Master Thesis
Master's Programme in Electronics Design, 60 credits
On-PCB Inverted-F Antenna Design For Wireless Sensor Nodes
Thesis in Electronics, 15 credits
Halmstad 2020-06-15
Syed Muhammad Asad Tirmizi, Okeke Amalachukwu
Abstract
The Internet of things (IoT) is a disruptive innovation which has shown the potential to change the way we live our lives. At the core of the IoT eco systems are the wireless sensor nodes. These are responsible for the sensing of the target environment parameters or situations and communication of such to the desired destinations. For this communication to happen, a good performing antenna is required.
This project explores the design of an on-PCB inverted-F antenna for a wireless sensor node. The Literature review and insight into wireless sensor nodes and antenna design are conducted. The designed antenna is stimulated using ADS simulation software and the performance analysis presented. Design description of a wireless sensor node is also made, which includes the high frequency PCB layout design for the sensor node. The antenna and PCB are both fabricated in the lab, while various antenna performance measurements and evaluation are done. The designed antenna showed good S11 parameter of -43dB on simulation and -20dB when fabricated.
Acknowledgements
Our profound gratitude goes to Björn Nilsson, for his supervision and guidance throughout this thesis work. A big thanks goes to Pererik Andreasson for his role as the examiner. We also like to thank Per-Olof Karlsson, who devoted his time to guiding us through the use of OrCAD cadence and for assisting with the fabrication of the PCBs. A big thanks to Per Sandrup, for his role in the use of ADS and the antenna testing.
We thank our project originator, Johan Malm for the wonderful learning opportunity. And finally, we are grateful to God, our families, and our friends for their support in different ways during this time.
Table of Contents
ABSTRACT ...II
ACKNOWLEDGEMENTS ... III
TABLE OF CONTENTS ... IV
LIST OF FIGURES ... V
1 INTRODUCTION ... 1
1.1 PROJECT AIM AND SCOPE 3
2 BACKGROUND ... 4
2.1 ANTENNA THEORY 4
2.2 ANTENNA PARAMETERS 5
2.3 IMPEDANCE MATCHING 8
2.4 ANTENNA ELECTRICAL SIZE 10
2.5 MONOPOLE ANTENNAS 11
2.6 MICROSTRIP ANTENNAS 12
2.7 MINIATURIZATION TECHNIQUES 12
3 DESIGN AND METHODOLOGY ... 14
3.1 INVERTED-FANTENNA DESIGN 14
4 WIRELESS SENSOR NETWORK ... 19
4.1 WIRELESS SENSOR NETWORK TYPES 19
4.2 WIRELESS SENSOR NODE HARDWARE DESIGN 21
4.3 PCB LAYOUT DESIGN 26
5 MEASUREMENT AND DISCUSSION ... 31
5.1 DISCUSSION 34
6 CONCLUSION ... 35
7 REFERENCES ... 36
8 APPENDIX ... 41
List of Figures
Figure 1. A typical λ/2 dipole configuration and the current distribution across the antenna
length. ... 4
Figure 2. Radiation pattern of a directional antenna. ... 5
Figure 3. Impedance change with series inductor(clockwise) and series capacitor(anti- clockwise) ... 9
Figure 4. Admittance change with shunt inductor (anti-clockwise) and shunt capacitor (clockwise) ... 9
Figure 5. Illustration of antenna dimensions within the radian sphere ... 10
Figure 6. Monopole antenna signal feed and image creation ... 11
Figure 7. Labelled antenna dimensions. ... 15
Figure 8. Layout figure of the designed antenna. ... 16
Figure 9. Defined substrate in ADS for the layout. ... 16
Figure 10. Simulated Return Loss of designed antenna. ... 17
Figure 11. Simulated Impedance of designed antenna. ... 17
Figure 12. Simulated radiation patterns of designed antenna; YZ-plane(left) and XZ- plane(right). ... 18
Figure 13. Star connection ... 19
Figure 14. Mesh topology ... 20
Figure 15. Hybrid network ... 21
Figure 16. Pin layout of SHT31-DIS temperature sensor ... 22
Figure 17. Basic connections of the sensor circuit ... 23
Figure 18. Basic connections of the microcontroller circuit ... 25
Figure 19. Basic connections of the transceiver circuit. ... 26
Figure 20. Board layout Plan ... 27
Figure 21. Component placement on transmission lines ... 29
Figure 22. Break in Ground Plane ... 30
Figure 23. PCB set up for the antenna parameter measurements. ... 31
Figure 24. The measured return loss of the fabricated antenna ... 31
Figure 25. Measured return loss of the 4mm shorter antenna... 32
Figure 26. Implemented L-network and illustration of impedance matching on a smith chart.
... 32
Figure 27. Measured return loss of the matched antenna compared to the return loss before matching. ... 33 Figure 28. Measured return loss of the matched antenna compared to the simulated return
loss. ... 33
1 Introduction
For many decades, the world has enjoyed the benefits and possibilities associated with the internet technology. Some of these benefits include the ease of information dissemination, ease of business, ease of education, movies, sports, and so many other ways that the internet has improved our lives. The internet is basically a connection of computer devices to servers, which are also computers. Nowadays, this connection has extended to other things that we use in our everyday lives outside the computers.
Things like our home refrigerators, doors, microwaves, physical mailboxes, and anything as desired. This new concept is known as the Internet of things (IoT).
Since its introduction, there has been numerous research activities in finding ways to improve this concept and make it more viable for use by businesses and individuals. The fate of the IoT technology has also been further enhanced by the advent of the 5G network, which when available guarantees strong enough connectivity.
An IoT system is comprised of several parts. Different authors have categorized the components of an IoT system in different ways. However, the processes in all of these ways remain the same. One of the most common categorizations, as seen in [1-3], divides the IoT system into layers. The perception layer which comprises of the sensors that measure the actual parameters, the network layer, the middleware layer, the application layer, and the business layer.
