qrn 0 Ce

fs ee
P4
c7B~o 0
o.
0
C3 cs
a-

5
0
0
0 c0

02
0

,
02
1.4
0 V
0
IC) a0 v
s-a
bO
-02 m
~a
- I ,0
C.) 0 ~ Q 0
Oao 8 .~~ Q
.4 0
'3L
.4
.)
v C.)

E0 0W
'4
0
uu
ml (1,
C)
-~ 0 cli
-Cou
- U
Q)
1-a
C2)
L
0' VQ .m
0D C.)
Ii

ao

9
1 a
C.
bp 0o I
4)
C.)

C) Sa
bO Z0 ).
o0
4 bO a
- s
12 n J
7
10 I:$ Ze
() g ,l C:=: ) O
M/
o
C4
4
.4

S.
C)
a4
-4
:Iu

iI--
a
a 0
cc Q) 0 -
>-~ to
D
12
-. 0 N 0 g .r
a; 5 =:
'.S " 4-a
-a *4 cc
r = Q. 4Q,
) II
P. 0
II.
;1

4
a~~~~c

P-' g r.c
e
1-1 -4 :

44
Ground plane Patch
(i::ilml plmic
I' ,idh

(a) Direct Feed

4.,t:,ji,1:
.
I,. '
(b) Coaxial Feed

2-

(c) Slot Feed

Figure 3-3: Examples of three typical feeding techniques

45
value of Er. Common dielectric materials used to fabricate microstrip antennas are
FR-4 (r = 4.2), GML1000 (r = 3.2), RT/Duroid (r = 2.2), and Teflon (r, = 2.1)

Element Width and Length

Patch width has a minor effect on the resonant frequency and radiation pattern of the
antenna. It affects the input resistance and bandwidth to a larger extent. A larger
patch width increases bandwidth, and radiation efficiency. With proper excitation
one may choose a patch width W (see figure 3-2) greater than the patch length L
without exciting undesired modes. It has been suggested that 1 < WIL < 2 [14,15].
The patch length determines the resonant frequency, and is a critical parameter
in design because of the inherent narrow bandwidth of the patch. To a zeroth-order
approximation, the patch length L for the TMlo0 mode is given by:

L= 2 (3.1)

The factor 1/VO' in (3.1) is due to the loading by the substrate and is strictly
valid for a very wide patch. In practice, a fraction of the fields lie outside the physical
dimensions Lx W of the patch. This is called the fringing field. The effect of such a
field along the edges y = 0 and y = W can be included through the effective constant
Crefor a microstrip line of width W on the given substrate. Thus equation (3.1)
becomes

L= (3.2)

The next improvement is obtained by including the effect of fringing fields at the
other ends of the patch, that is, along x = 0 and x = L. An additional line length
AL on either ends of the patch can describe this effect. Consequently:

'An incident wave of general polarization can be decomposed into two linearly polarized waves;
one with the electric field vector perpendicular to the plane of incidence which is the TE wave, and
one with the electric field vector parallel to the plane of incidence which is called the transverse
magnetic (TM) wave [16].

46
 C
fr = 2(L + 2AL)
C (3.3)
(3.3)
Another way to include the effects of fringing fields is to describe these in terms
of the effective dielectric constants for the width and length separately.

ere(W) 2 + 2 F(W/h) (3.4)

Ere(L)= + F(L/h) (3.5)

2 2

Where

F (h) = (1+ 12h/a)'/ 2 + 0.04(1- a/h)2, a/h<1

F (a) = (1 + 12h/a)'/2, a/h<l

Previous studies [17]show that AL calculated from

AL = 0.412h c,, + 0.300W/h + 0.264 (36)

,re- 0.258W/h + 0.813
Fits the real measured resonant frequency within 1.6% [18]

Radiation patterns and radiation resistance.

The radiation patterns of an antenna are of prime importance in determining most of

its radiation characteristics, which include beamwidth, beam shape, side lobe level,
directivity, polarization, and radiated power. The radiation from the patch can be
derived either from the E, field across the aperture between the patch and the ground
plane (using vector electric potential) of from the currents on the surface of the patch
conductor (employing vector magnetic potentials). The transmission line model and
cavity model employ the aperture field approach. In the, transmission line model, two
slots at the radiating edges are considered. The cavity model employs four slots along

47
the periphery of the patch. However, the contribution of the non-radiating slots to
the radiation pattern can be neglected for approximate calculations.

Radiation pattern based on a two-slot model

N
y
- h ..- i-. hk-

i
K-

! I
l
__q~

~*_ __ _ L ------ c

Figure 3-4: Radiating Slots Geometry

In this model, the radiation patterns for the TMlo mode are calculated by mod-
elling the antenna as a combination of two parallel slots of length W, width h, and
spaced a distance L apart, as suggested in figure 3-4. The radiation from the patch is
linearly polarized with the electric field directed along the patch length. If the voltage
across either radiating slot is taken as V0 , the radiation fields are obtained by multi-
plying the radiation pattern of one slot, with the array factor 2 cos(koL sin 9 cos b/2).
One obtains

e-jkor
Ee = -jAkoVoW i cos OFF2 (3.7)

o-e- jk o
E = -jkVoW 4r
47r
cos0sin FF 2 (3.8)

where

F = sinc(koh sin 0 cos 0/2)sinc(koL sin 0 sin 0/2)

48
 F2 = 2cos(koL sin 0 cos 0/2)

Equations (3.7) and (3.8) yield the following expressions for the radiation field in
the principal planes of a rectangular microstrip antenna operated in the TAMlomode.
For the k= 0 ° plane or E plane:

E,() = 0 (3.9)

e-jkor
Eo(O)= -jkoVoW 4irk FE(8) (3.10)

where

FE = sinc(koh sin 0/2)cos(koL sin 0/2)

For the b = 900 plane or H plane

Eo(O) = O (3.11)

e-jkor
E(O)=jkoVoW FH ( 6) (3.12)
47rr

where

FH = sinc(koW sin 0/2)cos0

The effect of the ground plane and substrate of the antenna on the radiation
patterns can be included by means of factors F3 (0) and F4 (0). For the E-plane
pattern:

2 cos 0V - sin
(3.13)
V,r- sin 2 0 - jr cos cot(kh/ - sin2U
)
and for the H-plane pattern:

49
 F4() = - 2cos (3.14)
Cos jc- - sin cot(koh/Er - sin 8)
The final radiation patterns are then, for the E-plane pattern

E@(@)
= -jkoVoW 2--r FE()F 3 (8) (3.15)

and for the H-plane pattern

e-jkor
E,(0) = jkoVoW 2-rr FH(O)F4 (O) (3.16)

Radiation pattern based on the electric surface current model.

In this model the patch metallization is replaced by the surface current distribution,
and the fields are solved for, taking into account the grounded substrate through an
exact Green's function. The radiation patterns are found to be [19,20]

EO(0,),) = sinO iJ47/e-jkorj(4(0) (3.17)

Eo(0,0)=-cos ( eikorJ(O, )F3 (8) (3.18)

where J(O, ) is the Fourier transform of the patch current. For the TM1 0 mode
with [20]

J(x,y) =x - sin(3x) for 0 < x < L, 0 < y < W (3.19)

The expression for J(O, k) is obtained as [20]

VO2\r erEcos(7rsin cos/(2V/6))sinc(0.5OW sin Osin O) (3.20)

Zo ko(sin 2 0 cos2 X- re)

where V/4 = 3I/ko is the effective dielectric constant of the microstrip line of
which the patch is a segment and Zo is the characteristic impedance corresponding

50
 -

to width W. For the TMiO mode the patch length is selected such that

L = A/2=okoL = /VE

The principal plane power patterns are then obtained as follows [20]. For the
= 0° plane or E-plane:

EO = 0 (3.21)

IEe(0)12 -cI[1 + e, cot 2 (khV)] cos2 (koL sin

2
8/2)
(cre - sin 8)2
(c, - sin2 8) cos2 8
(<ire
(Ere sin2
- sin )+
2 8) + E2COS2Cot2(h
cos2Ocot2(kohVE - in2
sin2) )

and for the ~ = 900 plane or H-plane

E =0 (3.23)

IE(8) 12=[1 + E cot2 (kohv/)]

cos 2 8/2) sinC2W sin (3.24)

(fre - sin2 8) cot2(kohv/ - Sin2+ cos2 n

The radiation patterns based on the two different radiation mechanisms described
above are similar. They have been compared with measured radiation patterns for
various values of substrate thickness with r = 2.5 [211.The aperture model, without
the factors F3 (8) and F4 (8) is found to give radiation patterns that compare well with
measurements.

Beamwidth.

The half power beamwidth of an antenna its equal to the angular width between
directions where the radiated field reduces to 1/V2/ of the maximum value. After

51
some approximations and algebra, one can obtain the following relationships:

OH = 2sin - 2 + koW (3.25)

E = 2sin- t f k0h )I/2 (3.26)

3ko2L2
' + koh2 (3.26)
Where OH and E are the half-power beamwidths in the H and E planes, re-
spectively. Choosing a smaller element, thus reducing L and W, can increase the
beamwidth of a microstrip element.

Directivity and Gain

The directivity is a measure of the directional properties of an antenna compared

of those of an isotropic antenna. The directivity is always greater that 1 since an
isotropic radiator is not directional. The directivity is defined as the ratio of the
maximum power density in the main beam direction to the average radiated power
density. The directivity of the patch antenna is expressed as:

D = 2Re(EHe-
1(IEeI E~H)|,o -
2 + IE12)327)=o
P /4irr2 Pr/47r
Where Pr is the radiated power, qo = 1207rf, and the radiation fields Ee and E,
where defined earlier. A simple approximate expression is

D 4(k W)2 (3.28)

iroGr
Where Gr is the radiation conductance of the patch, defined later.

Radiated power and radiation resistance.

