Project topics and materials
Get Complete Project Material File(s) Now! »
Proposed topologies for SI cancellation
Many designs have been proposed recently to implement FD wireless systems. All focus on achieving a high SI cancellation: some designs are based on SI canceling transceiver front-end structure [27]. Other designs are based on dual-polarized micro-strip patch antenna structures, complemented with an active analog cancellation network [28]. Other designs are based on an electrical balance duplexer circuit, which connects with a single port antenna [29]. The most significant ones are detailed below:
1) Antenna separation: This design is based on a variable phase shifter and a variable attenuator that are added between the two antennas (Figure 1.9). This path intends to provide a 180° phase difference and an identical signal level of the SI transmitted signal coupled at the receiver, in order to cancel it up at the receiver antenna [30]. In [30], a cancellation of about 33dB is achieved over a 100MHz bandwidth. The two antennas are separated from each other by a distance much greater than /2 in order to reduce the coupling level (direct SI) between TX and RX. On the other hand, this distance is not preferable for compact devices especially at lower operating frequency where the wavelength will be large.
2) Antenna placement: Two TX and one RX antennas are structured in a way to force the TX signals to cancel each other at the receiver position (Figure 1.10). The first antenna is set at a distance d from the receiver and the second antenna is at distance d + /2 from the receiver ( is the wavelength of the operational frequency). As a result, there will be a 180° phase difference between them, which results in a destructive interference at the RX level. In theory, a null is created at the position of the RX if the leakage power magnitude level from the pair of TXs are identical which is impossible to obtain in practical implementation. Thus, this method can only reduce the SI from about 30dB and with the help of digital cancelation it can reach 60dB at least but remains intrinsically narrowband [30].
3) MIDU topology: In MIDU (Enabling MIMO Full-Duplex) project, a symmetric placement of antenna is used to cancel SI in a FD system [31]. Antenna cancellation technique is duplicated at both transmit and receive stages. The two identical transmit antennas are separated by a distance d and transmit at the same time but with opposite phase; thus, for the receiver antenna, the potential null point will be located on the perpendicular bisector of the line connecting the two transmitter antennas Figure 1.11.
The same principle is applied at the receiver stage and can potentially eliminate the need of other forms of analog cancellation, thereby avoiding the need for delays and variable attenuator. In [31], the overall isolation is about 45dB over a bandwidth of 625 KHz. It should be pointed out that the radiation patterns of TX antennas shows a null on RX antenna axis and identically for RX antenna pattern on TX axis. This technique is potentially more wideband than the second one.
Antenna specifications for a compact FD system
The future communication generation (5G) and the strong demand for multi-function or multi-standard devices call for the use of wide-band/multi-band antennas. In Internet of Things (IoT), M2M (Machine to Machine) and medical sensors, antennas need to be implemented in small devices to ensure the interaction between the sensors and their environment and to allow a widespread use.
Thus, antennas are required to be small, compact and low-cost. Although the need for wide-band small antennas is spreading, the design and the matching of Small Antennas (SA) or Electrically Small antennas (ESA) is not an easy task especially in wide and low frequency band applications. Indeed, SAs famously suffer from high quality factor Q value since there is a high reactive energy stored in their near field and a small real power is propagated in their far field [32]. In other words, SA and ESA are intensely reactive, which results in a poor efficiency-bandwidth trade-off when using passive impedance matching [33]. This limit can be overcome by adding active elements, such as nonmatching circuit, which are characterized by their negative reactance [33]. We suggest exploring in this PhD the study and design of NF circuits, which will serve as key components to achieve a high SI cancellation between two SAs in order to realize a compact full-duplex system. The Non-Foster circuits will be implemented at different levels of the front-end and a dedicated front-end topology is put forward in the following section.
Motivation, Objectives and Proposed topology
past few years as it can play important role in the coming future 5G communications and IoT devices. However, there exist many challenges for deploying FD, the main one is the high SI cancellation between the TX chain and the RX chain. It was found out that for communication systems, we need around 110dB of SI cancellation, and this cancellation can only be achieved by two levels of SIC at least, i.e., analog and digital cancellation. Analog cancellation focuses on the Radio front-end and can proceed at antenna stage and/or in RF/analog-base band circuits, while digital cancellation is basically an algorithm that deals with all the distortions and noise that the signal faces along the RX chain and in NLOS channels. Many topologies have been recently studied to achieve FD system, where they can offer around 30dB – 47dB of SI cancellation at the analog/RF stage. The amount of SI cancellation needed depends mainly on the system scenario and application; and since our focus is on small devices (e.g. IoT), the need of compact devices is of prime interest. Then, our goal is to design a compact FD front-end based on two small antennas (TX and RX) working at a low frequency range and if possible over a wide frequency range. In that context, the two antennas have to be small and need to be really close to each other, i.e. much more closer than in previous FD front-end studies. It means that none of the aforementioned approaches is suitable for these tough specifications. Indeed, the close proximity of the antennas requires a specific decoupling topology where non-Foster circuits may bring some advanced solutions by cancelling or compensating the reactive components that model the antenna coupling. The Non-Foster circuits, through is a dedicated matching network, seem also to be a promising way to shift antenna operating band to low frequency which corresponds to compactness improvement. Moreover, a 2nd stage or complementary SI cancellation circuit is intended to achieve a higher SI cancellation level and we put forward the design of a 180° phase shifter based on non-Foster in that aim.
