UHF RFID Sensor Tag for Tire Monitoring THESIS

Transcription

1 UHF RFID Sensor Tag for Tire Monitoring THESIS Presented in Partial Fulfillment of the Requirements for the Degree Master of Science in the Graduate School of The Ohio State University By Navtej Singh Saini Graduate Program in Electrical and Computer Science The Ohio State University 2016 Master's Examination Committee: Dr. Robert J. Burkholder, Advisor Dr. John L. Volakis

2 Copyrighted by Navtej Singh Saini 2016

3 Abstract RFID (Radio Frequency Identification) technology finds application in different sectors ranging from item level identification & handling to smart tracking in the healthcare industry. One very useful application is placing passive RFID tags inside tires to obtain critical information about tire condition which will ultimately result in improved road safety and vehicular performance. The challenge is that passive RFID technology capable of smart sensing and data logging while being self-powered from local energy harvesting sources is non-existent. This research presents the design of a reliable, in-tire passive UHF RFID tag structure capable of smart sensing, processing and data logging over time. A proof-of-concept application for counting tire revolutions using a prototype RFID tag is demonstrated and one possible approach to self-powering the tag is presented. The focus of this thesis is the development of the Tag IC and not the tag antenna. The design of the tag antenna for reliable in tire operation was done by Shuai Shao, working on the same research problem. The antenna design has been summarized in appendix A. In this thesis, the existing RFID IC technology is first reviewed and tested for in tire data sensing and logging using multiple sensors. After coming to the conclusion that existing RFID ICs lack efficient data processing and logging abilities, a novel tag circuit is developed by integrating a commercial RFID IC with an onboard microcontroller. A ii

4 prototype battery operated tag board, is then developed and tested to demonstrate a proof of concept for in-tire data sensing and logging. by counting tire revolutions. This achievement is followed by miniaturizing the tag circuit and improving its sensing and logging functionality. Finally, an approach to self-power the tag using piezo sensors has been presented. iii

5 Acknowledgments First and foremost, I would like to thank my advisor Dr. Robert J. Burkholder and Dr. John Volakis for believing in me and giving me the opportunity to work on this interesting and challenging problem. Without their vision, guidance and continuous support, this project would not have been possible. Their unending encouragement, novel ideas, feedback and occasional scolding molded me from a student into a researcher. I will always look up to them. I would like to express my sincere gratitude to Bridgestone Americas for sponsoring this project and helping with conducting several drum tests for this project. I would also like to extend my gratitude to Terence Wei and Tom Sams from Bridgestone for their continuous feedback and invaluable support during the course of the project. I am extremely grateful to the ElectroScience Laboratory and its entire staff for providing a conducive work environment. I am also thankful to all my fellow researchers at the ElectroScience Laboratory for sharing their knowledge and providing feedback at critical moments during the course of the project. iv

6 I would also like to take this opportunity to thank the Ohio State University s Student Safety transport service and its staff without which I would not have been able to commute safely to lab for late night experiments. I will forever be in debt to my family for their continuous support and encouragement throughout my life. Without them nothing would have been possible. I would also like to thank Anna and all my friends for their constant support.. v

7 Vita July 2010 July B.Tech. Electrical and Electronics VIT University, Vellore, India. Aug 2014 Dec M.S. Electrical and Computer Engineering, The Ohio State University. Jan 2015 Dec Graduate Research Associate, The ElectroScience Laboratory, The Ohio State University Field of Study Major Field: Electrical and Computer Engineering vi

8 Table of Contents Abstract... ii Acknowledgments... iv Vita... vi List of Tables... xi List of Figures... xii Chapter 1: Introduction Motivation Concept of Operation Challenges... 5 Chapter 2: Literature Review... 7 Chapter 3: Theory RFID Overview Classification Components Operation: Passive UHF RFID vii

9 3.1.4 Application RFID for Data Sensing & Logging RFID for Tires Chapter 4: RFID Tag IC Development and Testing Basic Theory: Desired Tag Desired Tag Functionality Desired Tag Features RFID IC for In-Tire Operation RFID IC Technology Review and Selection SL900a RFID IC Main Features of the SL900a Tag IC: Testing SL900a IC Conclusions from Testing Chapter 5: Refining Tag Circuit Addition of Microcontroller (MCU) Selecting the Microcontroller Interfacing the Microcontroller and the RFID IC SPI Read Operation SPI Write Operation viii

10 5.3 Refined Tag Circuit Operation Breadboard Setup Development of Prototype Board Testing the Prototype Board Proof of Concept Test 1: Counting Bicycle Tire Revolutions Test 2: Counting Automotive Tire Revolutions Chapter 6: Miniaturization Board Design PCB Design Circuit Schematic Board Layout Board Fabrication Testing the Board The Hello World LED Test Revolution Counting for Varying Tire Load and Speed Conclusions Chapter 7: Self-Powered Tag Operation Energy Harvesting ix

11 7.1.1 Sources and Harvesters Piezo Sensors as Harvesters Tag Power Consumption Conclusion Chapter 8: Conclusions and Future Work References Appendix A: Tag Antenna Design Summary Appendix B: List of Abbreviations Appendix C: Electrical Model of the Piezo Sensor x

12 List of Tables Table 1: RFID frequency classification Table 2: Comparison of advanced RFID ICs Table 3: Settings for log test Table 4: Settings for log test Table 5: Settings for log test Table 6: Settings for SPI connection Table 7: Test conditions and results Table 8: Speed and load conditions for drum tests Table 9: Different values of tire speeds and loads for drum tests Table 10: effect of voltage & frequency on power consumption Table 11: Read range tests Table 12: Threshold test for the designed antenna xi

13 List of Figures Figure 1: RFID for tires... 2 Figure 2: Embedding RFID tags inside tire... 4 Figure 3: offloading tag using RFID reader... 5 Figure 4: Real time monitoring with tag on vehicle... 5 Figure 5: A RFID system Figure 6: Types of RFID readers Figure 7: A typical passive UHF RFID tag Figure 8: RFID tag classification Figure 9: Range of RFID applications Figure 10: A basic passive UHF tag IC Figure 11: Advanced RFID tag IC Figure 12: SL900a RFID IC Figure 13: SL900a block diagram Figure 14: AMS Radon RFID reader Figure 15: SL900a test board RFID tag Figure 16: Test setup Figure 17: Tag access process Figure 18: Reading tag ID xii

14 Figure 19: Real time temperature sensing Figure 20: Pin diagram and applied input signal Figure 21: External sensor readings Figure 22: SL900a pin connection for external sensing Figure 23: External sensor 1 input Figure 24: External sensor 2 input Figure 25: Readings from multiple external sensors Figure 26: Reading logged data Figure 27: Logged data inside SL900a user EEPROM Figure 28: Pin connections for log test Figure 29: Waveforms for threshold based data logging Figure 30: Connections for multi-sensing operation using SL900a Figure 31: EEPROM content with logged data from multiple sensors Figure 32: Block diagram of refined tag circuit (battery not shown) Figure 33: AVR ATMEGA328P as tag Microcontroller Figure 34: Interfacing MCU & RF chip over SPI Figure 35: Timing diagram for SPI read operation Figure 36: Timing diagram for SPI write operation Figure 37: Block diagram of tag circuit Figure 38: Lab test setup Figure 39: Prototype tag board Figure 40: Threshold detection and temperature sensing using prototype tag board xiii

15 Figure 41: Test setup and process for bicycle tire rev counting Figure 42: Test setup for automotive tire rev counting Figure 43: Process flow for rev count operation Figure 44: Read range for reading tag in tire Figure 45: Logged rev count inside SL900a EEPROM Figure 46: Schematic of tag circuit Figure 47: Board layout Figure 48: Final tag board next a US quarter Figure 49: Designed tag antenna for tires Figure 50: Programming the tag board Figure 51: Working tag board running the LED blink code Figure 52: Setup for PVDF piezo sensors inside a truck tire Figure 53: Process for regenerating tire test conditions for testing tag board Figure 54: Regenerated voltage profile for 20Kph, 5 KN load Figure 55: Regenerated voltage profile for 20Kph, 15 KN load Figure 56: Regenerated voltage profile for 80Kph, 15 KN load Figure 57: Modified tag to incorporate an energy harvesting block Figure 58: Energy sources to power the RFID tag circuit Figure 59: Arrangement of PVDF sensors inside tire Figure 60: Tire rim with external leads to connect to PVDF placed inside the tire Figure 61: Data collection setup for drum tests Figure 62: PVDF voltage profile for 5KN load and varying speeds xiv

16 Figure 63: PVDF voltage profile for 80Kph and varying loads Figure 64: PVDF open circuit waveform for 80Kph, 15KN load Figure 65: Post rectified PVDF open circuit voltage approximated to a pulsed signal Figure 66: RC circuit with square pulsed input Figure 67: Flexible e-fiber RFID antenna in polymer coating Figure 68: RFID antenna for tires Figure 69: S11 between antenna and RFID chip for different dielectrics Figure 70: Setup for read range test and tag placement on tire Figure 71: Designed e-fiber tag embedded in tire Figure 72: Electrical equivalent circuit of piezo sensor Figure 73: Piezo sensor connected to a load in series Figure 74: RC circuit for PVDF analysis Figure 75: PVDF sensor interfaced to a regulating circuit xv

17 Chapter 1: Introduction RFID (Radio Frequency Identification) technology has significantly improved in recent years. It is being implemented in various aspects of day to day life from inventory tracking & item level identification to storing and linking information via smart cards. RFID is also being used for biomedical applications in the healthcare sector [1]. The potential to expand this technology has prompted much research towards the development of novel applications aimed beyond standard inventory tracking and item level identification. Foreseeing the rise of the Internet of Things, RFID technology will have an even bigger role to play in the future where objects tied to one another over an intricate network will be able to sense and process multiple data signals before taking an action [2], [3]. With all the developments in antenna design & RF chip technology, RFID can be used to realize standalone applications and is capable communicating more than just returning an identification number. One such specific application is placement of RFID tags inside tires to monitor tire health & condition. Interfacing RFID technology with sensors inside a tire will help in retrieval of important information such as tire pressure, temperature, stress etc. This data can be communicated from within a tire back to the reader not only in real time but can be logged to monitor tire health over extended periods of time. 1

18 Figure 1: RFID for tires However the task of guaranteeing a reliable operation from a UHF RFID unit inside a tire, is a very challenging task. The extreme conditions inside a tire along with its nonuniform structural composition make it very hard for an antenna along with its electronics to operate reliably. 1.1 Motivation Road Safety & Improved Vehicular Performance Passive multifunctional RFID tags inside tires will help in retrieving critical information about tire condition which will ultimately result in improved road safety and vehicular performance. 2

