The research outlined in this thesis is involved with some aspects of power electronics. In particular, within the area of machine drives, a new type of inverter is investigated. This inverter, referred to as the delta inverter, is principally intended for incorporation into an ac drive system. The inverter appears to offer some advantage in circuit design over conventional inverters and this is discussed in the context of the inverters overall performance and characteristics. Chapter 1 presents a review of machine drives, both ac and dc systems, and, in accepting the relative merits of the ac squirrel cage induction machine, outlines the penalties generally paid in conventional ac inverter complexity. The delta inverter configuration, which appears to offer a reduction in circuit complexity, is then introduced. Chapter 2 investigates the general behaviour of the delta inverter, specifically with passive loads, and outlines a suitable control strategy for the inverter. Chapter 3 deals with the subject of harmonic improvement of inverter waveforms. Two methods which appear particularly attractive, selective harmonic reduction (SHR) and sinusoidal pulse-width modulation (PWM), are presented. Using the strategy suggested in Chapter 2, the implementation of methods of sinusoidal pulse-width modulation applicable to the delta inverter are derived in this chapter and a theoretical comparison is made with the conventional bridge inverter. Chapter 4 presents both experimental and computed results for a delta inverter ac drive system using the previously derived switching techniques. The results are compared with the bridge inverter performance and with a mains quality sinusoidal supply. This chapter concludes with some discussion of the delta inverter ac drive system and some general points are made. Chapter 5 summarises the main conclusions that have been established, in the course of the present research, concerning the delta inverter system and its feasibility as an alternative ac drive circuit.

This thesis describes an investigation into the causal mechanisms of anomalous non-linearity in quartz crystal resonators. In general terms anomalous non-linearity is taken to be any non-linearity which is greater than that produced by the higher order elastic constants of quartz. More specifically, anomalous non-linear behaviour is that which does not obey the conventional non-linear law of a fixed coefficient polynomial in crystal current or strain. Both types of anomalous behaviour are studied. Five aspects of the AT-type crystal unit are considered for potential non-linear mechanisms: i) surface contamination ii) bulk impurities and defects iii) internally coupled unwanted modes iv) electrode and mounting defects v) externally applied stress. The details of experiments performed to determine the influence of these factors on the non-linear behaviour of quartz crystal resonators axe presented and the interpretation of results discussed. Seven physical mechanisms are proposed as possible causes of anomalous non-linearity in quartz crystal resonators. The most significant of these is surface contamination. This is found to be temperature and frequency dependent. The effectiveness of current industrial surface cleaning techniques is assessed. It is shown that good correlation exists between surface contamination related non-linearity and anomalous activity and frequency variations. A simple model of the surface contamination mechanism is presented and it is demonstrated that microscopic particles of either quartz or electrode material, in conjunction with a surface fluid film, can produce activity and frequency variations of the magnitude and form of those measured in quartz crystal resonators. The non-linearity of the mechanism is associated with the complex non-linear rheological properties of the fluid film. It is suspected that this film is composed of a phenylated silicone oil which is commonly used in diffusion pumps.

This thesis is concerned with some aspects of losses produced in the inductors and transformers of switched-mode power supplies (SMPS). The main subject of study is the eddy-current losses due to non-uniform current distribution across foil windings. These are calculated and measured for the cases when the foil constitutes an air-cored inductor, and when it is used as the secondary of a ferrite cored transformer. A coupled circuit model is used in which the foil inductor is subdivided into sets of mutually coupled filament coils. A formal method of solution of the resulting sets of equations is presented, and alternative mathematical methods of calculating self and mutual inductances are also compared. In addition to this work two other topics are addressed briefly. First, the magnetic fields in and around magnetic cores are measured, and compared with computed results using a 3D magnetostatic package, TOSCA. Secondly, core losses in ferrite cores due to sinusoidal and non-sinusoidal excitation are measured and calculated.

No description available.

No description available.

No description available.

No description available.

No description available.

No description available.

No description available.