Externally indexed torrent
If you are the original uploader, contact staff to have it moved to your account
Textbook in PDF format
The design of power amplifiers (PA) for wireless applications has been a topic of great concern among the scientific community in electronics since the early 1990s. Power amplifiers dominantly determine the power efficiency and battery lifetime of modern mobile terminals. PA linearity is a key feature that limits the maximum allowed data rate of a radio link. The deployment of always more complex radio networks with a continuously growing throughput sharpens the specifications of power amplifiers. The market of wireless handsets (cellular phones, . . .) pushes manufacturers towards multi-standard capabilities (3G, Bluetooth, WIFI) and a growing level of integration. The Bill-of-Material and the consumed die area that are affordable are constantly reduced for cost purpose. The trade-offs that are inherent to PA design are therefore extremely stringent and no technical/technological solution can unanimously be regarded as a definitive contribution. To this date, the market of handset-dedicated power amplifiers is widely dominated by III/V technologies. However, throughout the following pages, we will try to highlight the benefits of PA integration on silicon. The architectures that will be proposed hereunder take advantage of silicon capabilities and strength, among others their relatively low cost and their ability to combine high power devices with low-power analog/digital control circuitry. Chapter 1 will first present the respective features of 2nd and 3rd generation cellular applications (GSM, DCS, EDGE, WCDMA. . .) and data transmission standards (WIFI, WIMAX, LTE). An overview of the most commonly employed RF power amplifier topologies will also be provided, with their advantages and drawbacks. Finally, the 0.25 μm BICMOS ST Microelectronics technology will be described and compared with III/V processes in the prospect of PA development. The fundamental features of power devices will be detailed and the most appropriate technological choice prior to the PA design itself will be discussed. In Chapters 2 and 3, several novel PA topologies will be proposed and discussed in terms of efficiency, linearity and complexity by means of mixed system/transistorlevel analyses. Chapter 2 will investigate three novel switched-mode power amplifier topologies. The first non-constant-gain principle is based on the power stage bypass/extinction and applied to a silicon HBT demonstrator. A silicon HBT demonstrator that was developed in the frame of RNRT ASTURIES project will be presented The other two topologies were proposed in the frame of the European Medea UPPERMOST project and deal with the dynamic control of a fragmented reconfigurable multi-cell power stage. First, open-loop power stage control will be considered. In this topology, power detection is carried out by a specifically dedicated digital Built-In Current Sensor whose behavior will be detailed. Second, a closed-loop power stage control system will be addressed and compared with the open-loop power stage control. Lastly, Chapter 3 will describe a PA Module that was developed in the frame of the European FP6 MOBILIS project. Furthermore, some mathematical developments and vector illustration will explain the theory of a continuously and dynamically power adaptive system that aims to combine linearity and efficiency even at low power levels. Moreover, frequency-dependent memory effects will be introduced in order to provide some further insight in the complex non-linear phenomenon’s that may alter the behavior of power amplifiers operating at wide channel bandwidths. An illustration of memory effects will be provided in the experimental section of this chapter. A PA module demonstrator based on integrated silicon and passive network dice will be presented, both in a stand-alone mode and in association with the other blocks of the MOBILIS transceiver. The final discussion will validate the proposed efficiency/linearity improvement principle under some conditions of channel bandwidth