The power radiated by an antenna can be obtained by integrating the real part of
the Poynting vector over the hemisphere above the patch, that is:

1
Pr r /2(IE2 + IE],2)r2sin ddk (3.29)

52
 If (3.22) and (3.24) are used for Ee and E*, one obtains [20]

Pr Vo2 60fre [2 [/2 COS2 ( sin cos /(2V-))

2
Pr0O2 Jo Jo (sin cos2 re,)2

sin(c2 ) (koW sin 0 sin 0/2)sinOdOdo

cos 2 0 sin2 0 (3.30)
(Er - sin2 0) cot2(koh e - sin 2 ) + cos2 0
(+r - sin 2 0) cos 2 0 cos 2 a
(E - sin2 0) + 2 cos 2 0 cot 2 (kohvE - sin2 )

A numerical integration is performed given the complexity of this equation. If

the effect of the substrate is neglected, an approximate closed-form expression can be
obtained for Pr. Thourode et al. [22] have obtained:

(Eoh)2 Ar 4
rA A2 B2
2
A A2
Pr= 23040 [(1-B) 1- + - - + l9 (3.31)

Where A = (VrW/Ao)2 , and the normalized resonant length B = (2L/Ao)2 . Equa-

tion (3.31) highlights the fact that the radiated power is directly proportional to
(koW) 2 and (koh) 2 .
The resonant radiation conductance Gr for a patch fed at an edge can be deter-
mined from power radiated Pr as follows:

Pr = G(EOh)2 = GrVo2 (3.32)

Rr=l/G, (3.33)
The expressions for Rr with an estimated accuracy of 10% average for h<0.03Ao
and e_ <10 are given by:

Rt =- 2 1201e (3.34)
2 Pr 12OI2

53
 Where Z0 is the characteristic impedance of the microstrip line of which the patch
is a segment, and for Er<5,

I2 = (koh)2[0.53 - 0.03795(koW/2)2 - 0.03553/Ere] (3.35)

Comparisons with numerical results for Rr have shown that (3.35) has better
accuracy for thin substrates with lower Er.
The radiation resistance is found to decrease with increase in substrate thickness
and patch width because radiated power increases. If the patch is fed at a distance
xf from one of the radiating edges, the input resistance is:

Ri = R, cos 2 (7rzf/L) (3.36)

The factor cos(7rxf/L) arises due to electric field variations for the dominant mode.

Radiation Efficiency Er

The radiation efficiency is defined as the ratio of the radiated power Pr to the input
power Pi, that is,

Pr
r P-i= (3.37)

The input power gets distributed in the form of radiated power, surface wave
power, and dissipation in the conductors and dielectric. Therefore, (3.37) can be
expressed as

Er P (3.38)
Pr + Pc+ Pd Pr
Here, Pd is the power lost in the lossy dielectric of the substrate, P is the power
lost due to the finite conductivity of metallization (also called copper loss), Pr is the
power radiated in the form of space wave, and Psur is the power lost in the form of
power carried away by the surface waves.
The dissipated power is generally small for low-loss substrates at microwave fre-

54
quencies, and one can write:

Er=
PrPr+P
+ Pur (3.39)

It has been observed that radiation efficiencydepends primarily on the substrate

thickness and permittivity, ar'd is not affected very much either by the patch shape
or the feed [231.Pozar has used an infinitesimal current source on the substrate and
obtained the following closed-form expressions for Pr and Psur [20]:

2 (1-
Pr = 40k2(koh) + 2 (340)

Psur= 30rk rX2 - 1) (3.41)

e[ + 2yj +koh 1+ ]

Where xo is the normalized phase constant for the TMlo surface wave. The
radiation efficiencyobtained using (3.40) is found to be accurate to at least 5% for
ho < /8.

Feed point Location.

After selecting the patch dimensions L and W for a given substrate, the next task is to
determine the feed point (o, yo) so as to obtain a good impedance match between the
generator impedance and the input impedance of the patch element. It is observed
in the literature [24] that the change in feed location gives rise to a chang: in the
input impedance and hence provides a simple method for impedance matching. If the
feed is located at x0 = xf and O<y<W, the input resistance at resonance for the
dominant TMlo mode can be expressed as:

R.,, = R, cos2 ('rxf/L), Rt_>Rn. (3.42)

Where xz is the inset distance from the radiating edge, and R. is the radiation
resistance at resonance when the patch is fed at a radiating edge. The radiation

55
resistance R, has been determined previously in (3.35). The inset distance x is
selected such that Ri,, is equal to the feed line impedance, usually taken to be 5052.
Although the feed point can be selected anywhere along the patch width, it is better
to choose y! = W/2 if W>L so that TM,0 (n odd) modes are not excited along with
1 mode. Kara [25] has suggested an expression for xf that does not need
the TA lo
calculation of radiation resistance. It is approximately given by:

Xf = L/2VcQ,7(L) (3.43)

Where ,,e(L) is defined by (3.5).

3.2 Rectifier Design

Detectors make use of the nonlinear characteristics of a solid-state device to bring
about frequency conversion to an input signal. The more nonlinear the device's I-V
characteristic curves are, the more efficient the detection process will be, that is, a
higher percentage of the signal power at the input frequency will be converted to
signal power at the output frequency.
The three basic types of frequency conversion circuits are

* A rectifier

* A detector

· A mixer

Each is described briefly as follows:

A rectifier is a circuit that converts the RF signal into a zero frequency signal
(i.e. a DC signal) with time and frequency-domain signals as shown in figure 3-5. The
rectifier is used for automatic gain control (AGC) circuits or power monitor circuits,
etc. A detector (also called a demodulator) is a circuit that demodulates a modulated

56
 Vc
0, 1-
Unmnodulalod
/o

o-.........
R
...
. octiftor
....
row-- ·-- · ----

Vi M. , A , Vo

V, , .. ,

Aol L' _

AO
_ ... I
I

Figure 3-5: Time and Frequency Domain Signals for a Rectifier

VIL
Modulated

V,
o-
A
D.-t...
I:eector ..
Je. , Vo

Vo A

iO .B
Figure 3-6: Time and Frequency Domain Signals for a Detector

carrier wave by discarding the carrier wave and outputting the modulating signal as
2
shown in figure 3-6. Detectors are used in circuits such as AM demodulators.
A mixer (also called a converter) is a circuit that converts an input signal to a
higher-frequency signal (called an up-converter) or to a lower-frequency signal (called
a down-converter) while ideally preserving all of the original signal characteristics
(such as sidebands, waveshape, etc.) This frequency conversion can be readily ob-
tained by mixing the input signal with another signal (called the Local Oscillator
signal, or LO for short) as shown in figure 3-7

2
A rectifier is a special case of a detector where proper filtering is used at the output to reject
all frequencies except for the DC component.

57
 r , fol
V I' fr * fo
1

IF
t Ifj RFl JN,s 0
A,~~~~~~/,, 1-1 1.-',

t
_I__. f~~

Figure 3-7: Time and Frequency Domain Signals for a Mixer

3.2.1 Small-signal Analysis of a Diode

In general, a diode can be considered to be a nonlinear resistor with its I-V charac-
teristic curve mathematically given by:

I(V) = I (e(V/nV
) - 1)
(3.44)

where:

* Vt=KT/q (Vt= 25mV at T = 293K)

* I8= Diode saturation current

* n =Ideality factor (1<n<2) which depends on the material and physical struc-
ture of the diode. For example, for a point-contact diode n = 2 whereas for a
Schottky barrier diode, n = 1.2, etc.

Figure 3-8 shows a sketch of a diode I-V characteristic curve, as described in

equation (3.44).
To perform a small signal analysis, we assume that the total voltage across the
diode (V) is composed of a DC bias voltage (Vo) and a small-signal RF voltage (v),

58
 IJ

+ L.

Figure 3-8: Voltage-Current Curves

that is:

V = VO+ v (3.45)

Substituting equation (3.45) in (3.44) and performing a Taylor series expansion

around the Q-point (Io, Vo)we obtain:

dl 1 2I
I = I(Vo + v) = I(Vo) + v lvo+ j2
-vVo+ ... (3.46)

where I(Vo) is the DC bias current given by:

Io = I(Vo) = Is (e(V /" )' - 1) (3.47)

The first order derivative corresponds to the dynamic conductance of the diode
Gd (the inverse of the junction resistance, Rj) and is given by:

I d=I Is /- Io+
+ Is (3.48)
Gd = R
Rj dv L
nV e(V/nv) n VT
The second-order derivative is given by:

22V, dd = G' 2 e(vOnvT) (3.49)

Idv dd (nVT)

59
 Is + Io Gd
Gd = (nVT) i = =-
2
nVT
Equation (3.46) can be now written as a DC current (lo) and an AC small signal
current (i):

I(v) = Io+ I (3.50)

where

i = vGd + 1/2(v 2 G'd) + ... (3.51)

The three term approximation for the diode current, as given by equation (3.50)
is known as the "small signal approximation". This means that for small signals (i.e.
v/nVT < 1), higher order terms (above the second order) for i may be truncated
without any loss of accuracy.
In practice, however, the diode's behavior also involves reactive effects caused by
junction and parasitic capacitances and lead inductances that are directly related to
the packaging as well as the structure of the diode. A typical equivalent circuit for
the diode is shown in figure 3-9. In this figure, (Cp, Lp) are the parasitic elements
due to packaging, Rs is the series resistance due to semiconductor neutral regions and
the contact areas, and (Rj, Cj) are the junction resistance and capacitance, which
are bias-dependent.

ILP

iL~
. .. .... Eq;wnt
Orcuit

..... ; ; ERs

C V)': ' R V)
Figure
D
9: iode Equivalent Circuit Model

Figure 3-9: Diode Equivalent Circuit Model

60
 Cm'a +'\F 4 t F t i S i i

'IN

Figure 3-10: Diode Detector Configuration

3.2.2 Diode Applications in Detector Circuits

A typical diode (as analyzed previously) can be used for many different applications,
particularly in a rectifier or a detector circuit, which is the primary focus of the
wireless power transmission system discussed so far.
A diode converts a portion of the input RF energy to a DC current that is propor-
tional to the input RF power. The type of detector circuit that uses an unmodulated
RF signal and converts it into a DC output signal may also be referred to as a recti-
fier. In this section, we will explore the magnitude of the DC current relative to the
input RF signal.
Let's consider a diode detector circuit (as shown in figure 3-10) and assume that
the diode is biased at a Q-point (10,VO)with an applied input small-signal RF voltage
(v) having a frequency (wo) and amplitude (vm) given by:

v(t) = mco(wot) (3.52)

From equations (3.50) and (3.51), the total current is composed of a DC bias (lo)
and an AC current given by:

I = I0o+ (3.53)

61
 where

i = GdVmcos(wot) + 1/2[G'dv cos2 (wot)] (3.54)

Using the identity

cos 2 (wot) 1 + cos(2wot)

2

we can write equation (3.54) as:

i = v2Gd/4 + vmGdCOs(wot)+ v2Gd cos(2wot)/4 (3.55)

Thus the total DC current is given by:

IDe = Io + v2Gd/4 (3.56)

If now the output RF signals of frequency (wot), (2wot), and other higher- order
harmonics, are filtered out using a simple low-pass filter, the remaining output term
will be composed of the bias current (0o) and a term equal to vGd'/4. The DC
rectified current is proportional to (2), which is the input RF power, desirable
in many applications such as automatic gain control, power monitors, and virtual
batteries [26]. Under this condition, the detector is said to operate in the square-law
region.