19 The ability of a passive tag to track the number of miles and relate it to tire wear and degradation will help the tire industry in generating proven reliable statistics to determine tire life. Other crucial factors signifying tire health could provide the vehicle owner an early heads up to replace the tire. Other advantages involve uploading the data gathered by the reader to the cloud. In this way the owner can be constantly updated on the tire condition, miles driven etc. through one simple passive tag. Improving Existing RFID Tag Technology Present RFID applications implement the sensing mechanism via changes measured in a passive tag s scattering properties [4], [5], while others incorporate real time sensors. The former being prone to unwanted disturbances, many RFID manufacturers have rolled out advanced RFID chips with an on-chip sensor, ADC, and dedicated memory for storing relevant data. But the problem with these chips is the lack of efficient data processing & logging. These drawbacks greatly limit the range of applications these tags can be used in. Self-powered RFID sensor technology being developed in this research will be adaptable to other applications and have an even broader impact on the rapidly growing RFID and wireless sensor market. An autonomous RFID based smart sensor with energy harvesting and data-logging capabilities, overcomes the power limitations of sensors and extends the functionality of passive RFID wireless communications. It is expected that once the 3

20 groundwork is laid in this project and a working prototype is demonstrated, many other RFID based sensor applications will be identified. 1.2 Concept of Operation The RFID tag must be compact and durable to be embedded inside a tire. Once embedded, the tag will operate as follows: a. During normal operation, sensor data will be processed and logged on the tag chip while the tire is in motion. Energy is harvested mechanically from tire movement/deformation. Revolutions, temperature, pressure, strain, etc., logged while chip is powered up. Figure 2: Embedding RFID tags inside tire b. Data is offloaded using an RFID reader while stationary. Energy is harvested from the RF signal. Simple instructions can be given to reset or refunction the chip. 4

21 Figure 3: offloading tag using RFID reader Note: Data can be read and logged using a RFID Reader as an additional energy source if the RFID readers are mounted inside the vehicle near the tire as shown in figure 4 Figure 4: Real time monitoring with tag on vehicle 1.3 Challenges Developing a RFID tag for an in-tire autonomous wireless data sensing and logging application is a very challenging task [9]. The tire s structural composition and extreme running conditions pose serious challenges to a RFID tag. Some of the critical design challenges are as follows: 5

22 a. Tire rubber attenuates the RF signal. b. Tires are often stacked or mounted on dual-wheel axles making it harder to read the tag. c. Truck tires have a steel radial ply from rim-to-rim. d. Construction and material properties are variable: Tires degrade with age and wear. The above issues are concerned with tag placement and wireless communication. The task of developing a tag circuit capable of multisensing and data logging faces additional challenges: e. Lack of commercial off the shelf RFID chips capable of data sensing and logging. f. Energy harvesting to power in tire tag electronics while sustaining reliable operation. g. Minimizing overall tag power consumption while performing sensing, processing, logging and wireless data transmission operations. 6

23 Chapter 2: Literature Review RFID technology has been in use since several decades. Technology similar to RFID and considered RFID s predecessor was in use during The Second World War. Since then RFID has come a long way. Today, RFID finds applications in a wide range of fields. Despite its widespread use, RFID has mainly been restricted to the identification and tracking side of applications. Developing self-powered RFID tags for in tire data sensing and logging application requires an in depth understanding of the effects of the tire environment on tag operation. Present state of RFID sensor technology has to be thoroughly reviewed to determine the possibilities and limitations. Before investing time, effort and resources into developing the tag, it is important to do a feasibility study. The first step is reviewing the state of existing RFID sensor technology. RFID has undergone a lot of development in both reader and tag technology. RFID ICs in [6], [28] and [29] are some of the commercial ICs offering several advanced features including an on chip memory, sensor support, and external peripheral connectivity. These ICs can be integrated with a suitable antenna to build a tag capable of data sensing & logging. The next step in the feasibility study is an analysis of the operation of a tag antenna inside tires. Commercial UHF RFID tags do not perform well inside tires. The study by T.Wei 7

24 et al [16], reports that commercial tags for tires do not operate well inside tires. The read range of such tags is considerably affected due the variations in tire structure and size. The work by S.Shao et al [9], [17] and [18], presents a UHF RFID dipole tag antenna specifically designed for reliable in tire operation. The designed tag [9], is a flexible broadband antenna capable of delivering reliable performance over a range of dielectric media. The focus of this thesis is the development of a UHF RFID tag IC for in-tire data sensing and logging. After reviewing one of the most advanced commercial RFID IC: SL900a by AMS technologies [6], it was concluded that despite its advanced features, it lacks the necessary data processing and logging abilities to operate reliably in remote conditions. The work presented in [27] brilliantly demonstrates a passive, single chip UHF RFID tag for a multi-sensing application. But the features of this tag are limited and satisfy the minimal needs of the listed application. Further literature was reviewed to explore the design of UHF RFID sensor tag electronics. In [30], RAMSES (RFID augmented module for smart environmental sensing) incorporates a computationally capable energy harvesting RFID sensor tag with a long read range but lacks the ability to flexibly log data on tag for an extended period of time, in a standalone environment. Moreover the tag logs the readable data in the EPC memory which is a limited memory space. SPARTACUS UHF RFID tag [31], presents the design of a passive sensor-logger tag. This tag combines the RFID IC with a 8

25 microcontroller and sensors on a neatly interfaced tag board. The tag capable of data sensing is powered using the reader s RF field energy by using a two antenna scheme. The waste mechanical energy is good source for powering in tire tag electronics. The work presented in [32] reviews the piezo electric energy harvesting (PEH) from piezo electric materials. Their work describes the advantages, disadvantages and the problems of PEH. The study by A. Majeed [33] talks about using PEH to power Micro Electro Mechanical Systems (MEMS). It covers the design of a complete energy harvesting system and talks about applications of a MEMS system powered by PEH. J.Liang and W.H. Liao [25] present an impedance matching analysis for improving the piezoelectric energy harvesting systems. The piezo structure is modelled both as a mechanical and an electrical system to understand its behavior as an energy harvester. The model presents the impedance of a piezo harvester and defines the approach for interfacing PEH s to a harvesting circuit with a good match. 9

26 Chapter 3: Theory 3.1 RFID Overview RFID stands for Radio Frequency Identification. RFID is a type of wireless automated data collection system. Unlike barcode scanning, RFID is a non-line of sight technology. The basic operation of a RFID system is shown in figure 5. Figure 5: A RFID system Any RFID system comprises of three components: a. RFID Reader b. RFID Antenna 10

27 c. RFID Tag - Tag Antenna - Tag IC Classification On the basis of frequency, a RFID system can be classified as: a. Low Frequency (135 KHz) b. High Frequency (13.56 MHz) c. Ultra-high Frequency ( MHz) d. Microwave (2.45 GHz to 5.8 GHz) Table 1 lists the classification of RFID based on frequency of operation. Table 1: RFID frequency classification Frequency range LF (135 KHz) HF (13.56 MHz) UHF ( MHz) Microwave (2.45 GHz 5.8 GHz) Typical Read Range (Passive Tags) Tag Power Source Shortest 1 12 Generally Passive tags powered through inductive coupling Short 2-24 Generally passive powered via inductive or capacitive coupling Medium 1-10 Passive, Semipassive or active tags powered via RF field or an on tag battery Longest 1-15 Active tags with integral battery or passive tags powered using RF field or capacitive storage Data Rate Slower Moderate Fast Ability to read metal or wet surfaces Better Moderate Poor Faster Worse 11

28 3.1.2 Components The Reader The RFID Reader is a RF system responsible for transmitting a modulated information signal at the carrier frequency and demodulating the received signal in accordance with the regulation standards. There are several regulation standards in place for RFID. Some of these are the ISO, IEC, ASTM and EPC Global standards. [10] & [26] briefly describe the various RFID standards in place. The most popular one is the EPCGlobal standard [11] and is being pushed worldwide to be the single standard for RFID. There is wide range of commercially available RFID readers. To begin with, the readers are separated into different classes based on the frequency of operation. These include the LF, HF, UHF and microwave frequency as mentioned in table 1. The RFID reader system is specifically designed for each operating frequency. For the same frequency of operation, the RFID readers come in range of shapes and sizes. Figure 6 shows the classification for RFID Readers. 12

29 Figure 6: Types of RFID readers Reader Antenna The Antenna is a very crucial part of any RFID system. A poorly designed or matched antenna can considerably degrade the performance of a RFID system, despite having the best reader and tag. Just like the RFID Reader, the antenna is designed for operation at a specific frequency. Even for the same operating frequency, the type and the size of the antenna can vary considerably based on the application of interest. The primary goal of the antenna is to illuminate the surroundings in accordance with the application requirements. RFID Tags The final component completing a RFID system is a tag. In fact there is not just one tag, there are several tags. A tag is the end point in the system that carries the information the reader seeks. Any tag has two major components: The antenna and the Integrated Circuit 13

30 (IC) as shown in figure 7. There may or may not be an on tag battery depending upon the type of RFID tag. Figure 7: A typical passive UHF RFID tag The RFID tag classification is depicted in figure 8. RFID tags come in wide range of shapes and sizes determined by the antenna. The antenna itself is determined by the application requirements. Just like the Reader antenna, the tag antenna is a critical part of the tag. A poorly designed or matched antenna connected to the best integrated circuit is not of much use. The final tag along with the antenna and IC is packaged neatly onto a substrate to impart robustness to the tag. 14

31 Figure 8: RFID tag classification Operation: Passive UHF RFID Ultra High Frequency (UHF) RFID operates in the 868 MHz 915 MHZ frequency band. In the US, the UHF band ranges from MHz with a center frequency of 915 MHz Any UHF RFID system has 3 components: a UHF RFID reader, one or more UHF Reader Antennas and UHF Tags. The Reader generates the UHF carrier signal in the UHF band and the modulator on the reader modulates the carrier signal using the ASK or PSK modulation schemes. The modulated information, the command sequence and rules governing the RFID modulation have been thoroughly covered in the EPCGlobal specification document [11]. The modulated signal is then transmitted via a reader antenna either embedded inside the reader or attached externally to it. The size and shape of the reader antenna depends upon the application. 15

32 The FCC limits the transmitted signal power to a maximum of 36dBm or 4W. This transmitted signal propagates through air or other media before reaching the RFID tag. The propagation losses significantly reduce the power available to the tag IC. The power received at the tag IC must be higher than its sensitivity value. The tag IC down converts and demodulates the received signal. The IC interprets the received command defined by the EPC protocol and responds by load modulating the original signal. This reflected or backscattered signal is picked up by the Reader antenna and demodulated. The signal strength of the backscattered signal picked by the reader antenna must be higher than the maximum sensitivity level of the RFID reader. The entire protocol related to timing, Collison, modulation etc. is covered in the EPC specification for UHF RFID Application RFID has been in use since several decades [12] and has primarily been used for object identification and tracking. Mario W. Cardullo received the first U.S. patent for an active RFID tag with rewritable memory on January 23, From then on numerous inventions circling around RFID have been made. Over the years, RFID technology has seen tremendous growth. They can be seen in use in day to day applications. Agriculture, livestock, retail, household, navigation systems, clothing, defense etc. employ RFID in one way or the other (figure 9). 16

