The thesis is concerned with the estimation of the sampled impulse-response (SIR), of a time-varying HF channel, where the estimators are used in the receiver of a 4800 bits/s, quaternary phase shift keyed (QPSK) system, operating at 2400 bauds with an 1800 Hz carrier. T= FIF modems employing maximum-likelihood detectors at the receiver require accurate knowledge of the SIR of the channel. With this objective in view, the thesis considers a number of channel estimation techniques, using an idealised model of the data transmission system. The thesis briefly describes the ionospheric propagation medium and the factors affecting the data transmission over BF radio. It then presents an equivalent baseband model of the I-IF channel, that has three separate Rayleigh fading paths (sky waves), with a 2Hz frequency spread and transmission delays of 0,1.1 and 3 milliseconds relative to the first sky wave. Estimation techniques studied are, the Gradient estimator, the Recursive leastsquares (RLS) Kalman estimator, the Adaptive channel estimators, the Efficient channel estimator ( that takes into account prior knowledge of the number of fading paths in the channel ), and the Fast Transversal Filter (F-FF), estimator (which is a simplified form of the Kalman estimator). Several new algorithms based on the above mentioned estimation techniques are also proposed. Results of the computer simulation tests on the performance of the estimators, over a typical worst channel, are then presented. The estimators are reasonably optimized to achieve the minimum mean-square estimation error and adequate allowance has been made for stabilization before the commencement of actual measurements. The results, therefore, represent the steady-state performance of the estimators. The most significant result, obtained in this study, is the performance of the Adaptive estimator. When the characteristics of the channel are known, the Efficient estimators have the best performance and the Gradient estimators the poorest. Kalman estimators are the most complex and Gradient estimators are the simplest. Kalman estimators have a performance rather similar to that of Gradient estimators. In terms of both performance and complexity, the Adaptive estimator lies between the Kalman and Efficient estimators. FTF estimators are known to exhibit numerical instability, for which an effective stabilization technique is proposed. Simulation tests have shown that the mean squared estimation error is an adequate measurement for comparison of the performance of the estimators.

This thesis presents the work performed by the author on the application of resonantmode techniques to commercially-orientated off-line converters. An extensive review of resonant-mode topologies leads to the development of a method of categorisation of these topologies which allows a greater comprehension of their properties. The categories of converter thus obtained are the conventional resonant converter, the quasi-resonant converter, and the gap-resonant converter. The gap-resonant converter is selected for further investigation. An analysis reveals the limited load and input voltage capabilities of this converter, and hence leads to the introduction of a pre-regulating converter to improve reliability and commercial viability. High-frequency techniques are explored and reported, and new techniques are developed in several areas in order to extend the concept of the gap-resonant converter to a realworld practical design. Subjects explored include the high speed driving of power MOSFETs, MOSFET and diode switching losses, high frequency magnetic materials and core losses, and skin and proximity effects. The techniques developed are used in the design of a 30OW, off-line converter with an input voltage range of 165V to 380V after rectification, and a ten-to-one output load range.

M.Ing. / Electromagnetic Compatibility (EMC) of electronic equipment is currently an important design parameter. Layout play a significant role in the EMC of power electronic converters. This thesis describes an investigation undertaken into the electromagnetic effects of converter layout. Typical detrimental effects identified during experimental work are presented. Possible causes for these effects are discussed. The experimental work is based on a systematic approach, which starts with a basic single switch chopper and ends in a split supply half-bridge converter. Interconnection modelling and SPICE simulations of layout affects are investigated next. The focus falls on analytical equations for extraction and simplified simulation circuits to make the process generally accessible. Typical resonant frequencies present in some of the experimental circuits are investigated with the help of analytically extracted parameters. The possibility of minimizing detrimental layout effects through impedance matching of interconnections and their terminations, is investigated next, since the previous section quantified layout parameters. Distributed vs. lumped element modelling of interconnections, and the boudary in between, are discussed. Simulation and experimental results are presented. Since maximum fuctionality and power, and minimum cost, per volume drives product development, all elements of a circuit should be investigated for the possibility of realizing secondary or even tertiary functions contributing to normal circuit operation. This is the focus of the last part of this thesis. Employing interconnections as low-pass or surge filters are investigated. Several waveforms are used to test experimental interconnection structures. Lumped and distributed modelling of these strucutres are discussed. The thesis concludes with a theoretical investigation into the possibility of dissipation of surge-energy instead of reflection utilizing interconnection-structures. One of these structures utilizes the skin- and proximity effect to realize low-pass behaviour.

M.Ing. / The focus of the project is on the design of a variable output current source applied in the testing of circuit breakers. The possibility of the use of high-frequency, switch mode converters for the application, will be investigated. The expectation is the improvement of the system currently in use, with the help of a power electronic converter. For the application, a 1 to 200A adjustable current source must be developed, which will be powered from the 220V, 50Hz network. A number of possible solutions to the problem will be investigated. One of the challenges of the project is that the output of the converter must be a current and not a voltage, which is normally the case. Based on mathematical calculations and practical results, the best possible solution to the problem is obtained. An optimum system is presented, which meets the desired specifications.