3.3 Transmission-Line Impedance Matching

Transmission lines are used for the transmission of power and information. For radio
frequency power transmission it is highly desirable that as much power as possible
is transmitted from the generator to the load and as little power as possible is lost
on the line itself. This will require that the load be matched to the characteristic
impedance of the line so that the standing-wave ratio on the line is as close to unity
as possible. For information transmission it is essential that the lines be matched

62
 l

Ri I

R1 2

Figure 3-11: Quarter-Wave transformer Coupling Sections

because reflections from mismatched loads and junctions will result in echoes and
will distort the information-carrying signal. In this section we discuss several methods
for impedance matching on lossless transmission lines. We note parenthetically that
the methods we develop will be of little consequence to power transmission at lower
frequencies, e.g. 60Hz lines, in as much as these lines are generally very short in
comparison to the millimeter wavelengths and the line losses are appreciable. Sixty-
hertz power-line circuits are usually analyzed in terms of equivalent lumped electrical
networks.

3.3.1 Impedance Matching by Quarter-wave Transformer

A simple method for matching a resistive load RL to a lossless transmission line of
a characteristic impedance Ro is to insert a quarter-wave transformer (as shown in
figure 3-11 with a characteristic impedance PR such that

Ri = ByoL (3.57)

Since the length of the quarter-wave line depends on wavelength, this matching
method is frequency-sensitive. Ordinarily, the main transmission line and the match-
ing line sections are essentially lossless. In that case, both Ro and Ro are purely real,
and equation (3.57) will have no solution if RL is replaced by a complex ZL. Hence

63
quarter-wave transformers are not useful for matching a complex load impedance to
a low-loss line.

3.3.2 Single-stub Matching

1 -- t d---

Figure 3-12: Single Stub Matching Configuration

We now tackle the problem of matching a load impedance ZL to a lossless line

that has a characteristic impedance Ro by placing a single short-circuited stub in
parallel with the line, as shown in figure 3-12. This is the single-stub method for
impedance matching. We need to determine the length of the stub, 1,and the distance
from the load, d, such that the impedance of the parallel combination to the right
of points B - B' equals Ro. Short-circuited stubs are usually used in preference
to open-circuited stubs because infinite terminating impedance is more difficult to
realize than zero terminating impedance for reasons of radiation from an open end
and coupling effects with neighboring objects. Moreover, a short-circuited stub of an
adjustable length and a constant characteristic resistance is much easier to construct
than an open-circuited one. Of course, the difference in the required length of an
open-circuited stub and that for a short-circuited stub is an odd multiple of a quarter-
wavelength.
The parallel combination of a line terminated in ZL and a stub at points B - B'

64
in figure 3-12 suggests that is advantageous to analyze the matching requirements in
terms of admittances, The basic requirement is

Y1 = YB + Y.
1 (3.58)
= Yo= Ro

In terms of normalized admittances, (3.58) becomes

1 = YB + Y (3.59)

where YB = RoYB for the load section and y, = RoY, is for the short circuited stub.
However,since the input admittance of a short-circuited stub is purely susceptive, y,
is purely imaginary. As a consequence, equation (3.59) can be satisfied only if

YB = 1 + jbB (3.60)

and

y, = -jbB (3.61)

where bB can be positive or negative. Our objectives, then, are to find the length
d such that the admittance, YB, of the load section looking to the right of terminals
B - B' has a unity real part and to find the length iB of the stub required to cancel
the imaginarypart.

65
Chapter 4

Design and Experimental Results

4.1 Antenna Design

Following the design procedures explained in chapter 3, the step-by-step procedure
for designing the rectenna is presented in detail. First, the microstrip antenna design
is discussed and analyzed, followedby the rectification section.

The first choice for the antenna design concerns the dielectric material. FR-4 has
been selected for its ready availability and moderate relative permittivity , = 4.6.
The selected substrate thickness is 0.062", so there's a good compromise between
radiated power and input impedance1 . The feeding technique was selected to be by
an inset on one of the antenna's edges, so during the design process there can be
more control on the input impedance of the antenna by varying the inset's width and
length.
As a starting point, the antenna's length is given by

L =-C c
2f, E/
Substituting the values c = 3x10 8m/s, ,f = 2.45x10 9 Hz and = 4.6 into the
equation we obtain:

'Cfi Section 3.1

66
 Patch Length [cm] Resonant Frequency [GHzJ
2.80 2.5755
2.85 2.5325
2.90 2.4909
2.95 2.4507
3.00 2.4118
3.05 2.3740
3.10 2.3375
3.15 2.3020
3.20 2.2676

Table 4.1: Design values for L and resulting resonant frequencies fr

2f, x10m/s
3x 108m/s
2x (2.45 x 109 Hz) vf4'
= 0.0285m
= 2.85cm

According to what it was discussed in section 3.1, to calculate the resonant fre-
quency with a ±1.6% accuracy, we have to use the following expressions:

2(L + 2)C
2(L+ 2AL)V
where

Ere(L)= 2
+ 2 F(L/h)
2

Programming this equations in MATLAB, we can select the value of L such that
the resonant frequency is 2.45GHz+39.2MHz. Table 4.1 contains the corresponding
values of L, and resulting fr where re = 4.2.
According Table 4.1, a close resonant frequency value to the desired 2.45GHz
frequency is obtained with a patch of a length L = 2.95crn. A more in depth analysis
gives us an exact solution with L = 2.9509cm. In practice, such a level of accuracy is

67
not easily obtained with conventional fabrication methods, therefore it was decided to
keep the original value of L = 2.95cm. If, during the lab tests, the resonant frequency
is out of the expected range, the antenna can be manually tuned by adding a small
tuning flag2 at one end of it, and by trimming it consecutively the desired resonance
can be achieved [24]. The width of the patch was selected to be W = 3.7cm to keep
a W/L relation of 1.25.
Another design issue was to select the proper feed point location (inset length)
such that the antenna's input impedance was 50Q. The proposed value is given by:

L
Xf

Using L = 2.95cm and a value of ere = 3.84, obtained from 3.5, the inset's length
is

L
Xf
0.0295m

= 0.00752m
= 7.52mm

Summarizing, the patch antenna to be analyzed with a simulation software will

have the followingphysical characteristics

* Length: 2.95cm

* Width: 3.7cm

* Inset length: 7.52mm

Figure 4-1 shows the dimensions of the designed antenna.

2
A tuning flag is a piece of transmission line terminated in an open circuit connected to one end
of the antenna.

68
 2.90 cm
E I

C i

4-Ab- 0.72 cm

Figure 4-1: Final Antenna Design

4.1.1 Maxwell 2D Modeling

Here the results of the simulation with Maxwell 2D [27 of the antenna previously
analyzed are presented. Using an edge port model, the resulting simulations for three
different patches are shown: a patch with the feed point located on the edge of the
patch (figure 4-2), a patch with the feed point located at the previously calculated
inset length (figure 4-3) and a patch with an optimized inset length for near-resistive
input impedance at 2.45GHz (figure 4-4).
The edge-fed patch results are shown in figure 4-2. It can be seen that this
antenna exhibited poor performance from its extremely high VSWR3 of 4.5, and the
non purely-resistive nature of the input impedance (Zi3 = 43 - j70[Q]) near the
frequency of interest contributes to this.
The performance for the proposed microstrip antenna patch is shown in figure 4-
3. Note that the resonant frequency is not located near the expected 2.45GHz value.
This can be seen in the Input Return Loss (Si) chart 4-3(a), where the resonance is
located around 2.475GHz. Also, the input impedance at this frequency is somewhat
high, about 70Q.
Under the simulator environment, the tuning was achieved by modifying both the

3
Voltage Standing Wave Ratio

69
 6,,a?

2.4 245 25 25

(a) SlI (b) VSWR

W*=

(c) Input Impedance

Figure 4-2: Simulation results for the edge-fed patch antenna

70
 s. pl

24 245 2 2

(a) Sll (b) VSWR

", I,~p

Z4 245 25

(c) Input Impedance

Figure 4-3: Simulation results for the inset-fed patch antenna

71
 Ste pot VW

t
O

IFfwq "

(a) Sll (b) VSWR

__e . I . I. . . . .

rI II I I I '!f

I
I
I

I IIS
I I
C
I

I
I
II

L . . . . .
Z4 Z41 242 Z43 244 245 246 247 248 249 25
FP-Iyq

(c) Input Impedance

Figure 4-4: Simulation results for the modified inset-fed patch antenna

72
 inset dimensions and the patch width. The results obtained from this simulation
are shown in figure 4-4. The new inset length is 7.2mm, and the separation from
the microstrip feeding line is about 2mm. It is important to also notice that the
patch width has been slightly modified, from 3.7cm to 3.73cm. It can be seen that
there is an important improvement in the resonant frequency; now the antenna is
close to 2.455GHz with an acceptable VSWR of approximately 1.2:1. The expected
bandwidth is about 40MHz, where the bounds are given by a VSWR of 2:1.

90 so FE pias
2: - --:; : Go--0 -- AX

240 300
270

Figure 4-5: Predicted Radiation Pattern for the proposed patch antenna. The calcu-
lated beamwidth is about 90°

Summarizing, the characteristics of the final antenna to be manufactured are

* Length: 2.95cm

* Width: 3.73cm

* Inset Length: 7.2mnm

The expected radiation pattern is shown in 4-5. The simulated gain of the antenna
is 6dB, a value that is typical for this kind of antennas.

73
)l:Reflectfon Log mg .t dl/ Ref 1,11 di I1: Reflectlon SWR 1.2 / Ref 1.111
02:Off n1
-. UT
,t'; T

2.-
2 7_ _ _ _ _ _ D H

1.'

