No description available.

This dissertation is focused on developing a new passband behavioral model in order to account for asymmetric intermodulation distortion resulted from memory effect. First, a measurement setup is prepared to measure the AM/AM, AM/PM distortion, magnitudes and the phases of intermodulation (IMD) and fundamental (FUND) components which are created by the amplifier where phase is calculated only by measuring magnitudes. Then, responses of a sample amplifier are measured for different excitation situations (center frequency and tone spacing are swept). A new modeling technique, namely Odd Order Modeling (OOM), is proposed which has unequal time delay terms. The reason of unequal time delay addition is the change of effective channel length according to the average power passing through that channel. These unequal delays create asymmetry in the IMD components. General Power Series Expansion (GPSE) model is also extracted, OOM and GPSE model performances are compared by using NMSE metric. In order to improve model performance, even order terms with envelope of input are added. It is mathematically proven that even order terms with envelope of the input have contribution to IMD and FUND components&rsquo / . This improved version of modeling is named as Even Order modeling (EOM). EOM model performance is compared with the others&rsquo / performance for two-tone excitation measurement results. It is shown that EOM gives the most accurate result. Model performance is checked for unequal four-tone signal as well. EOM model is applied to baseband DPD circuit after making some modifications. Model linearization performance is compared with the performances of the other memory polynomial modeling techniques.

Ce mémoire présente la conception, la réalisation et la caractérisation d'un micro-haut-parleur en silicium destiné à des applications électroniques portables, telles que les tablettes et les téléphones cellulaires. L'objectif est d'évaluer le potentiel des microtechnologies pour améliorer la qualité sonore et le rendement électroacoustique, qui sont deux points faibles majeurs des micro-haut-parleurs actuels.En analysant les paramètres dont dépendent le rendement et la qualité sonore, nous montrons que le silicium monocristallin présente des propriétés particulièrement intéressantes pour réaliser la surface émissive et la suspension du transducteur. Une microstructure de la partie mobile est proposée pour satisfaire la double exigence d'une surface émissive très rigide, nécessaire à la qualité sonore, et d'une masse très faible, permettant d'augmenter le rendement. Les aimants et la bobine, qui constituent le moteur électrodynamique, sont également optimisés en utilisant conjointement des modèles analytiques et à éléments finis. La microfabrication du transducteur MEMS est étudiée, étape par étape. Elle repose sur l'utilisation d'un substrat SOI (silicium sur isolant), qui sert de base à la structuration des différents composants, et sur lequel sont rapportés des aimants massifs. La caractérisation électroacoustique des échantillons réalisés montre une très bonne qualité de reproduction sonore. Un niveau sonore de 80 dB à 10 cm est obtenu pour une puissance électrique de 0,5 W, ce qui place le rendement au niveau des micro-haut-parleurs du marché. Ces travaux montrent en outre que les technologies MEMS offrent des possibilités d'augmenter très largement le rendement.

[pt] A análise de desempenho de sistemas de comunicações nos quais um grande número de portadoras, de diferentes tipos e níveis de potência, acessam uma mesma não-linearidade, esbarra, inicialmente, na dificuldade de se determinar, com precisão e tempo de computação aceitável, a distribuição de potência do ruído de intermodulação gerado. O problema torna-se ainda mais complexo nos casos em que as portadoras têm envoltória variável no tempo, quando, além de se tornar mais difícil a determinação do espectro de intermodulção, há o aparecimento adicional da transferência de modulação das demais portadoras para a portadora desejada. Com o enfoque voltado para o desempenho de sistemas digitais, mostra-se que, utilizando-se determinados algoritmos e aproximações, o tratamento analítico torna-se factível. Com base na teoria apresentada, foi desenvolvido um programa de computador que permite avaliação de desempenho em situações bastante gerais. / [en] One of the difficulties arising in a theoretical performance evaluation of communication systems in which several carriers of various types are transmitted through a cammon nonlinear amplifier, is related to the problem of calculating, with good precision and within reasonable computer time, the power spectrum of the intermodulation noise generated at the output of the nonlinear device. The analysis becomes more complex when the carriers have time- varyng envelopes, since in such situations, besides the difficulty involved in determining the power and spectrum shape of the output signals and intermodulation products, an extra system impairment should be taken into account. Namely, the performance degradation due to modulation transfer from all the other carriers into the desired carrier. With focus on digital systems, it is shoun that, by using certain algorithms and approximations, theoretical performance analysis is feasible. Based on the results presented, a computer program was developed to evaluate performance in fairly general situations.