Wireless sensor nodes (WSN) are part of the perception layer of the IoT system. Their primary function is to sense the external environment and measure the desired parameters periodically or continuously. The parameters can range from temperature, humidity, vibrations, pressure to any measurable physical quantity. The measured data are processed and transmitted wirelessly to the IoT base station or end devices as the case may be. To perform these functions, each sensor node is typically equipped with a sensing unit/device, a processing or computation unit, a radio transceiver or other wireless communication device, and a power unit to provide energy for the sensor system[4]. For adequate performance, and to ensure that the IoT network serves its desired purpose, the sensor nodes must be effective and reliable. To have a reliable wireless sensor node entails the optimization of performance at the different sections of the node. Various approaches have been adopted in improving this reliability.
The power efficiency is one of these approaches. On this approach, [5-8] Used the method of energy harvesting from neighbouring RF energies, since our atmosphere is filled with different RF waves from transmitters. The research tried to harness these neighbouring RF energies to be used as power source
for sensor nodes. By so doing, the battery life of the nodes is extended, or in some cases replaced.
Experimental results show that a sensor node needing about 5µW of power should be within 120km of a 150kW transmitter in other to use the energy of the transmitters [5].
In other research work, Abhishek [9] used the clustering method to reduce energy consumption and prolong battery life. Here, the sensor nodes are arranged in geographical clusters, and each node is made to send data to a cluster head, while the cluster head sends to the base station, by sending data to geographically near destination, less power is consumed. Even though the cluster head consumes more power, results show that the average energy consumption in the entire network is reduced.
Another part of the wireless sensor node which has also been an issue of concern is the security of sensor data and the security of the entire IoT system. On this area of the WSN, Sharma [10] proposed an approach of key distribution in the authentication of sensor nodes. Here the nodes are logically arranged in the form of a binary tree, and key authentication is only done by the parent node. This achieved good authentication as it is distributed and not centralised, however it is not as reliable when a node leaves the network.
These and many more ways are some of the different ways and efforts already put by researchers to improve the performance and reliability of the IoT system. Among many things, a wireless sensor node needs to consume as less power as possible to ensure the durability of the solution. At the same time, It needs to have wide coverage in connectivity so as to help the deployment of solutions in remote areas.
An innovative solution known as LORAWAN was introduced by Lora alliance to address this low power consumption and wide network coverage problem [11-13]. LORAWAN stands for Long Range Wide Area Network, it is a communication protocol which uses chirp spread spectrum modulation to send data across long range and at low power [14]. It also defines the system architecture of the network. The communication protocol and the system architecture have a huge determinant in the power consumption of a node. Hence the LORAWAN technology has the advantage of low power consumption, wide-coverage, network security and robustness. It has the capability of sending data to distances up to 10km [14].
This project design is based on a wireless sensor node using the LORAWAN protocol at 868MHz. Antennas are one of the most sensitive parts of the WSN hardware circuit. The rising desire to miniaturize electronic devices into small and compact form has grown the need to have on-PCB microstrip antennas for wireless sensor nodes[15]. Generally, it is a challenging task to design small-sized printed antennas that work below the 1GHz, this is due to the longer wavelengths when compared to the GHz bands[16].
Researchers have explored different ways to achieve small-sized antennas. A widely studied method as demonstrated in [17-20] is the use of metamaterials in antenna design. Metamaterials are artificially structured materials with novel physical properties such as negative permittivity, negative permeability
and negative refractive index[21]. These unusual properties enable its application in antenna size reduction when used in design. However, they have higher power loss, hence may not be most suitable for low power consumption goals, as applicable with the WSN [22, 23].
A more common approach is the use of microstrip antennas. They have lower cost, lighter weight and easier to manufacture using regular fabrication techniques[24]. The popular patch antenna topology was used by [24, 25] for the LORA hardware circuit. [15] proposed a planar topology, where the size was reduced by shortening the radiating part of the antenna to the ground plane. The resultant design was reduced in size with dimensions of 31mm by 13mm, although at the expense of the antenna gain.
[26]presented a miniaturised design inspired by the inverted-F topology, and also for the LORA hardware circuit. On this design, the antenna is made to have the shape of the university logo, thus giving it some aesthetic benefits while being reduced in size. Another widely studied topology is the planar inverted-F antenna (PIFA) as its regular size is quarter of a wavelength. [27-31] have all studied and presented the PIFA topology in different forms for use with the LORAWAN 868MHz frequency.
The antenna design of this project is based on the inverted-F topology. The design is simulated with ADS software and fabricated in the lab. The simulated and measured results are presented and compared with previous design projects based on similar topic. Performance reference is made to two of the similar designs as done in [27].
1.1 Project aim and scope
This project tries to improve on the reliability of a wireless sensor node through the design of an on-PCB inverted-F antenna. The project includes the following tasks:
● Literature survey on IoT system, wireless sensor nodes, and antenna design
● Background study on antenna design for wireless sensor nodes
● Inverted-F Antenna design and simulation using ADS software
● Design of a wireless sensor node using a reference design
● PCB design of the wireless using OrCAD
● Fabrication and testing of the PCB antenna.
2 Background
This chapter gives a comprehensive background knowledge for the topic of this project. A background summary of both antennas and wireless sensor systems are provided on this chapter.
2.1 Antenna Theory
In the modern wireless systems, the antenna is one of the most critical components. A good antenna design can improve the overall system performance and relax the other system requirements.
An antenna, as defined in the IEEE set of standard definitions, is “a means of radiating or receiving radio waves” [32]. Among other definitions, an antenna can also be defined as a dual-mode transducer that, in transmitting mode, transforms the guided signals in a transmission line into radiated free-space waves or transforms back the radio waves into guided signals, in the receiving mode [33, 34].
