|USFDC Home | USF Electronic Theses and Dissertations||| RSS|
This item is only available as the following downloads:
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001910790
007 cr mnu|||uuuuu
008 071001s2006 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001713
IEEE 802.11b wireless LAN sensor system and antenna design
h [electronic resource] /
by Leonard Guerra.
[Tampa, Fla] :
b University of South Florida,
Thesis (M.A.)--University of South Florida, 2006.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 62 pages.
ABSTRACT: A novel approach to miniaturizing an 802.11b WLAN card using folded-flex ultra-thin substrates is presented. A 73 percent reduction in size was realized using hybrid circuits on FR4 and polyimide. There is even more potential for further reduction if more copper layers are used. The miniaturized 802.11b WLAN cards were used to design 802.11b wireless sensor nodes. A research test-bed was setup to study how 802.11b networked sensor nodes could operate in the field. There are many applications for such sensor networks like habitat monitoring, object tracking, seismic detection, military surveillance, or fire detection to name a few. This investigation focuses on the requirements, design, and performance of a miniaturized 802.11b wireless LAN sensor node that is reliable, can be deployed in large-scale, and has the endurance long-lived for surveillance applications. An aperture coupled microstrip antenna is investigated for 2.44 GHz wireless local area networks (WLAN) which has the advantages of being low-profile and compact. The most important parameters for antenna optimization have been determined through extensive simulation using Ansoft's HFSS and experimental verification. As a result, an omnidirectional antenna with a size of 36.2 mm x 32 mm x 4.75 mm has been realized using Rogers 5880 duroid (permittivity = 2.20 ; loss tangent = 0.0004) with 4.4 dBi of gain, > 80 MHz, and a return loss > -32 dB. These types of performance characteristics make the antenna highly desirable for both 802.11b and Bluetooth applications.
Adviser: Thomas Weller, Ph.D.
x Electrical Engineering
t USF Electronic Theses and Dissertations.
IEEE 802.11b Wireless LAN Sensor System and Antenna Design by Leonard Guerra A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Electrical Engineering College of Engineering University of South Florida Major Professor: T homas Weller, Ph.D. Lawrence Dunleavy, Ph.D. Huseyin Arslan, Ph.D. Date of Approval: July 12, 2006 Keywords: patch, slot-coupled, arra y, beam steering, folded flex Copyright 2006, Leonard Guerra
ii TABLE OF CONTENTS LIST OF TABLES iv LIST OF FIGURES v ABSTRACT vii CHAPTER 1 INTRODUCTION 1 1.1 Overview 1 1.2 Preliminary Investigation 2 1.2.1 Wireless Sensor System Design 2 1.2.2 Slot-Coupled Microstrip Antenna Design 3 1.3 Research Accomplishments 4 1.4 Thesis Organization 5 CHAPTER 2 WIRELESS LAN SENSOR SYSTEM DESIGN 6 2.1 Introduction 6 2.2 System Requirements 7 2.3 System Design 9 2.3.1 System Overview 9 2.3.2 Power Management 11 2.3.3 Wireless LAN (802.11b) Radio 15 2.3.4 PICmicro Processor and Sensor Board 20 2.4 Packaging 22 2.4.1 Overview 22 2.4.2 Folded Flex Design Using Polyimide 23 2.5 System Performance 25 2.6 Chapter Summary 28 CHAPTER 3 SLOT-COUPLED MICROSTRIP ANTENNA DESIGN 29 3.1 Introduction 29 3.2 Design Requirements 29 3.3 Design Process 30 3.4 Antenna Simulation 34 3.4.1 Model Construction 34 3.4.2 Simulation Results 36 3.5 Design Modifications 42 3.6 Experimental Validation 49 3.7 Chapter Summary 57
iii CHAPTER 4 CONCLUSIONS AND RECOMMENDATIONS 58 REFERENCES 60 APPENDICES 61 Appendix A. Electrical Schematics 62
iv LIST OF TABLES Table 2.1 Current Consumption for PRISM 3 and PIC18F452 13 Table 2.2 Current Consumption at Startup for PRISM 3 and PIC18F452 13 Table 2.3 Battery Life Expectancy of WSN Under Various Scenarios 14 Table 2.4 Wireless Sensor Node Specifications 27 Table 3.1 Parameters of the Simulated Aper ture Coupled Microstrip Antenna 42 Table 3.2 Comparison Table of Simulation Results 49
v LIST OF FIGURES Figure 1.1 MICA2 Mote by Crossbow 2 Figure 1.2 Basic Components of the Aperture Coupled Antenna 4 Figure 2.1 Communication Block Diagram for Sensor Node 6 Figure 2.2 Basic Network Star Topology for 802.11b Systems 8 Figure 2.3 Graphic of Packaging Solution for Wi reless Sensor Node 9 Figure 2.4 Three Sections of the Wireless Sensor Node 9 Figure 2.5 System Diagram for the WSN 10 Figure 2.6 Going From Defin ition to Design Phase 11 Figure 2.7 PRISM 3 Power Consumption Versus RF Output Power 12 Figure 2.8 PRISM 3 Radio Block Diagram 16 Figure 2.9 Layer Stack Illustra ting the Thicknesses of Each Layer 17 Figure 2.10 Layer Stack of Ra dio Side of the 802.11b Board 18 Figure 2.11 Cross-section of a CPW Transmission Line 18 Figure 2.12 Top Layer of RF Section Highlighting CPW Transmission Lines 19 Figure 2.13 Schematic of PIC18L F6620 to PCMCIA Interface 21 Figure 2.14 Sensor Circuit Showi ng Connections to PICmicro 22 Figure 2.15 Cross-section of Folded Radio Design 23 Figure 2.16 Picture of Actual Radio and Sensor Boards 23 Figure 2.17 Damaged Fingers on Radio Board Tail 24 Figure 2.18 Stiffener on Bottom Side of Flex-Tail 25 Figure 2.19 Close-up of Damaged Fingers on Flex-Tail 25 Figure 2.20 Frequency Spectrum for WSN Radio 26 Figure 3.1 Basic Components of an Aperture Coupled Patch Antenna 30 Figure 3.2 Smith Chart Plot of the Impeda nce Locus as a Function of Frequency 33 Figure 3.3 3-D Representation of the Aperture Coupled Antenna in HFSS 34 Figure 3.4 Defined Excitation Port in HFSS Simulation 35 Figure 3.5 Return Loss Results for the Aperture Coupled Antenna 36
vi Figure 3.6 Smith Chart Plot of the Impedance Locus Versus Frequency 37 Figure 3.7 Smith Chart Plot of the Impe dance with Feed Line Length of 40.5 mm 38 Figure 3.8 Smith Chart Plot of the Impe dance Locus with 11 mm Slot Width 38 Figure 3.9 Phase Response of S11 Results 39 Figure 3.10 Phase Response of S11 Resu lts Using 43.5 mm Feed Line Length 39 Figure 3.11 Phase Response of S11 Re sults Using 11 mm Slot Length 40 Figure 3.12 VSWR Plot of the Aperture Coupled Antenna 40 Figure 3.13 E -plane (blue) and H -plane (red) Radiation Patterns 41 Figure 3.14 3-D Radiation Pattern as a Function of the Gain 42 Figure 3.15 Return Loss Results Usin g Modified Antenna Dimensions 43 Figure 3.16 Smith Chart Plot of Impedance Locus for Modified Antenna 44 Figure 3.17 VSWR Plot for Modified Antenna 44 Figure 3.18 Return Loss Results 3.175 mm S ubstrate Thickness 45 Figure 3.19 Smith Chart of Impedance Lo cus Using 3.175 mm Substrate Thickness 46 Figure 3.20 Phase Response of S11 Results Using 3.175 mm Subs trate Thickness 46 Figure 3.21 VSWR Plot of Usi ng 3.175 mm Substrate Thickness 47 Figure 3.22 E -plane (Blue) and H -plane (Red) Radiation Patterns 48 Figure 3.23 3-D Radiation Pattern of An tenna Using 3.175 mm Substrate Thickness 48 Figure 3.24 Antenna Patch on 3.175 mm Rogers Duroid 5880 Material 50 Figure 3.25 Coupling Aperture and Ground Plane on 1.575 mm FR4 50 Figure 3.26 Feed Line on Bottom Layer of 1.575 mm FR4 51 Figure 3.27 Entire Antenna Struct ure Including all Layers 51 Figure 3.28 Return Loss Comparison of Measured vs. Simulated Data 52 Figure 3.29 Smith Chart of Impedance Locus for Antenna 53 Figure 3.30 Phase Response of S11 Results for Antenna 54 Figure 3.31 VSWR Plot for Antenna 55 Figure 3.32 Simulation of S11 with 0.4 mm Air Gap and 1.3mm Slot Width 56 Figure 3.33 Smith Chart Plot with 0. 4mm Air Gap and 1.3mm Slot Width 56 Figure 4.1 Layout for First Prototype Board 59 Figure A.1 Sensor Board Electrical Schematic 62
vii IEEE 802.11b WIRELSS LAN SENSOR SYSTEM AND ANTENNA DESIGN Leonard Guerra ABSTRACT A novel approach to miniaturizing an 802.11b WLAN card using folded-flex ultra-thin substrates is pr esented. A 73 percent reduction in size was realized using hybrid circuits on FR4 and polyimide. There is even more potential for further reduction if more copper layers are used. The mi niaturized 802.11b WLAN cards were used to design 802.11b wireless sensor nodes. A re search test-bed was setup to study how 802.11b networked sensor nodes could operate in the field. There are many applications for such sensor networks like habitat m onitoring, object tracking, seismic detection, military surveillance, or fire detection to name a few. This investigation focuses on the requirements, design, and performance of a miniaturized 802.11b wireless LAN sensor node that is reliable, can be deployed in large-scale, and ha s the endurance long-lived for surveillance applications. An aperture coupled microstrip ante nna is investigated for 2.44 GHz wireless local area networks (WLAN) which has th e advantages of being low-profile and compact. The most important parameters for antenna optimization have been determined through extensive simulation using Ansofts HFSS and experimental verification. As a result, an omnidirectional antenna with a size of 36.2 mm x 32 mm x 4.75 mm has been realized using Rogers 5880 duroid (permittivity = 2.20 ; loss tangent = 0.0004) with 4.4 dBi of gain, > 80 MHz, and a return loss > -32 dB. These types of performance characteristics make the antenna high ly desirable for both 802.11b and Bluetooth applications.
