D.Ing. / Changes in the preferred technologies used in modem switched mode power supplies are driven by the need for a decrease in cost of manufacturing while maintaining the highest possible power density. Modem materials allow smaller components to be manufactured without affecting their electromagnetic properties but thermal problems still impair efforts to further reduce their size. Increased switching frequencies cause increased conductor and core losses in magnetic components such as inductors and transformers. Amongst other advantages the increased surface area of low-profile planar structures allow better removal of the heat generated in the component and this work reevaluates conductor optimization for specific planar inductor windings. Conductor optimization has been a topic thoroughly investigated over the last few decades and these techniques have been adopted for planar winding design. It will be shown that the process involved in the manufacturing of helical planar windings allows further exploitation of the basic optimization process. A per-layer optimization technique, earlier proposed by other researchers and deemed impractical at the time, is revised and it is shown here that it can be successfully implemented in the chosen winding type. The per-layer optimization is also extended to accommodate non-sinusoidal current waveforms in this work. Other facets of loss reduction are also addressed. Windings are shaped in the region of air gaps in magnetic circuits to reduce the losses caused by the fringing flux intersecting the conductors but since the shaping influences both the ac-resistance and dc-resistance of the winding an optimization process is required to find the optimal conductor arrangement. Contributions are made with regard to speeding up the process of finding the optimal layout of conductors around the air gap by devising a simple model for the flux distribution of the fringing flux. The effect of combining the per-layer optimization technique with these shaping techniques is exploited to its full potential in this work and contributes largely to loss reduction in helical inductor windings carrying currents containing ac- and dc components. The optimization time is reduced by allowing winding shaping while maintaining the dc-resistance of each layer. The optimal winding shape then becomes a function of the window constraints for a chosen core rather than finding it through conventional methods. Case studies, complete with FEM-simulations and experimental measurements, are offered in support of the proposed solutions. Manufacturing issues are dealt with and loss measurement techniques developed as part of this work.

M.Ing. (Electrical and Electronic Engineering) / The main purpose of the study is to propose an unique unity power factor AC-DC converter topology, with isolated output. The proposed converter is obtained by merging the unity power factor boost converter and the full bridge resonant transition converter. The converter is unique as resonant switching techniques are used to obtain a zero voltage switching boost stage, resulting in an all soft switching converter. The result is lower levels of conducted EMI, as well as lower switching losses. Conventional topologies utilize more than one converter stage to obtain the required results, whereas _ the proposed converter results in a single step solution. As the purpose of the proposed converter is power factor regulation resulting in lower levels of conducted EMI, the study is commenced with a review of the definitions and equations for both power factor and EM!. Thereafter sources and propagation of conducted EMI are discussed. To gain a better understanding of the operation of the proposed converter, the operation of the unity power factor boost converter, as well as the full bridge resonant transition converter are discussed in depth. A switching cycle of the proposed converter is divided into different intervals, and analytical equations for the relevant currents and voltages for each interval are derived seperately. Possible control schemes are introduced and where applicable, conditions for zero voltage switching are given. It is concluded that a constant frequency PWM controller is the most ideal control scheme to ensure proper operation of the converter. A simple design procedure is presented by which the size of the most important circuit parameters can be designed in terms of the desired switching frequency and permissable duty cycle variation. Finally, experimental results obtained from the practical converter, are given. Deviations from the ideal are interpreted and methods of improving the performance are suggested.

M.Ing. (Electrical & Electronic Engineering) / Induction heating is a relatively old and established process. New switching devices and ideas are, however, resulting in much improved performance at reduced cost. Some of these devices are investigated in this thesis, and some new ideas presented. From a literature study a general review of the current state of the technology is given. The fundamental principles and limits of induction heating are discussed. Power electronic topologies for induction heating, as well as switching devices which can be used in these topologies, are compared and evaluated. From the knowledge gained, possible fields for improvement are identified. One such field is that of semiconductor switch drives, and a high-performance gate drive was developed. Using such fast drives, special attention must be given to the electromagnetic layout of the circuit. Another new field is that of integrated resonant elements. These elements find application in most resonant converters, and offer reduced weight and losses compared to conventional technologies. They are evaluated both experimentally and theoretically. A prototype induction heater was built, using a newly released switching device named a zero turn-off thyristor. The high-performance gate drive developed earlier was tested in the inverter and gave excellent results.

M.Ing. (Electrical and Electronic Engineering) / The purpose of this investigation is mainly the study of various topologies of force commutated distortion compensators. This is done against the background of distortion in power systems in general. At the time of the investigation confusion reigned the domein of terminology and definition of power components and power compensators. Part of the investigation was the functional categorizing of power compensators. Forced commutated distortion compensators were then highlighted as element of this suggested frame work. There is a general increase of distortion in power networks which emphasises the limits of passive filters to compensate for non-active power. The use of power converters as forced commutated distortion compensators proves an effective means whereby to counteract these limits. In this document the different topologies of forced commutated distortion compensators are labelled according to the position where the converter is coupled to the network, as well as the characteristics of the energy storage element used. The topologies are studied through modelling, mathematical analysis and simulation. An experimental system of one of the topologies is presented. Control strategies for forced commutated distortion compensators are discussed with emphasis on the control of energy loss in the energy storage elements of the compensators. The investigation is concluded with proposed ideas for futher study as well as a philosophical approach to the role of forced commutated distortion compensators in the future.

