Filtre (électronique) | EPFL Graph Search

En électronique, un filtre est un circuit linéaire qui transmet une grandeur électrique (courant ou tension) selon sa répartition en fréquences. Le filtre transforme l'histoire de cette grandeur d'entrée (c'est-à-dire ses valeurs successives depuis un certain temps) en une grandeur de sortie. Pour raisonner sur les filtres électroniques, on les considère comme des quadripôles dont les grandeurs électriques d'entrée et de sortie seraient un signal, même quand celles-ci ne servent pas à transmettre de l'information (comme dans le cas des filtres d'alimentation). Cette approche exploite l'important effort mathématique fourni dans le domaine du traitement du signal. Les filtres électroniques analogiques utilisent, éventuellement en combinaison avec des composants actifs (amplificateurs), des résistances, qui sont des composants indifférents aux variations dans le temps, avec d'autres dipôles qui emmagasinent l'énergie pour la restituer ultérieurement, appelés réactances : ; sous forme é

Design of a 2.4 Ghz BAW-Based CMOS Transmitter

Matteo Contaldo

In recent years, bulk acoustic wave resonators (BAW) in combination with RF circuits have shown a big potential in achieving the low-power consumption and miniaturization level required to address wireless sensor nodes (WSN) applications. A lot of work has been focused on the receiver side, by integrating BAW resonators with low noise amplifiers (LNA) and in frequency synthesis with the design of BAW-based local oscillators, most of them working at fixed frequency due to their limited tuning range. At the architectural level, this has forced the implementation of several single channel transceivers. This thesis aims at exploring the use of BAW resonators in the transmitter, proposing an architecture capable of taking full advantage of them. The main objective is to develop a transmitter for WSN multi-channel applications able to cover the whole 2.4 GHz ISM band and enable the compatibility with wide-spread standards like Bluetooth and Bluetooth Low Energy. Typical transmissions should thus range from low data rates (typically tens of kb/s) to medium data rates (1 Mb/s), with FSK and GFSK modulation schemes, should be centered on any of the channels provided by these standards and cover a maximum transmission range of some tens of meters. To achieve these targets and circumvent the limited tuning range of the BAW oscillator, an up-conversion transmitter using wide IF is used. The typical spurs problems related to this transmitter architecture are addressed by using a combined suppression based on SSB mixing and selective amplification. The latter is achieved by cointegration of a high efficiency power amplifier with BAW resonators, which allows performing spurs filtering while preserving the efficiency. In particular the selective amplifier is designed by including in the PA analysis the BAW resonator parameters, which allows integrating the BAW filter into the passive network loading the amplifier, participating in the drain voltage shaping. Finally, the frequency synthesis section uses a fractional division plus LC PLL filtering and further integer division to generate the IF signals and exploit the very-low BAW oscillator phase noise. The transmitter has been integrated in a 0.18 µm standard digital CMOS technology. It allows addressing the whole 80 MHz wide 2.4 GHz ISM band. The unmodulated RF frequency carrier demonstrates a very-low phase noise of –136 dBc/Hz at 1 MHz offset. The IF spurs are maintained lower than –48 dBc, satisfying the international regulations for output power up to 10 dBm without the use of any quadrature error compensation in the transmitter. This is achieved thanks to the rejection provided by the SSB mixer and the selective amplifier, which can reach drain efficiency of up to 24% with integrated inductances, including the insertion losses of the BAW filter. The transmitter consumes 35.3 mA at the maximum power of 5.4 dBm under 1.6 V (1.2 V for the PA), while transmitting a 1 Mb/s GFSK signal and complying with both Bluetooth and Bluetooth Low Energy relative and absolute spectrum requirements.