1 CHAPTER 1 INTRODUCTION 1.1 Overview Wireless sensor networks (WSNs) are becoming more and more mainstream today and hold great promise for the future. WSNs are composed of a large number of nodes that sense and store information as data. They have some limited computing ability to process this inform ation and also have the capab ility of radio communication. They can operate in different kinds of environments doing tasks like environmental monitoring or surveillance. There are lite rally thousands of different ways sensor networks might be used. It is a technology th at has been around for a long time but is just now starting to create a new paradigm as nodes shrink in size and consume less power. In this work, a WSN is designed based on the 802.11b wireless protocol with an emphasis in miniature system packaging. An aperture-coupled antenna design for such a network is also investigated. The intended use for the WSN is for envi ronmental monitoring, marine ecosystem monitoring, and for harbor secu rity and defense. The ubi quitous 802.11b standard as the network protocol was chosen to easily crea te a link between the sensor data and the existing 802.11b network infrastructure at the un iversity. This made sensor information easily accessible via the internet for quick observation and analysis. Another great advantage of the design is the flexibility of being able to add practically any kind of sensor to a node. These include sensors that can detect chemical and bio-chemical traces in water, electro-chemical sensors that can de tect TNT in water, and even a CTD analyzer that senses pressure, conductivity, and temperature. Lots of work has been completed on WSNs by Berkeley, MIT, University of Virginia, and Ohio State University due to the strong support by the Defense Advanced Research Projects Agency (DARPA). Each of these universities has devoted extensive
2 research into different aspects of sensor network platforms. Some have focused on launching large-scale networks consisting of thousands of motes. Others have concentrated on creating simple yet powerful network protocols custom tailored for low data and low power use like Crossbows MI CA2 Mote shown in Figure 1.1. The MICA2 uses 2 AA batteries to provide power to the CPU and radio for up to 6 months. Work has also been concentrated in mi niaturization. Projects like Smartdust continue on in the endeavor for pin-size nodes all in tegrated on a single chip. Figure 1.1 MICA2 Mote by Crossbow 1.2 Preliminary Investigation 1.2.1 Wireless Sensor System Design The idea for a project invol ving the original design of a wireless sensor network (WSN) came out of the need for marine eco system monitoring, harbor security and defense, and environmental monitoring. Wirele ss networked sensors that can operate for extended periods of time and that could conne ct to existing 802.11b ne tworks on land are the perfect solution for these needs. However, for these networks to be financially and operationally viable there is the need for small, cheap sensors and networked devices which continues to fuel the research in th is field. Basing a sens or network on the 802.11b
3 protocol provides a solid st andard that can support the da ta communication of a large number of nodes while adding reliability to the wireless network platform, relative security, and ease of use. 1.2.2 Slot-Coupled Microstrip Antenna Design A slot coupled microstrip antenna makes an excellent radiating element for small wireless sensor nodes because they are low in cost, lightweight, compact, and can be fed in a variety of ways . They can also be easily integrated onto planar and non-planar surfaces. However, they do have some drawbacks. Microstrip antennas have many positive qualities like cost-effectiv e and ease of fabrication. However, they are plagued with limitations in their bandwidth characteri stics. The slot-coupled microstrip antenna offers an improvement in bandwidth at a cost of radiation efficien cy which results in a lower antenna gain. Its radiation pattern is normally more hemispherical than omnidirectional. The antennas ha ve significant back radiat ion which can be extremely undesirable in some applications. Side lobes also give out undesira ble radiation. This becomes an issue when the antennas length is greater than one wavelength. As the length of the antenna is increased it creates mo re side lobes. If the antenna is used to transmit a signal, the back and side lo bes result in more power loss . The slot-coupled antenna was first proposed by Pozar  in 1985. The antenna features a type of electro-magnetically coupled (EMC) feed. The RF energy from the feed line is coupled to the radiating element through a co mmon aperture usually in the form of a rectangular slot . These ba sic components that make up the antenna are shown in Figure 1.2.
4 Figure 1.2 Basic Components of the Aperture Coupled Antenna 1.3 Research Accomplishments In this project, wireless sensor nodes were designed, fabricated, tested, and integrated into a network for environmental monitoring. Accomplishing this task led to a much greater insight into seve ral different aspects of engine ering. These aspects included system integration, digital and analog design, RF design, antenna design, embedded design, and software design. Prototype boards were created for proof-of-concept purposes and testing. Testing included comp aring simulated and real-world measured data that was carefully analy zed and studied. The results of which correlated well with simulated data. Using these results, de sign enhancements were made and the miniaturization of the packaging commenced. These miniaturized designs were designed and manufactured using folded-flex technol ogy which uses ultra-thin substrates like polyimide or liquid crystal polymer (LCP). In addition to this, an aperture coupl ed microstrip antenna was designed, fabricated, and tested as well. The de sign was proven to meet all initial design requirements to function within the IEEE 802.11 network protocol. As digitized wireless communications like 802.11b, Bluetooth, Zigbee, and others continue to grow, so will the
5 need for antennas that have wider bandwid th with higher gain and omni-directional radiation patterns. This work presents the apertu re coupled microstrip antenna to have the necessary properties to fulfill these needs. 1.4 Thesis Organization Work on this thesis was organized into th ree major parts: a system design section detailing the digital/analog design and fabr ication of a wireless sensor node, the simulation and fabrication of a slot-coupled aperture antenna for wireless sensor nodes, and finally a conclusion summarizing the enti re project giving ideas, suggestions, and warnings about setting out to proceed with such a project.
6 CHAPTER 2 WIRELESS LAN SENSOR SYSTEM DESIGN 2.1 Introduction This chapter presents the requirements, design, construction, and testing of an 802.11b wireless sensor network (WSN). A WS N is a set of miniature sensor nodes networked together using RF connectivity. WSNs are deployed to monitor and collect data of external environmental variables. These networks can be deployed in remote and hostile areas for many different types of appli cations. To name a few, WSNs can be used for habitat monitoring, object tracking, seismi c detection, military surveillance, fire detection, or traffic monitoring for periods of time extending to weeks, months or even years. This chapter will be gin with the specific requireme nts of our WSN project. The design process is then presented. This involves using an 802.11b chipset reference design for the radio and networking protocol. A micro-controller is used to interface with the 802.11b chipset and process sensor data. The entire system is designed with compactness in mind, therefore folded-flex printe d circuits using ultrathin substrates are employed. Finally, performance results are give n and compared to those of the reference design. A block diagram of the system is shown in Figure 2.1 Figure 2.1 Communication Block Diagram for Sensor Node
7 2.2 System Requirements The motivating reason behind designing a WS N was to deploy sensors for marine ecosystem monitoring, harbor security and de fense, and environmental monitoring that could connect to existing 802.11b networks on land. Basing a sensor network on the 802.11b protocol provides a solid standard th at can support the data communication of a large number of nodes while adding reliability to the wireless network platform, relative security, and ease of use. It helps fulfill several requirements in this design by allowing the network to be able to s upport a large number of intelligen t nodes that are small in size and that will be able to gather, m easure, and control relevant data. One of the most important requirements is that the system be able to function with existing wireless networks already in place. By far, the most popular wireless protocol is the IEEE 802.11 specification. By using this pr otocol for the WSN, it can easily integrate itself with existing 802.11 wireless netw orks. The endpoints in an 802.11b network adhere to a star topology wher e all nodes are conn ected to a single hub. This is shown in Figure 2.2. The hub requires greater messa ge handling, intelligence, and routing capabilities than the other nodes. If a communication link is cut only one node is affected. A big disadvantage to this setup is if a hub is incapacitated the network is destroyed . To achieve scalability the node s in a WSN must be lo w-cost. Therefore, choosing low-cost components played an impor tant role in the design of the system.
