ELECTRONICS

Author: FRANCO PELÁEZ FRANCISCO JAVIER.
Year: 2004.
University: COMPLUTENSE DE MADRID.
Place of defense: FACULTAD DE CIENCIAS FISICAS.
Place of preparation: FACULTAD DE CIENCIAS FISICAS.

Summary: The memory of this thesis is facing the problem of the degradation of the operational amplifiers built in bipolar technology when they are irradiated with neutrons rápidos.Este type of radiation damage produced by posting on the Web cirstalina of semiconductor, which leads to the degradation of the components internos.Al be associated with them, provides for a modification of the electrical parameters of external operational amplifiers. This has been proven experimentally in amplicadores operational in a commercial source of neutrons especially dedicada.Para This will design an automated system characterization, which was optimized to conduct a comprehensive monitoring of the parameters during the irradiation process. It was also decided to extrapolate rasultados obtained amplifiers opreacionales those devices in which there was some integrated amplifier this tipo.Estos components were instrumentation amplifiers, consisting generally of three operational amplifiers; references tensiónk, inthe that amplirficador plays an important role in improving the characteristics of output; finally be estudaron converters D / A, in which there is both a amplificadro operational charge of converting flow in as an internal reference voltage of tensión.En general, theoretical and experimental results match perfectamente.Esta memory is supported by national and international publications.

Author: MARSAL COLL ANDREU.
Year: 2004.
University: BARCELONA.
Place of defense: FACULTAD DE FÍSICA.
Place of preparation: DEPT. ELECTRÓNICA, FAC. FÍSICA.

Author: GÒELL VILÀ FRANK.
Year: 2004.
University: BARCELONA.
Place of defense: FACULTAD DE FÍSICA.
Place of preparation: FACULTAD FÍSICA.

Summary: The materials doped with ions Tm3 + emerge as a solid-state lasers in the spectral region of 2um (micron), on the one hand because of the possibility of bombearlos with diode laser around the 800 nm (nanometers) and another by the wide tuning they offer. The potential applications are in the medical field and in the remote sensors or radar systems of the atmosphere, because the water represents strong absorption coefficients in this wavelength.

Author: CALVO IBAÑEZ OSCAR ALBERTO.
Year: 2004.
University: POLITÉCNICA DE CATALUÑA.
Place of defense: Escola Uni. I.T.I..
Place of preparation: C4 Despatx Direcció nord.

Summary: The unifying theme of this paper is the phenomenon of nonlinear electronic systems synchronization in the presence of white noise and obtaining experimental results. We have worked with systems with detection threshold, excitable and oscillators. In Chapter 1 is an introduction to the topic, describing the motivation, the state of the art, and the major contributions of the thesis. In Chapter 2 we study the phenomenon of stochastic resonance in the electronic circuit Chua operating in chaotic regime, with a high-frequency modulation, ie when the frequency is forcing the order of the natural frequency (Buddha) Chua. The case also analyze the frequency of forcing a much smaller (Buddha), noting a harmonic resonance frequency of forcing close to (Buddha) or one of its harmonics. In chapter 3 we show that it is possible to observe, in chaotic systems consistent with the phenomenon of resonance. In particular we show that the circuit Chua, operating under chaotic and in the presence of noise exhibited oscillations whose regularity is optimal for a value intermediate intensity noise. In Chapter 4 we extend the concept of stochastic resonance systems excitable, analyzing the response of a neuron forced simultaneously by two signals periodic noise and low intensity. We concentrate on frequency signals: kf0; (k + 1) f0; ... (n + k) f0 with k> 1. The output of the neuron is a sequence of pseudo pulses spaced intervals. An optimal in the SNR when pulses are spaced 1/f0, ie stochastic resonance occurs at a frequency not present in the excitement. In Chapter 5 demonstrates experimentally the regime of the resonance stochastic subarmónica in response to an electronic circuit type Schmitt Trigger, excited by noise and N frequency components: [kf0 + Ãf], [(k + 1) f0 + Ãf], â | [(n + k) f0 + Ãf]. We reproduce experiments psicoacústicos Detection ringer (pitch) using electronic neurons. The frequency at which produces the maximum resonance, there is power in the entry and call stochastic resonance ghost. In Chapter 6 we continue the study of resonance subharmónica stochastic into the channels formed by pll (phase locked loops). We started the study results were analyzed in an experimental prototype mail. This analog circuit implemented on the basis of PLLs forced quasi periodically investigate the structure resonances in the dynamics. In Chapter 7 examines a phenomenon that occurs in certain chaotic oscillators forced synchronization phase, in which two different systems ranging unabated. The imperfect synchronization phase appears in systems in which the oscillators can present times of infinite return, eg Chaotic oscillators with a fork type saddle, as in the case of Lorentz oscillator. We demonstrate this phenomenon by implementing electronic circuit Lorenz. In Chapter 8 we look at a different kind of synchronization between two dynamic systems: study the synchronization scheme described recently anticipated for several dynamic systems, including systems chaotic and in the presence of noise. We use simple models to show the minimum conditions to reproduce the phenomenon and the most significant. In Chapter 9 we apply the concepts introduced in the previous chapter to a case study: non-linear systems. We studied the regime of sync early in neurons coupled unidireccionalmente and subject to a forcing aperiódico that produces unpredictable behavior.