The radiation is achieved due to the oscillation of charges in the conductor. In a typical dipole, as shown in Figure 1, the conductor length equals half wavelength (HWL) of the transmitting signal, and the traveling current wave creates a standing wave at the terminal with maximum current in the centre and minimum at the edges. The frequency is called resonance frequency, as maximum power is radiated from the antenna.
Figure 1. A typical λ/2 dipole configuration and the current distribution across the antenna length.
The history of antenna dates to 1886, when Professor Heinrich Rudolph Hertz assembled the apparatus that later became known as a “radio system.” It had a half-wave dipole as the transmitter, and a resonant square loop placed nearby to receive. Furthermore, it was not until 1901, when Gulielmo Marconni claimed to successfully be able to send signals over large distances. Specifically, receiving radio signals at St. John’s, Newfoundland, sent across Atlantic from Poldhu in Cornwall, England [35].
The antenna technology was primarily focused on wire-based radiating elements until 1940s, and around the same time as World War 2, modern antenna technology was launched and new elements such as, waveguide apertures, horns, reflectors, etc were introduced [34].
2.2 Antenna Parameters
The antenna performance can be defined by a set of various parameters. Some of the parameters are interrelated which is discussed in the following sections. So not all of them need to be specified to describe an antenna’s performance.
2.2.1 Radiation Pattern
An antenna radiation pattern or antenna pattern is defined as a graphical representation of the radiation properties of the antenna in spatial coordinates [32]. Radiation properties include radiation intensity, field strength, phase and polarization. Usually, the radiation pattern is determined in the far-field region, and plotted on a logarithmic scale as decibels dB [33].
Figure 2. Radiation pattern of a directional antenna.
As shown in Figure 2, the pattern may have several maxima which are called lobes. The lobe with the maximum value is called the main lobe or main beam, while lobes with lower values are side lobes [36].
2.2.2 Directivity
A fundamental parameter of an antenna is its ability to focus the radiated or received power in a particular direction. Antennas with constant radiation pattern in the azimuthal plane are called omnidirectional as they can radiate and receive signals in all directions equally and are used in broadcast and handheld applications. Applications such as radar or point to point radio links use pencil beam antennas with narrow and directional beams [36].
Directivity is a measure of the antenna’s focusing ability and is defined as the ratio of radiation intensity in a given direction U to the radiation intensity averaged over all directions Uavg [32]. The average radiation intensity is total power radiated from the antenna Prad divided by 4π.
Mathematically, it can be expressed as,
D = UU
avg = 4πU
Prad [33] (1)
2.2.3 Efficiency and Gain
Antenna efficiency, as with many other electrical components, is a ratio of radiated power to the total power fed into the antenna.
η = PPrad
in (2)
Resistive losses are present in all practical antennas due to imperfect conductors and dielectric materials.
External factors such as impedance mismatch at the feeding terminal or polarization mismatch at the receiving terminal can also contribute to the effective power loss. However, they are not attributed to the antenna itself as inherent dissipative losses.
The gain of the antenna is closely related to the directivity. Since directivity only describes the directional properties of the antenna, Gain takes the efficiency of the antenna into account, and can be defined as a product of Antenna efficiency and Directivity,
G= η*D (3)
Thus, the Gain can only be less than or equal to Directivity and is usually expressed in dB. Reflection losses due to impedance mismatch are not included in this measure, the Gain, including this effect is referred to as Realized Gain [36].
2.2.4 Bandwidth
The bandwidth of an antenna is defined as the range of frequencies within which the performance of the antenna, with respect to some characteristic, conforms to a specified standard [32]. It is a range of frequencies on either side of the centre frequency (resonance frequency), where the antenna performance characteristics such as radiation pattern, gain, polarization, impedance, etc. are close to the acceptable values as those at the centre frequency.
Bandwidth can be expressed as a ratio of upper to lower frequencies of the operable band for wideband/broadband antennas, or as a percentage of the difference from centre frequency for narrowband antennas [33].
2.2.5 Input Impedance and Return Loss
The impedance present at the terminals of the antenna is called the input impedance [32]. At the resonance, if the input impedance does not match with the impedance of the rest of the system, it creates the return loss, which is caused by the reflection of the incident signal from the antenna terminal. The return loss, interchangeably used with the term S11, is the logarithmic expression of reflection coefficient L, which is given by [33].
Γ = 𝑍𝑍𝑖𝑛−𝑍𝑜
𝑖𝑛+𝑍𝑜 (4)
Where Zin is the impedance at the antenna terminal and Zo is the characteristic impedance of the system.
Another measure of the impedance mismatch, Standing Wave Ratio (SWR), also sometimes known as Voltage Standing Wave Ratio (VSWR), is given as,
SWR = 1+Γ1−Γ (5)
A return loss of >10dB (or <-10dB as S11 is just the negative of return loss) is widely considered good enough for antenna performance. Table 2.1 compares the transferred power (in percentage) to the corresponding return loss figure.
Table 2.1 A Comparison of transferred power to corresponding return loss [37]
Return Loss dB Power transferred %
20 99
10 90
3 50
1 21
2.3 Impedance matching
The basic idea of impedance matching is to place a matching network between the transmission line and the load(antenna) to minimize the return loss which is discussed in the previous section. Ideally, the matching network has to be lossless, to avoid unnecessary power loss, and is designed to get the characteristic impedance Zo when looking into the network. This procedure is also referred to as tuning.
Impedance matching or tuning is an important design step for maximum transfer of power to the load (assuming generator is matched to the line) and the improvement of signal to noise ratio of the system.
Any impedance Z = R ± j X can be matched to the characteristic impedance using reactive elements.
