Project topics and materials
Get Complete Project Material File(s) Now! »
Introduction to electronic filters
A filter is an electrical network that changes the amplitude or phase characteristics of a signal, with respect to frequency. Additional frequency components should ideally not be added to a signal by a filter, and it should only change the amplitude or phase characteristics of the existing frequency components. In this way, filters can be used to emphasize or reject certain frequency components, depending on the system specifications. This is done by altering the gain (or attenuation) of the filter at certain frequencies.
Typical requirements of filters are the frequency response, phase shift (group delay), and impulse response. A filter must be causal (only be dependent on current and past inputs) and stable (to prevent oscillation). The filter should not be computationally complex, and implemented in either software or hardware, depending on the application. Other considerations include whether the filter should be analogue, analogue-sampled, digital (finite impulse response or infinite impulse response), or mechanical. The linearity of the filter must be accounted for. Finally, the filter can be either active (introduce amplification) or passive (no amplification).
Several filter types exist, each with specific characteristics for amplitude response and phase shift (group delay). The choice of filter therefore depends on the application. Butterworth, Chebyshev, Bessel, Elliptical, and Gaussian filters are a few examples of filter types, with variations on each type also possible. Butterworth filters produce the flattest pass-band amplitude response of all the filters, but at the expense of a relatively slow roll-off between pass-band and stop-band. Butterworth filters also present an unwanted phase shift which can result in distortion of the output signal. Chebyshev filters improve somewhat on the sharpness of the roll-off of Butterworth filters, but at the expense of a small output ripple in the transfer characteristic amplitude response. Bessel filters address the maximum flatness of the time delay within it the pass-band (in contrast to the Butterworth filter maximum flatness of amplitude response) but again, at the expense of large ripples in the pass-band, and also a lower roll-off in the frequency domain.
Electronic filters can operate in the digital (digitized) domain and in the analogue (continuoustime) domain. Each domain has its own advantages and disadvantages depending on the application. Analogue filters are most commonly used in high-frequency, RF applications such as electronic frequency mixers and RF modulators, whereas digital filters are preferred for high precision and high-accuracy applications. Digital filters generally operate far below the of the transistors used to design these filters to avoid aliasing. Analogue filters present higher bandwidth capabilities compared to its digital counterpart (as there is no need for analogue to digital conversion). Analogue filters are also relatively easier and cheaper to implement as there are less overhead circuitry needed. For this research, analogue filters are used due to its higher bandwidth capabilities.
Implementing a passive filter (or active filters, depending on the frequency of operation and limiting parameters of active components) on an IC requires knowledge of process parameters to determine the feasibility of designing these filters on-chip. Size, speed, and the addition of noise to these circuits are considered and comparisons to find the optimal process in terms of performance and cost-effectiveness are crucial.
Passive filters
A passive filter is a filter that uses only passive components (capacitors, inductors, and resistors) and has no amplification properties (gain), and ideally only attenuation outside the pass-band. Active filters employ amplification elements, such as transistors and operational amplifiers, and use the passive components in the feedback loop. The distinct advantage of passive filters over active filters is that it does not require any power supply to operate, and thus does not consume power. There are also no bandwidth limitations incurred from active components, and it can operate at higher frequencies compared to active filters, which suffer from gain-bandwidth limitations. Passive circuits generate low noise (only thermal noise is present from the resistive components) and no active noise is present, the circuits are reliable even under high-current operation as no active devices are present. The lack of gain in these circuits can in some applications be considered a disadvantage, where active filters might be more practical. Passive filters can also present too low input and too high output impedances, and additional buffer circuitry may be required. Active filters present high input impedance, low output impedance and can be designed to have almost any desirable gain. The use of inductors in passive filters increase the real-estate needed to implement the circuits (increased cost in IC fabrication). The fact that active filters do not need inductors is one of the more distinct advantages of these types of filters (especially at lower frequencies where inductor sizes tend to be physically impractical). Large order filters (to attain sharp roll-off and attenuation slopes) can become very time-consuming to design, and limit tuning due to its mathematical complexity. Active filters are generally easier to design compared to passive filters. In high-frequency applications, standard CMOS and BiCMOS integrated filters that employ on-chip passive devices suffer from large substrate losses and therefore exhibit generally low quality factors and hence low-bandwidth capabilities. To counter this effect, active filters can be used to enhance the quality factor of the circuit, but an increase in noise factor (NF), non-linearities, and direct current (DC) power consumption are exhibited. RF (onchip) filters can also be implemented as digital or analogue filters.