8 Figure 2.2 Basic Network Star Topology for 802.11b Systems There must be sufficient bandwidth to support a large number of nodes. The 802.11b specification uses direct sequence sp read spectrum technology to support data rates up to 11 Mbps in the 2.4 GHz frequenc y ISM band . Supporting such fast data rates consumes more energy but since data transmissions ar e fast, the nodes spend more time in sleep mode than in awake mode. One of the most challenging, but necessar y, requirements of this project was to miniaturize it as much as possible. From the beginning, a folded-flex design was envisioned using extremely thin substrates like liquid crystal polym er or polyimide. A graphic representation of this goal is shown in Figures 2.3 and 2.4. As shown in these figures, the node is folded into three separate sections. These three sections are separated by a thin folded-flex layer which connects the sensor section, 802.11b digital processing section (Media Access Control Processo r), and the 802.11b RF section (radio and antenna). Network Coordinator Endpoint Internet Communications Flow
9 1.75" 1.25" 0.31"P o ly i mi d e f l e x r e g i o n Figure 2.3 Graphic of Packaging Solu tion for Wireless Sensor Node 2 Layer Polyimide 2 Layer FR-4 Folded Flex Interconnects BGA Components SMD Devices Double-Sided Board Figure 2.4 Three Sections of the Wireless Sensor Node 2.3 System Design 2.3.1 System Overview The system is made up of an Intersil PRISM 3 chipset, a Microchip PICmicro (PIC18F452) micro-processor, ex ternal memory, and onboard sensors. The sensors were made modular by placing the PIC microcontroller on a separate board than the radio. In this manner, different sensor boards could be attached to the same radio. Power regulation is handled on the sensor board as well. Power is supplied by a small 1600mAh lithium-ion polymer batter. Finally, to cons erve space, small dual chip antennas were used to connect to the 802.11b radio. Th e block diagram in Figure 2.5 shows the integrated platform.
10 text PIC Microcontroller 10-bit A/D Converter 32k On-Chip Program Memory Power Saving Sleep Mode Sensor 1 Baseband Processor/ MAC Up/Down Converter Power Amp Sensor 2 Lithium Ion Polymer Battery Sensor ...802.11b Radio W ireless Sensor Node Dual Internal Antennas w/ Diversity External Antenna Connector Optional High Gain External Antenna 3-wire SPI RS-45 or RS-232 USART Parallel Slave Port In-Circuit Debug (ICD) Figure 2.5 System Diagram for the WSN At the definition phase of the project, the block diagram was drawn after some research to see if the PIC had sufficient resources to drive the PCMCIA WLAN card as well as support the number of sensors neede d. As more research was conducted each block was converted into a schematic piece-w ise until the design phase of the project was completed. An example of this is shown in Figure 2.6.
11 PIC18LF6620 Photodiode Temperature Sensor PRISM 3 802.11b Chipset Battery Level 3.3V Regulator Battery Buttons Reset User Sel. 8 Power LED RX/TX LED Oscillator AN6 AN5 OSC1 OSC2 RA5 RG4 RD0 RD7 RB0 RB7 /Status Figure 2.6 Going From Defi nition to Design Phase 2.3.2 Power Management Power is the scarcest resource for wirele ss sensor networks. Wireless sensor networks are primarily designed to monitor data in remote and hostile environments. Low power consumption to achieve maximum de vice lifetimes is imperative. For this reason, careful design to conserve maximum power is necessary. In Figure 2.7, a graph is shown that i ndicates how controlling the RF output power affects the power consumed by the PR ISM 3 wireless LAN card. It is obvious from the results that changes in the RF pow er level have a large effect on the overall power consumption. Therefore, by using th e Intersil MAC to cont rol the output RF power, the overall power consumption can be decreased by more that a factor of 2.
12 Figure 2.7 PRISM 3 Power Consumption Versus RF Output Power Designing an 802.11b radio to run off of ba ttery power for extended periods of time is difficult due to its large consumpti on of current. 802.11b chipsets consume a lot of current while processing data and even more while transmitting data. It is not unusual for some chipsets to consume close to 400mA while transmitting at maximum data rates. Current measurements were taken using an off-the-shelf 802.11b PCMCIA card with an Intersil PRISM 3 chipset to determin e how much power an 802.11b WSN node would require. The measurement results for the PCMCIA card are shown in Table 1. The results were better than expected. The cu rrent consumption of a Microchip PIC microprocessor was also added because it would be used to interface with the sensors and PRISM 3 chipset. Further data was also tabulated to get an estimated unit life expectancy. This data is shown in Table 2. Using all this data along with certain assumptions like the amount of sensor da ta transmitted and battery capacity, it was determined that using a 1600mAh battery would yield a life expectancy of approximately 5 months if it was turned on on ce an hour to check fo r connectivity. This data is shown in Table 3 along w ith different operating scenarios.
13 Table 2.1 Current Consumption for PRISM 3 and PIC18F452 # Description Unit current drain (mA) Standby Rx Tx 1 PRISM 3 WLAN Radio (3.3V) 60 1 Mbps N/A 167 167 2 Mbps N/A 176 176 5.5 Mbps N/A 205 205 11 Mbps N/A 230 230 2 ISL3984 Power Amplifier 2 2 100 3 PIC (PIC18F452) 0 4 4 Total current drain (mA) TOTAL: 62.00 1 Mbps N/A 173 271 2 Mbps N/A 182 280 5.5 Mbps N/A 211 309 11 Mbps N/A 236 334 Table 2.2 Current Consumption at Startup for PRISM 3 and PIC18F452 System Parameters: Current (mA) Time (s) 1 Power Up / Initialization (PRISM) 135 1 2 Association (PRISM) 165 2 3 Power Down (PRISM) 130 1 4 PIC (sensor) 0.4 always on 5 Actual data rates (Mbps) 1 Mbps 0.7 2 Mbps 1.3 5.5 Mbps 2.8 11 Mbps 5.5
14 Table 2.3 Battery Life Expectancy of WSN Under Various Scenarios Battery Life 1 Scenario: Radio left on permanently in the following modes: Standby (hrs) Rx (hrs) Tx (hrs) 25.81 N/A N/A 1 Mbps N/A 9.25 5.90 2 Mbps N/A 8.79 5.71 5.5 Mbps N/A 7.58 5.18 11 Mbps N/A 6.78 4.79 2 Scenario Radio turns on once/day at a specific time to send data. Operating Lifetime Days Download Time (s) 1 Mbps 158.4 5.71 2 Mbps 161.4 3.08 5.5 Mbps 163.3 1.43 11 Mbps 164.3 0.73 3 Scenario Radio turns on twice/day at specific times to check for connectivity 1 Mbps 156.68 5.71 2 Mbps 159.65 3.08 5.5 Mbps 161.53 1.43 11 Mbps 162.44 0.73 4 Scenario Radio turns on once/hr to check for connectivity 1 Mbps 135.29 5.71 2 Mbps 137.50 3.08 5.5 Mbps 138.89 1.43 11 Mbps 139.56 0.73 The smallest batteries ava ilable that have the necessary capacity of 1600 mAh are lithium-ion polymer types. Due to the need to have a battery with such a large capacity, this would determine the overall size of the de vice since the smallest batteries available were approximate 54 mm x 32 mm x 6 mm. More was done in the design of the wire less sensor node to conserve power than simple shut-down schemes as described above. A low dropout power regulator was chosen that was efficient, had a low quiescent current, and gave off little heat. It was also chosen because it required very little external circuitry to help conserve precious PCB space. Another thing that was done was to increase the values of pull-up resistors throughout the design. Usually, 10 kOhm pullup resistors are used throughout a design. However, if high value resistors in the mega -Ohm range are used current consumption is reduced. Finally, the PIC processor and the PRISM 3 chipset were placed in sleep mode whenever possible.
15 The Intersil PRISM chipset offers a power management option that can cut its power consumption in half while in infrastruc ture mode . With this option enabled, it was observed that there was a noticeable slowing of network activity. This was mostly noticed by a lengthening of the ping respons e time. This caused the chipset to have very long startup times which resulted in higher overall current consumption. It was decided that it was best to power down the ra dio completely. This capability was added to the PIC sensor board. However, it was obs erved that the radio must be reset first. Otherwise, it would still draw power. 2.3.3 Wireless LAN (802.11b) Radio The radio chosen to be used in the WSN design was an Intersil PRISM 3 WLAN PC card utilizing the IEEE 802.11 Direct Sequence spec ification. It provides data rates of 1, 2, 5.5, and 11 Mbps. The radio is design ed to operate in the 2.4GHz ISM frequency band using channels 1 to 11, as specified by the FCC. Acting as a receiver, the 2.4 GHz signal enters via the antennas and the diversity switch. Then it goes through the band pass filters and transmit/rece ive switch. Finally, the signa l goes through the direct up/down converter where it is converted to RX IQ signals. The IQ signals are converted into data bits by the base band processor (BBP ) which also controls the antenna diversity switch and transmit/receive switch. The data bits are processed by the BBP on the 802.11b protocol level which also provides the PCMCIA interface to the PICmicro. The path is reversed if the radio is setup as a transmitter. The block diagram is shown in Figure 2.8.