.. V

Start 2 )3. I MHz Stop 2 911. MHz Start 2 11DN "Hz2 Stop2 611.11 I4z

(a) Input Return Loss (b) VSWR

1I:Ref lectlon Smith I U FS

>2:Off
I8I MHz
450
. mo
I 3pH

Start 2 318. 00 MHz Stop 2 89.0888MNz

(c) Input Impedance

Figure 4-6: Measured values for the designed inset-fed patch antenna

74
 ]:RsflecLTom Log
g l.l d/l Ref 1I, dl )t:Reflectlon SOR 1.2/ Ref 1.111
I 2:of¢
d

2.. - - 1 . j

I /.

I .

Start 2 3NI.1I1 Hz Stop 2 6l.1iH Pz

6: _AW\/
Start 2 31.lw Hz Stop2 611.lI z

(a) Input Return Loss (b) VSWR

*1:Ref lectlon Smith I U FS

>2: f
BII MHz
830
87o
. SpH

Start 2 28.808 MHz Stop 2 78188.8 MHz

(c) Input Impedance

Figure 4-7: Measured values for the tuned inset-fed patch antenna

75
 Figure 4·
1S: MilllI"Ifactured Antennatc'll);

4.1.2 Measured Antenna Performance

Th' I)liysi(ia ll ilit'lilla that w.s b)ililt is slowil ill irll(e -1-8. l)i t,, t la k (' anl RF

al,((li)i (ill be)r. te Iln'asillre('illt ()f t' alit'lilla's ra(liatio 1rns

pa)l
tte l(1 )t 1,
( i
'.
Il; ( Boh,tIlV.SW ;lll l ll t lI I('(
,I' n()})llir.('l
;illt', *' , I.t'l
( dtli , ~il, ,,t 'l' i,,',Al(,-
gi(s , 1 ( N(twOIwk AnalYzcr. Thl l'flsirellit If these (li.t (altl
an ovi(le liM'tii

il iat i(ti aI)(

it) t i (' x) ( t (l ailit'llll as lp rf(i'lial( e. Figili( 41-6s()ws t, Inc.a-

sllre(l Ilit Return Loss, VS\WR al(l U i i)lex li)lit 11iledilie of theailitellia that
W;L a.alvz(d(l nd (ilis('se((l I'( iOsllV. It (an )(' ()s( ,rv(d(lthat t' (ceiter f'(ell(v-y

is loca.t liat2.54GFHz. ot at te dlesire( 2 15G 1z vilhi. 'l'leref .. tie, ne (1 f'(, ia1 -

.ia1v tiliiiing antellna

tlhe' aris'es. Usinllg a tlinlg flag il )1, (',(l ()f ti,' aitlitena [2-1 (f

a fixe(l l()migitil(e.}v ('(lise(litive'v triniinig it. the ait(eniia ( i)(be tne(l to illgher

freqiieiicy. F(,llowing this l)roce((llre. te resiilts slw i .1-7 wNere()ttaie(1. N()w

th atennia is till(ed t 2.45GHz. withi an It)llt Rtirn Lo)ss (lose to -20()(113.an an

a)l)roxilmate ,)al(lwi(lt ()
of 3(NIHz. The ilnlt inil)eda('e it te (enter fr(qille ('l was
i

iil(Slre((l to e verv ()se to t}hede(sired 5)() vail, (55 + */).

4.2 Rectifier Design

lW'' rol)s((l r,(tifi(
eati(,i ('ir('llitrv is sow i figure('4-9

7(i
 RL
2k-l100k

Figure 4-9: Proposed Rectification Circuit

The diodes are placed in a voltage-doubler configuration. The rectifiers are Schottky-
barrier diodes, manufactured by Agilent technologies. These diodes are optimized for
high-frequency operation (f > 1GHz). Two types of diodes will be analyzed: the
HSMS8202 and the HSMS2852. They were selected following recommendations given
in a few Agilent Technologies Application Notes [28-30].
To simulate the diodes, their SPICE parameters are needed. These are summa-
rized in table 4.2

Parameter Units HSMS-2852 HSMS-8202

By V 3.8 7.3
CJo pF 0.18 0.18
EG eV 0.69 0.69
IBV A 3E-4 10E-5
Is A 3E-6 4.6E-8
N - 1.06 1.09
Rs S 25 6
PB(VJ) V 0.35 0.5
PT(XTI) - 2 N/A
M - 0.5 0.5

Table 4.2: SPICE parameters for the proposed diodes

Using the SPICE parameters, the first analysis of rectification properties of an

HSMS8202diode in a single detector configuration was made using Agilent Technolo-
gies AppCAD design software. The resulting performance is shown in 4-10(a). Using
the HSMS2852, the obtained simulated performance is presented in 4-10(b)

77
 Inp Power [dBm

(a) HSMS8202

&
I
I

putPoWaIdBm)

(b) HSMS2852

Figure 4-10: Voltage Outputs for the two proposed diodes

78
 14

12

10 ,

2
Id
O I
2 4 6 8 10 12 14 16 18 20
InptA
Po [dml

(a) HSMS8202

&
$
J.

InA Powe Idm

(b) HSMS2852

Figure 4-11: Voltage-doubler Outputs for the two rectifiers

79
 Note that the presented voltage values correspond to a rectifier diode in a single
configuration. In theory, the voltage-doubler configuration doubles the output voltage
compared to the single configuration at the same input power [31]. Thus the expected
performances are presented in figures 4-11(a) and 4-11(b). According to these results,
so far no significant difference can be seen between in either one of the diodes. Thus
any further analysis (unless stated otherwise) is assumed to had been made using an
HSMS8202 for the sake of simplicity.
Considering that the SPICE model assumes a perfect match between the antenna
and the rectification circuitry, the efficiency can be as high as 100% (ideal case) or,
assuming a worst-case scenario with a poor match at the rectifier's input, the overall
efficiency of the detector can be as low as 50% [6, 26, 32]. The resulting power outputs
curves are then plotted in 4-12 for four different possible efficiencies according to
literature [26].

II

2 4 6 8 to t2 4 o lo 20
W4 P I.dBE.

Figure 4-12: Output power for different rectification efficiencies

Combining the previously obtained output voltages and output powers, and also
remembering that P = VI, a final set of curves related to the current that can be
delivered to the matched load given a specific input power is shown in figure 4-13.
This set of curves characterizes the performance of the wireless Rectenna system.

80
 InpAPower(dm

(a) Unregulated

Poor ldmJ

(b) Regulated (V,,t = 3.3V)

Figure 4-13: Current Outputs for the proposed rectifier: a) with no voltage regulation
and b) with a fixed voltage output

81
 Frequency OdBm 5dBnm 1OdBm
2435 7-88.4j 8.4-92j 9.4-95j
2440 7.5-89j 8.6-92j 9.7-95j
2445 7.6-89.3j 9-93j 10-95.8j
2450 7.9-89.7j 9.2-93.4j 10.3-96j
2455 8-90j 9.3-93j 10.5-96.7j
2460 8.2-90j 9.7-94j 10.8-97j
2465 8.5-90.8 9.9-94.5j 11-97.7j

Table 4.3: HSMS8202 Input impedances at different power input levels, RL = 100kQ

Frequency OdBm 5dBm 10dBm

2435 30-101j 40.8-101.8j 49.5-101j
2440 30.5-101.8j 41.4-101.6j 50-lO1j
2445 31.2-102.5j 42.2-102.2j 51-101j
2450 31.5-102.6j 43-102.4j 52-101.9j
2455 32.2-103.2j 43.5-102.7j 52.5-102j
2460 32.8-103.6j 44.3-103.3j 53.6-102.4j
2465 J 33.4-104j 45-103.3j 54.2-102.5j

Table 4.4: HSMS8202 Input impedances at different power input levels, RL = 2kQ

4.2.1 Diode's Impedance Matching

To ensure maximum power transfer, a good impedance matching network has to
be realized between the antenna and the rectifier's input. First, the diode's input
impedance has to be measured at the expected frequency range (2.4-2.5GHz). For
the HSMS8202diode pair, the results obtained with two different loads are shown
using two different load values: 100kQ (table 4.3) and 2kQ (table 4.4).

82
 For the HSMS2852the corresponding obtained impedances are:

Frequency OdBm 5dBm 10dBm

2435 8.5-66.9j 11.0-68.7j 28.7-62.9j
2440 8.7-66.9j 11.4-68.6j 28.7-63j
2445 8.9-67j 11.6-68.9j 29-63. lj
2450 9-67.2j 11.8-68.9j 29.0-63j
2455 9.1-67.2j 12-69j 29.1-63j
2460 9.3-67.4j 12.2-69j 29.3-63.1j
2465 9.5-67.4j 12.37-69j 29.3-63j

Table 4.5: HSMS2852 Input impedances at different power input levels, RL = 100kQ

Frequency OdBm 5dBm 10dBm

2435 22.8-44.6j 26-45j 30-44j
2440 23.0-44.6j 26.2-44.8j 30.3-44j
2445 23.1-44.6j 26.4-44.8j 30.5-43.9j
2450 23.2-44.6j 26.6-44.8j 30.7-43.6j
2455 23.4-44.6j 26.7-44.6j 30.8-43.7j
2460 23.6-44.6j 26.9-44.7j 31-43.6j
2465 23.7-44.5j 27-44.6j 31.3-43.4j

Table 4.6: HSMS2852 Input impedances at different power input levels, RL = 2kQ

To achieve a good impedance match between the elements, a single-stub matching

network was designed using microstrip techniques, as previously discussed in section
3.3.2. Figure 4-14 shows the proposed matching networks for both rectification diodes.
Given the fact that the diode's input impedance is somewhat sensitive to the
input power, as it was shown previously in 4.2.1, the design was more focused on
the medium-to-low end of the power range, which is 5dBm to OdBm. Thus, it was
assumed that Pin=5dBm for the proposed matching sections shown in 4-14. The
assumed input impedances correspond to the values obtained in tables 4.4 and 4.6,
that correspond to input impedances with a 2kQ output load. The Smith Chart in
figure 4-15(a) shows the expected performance for the HSMS8202 matching section
at OdBm, whereas figures 4-15(b) and 4-15(c) show the resulting matched impedance

83
 d10.127k HSMS-8202
I C_ ·
rJ.