M.Ing. / Low profile power electronics components are currently in great demand. The rapid advances in semiconductor and micro-electronics technology during the last ten years have played a major role in the creation of this demand. These advances are in turn driven by the need for compact design in computing, communication, consumer electronic goods and control systems with direct consequences in power supply design and manufacture. The study covered by this thesis involves the design, manufacture and thermal analysis of a planar inductor, which is a typical planar power electronics component. First, a throughout literature survey of planar magnetics revealed that satisfactory experimental procedures for the thermal analysis of passive power electronic components under operating conditions representative of modern applications are seldomly applied. Secondly, a computer based field-solver program and analytical methods are used to design and analyse a planar inductor. The applicability of different methods for determination of low power loss in passive components is discussed next. Finally, an experimental method suitable for low power loss determination is proposed and investigated. This method can be used in the analysis of inductors or capacitors of different sizes. It has a wide spectrum of application due to the advantages of frequency independence and different possible power levels.

On-device audio recognition, in particular keyword spotting, will be instrumental to realizing the promise of pervasive intelligence. On-device operation demands ultra-low power and compact area. The state of the art in fully-integrated keyword spotting chips reveals that the power and area bottleneck is not the backend keyword spotting classifier, but rather the frontend audio feature extractor, motivating research into frontend audio feature extraction that is both power- and area-efficient. After, first, introducing the topic of ultra-low power audio feature extraction using time-mode analog signal processing circuits, we, second, present an analog audio feature extractor chip that achieves the lowest power/feature and area/feature, as compared, respectively, to the most area-efficient and power-efficient published analog audio feature extractor chips. Despite the chip's state-of-the-art efficiency, competitive keyword spotting accuracy is maintained when interfacing the chip with a standard, small-footprint, software backend neural network. The chip's efficiency is due to a pair of novel circuit techniques we developed. The techniques are based on time-mode analog signal processing. This is a paradigm favored by technology scaling, in which analog information is encoded in the timing of digital edges, enabling digital gates to perform analog signal processing. Third, we present a theory-based analysis of one of the two circuit techniques. Fourth, we present theory- and simulation-based progress towards what would be a novel type of analog filtering, ``Time-Mode Analog Filter." Such a filter would use only the horizontal time axis to represent and process continuous-valued information, and would be built out of nothing more than digital gates. Fifth, and finally, we present a simulation-based study that finds that in state-of-the-art analog audio feature extractor chips, the power consumption of the critical block, the analog filterbank, can be reduced by one-and-a-half orders of magnitude, while degrading downstream keyword spotting accuracy by only a couple percent, paving the way towards more rigorous system-level design of audio recognition systems.

Note:

No description available.

The performance of the clamped-mode series resonant converter operating at a fixed frequency is studied for off line applications. A new set of characteristics for the converter operating above and below resonant frequency has been developed by including the effect of losses in the analysis. Based on the analytical results, design guidelines are established and two prototypes were built to operate below and above resonant frequency respectively. The advantages and limitations of the two breadboards are assessed and their major sources of loss identified. / Master of Science

Numerical frequency-domain techniques are widely used for the a.c. steady-state analysis of nonlinear electric circuits. Such techniques require that one compute the Fourier series for the response of each nonlinear circuit element, given a known excitation. Current approaches to this computation encounter difficulty when the response is almost periodic (that is, when the frequencies in its Fourier series are not all harmonically related), especially when the nonlinear characteristic is abrupt and the Fourier series for the response contains many significant terms. This dissertation develops an alternative approach that is theoretically sound and computationally efficient, for the important special case of a memoryless nonlinearity described by a continuous, bounded function. To begin the development, basic properties of almost periodic functions are presented. It is proven that the response of a memoryless nonlinearity is almost periodic whenever the excitation is. Next, the concept of a basis for a set of frequencies is introduced. The frequency content of the response is investigated, and it is proven that the frequencies in the response have the same basis as those in the excitation. The Fourier series for an almost periodic function is discussed, and its coefficients are expressed as mean values taken over an infinite interval. Results are given for the summability of the series. Starting with a theorem from Diophantine Approximation, it is proven that the normalized (Hertzian) phases corresponding to a set of M basis frequencies have their fractional parts uniformly distributed in an M-dimensional unit cube. This property of uniform phase distribution is then used to convert the single-dimensional integral for the Fourier series coefficients into a multiple integral over the unit cube, with the dimension of the integral equal to the number of basis frequencies in the Fourier series. A multi-dimensional extension of the Discrete Fourier Transform is used to evaluate the multiple integral, and expressions for aliasing are derived. It is shown that the multiple integral formulation compares favorably with existing approaches, and several numerical examples are presented to illustrate this formulation's capabilities. / Ph. D.