16 Figure 2.8 PRISM 3 Radio Block Diagram The printed circuit board (P CB) for the radio section of the WSN node had to be as small as possible. To this end, we decide d to use an 8 layer flexible printed circuit design. The layer stack is shown is Figure 2.9. Essentially, it is a PCMCIA card that has been shrunk and folded over. The board dimensions are 1.75 x 1.25 x 0.31 (L x W x H). Since the flexible section essentially di vided the board in two, one section was left for the RF and the other for the digital.
17 Figure 2.9 Layer Stack Illustrating the Thicknesses of Each Layer The RF section was designed first since the layout would have to be designed manually due to the RF transmission lines and the high level of mini aturization needed. The RF section is made up of a down-convers ion transceiver, a 5 Ghz voltage controlled oscillator (VCO), power amplifier, voltage regulators, antennas, and external supporting circuitry (mostly filters and capacitors). Care ful attention was given to the RF section of the design to achieve both maximum miniaturiz ation and performance. All microwave signal transmission lines are coplanar wavegui de with 50 Ohms impedance. Coplanar waveguide is used because it offers low loss a nd a relatively high degree of shielding. A more detailed layer stack-up is shown in Fi gure 2.10. Coplanar waveguide (CPW) uses a ground conductor that is coplanar with the signal conductor. The impedance is controlled by the thickness of the substrate material, it s dielectric constant, th e signal line width and to a lesser degree, the ground gap . A CPW transmission lines cr oss-section is shown in Figure 2.11. Its physical characteri stics include the conductor width ( ), the conductor thickness ( ), the slot width ( s ), the substrate height ( d ), and the relative permittivity ( ). The values for these variables were found to be .0007 inches for the conductor thickness,
18 .005 inches for the substrate height, and 4.3 for the relative permittivity of FR4. The conductor width was calculated to be .0075 inches and the slot width to be .010 inches referenced to ground on layer 2 . Because the impedance of the a coplanar waveguide printed circuit trace depends so strongly on the dielectric constant of the substrate material and its thickness, it was extremely im portant that the FR4 PCB material did not deviate from 50 Ohms. For this reason, a P CB manufacturer was chos en that offered a controlled impedance of 50 +/10%. This is of utmost importance to ensure proper propagation of RF signals throughout the boa rd. The top and bottom ground planes are connected by an array of ground vias which ensure a low impedance ground connection between layers. This also prevents a higher order propagation mode and improves isolation . The total width of the top and bottom ground are not extremely critical as long as they are wide compared to the width of the signal line. The top layer for the final layout of the RF section of the board with the CPW transmission lines highlighted is shown in Figure 2.12. Also shown in this figure is how each section is isolated to reduce noise and cross-talk. The Zero-IF chip, VCO, power amplifier, and 2.4 GHz filter are all blocked off into separate sections. Top Layer Core (5mil) G round Plane ((Multiple Nets) ) Prepreg (3mil) Polyimide Layer 1 Core (2mil) Polyimide Layer 2 Prepreg (3mil) Power Lines Layer Core (5mil) MidLayer6 Prepreg (3mil) MidLayer7 Core (5mil) Bottom Layer Figure 2.10 Layer Stack of Ra dio Side of the 802.11b Board Figure 2.11 Cross-section of a CPW Transmission Line
19 Figure 2.12 Top Layer of RF Section Hi ghlighting CPW Transmission Lines (surrounding ground pour is not shown for clarity) The ANT-2.45-CHP chip antennas by Li nx Technologies were chosen based on their small form factor and their relatively good performan ce. They have a 0.8 dBi of gain and a beamwidth of 105 degrees. An antenna can be connected ex ternally via a coax connection or the onboard chip antennas can be used for ope ration. Two antennas are not needed since one of them is used only for sp atial diversity (receive only). The chipset has the capability to choose between the two antennas to achieve the best transmission and reception. The digital section of radio is attach ed to the analog section by a folding polyimide substrate that makes up layers three and f our of the printed circuit board. This section of the board is made up of Inters ils baseband processor/MAC chip, 1Mbit of flash memory, 4Mbit of SRAM, a 44 MHz osci llator, a 32 kHz crystal, and three voltage regulators. Due to the 802.11b standard requirem ent, the digital subsystem of the radio is very similar to the reference design that is given by Intersil under a non-disclosure agreement. There were two minor changes th at were made. One change was that a 3.3V regulator was substituted for a 1.8V regulator because the only SRAM that could be
20 sourced had a minimum voltage requirement of 2.8V. The other change was the addition of the folded flex connectors to connect th e external sensor board. The major changes were concentrated on the layout to miniaturize the design to the fullest extent. 2.3.4 PICmicro Processor and Sensor Board The PIC18LF6620 processor was chosen pr imarily because the only development kit available to work with 802.11b cards came with a PIC processor. However, the PIC was a good choice for the intended application; it consumed very low power and had the necessary resources to connect several sensor s. The processor has 16K words of flash program memory, 1536 bytes of data memory, and 256 bytes of data EEPROM. It has in-circuit reprogramming ab ility via the standard Mi crochip ICD connector. Power is supplied using a lithium ion polymer battery. These batteries are very slim, extremely lightweight, and flexible. Each cell outputs 3.7V at 1600mAh while delivering 2C continuous discharge with a maxi mum discharge of 3C. They also have excellent long-term self-discharge rates at less than 8% per month. Power is regulated using a National Semiconductor Low Dropout vo ltage regulator. The regulator supplies the 3.3V to the 802.11b radio at battery supply voltages as low as 2V. The PIC processor communicates with the Intersil chipset through the PCMCIA interface by way of a device driver that is supplied by the development kit from Iosoft. PCMCIA interface chips are targeted only for 32 bit laptop applications. Therefore, they are too complicated for use with an embedde d controller. However, communication with the Intersil PCMCIA chipset is possible by us ing either a 8 bit or 16 bit bus, address lines, and memory I/O lines. The schematic of this interface is shown in Figure 2.13. Once the interface is initialized, the PCMCIA interface is transpar ent, so the I/O cycles to the PRISM closely resemble those to an ethe rnet controller. There is also a memory interface on the card which is re ferred to as attribute memory . This gives information about the chipset in st andardized tuple format. A 10-b it address and 8-bit data bus is connected to the PICmicro, together with f our strobe lines; two to read or write to attribute memory, and two to read or write to the I/O devices. Further details on how the PIC achieves communication with the MAC cannot be disclosed due to the restrictions of