I
Rs 1 10( D2
C2 RL
D1
-

'ta) I Id2O.88X 100 pF 2k-100k

I
." . I
(a) HSMS8202

dl=0.051x HSMS-2852
1[= aC
t--- Nir
... I, , -' , - .,--

Rs i 10(c D2

L-
1~ C2 - vW RL
d2=0.100i D1 100pF 2k-100k

4 l , _ __,~ h

I
*

lb) HSMS2852

Figure 4-14: Proposed Matching Sections. Stub dimensions are given in wavelengths.

84
for the other two power levels of 5dBm and 10dBm respectively. Figure 4-16 shows
the simulated impedance match for the HSMS2852.

4.2.2 Measured Rectification Data

Impedance Matching

The tuning of the matching circuit was made manually with the aide of a network
analyzer. The rectification and proper tuning circuit can be seen in 4-20. Given the
fact that the input impedance of the diodes is sensitive to the input power, as it was
observed in the previous section, the matching had to be made with a fixed input
power. It was decided then to perform manual tuning with an input power of 5dBm,
as was observed that as the input power increased, the performance improved even
without a proper matching section. Thus for the input power levels starting at OdBm,
the impedance match will improve as the power gets close to 5dBm, and for higher
powers it should behave similarly.
For the HSMS2852 the results are shown in 4-17, 4-18, and 4-19. It is important
to notice the poor matching achieved here. The graphs 4-21, 4-22, and 4-23 show
the coupling obtained in the lab with the HSMS8202diode. The VSWR obtained
is 1.15:1, it also is interesting to notice that this value goes as low as 1.25:1 at the
highest input power value (lOdBm). This is a more than acceptable impedance match
over the whole power range.

Voltage Conversion

Table 4.8 shows the obtained voltages using the HSMS8202. The rectification effi-
ciency is higher around the medium power levels, which is expected due to the fact
that the matching section is optimized at those levels. The efficiency values range
from 50% in the lower power end and up to 70% near the medium-to-high power
levels.
It is important to notice that the other diode pair, the HSMS2852, had a poor
performance for the sane levels. This results are expected, given the poor matching

85
 (a) OdBm (b) 5dBm

(c) lOdBm

Figure 4-15: Simulation results for the HSMS8202Matching Circuit Section, RL

2kQ

86
 (a) OdBm (b) 5dBm

(c) 10dBm

Figure 4-16: Simulation results for the HSMS2852Matching Circuit Section, RL =

2kf

87
 I
d

3
d
I:Rof lectlon
.hm"

r-

-
-
_
Log Meg 13.

"agI:
d Ref .H3 dl

II 2 53.3I
12.43
PHa
dl
d
is

N-
PI:Rf lect on
&4

r~,, u! T
nmm

- -A
Lao Moo 16.0 dl/ Rat

_
fkje:r
1.11 d

21

I
.

-2

_4 _ _ -
4
1

Start 2 21l.3i3 MHZ Stop 2 71.13 M14z Start 2 211.1 I4HZ Stop 2 711.3# wlz
5.

(a) OdBm (b) 5dBm

ei
MP:Reflectfor Log Mg 1.3 di/ RQe 1.3 di
,. -- l ,
.

Hi
I -_~wl 2' B.11. I HZ
d
is 62 Wd
4
7-
I

21

1:-
71 __ I
I
_2,

-3
,t

---
Start 2 21. SU IMHz Stoo 2 71.111 PMHz

(c) 10dBm

Figure 4-17: Measured Input Return Loss for the HSMS2852 Matching Circuit Sec-
tion, RL = 2kQ

88
 1:Reflection SVR 1.2 / Ref IIN I :Rflection SR 1.2 / Ref 1.111
2:off
M/Eul:Nlr 251.A MHz
I ___
2.&3
2

2.2 e_.- - _ -
2

t A.-

1.* - -,. I - ~ S--

-

1
Start 2 2.HII RMz Stop 2 701. IN uz

(a) OdBm (b) 5dBm

>I:RelectIon SVR 1.2 / Ref t.11

>2: Off

Start 2 21 N I4z Stop Z 7IN.lU pz

(c) 10dBm

Figure 4-18: Measured VSWR for the HSMS2852 Matching Circuit Section, RL =
2kQ

89
 *: Rfilectlon SmIth I U 01:Rsflectlon Samth I U FS
02:Off "2:Off
a MHz In MHz
12a 71a
Sk0 M6o
s]pr Isp,

Start 2 il0tM I MHz Stop I3ll. I*4z Start 2 11.m PlHz SWo)Ill 11 tz

(a) OdBm (b) 5dBm

S. th I rs
>2: ff
11 MHz
SSo
6So
isf

Start 2 1 1N Mz Stoe 11.11 z

(c) lOdBm

Figure 4-19: Measured Input Impedance for the HSMS2852 Matching Circuit Section,
RL = 2kQm

90
 (a) Rectifier circuit (b) Detail of the Matching Section

(c) Complete Rectenna

Figure 4-20: Rectification Circuit Board for the HSMS8202 and Final Rectenna Sys-
tem.

91
 I:Rf Iectlon Log Nrg 1.I dl/ Ref I. dl I:Reflectlon Log Nbg 1.1 dl/ Ref I.II dl
P'2:Off ta7n··

dB _ _ I I rl:JI 24 5.l IiiZ

1 - _ _2
- - _6.- - 5 dl

-- - -
- -. -, . -

-. - - - - - - - - I I
_ -:" - - _ _
Start2 21 .11l Mz Sto 2 7.111 KN Start 2 211.111 Hz Sto 2 711.1 Wa

(a) OdBm (b) 5dBm

l: Re iecto Lc Pr.e ;1. di/ Ref 1. di

>2: ; _I

dl
z9. odl

2i- -
i t - - 4 [ 4I i- - t I

I~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

2c
1 1 I" 1 /
I· -- L
J8 -'

St1rt2 211.111 Nz7 Stop 2 71.111 H4z

(c) OdBm

Figure 4-21: Measured Input Return Loss for the HSMS8202 Matching Circuit Sec-
tion, R = 2kQ

92
 )l:Rflectlon SW 1.2 / Ref 1.H l:Rmflectlon soR 1.2 / Rat 1.111
'2:Dff
.... t:m1 l !11

2.2 _ ____

1. - - -

I - - I -
1
Start 2 21.3l K rz Sto 2 71.1t Rmz Start 2 211.11114z Stop 2 Ill7N.i Hz

(a) OdBm (b) 5dBm

1I:, fieto SWR 3.2 / e'f .1f1

1. 21

2.

2.4

2. - \

1.2 - -

S'.art 2 2.11M. Mz Stop 2 7ll.l1 mHi

(c) OdBm

Figure 4-22: Measured VSWR for the HSMS8202 Matching Circuit Section, RL =

2kQ

93
 bl:RefIletlon Silth I U FS Il:Reflectlon SRtIth I U rS
:Off 2: Off

II I
94a0
ItpF .2pH

Start 2 2.0 ilHz Stop 2 711.39 4az Start 2 211.31 tMHz Stop 2 7110.3 1iz

(a) OdBm (b) 5dBm

s.lth I u rS
i2:rf
tU MHz
.So
120
. iy

Start 2 21. 011 WI Stop 2 711.111Oz

(c) lOdBm

Figure 4-23: Measured Input Impedance for the HSMS8202 Matching Circuit Section,
R = 2kQ

94
Pi. [dBm] P [mW] Vt V] P,, [mW] P,,t [dBm] Efficiency [%]
RL = 2kQ - (Pin/Pot)x 100
0 1.0000 0.36 0.0648 -11.8842 6.4800
2 1.5849 0.44 0.0968 -10.1412 6.1077
4 2.5119 0.56 0.1568 -8.0465 6.2423
6 3.9811 0.70 0.2450 -6.1083 6.1541
8 6.3096 0.90 0.4050 -3.9254 6.4188
10 10.0000 1.16 0.6728 -1.7211 6.7280
12 15.8489 1.32 0. 712 -0.5988 5.4969
14 25.1189 1.63 1.3285 1.2335 5.2887
16 39.8107 1.92 1.8432 2.6557 4.6299
18 63.0957 2.16 2.3328 3.6788 3.6972
20 100.0000 2.16 2.3328 3.6788 2.3328

Table 4.7: Measured Voltages for the HSMS2852 rectification circuit

P,,n [dBm] Pin [mW] V [V] P,t [mW] Pt [dBm] Efficiency [%]
RL = 2kQ = (P/Pt)x 100
0 1.0000 1.00 0.5000 -3.0103 50.0000
2 1.5849 1.33 0.8845 -0.5333 55.8050
4 2.5119 1.70 1.4450 1.5987 57.5265
6 3.9811 2.22 2.4642 3.9168 61.8979
8 6.3096 3.00 4.5000 6.5321 71.3202
10 10.0000 3.84 7.3728 8.6763 73.7280
12 15.8489 4.56 10.3968 10.1690 65.5994
14 25.1189 5.60 15.6800 11.9535 62.4232
16 39.8107 6.63 21.9785 13.4200 55.2074
18 63.0957 7.56 28.5768 14.5601 45.2912
20 100.0000 7.56 28.5768 14.5601 28.5768

Table 4.8: Measured Voltages for the HSMS8202 rectification circuit

95
achieved during the laboratory tests. This can be seen in the low efficiencies stated
in Table 4.7. In previous literature [26] the same behavior was observed, even with
a better matching section. Thus, as a final decision for the rectenna, the HSMS8202
will be used as part of the rectification section. The complete rectenna system can
be seen in 4-20(c).