Various factors are considered while designing a matching network, such as complexity: The simplest design is always preferred. However, it may not be the best solution for a situation where the network has to function as a specific filtering circuit as well. Bandwidth: In some situations, a network is required to match over a band of frequencies. While this is achievable, it adds to the complexity of the network.
Implementation: A tuning stub, a fraction of the transmission line attached to the main transmission line, could be more suitable for a particular design compared to the lumped components i.e. inductors and capacitors [36, 37].
2.3.1.1 Matching with lumped elements
The simplest network comprising of two reactive components, arranged in such a way that one component is serially connected while the other is shunted, is called the L-network. The values of inductors or capacitors and their arrangement in different network topologies determine the impedance change. The effects of components are illustrated in Figure 3 and Figure 4 with the help of the smith chart.
Smith chart is a visualization tool developed by P. Smith in 1939 to understand transmission line phenomenon without the need for detailed and complex numerical calculations. It is based on the polar plot of the reflection coefficient Γ, but the real utility is being able to convert the reflection coefficient into normalized impedance (or admittance) hence to be able to characterize a transmission line [36].
Serially connected components move the impedance along the constant resistance circles (red). Inductors move the impedance clockwise while capacitors move the impedance anti-clockwise on the chart, the reactances XL and XC are deducted from the expressions below,
L = 2πf𝑋𝐿 (6)
C = 2πf𝑋1
𝐶 (7)
Figure 3. Impedance change with series inductor(clockwise) and series capacitor(anti-clockwise)
Shunt components move the admittance along the constant conductance circles (blue). Inductors move the admittance anti-clockwise while capacitor moves it clockwise, the susceptances YL and YC are deducted from the expressions below,
L = 1
2πf𝑌𝐿 (8)
C = 2πf𝑌𝐶 (9)
Figure 4. Admittance change with shunt inductor (anti-clockwise) and shunt capacitor (clockwise)
In practice, the final achieved impedance is typically not known during the design step. Therefore, provision is made in the design, for a three-component network such as a pi-network or a T-network which can be used to implement any suitable topologies during implementation [37].
2.3.1.2 Stub tuning
A stub is a short/open-circuited piece of transmission line connected either in series or shunt with the transmission at a certain distance from the load. It is a popular matching technique due to its convenience in fabrication as part of the transmission line. Lumped components are avoided as the open-circuited stubs behave as capacitors, and short-circuited stubs behave as inductors.
For microstrip lines, the shunt stub is preferred, while the series stub is preferred for slotted lines or waveguides [36].
However, stub tuning is dependent on the frequency of the system. At lower frequencies, stubs could be large enough to become detrimental to the overall system performance by influencing other traces being in proximity and contributing to the overall spurious emissions.
2.4 Antenna Electrical Size
Antennas with dimensions much less than the wavelength are subjected to limitations that are the same for a capacitor used as an electric dipole and an inductor(loop) used as a magnetic dipole, given that they occupy equal volumes [38].
Figure 5 shows a small electrical antenna which has dimensions within the radian sphere, where radian length a is given by,
a = 2πλ (10)
Figure 5. Illustration of antenna dimensions within the radian sphere
A lossless antenna is expected to radiate the maximum energy received from the source, independent of the size. A small antenna exhibits less dissipative loss due to the smaller resistance by virtue of its size.
However, less resistance delivers less voltage for the same amount of available power, thus reducing the efficiency of the antenna significantly.
Another aspect of small antennas is the Q-factor, which is addressed by various researchers. The small antenna requires a high Q-factor to deliver its power. This affects the bandwidth profile of the antenna as Q inversely relates to bandwidth [39-41].
2.5 Monopole Antennas
An antenna with a quarter-wavelength (QWL) of radiating wire or trace positioned over a ground plane is called a monopole antenna. Most PCB based antenna designs use the QWL monopole configuration due to the pre-existing ground plane. Figure 6 shows the basic structure of monopole antennas
Figure 6. Monopole antenna signal feed and image creation
2.5.1 Effect of ground plane
Unlike a dipole, the signal is fed to a single element and grounded to a conducting plane below (Figure 6).
The ground plane behaves as an electromagnetic mirror and creates an image of the monopole, making the antenna a virtual dipole. The performance of the antenna is that of a half dipole, where the gain of the antenna is doubled since all of the energy is radiated into one hemisphere.
Generally, the ground plane should be large. However, this is not a design freedom as it is one of the decisive factors influencing the antenna’s bandwidth, radiation, and efficiency[42].
2.6 Microstrip Antennas
Microstrip antennas have become widely used in recent years for high-performance applications where the major constraints are size, weight, cost of manufacturing, ease of implementation, etc. These antennas are low profile, conformable to planar and non-planar surfaces, inexpensive to manufacture using modern printed circuit technology, and depending on the shape of the patch, they can be versatile in terms of resonant frequency, impedance, pattern, etc. [33, 43].
The electrical performance of the basic microstrip antennas suffers from significant drawbacks including, narrow bandwidth, high feeding line losses, and low power handling capacity. However, with the progress in theory and technology, some of the drawbacks have been overcome or minimized.
The radiating element and the feed line are fabricated on a dielectric substrate. Microstrip antennas are also often referred to as patch antennas [43].
2.7 Miniaturization techniques
Antenna miniaturization is the process of reducing antenna dimensions while maintaining reasonable level of original radiation characteristics. This is not easily achieved, and compromises are made between size reduction and antenna performance. But throughout the years, a number of creative approaches have been developed, such as loops, monopoles, slots, and various integrated antennas.