NIC and NII basics
NF elements can be implemented by two groups of circuits called Negative Impedance Converters (NICs) and Negative Impedance Inverters (NIIs). As seen in Figure 2.3. NIC is a two-port network terminated at one port by an impedance, and the other port presents the negative value of that impedance multiplied by a scaling factor The principle of operation of a NIC, configured in one of the earlier topologies by using two cross-coupled Bipolar Junction Transistors (BJT), is shown in Figure 2.4. The red arrows showing voltage flow and blue arrows showing Current flow generate an impedance ZNIC = −|K|ZL as seen across source terminals. This phenomenon is possible due to a voltage inversion at the load impedance (ZL) terminals. On the other hand, NII is also a two-port network terminated at one port by an impedance, but by looking at the other port we will see the inverse of the impedance with a negative scaling factor. For example, if the load impedance was a capacitor, then at the other port we will have a negative inductor.
NIC and NII can be represented using [ABCD] matrix as shown in Table 2.1, and satisfying the following conditions of (k1k2 < 0 and G1G2 < 0). In addition, depending on the sign of k1 and k2, NIC can change either the direction of the load current hence called current inversion NIC (INIC) or change the polarity of the load voltage hence called voltage inversion NIC (VNIC). The same thing is also applied for NII.
Non-Foster circuits Based on a Negative Group Delay NGD
This method introduces a new view in realizing NF elements based on wave propagation theory. In addition, this configuration solves the electrical stability problems of the traditional designs. This method compares between NF elements and NGD networks, and shows that both share the same influence on propagating waves [14]. In general, the group delay in propagation τD can be determined by examining the phase response ϕ of a network using the following equation: According to this equation, the group delay of NGD is equivalent to an increasing phase response with frequency, which is the same behaviour of NF. NGD networks can be designed using series and parallel RLC resonators operating around their resonance frequencies as shown on Figure 2.7. However, NGD suffers from losses due to the use of resistive elements. In fact, losses in NGD can be compensated by the use of transistors [15].
Non-Fosters Based on Distributed Amplifier
An unconditionally stable NIC circuit can also be built using two-stage microwave distributed amplifier that is based on negative group delay. The NGD circuit is built by using two stage distributed amplifier based on microwave transversal-filters. The trans-conductance of each stage can be adjusted by tuning its gate voltage, and the input/output phase responses can be tuned by two varactors, thus realizing tuneable NF characteristics between its reverse port (P2) and input port (P1), as shown in Figure 2.8. This topology provides good input and output impedance matching and can offer an unconditional stability[16]
Table of contents :
ACKNOWLEDGMENTS
SUMMARY
TABLE OF FIGURES
LIST OF TABLES
GENERAL INTRODUCTION
CHAPTER 1
1. INTRODUCTION
1.2 DUPLEXING TECHNIQUE IN MODERN WIRELESS LINKS
1.2.1 TIME DIVISION DUPLEX (TDD)