4.3 Radio Link calculation

The calculations for the basic radio link that will provide the wireless power to the
receiver are presented. The basic Friis equation explained in section 2.3 has been
used to calculate the expected radio link power levels at different distances between
the transmitter and receiver.
First, it is important to notice the constraints given by the IEEE [10]and FCC [33]
concerning the amount of power and power densities a transmitter in the ISM band
can handle. According to FCC regulations, the maximum transmitter power is 1W,
therefore in all foregoing calculations Pt = 1W. Also, the design must be made taking
into account that the power density has to be within regulations, that is 2mW/cmr2 . or
equivalently 20W/m 2 . For this purpose. equation 2.2 is used so the minimum safety
distance where the power density is within specification can be calculated. Thus,
solving for R

Rmn- 4rS [m]

According to the antenna design, the patch antenna has a gain of 6dB. This, in
linear quantities, is equal to 106/1' = 3.98. In this first approach, let's consider that
both the transmitting and receiving antenna are the same. Therefore

96
 GtPt
Rin= 4S 4 ir [m]

/(3.98)(1W)
(47r)-(20W/m 2 )
= 0.126m
= 12.6cm

As it can be seen, the minimum distance is 12.6cm. Shorter distances imply higher
densities.
The ends of the ISM band will give the shortest and longest wavelengths that can
be used in the system. For the low frequency end (long wavelength)

c
X=--
f
3 x 10l[m/s]
2.4 x 10911s]
= .125m
= 12.5cm

For the high frequency end (short wavelength)

f
3 x 108[m/s]
2.5 x 109[1/s]
= 0.12m

= 12cm

For simplicity, further calculations cxemade using the wavelength mean of 12.25cm
that corresponds to a frequency of 2.45GHz.
Plugging all the previous values into the Friis equation, the maximum power at

97
the receiver's end, obtained with the receiver closest to the transmitter, is

GtGrA 2
r (4rR) 2 t

(3.98)2(0. 1225m) 2 (1W)

(47r.0.126m)2
= 94.81 x 10-3W

-95mW

The minimum amount of power is received at the maximum distance, R = lm.

Therefore

GtGrA2
(47rR)2
(3.98) 2 (0.1225m) 2 (1W)
(4ir.lm)2
= 1.5 x 10-3 W
= 1.5mW

The power range is then from 95mW at short ranges and up to 1.5mW at the
maximum specified distance. In a decibels scale, the range goes from -20dBm to
-1.8dBm. This is shown in figure 4-24
If an antenna of bigger gain is used, say an array of two patches on the transmitter's
side, it is possible to receive more power within the desired distances. The array
equivalent gain will be around 12dB according to the theory associated with antenna
arrays [34]. The minimum safety distance will then be modified to:

98
 35i - · -· · · · ·
I I I

30

25

20

a
a 15
.z
a
10

I I I
-0 '2 0.4 0.6 0.8 1 . 1.2 1.4
Distance [m]

Figure 4-24: Radiolink performance for Pt. = 30OdBm,Gt. = 6dB

Rmi = 4rSvg [m]

/ (15.85)-(1W)
V(47r)(20W/m2)

= 0.25m
= 25cm

And the modified ranges are given by

GtGrA22 p
P= (4irR) Pt

(15.85) (3.98) (0 1225m)2 (1W

(47r-0.25m) 2

= 95.91 x 10-3 W

-96mW

99
 GtGrA 2
(47rR)2
(15.85).(3.98)(0.1225m) 2 1)
2
(47r.1m)

= 5.99 x 10-3W

-6mW

Therefore, for this kind of radio link we have a range from -20dBm to 8dBm,
6dB more than the previous link with a transmitting antenna of lesser gain. Figure
4-25 shows these results.
40 ! .....
I

35

30

25

a 20

- 2

15

10

At
0 0.2 0.4 0.6 0.8 1 1.2 1.4
Distance[m]

Figure 4-25: Radiolink performance for Pt. = 30dBm, Gtx = 12dB

4.3.1 Measured Link Data

The laboratory setup used to characterize the Radio Link is shown in 4-26. The
physical length of the test cables restricted the laboratory test to Rmax = 40cm.
The amount of transmitted power was also restricted by the laboratory equipment,
where the HP Function generator could only provide signal powers up to 20dBm. The

100
 Antennas, Gx=GRx [6dB]

Signal : L- Spectrum
Generator i Anayzer
L=6dB L=0.5dB
P=20dBm
4---
R

Figure 4-26: Block Diagram for the Power Link Test

obtained values for four fixed distances are summarized in table 4.9.
Distance [cm] Expected Power [dBm] Measured Power [dBm]
10 12 7
20 6 -2
30 3 -13
40 0 -7

Table 4.9: Measured and expected Powers for the experimental Radio Link

The Radiolink performance is poor compared to that expected from the theory
obtained in 4.3. This can be due to several reasons. Attenuation effects are discarded
because the laboratory environment is somewhat free of humidity and particles. One
explanation could be that the radiation properties of the antennas used so far are not
the expected ones, either these antennas have a lower gain than the expected 6dB
value or the radiation efficiencyfor the FR4 material is very low. Another important
consideration is the unexpected high attenuation of the transmitting cable, as shown
in 4-26. Also, the radiation patterns of the antennas are broad, and multipath effects
could exist that create destructive interference of the signals arriving at the antenna
on the receiver's side. It is important to notice that phase plays an important role,
this can be seen in the table 4.9 at L = 30cm, where we have read a minimum in
the power level. For future improvements, a more complete radio link model analysis
should be made, where phase delays and multipath are taken into consideration.

101
 .?
'I •
I 1
I

Figure 4-27: Actual Power Transmitter. From left to right: 2.45GHz tuned VCO.
RF2189 medium power amplifier and RF2126 high power amplifier

4.3.2 Power Base Station

Using off-the-shelf components. alpower station was assembled and tested. The actual
tranmilitter is shown in figllre 4-27. Tlhis high-glain, high-power base station generates
a purle 2.45GHz 3dlBIll tone that is being b)roadlcaste(l to the environment. As can
be seen in 4-27, the power base station consists of three major elements: a Voltage
Controlled Oscillator (VCO) from Nini Circuits® that generates the 2.45GHz tone.
The RF2189 is a nediumn gain. inedium power stage with a gain of 20dB and mxxinlinm
power handling of 251dBm. and the fina-lstage is an RF2126 medilum gain. high power
stage with 12dB snall signal gain and maximlm power handling of 31dBm. The
manufacturer of both amplifiers is RF 5Micro Devices®. Both amplifiers also have
power control pins that allow control of the gain in each section and also provide
a sht-down mode. A microcontroller can make ulse of those pins to automatically
control the power base station.

102
 Chapter 5

Scenario Discussion

5.1 Interactive Necklace

(a) Necklace v2.0 (b) Beads Detail

Figure 5-1: Interactive Necklace System v2.0

As a first application example for the proposed Wireless Transmission of Power

discussed through the present dissertation, the platform known as the Interactive
Necklace (figure 5-1) will be used. The first version of the necklace was designed by
Ted Selker [35]in the year 1999 and consists of a group of electronic beads that make
sound and turn lights on when a DC motor wheel is turned. This motor is present as
another bead in the necklace. The DC motor provides the energy that the rest of the
beads need to operate, so no batteries are present in the necklace. The main objective

103
is to implement a new series of beads that allow the necklace to have interactions with
elements that may be present in the environment, whether they be other necklaces,
computers, etc., and also that these external devices can somehow interact with the
necklace itself, even at a physical level.
To achieve this objective, as a first step, an RF transmitter bead was implemented.
Such transmitter is a modified version of the Tx99 v3.0 transmitter manufactured by
Ming Microsystems; the modification was made by Mark Feldmeier [36] in the MIT
Media Laboratory Responsive Environments Group. Such a transmitter works within
the 300MHz band. Also, a base station was implemented using the RE-99 v3.OA
300MHz receiver from Ming Microsystems and an IRx board [37]. It was decided
that, whenever there was activity in the necklace (i.e., the wheel was being turned)
the base station would detect this, which gives a bit of information (activity/no-
activity) that can be used for some applications. Images of both Transmitter and
receiver electronics can be seen in figure 5-2.

(a) 300MHz Transmitter Bead (b) 300MHz Receiver Station

Figure 5-2: Transmitter and Receiver pair used in the Interactive Necklace Scenario.

One of the intended new applications for the necklace is a basic Friend/Foe game
implemented with a computer. This has the purpose of enhancing the social inter-
actions of the necklace bearer, with the computer playing the role of another entity
with whom to have interactions. The main idea of this game is that, whenever the
necklace is activated in front of a friendly entity (Friend) there won't be any problems

104
 in operating the necklace, this means that the torque of the DC motor wheel would
be low; when the necklace is being activated in front of an unfriendly entity (Foe),
the wheel will have a higher torque and thus become harder to spin for the user,
indicating a certain degree of rejection.
From the electronic point of view, it can be explained like this: whenever there is a
friendly entity present in front of the necklace, the impedance between the terminals
of the DC motor will be high, thus allowing a normal operation of this DC voltage
source. In the presence of a foe, this impedance has to be very low such that there is
a short circuit between the motor DC terminals, which blocks the wheel and presents
a higher torque. This electrical behavior was done using the circuit shown in 5-3

RI

LEGC,
MOCTOCF:

Figure 5-3: Electronic Switch Schematic

As one can see, the circuit is a very basic transistor-based electronic switch. The
explanation of its operation follows. As long as there is no current present in the
base of the transistor Q1, the Collector and Emitter will be operating in the Cutoff
region, hence no current can circulate through them and the DC motor will see a high
impedance that will not interfere with its normal operation. This only happens in a
given spinning direction, due to the fact that the collector-emitter junction can be
treated as a diode in the sense that current can just flow in one direction, collector-
emitter, and not the other way around. This implies that the DC motor will now
have a sort of polarity.
Now, when there's a current flowing through the transistor's base, the collector-

105
emitter junction enters into the Saturation region. This will make the motor terminals
to be close to a zero volt potential, and now all the current from the DC motor will
tend to flow through this new low impedance path which will automatically turn off
all the associated electronics present in the necklace. There are also repercussions on
the mechanical level, because the low impedance path will force the wheel to present
a much higher torque than in the free operation state, thus for the user it will imply
a sensible physical blockage (braking) of the wheel.

:"!Q-:
·

r:1·

Figure 5-4: Rectenna Implemented in the Interactive Necklace.