33 Figure 9: Range of RFID applications 3.2 RFID for Data Sensing & Logging RFID has primarily been used for object tracking and identification. However with a better understanding of antennas and rapid improvements in IC design and fabrication technologies, RFID is now capable of a lot more than basic identification [13]. RFID sensing has been generating a lot of excitement. There are several reasons for the same. Power, cost and size are key considerations for any commercial product. Passive RFID tags harvest power, are cheaper and fabricating a RFID tag is not a cumbersome process. Semi-passive tags do carry a battery but consume little power since the mode of wireless transmission is still backscattered modulation. 17

34 There are two components that can be modified to realize sensing applications using RFID. Since any RFID tag has two components: the antenna and the IC, the only way to incorporate data sensing is to modify either or both of these components. There is a lot of published research on RFID antennas being used for data sensing [4] and [5]. This involves sensing the change in tag antenna properties such as gain, phase, signal strength etc. and relating it to a change in temperature, humidity or another parameter. This methodology is not accurate as several outside factors end up skewing the results. The second method involves the tag IC. Recent developments in the RFID IC technology has seen incorporation of higher on chip memory, sensors, ADC and external communication peripherals. This method of sensing highly accurate since the IC interacts with actual data sensors. However ICs with additional features lead to increased size, cost and power requirements. 3.3 RFID for Tires Tire companies have to track tires during the manufacturing and distribution process. Keeping an inventory of tires is crucial since vehicle tires are costly. One method is using the barcode technology. Replacing bar code with RFID tags would not only speed up the tracking process but drive down the labor costs. RFID has several advantages over barcode technology [14]. Another application of embedding RFID tags inside tires is automotive data sensing. Current tire sensing solutions involve deployment of individual sensors that are bulky and 18

35 expensive. An ideal solution is to deploy compact multisensing and data-logging wireless modules inside tires. However deployment of RFID tags inside tires faces several challenges as mentioned earlier. The research in [15] highlights the detuning effects of tire environment on commercial RFID tags. Generic RFID tags do not work well in tires. RFID tags for in tire applications need to be designed keeping the detuning effects in mind. Patch Rubber Company sells commercial RFID tags for tires with the name Speedy Core. However the read range of these tags when placed in a truck tire with steel belts and steel body plies, read by a RFID reader transmitting an RF signal at 35dBm is less than a meter [16]. This range is too low to be feasible for any practical tire application using RFID. Research work by S. Shao et al [17], [18] presents a RFID tag specifically developed to address in tire tag detuning effects. The antenna in [17] is a broadband end loaded meander line dipole antenna designed to operate over a range of dielectric media. Moreover the tag being embroidered using conductive e-fiber strands is robust and flexible. This helps to withstand the tire curing and retreading process. The tag design has been summarized in appendix A. 19

36 Chapter 4: RFID Tag IC Development and Testing Any RFID tag has two major components: A tag antenna and a tag IC. This project report is focused on the development of the tag circuit or tag electronics capable of multi data sensing and logging. 4.1 Basic Theory: Any RFID tag has two components: a. The antenna b. The Integrated Circuit (IC). The antenna is responsible for receiving the RF signal and the IC demodulates the information on the carrier signal, executes the command sent by the reader and returns the information back to the reader. All of this is done in compliance with the RFID EPC Protocols. RFID ICs can be classified as follows: Based upon the frequency of operation, RFID is classified into three types: 20

37 a. Low Frequency (LF) b. High Frequency (HF) c. Ultra High Frequency (UHF) Based upon the mode of data transmission, RFID is classified into three types: a. Passive b. Semi Passive c. Active This chapter will focus on the RFID IC technology. The goal is to develop a UHF RFID tag capable of multi-sensing, data processing and logging. Hence the first step is to select the right RFID chip. 4.2 Desired Tag RFID is not a one size fits all technology. Before diving into the process of selecting the right RFID IC, it is important to understand the application and its specific requirements Desired Tag Functionality Data Sensing: Ability to sense changes in the surrounding environment. 21

38 Data Processing: The tag must be able to process the raw sensor data into meaningful information. For in-situ tire operation, the following information is considered as meaningful: Average, minimum and maximum temperature values over time Average, minimum and maximum stress & strain values over time Number of miles travelled Average, minimum and maximum Pressure over time Data Logging: The tag IC must be able to store the processed sensor information in an efficient manner. The logging feature is critical for remote operation of the RFID tag in the absence of a reader. The end goal is to use a single tag circuit to generate a wide range of information on tire health and condition. Existent tire sensing solutions generate information related to one specific parameter. For example tire pressure monitoring systems for tires solely monitor tire pressure and are big in size carrying an onboard battery Desired Tag Features Self-Powered Having a battery embedded inside a tire is not a feasible option for two reasons. One, the extreme in-tire conditions can lead to undesired behavior from the batteries such as chemical leakage and fires. The other reason is the ordeal of replacing a 22

39 battery once it dies out. Hence some sort of self-powered mechanism is required to power the tag. Small Form Factor The entire tag system has to fit inside the tire and in no way affect the tire s orientation and natural behavior. Robust and flexible Extreme conditions inside a tire demand that the tag be able to handle high g forces and high temperatures in excess of 100 degree Celsius. The tag must under all conditions maintain its integrity while performing the desired operations without fail. 4.3 RFID IC for In-Tire Operation The focus for in-tire multisensing application will be on UHF RFID ICs operating in the US MHz band. The reason for choosing UHF is longer range and faster data rates. The tag IC must be passive or semi-passive to minimize power requirements. Under the UHF classification, there is a range of available RFID ICs capable of a multitude of operations. Almost all UHF RFID ICs are based on the EPC Class 1 Gen2 Standard [11] which implies that these ICs respond to a pre-defined set of commands sent by the reader. Based on the EPC Class1 Gen 2 standard and the structure of RFID ICs, the application of most RFID ICs is limited to object identification. If the tag IC does support sensing, there are limitations in terms of data processing, memory space and data logging. 23

40 Simplest RFID ICs have benefits being smaller and cheaper. However, they are mainly used for object tagging for identification purposes. Figure 10 below highlights the simplest RFID IC on a UHF tag. Figure 10: A basic passive UHF tag IC However, with the growing interest in RFID, there have been developments in the tag IC technology. Newer and advanced commercial ICs as in [6],[7] and [8] have internal sensors, on chip ADC s, support for external sensors, increased memory and advanced data logging options. These ICs still adhere to the EPC Class1 Gen 2 standard but come with additional features and support. The drawback of these ICs is increased size, cost and power requirements to support the enhanced functionality. 24

41 Figure 11: Advanced RFID tag IC 4.4 RFID IC Technology Review and Selection For the in tire application of remote multisensing, data processing and logging, an RFID IC with the following features is required: a. On chip sensing and data logging b. Data processing c. Significant memory space d. Small size e. Low power requirements f. Robust and durable to handle harsh in-tire environment 25

42 There are several commercially available off the shelf RFID ICs, offering the aforementioned features desired for in-tire operation. The review and selection process is listed in the table below: Table 2: Comparison of advanced RFID ICs IC On chip ADC Data Logging External Sensors Memory External Interface NXP SL3ICS3001 No Yes Yes (Digital) 2Kb Yes : I2C Farsens Andy 100 No Yes Yes (Digital) Less than 1 Kb Yes: SPI Monza x 2K No Yes Yes (Digital) 2K Yes : I2C Monza x 8K No Yes Yes (Digital) 8K Yes: I2C RAMTRON WM72016 No Yes Yes (Digital) 16Kb FRAM Yes: SPI SL900a Yes (10 bit ADC) Yes (Multi-mode) Yes (Analog & Digital) 9Kb Yes: SPI Table 2 lists the key performance specifications for choosing the tag IC. There are though several other considerations involved in selecting a tag IC, few of which are the read and write sensitivities, up link and down link data rates, modulation schemes etc. However the above listed specs were key points in selecting the tag for in-tire multisensing application. 26

43 Based on the above review process, SL900a by AMS [6] was found to be the most suitable tag for in-tire sensing application. Unlike all the commercially available tags, the SL900a has an internal temperature sensor and an on chip 10 bit ADC. Having an on-chip ADC enables the chip to interface with analog sensors unlike other chips which interface only with digital sensors. In terms of data storage, the SL900a offers 9Kb of user memory with multiple data logging modes. The SL900a can connect to an external processor or an external digital sensor using the SPI interface. 4.5 SL900a RFID IC Figure 12: SL900a RFID IC 27

44 4.5.1 Main Features of the SL900a Tag IC: Operation modes a. Frequency: MHz b. Fully Passive mode and BAP mode of operation Compatibility a. Worldwide EPC Compliant b. EPC Class 1 c. EPC Class 3 Data Sensing a. On chip 10 bit ADC b. Internal temperature sensors c. 2 external sensors d. Both analog and digital sensors e. Multiple sensing modes including interrupts. Data Logging a. Multiple data logging modes b. On chip 9Kb EEPROM Power: a. Fully Passive Mode: Energy Harvesting from Reader field b. BAP Mode: 1.5 V 3 V 28

45 The SL900a is packaged as a QFN 16 type IC. Block Diagram of SL900a Figure 13: SL900a block diagram Testing SL900a IC Equipment Required a. UHF RFID Reader Commercial RFID tags interact with the reader in accordance with the EPC Class1 Gen2 UHF Protocol. This protocol covers the basic commands: acknowledge, query, inventory 29

46 etc. of RFID tags. The SL900a IC offers advanced features which require responding to a specific set of commands known as the cool log commands. These advanced operations cannot be performed using the standard EPC commands. Only the RFID readers supporting the cool log command set can be used to access the advanced features on these chips. For this reason, the Radon RFID reader by AMS [19] has been used to access the advanced features offered by the SL900a IC. Figure 14: AMS Radon RFID reader AMS Radon Reader key features: - 32 dbm output power - In-circuit test points for digital, analog and RF signals - Two antenna output ports - USB/UART interface 30

47 - Compact size - Fixed / Battery Type Reader b. UHF RFID Tag (SL900a Tag IC matched to an antenna) For testing purposes, the RFID chip has been used on a test board (figure 15) interfaced to a pre-built dipole antenna and 3V Li-ion CR-2032 coin cell as shown in the figure below. The test board has an inductor connected between the chip and the antenna for matching. The testboard tag is a battery assisted passive (BAP) tag. The battery however can be removed. Figure 15: SL900a test board RFID tag The SL900a IC offers several data sensing and logging features. The SL900a was tested keeping the tire application in mind. The IC was tested for concurrent data sensing and logging from multiple sensors. Different calibration settings, sensing and logging modes were used while testing the RFID Chip. 31

