Modelling and Simulation of the Diode Split Transformer

The development of horizontal deflection circuits for television receivers is commonly based on experiments and trial-and-error methods. This design practice is expensive due to the long development process, and the designers of the circuit face the problem of reducing the development time and simultaneously meeting increasing demands on the functionality and performance. Consequently, the designers need to improve their knowledge of the circuit and have better tools to help them in the design process. The complexity of the circuit is high and the stray effects in the power components used have a significant influence on the picture quality. The most critical component is undoubtedly the diode split transformer (DST). Therefore, if developing a simulation model of the DST is possible, a significant step has been taken in the attempt to model the entire horizontal deflection circuit and to obtain a design tool. The objective of the research project documented in this thesis is to develop such models of the DST and to verify them by actual measurements.

Prior to the development of the simulation models, the most important properties of the DST are presented and discussed. This includes its interaction with the rest of the horizontal deflection circuit and its characteristic frequency behaviour. Also, a survey of the physical properties of the windings, the ferrite core, and the discrete components in the DST is given. Based on this information, the overall requirements of the model are formulated. To fulfill these requirements, two different model structures are developed and applied to a layer-wound DST. The first model is simple two-winding model with parameters found from a few measurements, while the other is a multi-winding model using parameters found from the geometry and material characteristics of the DST. Thus, the first model requires that the DST has already been designed and constructed, while the second model can be set up prior to the actual construction of the DST.

The two-winding equivalent circuit diagram is derived from a two-port description of the inductive, the resistive, and the capacitive effects in the DST. The methods are given to measure the inductive parameters, while the capacitive parameters are calculated from analytical equations by which the values of the characteristics resonance frequencies are modelled exactly. The resistive model parameters are frequency dependent and are found by an optimization algorithm that minimizes the difference between the measured and the calculated damping in the frequency-domain. The obtained results are in excellent agreement with the measurements in both the frequency-domain and the time-domain. The model is simple to set up and requires very few measurements; to find all the parameters only the open-circuit primary impedance and the number of turns in the windings must be known. An extended method is given in which the characteristics resonance frequency in the primary impedance is also measured when one or more layers in the high-voltage winding are short circuited. Using the value of the resonance frequency in the capacitive model parameter calculations, a slightly better accuracy of the model in short-circuit conditions is given.

In the development of the multi-winding model, the focus is on a model of the winding structure of the DST, while a constant magnetizing inductance and an average core-loss model describes the ferrite core. Two different approaches have been used to model the frequency-dependent inductive and resistive effects in the winding structure. The first approach uses model parameters found from two-dimensional numerical calculations, while the other approach uses model parameters found by one-dimensional analytical equations. The capacitive effects in the winding structure of the DST are modeled using analytical equations and methods to connect the calculated distributed capacitances to the nodes of the inductive and resistive model are developed. The average core-loss model is set up using only data-sheet information and added to the equivalent circuit model of the winding structure by a single resistor. As for the two-winding model, the results obtained with the multi-winding model are in excellent agreement with the measurements in both the frequency and the time-domain.

To verify the developed models of the DST, measuring systems have been constructed. The systems for frequency and time-domain measurements are well-known, while the overall power loss in the DST is measured using calorimetric wattmeters. Two different calorimetric wattmeters are designed and constructed. The first wattmeter is simple and fast to build, while the other wattmeter, constructed in a parallel research project, requires much more time to build. The basic difference between the two wattmeters is that in the advanced calorimetric wattmeter, the ambient test temperature can be specified. All the measuring systems have proven to be reliable and to give accurate results.

By the research documented in this thesis a significant step towards a complete simulation model of the horizontal deflection circuit is made. The developed models of the layer-wound DST are fast to set up and give accurate results. The two-winding model is well-suited when the designers of the horizontal deflection circuit need a model of a new DST that has already been designed, while the multi-winding model is excellent when the designers are involved in the in the design process of the DST. In the process of developing the models, many problems related to the model wire-wound magnetic components have been investigated and new solutions have been proposed to some of the problems. Although the investigations have focused on the actual layer-wound DST, the modelling methods can directly or with some modifications be applied to many other wire-wound magnetic components.