The miniaturization involves changing the geometry of the structure, adding additional components, or altering the material characteristics. Miniaturized antennas are different from the small antennas which are compromised in most if not all the usual performance characteristics. Although the performance characteristic need not be as good as the original antenna, it must be better than that of the unmodified antenna reduced to the same size [33, 44]. Basic designs are described below,
2.7.1 Impedance loading
Among the simplest designs to miniaturize monopoles and dipoles is the addition of lumped impedance elements along the antenna length. By inserting an inductor coil of a given Q, the antenna can be made short. However, the integration of the coil affects the antenna performance, decreasing the radiation resistance and efficiency.
Another design is to use a shorting strip for the microstrip/patch antennas as a form of impedance loading.
This can lead to the length of QWL instead of HWL, a reduction factor of 2. This method is utilized in the design of IFA and PIFA.
The trade-offs due to miniaturization by the impedance loading method are illustrated through quantitative examples [44].
2.7.2 Materials loading
In the 1970s, researchers explored the possibility of miniaturizing monopole and dipole antennas by surrounding them with dielectric materials. The effects of material type and the surrounding volume can lead to improved performance characteristics for the miniaturized antenna [44].
2.7.3 Folding
Another miniaturization technique involves folding of the monopole and dipole arms into more compact structures to reduce the overall size. This can be achieved by keeping roughly the same length for the current path and maintaining the resonance, while shortening the length of the antenna. Typically, the radiation resistance, bandwidth, and efficiency are reduced. Researchers have proposed different folding patterns such as zig-zag antennas or meandered dipole antennas, both in two dimensions and three dimensions and presented the resulting trade-offs between size reduction and antenna performance [44].
3 Design and Methodology
This chapter highlights the steps taken in the design of the inverted-F antenna.
3.1 Inverted-F Antenna Design
Inverted F Antenna (IFA) is a type of quarter-wavelength monopole antenna, having the radiating arm/wire bent horizontally to be parallel to the ground plane. This structure makes IFA or L antenna more conformal to tighter design space than usual monopoles.
Despite the reduced physical structure, the 3D IFA designs still have significant space volume that may sometimes not be fit for small applications like handheld wireless devices [45, 46]. Printed antennas are the most popular choices in wireless and mobile applications due to their compact size, more conformity to design space restrictions and ease of manufacturing[47-49].
3.1.1 Operating mechanism
Matching blocks such as pi or L-filters are placed at the feeding points of the antennas to achieve the impedance matching. A single lumped component, such as a shunt inductor or capacitor, can be used to impedance-match the L antenna. However, using a shorting pin to the ground at the feeding point transforms it into the IFA. The antenna shape looks like the letter F, hence the name Inverted F.
The shorting pin behaves as an intrinsic shunt inductor, which makes it possible to impedance-match the antenna only by adjusting the distance between the feeding point and the shorting pin. The open end of the antenna contributes capacitive reactance to the total impedance since it is parallel to the ground plane and separated by a dielectric. If the separation between the antenna arm and the ground plane becomes small, the shorting pin should be placed closer to the feeding point to compensate for the impedance change[42].
3.1.2 Design steps
The resonant frequency of the IFA is determined by the length of its printed radiating arm, which is the quarter-wavelength of the central frequency Fc. For our application,
Fc = 868MHz
The antenna is printed on the RF community favourite, FR4 substrate with 1.5mm thickness, and a typical dielectric constant ϵr = 4.6.
The wavelength(λ) in free space at Fc = 868MHz,
λ = cf = 345mm (11)
The effective wavelength traveling through the printed antenna on FR4 is approximately given as,
λeff = f√ϵc
r ~ 0.6*λ = 207mm (12)
The approximate length of the printed antenna arm is,
l = λeff4 = 51.75mm (13)
The antenna design is inspired by studying the documents mentioned above [45-50] and the reference design [51]. Total antenna dimensions are fixed at 80x10 mm2 at one end of the PCB with an 80x130 mm2 ground plane (Figure 8). Other dimensions are iteratively chosen to be (Figure 7),
Radiating arm length LR = 59mm Radiating arm width WR = 2.5mm Ground/radiator separation dR = 7mm Ground/feed pin separation dF = 7mm
Figure 7. Labelled antenna dimensions.
Figure 8. Layout figure of the designed antenna.
The simulations are performed in ADS Momentum with MoM-microwave setup.
3.1.3 Results from ADS
A two-layer FR4 substrate was defined for the layout, as shown in Figure 9. The conducting layers cond, cond2, and via were defined as copper, and the dimensions were defined as,
h = 1.5mm t = 0.035mm
Via diameter d = 0.6mm
Figure 9. Defined substrate in ADS for the layout.
The S11 results for a frequency sweep from 840MHz to 900MHz, as seen in Figure 10 and Figure 11, show excellent resonance at the desired frequency with the peak value of -43dB at exactly 868MHz.
Figure 10. Simulated Return Loss of designed antenna.
Figure 11. Simulated Impedance of designed antenna.
The simulated parameters of the antenna at 868MHz are shown in Table 3.1, and the corresponding radiation patterns in XZ and YZ planes are shown in Figure 12.
Table 3.1 Simulated Antenna Parameters of the designed antenna
Figure 12. Simulated radiation patterns of designed antenna; YZ-plane(left) and XZ-plane(right).
3.1.4 Design Optimization
The following are the iterative steps and observations to reach the optimum result after achieving resonance,
1- The width of the antenna arm contributes to the capacitive reactance. It was increased from 1mm to 2.5mm to make an optimal match.
2- The distance between the feeding pin and the shorting pin influences the impedance as well as the resonance. It was decreased from 8mm to 7mm to get the optimal match.
3- The separation between the ground plane and the antenna arm influences the impedance and the sensitivity of the antenna. In a trade-off between sensitivity and impedance match, the separation is neither too small nor too large. The antenna arm was moved 1mm closer to the ground plane to get better impedance matching without antenna getting coupled to the ground plane.