48 Testing Environment & Setup All the tests were performed in the RFID lab at The ElectroScience Laboratory at The Ohio State University. AMS Radon reader is attached a single patch antenna fixed on a movable setup. The tag is mounted as shown below. Figure 16: Test setup Figure 17: Tag access process 32

49 Test 1: Accessing the RFID Tag All RFID tags need to be in the reader s RFID field to be accessed by the reader. Each RF chip has a sensitivity level or a minimum threshold level of detection. The SL900a has a detection threshold of -15dBm. This means that the RF chip will be activated only as long as the RF power from the reader received by the RF chip after transmission, polarization and mismatch losses is higher than -15dBm. Activating the chip itself is not enough. The RFID operation is considered complete only when the RF chip responds to the reader s query/command. The backward link (Tag to reader) involves backscattering of the load modulated RF signal. Backscattering involves further drop in the dbm level of the power signal. Hence a RFID reader with a high sensitivity is desirable to detect the backscattered signal. 33

50 Figure 18: Reading tag ID Test 2: Real Time Data Sensing The SL900a RFID chip has an on chip temperature sensor and a 10 bit ADC. The on chip ADC allows the IC to communicate directly with analog sensors. External digital sensors can also be connected using the SPI interface on the chip. The IC also allows connection of up to two external sensors. The internal architecture of the IC has been designed to support a range of sensors on these external pins. These sensors include: 34

51 a. Linear resistive sensors b. Linear conductive sensors c. DC voltage source sensing d. Capacitive sensors e. AC resistive sensors f. DC current source sensing g. Optical sensors The chip can be programmed and calibrated to operate with specific sensors in different modes for a given application. Sensing Application 1: Internal Temperature sensing Description: This test demonstrates the real time temperature sensing using the ICs on chip temperature sensor. Before using the IC for this purpose, it is necessary to send several commands using the reader to select the temperature sensor and set the ADC voltage reference and calibration bits. Once the settings are correctly programmed onto the chip, a simple conversion formula can be used to obtain the temperature values. T( C) = code 0.18 C 89.3 C (4.1) 35

52 The digital code is a decimal digital number generated by the on chip ADC. The above equation holds true for a given calibration setting on the RFID chip. The general equation is: T( C) = Vo2(mV).(code+1024) code.vo1(mv) (4.2) Vo1 and Vo2 are adjustable voltage range values for the on chip ADC. Vo2 is the reference voltage of the ADC. Vo1 and Vo2 can be varied to operate the ADC in a specific voltage range. The data registers are calibrated as per the data sheet [6]. Setup: The test setup is the same as shown in figure 17. After programming the IC with appropriate settings, another RFID command is sent to the IC to read the real time temperature value as shown in figure 19, right hand side. Figure 19: Real time temperature sensing 36

53 Sensing Application 2: External sensor testing: Description: This test demonstrates the IC s external sensor capability. For this test, an arbitrary DC voltage source connected between the EXT1 and ground pins has been used to simulate an external sensor. Setup: The test setup is the same as shown in figures 16 and 17. The pin connections are shown in figure 20. The DC voltage waveform fed into the IC is also shown in figure 20. The real time sensed values are shown in figure 21. Figure 20: Pin diagram and applied input signal 37

54 Results: The RFID reader s GUI shown in figure 21 depicts the analog DC voltages sensed by the RFID IC. The sensor value depicted is a decimal digital code generated by the 10 bit ADC. The ADC can be calibrated differently to increase/decrease the sensitivity of the sensed voltage inputs. Figure 21: External sensor readings Sensing Application 3: Multisensing: Description: This test demonstrates the multisensing abilities of the RF chip. For this test, the RFID IC has been programmed to utilize all 3 sensors (2 external & 1 internal) to simultaneously sense and return sensor data. 38

55 Setup: The lab setup is the same as shown in figures 16 & 17. The pin connections for the concurrent real time data sensing test are shown in figure 22. The calibration settings are done in accordance with the sensors attached to the IC. Figure 22: SL900a pin connection for external sensing Figures 23 & 24 show input from arbitrary DC voltage sources acting as external sensors. Figure 23: External sensor 1 input 39

56 Figure 24: External sensor 2 input Results: Figure 25 shows the real time sensor data recorded concurrently from the three sensors. The returned data for external sensor is a decimal number between 0 to This can be converted to a sensed voltage value using the following formula: Sensed value (mv) = (Vo2 Vo1) (code) + Vref (4.3) 1024 Figure 25: Readings from multiple external sensors 40

57 Test 3: Data Logging The SL900a IC offers several data logging modes: Based on number of number sensors: a. Single sensor logging b. Multi sensor data logging c. Interrupt detection data logging Based on the sensed value: a. Logging all sensed data b. Logging sensed data based on limits detection c. Logging sensed data based on interrupts detection Based on timing of logging interval: a. Delayed start of data logging b. Logging after fixed time intervals The SL900a IC was tested for all possible data logging combinations. The key ones relevant to the application are presented below: a. Single sensor data logging with predefined delay settings b. Single sensor interrupt based data logging: c. Concurrent Multisensing and data logging using external sensors 41

58 The test setup for each of these tests is the same as shown in figure 16 & figure 17. The pin connections and the calibration settings are changed in accordance with the sensor/s connected to the IC. Logging test 1: Single sensor data logging with predefined delay settings Description: For this test, real time temperature values were read from the SL900a IC on test board. The test board was kept in the Reader s field. Commands were sent to log data with a logging interval of 1 second. After 15 seconds, a stop log command was sent. IC Settings: Table 3: Settings for log test 1 Setting Type Description Sensor used Delay interval Delayed start of logging mode Logging mode Interrupt mode Limits based mode Testing interval Internal Temperature sensor 1 second No Dense (Single Sensor mode) No No 15 seconds 42

59 Results: Figure 26 shows the number of logged values in memory. Figure 26: Reading logged data As expected, 15 values were logged for a 15 second interval with a pre-defined 1 second log interval. Figure 27 shows the memory snapshot of the SL900a tag IC containing the stored temperature values as hexadecimal numbers. The figure just by looking does not make much sense. Each sensed value is a 10bit number. Each value in the figure (7c, 44, 7c, 43 etc.) is a hexadecimal 8 bit number. 10 continuous bits form a sensed value. Memory Map of the SL900a IC consisting of logged temperature values Figure 27: Logged data inside SL900a user EEPROM 43

60 Logging Test 2: Single sensor interrupt based data logging: Description: This test is used to demonstrate the ability of the RFID IC to log data for crossing a preprogrammed voltage threshold value. The IC offers 4 threshold detection points as percentage of the IC supply voltage. For this particular test the supply voltage is 3V and the set threshold percentage is 25% or 750mV. Whenever the input voltage crosses this detection point, the ADC is activated and the IC logs the value while simultaneously incrementing a counter. IC Settings Table 4: Settings for log test 2 Setting Type Sensor used Delay interval Delayed start of logging mode Logging mode Interrupt mode Limits based mode Testing interval Description Sinusoid sensor signal 0 seconds No Interrupt based Single sensor logging Yes No 15 seconds 44

61 Test Setup: The setup is the same as shown in figures 16 & 17. The pin connections are shown in figure 28. Figure 28: Pin connections for log test 2 Results: The ADC was calibrated to operate in the range from 610mV 1220 mv range and the threshold point was set as 775 mv. The waveform recorded at the oscilloscope (figure 29) shows the ADC being activated (orange spikes) at each threshold crossing. Each time the spike occurs, a data is logged into the memory. 45

62 Figure 29: Waveforms for threshold based data logging Logging Test 3: Concurrent Multisensing and data logging using external sources Description This test involves testing the SL900a IC for concurrent sensing and data-logging. Two sensors are connected to the IC as shown in figure 30. The Piezo sensor is disturbed manually by hand flicking it randomly while the photo-resistor is stimulated by shining light on it. The piezo sensor converts mechanical energy to electrical energy or force to voltage. The piezo sensor s output is fed to a full bridge rectifier circuit before being fed to the EXT2 pin of the SL900a chip. This voltage level is then detected by the ADC of the SL900a and 46

63 converted to a number between 0 and This number is then retrieved as a hexadecimal value when read by the reader. The second sensor is a photo-resistor which is basically a light dependent resistor. Variation in the light intensity produces a change in the resistance. This resistance change produces a variation in the analog voltage at the input of the EXT1 pin. This value is then mapped to a corresponding digital value using equation 4.3. Vo2 and Vo1 in equation 4.3 are the internal voltage settings that determine the voltage range of the ADC. The voltage range of the ADC is adjustable between 0 and 1.2V Settings: Table 5: Settings for log test 3 Setting Type Description Sensor used Delay interval Delayed start of logging mode Logging mode Interrupt mode Limits based mode Testing interval Piezo (PVDF ) Sensor and Photo-resistor 1 seconds No Multi-sensing dense mode No No Not defined 47

64 Setup: The test setup for the test board is the same as shown in figure 16. The circuit for the sensors interfaced to the ICs external sensor pins EXT1 and EXT2, is shown below in figure 30. Figure 30: Connections for multi-sensing operation using SL900a Results: Figure 31 shows the memory map of the SL900a tag after finishing the data sensing & logging operation. Each logged value comprises of the data from the 2 external sensors along with the timing information. Each sensor value is 2 bytes long with hexadecimal 48

65 representation. Each time stamp is (highlighted blue in figure 31) 4 bytes long making each logged value 8 bytes long. The IC can log 131 such values. Figure 31: EEPROM content with logged data from multiple sensors Conclusions from Testing Power: All the tests were done using a tag board in Battery Assisted Passive (BAP) mode. Most RFID chip operations can be performed using the RFID reader in passive mode at the expense of read range, but data logging requires the presence of a battery. There is no getting around an external power supply if the tag has to operate in the absence of a reader. 49

66 Multisensing & Data logging The multi-sensing and data logging capabilities of SL900a RFID IC have been established through various tests. The IC, with its on board ADC supports both analog and digital sensors. For in-tire sensing application, a piezo sensor, pressure sensor, temperature sensor or strain sensor can easily be interfaced to the IC. Moreover the tag has the ability to concurrently sense and log data from multiple sensors. The various logging modes offer flexibility with data storage. However the IC has drawbacks that make it impossible to realize an efficient multisensing UHF RFID remote data sensing and logging tag. Limitations of SL900a RFID tag IC Problem 1: Logged values do not convey useful information How would one reading the logged IC memory understand the data? The values logged into the IC EEPROM are nothing but voltage values collected from some sensor. These are raw voltage numbers converted to a digital code and logged without any sort of interpretation. These voltage values mean nothing unless processed. If the sensed values were to be read in real time, there would not be any issue, since the values read by the reader GUI could be processed through programs written on the PC in real time. However, for remote data sensing and logging inside a tire, these values need to be processed before being logged. 50