21 the Intersil non-disclosure agreement (NDA), no information can be provided on the internal architecture and opera tion of the PRISM chipset. RA4 RB0 RA5_LED_Switch RB3 RB2 RB1 RD1 HRESET RC2_ALL_PE RB3 RA0 RB1 PB7_PGD RD4 RD3 RD2 RC0 RD7 RA4 RA3 RB0 RB6_PGC RD0 RB4 RD5 RD6 RA1 RA2 RC1 RB5 RB2 RB5 PB7_PGD RB6_PGC RB4 RC7_RX RD7 RC6_TX RC2_ALL_PE MCLR RA3 RA0 RA1 RA2 LIGHT Temperature 1.8V_REF D8 D9 RC5 D10 D11 D12 D13 D14 D15 HCE2 HWAIT HNPACK HSTSCHG HRESET RD6 RD0 RC0 RD1 RD2 RD4 RD3 RC1 RD5 HA9 1 HA8 2 HA7 3 HA6 4 HA5 5 HA4 6 HA3 7 HA2 8 HA1 9 HA0 10 HD7 11 HD6 12 HD5 13 HD4 14 HD3 15 HD2 16 HD1 17 HD0 18 HOE 19 HREG 20 HIORD 21 HIOWR 22 HCE1 23 GND 24 3.3V 25 HIREQ 26 ALL_PE 27 GND_ 28 HWE 29 RESET 30J1 Conn Top contact RE2/CS 64 RF1/AN6/C2OUT 17 RF0/AN5 18 AVdd 19 AVss 20 RA3/AN3/Vref+ 21 RA2/AN2/Vref22 RA1/AN1 23 RA0/AN0 24 Vss2 25 Vdd2 26 RA5/AN4/LVDIN 27 RA4/TOCKI 28 RC1/T1OSO/T13CLK 29 RC0/T1OSO/T13CLK 30 RC6/TX1/CK1 31 RC7/RX1/CT1 32 RE3 63 RE4 62 RE5 61 RE6 60 RE7/CCP2 59 RD0/PSP0 58 Vdd5 57 Vss5 56 RD1/PSP1 55 RD2/PSP2 54 RD3/PSP3 53 RD4/PSP4 52 RD5/PSP5 51 RD6/PSP6 50 RD7/PSP7 49 RE1/WR 1 RE0/RD 2 RG0/CCP3 3 RG1/TX2/CK2 4 RG2/RX2/DT2 5 RG3/CCP4 6 MCLR/Vpp 7 RG4/CCP5 8 Vss 9 Vdd 10 RF7/SS 11 RF6/AN11 12 RF5/AN10/CVref 13 RF4/AN9 14 RF3/AN8 15 RF2/AN7/C1OUT 16 RB0/INT0 48 RB1/INT1 47 RB2/INT2 46 RB3/INT3 45 RB4/KB10 44 RB5/KBL1/PGM 43 RB6/KBL2/PGC 42 Vss3 41 OSC2/CLK0/RA6 40 OSC1/CLK1 39 Vdd3 38 RB7/KB13/PGD 37 RC5/SDO 36 RC4/SDI/SDA 35 RC3/SCK/SCL 34 RC2/CCP1 33U1 PIC18LF6620 R4 10.0K 1 2Y1 19.6608MHz 3.3V 3.3V 3.3V 3.3V 3.3V 3.3V HE1 REG Figure 2.13 Schematic of PIC 18LF6620 to PCMCIA Interface The wireless sensor node also includes a temperature sens or and light sensor each connected to an ADC channel on the PIC. Th e temperature sensor is a 0603 sized Murata NCP18XH103J03RB thermistor. It allows the sensor node to measure ambient temperature. It can be used to power down the node if the ambient temperature surpasses the nodes operating range. The thermistor form s the top half of a voltage divider that is connected in series with a 10 kOhm resistor Higher temperatures result in a higher thermistor resistance which corresponds to a lower ADC value. The light sensor is BurrBrown OPT101 photodiode. The light sensor ca n be used to measure clarity level in water, predict the expected number of day light hours, or discriminate between cloud
22 cover and nightfall. The output voltage of the sensor increases linearly with light intensity. The sensor circ uit is shown if Figure 2.14. RA4 RB0 RA5_LED_Switch RB3 RB2 RB1 LIGHT Temperature RB5 PB7_PGD RB6_PGC RB4 RC7_RX RD7 RC6_TX RC2_ALL_PE MCLR RA3 RA0 RA1 RA2 LIGHT Temperature 1.8V_REF D8 D9 RC5 D10 D11 D12 D13 D14 D15 HCE2 HWAIT HNPACK HSTSCHG HRESET RD6 RD0 RC0 RD1 RD2 RD4 RD3 RC1 RD5 R14 10.0K 0.1% t RT1 Thermistor 10K C14 .1uF C16 1500pF OMIT C15 .1uF OMIT R12 200Kohm Vs 1 -In 2 -V 3 FB 1Mo 4 Out 5 NC 6 NC7 7 GND 8U4 OPT101 C17 .1uF OMIT RE2/CS 64 RF1/AN6/C2OUT 17 RF0/AN5 18 AVdd 19 AVss 20 RA3/AN3/Vref+ 21 RA2/AN2/Vref22 RA1/AN1 23 RA0/AN0 24 Vss2 25 Vdd2 26 RA5/AN4/LVDIN 27 RA4/TOCKI 28 RC1/T1OSO/T13CLK 29 RC0/T1OSO/T13CLK 30 RC6/TX1/CK1 31 RC7/RX1/CT1 32 RE3 63 RE4 62 RE5 61 RE6 60 RE7/CCP2 59 RD0/PSP0 58 Vdd5 57 Vss5 56 RD1/PSP1 55 RD2/PSP2 54 RD3/PSP3 53 RD4/PSP4 52 RD5/PSP5 51 RD6/PSP6 50 RD7/PSP7 49 RE1/WR 1 RE0/RD 2 RG0/CCP3 3 RG1/TX2/CK2 4 RG2/RX2/DT2 5 RG3/CCP4 6 MCLR/Vpp 7 RG4/CCP5 8 Vss 9 Vdd 10 RF7/SS 11 RF6/AN11 12 RF5/AN10/CVref 13 RF4/AN9 14 RF3/AN8 15 RF2/AN7/C1OUT 16 RB0/INT0 48 RB1/INT1 47 RB2/INT2 46 RB3/INT3 45 RB4/KB10 44 RB5/KBL1/PGM 43 RB6/KBL2/PGC 42 Vss3 41 OSC2/CLK0/RA6 40 OSC1/CLK1 39 Vdd3 38 RB7/KB13/PGD 37 RC5/SDO 36 RC4/SDI/SDA 35 RC3/SCK/SCL 34 RC2/CCP1 33U1 PIC18LF6620 R4 10.0K 1 2Y1 19.6608MHz 3.3V 3.3V 3.3V 3.3V 3.3V 3.3V 3.3V Figure 2.14 Sensor Circuit Show ing Connections to PICmicro 2.4 Packaging 2.4.1 Overview The primary purpose of the packaging is to protect the electronics from its external environment, but it also limits, and is limited by, the circu it board size, antenna design, and sensor placement. To this e ffect, an 802.11b radio was designed onto a compact folded flex package to interface with a separate sensor module.
23 2.4.2 Folded Flex Design Using Polyimide Flexible circuits on polyimide were used to miniaturize the design as well as make it conformal. Flexible circuits are great for applications where lightweight, compact design, and reliability ar e all critical because a flex circuit is built to bend, fold, twist, and wrap in extremely tight areas. In our design, we used these features to fold the entire printed circuit boa rd over itself as shown in Figure 2.15. An actual picture of the first version of the radio is shown in Figure 2.16. Battery Multilayer polyimide (0.2mm) w/ embedded passives Single layer flexible polyimide (0.05mm) Integrated circuits 30mm 8mm Multiple Sensors 802.11b Radio Figure 2.15 Cross-section of Folded Radio Design Figure 2.16 Picture of Actual Radio and Sensor Boards Although rigid-flex designs are not new, we soon discovered they had limitations. On the first design of the radi o board, the flex section of the board would extend out and connect to the sensor module as shown in Fi gure 2.17. In order for the flexible extension to meet the specifications to connect to the Hirose FH12A-30S-0.5SH connector a
24 stiffener had to be added underneath to m eet the required thickness as shown in Figure 2.18. However, after a few repeated insertions into the connector, some of the fingers would lose contact to the connector due to th e stiffener flexing under the pressure of the connector. In an attempt to remedy this, a thin layer of solder was added to the fingers to add thickness to make up for flexing of th e stiffener. However, the solder would eventually give and lose cont act again. Repeating the pr ocess over and over again had the result of damaging the fingers of the ra dio board as shown in Figure 2.19. The only solution was a redesign of the board complete ly eliminating the flex-tail and adding a flex connector on either end as shown before in Figures 2.16. Figure 2.17 Damaged Fingers on Radio Board Tail
25 Figure 2.18 Stiffener on Bottom Side of Flex-Tail Figure 2.19 Close-up of Dama ged Fingers on Flex-Tail 2.5 System Performance Radio performance was tested by compari ng the performance of an off-the-shelf Intersil PRISM 3 reference design to the WS N design. Initially, the new design was used in a PC laptop which was setup as a host. It was used like an y other 802.11b PCMCIA card to connect wirelessly with an access poi nt. After extensive us e in both indoor and
26 outdoor environments the performance of the WSN radio was identical to the reference design. The specifications of the refere nce design are given in Table 2.4. To place our initial impressions to the te st, the RF output power delivered to the antenna was measured and compared to the reference design 802.11b WLAN card. The amount of output power sent out from the tr ansmitter has a large effect in the range and performance of the network system. To meas ure the output power, we connected the RF output jack to the spectrum analyzer us ing the Murata MM126036 coax connector. Using test software provided by Intersil the WSN radio was placed in continuous transmit mode on channel six. The spectrum an alyzer was set to a center frequency of 2.437 GHz, a span of 80 Mhz, and an amplitude of +12 dBm. All other settings were left to their default settings. According to In tersil, the bounceback in signal level appearing on either side of the main spectrum, which is called regrowth, is caused by nonlinearities in the transmitter . For a compliant IEEE 802.11b transmitter, the first regrowth should be at least -30 dBc and the second regr owth at least -50 dBc. The spectrum for the WSN design is shown in Figure 2.21. It can be seen from the spectrum that the output power of the WSN design is we ll within the IEEE 802.11b specifications. Figure 2.20 Frequency Spectrum for WSN Radio
27 Table 2.4 Wireless Sensor Node Specifications Standard IEEE 802.11b Direct Sequence Specification Frequency 2.412 2.462 GHz Modulation DBPSK, DQPSK, and CCK Channels 11 Data Rate 1, 2, 5.5, and 11 Mbps MAC CSMA/CA, WEP, RTS/CTS RF Power 13.5 dBm Range 350 ft indoors, 900 ft outdoors Sensitivity 1 Mbps, 8% PER: -91 dBm 2 Mbps, 8% PER: -88 dBm 5.5 Mbps, 8% PER: -87 dBm 11 Mbps, 8% PER: -84 dBm Security 64/128 bit WEP encryption Antenna Internal diversity w/ conn ector for optional external antenna Voltage 3.0V 3.6V Current Sleep: Quiescent: 40 mA Transmit: 280 mA Receive: 260 mA Temp -40 +75C I/O Interface Serial RS232 Serial Interface Asynch,. 3.3V level PCB Flex FR4/Polyimide hybrid Weight 30 grams
28 2.6 Chapter Summary The experiences in designing an 802.11b wi reless sensor node using folded-flex technology has been reported. We first presen ted the requirements needed of the sensor node. These requirements included bei ng able to function within existing 802.11b networks, being able to suppor t a large amount of nodes, bein g able to function on battery power, and having a compact form factor. In the design phase of the project, the goal of achieving each individual requirement was met. An Intersil 802.11b chipset was used in a new folded-flex package and was made modul ar by making the sensor interchangeable. Finally, the radio was used and tested al ongside off-the-shelf reference design WLAN cards to test for reliab ility and performance.