Author: CONESA ROCA ALFONSO.
Year: 2005.
University: POLITÉCNICA DE CATALUÑA.
Place of defense: Sala d'Actes, Edifici EUETIB.
Place of preparation: C4 Jordi Girona, 1-3, campus nord 08034.

Summary: In this thesis examines the implementation of the rectification controlled converters DC-DC several switches tank resonant series. The converters are considered the PRC, the SRC and SPRC. In these structures discusses the possibilities offered by these regulatory converters, in order to maximize the robustness to changes in load and the input voltage. After the first introductory chapter to the topic of the thesis, in the second chapter are deducted new converters derived from the classics. Get mathematical models of converters resulting from the replacement of the grinding stages and summarizes the new structures grinding. The new converters are called PRC-CR, SRC-CR and SPRC-CR, in a clear allusion to its new phase of correction controlled (Controlled Rectification). The third chapter introduces the drivers developed for the new converters. The control of the converter is operating in the block rectifier through two basic types of action. The first feedback variable state resonant tank, through a controller nonlinear One-Cycle. The second feedback with variable output state, through a linear regulator IP. The fourth chapter examines the structure of the converter PRC-CR using two methods of analysis. The first method is the proposed alignment of the first harmonic, and its analysis will be deducted interesting conclusions. First points of work to enable optimal regulation in the converter given load conditions thereof. Is obtained dependence on the variable itself out of each converter connection with the relationship of conduction block rectifier and propose design criteria converter. The proposed methodology is verified by simulation of the converter. The second method of analysis is based on the solution of the mathematical model in state space converter. Under this analysis will be obtained again the relationship of the output voltage with respect to the relationship driving the grinding stage and the parameters of the converter. Finally, it compares the results of both methods of analysis. In the fifth and sixth chapter discusses converters SRC-CR and SPRC-CR respectively equally to the proposal in the fourth chapter of PRC-CR. In the seventh chapter presents the experimental results obtained from the prototype converters developed. Finally in the eighth chapter presents the findings from the doctoral thesis. It assesses the benefits and limitations observed in the structures developed. The proposed approach to possible improvements in terms of the non idealidades converter and the controller proposed opening future lines of investigation that will complement this work.

Author: MASSANA HUGAS IMMACULADA.
Year: 2006.
University: POLITÉCNICA DE CATALUÑA.
Place of defense: SALA D'ACTES DE L'FME.
Place of preparation: U FACULTAT DE MATEMATIQUES I ESTADISTICA SUD.

Author: MENDEZ VILLEGAS MIGUEL ANGEL.
Year: 2006.
University: POLITÉCNICA DE CATALUÑA.
Place of defense: AULA MÀSTER, EDIFICI A3 CAMPUS NORD.
Place of preparation: Edifici C4 DESPATX 310 ND.

Summary: The evolution of the semiconductor technology has enabled during the last decade the realization of wireless communication systems with an increasing degree of integration. For some standards the single-chip integration of the whole communication system (SoC paradigm) has been a reality. Such SoCs integrate the digital section, which implements the baseband functionalities, together with the radio, which operates at high frequencies (RF). The realization and deployment of these Communications Systems-on-Chip (Com-SoC) requires low cost, low power, high performance, high integration density, and small size manufacturing technologies. In order to meet all these requirements the CMOS technology is usually selected. However, the single chip integration of mixed-signal circuits in CMOS technology faces severe problems due to signal integrity issues, which are aggravated as the technologic scales down. Among these problems, the substrate noise and its effects on the analog/RF sections has become a critical limiting factor for the implementation of Com-SoCs. This thesis work boards two fundamental aspects of the substrate noise problematic: the characterization of the substrate noise generation, with emphasis in the frequency domain characteristics, and its impact on some RF CMOS circuits. In our work the study of the substrate noise spectrum is addressed in order to determine its content in frequency bands where it can be potentially harmful for the correct operation of RF/analog circuits. The substrate noise spectrum characteristics are derived from the characteristics of the digital supply current (the dominant source of substrate noise) in combination with what is named the substrate coupling transfer function (a linear transformation that models the propagation of noise from the digital section to any substrate node of interest on the Com-SoC surface). The digital supply current spectrum is analyzed by considering it a stochastic process. This approach has fully allowed characterizing the digital current spectrum with a reduced set of parameters. The substrate noise coupling transfer function is obtained from the chip level description of the Com-SoC. The identification of the electrical elements that compose such transfer function, their relevance and a methodology to extract them are also addressed in this work. Several mixed-signal test chips, implemented in a CMOS process on high resistive substrate, are used to illustrate and analyze practical examples of substrate noise generation and its dependence with some parameters (clock frequency, digital circuit timing operation, etc). The study of the substrate noise impact on RF CMOS circuits is focused on voltage controlled oscillators (VCOs) and active mixers. An analysis of the VCO phase noise degradation caused by substrate noise is performed, starting with an experimental approach using digital circuits as realistic on-chip noise generators. Next, theoretical models are derived, which allow the identification of the relevant noise impact mechanisms and its dependence on several operating conditions of the VCO. The effect of substrate noise found at different frequency bands and its up-conversion or down-conversion to sidebands around the oscillator fundamental are specifically boarded and explained. The proposed expressions for the substrate noise impact on the LC-tank VCOs are validated experimentally using the same test chips used for substrate noise generation analysis. Regarding the impact of substrate noise on mixers, the same approach is applied. The results show how the presence of substrate noise in different frequency bands can affect the intermediate frequencies (IF) band. The dominant substrate noise impact mechanisms are identified. All these investigations pave the way to a new approach for substrate noise aware design, not based only on layout isolation techniques, but also on clever circuit level design of the RF section of the Com-SoC.