67 Problem 2: Memory access Even though there are several data logging modes available for the SL900a IC, there is no efficient way to access the EEPROM. It is impossible, using RFID cool log commands to program a tag working in the absence of a reader, to access a particular memory location. IC Memory is accessed in a particular order when programmed using cool log commands. Problem 3: Efficient data logging Even though the SL900a offers a considerable amount of memory (9Kb), it is not enough when data needs to be sensed at frequent intervals in the absence of a reader. Without processing and selective data logging, one would rapidly run out of the 9Kb user EEPROM space. Overcoming the Limitations One great feature of RFID chip as mentioned in section is its ability to interface with external devices using the Serial Peripheral Interface (SPI) protocol. Interfacing the RFID IC to an external microcontroller would solve the issues related to data processing and efficient logging. Moreover, an externally interfaced microcontroller would enhance the tag s overall memory and increase the number of sensors that can be interfaced. Obviously, there is a big drawback in terms of increased size and power consumption but the application requirements make it necessary. 51

68 The next chapter deals with development of tag circuit through the addition of a microcontroller. 52

69 Chapter 5: Refining Tag Circuit Addition of Microcontroller (MCU) As previously stated, the SL900a RF chip is wonderful piece of technology, but falls short of fulfilling the role of a chip able to support smart and efficient data processing and logging. To achieve this functionality from the tag, a microcontroller circuit was added to the existing tag structure. A diagram of the upgraded tag structure is shown in figure 32. Figure 32: Block diagram of refined tag circuit (battery not shown) 53

70 5.1 Selecting the Microcontroller A microcontroller is a small computer on a chip with input & output peripherals, capable of processing and storing information. There is a vast range of microcontrollers available in the market. The features offered by microcontrollers cover a wide spectrum of applications. The primary criteria for choosing a microcontroller was the rapid development of a prototype tag to demonstrate a proof of concept. For this reason, the AVR ATMEGA328P [20], was chosen as the initial microcontroller. The ATMEGA328P MCU is the same one on the Arduino Uno development board. There is a huge amount of useful literature and open source support for this processor. This makes it easy to get started and rapidly begin prototyping an application. Besides being popular with students and hobbyists, the AVR ATMEGA328P is also one of the highly rated Microcontrollers in the market in terms of instruction execution rate and power consumption levels. Figure 33: AVR ATMEGA328P as tag Microcontroller 54

71 The approach is to use the ATMEGA328P MCU to develop the initial tag circuit. Once a proof of concept has been demonstrated, more effort and time will be invested in choosing a MCU that is smaller, efficient and more specific to need. 5.2 Interfacing the Microcontroller and the RFID IC The SL900a IC has an inbuilt Serial Programming Interface (SPI) module which enables it to communicate with the outside world. Outside world implies external circuits/chips also sporting a SPI module. The ATMEGA328P microcontroller has an inbuilt SPI module. This will be used to interface it with the SL900a RFID IC. SPI is a synchronous serial communication protocol developed by Motorola [21]. One device acts as a master and controls one or more devices referred to as slaves. The level of control the master has over the slave depends upon the functionality supported by both the master and the slave. The SPI is a 3 wire communication protocol. It sometimes requires a 4th wire known as Chip Select (CS) or Serial Enable (SEN) as a control signal to start/stop the data transfer. Figure 34 shows the block diagram of the MCU & RF chip interfaced using SPI. 55

72 Figure 34: Interfacing MCU & RF chip over SPI Each device with an SPI module has its own clock frequency requirements, clock settings and data transfer rules. Hence the SPI on the ATMEGA328P and SL900a have to be carefully configured before they can communicate. The verification of the correctness of the SPI operation can be done using a logic analyzer. In the RFID IC to MCU SPI communication, the ATMEGA328P acts as the master controlling the slave chip SL900a. The range and type of operations that the MCU can perform on the RF chip over the SPI is limited by the inherent SPI functionality of the SL900a IC. SL900a offers many SPI functions as mentioned in [6] but the focus will mainly be on the read and write operations SPI Read Operation The read operation allows the microcontroller to access any memory location on the SL900a EEPROM. The read command sequence is defined in the datasheet [6] of the 56

73 SL900a IC and is depicted in figure 35. The SPI configuration settings for the read command are shown in table 6. An embedded C code was written to execute the read instruction. Figure 35: Timing diagram for SPI read operation Table 6: Settings for SPI connection Setting SPI Clock Set Value 125 KHz SPI mode CPOL = 0 CPHA = 1 Serial Enable/Chip select Master Slave Active HIGH ATMEGA328P MCU SL900a RFID IC 57

74 5.2.2 SPI Write Operation Similar to the read operation, the MCU can write data to any memory location on the SL900a IC using the write command.. The SPI configuration for the write operation is the same as for the read operation listed in table 6. Write command, like the read command is a sequence of data bits comprising of command code, address bits and data bits forming a 2 byte sequence. The write sequence is shown in figure 36 Figure 36: Timing diagram for SPI write operation 5.3 Refined Tag Circuit The refined tag circuit incorporates a microcontroller interfaced through the SPI to the SL900a RFID IC. External sensors are hooked up directly to the MCU I/O peripheral instead of the SL900a s EXT1 and EXT2 external sensor pins. The reason for the same is to have the sensor data directly passed on to the processor rather than first going through 58

75 the RFID IC, which would lead to increased power consumption and processing time. The block diagram of the tag circuit with attached sensors is shown in figure 37. Figure 37: Block diagram of tag circuit Operation Figure 37 shows the block diagram of the updated tag structure. The process flow is as follows: a. The sensors interact with the outside environment and generate either a voltage/current based on change in the environment. b. The analog signal is detected by the analog module inside the MCU chip and fed to the ADC which generates a corresponding digital number between c. This number is then processed with the help of a targeted code to generate the desired information. 59

76 d. This information is then logged into the memory of the RF chip via SPI communication with the MCU. e. This information can be retrieved at any later point in time by bringing the tag in the range of the RFID reader. The sensors can be connected to the RF chip too as long as the sensed data is transferred to the MCU for processing. It is not a very smart idea to connect the sensors to the RF chip for this configuration, since the sensed data has to be continuously transferred to the MCU via the Serial Peripheral Interface (SPI) protocol before it can be processed. This repeated process of SPI data transfer is undesired due to increased power and time consumption Breadboard Setup The tag circuit setup is shown in figure 38. The setup comprises of the SL900a test board tag on a breadboard interfaced to the ATMEGA328P MCU. The PVDF (Polyvinylidene fluoride) [22] piezo sensor used as an external analog sensor, is connected to the I/O pin of the microcontroller. Figure 38 also depicts the flow of voltage signals starting from the sensor and ending with the processed information logged inside the SL900a EEPROM. To read the logged information, the SL900a tag board is illuminated by the RF field of the RFID reader. 60

77 Figure 38: Lab test setup The size of the tag circuit in figure 38 is huge and absolutely not feasible for demonstration of a proof of concept, much less for actual deployment inside a tire Development of Prototype Board To be able to demonstrate a proof concept of tag operation inside a tire, a compact version of the RFID tag is required. The entire setup of figure 38 is transferred to a through-hole PCB as shown in figure

78 Figure 39: Prototype tag board The assembled prototype board consists of the SL900a RFID test board integrated to a through-hole PCB, comprising of a microcontroller, interfaced sensors and an indicator LED. The board is powered by a 3V Li ion battery. Features of the prototype board: a. Supports multiple sensors b. Data Processing c. Data Logging d. Data retention until overwritten e. Wireless data transmission f. Reprogrammable 62

79 5.3.4 Testing the Prototype Board Description: The prototype tag board was tested for concurrent multisensing, data processing and logging in free space. A temperature sensor and PVDF piezo sensor were used for demonstration. Process The process is preprogrammed to run for 30 seconds in the absence of a reader. The tag records temperature values and processes them to generate maximum, minimum and average temperature values as shown in figure 40. The tag simultaneously keeps track of the pressure applied to a PVDF sensor via a threshold detection algorithm as shown in figure 40. At every threshold crossing, the count is updated. After 30 seconds, the temperature data and count are logged into the RFID chip memory. Only 4 bytes of memory were consumed. This data can later be retrieved via the RFID reader. Of the four bytes, the first three bytes store the temperature information, while the remaining byte stores the count. 63

80 Figure 40: Threshold detection and temperature sensing using prototype tag board Table 7 lists the operational voltage, average power draw and the read range of the tag Table 7: Test conditions and results Parameter Value Voltage Power Draw Battery type Read Range (free space) 3 Volts mw Lithium ion coin cell CR dBm 64

81 5.4 Proof of Concept With the prototype board correctly configured and programmed, the next step is counting the number of revolutions of a tire. Two tests were performed: Test 1: Logging bicycle tire revolutions Test 2: Counting automotive tire revolutions For each of these tests, the tag was preprogrammed to count revolutions for a set time interval, store the count in the MCU EEPROM and then transfer this count to the memory of the SL900a RF chip, which could then be read later at any point in time using a RFID reader. The tag uses only 4 bytes of memory space to store up to 2 32 or around 4 billion counts. This is where the microcontroller plays a key role. It allows for efficient and controlled data logging. This feature is not supported by any commercial UHF class 1 Gen 2 RF chip Test 1: Counting Bicycle Tire Revolutions The entire setup and process flow is shown in figure 41. The process involves mounting the tag around the center of the tire to keep it stationary, while the PVDF piezo sensor is fixed onto the outer surface of the tire. 65

82 Figure 41: Test setup and process for bicycle tire rev counting The sensor was activated in two ways. In the first case I actually rode the bike to count revolutions and in the second case, as shown in figure 41, the sensor was activated by rotating the tire freely by hand and using a steel plate to simulate the tire-road interaction Test 2: Counting Automotive Tire Revolutions Setup The setup comprises of a tire attached to a mechanical fixture as shown in figure 42. This mechanical fixture allows the tire to be easily rolled along the ground. The tag board was then duct taped to the inner wall of the tire which was then moved to and fro to simulate road-tire interaction. 66

83 Figure 42: Test setup for automotive tire rev counting Process This setup was used to count tire revolutions for a preprogrammed interval of 30s. The MCU connected to the piezo sensor processes the generated voltage to determine a revolution and updates the count in the MCU memory. When the tag finishes the operation, the MCU transfers this logged count to the memory of the RF chip which is later retrieved using an RFID reader. The memory can be cleared and a new logging process be started using the reader. The flow of the entire process is depicted in figure

84 Figure 43: Process flow for rev count operation Results & Conclusions a. Correctness of operation: The tag board was tested multiple times for the same operation and gave the correct revolution count each time. b. Read Range: The tag s memory was read when placed 2.5 feet away from the reader antenna. c. Data Sensing & Processing: The tag processes sensed data to extract useful information d. Flexible and Controlled Data Logging: The tag was smartly programmed to use n bytes of memory to store 2 8*n revolutions. As shown in figure 45, only 4 bytes of the 9Kb EEPROM of RF chip were used to store 2 32 revolutions (approx. 4 billion). 68