3.1.5 Conclusions
The simulation results are comparable to the results presented in the papers [47-52]. An additional feature achieved is an approximately 130MHz bandwidth under the -10dB mark, though it does not affect or contribute to our design. The design is acceptable for fabrication and to be tested for real performance.
4 Wireless Sensor Network
Wireless sensor network is a wireless interconnection of sensor nodes. They are a group of sensor nodes interconnected with each other and a central base station, in other to monitor physical conditions such as pressure, motion, temperature, humidity, etc. The sensed information from these sensors are sent wirelessly to the base station where they can be studied and analysed. The processed data from these networks can aid individuals and businesses to make vital decisions and conclusions.
4.1 Wireless Sensor Network Types
The different sensor nodes in a wireless sensor network can be interconnected in different ways. These variations in connection pattern give rise to the different topologies of the wireless sensor network.
4.1.1 Star network topology
This is the type of connection where all the sensor nodes are connected directly to the base station. In this type of topology, the individual sensor nodes do not send data to each other. The communications are done between each node and the base station. Since communication is only done to the base station, this makes the sensor nodes to consume less power. Therefore, low power consumption is an advantage of this network topology. The latency of this topology is also low since the communication to the base station is direct. For this topology to be effective, the base station needs to be within the radio transmission range of all the sensor nodes in the network. Figure 13 shows a representation of node connections in the star topology.
Figure 13. Star connection
4.1.2 Mesh network topology
This topology allows communication between different nodes. Therefore, a node can send data to another node, which in turn sends the data to the destination; this method of communication is called multi-hop communication. The multi-hop style of communication allows a node to be located farther away from the base station. Therefore, it is easily scalable and can cover wider range of space. Since each node can receive data from multiple nodes, it increases the power consumption of the sensor nodes. The latency is also higher as the communication process is longer. Figure 14 shows a representation of node connections in the mesh topology.
Figure 14.Mesh topology
4.1.3 Hybrid Star-Mesh network topology
This topology combines the advantages of the star and the mesh topology. In this topology, only dedicated nodes in the network can receive data from other nodes. Since these nodes consume more power, they can be powered through the mains while the rest of the nodes are battery powered. The low power consumption of the star method is achieved here, and the wider reach of the mesh topology is also achieved. Figure 15 shows a representation of node connections in the hybrid star-mesh topology.
Figure 15. Hybrid network
4.2 Wireless Sensor Node Hardware Design
A hardware circuit for the wireless sensor node is required together with the antenna for a fully functional wireless sensor node. This section gives a description of the design procedure for the hardware circuit needed for the wireless sensor node. The software programming aspect of the design is not covered in this text. Generally, the emissions from the circuit components, the trace routing through the ground plane, and the proximity of components to the antenna all have the potentials to affect the antenna performance. While the antenna gives good results from the ADS simulations and fabrication, a more viable evaluation is the testing carried out on a printed circuit board with the real ground plane of the circuit. This represents a closer scenario to the reality of which the antenna would be used. The sensor employed for this purpose is a temperature/humidity sensor. The primary aim of this sensor node is to sense the temperature and humidity of the surroundings and send this information to the base station.
The hardware circuit for the sensor node contains circuit designs that enable adequate operation of the sensor, the processing of the information, modulation of the data, and the transmission of the information to the destination. This is achieved by different sections of the node. The basic requirements of a WSN circuit include low cost, good performance, and low power consumption[52]. The circuit is divided into three sub-circuits: The sensor circuit, the microcontroller circuit, and the transceiver circuit. Reference designs are used as starting points where applicable. In practice, a power circuit is needed to ensure proper power management and for the durability of the product. However, for the purpose of lab experiments, the node is powered by two pcs of 1.5V AA batteries connected in series to supply 3V to the circuit.
4.2.1 Sensor Circuit
The sensor device used for this circuit is the SHT31-DIS F2.5KS temperature and humidity sensor manufactured by SENSIRION. It measures the temperature and humidity of the surroundings and sends the data digitally to the microcontroller. It uses the I2C data communication protocol to send and receive data from the microcontroller. It comes in an 8-pin SMD package, as shown in Figure 16. To correctly set up the circuit for this sensor, the functionalities of each of the pins need to be well understood. The pin functions as described by the manufacturer are summarised below:
Figure 16. Pin layout of SHT31-DIS temperature sensor
Pin 1 (SDA) – This is the serial data input/output pin. It carries the digital signal containing values of the measured temperature and humidity. Therefore, it is connected to the microcontroller’s input/output pin where the data is received.
Pin 2 (ADDR) – This pin enables the address configuration of the sensor. It is either connected to a logic high or a logic low. Each of the two connections represents a specific address. Hence this enables the sensor to be used alongside another sensor device on the same bus. For the sensor circuit of this project, the ADDR pin is connected to logic low (GND), this is also taken note of in the microcontroller programming.
Pin 3 (ALERT) – This pin is used to trigger an alert to the microcontroller when the pre-programmed limits of the temperature and humidity readings are exceeded. It is connected to the interrupt pin of the microcontroller. On the sensor circuit, this pin is connected to the base of a transistor, with an LED on the collector of the transistor. An alert signal on the transistor base turns the transistor on and completes the LED circuit to be turned on. This gives a visual signal to the user when the temperature limits are being exceeded. The alert signal is still connected to the microcontroller interrupt pin from the transistor collector. Hence, while the LED shows that an alert has been triggered, the microcontroller also receives the signal for the corresponding action.
Pin 4 (SCL) – This receives the serial clock of the I2C communication protocol. It synchronizes the communication between the sensor device and the microcontroller. It can receive clock frequencies up to 1000KHz. This pin is connected to one of the I2C clock output pins of the microcontroller.