Author: MINO AGUILAR GERARDO.
Year: 2006.
University: POLITÉCNICA DE CATALUÑA.
Place of defense: SEMINARIO I.1, ETSEIB.
Place of preparation: C4 Jordi Girona, 1-3, campus nord 08034.

Summary: The induction motor has been and will remain one of the electric motors with more use in many areas. Thus has been the subject of study over the past few decades. The form of control has evolved tremendously from the connection and disconnection from direct energy through the typical voltage control frequency and reaching the controls as the type vector control field oriented direct and indirect or direct control torque. Also, because of their high performance, these have evolved to be a high performance engines. However, in most cases, their use is not done in the area of maximum benefits and burdens that are managed and reduced speeds, causing its performance is significantly reduced and consume more energy that is required mainly in small machines . That is why many studies have focused on improving the performance of these engines in permanent regime, since most of time of use takes place in that scheme. In summary, the objective of this line of research, better known as optimization of efficiency, should enable a dynamic response stable, with a couple of quick off load, but mostly there is a significant increase in the performance of the engine. This operation is carried out by reducing the flow engine to find the minimum loss for permanent regime. But still there is a method that can be generalized. It requires optimal control system to provide dynamic in a decrease in energy consumption reflected in the reduction of losses. There are currently two approaches with which this succeeds this point of minimum losses: search methods point minimum loss and optimization model for losses. The control schemes for high performance of an induction motor invariably require knowledge of at least one of the parameters of the engine. The values of the parameters of the engine used in the controller, and therefore, they have to be identified offline during the adjustment drive. However, since all parameters inevitably vary during operation of the actuator, it is desirable to improve its operations by adding a parameter estimator online. Such a situation has led to the development of a large number of identification methods offline and estimation of parameters in line during the past two decades. This thesis presents an estimator flow engine for optimization of efficiency based on the model of losses engine using neural networks and includes an estimate of some engine parameters that impact directly on the control and optimization of efficiency. It has been necessary to analyze the approach search methods starting with the definition of algorithm Rosenbrock. In addition to an analysis of sensitivity of the parameters involved in the control. Here arises an engine model and defines the loss, finishing with a loss model for the optimization. Estimators are introduced to the inductance of magnetization, the resistance of the rotor and the stator in the control loop. Finally he trains a neural network for the estimated flow to optimize efficiency. The operation of these methods and their benefits have been verified by simulation results and experimental results.

Author: PARRILLA SANCHEZ MANUEL.
Year: 2006.
University: NACIONAL DE EDUCACIÓN A DISTANCIA.
Place of defense: E.T.S. INGENIERÍA INFORMÁTICA.
Place of preparation: E.T.S. INGENIERÍA INFORMÁTICA.

Summary: The design of robust multivariable control systems is a complex task that requires a great knowledge of the subject by the designer. Moreover, in many cases will be required iterative procedures of trial and error, which is very important experience. On the other hand, evolutionary algorithms and have shown a very effective technique in solving optimization problems multiobjetivo, within which can encuadernarse control problems. Using evolutionary algorithms are achieved, in a more simple, very acceptable results. In this thesis, developed a methodology for broad application for the design of control systems using evolutionary algorithms. The result was a very flexible approach, allowing work from a nearby user, deal directly with the design specifications, and use of parameters simultaneously own domain timeline and the frequency. The methodology was applied designed to three specific control problems: The first one was designed for a driver LQ a commercial airliner. The algorithm was responsible in this case for the determination of the nuances of weight QyR own method LQ. In the second problem, the algorithm performed the direct line of a predetermined structure controller for the same commercial aircraft that the driver LQ already discussed, the difference is that now done directly tune the algorithm, without the support of any other method . Finally, we applied the methodology for determining the order of the transfer functions woe to his tune, a driver for him reduce rates in a way high-speed vessel. The implementation of the algorithm to the latter problem posed a high cost computations, and the algorithm is paralelizó and was conducted in a multi-processor supercomputer. The success stories in the three problems treaty confirmed the validity of the method used. Given the high computational cost that, as already mentioned, is the application of the methodology developed for complex control problems, it was concluded with the thesis a study on the stay of the algorithm. The algorithm, once paralyzed, was carried on a cluster of PCs. The results showed that in this case the algorithm was easily paralelizable and that the use of PC clusters can be achieved significant acceleration at an affordable cost.