85 Figure 44: Read range for reading tag in tire Assuming tire diameter = 1m, 4 bytes of memory can store revolutions for over a distance of approx. 13,493,037 kilometers or 8,384, miles (8.3 million miles). Remaining 1047 bytes can be used for storing average speed, average temperature, maximum and minimum temperature values etc. This clearly highlights the importance of smart data sensing & logging using RFID chips. e. Reconfigurable tag: The tag can be reprogrammed at any later point of time. 69

86 Figure 45: Logged rev count inside SL900a EEPROM Drawbacks: a. The prototype tag board is still too bulky to be embedded inside a tire. b. The revolution count experiment was done by manually rotating the tire using the setup shown in figure 42. Since the speed of rotation and the tire load is fixed and considerably low, the effectiveness & accuracy of the algorithm is not very reliable. c. The prototype board is powered by a battery, which is not an ideal component in a harsh tire environment. Moreover no one would want to replace the battery on an embedded tag. Chapter 6 will focus on overcoming the first two drawbacks. 70

87 Chapter 6: Miniaturization Board Design Chapter 5 demonstrates the proof of concept of the operation of a RFID tag for in-tire multisensing and data logging applications. Tire revolutions were counted using the tag board as a part of the demonstration process. However, as mentioned towards the end of chapter 5, there are drawbacks of the prototype board that prevent it from being used as a commercial in-tire multisensing wireless tag. This chapter focuses on the work done to overcome the drawbacks related to tag size. Furthermore, the tag s rev counting operation is tested for varying tire loads and speeds. 6.1 PCB Design With a working prototype, the next step is tag miniaturization. The goal is to assemble the same circuit using smaller components, all neatly interfaced on a compact board Circuit Schematic The first step is to create a schematic for the tag circuit in a PCB CAD tool. The PCB design tool used was EAGLE CAD. There were a few reasons behind choosing EAGLE CAD as the design tool. One, the board design is not complicated and easily achievable 71

88 through a two layer design. Secondly, there are no high frequency routing lines and hence no need for an advanced PCB design software for modelling and simulating high frequency routing effects. Finally, EAGLE CAD has massive open source support and is backed by considerable online literature. Figure 46 shows the schematic of the tag electronics prepared in the EAGLE CAD environment. Figure 46: Schematic of tag circuit The major components on the schematic are: a. SL900a RFID IC 72

89 b. ATMEGA328P SMD version c. 6 pin ISP Programming Header d. 6 I/O through hole solder pads e. Passive components resistors, inductors and capacitors f. LED for output indication and debugging Most of the footprints are available in the standard library of the CAD tool. However for this particular design, the EAGLE library did not incorporate the footprint of the SL900a IC by AMS. Even though the IC is 5mm x 5mm QFN 16 footprint, it is not wise to substitute this footprint with a footprint of a similar IC. There can be minor mismatch between the two footprints and this can later result in higher costs, wasted time and efforts. Hence a library was created in EAGLE CAD containing the footprint of the SL900a IC Board Layout Once the schematic of the board was complete, the next step was the generation of the board layout. The CAD tool generates the board layout automatically from the schematic. However component placement and routing has to be done carefully to reduce the board size while avoiding electrical faults. The finished layout of the circuit board is shown in figure

90 Figure 47: Board layout Once the board layout was finished, DRC checks and electrical fault checks were performed before generating the Gerber files. The final board dimensions are: 43mm x 36mm Board Fabrication After full verification and the generation of Gerber files, the next step is the fabrication of the actual board. For this process, the Gerber files, along with the component list and component placement files were sent to the PCB fabrication house. 74

91 The final board after fabrication is shown in figure 48. Figure 48: Final tag board next a US quarter Notes: The above board has two slots in the upper region on top of the RFID chip to connect to the antenna. The antenna deigned for this application is shown in figure 49. For more details on the antenna design refer to appendix A. 75

92 Figure 49: Designed tag antenna for tires The Microcontroller can be programmed by hooking the ISP programmer via the 6 pin ISP header. 6.2 Testing the Board The Hello World LED Test LED blink test is the hello world equivalent for testing hardware systems. The motive behind this test is to ensure that the fabricated board operates as desired without any issues. The embedded C code was written in the Atmel studio environment. The code hex files were flashed using the extreme burner software. The board was programmed using the USBasp programmer. 76

93 Figure 50: Programming the tag board An ISP6 header on the board is used to interface with the external USBasp programmer. The LED blinks on and off at a rate of 500ms. A snapshot of the board running the code is shown below in figure 51. Figure 51: Working tag board running the LED blink code 77

94 6.2.2 Revolution Counting for Varying Tire Load and Speed In Chapter 5, a proof of concept of tire revolution counting application using the RFID prototype tag board was presented. However due to setup limitations, the applied tire speed and load were quite low and could not be varied. To ascertain the correctness of the RFID tag for counting revolutions under varying load and speed conditions, actual drum tests were conducted at Bridgestone s tire facility in Akron, Ohio. However for these tests, the board wasn t embedded inside the tire, but the PVDF sensors were attached to the inner tire surface in different configurations as shown in figure 52. The signals generated by the PVDF sensors were captured using National Instrument s data acquisition system and replicated using a waveform generator. These signals were then fed back to the RFID board to determine the accuracy of the revolution count algorithm for varying speed and load conditions. Figure 52: Setup for PVDF piezo sensors inside a truck tire 78

95 Figure 53 shows the process of capturing the signals from drum tests and regenerating using a waveform generator. These signals are then fed back into the tag board for testing. The drum tests were performed for combination of different speeds and loads as shown in table 8. Table 8: Speed and load conditions for drum tests Speed (Kph) Load (KN) Figure 53: Process for regenerating tire test conditions for testing tag board 79

96 Test 1: Rev Count for 5KN Load Voltage profile from drum testing at 20 Kph, 5 KN load was regenerated using an arbitrary waveform generator. The profile is shown in figure 54. In the figure, the waveform duration is one second. It is clearly visible from the waveform that two revolutions occurred during the 1 second time period. Figure 54: Regenerated voltage profile for 20Kph, 5 KN load The waveform was then fed to the RFID tag board running the rev. count algorithm. The algorithm ran for duration of 10 seconds and returned an expected count of 20. In a few cases the recorded count was 21, but this is can be explained with the reasoning that at 80

97 time t = 0 or time t = 10s, the circuit probably encountered a revolution still in transition. For several repeated tests, the revolution count was consistently accurate. Test 2: Rev Count for 15KN Load Keeping the speed constant while varying the load to 15 KN, has no significant impact on the generated voltage signal. In fact it looks exactly the same. The signal is shown in figure 55. Figure 55: Regenerated voltage profile for 20Kph, 15 KN load 81

98 The signal fed into the RFID circuit for a 10 second processing interval consistently generated the correct revolution count of 20. Test 3: Rev Count for 15KN Load The waveform for the maximum speed and peak load conditions is shown in figure 56. Figure 56: Regenerated voltage profile for 80Kph, 15 KN load The signal when fed into the RFID tag circuit generated a consistent rev count of Based on the above signal, the expected rev count for 10 second is 70. These tests establish the correctness of the rev count algorithm for varying speed and load conditions. 82

99 6.3 Conclusions Two major problems have been solved. One the tag has been miniaturized to a compact form, considerably small and ready to be used for in tire applications. The other conclusion is that the tag circuit s rev count algorithm works correctly for variations in tire speed and load conditions. Despite the above achievements, one final hindrance to tag deployment inside tire is the on board battery running the tag circuit. A battery is undesired for a couple of reasons. One is that it is unsafe to run a circuit with a battery operating in extreme temperature and pressure conditions. The second is the ordeal of replacing a battery on a tag embedded inside a tire. The next chapter focuses on resolving this issue. An approach to self-power the tag by harvesting the mechanical energy from the tire using PVDF piezo sensors will be covered in the next chapter. 83

100 Chapter 7: Self-Powered Tag Operation Until this point, a compact RFID tag board capable of in tire data sensing and logging has been developed. The power source for the tag however, is a 3V Li-ion coin cell. The focus of this chapter is to eliminate the battery and power the tag electronics using harvested energy. The modified tag circuit s block diagram is shown in figure 57. The tag circuit is now composed of a RFID IC, a MCU and power conditioning circuit replacing the battery. Figure 57: Modified tag to incorporate an energy harvesting block 84

101 Before beginning the development of a self-powered tag, the following three things need to be analyzed: a. Source of energy for power harvesting b. Power conditioning or regulating circuit c. Power consumption of the tag circuit Firstly, an energy source needs to be identified. Once this is done, a thorough quantitative analysis is required to determine the available harvested power. Next step is to quantify the power consumption of the existing tag circuit. If it is too high to be powered by the harvested energy, steps have to be taken to bring down the consumed power. Finally, if the power generation and power consumption values are not too far off numerically, a suitable power conditioning circuit has to be chosen and interfaced between the source and the load. Since the harvested power is unregulated and inconsistent, a storage unit along with the regulating circuit will be required Energy Harvesting Sources and Harvesters For in tire RFID s, there are two sources of available energy a. Energy from the electromagnetic field of the reader b. Mechanical energy 85

102 Based on the concept of operation mentioned in chapter 1, mechanical energy will be utilized to power the tag during run time. During this period, the tag will perform data sensing and logging operations. When the vehicle comes to a halt, there is no source of mechanical energy to power the tag. However, there is no need to power the tag when stationary unless the information has to be off-loaded. In that case, the RFID reader s electromagnetic field will be used to power the circuit and off load the necessary information. Figure 58 depicts the energy harvesting approach. Figure 58: Energy sources to power the RFID tag circuit To harvest the available free energy, devices known as energy harvesters are required. In our case, piezo sensors are used as harvesters. Piezo sensors convert waste mechanical 86

103 energy to electrical energy. The other harvesting source is the RFID IC. It has an on chip rectifier to harvest the reader s electromagnetic field by converting it to usable DC voltage to power the tag circuit Piezo Sensors as Harvesters Piezo materials convert mechanical energy to electrical energy by converting applied force to voltage. There are several kinds of piezo substances. For this project, PVDF [22] type piezo sensors have been chosen. The reasons for the same are: Decent Energy Harvesters Good sensors Ability to withstand high temperatures (up to 100ºC) Test 1: Drum Tests using PVDF Sensors For this test, four PVDF sensors were placed inside a truck tire. The sensors were set up as shown in figure 59. Each sensor is covered with a tire patch. Of the four PVDF sensors, two are operating individually while the other two are stacked in either in series / parallel configuration. 87