Pin 5 (VDD) – This is the input supply voltage of the device. It is expected to work between the voltages of 2.15V and 5.5V, with a typical value of 3.3V. 3V from the batteries is connected to this pin. To filter out noise, this pin is decoupled with a 100nF capacitor placed close to the pin.
Pin 6 (nRESET) – This is used to receive a reset signal to reset the sensor device. It is internally connected to VDD through a pull-up resistor. Therefore, pulling this pin low generates a hard reset for the device.
Pin 7 (R) – This pin does not have an electrical function. It serves for the mechanical balance of the package. It is connected to ground.
Pin 8 (VSS) – This is the ground pin of the device, and is connected to the circuit GND.
Pull-up Resistors.
The SDA and SCL lines of the device are open drain. Hence pull-up resistors are needed to enable the digital communication. The resistors are connected between the pins and the VDD. The value of the resistors is chosen to be in between the maximum and minimum values, as indicated on the respective graphs from the datasheet. 10k pull-up resistors are used on this circuit. Figure 17 shows the basic connections of the sensor circuit.
Figure 17. Basic connections of the sensor circuit
4.2.2 Microcontroller Circuit
The microcontroller is responsible for the reception of data from the temperature/humidity sensor, the processing of this data and the communication of processed data to the transceiver for the RF transmission. The microcontroller circuit enables the microcontroller to perform its function. The STM32L073RZT6 32-bits microcontroller was used, in a 64-pin LQFP-64 package. A reference design from an evaluation board was used in the circuit configuration. The reference design is configured to the purpose of this project. The STM32 Nucleo-64 reference design was used. Below steps were taken in the configuration of this reference design for the project goal.
• Power Supply – The circuit is powered externally by two pcs of 1.5V AA batteries. Therefore, the voltage regulator circuit of the reference design is by-passed, and the 3V from the batteries is supplied directly to the supply pins of the microcontroller. The supply pins VDD include (Pin 1, Pin 13, Pin 19, Pin 32, Pin 48, Pin 64). The analog supply pin AVDD is fed through a noise filter comprising of inductor bead and a capacitor. Other supply pins are protected from noise through decoupling capacitors connected to each pin and placed as close as possible to the pins.
• Oscillators – External oscillators are used to generate the clock signal of the circuit. An 8MHz crystal is used for the HSE, while a 32.7KHz crystal is used for the LSE. The load capacitors of the oscillator circuit are chosen according to the specified load capacitance of the crystal and assumed parasitic capacitance of 4pF.
𝐶𝐿= 𝐶𝐿1𝑥 𝐶𝐿2
𝐶𝐿1+ 𝐶𝐿2+ 𝐶𝑆 (14)
• Temperature Sensor communication – The MCU communicates with the sensor device through an I2C communication protocol. The clock signal for the synchronisation is also provided by the MCU. The MCU has several I2C communication pins, one of these is used. Pin 62, PB9 is used for the SDA of the I2C communication, while pin 61, PB8 is used for the corresponding clock signal.
• Transceiver communication – The transceiver uses an SPI communication protocol. This protocol uses 4 pins for communication, the MISO, the MOSI, the NSS, and the SPI clock. The MCU also has assigned pins that can be used for this type of communication. Pin 23 PA7 is used as the MOSI, pin 22 PA6 is used for the MISO, pin 58 PB6 is used for the NSS function, while pin 21 PA5 is used for the SPI clock signal. The transceiver and the MCU use other I/O pins for digital communication.
Five of the MCU I/O pins are assigned for this purpose. These include PB3, PB4, PB5, and PA10, corresponding to Pins 55, 56, 57, and pin 43, respectively. This digital communication enables control of the transmission and receiving processes. It involves commands like TXReady, RXReady,
Timeout, etc. For switching between transmitting mode and receiving mode of the transceiver, the MCU uses pin 9 PC1. Pin 14 PA0 is used to receive interrupt signals from the transceiver for the reset function.
• Programming and Debug interface – This is used to establish a connection to the MCU for software programming. The pins are assigned appropriate functions during programming. PA13 and PA14 are used for data communication and clock synchronisation, respectively. External reset option is provided through pin 7 NRST. These pins are extended through a header connector. The interrupt signal from the alert pin of the sensor is connected to pin 54 of the MCU, PD2. Pin 59 PB7 is used for EVENT_OUT and connected to the reset pin of the sensor. Figure 18 shows the basic connections of the microcontroller circuit, as described above.
Figure 18. Basic connections of the microcontroller circuit
4.2.3 Transceiver Circuit
The transceiver circuit transmits the modulated data through the transmission lines and antenna. It is also responsible for receiving and demodulating RF signals. This function in a sensor node has been achieved by designers with solutions like zigbee, xbee, Sigfox etc [53]. For this circuit, the LORA transceiver chip is
used. Among other things, it has the advantage of low power consumption and longer range coverage, which is crucial in a wireless sensor node[24].
The circuit uses an external clock source with an oscillator circuit of 32MHz frequency. The radio frequency signal is transmitted out through the RFO pin, while the received signal from the antenna goes into the chip through the RFI pin. Therefore, the transmission lines from these two pins are fitted with appropriate filter circuits to remove unwanted frequencies. An RF switch is used to switch between receiving mode and transmitting mode. The switch is controlled through a logic inverter. A matching network is fitted at the end of the line to match the line impedance to the antenna and ensure minimal reflection.
• Microcontroller communication – The transceiver communicates with the microcontroller through its SPI interface and digital I/O pins. The processed sensor data is relayed to the transceiver through the SPI pins connected to pins 17,18,19 and 20 of the transceiver. The five digital I/O pins send status commands to the respective pins of the microcontroller. Figure 19 shows the basic connections of the transceiver circuit.
Figure 19. Basic connections of the transceiver circuit.
