【文件名】:07110@52RD_Analysis.doc
【格 式】:doc
【大 小】:1954K
【简 介】:
Joel Vuolevi (Distortion in Rf Power Amplifiers 作者)关于 PA线性与失真方面的大作,不知道内容是否和Distortion in Rf Power Amplifiers一样, 请自己对照目录,以决定是否需要下载.下面是英文的简介:
The main emphasis in modern RF power amplifier (PA) research is on improving linearity while at the same time maintaining reasonably good efficiency, for which purpose external linearization in the form of feedforward or predistortion is often used. Linearity and linearization can be considered from both a fundamental signal (amplitude and phase conversions, AM-AM & AM-PM) and an intermodulation distortion (IMD) regeneration point of view, and since a study of intermodulation gives more information on the behaviour of an amplifier, linearity is studied in this thesis by analysing the amplitude and phase of IM components under varying signal conditions, i.e. as functions of temperature, modulation bandwidth and amplitude.
To study the behaviour of IM components analytically, a Volterra model including electro-thermal distortion mechanisms is developed and a simulation technique is introduced to determine how easily the amplifier can be linearized. An S-parameter characterization method for extracting the Volterra model and the simulation model is developed, and the amplitude and phase dependences of the IM components are shown by means of measurements performed by a novel technique developed here. The results show that the behaviour of IM components is more complicated than had commonly been expected.
Three techniques are developed for eliminating the frequency dependence of IM components, impedance optimization, envelope filtering and envelope injection. In the envelope injection technique, a low frequency envelope signal is added to the input of the amplifier in order to improve both the bandwidth and amplitude range of the memoryless predistortion. The functionality of envelope injection is demonstrated by Volterra calculations, simulations and measurements, and the technique is applied to 1W, 1.8 GHz common-emitter BJT and common-source MESFET amplifiers. IM cancellation better than 20 dB is achieved over a wide range of bandwidths and amplitudes.
It is concluded that an inherently linear amplifier is not necessarily easy to linearize any further using external techniques, but that the part of the distortion that varies with bandwidth and amplitude can be cancelled out using envelope injection and the remaining memoryless distortion by means of a simple polynomial RF predistorter. This results in good cancellation of distortion, and since both envelope injection and RF predistortion consume little power, both good efficiency and linearity can be achieved.
【目 录】:
Table of Contents
Preface
List of symbols and abbreviations
List of contributions
1. Introduction
1.1. Efficiency
1.2. Nonlinearity and intermodulation distortion
1.3. Linearization
1.4. Outline of the thesis
2. Memory effects in theory
2.1. Definition
2.2. Electrical memory effects
2.3. Thermal memory effects
2.4. Amplitude domain effects
2.4.1. Narrowband fifth order composition
2.4.2. Fifth order composition with memory effects
2.5. Conclusions
3. The Volterra model
3.1. Nonlinear simulation models
3.1.1. Empirical black-box models
3.1.2. Physical circuit-level models
3.2. Principles of Volterra series
3.2.1. Small-signal and Volterra series analysis
3.2.2. Direct calculation of nonlinear responses
3.3. The BJT/HBT model
3.3.1. Linear analysis of a CE BJT amplifier
3.3.2. The effects of output conductance, cross-terms and temperature
3.3.3. IMD distortion in a CE BJT amplifier
3.4. Calculating IM3 responses
3.4.1. BJT as a cascade of two nonlinear blocks
3.4.2. Detailed BJT analysis
3.5. MESFET model and analysis
3.6. Conclusions
4. Characterization of the Volterra model
4.1. Characterization difficulties
4.1.1. Effects of parasitics
4.1.2. Problems of self-heating
4.1.3. Effects of pulse length
4.1.4. Effects of fitting range
4.2. DC characterization
4.2.1. Principle and test set-up
4.2.2. Extracted nonlinearity coefficients of BFG11 BJT
4.3. AC characterization
4.3.1. Small-signal S-parameter characterization
4.3.2. Nonlinear S-parameter characterization
4.3.3. Test set-up for pulsed S-parameter measurements
4.3.4. Extracted nonlinearity coefficients
4.4. Conclusions
5. Linearization and memory effects
5.1. Linearization techniques
5.2. Simulating memory effects
5.2.1. Normalization of IM3 components
5.2.2. 2D sweep
5.2.3. Conclusions
5.3. Measuring the memory effects
5.3.1. Test Set-up and Calibration
5.3.2. Accuracy, speed and automation
5.3.3. Memory Effects in a BJT PA
5.3.4. Memory Effects in a MESFET PA
5.3.5. Summary
5.4. Conclusions
6. Linearization circuits with memory
6.1. Impedance optimization
6.1.1. Active load principle
6.1.2. Test set-up and its calibration
6.1.3. Optimum Zin at the envelope frequency without predistortion
6.1.4. Optimum Zin at the envelope frequency with predistortion
6.1.5. Summary
6.2. Envelope filtering
6.3. Envelope injection
6.3.1. Modulation frequency domain compensation
6.3.2. Amplitude domain compensation
6.4. Conclusions
7. Discussion
7.1. Contributions
7.2. Future work
References
A. Comparison between two Volterra modelling approaches
B. IM3 equations for cascaded 2nd order distortion mechanisms
IP卡
狗仔卡
发表于 2007-1-10 16:47:05
提升卡
置顶卡
沉默卡
喧嚣卡
变色卡
显身卡