Examples of passive filters include inductor-capacitor (LC) configurations, RF microelectromechanical systems (RF-MEMS), and electro-acoustic filters. LC filter configurations are the most generally used filters, however, at higher frequencies the bandwidth of the filter depends on the -factor of the inductor and capacitor, which could be a limiting factor depending on the technology process. Several new methods are being investigated to improve the quality of the passive components, such as the works presented in [43], [44], and [45]. RFMEMS filters can be implemented on-chip as they have good linearity, low-power consumption, and low losses. RF-MEMS filters do, however, suffer from high thermalmechanical noise due to Brownian motion [46], resulting in additional noise at the output of these circuits. A study on the linearity of RF-MEMS (tuneable) filters are done in [47] by using the pole-perturbation approach. Acoustic wave filters employ an acoustic component, not mounted on integrated level, and is not considered for this research application.
The following section is an introduction to antenna theory, with specific reference to equations used to design an equivalent circuit model used in this research. Important considerations such as the directivity and radiation intensity, definition-of-field regions, the gain and efficiency of an antenna, as well as the impedance and aperture characteristics of antennas are described here.
CHAPTER 1: INTRODUCTION
1. CHAPTER OVERVIEW
1.1 BACKGROUND TO THE RESEARCH
1.2 HYPOTHESIS AND RESEARCH QUESTIONS
1.3 JUSTIFICATION FOR THE RESEARCH
1.4 RESEARCH METHODOLOGY
1.5 DELIMITATIONS AND ASSUMPTIONS
1.6 CONTRIBUTION
1.7 PUBLICATIONS FROM THIS RESEARCH
1.8 OUTLINE OF THE THESIS
1.9 CONCLUSION
CHAPTER 2: LITERATURE REVIEW
2. CHAPTER OVERVIEW
2.1 PATH-LOSS PREDICTION MODELS
2.2 ELECTRONIC FILTER THEORY
2.3 ANTENNA THEORY
2.4 DIPOLE MODELLING
2.5 MATCHING NETWORKS
2.6 PASSIVE DEVICE MODELLING
2.7 CONCLUSION
CHAPTER 3: RESEARCH METHODOLOGY
3. CHAPTER OVERVIEW
3.1 DEVICE SPECIFICATIONS
3.2 TECHNICAL PACKAGES
3.3 MATHEMATICAL MODELLING
3.4 ELECTROMAGNETIC SIMULATIONS
3.5 SCHEMATIC DESIGNS AND SPICE SIMULATIONS
3.6 CIRCUIT LAYOUT AND VERIFICATIONS
3.7 PCB DESIGN AND LAYOUT
3.8 MEASUREMENTS
3.9 CONCLUSION
CHAPTER 4: MATHEMATICAL MODELLING AND SIMULATIONS
4. CHAPTER OVERVIEW
4.1 DIPOLE DESIGN
4.2 SIMULATION RESULTS
4.3 DIPOLE MODEL
4.4 PATH-LOSS MODELS
4.5 PASSIVE FILTER DESIGN (MATHEMATICAL APPROXIMATION)
4.6 PASSIVE FILTER DESIGN (SIMULATED RESULTS)
4.7 MATCHING NETWORK DESIGN
4.8 CONCLUSION
CHAPTER 5: MEASUREMENT RESULTS
5. CHAPTER OVERVIEW
5.1 DIPOLE MEASUREMENTS
5.2 PASSIVE FILTER MEASUREMENTS
5.3 MATCHING NETWORK MEASUREMENTS
5.4 CONCLUSION
CHAPTER 6: CONCLUSION
6. CHAPTER OVERVIEW
6.1 INTRODUCTION
6.2 CRITICAL EVALUATION OF HYPOTHESIS
6.3 SCOPE LIMITATIONS AND ASSUMPTIONS
6.4 FUTURE WORK AND POSSIBLE IMPROVEMENTS
REFERENCES