29 CHAPTER 3 SLOT-COUPLED MICROSTRIP ANTENNA DESIGN 3.1 Introduction The slot-coupled antenna presented in this thesis was designed at a frequency of 2.4 GHz for use in 802.11b applications. The positive features offered by this antenna make it a good candidate for use in wireless sens or networks (WSNs) that operate in the 2.4 GHz frequency band. This includes having the ability to use different materials for the antenna and feed substrates and the promis e of wide impedance bandwidths of 5% to 50% . This chapter begins with the specific requirement s that are inherent to 802.11b applications. The design process is then discussed followed by the simulation results of the design using Ansofts HFSS and Agilents Momentum as part of the Advanced Design System software. Finally, measurement results are given and compared to results acquired by simulation. 3.2 Design Requirements Based on the testing of the 802.11b WSN de scribed in chapter 2, an antenna optimized for use in this system would have to adhere to specific design requirements. The 802.11b specifications are shown in Table 3.1. Form must follow function and the size and shape of the antenna must align itsel f well with the form factor of a miniature sensor node. To this end, the antennas feed-l ine substrate must be the same as the sensor nodes substrate. No thickness limit was placed on the patchs substrate. It must have a width and length no larger than 33 millimeter s and 48 millimeters respectively. Also, the antenna must be largely insensitive to orient ation. Furthermore, it must have a minimum gain of 3 dBi with a minimum impedance bandw idth of 3.5% with a VSWR less than 2:1. This allows the real part of the antenn a impedance to fall between 25 and 100 Ohms
30 which are the intersections of 2:1 VSWR circle on the Smith Chart with the real axis. This bandwidth figure is expressed as a per centage of the frequency difference over the center frequency of the bandwidth . 3.3 Design Process Figure 3.1 shows the basic st ructure of the aperture coupled microstrip patch antenna. The feed line is on the bottom of the feed substr ate while the radiating patch element is located on top of the antenna subs trate. The top and bottom thicknesses of the substrates will determine the antennas radi ation pattern and bandwidth. According to Pozar , thinner feed line substrates offer less spurious radiation at a cost of higher loss. The recommended thickness is 0.01 to 0.02 Due to our design c onstraints, the feed line substrate has to be FR4 which has a permittivity of 4.3 and a thickness of 1.575 millimeters which falls within the recommended range. Increasing the antenna substrate thickness increases bandwidth and the coupling level while a lower permittivity will give a wider impedance bandwidth and reduced surface wave excitation . Rogers RT/Duroid 5880 substrate material was used due to its low permittivity of 2.2 and a thickness of 1.575 millimeters. Figure 3.1 Basic Components of an Aperture Coupled Patch Antenna
31 The dimensions of the patch antenna were determined using the transmission-line model technique to analyze different patch sizes. Although it does not provide the most accurate results, it is quick to use and sheds phys ical insight into the design . The patch length determines the resonant frequency of the antenna. Resonance occurs when the input impedance is purely real . To obtain an initial value of the length (l) the following equation was used (fr = 2.45 GHz): Using equation 3.1, the effective relative permittivity (eff) can be calculated as follows: The patch width will affect the resonant resist ance of the antenna. The wider the patch is the lower the resistance of the antenna. The width should be less than the length to avoid having a square patch which may result in the generation of high cross polarization levels . The width of the patch was ca lculated using the following equation: The patch length was calculated to be 39 m illimeters and the patch width was calculated to be 32 millimeters. The feed line and slot dimensions were calculated next. The feed line width not only determines the characteristic impedance of the line but also the coupling to the slot. The thinner the feed line is the stronger the c oupling to the slot . For a characteristic impedance of 50 Ohms (Z0), the width of the feed line was calculated to be 2.9 millimeters using the following equations: (3.2) (3.3) (3.1)
32 The slot length affects the level of coupli ng to the patch and also affects the amount of back radiation. For this reason, the slot shou ld not be made any larger than is required to get a good impedance match. The slot width also affects the level of coupling but to a lesser degree than the slot length . A good rati o of slot length to slot width is 1/10 . The slot length used as a starting point was 12 millimeters and the sloth width was 1.2 millimeters. Finally, the stub length is used to tune the excess reactance of the antenna. According to Pozar , the length of the stub is usually slightly less than g/4 in length. If the stub is shortened, it will move the impedance locus on the Smith chart in the capacitive direction. The total feed line le ngth including the stub th at was used was 42 millimeters. A Smith chart of the impeda nce locus for a typical aperture coupled microstrip antenna is shown in Figure 3.2. (3.4) (3.5)
33 Figure 3.2 Smith Chart Plot of the Impe dance Locus as a Function of Frequency All of the above calculated values were inserted into Ansofts high frequency structure simulator (HFSS) software. The 3D representation of the antenna is HFSS is shown in figure 3.3.
34 Figure 3.3 3-D Representation of the Aperture Coupled Antenna in HFSS 3.4 Antenna Simulation 3.4.1 Model Construction The aperture coupled antenna was simulated in HFSS as structures with air as the surrounding medium. A port face was defined at the input end of the feed line as the location where the fields will enter the stru cture. The port had to be large enough to accommodate the field pattern of the TEM mode common to microstrip so it is larger than the end of the feed line as shown in Figure 3.4.
35 Figure 3.4 Defined Excitati on Port in HFSS Simulation Spending the time to carefully set up all of the parameters in HFSS is essential to get the most accurate results while drasti cally speeding up the time it takes to reach convergence. All of the materials in the de sign must be set up in HFSS and assigned to their respective locations. Copper was assigne d to the metalization, air to the large box surrounding the substrate, Rogers duroid 5880 to the antenna substrate, and FR4 to the feed substrate. Also, the boundaries between these substances must be defined. The copper face was defined as Perfect E meani ng that the metal is a perfect conductor and the radiation boundary was setup as an air box. Careful attention must also be paid to how the analytic mesh is constructed. HFSS will perform an adaptive solution proce ss where it automatically defines and redefines the mesh dependent on the results ob tained in the first of multiple passes. Because HFSS tends to concentrate refinement where the fields are most prevalent, it may not automatically give the best results for an antenna. In the case of an antenna the most important fields, which are not necessari ly the strongest, lie in the surrounding air or substrate, not in the circuit itself. To enc ourage HFSS to refine the mesh in the correct areas, the mesh was concentrated in the uppe r region of the air box. Also, the HFSS was setup to initially calculate the mesh using free space lambda. As a result, the time it takes to reach convergence is minimized while increasing the accuracy of the simulation.
36 3.4.2 Simulation Results From the beginning, the HFSS simulation results looked very good needing very little tweaking of the calculated antenna dimensions. The return loss for the antenna with the parameters specified above is shown in Figure 3.5. The Smith chart plot of the impedance locus versus frequency for the aper ture coupled microstrip antenna is shown in Figure 3.6. Figure 3.5 Return Loss Results fo r the Aperture Coupled Antenna
37 Figure 3.6 Smith Chart Plot of the Impedance Locus Versus Frequency Lengthening the tuning stub has the effect of moving the impe dance locus in the inductive direction on the Smith chart. As an example, Figure 3.7 shows the impedance locus for a feed line length of 40.5 millimeters (lengthened by 1.5 millimeters). It is easy to see that by adjusting this parameter th e antenna can be tuned to remove the excess reactance. Adjusting the slot length has the effect of cha nging the size of the impedance locus. Increasing the slot length increases the diameter of the locus. To illustrate this point, Figure 3.8 shows the Smith chart of the impedance locus for a slot width that has been shortened to 11 millimeters (shortened by 1 millimeter). Unfortunately, this also had the negative effect of increasing the re sonant frequency to 2.47 GHz. Obviously, the slot width is frequency depe ndent. Therefore, the length of the patch must also be adjusted to compensate for this. Nevertheless, this is another parameter that is used to properly tune the antenna to reach optimum matching.