1.2.2 FREQUENCY DIVISION DUPLEX (FDD)
1.2.3 OTHER DUPLEXING/MULTIPLEXING TRANSMISSION TECHNIQUES
1.3 ADVANTAGES AND DISADVANTAGES OF FULL-DUPLEX TECHNIQUE
1.4 SOURCES OF SELF-INTERFERENCE
1.5 FD REQUIREMENTS
1.5.1 PASSIVE SELF-INTERFERENCE SUPPRESSION
1.5.2 ANALOG SELF-INTERFERENCE CANCELLATION
1.5.3 DIGITAL SELF-INTERFERENCE CANCELLATION
1.6 DISTORTION AND NOISE ISSUES IN FD
1.7 PROPOSED TOPOLOGIES FOR SI CANCELLATION
1.8 ANTENNA SPECIFICATIONS FOR A COMPACT FD SYSTEM
1.9 MOTIVATION, OBJECTIVES AND PROPOSED TOPOLOGY
1.10 ORGANIZATION OF THE THESIS
1.12 REFERENCE
CHAPTER 2
2.1 INTRODUCTION TO FOSTER AND NON-FOSTER ELEMENTS
2.3 NIC AND NII BASICS
2.4 TOPOLOGIES OF NIC
2.5 OTHER NON-FOSTER TOPOLOGIES
2.5.1 NIC BASED ON AMPLIFIERS
2.5.2 NIC BASED ON A TRANSFORMER
2.5.3 NON-FOSTER CIRCUITS BASED ON A NEGATIVE GROUP DELAY NGD
2.5.4 NON-FOSTERS BASED ON DISTRIBUTED AMPLIFIER
2.7 NICS CHOSEN CIRCUIT
2.7.1 CASE R0 INFINITE:
2.7.2 CASE R0 IS FINITE
2.8 STABILITY WITHIN NICS
2.9 LINVILL ANALYSIS LIMITATION
2.10 CONCLUSION
2.11 REFERENCES
CHAPTER 3
3.1 PHASE SHIFTER
3.2 TYPES OF PHASE SHIFTERS
3.2.1 ACTIVE AND PASSIVE PS
3.2.2 MECHANICAL OR ELECTRONIC PHASE SHIFTERS
3.2.4 FIXED OR ADJUSTABLE PHASE SHIFTERS
3.2.3 DIGITAL OR ANALOG TUNABLE PHASE SHIFTERS
3.2.5 RECIPROCAL OR NONRECIPROCAL PHASE SHIFTERS
3.2.6 PURE PHASE SHIFTERS
3.3 1ST DESIGN OF PS TUNABLE AROUND 180° USING NF CIRCUIT
3.3.1 PROPOSED TOPOLOGY
3.3.2 NF CAPACITANCE DESIGN AND MINIMIZATION OF RESIDUAL RESISTIVE PART .
3.3.3 IMPROVED REDUCTION OF RESISTIVE PART OF NF CAPACITANCE AT THE S-PARAMETERS LEVEL
3.3.4 ACTIVE TUNABLE PURE PHASE SHIFTER SIMULATION RESULTS
3.4 STABILITY STUDY
3.4.1 STABILITY STUDY BASED ON NDF
3.4.2 STABILITY STUDY: INFLUENCE OF DC BIAS
3.4.3 STABILITY STUDY: INFLUENCE OF TRANSMISSION LINE
3.5 DESIGN OF NON-FOSTER COMPONENTS USING TRANSISTOR WITH LOW TRANSITION FREQUENCY
3.5.1 MEASURED RESULTS
3.5.1.1 Prototype 1
3.5.1.2 Prototype 2
3.5.1.3 Prototype 3
3.5.2 2ND DESIGN OF TUNABLE PS
3.6 CONCLUSION
3.7 PERSPECTIVES ON PS
3.8 REFERENCES
CHAPTER 4
4.1 INTRODUCTION
4.1.1 DEFINITION OF AN ESA
4.1.2 TYPICAL APPLICATIONS OF ESAS
4.1.3 CHARACTERISTIC IMPEDANCE OF ESA
4.1.4 FUNDAMENTAL PARAMETERS OF ESA
4.1.4.1 Quality factor Q
4.1.4.2 Bandwidth
4.1.4.3 Gain-Bandwidth Limitation of Lossless Passive Matching Networks
4.2 ACTIVE CIRCUITS APPLIED TO IMPEDANCE MATCHING
4.3 METHODOLOGY OF COMBINED NON-FOSTER AND PASSIVE BROADBAND MATCHING NETWORKS
4.3.1 CONVENTIONAL MONOPOLE ANTENNA
4.3.2 FIRST MATCHING TOPOLOGY: NF-PASSIVE-ANTENNA
4.3.3 SECOND MATCHING TOPOLOGY: PASSIVE-NF-ANTENNA
4.3.4 PROPOSED MATCHING TOPOLOGY: PASSIVE-NF-PASSIVE-ANTENNA
4.3.5 TOPOLOGIES COMPARISON AND DISCUSSION
4.4 DESIGN OF A FULL-DUPLEX FRONT-END DEDICATED TO CLOSE AND ELECTRICALLY SMALL ANTENNAS (ESA) BY USING NF CIRCUITS
4.4.1 FULL DUPLEX APPROACH USING PASSIVE ELEMENTS
4.4.1.1 Analysis of topologies
4.4.1.2 Simulations and Measured Results
4.4.2 FULL DUPLEX TOPOLOGY USING NON-FOSTER ELEMENTS
4.4.2.1 Non-Foster matching of one monopole antenna
4.4.2.2 Non-Foster Matching and decoupling
4.4.2.3 Combining Two Levels of Self-Interference Cancellation
4.5 CONCLUSION
4.6 REFERENCES
CHAPTER 5
APPENDICES
APPENDIX A
APPENDIX B
APPENDIX C
APPENDIX D
COMMUNICATIONS AND PUBLICATIONS