The energy source that will power the transistor's base is the rectenna system that
has been proposed and characterized in the present thesis. Due to the fact that the
base-emitter junction turns on at 0.7V and the nature of such junction is basically a
diode, this voltage will be constant. That implies that most of the power delivered
from the rectenna to the transistor will be in the form of electrical current, which
makes the rectenna a Power Controlled Current Source.
The free spinning of the motor (i.e. no load connected) generates peak voltages of
about 18V, with associated peak currents of about 8mA which means that the motor
can deliver to a load as much as 144mW of power at maximum spinning. To provide
a quantitative measurement of the switch operation, when the base-emitter junction
voltage reaches 0.75V and the base current is about 700uA, the motor's voltage
performance is reduced about 60%, with output voltages of 2.5V and currents about

106
60mA. With base currents of about 3mA the voltage performance is further reduced
to 90%, that is 1.8V at 85mA. In both cases, the mechanical equivalence is an increase
of the motor's torque. From a qualitative point of view, the wheel is harder to turn
when there's more current flowing into the transistor, compared to the open switch
state (no base current flowing). It was observed that the maximum distance for power
transmission where there still was a significant torque were 50cm.
The complete scenario is the following. The 300MHz base station connects to a
host computer using the serial port and also to a 2.4GHz base station that provides
the Wireless Power. When the necklace has the 300MHztransmitter attached to it, it
generates a beacon that is detected by the base station, and this beacon is associated
with the Foe state. Then the computer displays an alarm and a rejection message;
also it activates the 2.4GHz wireless power source. The rectenna attached to the
necklace then harvests the power from the wireless power source and thus blocks the
necklace. After a programmable period of time (between 6 to 10 seconds) the base
station deactivates the wireless power source and the computer presents to the user
information available only to "New" Friends.

5.2 Wireless Sensors

The next step is to create a group of wireless and battery-less sensors that would be
installed in different environments. This way, the sensors relay the information of the
sensed variables, output can range from visual feedback using LEDs, audio feeback
using buzzers, and other types of alarms. The sensors can even wirelessly relay the
information to somewhere else.
The first example is a stapler sensor. The circuit is a low power CMOS 555 timer,
programmed to generate a square signal of a fixed 10Hz frequency and 10% duty
cycle. This signal is used to turn on or off an LED. The mechanism inside the stapler
that pushes the staples forward is the mechanical switch of this system. A pair of
connectors was placed inside the stapler, one on the front inner side of it and the other
on the sliding piece of metal that pushes the staples. When the stapler runs out of

107
staples, both connectors make contact and the LED starts flashing, giving a visual
feedback to the user that indicates a refill state. This flashing circuit drains power that
can be as low as 500piW, with 1V@500/zA for assured operation. The power peaks
present whenever the LED flashes, where the circuit drains 1.5mW (1.5V@lmA).
These low power levels can be easily handled by the designed rectenna. In a laboratory
setup, the maximum observed distance where the sensor's LED stopped blinking were
near 60cm when using the rectenna system as a power source. Figure 5-5 shows the
complete Circuit Schematic.

Figure 5-5: CMOS Timer implemented in the Staples Sensor

The same principle can be applied to a piezo-buzzer. The circuit is shown in figure
5-6. The power demand of such a circuit is extremely low, with an weak but audible
tone starting at just about 2251pW (1V(225 uA),
1 reaching a high audible level at a
mere 1.4mW (2V©700/pW). The maximum distance of operation with the rectenna
as a feeding element was the longest one achieved from all the sensors, 80cm.
Another more elaborate sensor is a humidity sensor. This sensor is placed in
a flowerpot, and it senses the humidity level of the soil. The sensor is basically
a variable capacitor made with two copper electrodes separated by a hygroscopic
material (special paper used to clean lenses). The complete Schematic is shown in
5-7(a). When the sensing capacitor is dry, the capacitance value is about 4-6pF;
when the piece of paper that separates the electrodes is wet the value jumps to
about 20-30pF. This capacitor is part of a tank circuit in the 300MHz transmitter

108
 ?EF TENNit

R3
f-'0

Figure 5-6: Piezo-buzzer circuit implemented

discussed previously. Thus when the sensor is dry, the tank circuit will be tuned to
the desired frequency of 300MHz and the signal generated by the oscillator can be
detected by a wireless station tuned to the same frequency. When the sensor is wet
the frequency drops around 200MHzdue to the higher capacitance value, and given
the fact that the base station only can be tuned to detect signals around 290-310MHz,
the generated beacon will not be detected. The base station can be programmed such
that, whenever it detects the 300MHz beacon, a signal is sent to the host computer,
using the IRx [37]232 serial port or maybe other means, so it can alert the user about
the state of the plant (i.e. sending the user a courtesy notice email).
Laboratory tests have shown that the 300MHz transmitter can operate as low
as with 1.2V@1.5mA, which means a power demand of 1.8mW. The associated
square wave generator (the same low-power CMOS timer used for the stapler) only
drains 500pW; taking both circuits into account, the transmitter needs only 2.3mW
to operate. The rectenna system provided enough energy to make the circuit to work
at a distance ranging up to 45 cm. At longer distances, the signal's level was so weak
that the base station could not detect the beacon. This distance could be improved
if the base station is designed to have a higher sensitivity.

109
 pF

(a) Schematic Circuit

BI 14tz'·rtt ;.
:
i
-· '': i_··f;

I--
.·r,,9
i: i

,.·· ,··.- .·.

r"' i·:· :· ·
----- .·.·. ··- i,
:.i·.··p;
· .. ·.. ,. ,
m " "::I
i .·x ::

(b) Actual Sensor

Figure 5-7: Humidity Sensor.

110
 Chapter 6

Conclusions

A Wireless Transmission of Power for Sensors (WTPS) system has been analyzed,
discussed, designed, and tested in the present dissertation, to explore the idea of
getting rid of cables and batteries.
First, a low profile antenna with compact size was designed, simulated and imple-
mented. It was observed that the theoretical design procedure was straightforward;
the simulations obtained from this design shown very acceptable performances in
the frequency band of interest. In the implementation level, a manual tuning of the
antenna had to be done to obtain the desired radiation properties at the frequency
objective of 2.45GHz. A very simple procedure for tuning followed,where, with the
addition of a small tuning flag, a direct control of the antenna's resonant frequency
was obtained. The fine segmentation of this short tuning flag allowed precise fre-
quency tuning, with an accuracy of a few megahertz.
Without the proper equipment and facilities to measure the radiation properties of
the antenna, the design and testing relied entirely on other properties of the antenna
such as Voltage Standing Wave Ratio (VSWR) and Input Impedance. The resulting
measurements of these properties were more than satisfactory. In a qualitative way,
the radiation properties were obtained for the microstrip antenna by means of a
small laboratory radio link setup. It was observed that the FR4 material selected
as the dielectric of the antenna had a poor radiation performance, presenting high
radiation losses. Future improvements of the antenna can take this into account, and

111
other materials can be used, such as duroid or GML1000, to improve the radiation
properties and thus obtaining better performance. It is important to note that if the
substrate material is changed, the final design presented in this dissertation cannot
be used and changes in the antenna's dimensions have to be made, following the
procedure shown in section 3.1.
Little or no attention was paid to the associated ground plane of the antenna.
Increasing its size will definitely improve the front-to-back ratio of the antenna, which
can increase the endfire properties and provide higher gains in the direction of interest.
The theoretical radiation pattern presents a fan-beam shape, which is desired for
the type of applications sought here, which have a set of sensors extended all over a
wide area, with each sensor will be harvesting the incoming energy from a power base
station. Nonetheless, a more directive antenna with a narrower beam could be used
to concentrate more energy within a small area, thus having a high power density that
the designed rectennas can convert to usable DC power. The use of such a directional
antenna can add directionality to the WTPS system, hence certain applications that
can take advantage of this property should have to be analyzed in any future work.
A 2.4GHz rectifier was also analyzed, designed, and implemented. Two different
types of diode pairs were used. It was observed that, from the theoretical point of
view, both diodes presented very similar rectification properties. In the coupling
section, the obtained stub length values for the HSMS2852were low, so their manual
implementation in the lab presented certain complications that didn't allow a proper
match to be achieved. On the other hand, the HSMS8202 diode pair had better
performance in that sense. The stub lengths were acceptable, and in the lab bench,
the match was achieved with very good results. The impedance match of both diodes
was key when measuring their rectification properties. The HSMS8202got the higher
efficiencies, some of them above 70% as stated in 4.8. Note, however, that the diode
entered into saturation with moderately high inputs (2OdBm), which places a top
boundary on the whole range of operation of the rectenna system.
A set of low-power circuits was also designed to serve as a test platform of the
whole WTPS system. Very basic sensors where designed and implemented in common

112
environments, such as staplers and flowerpots. The information obtained from the
sensors provided just one bit, but given the intended applications for these sensors,
that proved to be enough to give the user meaningful information. More complex
sensors can extent the capabilities of the WTPS system. A design of more evolved
sensors can follow in the future, allowing more information to be extracted from the
environment and thus gaining more complex knowledge of the sensors' surroundings.
Because of the nature of the microwave power links, these sensors have to operate
under low power conditions, hence a limit of P < 5mW is suggested to take good
advantage of the system presented in this dissertation.
Finally, the complete WTPS system was used to energize this set of sensors in
a remote fashion, which was the primary objective of the system. The maximum
distances obtained in lab tests (lm)certainly were less than those expected in the-
ory. This could be explained given the very basic link model that was used, where
multipath effects and fading due to elements in the line-of-sight were not taken into
account. A more complex model that considers all these effects can be realized and
used to predict the real performance of a system like the WTPS.

6.1 Future Work

Every day more information is going into computational systems. Accordingly, infor-
mation coming from sensors deployed in the external world is in high demand, and
such sensor systems can help the computer to make important decisions on what the
user needs are. By following somewhat complex rules and user-task models, com-
puters can become more aware of their context and also the user's context as well.
To obtain reliable user task models, the sensors have to be simple yet provide useful
information to the system.
If one thinks about the possible future of a wireless power system like the one
presented in this dissertation, the possibilities appear attractive when focused on
electronic sensors systems. First, one could think of a world where low power bat-
teries, like the coin-cell type ones, have evolved in a way such that there'is no need

113
to replace them. They could become permanent recharging batteries thanks to an
existing wireless power source in the surroundings. This has important repercussions
in different fields; from an economic point of view there would be no need to buy new
batteries, and from the environmental point of view less disposed batteries mean less
pollution and less poisonous materials exposed to Nature.
Also, getting rid of cables is an attractive option for the automotive industry.
Current research and discussionswith representatives in the field have expressed their
interest in using wireless power to feed the electronic door switches of an automobile.
In an enclosed space, such as an automobile's door, the electromagnetic fields could
certainly be more easily guided to the area of interest, thus allowinglarger concentra-
tions of energy within the rectenna's area and, as a consequence, better performance
can be achieved for not only activating door switches, but for possibly allowing the
existence of sensors inside the door that can provide useful information to the user.
As a few examples of this concept are door temperature sensors that could help reg-
ulate the driver's environment and water sensors that could trigger the side windows
when it is raining.
It would also be interesting to explore the idea of an array of power transmitters,
such that within a specific area, a uniform power distribution can be obtained. This
idea can be applied in household environments or even in office environments. For
example, to have entire walls covered with power transmitters (or even floor mats)
could provide a good platform to power and implement sensors in diverse low-tech
devices like chairs and tables [38].
The WTPS system could extend the capabilities of modern RFID systems, allow-
ing more active electronics to be embedded in different appliances that could extent
their native capabilities and thus provide a test-bed for future more complex Context
Aware Spaces. Right now, RFID systems do not have the capabilities of enabling
active electronics to perform common tasks because of the low power levels involved.
By creating sounds with off-the-shelf components, by having visual feedback with
lights and, even more, by having active sensorial systems relaying the information
they generate one can envision a first step towards a new extended RFID system.