EPFL2011

Photocarrier-induced band-gap renormalization and ultrafast charge dynamics in black phosphorus

Majed Chergui, Gianmarco Gatti, Michele Puppin

With its tunable band-gap and its unique optical and electronic properties black phosphorus (BP) opens exciting opportunities for optoelectronic nanotechnology. The band-gap extends from the visible to the mid-infrared spectral range, as a function of sample thickness and external parameters such as electric field and pressure. This, combined with the saturable absorption and in-plane anisotropic optical properties, makes BP a versatile platform for realizing polarization-sensitive photodetectors and absorbers. Although its near-equilibrium properties have been intensively studied, the development of efficient ultrafast optical devices requires detailed knowledge of the temporal dynamics of the photoexcited hot-carriers. Here we address the electronic response of BP to an ultrafast laser excitation, by means of time-and angle-resolved photoelectron spectroscopy. Following the optical excitation, we directly observe a shift of the valence band (VB) position, indicative of band-gap renormalization (BGR). Our data also show that the hole population in the VB relaxes with a characteristic time tau(VB) = 10.2 +/- 1.0 ps, while the lifetime of the electrons accumulated at the minimum of the conduction band is tau(CB) = 1.9 +/- 0.2 ps. The experimental results are well reproduced by ab initio calculations of the out-of-equilibrium electronic properties. Our study sets the reference for the ultrafast carrier dynamics in BP and demonstrates the material's ultrafast BGR, which is promising for optoelectronic switches.

2019

CMOS ASIC Design of Multi-frequency Multi-constellation GNSS Front-ends

Saeed Ghamari

With the emergence of the new global navigation satellite systems (GNSSs) such as Galileo, COMPASS and GLONASS, the US Global Positioning System (GPS) has new competitors. This multiplicity of constellations will offer new services and a much better satellite coverage. Public regulated service (PRS) is one of these new services that Galileo, the first global positioning service under civilian control, will offers. The PRS is a proprietary encrypted navigation designed to be more reliable and robust against jamming and provides premium quality in terms of position and timing and continuity of service, but it requires the use of FEs with extended capabilities. The project that this thesis starts from, aims to develop a dual frequency (E1 and E6) PRS receiver with a focus on a solution for professional applications that combines affordability and robustness. To limit the production cost, the choice of a monolithic design in a multi-purpose 0.18 µm complementary metal-oxide-semiconductor (CMOS) technology have been selected, and to reduce the susceptibility to interference, the targeted receiver is composed of two independent FEs. The first ASIC described here is such FEs bundle. Each FE is composed of a radio frequency (RF) chain that includes a low-noise amplifier (LNA), a quadrature mixer, a frequency synthesizer (FS), two intermediate frequency (IF) filters, two variable-gain amplifiers (VGAs) and two 6-bit flash analog-to-digital converters (ADCs). Each have an IF bandwidth of 50 MHz to accommodate the wide-band PRS signals. The FE achieves a 30 dB of dynamic gain control at each channel. The complete receivers occupies a die area of 11.5 mm2 while consuming 115 mW from a supply of a 1.8 V. The second ASIC that targets civilian applications, is a reconfigurable single-channel FE that permits to exploit the interoperability among GNSSs. The FE can operate in two modes: a ¿narrow-band mode¿, dedicated to Beidou-B1 with an IF bandwidth of 8 MHz, and a ¿wide-band mode¿ with an IF bandwidth of 23 MHz, which can accommodate simultaneous reception of Beidou-B1/GPS-L1/Galileo-E1. These two modes consumes respectively 22.85 mA and 28.45 mA from a 1.8 V supply. Developed with the best linearity in mind, the FE shows very good linearity with an input-referred 1 dB compression point (IP1dB) of better than -27.6 dBm. The FE gain is stepwise flexible from 39 dB and to a maximum of 58 dB. The complete FE occupies a die area of only 2.6 mm2 in a 0.18 µm CMOS. To also accommodate the wide-band PRS signals in the IF section of the FE, a highly selective wide-tuning-range 4th-order Gm-C elliptic low-pass filter is used. It features an innovative continuous tuning circuit that adjusts the bias current of the Gm cell¿s input stage to control the cutoff frequency. With this circuit, the power consumption is proportional to the cutoff frequency thus the power efficiency is achieved while keeping the linearity near constant. Thanks to a Gm switching technique, which permit to keep the signal path switchless, the filter shows an extended tuning of the cutoff frequency that covers continuously a range from 7.4 MHz to 27.4 MHz. Moreover the abrupt roll-off of up to 66 dB/octave, can mitigate out-of-band interference. The filter consumes 2.1 mA and 7.5 mA at its lowest and highest cutoff frequencies respectively, and its active area occupies, 0.23 mm2. It achieves a high input-referred third-order intercept point (IIP3) of up to -1.3 dBVRMS.

EPFL2016