M.Ing. (Electrical and Electronic Engineering) / In order to improve the industrialisation of modern power converters, its physical size must be reduced, material must be utilised more efficiently and its manufacturability improved. Consequently, many converters are operated in the ultrasonic frequency range, where the size of its constituent components are considerably reduced. Even at high frequencies of operation, however, magnetic components are still the largest and heaviest components in converters, and do not utilise materials to their maximum benefit. The aim of this work is to investigate the possibility of improving the industrialisation of power electronic converters by making use of a novel, multiple core magnetic component structure, called the distributed transformer. Existing magnetic component structures are evaluated from a materials point of view and the concept of distributed magnetic components introduced. A generalised transformer model is developed for the distributed transformer and pertinent design considerations discussed. A 1.5 kw battery charger industrialised by Spoornet is based on the CukDC-DC converter and is used as a case study. A conventional monolithic isolation transformer is used in its construction after a complete mathematical analysis of the converter topology, and compared with the equivalent distributed transformer configuration. The distributed transformer is then compared with the conventional monolithic transformer and evaluated in terms of its physical and electrical performance, design and ease of manufacture.

This thesis presents new configurations for three-phase AC-to-DC single-stage, softswitched, high frequency (HF) transformer isolated converters with power factor correction (PFC) and low harmonic distortion. Four different configurations are presented. Topology of all these four configurations is based on integration of a front-end DCM boost with a soft switching HF transformer isolated DC-to-DC PWM converter with fixed frequency. DCM operation of the front-end boost provides natural PFC with low total harmonic distortion (THD) and the DC-to-DC HF transformer isolated soft switching PWM converter with an appropriate gating scheme provides output voltage regulation. A double switch AC-to-DC converter is presented in Chapter 2. Due to unsymmetrical gating scheme, DC blocking capacitors are required to avoid transformer saturation. To reduce this problem, a new gating scheme is proposed in Chapter 3, which can be used in full bridge converters providing ZVS. This gating scheme is first used in a DC-to-DC bridge converter. In the next three chapters, this new gating scheme is applied to three different types of single-stage AC-to-DC boost integrated fixed-frequency bridge converters. These configurations are: (i) boost integrated single inductor linear current DC-to-DC PWM bridge converter, (ii) boost integrated series resonant DC-to-DC bridge converter and (iii) boost integrated parallel resonant DC-to-DC bridge converter. The steady state operation of each converter and modes of operation are explained with equivalent circuits for each interval of HF cycle. The general solutions for all the intervals are derived and design curves are obtained based on steady state relations. The design procedure is illustrated with a design example. Detailed PSPICE simulation results and experimental results obtained from a laboratory prototype model are given for all the converters to verify the theory and analysis. THD of the line current without any complex control circuit remains in a reasonable range of 8% to 13% for the total range of operation. Input line current waveforms for all suggested converters shows a low harmonic distortion similar to a single 3-Φ DCM boost. The difference would be in increase or decrease of DC bus voltage in each case, which can affect THD of the boost converter. Three switches in the full bridge converter operate with zero-voltage switching (ZVS) while the main switch operates with ZVS at full load, minimum line voltage and with ZVT at lower loads conditions. Soft switching of all the switches helps in lower loss. Chapter 6 presents the contributions of this thesis, summarizes the advantages and disadvantages of each configuration and gives some suggestions for future work. / Graduate

To date, measurements of tunnel diode noise have dealt mainly with the negative conductance region, because the tunnel diode is an active circuit element only in this region. The noise has not been measured for reverse or near-forward biases due to the difficulties involving excessively low diode impedances in these regions. The purpose of this thesis is to show that, from the Esaki formulation for the direct-tunneling currents of a tunnel diode, in the bias regions where the electronic bands overlap, a simple theory can be developed relating the power spectrum associated with the direct-tunneling current noise to the direct current passing through the diode. This theory assumes that the two oppositely-flowing direct-tunneling currents in the Esaki junction are uncorrected and that both contribute full shot noise. The theory can be critically tested only in the bias regions where the noise is yet unstudied, and at sufficiently high frequencies that no contaminating 1/f noise exists. These conditions have been met experimentally and the noise measured quantitatively over the entire reverse and near-forward regions at a frequency of 4 Mc/s. Impedance-transforming networks and a very low-noise preamplifier suitable to the particular source strengths and impedances presented by the tunnel diode are developed for these measurements. A noise measurement technique is chosen from among several possible ones for the high degree of accuracy and smallest dependence on a good noise figure required for the tunnel diode source. The experimental results agree with the theory and vindicate the usual assumption that the two oppositely flowing direct-tunneling electron currents between two bands of a degenerately-doped semiconductor are uncorrelated. Noise measurements in the "valley" and far-forward region of the tunnel diode characteristic, where the diode current is not due to direct tunneling, do not agree with the simple two-current shot noise theory for direct-tunneling electron currents. Possible reasons for the enhanced noise measured in this region are advanced in the form of two models based on indirect-tunneling electrons via traps as the most important mechanism describing the excess or valley current. These models offer a possible explanation of the observed phenomena, but noise measurements alone appear insufficient to demonstrate unambiguously the detailed mechanisms producing either the excess current or the associated enhanced noise found throughout the valley and far-forward regions. / Science, Faculty of / Physics and Astronomy, Department of / Graduate

This thesis presents the design and construction details of a physical model of a six-pulse HVDC converter suitable for steady-state studies. The converter rack has been built to facilitate simple testing and breadboarding. A digital equidistant firing angle controller has been constructed for the converter-rectifier current control. An electronic phase-locked loop has been used to generate an accurate control frequency. Modifications have been suggested to make the rectifier, fully operational. The control system has been constructed for future inverter control circuits and possible computer control. / Applied Science, Faculty of / Electrical and Computer Engineering, Department of / Graduate

D.Ing. (Electrical & Electronic Engineering) / Please refer to full text to view abstract

No description available.