Passive Intermodulation is a type of distortion affecting the sensitive receiving signals in a cellular network, which is a growing problem in the telecommunication field. One way to mitigate this problem is through Passive Intermodulation Cancellation, where the predicted noise in a signal is modeled with polynomials. Recent experiments using neural networks instead of polynomials to model this noise have shown promising results. However, one drawback with neural networks is their lack of explainability. In this work, we identify a suitable method that provides explanations for this use case. We apply this technique to explain the neural networks used for Passive Intermodulation Cancellation and discuss the result with domain expertise. We show that the input space as well as the architecture could be altered, and propose an alternative architecture for the neural network used for Passive Intermodulation Cancellation. This alternative architecture leads to a significant reduction in trainable parameters, a finding which is valuable in a cellular network where resources are heavily constrained. When performing an explainability analysis of the alternative model, the explanations are also more in line with domain expertise. / Passiv Intermodulation är en typ av störning som påverkar de känsliga mottagarsignalerna i ett mobilnät. Detta är ett växande problem inom telekommunikation. Ett tillvägagångssätt för att motverka detta problem är genom passiv intermodulations-annullering, där störningarna modelleras med hjälp av polynomiska funktioner. Nyligen har experiment där neurala nätverk används istället för polynomiska funktioner för att modellera dessa störningar påvisat intressanta resultat. Användandet av neurala nätverk är dock förenat med vissa nackdelar, varav en är svårigheten att tyda och tolka utfall av neurala nätverk. I detta projekt identifieras en passande metod för att erbjuda förklaringar av neurala nätverk tränade för passiv intermodulations-annullering. Vi applicerar denna metod på nämnda neurala nätverk och utvärderar resultatet tillsammans med domänexpertis. Vi visar att formatet på indatan till neurala nätverket kan manipuleras, samt föreslår en alternativ arkitektur för neurala nätverk tränade för passiv intermodulations-annullering. Denna alternativa arkitektur innebär en avsevärd reduktion av antalet träningsbara parametrar, vilket är ett värdefullt resultat i samband med mobilnät där det finns kraftiga begränsningar på hårdvaruresurser. När vi applicerar metoder för att förklara utfall av denna alternativa arkitektur finner vi även att förklaringarna bättre motsvarar förväntningarna från domänexpertis.

This work investigates the fundamental device limits related to operational voltage constraints and linearity in state-of-the-art silicon-germanium (SiGe) heterojunction bipolar transistors (HBTs) in order to support the design of robust next-generation high-frequency transceivers. This objective requires a broad understanding of how much "usable" voltage exists compared to conventionally defined breakdown voltage specifications, so the role of avalanche-induced current-crowding (or "pinch-in") effects on transistor performance and reliability are carefully studied. Also, the effects of intermodulation distortion are examined at the transistor-level for new and better understanding of the limits and trade-offs associated with achieving enhanced dynamic range and linearity performance on existing and future SiGe HBT technology platforms. Based on these investigations, circuits designed for superior dynamic range performance are presented.

<p>Most work on amplifier linearity has focused on the transconductance (gm) linearity, but there is increasing evidence that the voltage-dependence of the gate-source capacitance (Cgs) plays an important role in the linearity of emerging devices. This work addresses the capacitance contribution by incorporating the nonlinearities attributed to the voltage dependency of Cgs of a general FET on a circuit-level Cg amplifier model.</p> <p>An amplifier model including a voltage-dependent Cgs, and a voltage-dependent gm is studied using harmonic analysis and Volterra series. A closed form expression for the  third-order intercept point (IP3) of the amplifier, which depends on the nonlinear coefficients of Cgs, is obtained. A simple design rule, and a formula for the reduction of the IP3 due to the voltage-dependent Cgs are also presented. </p> <p>As application examples, the linearity of an amplifier based on a specific device is analyzed for two cases by extracting the nonlinear circuit parameters of the device. First for an analytic model of a bulk mosfet. Second for a one-dimensional, ballistic, coaxially gated Si nanowire. For low frequencies of design, the distortion introduced by gm is predominant, but for high frequencies it is obscured by the distortion coming from Cgs.</p> <p>We conclude that taking into account the voltage-dependence of Cgs is crucial when predicting the linearity behavior of a Cg amplifier, either designed for high-frequency operation, or based on a device operating near the quantum capacitance limit. </p>