38 Figure 3.7 Smith Chart Plot of the Impe dance with Feed Line Length of 40.5 mm Figure 3.8 Smith Chart Plot of the Impe dance Locus with 11 mm Slot Width
39 Another way of looking at the return loss data is to di splay the phase response. The phase response using the original antenn a parameters is shown if Figure 3.9. The phase response with the tuning stub lengthe ned to 43.5 millimeters is shown in Figure 3.10. Finally, the phase response with the slot length shortened to 11 millimeters is shown in Figure 3.11. The change in the resona nt frequency can be clearly seen in the phase response of the antenna. Figure 3.9 Phase Response of S11 Results Figure 3.10 Phase Response of S11 Resu lts Using 43.5 mm Feed Line Length
40 Figure 3.11 Phase Response of S11 Results Using 11 mm Slot Length The VSWR of the antenna is shown in Figure 3.12. The VSWR 2:1 bandwidth of the antenna is 2.5% from 2.41 to 2.47 Ghz which falls below the 3.5% requirement. The calculated maximum VSWR is 3.9:1 at 80 MHz which must be improved upon. Figure 3.12 VSWR Plot of th e Aperture Coupled Antenna The radiation pattern plot for the aperture coupled microstrip antenna is shown in Figure 3.13. The forward radiation pattern looks identical to that of regular microstrip
41 antenna patches. What is not typical is th e back radiation lobe which is caused by the radiation from the coupling slot. However, it is kept to a minimum by using the smallest slot size possible. A 3-D radi ation pattern as a f unction of gain is shown in Figure 3.14. The peak gain was found to be 4.14 dBi while the efficiency of the antenna was found to be .86. Table 3.1 summarizes the simulated results for the aperture coupled antenna. Figure 3.13 E -plane (blue) and H -plane (red) Radiation Patterns
42 Figure 3.14 3-D Radiation Pattern as a Function of the Gain Table 3.1 Parameters of the Simulated Aperture Coupled Microstrip Antenna Simulation Results Patch Width Patch Length Feedline Length Slot Width Slot Length Sub. H (FR4/Duroid) 0 (FR4/Duroid) fr GHz VSWR BW 27 mm 38 mm 42 mm 1.2 mm 12 mm 1.575 / 1.575 4.3 / 2.2 2.4 2:1 2.5% 3.5 Design Modifications Due to the results of the previous simula tions, in order for the antenna to operate within the entire ISM band of 2.40 GHz to 2.4835 GHz, the bandwidth would have to be increased. Some modifications to the aper ture coupled antenna design would become necessary in order to try to achie ve a 3.5% bandwidth (VSWR 2:1). Several modifications to the antenna were investigated in an attempt to address this issue. One change that was made was to increase the slot size while decreasing the
43 feed line width to increase c oupling. Another change was to increase the thickness of the antenna substrate to 3.175 millimeters to see how much of an increase in bandwidth this would give. The slot size was increased to 14 millimeters (1 millimeter increase) while the length of the patch was decreased to 37.5 millimeters (0.5 millimeter decrease) to compensate for change in the resonant freque ncy. The tuning stub was also shortened to 41.3 millimeters (0.7 millimeter decrease). The return loss for the modified antenna can be seen in Figure 3.15. The Smith chart plot of the impedance locus versus frequency is shown in Figure 3.16 and the VSWR plot is shown in Figure 3.17. The changes made to the antenna resulted in a negligib le increase in bandwidth. Figure 3.15 Return Loss Results Usin g Modified Antenna Dimensions
44 Figure 3.16 Smith Chart Plot of Impe dance Locus for Modified Antenna Figure 3.17 VSWR Plot for Modified Antenna
45 The optimizer in HFSS was also used to try to increase the bandwidth. This feature, which HFSS calls optimetrics, was se tup so that a VSWR of 2:1 throughout the 2.40 GHz to 2.48 GHz was the goal. The path length, slot width and length, and stub length were set as variables for the optimet rics software to change. After over 500 iterations were tried, the final outcome was almost identical to the parameters we had started out with. The antenna substrate thickness was th en increased to 3.175 millimeters to increase the bandwidth of the aperture coupled antenna. The return loss for this antenna is shown in Figure 3.18. It is apparent that a better impedance match was achieved with this antenna setup. The Smith chart of the impedance locus ve rsus frequency is shown in Figure 3.19. This confirms the good impedance match of the antenna where the locus is just large enough to pass thr ough the center of the Smith char t. In Figure 3.20, the phase response of the S11 results is given. Fina lly, the VSWR plot is shown in Figure 3.21. This shows the vast improvement in impe dance bandwidth due to the increase in substrate thickness. The calcu lated VSWR 2:1 bandwidth of the antenna is 3.7% which meets the design requirements for operat ing within the 802.11b operational frequency band. Figure 3.18 Return Loss Results 3.175 mm Substrate Thickness
46 Figure 3.19 Smith Chart of Impedance Locus Using 3.175 mm Substrate Thickness Figure 3.20 Phase Response of S11 Resu lts Using 3.175 mm S ubstrate Thickness
47 Figure 3.21 VSWR Plot of Usi ng 3.175 mm Substrate Thickness The radiation pattern of the antenna was studied next. Figure 3.22 shows the radiation pattern plot of the antenna. The beam is slightly more directional and a slight increase in gain is also apparent. Unfortunate ly, the price of using a bigger slot to match the impedance of the antenna properly was accomplished at the cost of greater back radiation. Figure 3.23 shows the 3-D radiation pattern of the aperture coupled antenna as a function of the gain. The peak gain was cal culated to be 4.28 dBi and the efficiency of the antenna 0.89. The better efficiency is most likely due to the better impedance match than the increase of the Rogers Duroid substrate thickness.
48 Figure 3.22 E -plane (Blue) and H -plane (Red) Radiation Patterns Figure 3.23 3-D Radiation Pattern of An tenna Using 3.175 mm Substrate Thickness
49 Table 3.2 shows a comparison of the three different results achieved for each of the antenna designs that were investigated. Based on the si mulation results, the antenna with the greater bandwidth was chosen fo r production. It met all of the design requirements for use within an 802.11b wireless network. Table 3.2 Comparison Table of Simulation Results Simulation Results Comparison Patch Width Patch Length Feedline Length Slot Width Slot Length Sub. H (FR4/Duroid) 0 (FR4/Duroid) fr GHz VSWR BW 27 mm 38 mm 42 mm 1.2 mm 12 mm 1.575 / 1.575 4.3 / 2.2 2.4 2:1 2.5% 27 mm 37.5 41.3 1.4 mm 14 mm 1.575 / 1.575 4.3 / 2.2 2.4 2:1 2.6% 32 mm 36.2 mm 46 mm 1.5 mm 15.4 mm 1.575 / 3.175 4.3 / 2.2 2.4 2:1 3.6% 3.6 Experimental Validation The objective in this section is to veri fy the results in section 3.5 by comparing the simulation and measurements of an apertu re coupled microstrip antenna. For this purpose, a structure was fabricated exactly as described in the prior section. The antenna structure was fabricated usi ng a wet etching process. This method involved a process flow that involved a photoresist exposure, development, and etching of the copper. The top layer of the antenna that included the patch on the 3.175 millimeter thick Rogers Duroid 5880 materi al is shown in Figure 3.24. The coupling aperture, ground plane, and f eed line were place on the bottom two layers of the 1.575 millimeter FR4 substrate material. The ground plane and slot layer are shown in Figure 3.25 and the feed line is shown in Figure 3.26. Notice that the etching process did not do a good job of etching all of the copper off of the slot. The entire structure is shown in Figure 3.27.
50 Figure 3.24 Antenna Patch on 3.175 mm Rogers Duroid 5880 Material Figure 3.25 Coupling Aperture and Ground Plane on 1.575 mm FR4
51 Figure 3.26 Feed Line on Bottom Layer of 1.575 mm FR4 Figure 3.27 Entire Antenna Stru cture Including all Layers
52 The return loss comparison between the meas ured and simulated data is shown in Figure 3.28. It was observed that the fabricated antenna reso nated at a frequency of 2.53 GHz. The small increase in resonant frequency is attributed to two things: One thing is that the chemical etching process did not make the slot thinner on one side. This not only affects the resonant frequency of the antenna but the coupling leve l as well. The other reason is that although the two la yers are held together by ad hesive and pressed together well, there is a slight air gap within th e slot cavity that was not simulated. The Smith Chart plot in Figure 3.29 showed that the antenna exhibited a very good input impedance match to 50 Ohms. However, the impedance locus has rotated towards the inductive side of the Smith Chart co mpared to the simulated results. This can also be seen in the measured phase res ponse plot that is shown in Figure 3.30. -50 -45 -40 -35 -30 -25 -20 -15 -10 -5 0 22.214.171.1242.352.402.452.502.552.602.652.702.752.80 Frequency (GHz)Magnitude (dB) Measured Data Simulated Data Figure 3.28 Return Loss Comparison of Measured vs. Simulated Data
53 Figure 3.29 Smith Chart of Impedance Locus for Antenna
54 Figure 3.30 Phase Response of S11 Results for Antenna Finally, the measured VSWR plot is shown in Figure 3.31. This shows that the measured data correlated well with th e simulated data. The vast improvement in impedance bandwidth due to th e increase in substrate thickness is shown here. The calculated VSWR 2:1 bandwidth of the antenna is 3.7% which meets the design requirements for operating within the 802.11b operational frequency band.
55 Figure 3.31 VSWR Plot for Antenna Another simulation was conducted to confirm that the reason for the increase in the resonant frequency was due to an air gap in the slot and the smaller slot size. An air gap of 0.4 millimeters was placed in HFSS and the slot width was decreased by 0.2 millimeters. The return loss is shown in Figure 3.32 and confirms that the factors mentions did indeed cause the difference in antenna performance. The Smith Chart plot is shown in Figure 3.32
56 Figure 3.32 Simulation of S11 with 0.4mm Air Gap and 1.3mm Slot Width Figure 3.33 Smith Chart Plot with 0.4mm Air Gap and 1.3mm Slot Width
57 3.7 Chapter Summary Simulation results indicate that aperture coupled microstrip antennas can work very well in 802.11b applications. In addition, it is apparent that sl ight changes in the antenna dimensions, such as the tuning stub and slot length, can ha ve large effects in impedance matching. Back radiation was foun d to be lower than expected with a high front-back radiation ratio wit hout using a reflecting plate. These are all promising features that make the aperture coupl ed microstrip antenna a good match for 802.11b applications in general and wirele ss sensor networ ks as well.