104 Figure 59: Arrangement of PVDF sensors inside tire The tire, along with the sensors was then mounted on a wheel rim as shown in figure 60, and subjected to drum tests. Figure 60: Tire rim with external leads to connect to PVDF placed inside the tire The drum tests were performed for combinations of speed and load as shown in table

105 Table 9: Different values of tire speeds and loads for drum tests Speed (Kph) Load (KN) Figure 61 depicts the drum test in action (left) and the NI data acquisition module (right) used for data collection. Figure 61: Data collection setup for drum tests The voltage signals generated by a single PVDF sensor across an open circuit are shown in figure 62 & 63. Figure 62 shows the voltage signals generated for a constant load of 89

106 5KN and speed values of 20, 40, 60 and 80Kph. Figure 63 shows the voltage signals generated for a constant speed of 80Kph and load values of 5, 10 and 15 KN. One thing to be noted is that the magnitude of the voltage shown in the figures is 1/12 th of the original value. This has been done using a voltage divider network to ensure that the voltage entering the NI DAQ system is inside safe limits. Figure 62: PVDF voltage profile for 5KN load and varying speeds 90

107 Figure 63: PVDF voltage profile for 80Kph and varying loads Estimating Power Output In the previous section we tested the PVDF sensors inside a tire. In this section we will use the voltage generated by a single PVDF sensor inside a tire running at 80Kph, to estimate the power dissipated across a load resistor connected to it. The voltage profile across the PVDF with no load at 80Kph, 15 KN is shown in figure

108 Figure 64: PVDF open circuit waveform for 80Kph, 15KN load For analysis, this voltage post rectification can be approximated to a square type pulse waveform as shown in figure 65. Figure 65: Post rectified PVDF open circuit voltage approximated to a pulsed signal 92

109 The circuit can be modelled as a simple RC circuit as shown in figure 70. Figure 66: RC circuit with square pulsed input From the waveforms in figure 64 and 65, we have: Time period T : 150ms Ton : 50ms, Toff : 100ms Frequency f: 7Hz V H : 12V, V L = 0 Based on the energy & power calculations described in Appendix C; for a simple RC circuit, the power or rate of energy dissipation across the load directly connected to a PVDF sensor can be computed as: Pr = CVin 2 f = = 1.31 uw [7.1] 93

110 The equation E = CVin 2 f assumed that the capacitor was fully charged and discharged during each on and off cycle. The Energy dissipated across the resistor during charging phase can be written as: Er = Vin2 Ton e 2t Γ R 0 = CVin2 2 [ 1 e 2Ton Γ ] [7.2] Based on the above equation, for a duration of Ton > 2Γ, the capacitor would have charged more than 98.1%. Same concept applies for the discharging period. For the PVDF pulse-approximated signal in figure 65, Ton = 50ms and Toff = 100ms. For C = 1.3nF, and 2Γ < Ton, R should be less than 19.3 M. However the average power dissipated across the resistor is still CVin 2 f = 1.31uW. To increase the average power dissipated across the load resistor, either or the entire C, V and f parameters should be increased. C is limited to 1.3nF by design, f and V values of 7Hz and 12 V are for a tire running at 80Kph and 15 KN load. 94

111 7.2 Tag Power Consumption The tag circuit shown in figure 42 of chapter 5 runs on a 3V battery. The proof of concept application using the tag circuit to count revolutions consumed an average power of 13.53mW with an average current draw of 4.1mA. To reduce the power consumption of the tag circuit, the following measures were taken: a. Reduction of operating voltage b. Reduction of microcontroller s operating frequency c. Elimination of external floating pins d. Effective Utilization of Microcontroller low power modes Table 11 lists the effect of reduced voltage and frequency on the average power consumed by the tag circuit. Reducing the voltage from 3.3V to 1.8V and frequency from 8 MHz to 1 MHz brought down the circuit power consumption by a factor of 19 or by 139%. Table 10: effect of voltage & frequency on power consumption Voltage & frequency Average Power consumption Average current draw 3.3V, 8 MHz 13.53mW 4.1 ma 3.3V, 1 MHz mw 0.65 ma 1.8V, 1 MHz 0.72 mw 0.4 ma 95

112 The tag circuit was retested for multisensing and data logging operation at reduced voltage and frequency levels. The tag circuit worked as expected. Efforts were made to further reduce the power by reducing the operating MCU frequency. However the clock on the ATMEGA328P MCU is limited to only a few discrete values. The lowest frequency setting on the MCU is 125 KHz (1/8 th of 1MHz). The tag did not function as expected at 125 KHz. Further details on ATMEGA328P s clock configuration settings can be found in [20]. When the tag s supply voltage level was reduced to 1.65V, the tag circuit stopped working. The voltage and frequency adjustments were made by changing the fuse bits on the microcontroller. One can refer to [23] for detailed information on AVR fuse settings. Another technique commonly used for reducing power consumption of the MCU is selectively switching on and off different modules on the controller. ATMEGA328P [20] offers several low power modes and sleep modes. However for this application, these modes won t do much good since the tag circuit needs to constantly track the voltage signal from the sensor to count revolutions. Shutting on and off the MCU modules will lead to inaccurate and undesirable results. 96

113 7.3 Conclusion At this point, we have quantified the energy available at the input of the regulating interface circuit and the energy requirements of the tag circuit. At this moment, the minimum power requirement for the tag to operate reliably is around 700uW, while the theoretically estimated power at the storage unit of the harvesting circuit is around 1-2 uw per PVDF sensor. This enormous disparity makes self-powered operation unfeasible at this point in time. The behavior of piezo materials is a strong function of the frequency of operation and mode of excitation. The work in [24] analyzes the operation of piezo sensors and provides a mathematical model for understanding the load matching requirements for piezo sensors. Optimizing the PVDF sensor s operation inside a tire could help generate higher power levels. To have a realizable self-powered operation, the energy consumption has to be minimized while maximizing harvested power. To reduce the tag s power consumption, one possible approach is to alter the tag operation to reduce the active run time. Another approach is to replace the current MCU with a better one. For example, MSP430 [25] by Texas Instruments is well known for its use in low power applications. Thus moving forward involves reducing the disparity between the available and required power. It is therefore important to first address this issue before diving into an extensive review of available literature on power regulating and conditioning circuits. 97

114 Chapter 8: Conclusions and Future Work A compact battery operated RFID tag circuit has been presented for in-tire data sensing and logging applications. A proof of concept has been demonstrated by using the prototype tag board to count tire revolutions. The developed tag circuit needs some structural modifications before deployment inside tires for commercial applications. The developed tag board is actually a compact reconfigurable RFID development module. With an onboard programmable microcontroller unit, the board can be configured for use in different environments using multiple sensors. An external, specific to application antenna can be designed and attached to the tag board. Future improvements for the board will involve size reduction, a robust & flexible PCB substrate and a custom software interface to rapidly configure the board for a range of applications. The developed tag board currently requires an onboard battery. This is undesired for multiple reasons: increased size, unsafe in-tire operation and replacement issues. Efforts were put towards eliminating the battery by using PVDF piezo sensors to harvest in tire waste mechanical energy and power the tag. In an effort to achieve self-powered tag operation, several drum tests were done to quantify the harvested and the required power. The tag power consumption for the final tag board was reduced by 139%. Despite all the 98

115 efforts, the gap between the available & required power is still too high. As a result, at this time an onboard battery is being used to power the tag board. Future work will be focused on increasing the harvested power while decreasing the consumed power. Once the battery can be completely eliminated from the circuit, the size will be considerably reduced making the tag ready for commercial in tire deployment and testing. 99

116 References [1] M. Brandl, K. H. Kellner, G. Grabner and J. Grabner, "Utilizing RFID Technology for Biomedical Sensor Applications", in World Congress on Medical Physics and Biomedical Engineering, vol. 25/11 of IFMBE Proceeding pp [2] L. Battle, G. Cole, K. Gould, K. Rector, S. Raymer, M. Balazinska and G. Borriello, Building the Internet of Things Using RFID: The RFID Ecosystem Experience, in Internet Computing, IEEE,, vol.13, issue: 3, [3] S. Dominikus, J.M. Schmidt, Connecting Passive RFID Tags to the Internet of Things, in Interconnecting Smart Objects with the Internet Workshop, Prague, [4] A.A. Babar, S. Manzari, L. Sydanheimo, and A.Z. Elsherbeni, Passive UHF RFID tag for heat sensing applications, in IEEE Trans. Antennas Propag., vol. 60, no. 9, pp , Sep [5] C. Occhiuzzi, C. Paggi, and G. Marrocco, Passive RFID strain-sensor based on meander-line antennas, in IEEE Trans. Antennas Propag., vol. 59, no. 12, pp , Dec [6] AMS SL900a RFID Tag, Available at: Interface-and-Sensor-Tag/SL900A. [7] PE3001 RFID IC, Available at: [8] Caen RFID sensor tag, Available at: [9] S. Shao, L.Z. Lee, R. Burkholder, J.L. Volakis, Embedded UHF RFID Tag Antennas for Automotive Tire Sensing in IEEE Trans Antenna and Propagation, page , Jul

117 [10] A summary of RFID Standards, Available at: [11] EPC Global Standard Specification, Available at: pdf. [12] B. Fennani, H. Hamam, A. Dahmane, "RFID overview", in Microelectronics (ICM), Accession Number: [13] C.M Roberts, Radio Frequency Identification (RFID), in Computers & Security, vol. 25, issue 1, February 2006, Pages [14] G.White, G.Gardiner, G.P. Prabhakar, and A. Razak, "A comparison of barcoding and RFID technologies in practice", in Journal of Information, Information Technology and Organizations, 2. pp [15] J.Grosinger, L.W Mayer, C.F. Mecklenbraeuker and A.L Scholtz, Input Impedance Measurement of a Dipole Antenna Mounted on a Car Tire, in Proceedings of the International Symposium on Antennas and Propagation (ISAP), Bangkok, Thailand, October 2009, pp [16] T. Wei and P. Wilson, "Read Range Sensitivity of Embedded RFID Tags in Commercial Tires", in Tire Science and Technology: April-June 2016, Vol. 44, No. 2, pp [17] S. Shao, A. Kiourti, R. Burkholder and J.L. Volakis, Broadband and flexible textile RFID tags for tires, in IEEE Trans Antenna and Propagation, page 1507, Jul [18] S. Shao, R. Burkholder, J.L. Volakis, Physics-based approach for antenna design optimization of RFID tags mounted on and inside material layers, in IEEE Trans Antenna and Propagation, page , Jul [19] RFID Reader: AMS AS3993 Radon, Available at: Demo-Kit-Radon [20] AVR ATMEGA328P datasheet, Available at: bit-AVR-Microcontroller-ATmega48A-48PA-88A-88PA-168A-168PA P_datasheet_Complete.pdf 101