This thesis presents novel radio frequency microelectromechanical (RF MEMS) circuits based on the three-dimensional (3-D) micromachined coplanar transmission lines whose geometry is re-configured by integrated microelectromechanical actuators. Two types of novel RF MEMS devices are proposed. The first is a concept of MEMS capacitors tuneable in multiple discrete and well-defined steps, implemented by in-plane moving of the ground side-walls of a 3-D micromachined coplanar waveguide transmission line. The MEMS actuators are completely embedded in the ground layer of the transmission line, and fabricated using a single-mask silicon-on-insulator (SOI) RF MEMS fabrication process. The resulting device achieves low insertion loss, a very high quality factor, high reliability, high linearity and high self actuation robustness. The second type introduces two novel concepts of area efficient, ultra-wideband, MEMS-reconfigurable coupled line directional couplers, whose coupling is tuned by mechanically changing the geometry of 3-D micromachined coupled transmission lines, utilizing integrated MEMS electrostatic actuators. The coupling is achieved by tuning both the ground and the signal line coupling, obtaining a large tuneable coupling ratio while maintaining an excellent impedance match, along with high isolation and a very high directivity over a very large bandwidth. This thesis also presents for the first time on RF nonlinearity analysis of complex multi-device RF MEMS circuits. Closed-form analytical formulas for the IIP3 of MEMS multi-device circuit concepts are derived. A nonlinearity analysis, based on these formulas and on  measured device parameters, is performed for different circuit concepts and compared to the simulation results of multi-device  conlinear electromechanical circuit models. The degradation of the overall circuit nonlinearity with increasing number of device stages is investigated. Design rules are presented so that the mechanical parameters and thus the IIP3 of the individual device stages can be optimized to achieve a highest overall IIP3 for the whole circuit.The thesis further investigates un-patterned ferromagnetic NiFe/AlN multilayer composites used as advanced magnetic core materials for on-chip inductances. The approach used is to increase the thickness of the ferromagnetic material without increasing its conductivity, by using multilayer NiFe and AlN sandwich structure. This suppresses the induced currents very effectively and at the same time increases the ferromagnetic resonance, which is by a factor of 7.1 higher than for homogeneous NiFe layers of same thickness. The so far highest permeability values above 1 GHz for on-chip integrated un-patterned NiFe layers were achieved. / <p>QC 20140328</p>

