Includes bibliography. / Two new hybrid companding delta modulators for speech encoding are presented here. These modulators differ from the Hybrid Companding Delta Modulator (HCDM) proposed by Un et al in that the two new encoders employ Song Voice Adaptation as the basis of the instantaneous compandor, rather than Constant Factor adaptation. A detailed analysis of the performance, both objective and subjective, of these hybrid codecs has been carried out. Results show that overall the two codecs developed as part of this project are better than the HCDM codec. In addition the new codecs offer simpler implementation in digital hardware than the HCDM. A Computer Aided Test (CAT) system has been developed to simplify the design and test processes for speech codecs.

Includes bibliographical references. / Harmonic distortion of three phase power systems is caused by large amounts of electric power consumers using power electronic installations (e.g. drives and rectifiers). It has a number of undesirable consequences and has become an increasingly important problem with the rapid growth and ready availability of the power electronic devices and installations. A famous mathematical technique called "symmetrical components" was used to model the fundamental components of three phase systems. This method was extended to various harmonics. The aim of this thesis is to measure both the magnitude and phase angles of various harmonics which exist in three phase power systems, and to express them in symmetrical components. A digital electronics and microprocessor-based instrument was designed and constructed using the novel signal modulation technique called "+1 and -1" which was developed by the author of this thesis. This instrument was employed to achieve the above objectives.

Bibliography: leaves 175-177. / This thesis is concerned with Gallium Arsenide Metal Semiconductor Field Effect Transistor Microstrip Integrated Circuit Dielectric Resonator Oscillators (GaAs MESFET MIC DROs) - the different types, their design and their performance compared to other high Q factor (ie narrowband) microwave oscillators. The thesis has three major objectives. The first is to collate the information required to build microwave DROs. The second is to present the practical results obtained from Dielectric Resonator Bandreject and Bandpass filters (DR BRFs and DR BPFs). The last is to present and compare results from a DR stabilised microstrip oscillator and three types of series feedback DROs. Narrowband oscillators are usually evaluated in terms of their frequency stability, reliability, size, cost, efficiency and output power characteristics. In terms of these parameters DROs outperform Gunn cavity oscillators and are only bettered by crystal locked sources in terms of frequency temperature stability and long-term stability. The components of a GaAs MESFET MIC DRO possess ideal properties for the construction of a narrowband source with the exception of the long term stability of the GaAs MESFET. GaAs MESFET•DROs have the best published DRO results for efficiency, output power, power temperature stability and external Q factor. Basic oscillator theory derived by Kurokawa can be applied to both negative resistance and feedback oscillators. Impedance locus, device-line and operating point concepts provide a convenient framework for understanding hysteresis in microwave oscillators. The work by Kurokawa can also be translated into the S-parameter domain which has proved convenient for the design of microwave oscillators.

Includes bibliographical references. / The work in this thesis is primarily directed toward the design, construction and testing of an experimental multilevel partial response signalling baseband system. The system will find practical application in existing frequency division multiplexed-frequency modulated microwave links. The basic supergroup bandwidth of these links is 240 kHz. The design requires a transmission rate of 1.024 Mb/s in this bandwidth. Class-4 15 partial response signalling is the coding technique suitable to achieve this. A pilot tone scheme is used to facilitate symbol timing recovery at the demodulator. A sixth order Butterworth low pass filter approximates the ideal raised-cosine Nyquist channel. A theoretical discussion on impairments caused by deviation from this channel is given. Since the experimental system was non-ideal, it produced a degradation in the channel signal to noise ratio. This degradation, coupled with other factors, showed that further development was necessary for the system to be suitable for connection into an existing microwave link.

Includes bibliographical references. / Dense Medium Separation (DMS) is a process used extensively in the minerals processing industry to separate dense from less dense material in a dense fluid. It may be considered to be a simple “sink-float” separation process. DMS is used on a large scale in South African coal, iron ore and diamond operations. There are, however, no commercially available systems that can determine the separation efficiency of a DMS process on-line. This presents severe problems to those operating DMS processes. The present study attempts to provide a measurement technique for on-line application.

Includes bibliographical references. / This dissertation examines the problem of on-line measurement of complex dielectric constant for the purpose of dielectric discrimination or product evaluation using microwave techniques. Various methods of signal/sample interaction were studied and consideration was given to the problem of sorting irregularly shaped discrete samples. The use of microwave transmission and reflection measurements was evaluated. The signal reflection methods were deemed to be best suited to applications with constant geometry feed presentation ( ie. a continuous, homogeneous product stream with little variation in surface geometry).

Bibliography: leaves 127-128. / There is a great need for real-time non-intrusive measurements in industry. A short-range radar system can be used to make these measurements. A standard requirement for these type of applications is high resolution. This is a standard problem in radar. Using classical signal processing techniques, the range resolution is proportional to the bandwidth of the transmitted signal. This poses a serious problem in radar as very large bandwidths are required - typically lSOGHz for 1 mm range resolution. Alternative techniques have been sought which do not rely on large transmitted bandwidths, but which rely on large signal-to-noise ratio (SNR). Such techniques exist in modem spectral analysis eg. auto-regressive techniques. These techniques model the data. In other words, they assume a priori information. Linear frequency-modulated continuous-wave (FMCW) radar was utilized, since a pulsed radar would require very precise time measurements due to the short range (a few ns). The FMCW radar would have to be very linear for the modelling process to work properly. The frequency domain measurement of the received system data would then be proportional to range. An FMCW radar system was built and tested. The modem signal processing techniques were found to work well when injected with sinusoidal signals from signal generators. The hardware was also found to perform satisfactorily. However, amplitude modulation was observed in the mixing process and subsequently, the modelling process did not perform satisfactorily when interfaced to the hardware. Due to the amplitude modulation problem, two closely-spaced targets disrupted the high resolution properties of the modelling process. Nevertheless, a single target could be resolved within a resolution bin of better than 1 cm. A solution is proposed in chapter eight, however, it is out of the scope of this thesis.