114
 It must be noted that various obstacles exist to this scenario, including a public.
that (justified or not) is very wary of the low power levels coming from cell phones,
power lines, and microwave oven leakage.

115
Bibliography

[1] H Hertz. Dictionary of scientific biography. New York: Scribner, 340-349.

[2] J. J. O'Neill. Prodigal genius - the life of nikola tesla, 1944. New York: Wash-
burn.

[3] M. Cheney. Tesla, Man Out of Time. Prentice Hall, Englewood Cliffs, NJ, 1981.

[4] Robert Buderi. The Invention That Changed the World: How a Small Group
of Radar Pioneers Won the Second World War and Launched a Technical Rev-
olution. Touchstone Books, March 1998.

[5] K.V.S. Rao. An overview of back scattered radio frequency identification systems
(RFID). IEEE Transactions, pages 746-749, 1999.

[6] W.C. Brown. The History of Power Transmission by Radio Waves, volume
MTT-32 no 9. IEEE Press, Sept 1984.

[7] G. Goubau and F. Schwering. On the guided propagation of electromagnetic

wave beams. IRE Trans. Antennas Propagat., Vol. AP-9:248-256, May 1961.

[8] H . Kogelnik and T. Li. Laser beams and resonators. Proc. IEEE, Vol. 54 no.
10, Oct 1966.

[9] Raytheon. Electronic and mechanical improvement of the receiving terminal of

a free-space microwave power transmission system. Raytheon Contractor Rep.
PT 4964, CR-135194, Aug. 1977.

116
[10] IEEE Standards Board. IEEE standard for safety levels with respect to human
exposure to radio frequency electromagnetic fields, 3kHz to 300GHz. Technical
Report IEEE C95.1-1992, IEEE Standards Coordinating Committee 28, 1992.

[11] Eleanor R. Adair. Biological effects of radio-Frequency/Microwave radiation.

IEEE Transactions On Microwave Theory And Techniques, Vol. 50, No. 3:953-
962, March 2002.

[12] James J. and P.S. Hall. Handbook of Microstrip Antennas. Peter Peregrinus,
1989.

[13] D.M. Pozar. Considerations for millimeter wave printed antennas. IEEE Trans.
on Antennas and Propagation, AP-31:740-747, 1983.

[14] Y.T. Lo Richards W.F. and D.D. Harrison. An improved theory for microstrip
antennas and applications. IEEE Transactions on Antennas and Propagation,
AP-29:38-46, 1981.

[15] D. Solomon Y. T. Lo and W.F. Richards. Theory and experiment on microstrip

antennas. IEEE Trans. on Antennas and Propagation, AP-27:137-145, 1979.

[16] Jin Au Kong. Electromagnetic Wave Theory. EMW Publishing, 2000.

[17] E.O. Hammerstad. Equations for microstrip circuit design. 5th European Mi-
crowave Conference, pages 268-272, 1975.

[18] M. Kara. The resonant frequency of rectangular microstrip antenna elements

with various substrate thicknesses. Microwave and Opt. Technol. Lett., vol 11:55-
59, 1996.

[19] D.R. Jackson and N.G. Alexopoulos. Simple approximate formulas for input
resistance, bandwidth and efficiencyof a resonant regular patch. IEEE Trans.
on Antennas and Propagation, Vol. AP-39:407-410, 1991.

117
 [20] Perlmutter P. Shritkman S. and D. Treves. Electric surface current model for the
analysis of microstrip antennas with application to rectangular elements. IEEE
Trans. on Antennas and Propagation,AP-33:301-311, 1985.

[21] M. Kara. Calculation of the radiation patterns of rectangular microstrip antenna

elements with various substrate thicknesses. Microwave and Optical Technol.
Letters, Vol. 13:221-226, 1996.

[22] D. Thoroude. CAD-oriented cavity model for rectangular patches. Electron.

Lett., Vol. 13:842-844, 1990.

[23] D.M. Pozar. Rigorous closed-formexpressions for the surface wave loss of printed
antennas. Eletronic Letters., Vol. 26:954-956, 1990.

[24] Ramesh Garg. Microstrip Antenna Design Handbook. Artech House, 2001.

[25] M. Kara. Formulas for the computation of the physical properties of rectangular
microstrip antenna elements with various substrate thicknesses. Microwave and
Opt. Technol. Lett., Vol 12:234-239, 1996.

[26] Agilent Technologies. Designing the virtual battery. Technical Report Applica-
tion Note 1088, Aggilent Technologies, 1999.

[27] Ansoft Ensemble SV. Maxwell 2d ver. 2.0.57. http://www.ansoft.com.

[28] Agilent Technologies. Designing detectors for RF/ID tags. Technical Report
Application Note 1089, Agilent Technologies, 1999.

[29] Agilent Technologies. Impedance matching techniques for mixers and detectors.
Technical Report Application Note 963, Agilent Technologies, 1999.

[30] Agilent Technologies. The zero bias schottky detector diode. Technical Report
Application Note 969, Agilent Technologies, 1999.

[31] R. Boylestad and L. Nashelsky. Electronic Devices and Circuit Theory. Prentice-
Hall, 1987.

118
[32] W.C. Brown. An experimental low power density rectenna. IEEE MTT-S Digest,
April:197-200, 1991.

[33] Federal Communications Commission Office of Engineering & Technology. Ques-

tions and answers about biologicaleffects and potential hazards of radiofrequency
electromagnetic fields, OET bulletin 56 Fourth Edition August 1999.

[34] Constantine Balanis. Antenna Theory: Analysis and Design. Bk & Dk 2nd

edition, May 1996.

[35] Ted Selker and Win Burleson. Context-aware design and interaction in computer
systems. IBM Systems Journal, Vol. 39, no. 3 & 4, 2000.

[36] Mark Feldmeier and Joe Paradiso. Ultra-low-cost wireless motion sensor for mu-
sical interaction with very large groups. UBICOMP 2001 Workshop on Designing
Ubiquitous Computing Games, ACM UBICOMP Conference Proceedings, 2001.

[37] Rob Poor. irx2 prototyping board programming seminar.

http://web.media.mit.edu/ r/projects/picsem/.

[38] E. Arroyo W. Burleson, T. Selker. Chameleon tables: Using context informa-

tion in everyday objects. Conference on Human Factors in Computing Systems,
Extended Abstracts:580-581, 2002.

[39] Agilent Technologies. Surface mount microwave schottky mixer diodes. Technical
Report Technical Data HSMS-820x Series, Agilent Technologies, 2001.

[40] Matthew M. Radmanesh. Radio Frequency and Microwave Electronics. Prentice

Hall, 2001.

[41] Agilent Technologies. Surface mount RF schottky diodes. Technical Report

Technical Data HSMS-282x Series, Agilent Technologies, 2000.

[42] Association of the Automatic Identification and Data Capture Industry. Draft
paper on the characteristics of RFID systems. Technical Report AIM FF
2000:001, AIM Frequency Forums, 2000.

119
[43] Agilent Technologies. Square law and linear detection. Technical Report Appli-
cation Note 986, Agilent Technologies, 1999.

[44] Agilent Technologies. Surface mount zero bias schottky detector diodes. Tech-
nical Report Technical Data HSMS-285x Series, Agilent Technologies, 1999.

[45] Dong-Gi Youn. A study of the fundamental transmission experiment for wireless
power transmission system. IEEE TENCON, pages 1419-1422, 1999.

[46] A. Papiernik S. Drabowitch. Modern Antennas. Chapman & Hall, 1998.

[47] J. Kymissis C. Kendall J. Paradiso and N. Gershenfeld. Parasitic power harvest-

ing in shoes. Proc. of the Second IEEE International Conference on Wearable
Computing, (ISWC), IEEE Computer Society Press:132-139, 1998.

[48] David Pozar. Microwave Engineering. John Wiley & Sons, 1997.

[49] David Pozar. Microstrip Antennas: The Analysis and Design of Microstrip
Antennas and Arrays. IEEE, July 1995.

[50] W.C. Brown. Beamed microwave power transmission and its applications to
space. IEEE Transactions on Microwave Theory and Techniques, Vol 40 no
6:1239-1250, June 1992.

[51] David K Cheng. Field and Wave Electromagnetics. Addison-Wesley, 1989.

[52] Paul Horowitz. The Art of Electronics. Cambridge University Press 2 nd edition,
1989.

[53] John Kraus. Antennas. McGraw Hill 2nd edition, 1988.

[54] W.C. Brown. Experiments involving a microwave beam to power and position a
helicopter. IEEE transactions Aerosp. Electron. Sys., AES-5:692-702, 1969.

120
 THESIS PROCESSING SLIP

FIXED FIELD: ill. name

index biblio

COPIES: Archives ,? Aero Dewey Barker Hum

Lindgren Music :Rotch Science Sche-Plough

TITLE VARIES: -'nt..,.

NAMEVARIES:-_ _

IMPRINT: (COPYRIGHT)

I COLLATION:

b- nn nrDP=-
v' r
TV. w " . &nDT -
' r '

· Ann-:
" ~ _'DFGRFF:
" . * Di:PT
"__,,__.

Q1
our
PFFov1QnPQ-
_ ww _ be

NOTES:

YEAR
, DEPT:

*NAME: I
!

"_2_i
i

·
J.A.
-',,

!k4;
D. A
cat'r: date:
page:

-,' ?DEGREE:

.r .. i