ITEM METADATA RECORD
Title: Combined Low- and High-Frequency Nonlinear Characterization and Modeling (Gecombineerd laag- en hoog-frequent niet-lineaire karakterisatie en modellering)
Other Titles: Combined Low- and High-Frequency Nonlinear Characterization and Modeling
Authors: Avolio, Gustavo
Issue Date: 16-May-2012
Abstract: Accurate characterization and modeling of transistors are essential for the successful design of electroniccircuits. In many microwave circuit applications, transistors experience high-frequencyelectrical signals with large excursions. These signals drive the devices intoa nonlinear regime of operation. With the introduction of new materials in thetransistor fabrication process, additional effects, such as low-frequencydispersion, must be taken into account for a more complete investigation oftransistor characteristics. This work has aimed in the first place at characterizing transistors through nonlinear vector measurements. A time-domain based high-frequency large-signal network analyzer (LSNA) has beenextended towards the low-frequency range. When modulated excitations areapplied, the low-frequency part operates synchronously with the core RF LSNA. Otherwiseit can work like a stand-alone low-frequency LSNA. With this instrument, low-frequencydispersive effects under large-signal operation can be captured more accuratelythan before. Next, the calibrated large-signal voltage and current waveforms acquired by the combined low- andhigh-frequency LSNAs have been exploited to generate nonlinear transistormodels. Furthermore a waveform based nonlinear de-embedding procedure has beendeveloped. This technique can be adapted for improving waveform engineeringtechniques aimed at the design of high-frequency power amplifiers. Both themodeling and the de-embedding procedure are independent of the investigatedtransistor technology and therefore can be - in principle - adopted for anytype of active device.
Table of Contents: Table of Contents

Acknowledgements i
Abstract iii
Samenvatting iv
List of symbols ix
List of acronyms xi
List of Figures xiii
List of Tables xx

1 Introduction and outline of the work 1
1.1 Maxwell, Fourier, and the invention of the transistor 1
1.2 Non-linear distortion in high-frequency electronics 2
1.3 Down to transistor level 3
1.4 Transistor architectures: an overview 4
1.4.1 High electron mobility transistor 4
1.4.2 Multiple-gate Field Effect Transistor (MuGFET) 6
1.4.3 Low-frequency dispersion in compound-based transistors 7
1.5 Nonlinear characterization of high-frequency transistors 9
1.5.1 Pulsed measurements 9
1.5.2 High-frequency large-signal network analysis 9
1.5.3 Low-frequency large-signal network analysis 10
1.6 Thesis objectives and outline 11
References 13
2 Characterization system 15
2.1 Overview 15
2.2 LF LSNA (dynamic-bias) 15
2.2.1 Extending the RF LSNA to low-frequency: why? 15
2.2.2 System architecture 17
2.2.3 LF LSNA configurations 19
2.2.3.1 Low-frequency large-signal network analyzer 19
2.2.3.2 Combined low- and high-frequency nonlinear characterization 19
2.3 Calibration 20
2.3.1 RF LSNA calibration 22
2.3.2 LF LSNA calibration 22
2.3.3 LF-RF LSNA calibration 24
2.3.3.1 Procedure 24
2.3.3.2 RF and LF calibrations alignment 25
2.4 Examples 26
2.4.1 Connectorized power amplifier 27
2.4.2 On-wafer transistors 31
2.4.2.1 LF load-pull 31
2.4.2.2 RF modulation 32
2.5 Measurement system customization 34
2.5.1 Bandwidth extension 34
2.5.2 Thermal capabilities 35
2.6 Conclusions 36
References 37
3 Experimental nonlinear characterization 40
3.1 Introduction 40
3.2 Modulated excitations: two-tone test 41
3.2.1 Two-tone experiment 41
3.2.2 Nonlinear model validation 44
3.3 Load-pull 47
3.3.1 Low-frequency load-pull 48
3.3.2 Low-frequency large-signal transistor characterization 49
3.3.3 Experimental results 51
3.3.4 High-frequency load-pull 54
3.4 Thermal characterization at microwave frequencies 55
3.5 Conclusions 58
References 59
4 Linear and nonlinear modeling of transistors 62
4.1 Introduction 62
4.1.1 Measurements based linear and nonlinear modeling 62
4.1.1.1 Bias-dependent small-signal models 63
4.1.1.2 Large-signal models 63
4.2 Linear modeling 64
4.2.1 Multivariate orthonormal vector fitting (MOVF) 64
4.2.2 Modeling results 65
4.3 Nonlinear modeling 67
4.3.1 Nonlinear model mathematical description 69
4.3.2 Large-signal model extraction 71
4.3.3 Moving towards low-frequency 71
4.3.4 Model parameters estimation procedure 73
4.4 Modeling of the I-V function 74
4.4.1 High-frequency large-signal identification 74
4.4.2 Low-frequency large-signal identification 78
4.5 Modeling of the I-V and Q-V functions 85
4.5.1 Combined low- and high-frequency large-signal identification 85
4.5.2 GaN HEMT and GaAs pHEMT modeling 86
4.6 Conclusions 90
References 91
5 Nonlinear de-embedding of transistor waveforms 94
5.1 Motivations 94
5.2 Proposed approach 95
5.2.1 Nonlinear de-embedding 96
5.2.2 Nonlinear embedding 101
5.2.3 The role of the feedback capacitance CGD 102
5.3 Examples of nonlinear de-embedding 103
5.3.1 FinFET 103
5.3.2 AlGaN/GaN HEMT 106
5.4 Embedding versus De-embedding 110
5.5 Fully waveforms-based nonlinear de-embedding 112
5.5.1 Procedure 112
5.5.2 Application to AlGaN/GaN HEMT waveforms 113
5.6 Conclusions 116
References 117
6 Conclusions and outlook 120
6.1 Main achievements of this work 120
6.2 Future work 122

Appendix A 124

Curriculum Vitae 140
List of publications 142
ISBN: 978-94-6018-509-0
Publication status: published
KU Leuven publication type: TH
Appears in Collections:ESAT- TELEMIC, Telecommunications and Microwaves
Electrical Engineering - miscellaneous
Faculty of Engineering Science - miscellaneous

Files in This Item:
File Status SizeFormat
Thesis_Gustavo_Avolio.pdf Published 7936KbAdobe PDFView/Open

 


All items in Lirias are protected by copyright, with all rights reserved.