Nanocomposites of polyethylene and metal oxide nanoparticles have shown to be a feasible approachto the next generation of insulation in high voltage direct current cables. In order to reach an operationvoltage of 1 MV new insulation materials with reduced conductivity and increased breakdown strengthas compared to modern low-density polyethylene (LDPE) is needed.In this work polyethylene MgO nanocomposites for electrical insulation has been produced andcharacterized both from an electrical and material perspective. The MgO nanoparticles weresynthesized into polycrystalline nanoparticles with a large specific surface area (167 m2 g–1). Meltprocessing by extrusion resulted in evenly dispersed MgO nanoparticles in LDPE for the silane surfacemodified MgO as compared to the unmodified MgO. All systems showed a reduction in conductivityby up to two orders of magnitude at low loading levels (1–3 wt.%), but where the surface modifiedsystems were able to retain reduced conductivity even at loading levels of 9 wt.%. A maximuminteraction radius to influence the conductivity of the MgO nanoparticles was theoretically determinedto ca. 800 nm. The interaction radius was in turn experimentally observed around Al2O3 nanoparticlesembedded in LDPE using Intermodulation electrostatic force microscopy. By applying a voltage on theAFM-tip charge injection and extraction around the Al2O3 nanoparticles was observed, visualizing theexistence of additional localized energy states on, and around, the nanoparticles. Ptychography wasused to reveal nanometre features in 3D of electrical trees formed under DC-conditions. Thevisualization showed that the electrical tree grows by pre-step voids in front of the propagatingchannels, facilitating further growth, much in analogy to mechanical crack propagation (Griffithconcept). An electromechanical effect was attributed as possible mechanism for the formation of the voids. / Nanokompositer av polyeten och metalloxidpartiklar anses vara möjliga material att använda i morgondagens isolationshölje till högspänningskablar för likström. För att nå en transmissionsspänning på 1 MV behövs isolationsmaterial som i jämförelse med dagens polyeten har lägre elektrisk ledningsförmåga, högre styrka mot elektriskt genomslag och som kan kontrollera ansamling av rymdladdningar. De senaste årens forskning har visat att kompositer av polyeten med nanopartiklar av metalloxider har potential att nå dessa egenskaper. I det här arbetet har kompositer av polyeten och nanopartiklar av MgO för elektrisk isolation producerats och karaktäriserats. Nanopartiklar av MgO har framställts från en vattenbaserad utfällning med efterföljande calcinering, vilket resulterade i polykristallina partiklar med en mycket stor specifik ytarea (167m2 g-1). MgO-nanopartiklarna ytmodifierades i n-heptan genom att kovalent binda oktyl(trietoxi)silan och oktadekyl(trimetoxi)silan till partiklarna för att skapa en hydrofob och skyddande yta. Extrudering av de ytmodifierade MgO nanopartiklarna tillsammans med polyeten resulterade i en utmärkt dispergering med jämnt fördelad partiklar i hela kompositen, vilket ska jämföras med de omodifierade partiklarna som till stor utsträckning bildade agglomerat i polymeren. Alla kompositer med låg fyllnadsgrad (1–3 vikt% MgO) visade upp till 100 gånger lägre elektrisk konduktivitet jämfört med värdet för ofylld polyeten. Vid högre koncentrationer av omodifierade MgO förbättrades inte de isolerande egenskaperna på grund av för stor andel agglomerat, medan kompositerna med de ytmodifierade fyllmedlen som var väl dispergerade behöll en kraftig reducerad elektrisk konduktivitet upp till 9 vikt% fyllnadshalt. Den minsta interaktionsradien för MgO-nanopartiklarna för att minska den elektriska konduktiviten i kompositerna fastställdes med bildanalys och simuleringar till ca 800 nm. Den teoretiskt beräknade interaktionsradien kompletterades med observation av en experimentell interaktionsradie genom att mäta laddningsfördelningen över en Al2O3-nanopartikle i en polyetenfilm med intermodulation (frekvens-mixning) elektrostatisk kraftmikroskop (ImEFM), vilket är en ny AFM-metod för att mäta ytpotentialer. Genom att lägga på en spänning på AFM-kantilevern kunde det visualiseras hur laddningar, både injicerades och extraherades, från nanopartiklarna men inte från polyeten. Det tolkades som att extra energinivåer skapades på och runt nanopartiklarna som fungerar för att fånga in laddningar, ekvivalent med den gängse tolkningen att nanopartiklar introducera extra elektronfällor i den polymera matrisen i nanokompositer. Nanotomografi användes för att avbilda elektriska träd i tre dimensioner. Avbildningen av det elektriska trädet visade att tillväxten av trädet hade skett genom bildning av håligheter framför den framväxande trädstrukturen. Håligheterna leder till försvagning av materialet framför det propagerande trädet och förenklar på det sättet fortsatt tillväxt. Bildningen av håligheter framför trädstrukturen uppvisar en analogi till propagering av sprickor vid mekanisk belastning, i enlighet med Griffiths koncept. / <p>QC 20161006</p>

The growth in the number of wireless devices and applications underscores the need for characterizing and mitigating interference induced problems such as distortion and blocking. A typical interference scenario involves the detection of a small amplitude signal of interest (SOI) in the presence of a large amplitude interfering signal; it is desirable to attenuate the interfering signal while preserving the integrity of SOI and an appropriate dynamic range. If the frequency of the interfering signal varies or is unknown, an adaptive notch function must be applied in order to maintain adequate attenuation. This work explores the performance space of a phase cancellation technique used in implementing the desired notch function for communication systems in the 1-3 GHz frequency range. A system level model constructed with MATLAB and related simulation results assist in building the theoretical foundation for setting performance bounds on the implemented solution and deriving hardware specifications for the RF notch subsystem devices. Simulations and measurements are presented for a Low Noise Amplifer (LNA), voltage variable attenuators, bandpass filters and phase shifters. Ultimately, full system tests provide a measure of merit for this work as well as invaluable lessons learned. The emphasis of this project is the on-wafer LNA measurements, dependence of IC system performance on mismatches and overall system performance tests. Where possible, predictions are plotted alongside measured data. The reasonable match between the two validates system and component models and more than compensates for the painstaking modeling efforts. Most importantly, using the signal to interferer ratio (SIR) as a figure of merit, experimental results demonstrate up to 58 dB of SIR improvement. This number represents a remarkable advancement in interference rejection at RF or microwave frequencies.