4.3 PCB layout Design
The PCB layout design is an important aspect of the WSN circuit. The efficiency of an electronic circuit
The entire WSN circuit contains digital signal traces, analog signals from the power traces and radio frequency signals. It is, therefore a mixed-signal circuit. Mixed-signal PCBs are prone to a number of technical issues if not properly routed. Some of these issues include interference, cross talk, ground bounce, etc. This is in addition to the RF technical constraints like return loss, propagation delay, characteristic impedance, etc. In other to achieve optimum performance in a mixed-signal circuit and with low power consumption, [54] noted a few design targets to have in mind. Some of these are as follows:
• The noise generated from the board should be minimal
• Crosstalk between adjacent traces should be reduced
• The ground bounce effect should be reduced. This is caused by varying ground potentials
• Correct signal line termination should be provided
It is always recommended to keep the analog signal return path separate from the digital signal return path. This is to ensure that the interference is reduced. This is either achieved through split ground planes or a study of the signals return path for good placements of the circuit components. Since this project is limited to a 2-layer board, the placement of the circuit blocks was done with the current paths in mind.
This is to ensure that the digital signals of the MCU and sensor traces do not interfere with the RF signals of the transceiver. The antenna is also kept away from the digital signals.
Figure 20shows the physical layout arrangement of the circuit blocks that make up the wireless sensor node.
Figure 20. Board layout Plan TEMPERATURE
SENSOR CIRCUIT
ANTENNA
MICROCONTROLLER CIRCUIT
TRANSCEIVER CIRCUIT
POWER SUPPLY
4.3.1 Sensor circuit layout
The sensor circuit has smaller number of components. Hence the PCB layout poses little challenge. It also has only two high-speed digital pins, which are the I2C communication pins SDA and SCL. The alert pin is digital; however, it is not high speed as alerts do not trigger all the time. The sensor device is oriented such that the side which contains the SDA and SCL pins face the microcontroller. This is done to ensure shorter trace lengths between the sensor device and the MCU. The decoupling capacitors are placed as close as possible to the power pins. The trace for the power supply VDD is kept away from the traces for SCL and SDA.
4.3.2 Microcontroller circuit layout
The microcontroller circuit has higher number of components, comprising of both a higher number of high-speed digital traces and analog traces. The microcontroller is also oriented such that the pins connected to the sensor are faced towards the sensor. Since the pins connecting to the transceiver do not occupy just one side of the MCU package, the shortest trace paths from those pins to the transceiver are chosen. The MCU circuit also contains two crystal oscillators, the 8MHz and the 32.7KHz crystals. These crystals are sources of high-frequency signals, hence are placed away from other components. The two push-button switches B1 and B2 are placed further away from the MCU since their operations are infrequent. This is done to accommodate components like the decoupling capacitors which are meant to be placed close to the respective pins. High-speed traces which include the SPI communication traces, the oscillator circuit traces are well spaced out while routing. This is to reduce the effect of crosstalk. Crosstalk is reduced by making the spaces between traces about 3 times or 4 times the width of trace[55]. The chosen trace width on this design is 12mil. Hence the spacing between the high-speed traces and adjacent traces are a minimum of 24mil. The clock traces for the SPI and I2C communication are kept away from noisy traces.
4.3.3 Transceiver circuit layout
The transceiver circuit contains both digital signal traces and RF signal traces. One side of the transceiver chip is dedicated to the RF traces RFO and RFI, these stand for Radio Frequency Output and Radio Frequency Input respectively. The decoupling capacitors are placed very close to the power input pins of the chip. The antenna is well away from the rest of the circuit with the matching network placed close to it. The width of the transmission line traces is chosen to maintain a 50Ω characteristic impedance. This is considered together with the spacing from the ground plane. Values for the trace width and clearance from the ground plane are obtained from ADS software as 1.41mm and 0.3mm respectively. These dimensions are maintained all through the trace length to avoid variations in characteristic impedance
which would cause reflection. It is a good practice to avoid stubs on the transmission line while connecting filter components or other components on it [37]. The pads for the components are preferably placed on the transmission line. Figure 21 shows an illustration of a preferred component placement on the transmission line.
Figure 21. Component placement on transmission lines
4.3.4 Overall layout
While the placement of components is crucial, the routing of traces is also important. In a standard mixed- signal PCB design, one or two layers are dedicated to the signals, while the power and ground have dedicated planes and layers. In such layout, the performance is improved as the signals are run on strip line traces. This is also good for EMI standards since most of the electric and magnetic fields are contained within the planes above and below and are therefore not radiated outside[55]. Trace paths are routed to avoid big loops that would increase the inductance and emission. Due to the limitations by the available milling machine, and the border closures at the time of this report, only 2-layer PCB can be made for this circuit. Therefore, both the top and bottom parts of the board are used for signals and ground. The RF signals are only restricted to the top layer. Since it is a 2-layer board, only microstrip lines are used all through. A power plane is always good for good performance. On this project, due to the limitations, power plane was not used. Power is supplied at a central position on the PCB and is routed in a star configuration to the respective pins. On the RF circuit, the power for the transceiver chip is traced straight from the power supply, while the analogue power for the RF switch and logic inverter is also separately traced from the power supply. This is to ensure that the RF power does not introduce noise to the analogue power. The power trace widths are bigger than the signal trace width so as to reduce resistance.
The power trace widths are 0.5mm while the signal trace widths are 0.3 mm. The milling machine available can drill a minimum of 0.6mm via diameter. Therefore, the used Vias have diameters of 0.6mm for signals and 0.8mm for the power traces. It is a good design practice to keep buses the same length, example the SPI interface. This is particularly important for differential pair signals. The SPI interface does not have a differential pair; on this circuit, it has only one master and one slave. Therefore, the difference in length