58 CHAPTER 4 CONCLUSIONS AND RECOMMENDATIONS The purpose of this project was to dem onstrate that a wire less sensor network could be designed to a) monitor maritime activ ity, b) have a long operational life, c) use existing 802.11b infrastructures, and d) have all this functionality in one small form factor. It was demonstrated that these featur es are not necessarily mutually exclusive. Through careful digital design to conser ve power and limit radio operation, 802.11b chipsets can be used in low power embedded systems. One recommendation would be to always pl an on having at leas t two revisions of a printed circuit board design. Doing so, you ca n plan the design of each board ahead of time. In the design of the radio, the first prototype board was de signed as a proof of concept. There was no miniaturization or use of flexible substrates like polyimide. The entire design was laid out in FR4. This allo wed us the flexibility to design the board with test points and add extra space for probes us ed in troubleshooting the design. The layout of the prototype board is shown in Figure 4.1. Basically, it was an Intersil PCMCIA reference design with a PIC processor interf ace and added power regulators. Much was learned about the challenges that can be encountered when modifying and designing complex design like the 802.11b chipset refere nce design. One issue was the extremely sensitive reset line of the Intersil MAC. New firmware could not be loaded into the MAC due to it constantly reset ting itself due to crosstalk in the lines. By keeping other digitally switching lines away from the reset line and making it as sh ort as possible this issue was avoided when going to the next boa rd revision which incl uded using polyimide as a substrate.
59 4.1 in. 1 8 in Figure 4.1 Layout for First Prototype Board One more recommendation would be to sp end at least 50 percent of the total allotted project time to the design phase of th e project. Start out with a very basic block diagram and then build upon those blocks by making them into smaller blocks. One block diagram showed the inputs of the micr ocontroller by the hardware function of the input. Other hardware blocks are then tu rned to individual ci rcuits. This makes it extremely easy to transition the design into a working schematic or simulation software like PSpice. Finally, by breaki ng up the project into smaller chunks, it easy to see ahead of time what ideas might work and which ones might not. In regards to the aperture coupled ante nna, it was first designed on paper and then simulated in Ansofts HFSS. The results from the simulation were positive and showed that aperture coupled antenna could perform very well in 802.11b applications. Using HFSS was also useful in making obvious was ch anges to the antenna features would most affect its performance. For example, knowi ng that the slot width shifted the tuning frequency, two antennas were fabricated usi ng slightly different widths to tune the antenna to 2.44 GHz. Also, it is obvious that th e simulator allows one to make an infinite amount of adjustments to these features like the length of the tuni ng stub or the patch length to fine tune the pe rformance of the antenna.
60 REFERENCES  Wireless LAN Medium Access Control and Physical Layer Specifications, IEEE Press, Jan. 14, 1999.  D.M. Pozar, Microwave Engineering, 2nd Edition New York: John Wiley & Sons, Inc., 1998, pp. 175-176.  R.E. Collin, Antennas and Radiowave Propagation New York: McGraw-Hill, Inc., 1985, pp. 261-265.  D. Anderson, PCMCIA System Architecture: 16-Bit PC Cards (2nd Edition) New York: Addison-Wesley Professional, September 25, 1995, pp. 144-155.  M. Gast, 802.11 Wireless Networks: The Definitive Guide, Sebastopol, CA: OReilly, 2005, pp. 254-261.  D.M. Pozar, A Review of Aperture C oupled Microstrip Antennas: History, Operation, Development, and Applica tion, IEEE Press, New York, May 1996.  C.A. Balanis, Antenna Theory: Analysis and Design, 2nd Edition New York: John Wiley & Sons, Inc., 1997, pp. 724-752.  I.J Bahl and P. Bhartia, Microstrip Antennas Design Handbook, MA: Artech House, 1980.  D.M Pozar, A Microstrip Antenna Apertu re Coupled to a Microstrip Line, Electronics Letters, Vol. 21, pp. 49-50, Janurary 17, 1985.  I.J. Bahl and P. Bhartia, Microstrip Antennas. Dedham, MA: Artech House, 1982.  D.H. Schaubert, Microst rip antennas, Electromagnetics, vol. 12, pp. 381-401, 1992.  Y. Yashimura, A microstrip slot antenna, IEEE Trans. Antennas Propag., vol AP-29, pp. 2-24, Jan. 1981.
62 Appendix A. Schematic for Sensor Board Title SizeDocument Number Rev Date:Sheet of 1 1 A TitleBlock0RA4 RB0 RA5_LED_Switch RB3 RB2 RB1 LIGHT 1.8V_REF Temperature RD1 HRESET RC2_ALL_PE RB3 RA0 RB1 PB7_PGD RD4 RD3 RD2 RC0 RD7 RA4 RA3 RB0 RB6_PGC RD0 RB4 RD5 RD6 RA1 RA2 RC1 RB5 RB2 PB7_PGD RB6_PGC RC7_RX RC6_TX RC5 MCLR RA5_LED_Switch LEDK RB5 PB7_PGD RB6_PGC RB4 RC7_RX RD7 RC6_TX RC2_ALL_PE MCLR RA3 RA0 RA1 RA2 LIGHT Temperature 1.8V_REF D8 D9 RC5 D10 D11 D12 D13 D14 D15 D10 D11 D12 D13 D14 D15 D8 D9 HCE2 HWAIT HNPACK HSTSCHG HCE2 HWAIT HNPACK HSTSCHG HRESET RD6 RD0 RC0 RD1 RD2 RB6_PGC RD4 RD3 RC1 RD5 R14 10.0K 0.1% t RT1 Thermistor 10K C14.1uF C16 1500pF OMIT C15 .1uF OMIT R12 200Kohm Vs 1 -In 2 -V 3 FB 1Mo 4 Out 5 NC 6 NC7 7 GND 8U4 OPT101 C17.1uF OMIT C18 .1uF 1 2J4 HDR 2 + 1 2C1 100uF 6.3V C2 .1uF HA9 1 HA8 2 HA7 3 HA6 4 HA5 5 HA4 6 HA3 7 HA2 8 HA1 9 HA0 10 HD7 11 HD6 12 HD5 13 HD4 14 HD3 15 HD2 16 HD1 17 HD0 18 HOE 19 HREG 20 HIORD 21 HIOWR 22 HCE1 23 GND 24 3.3V 25 HIREQ 26 ALL_PE 27 GND_ 28 HWE 29 RESET 30J1 Conn Top contact C3 .1uF Vout 5 Enable 3 GND 2 Ref 1 Vin 4U6 LM4120AIM5-1.8V RE2/CS 64 RF1/AN6/C2OUT 17 RF0/AN5 18 AVdd 19 AVss 20 RA3/AN3/Vref+ 21 RA2/AN2/Vref22 RA1/AN1 23 RA0/AN0 24 Vss2 25 Vdd2 26 RA5/AN4/LVDIN 27 RA4/TOCKI 28 RC1/T1OSO/T13CLK 29 RC0/T1OSO/T13CLK 30 RC6/TX1/CK1 31 RC7/RX1/CT1 32 RE3 63 RE4 62 RE5 61 RE6 60 RE7/CCP2 59 RD0/PSP0 58 Vdd5 57 Vss5 56 RD1/PSP1 55 RD2/PSP2 54 RD3/PSP3 53 RD4/PSP4 52 RD5/PSP5 51 RD6/PSP6 50 RD7/PSP7 49 RE1/WR 1 RE0/RD 2 RG0/CCP3 3 RG1/TX2/CK2 4 RG2/RX2/DT2 5 RG3/CCP4 6 MCLR/Vpp 7 RG4/CCP5 8 Vss 9 Vdd 10 RF7/SS 11 RF6/AN11 12 RF5/AN10/CVref 13 RF4/AN9 14 RF3/AN8 15 RF2/AN7/C1OUT 16 RB0/INT0 48 RB1/INT1 47 RB2/INT2 46 RB3/INT3 45 RB4/KB10 44 RB5/KBL1/PGM 43 RB6/KBL2/PGC 42 Vss3 41 OSC2/CLK0/RA6 40 OSC1/CLK1 39 Vdd3 38 RB7/KB13/PGD 37 RC5/SDO 36 RC4/SDI/SDA 35 RC3/SCK/SCL 34 RC2/CCP1 33U1 PIC18LF6620 R310.0K 1 2S1 Pushbutton 1 2 3D1 GRN R2 1.2K 1 2 3 4 5 6 7 8 9 10J5 Conn R4 10.0K 1 2Y1 19.6608MHz R15 10.0K 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20J6 Conn BOT contact 5V 5V 5V 5V 5V 5V 5V 5V 5V 5V 5V 5V 5V 5V 5V 5V 5V 5V HE1 REG C14 Not assemble becauseof conflict with J5 LED/Switch Figure A.1 Sensor Board Electrical Schematic