Includes bibliographical references. / This describes a mathematical approach to solving the pulsed magnetic field for an N-turned magnetising coil. This coil is assumed to be made from a copper conductor with rectangular cross-sectional area. The algorithm accounts for each turns physical dimensions and spatial coordinates with respect to other turns, by simulating each winding with a model helical turn. A boundary matching method is used to determine the current density distribution in the 2D plane of the call conductor and the changed impedance due to "skin-effect". By assuming this distribution and changed impedance applies for the N-turned coil volume, the effects of eddy current losses are approximated inside the call. Results of comparing field calculations with several analytical field solutions for static fields, and by measurements for pulsed fields, confirms the accuracy of the field algorithm in approximating a real coil field to within 10% error. This algorithm can be used to aid the design for generation of saturation fields for magnetisation of different ferromagnetic materials.

Includes bibliographical references. / A Ground Penetrating Radar (GPR) sensor is required to provide information that will allow the user to detect, classify and identify the target. This is an extremely tough requirement, especially when one considers the limited amount of information provided by most GPRs to accomplish this task. One way of increasing this information is to capture the complete scattering matrix of the received radar waveform. The objective of this thesis is to develop a signal processing technique to extract polarimetric feature vectors from Stepped Frequency Continuous Wave (SFGWV) GPR data. This was achieved by first developing an algorithm to extract the parameters from single polarization SFCW GPR data and then extending this algorithm to extract target features from fully polarimetric data. A model is required to enable the extraction of target parameters from raw radar data. A single polarization SFCW GPR model is developed based on the radar geometry and linear approximations to the wavenumber in a lossy medium. Assuming high operating frequencies and/or low conductive losses, the model is shown to be equivalent to the exponential model found in signal processing theory. A number of algorithms exist to extract the required target parameters from the measured data in a least squared sense. In this thesis the Matrix Pencil-of-Function Method is used. Numerical simulations are presented to show the performance of this algorithm for increasing model error. Simulations are also provided to compare the standard Inverse Discrete Fourier Transform (IDFT) with the algorithm presented in this thesis. The processing is applied to two sets of measured radar data using the radar developed in the thesis. The technique was able to locate the position of the scatterers for both sets of data, thus demonstrating the success of the algorithm on practical measurements. The single polarization model is extended to a fully polarimetric SFCW GPR model. The model is shown to relate to the multi-dimensional exponential signal processing model, given certain assumptions about the target scattering damping factor. The multi-snapshot Matrix Pencil-of-Function Method is used to extract the scattering matrix parameters from the raw polarimetric stepped frequency data. Those Huynen target parameters that are independent of the properties of the medium, are extracted from the estimated scattering matrices. Simulations are performed to examine the performance of the algorithm for increasing conductive and dielectric losses. The algorithm is also applied to measured data for a number of targets buried a few centimeters below the ground surface, with promising results. Finally, the thesis describes the design and development of a low cost, compact and low power SFCW GPR system. It addresses both the philosophy as well as the technology that was used to develop a 200 - 1600 MHz and a 1 - 2 GHz system. The system is built around a dual synthesizer heterodyne architecture with a single intermediate frequency stage and a novel coherent demodulator system - with a single reference source. Comparison of the radar system with a commercial impulse system, shows that the results are of a similar quality. Further measurements demonstrate the radar performance for different field test cases, including the mapping of the bottom of an outdoor test site down to 1.6 m.

Includes bibliography. / The subject of this thesis concerns the development of a Local Area Network (LAN) for the Department of Electrical and Electronic Engineering at the University of Cape Town. Motivation for this project was as a result of the ever increasing demands placed on the department's micro-computer training facilities by larger student intakes. The original training system consisted of a PDP 11/23 mini-computer connected via 9600 baud asynchronous links to 11 U.C.T. built micro-computers. This network topology was limiting in three ways: 1. It was slow because of the 9600 baud links and because the PDP was doing a large proportion of the processing.2. High-leve 1 software development tools for the PDP were too expensive and would over-load the computer. Because the micro-computers have no operating system but only an "in-house" monitor program which is not able to support any high-level language utility, all high-level software tools would have to be individually developed for this particular environment. 3. Switching was impractical. Because the PDP was the hub of the network all communication between computers had. to pass through it. This switching would lead to a greater processing load on the PDP, thus further degrading its performance. A two pronged attack was used to overcome these weaknesses: firstly, by designing a high-speed (1 Mbps) LAN to provide communications between a PDP 11/23 and up to 30 U.C.T. built micro-computers, faster inter-computer communication as well as switching and resource sharing was facilitated. Secondly, by customizing an operating system for the micro-computers, standard high-level software development tools could be used on these computers, consequently reducing the PDP's processing load.