118 [21] SPI of the AVR, Available at: [22] LDT1-028K PVDF Piezo Sensor, Available at: DT0-028K.pdf [23] AVR Fuse Settings, Available at: 328p-fuse-settings/ [24] J. Liang, W.H. Liao, "Impedance matching for improving piezoelectric energy harvesting systems", in The International Society for Optical Engineering, March 2010,vol [25] MSP430 Microcontroller, Available at: [26] Understanding RFID Standards, Available at: [27] J.F. Salmeron, A.Rivadeneyra, F. Martinez-Marti, L.F. Vallvey, A.J. Palma and M.A. Carvajal, Passive UHF RFID Tag with Multiple Sensing Capabilities, in Sensors 2015, 15, [28] Farsens RFID IC, Available at: [29] Ramtron WM7201, Available at: dg/wireless-memory-fram-16kb-8udfn/dp/ [30] D. De Donno, L.Catarinucci, and L. Taricone, RAMSES: RFID augmented module for smart environmental sensing, in IEEE Trans. Instrum. Meas., vol.63, no.7, pp , Jul [31] R. Colella, L. Tarricone and L. Catarinucci, "SPARTACUS: Self-Powered Augmented RFID Tag for Autonomous Computing and Ubiquitous Sensing," in IEEE Transactions on Antennas and Propagation, vol. 63, no. 5, pp , May [32] S.M. Taware, S.P. Deshmukh, "A Review of Energy Harvesting From Piezoelectric Materials", in IOSR Journal of Mechanical and Civil Engineering, ISSN(p): X, pp:

119 [33] A. Majeed, "Piezoelectric Energy Harvesting for Powering Micro Electromechanical Systems (MEMS)", in Journal of Undergraduate Research 5, 1 (2015) [34] Piezo Film Sensors Technical Manual, Available at: 103

120 Appendix A: Tag Antenna Design Summary The research presented in this thesis focuses on the RFID tag IC development and not the tag antenna. A RFID system however, is incomplete without a working tag antenna. For this in tire application, a robust and flexible broadband dipole antenna was developed at the Ohio State University s ElectroScience Laboratory. The publications [9], [17] & [18] by S.Shao et al, cover the compete design process of the RFID tag for automotive applications. This particular section summarizes the features of the designed antenna. Antenna for Tires An antenna that is broadband and mechanically robust is required for application inside the tires. The following section covers the features of the designed antenna as shown in figure 67 and 68. Figure 67: Flexible e-fiber RFID antenna in polymer coating 104

121 Broadband Antenna The designed antenna is a modified end loaded meander line dipole antenna. A clear picture of the antenna is shown in figure 68. The end loops make the antenna broadband and circular shape allows for uniform stretching while maintaining electrical performance. The center loop adds inductance to the antenna which compensates the capacitance of the RF chip. The meandering dipole arm reduces the overall length of the antenna and facilitates in flexing. This design requires a single thread of antenna embroidered to take the desired shape and hence fabrication and mass reproduction is easy. Figure 68: RFID antenna for tires Flexible Textile Design In-situ tire placement of tags demands a robust antenna structure capable of withstanding extreme variations in temperature, pressure and stress. Commercially printed and wired tags are not designed to handle such environments and will eventually breakdown. The antenna presented in figure 67 is embroidered using electrically conductive metalpolymer fibers embedded inside a polymer coating. The e-fibers uphold the electrical 105

122 property of the antenna while the polymer coating imparts mechanical integrity to the tag structure and assists in placement of the tags on the inner wall of the tire. The e-fiber used for this design is a 332-strand silver coated Amberstand. Each single e-fiber structure is composed of a 10um high strength polymer core with a 2-3 um thick silver coating. This structural composition imparts to these fibers excellent mechanical strength and flexibility and a low DC resistivity (0.8Ω/m). Performance Tags in tires are required to deliver reliable electrical and mechanical performance. Read range, bandwidth and radiation pattern are important metrics for an antenna s electrical performance. A. Electrical Performance The designed antenna is small in size with dimensions 87mm x 19mm and produces an Omni-directional radiation pattern. This antenna was tested for power reflection coefficient in free space and when mounted on a 4 mm thick dielectric with ɛr = 10, 13 and 16 respectively. The tag delivers a 3dB bandwidth of 263 MHz in free space and a sufficient bandwidth of 171 MHz for ɛr = 10 over the RFID band ( MHz) thereby allowing for variability in the surrounding material. Figure 69 shows the plot for reflection coefficient between antenna and RF chip. 106

123 Figure 69: S11 between antenna and RFID chip for different dielectrics The tag was subjected to read range and threshold tests. The setup for the test is shown in figure 70. The tag was mounted on outer surfaces S1, S2, S3 and S4 of two truck tires at six different locations P1-P6. The space (S) between tires is 15 cm as between two adjacent tires in a truck. The reader power was set to 30dBm and the reader was moved until it could no longer detect the tag. The measurements results are shown in table 12. Figure 70: Setup for read range test and tag placement on tire The threshold test was also performed to determine the antenna s performance. For this test the four antennas with the same design but fabricated differently were mounted on surface S1 and read from a fixed distance of 5 feet. The reader power was decreased until the tag couldn t be read. The lower the threshold power, the better the antenna. The 107

124 measurement results are recorded in table 13. The threshold value is almost the same for all the antennas which establishes the reliability of the design Table 11: Read range tests Antenna/Surface S1 S2 S3 S4 Copper Antenna with Polymer coating Table 12: Threshold test for the designed antenna Antenna/Location P6 P4 Copper Wire Embroider (Single & Double Thread) Copper Wire W/ coating Embroider W/ coating (Single & Double thread) B. Robustness During the process of embedding tags inside tires, the tag will suffer deformation along the longitudinal direction. Hence it is very important that the tag maintains reliable performance once embedded inside the tire. To ensure this, the tag was subjected to a stretch test and maintained the same performance after being stretched by nearly 10%. Performance inside tires The final performance analysis involved testing the tags embedded in tires. Figure 71 shows the fabricated textile tag embedded in the side wall of a tire. The designed tag and a commercial tag placed inside tire were subjected to read range tests. The maximum read range of the commercial tag is 1.2 m which is much less than the read range of 2.8m recorded using the designed tag inside a tire 108

125 Figure 71: Designed e-fiber tag embedded in tire 109

126 Appendix B: List of Abbreviations ADC: Analog to Digital Converter ASK: Amplitude Shift Keying BAP: Battery Assisted Passive CAD: Computer Aided Design EEPROM: Electrically Erasable Programmable Read Only Memory EPC: Electronic Product Code I2C: Inter Integrated Circuit IC: Integrated Circuit ISP: In System Programmer MCU: Microcontroller PCB: Printed Circuit Board PSK: Phase Shift Keying PVDF: Polyvinylidene fluoride QFN: Quad Flat No-Lead RFID: Radio Frequency Identification RSSI: Received Signal Strength Indicator SPI: Serial Peripheral Interface UHF: Ultra High Frequency 110

127 Appendix C: Electrical Model of the Piezo Sensor The piezo sensor can be represented using its electrical equivalent circuit. The circuit is a voltage source in series with a capacitance as shown in figure 72. Figure 72: Electrical equivalent circuit of piezo sensor For a piezo sensor, the capacitance results from the piezo film and the conductive plates formed from conductive electrodes printed or metallized on each surface of the film. This capacitance can be estimated using the following equation: C = Aε t [C.1] 111

128 Where: C is the film capacitance ε is the permittivity. ε can be expressed as ε r* ε o, where ε r is the relative permittivity and ε o is the permittivity of free space (8.854 x F/m). A is the active overlap area of the film s electrodes t is the film thickness. The piezo sensor used for this project is the LDT1-028K PVDF sensor. PVDF has permittivity ε = 12 and for the given dimensions, the material has a capacitance of 1.3 nf. The detailed technical information on the sensor can be found in [34]. Effect of Loading on the PVDF Sensor Figure 73 represents a PVDF sensor connected to a load. Together the load and the intrinsic capacitance of the PVDF form a voltage divider network. Figure 73: Piezo sensor connected to a load in series 112

129 With the output taken across the load, the circuit acts as a high pass filter with the cutoff frequency given by: fc = 1 2πRC [C.2] For example if the load impedance is taken to be 1MΩ, C = 1.3nF, the cutoff frequency turns out to be Hz. To maximize the voltage available form the PVDF at the resistor, frequency of the PVDF signal must be greater than the cutoff frequency. Estimating Power across a load resistor connected directly to the PVDF sensor The electrical model of a PVDF sensor connected to a load resistor is shown in figure 74. This is a basic RC circuit. In this section, we will analyze an RC circuit with a square pulsed input signal. The circuit is shown below in figure x. Figure 74: RC circuit for PVDF analysis 113

130 The charging phase As the input voltage transitions from low to high, the capacitor is charged by the energy supplied by the voltage source. The voltage and current equations during the charging phase are given as follows: Voltage across capacitor: Vc(t) = Vin (1 e t Γ ) [C.3] Voltage across resistor: Vr (t) = Vin e t Γ [C.4] Current in the circuit: i(t) = Vin R e t Γ [C.5] Where Γ is the circuit time constant = RC Energy delivered by source (E) = Energy dissipated across Resistor (Er) + Energy stored across Capacitor (Ec) At t = 0, Capacitor is completely discharged and Ec =0. At t =, capacitor is fully charged and Ec = C Vin2 Total energy supplied by the battery E: E = Vin i(t) dt = Vin i(t)dt [C.6] E = Vin2 R 0 e t E = CVin 2 Γ dt 114

131 Energy dissipated across R = E Ec = CVin2 2 [C.7] Discharging Phase The supply is turned off and the capacitor discharges through the resistance. The voltage and current across the circuit during discharge transition are as follows: Voltage across capacitor: Vc(t) = Vin e t Γ [C.8] Current in the circuit: I(t): Vin R e t Γ [C.9] Energy stored in the capacitor at t =0 is Ec = CVin2 Energy stored in the capacitor at t =, Ec = 0 2 Energy supplied by the source = Vin i(t) dt = 0 0 [C.10] Energy stored across the capacitor is dissipated across the resistor during discharge. This dissipated energy is equal to: Er = CVin2 2 [C.11] 115

132 For a complete switching cycle, the resistor dissipates CVin 2 energy. If the switching frequency of the pulse is f, then the average power dissipation across the resistor is equal to: Pavg = CVin 2 f [C.12] PVDF Sensor Interfaced to a Harvesting Circuit Figure 75 shows the model for a Piezo sensor interfaced to a regulating circuit to power the main tag circuit. Figure 75: PVDF sensor interfaced to a regulating circuit The signal generated by the PVDF passes through its own internal impedance, the impedance at the input of the harvester circuit and the full bridge diode rectifier circuit before it can charge the storage capacitor. The measured power at the storage capacitor will be lower than the generated power depending upon the magnitude of the loading effect and the losses across the full bridge rectifier circuit. 116