rf amplifiers - bienvenue au site web biblioth¨que et archives
TRANSCRIPT
Multiharmonic Tuning Behavior of
MOSFET RF Power Amplifiers
by Yucai Zhang
A thesis submitted in conformity with the requirements for the degree of Master of Applied Science
Edward S. Rogers Sr. Department of Electrical and Computer Engineering University of Toronto
2001
O Copyright by Yucai Zhang 2001
National Library Bibliothéque nationale du Canada
Acquisitions and Acquisitions et Bibliographie Services services bibliographiques
395 Wellington Street 395. rue Wellington Onawa ON K1A ON4 OttawaON K l A W Canada Canada
The author has granted a non- exclusive licence aliowing the National L i b r q of Canada to reproduce, loan, distribute or seU copies of this thesis in microform, paper or electronic formats.
The author retains ownership of the copyright in this thesis. Neither the thesis nor substantial extracts fiom it may be printed or otherwise reproduced without the author's permission.
L'auteur a accordé me licence non exclusive permettant à la Bibliothèque nationale du Canada de reproduire, prêter, disûibuer ou vendre des copies de cette thèse sous la foxme de microfiche/film, de reproduction sur papier ou sur format électronique.
L'auteur conserve la propriété du droit d'auteur qui protège cette thèse. Ni la thèse ni des extraits substantiels de celle-ci ne doivent être imprimés ou autrement reproduits sans son autorisation.
Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers
Master of Applied Science, 200 1
Yucai Zhang
Edward S. Rogers Sr. Department of Electrical and Cornputer Engineering
University of Toronto
Abstract
This thesis investigates multiharmonic tuning of RF power amplifiers using power
MOSFETs implemented in bulk silicon CMOS technology. The use of this technique may
lead to the low-cost implementation of the RF power amplifier integrated on the sarne chip
as the rest of the wireless transceiver.
The work proposes a complete classification of multiharmonic tuning into fow basic
modes: both odd/even harmonics SHORT (SS), odd harmonics SHORT and even harmonics
OPEN (SO), odd harmonics OPEN and even harmonics SHORT (OS), and both oddeven
harmonics OPEN (00) . Conventional power amplifiers c m then be characterized using
these modes of operation in so far as multiharmonic tuning is concerned. A systematic
multiharmonic tuning optimization procedure is introduced to find the optimal harmonic
terrninations.
The newly proposed 00 mode features a sinusoidal drain curent waveform containing
no harmonics, resulting in little or no energy wasted at harmonic fiequencies and yielding
high eficiency.
To study the multiharmonic tuning behavior of MOSFET RF power amplifiers, power
MOSFETs were implemented in a 0.25pm silicon CMOS process. For power amplifiers
using these MOSFETs, at 1.88GH2, the 00 mode yields the highest efficiency (PAE4lYo)
with a 23.3dBm output power at a 12dBm input power and at a 2.OV supply voltage.
Acknowledgrnents
1 would like to express my sincere gratitude to Professor C.A.T. Salama for his
insightful guidance and invaluable assistance throughout the course of this work.
1 am indebted to Mr. J. Illowski, Mr. P. Watson and Mr. M. Stubbs from Nortel
Networks for their technical advice and assistance with load-pull measurements.
My appreciation extends to al1 the staff and students in the Microelectronic Research
Laboratory. 1 am specially grateful to Jaro Pristupa for his assistance with CAD tools and
Dana Reem for her technical assistance during the chip testing. Thanks go to Anthoula
Kampouris, Richard Barber, Milena Khazak, Farhang Vessal, Mathew Atekwana Amberetu,
Dusan Suvakovic, Mehrada Ramezani, Sotoudeh Hamedi-Hagh, John Ren, Namdar Saniei
for al1 their help.
Thanks also to my wonderfùl fiiends who made my Iife at Uofï pleasant and
unforgettable. Especially, 1 would like to express my appreciation to Song Ye, Zhixian Jiao,
Rick Kubowicz and Wei An, for valuable discussion both technically and personally, and
the rest of my fî-iends: Hongfei Lu, I-leng Jin, Shuo Chen, Wei Yang, Mike Sheng, Edward
Chun Keung Yu, 1-Shan Michael Sun for constructive discussions and cheerfbl chats.
Special thanks to my other fnends, Ting Lu, Jun Zhang, Tingju Zhu, Yajuan Su,
Mengsi You, Jian Yang, for sharing both my hard tirne and good time.
My deepest appreciation goes to my parents and sister for their constant support and
encouragement.
This work was supported by the Natural Sciences and Engineering Research Council of
Canada, Micronet, CITO, Gennwn, Mitel, Nortel Networks and PMC Sierra.
Table of Contents
Page
............................................................................................... CHAPTER 1 Introduction 1
I . I RF Power Amplifiers for Wireless Communications ............................................. 2
.............................................................................................. 1.2 RF Power Amplifiers 4
.................................................................... 1.3 Objectives and Outline of the Thesis 1 1
CHAPTER 2 Theoretical Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers .....~..~.........b........................................ 14
2.1 Introduction ........................................................................................................... 14
................................................................... 2.2 Analysis of Power Amplifier Modes 1 4
....................................................... 2.2.1 Conventional Power Amplifier Modes 1 4
....................................... 2.2.2 Classification of Multiharmonic Tuning Behavior 21
................................................... 2.3 Multiharmonic Tuning Optimization Procedure 23
2.4 Multiharmonic Tuning Behavior of MOSFET PAS .............................................. 25
............................................................................................... 2.4.1 Device Design 25
2.4.2 MHT Optimization for PAE ........................................................................ 28
2.5 Summary ................................................................................................................ 34
CHAPTER 3 Experimental Results ........................................................................... *..38
3.1 Power MOSFET Implementation ......................................................................... 38
............................................................................... 3.2 Power Device Characteristics 41
................................................................................. 3.3 MOSFET PA Characteristics 45
3 -4 Summary ............................................................................................................... -49
........................................................................................... CHAPTER 4 Conclusions ...51
APPENDIX A Harmonic Load Pull Measurement ................................................ ....*53
iii
List of Figures
Page
............................................... . Fig 1 . 1. Block diagram of a generic digital RF transceiver 2
............................. Fig . 1.2. A generic schematic of Class A.B. AB or C Power Amplifier 6
Fig . 1.3. RF power and drain eficiency as a lùnction of conduction angle ....................... 7
Fig . 1.4. Class F power amplifier with even harmonic trap ............................................... 8
Fig . 1.5. Power and Efficiency contours vs . phases of TL of the MESFET PA ................. 9
................................................................................. . Fig 1.6. Inverse Class F wavefoms 10
............................................................................... Fig . 2.1 : Waveforms of Class A/B/C 1 5
Fig . 2.2. Waveforms of ideal Class F mode ..................................................................... 17
....................................................... Fig . 2.3. Conceptual "target" waveforms of Class E 1 9
Fig . 2.4: (a) Schematic of basic (low-order) Class E power amplifier
...................................................................................................... (b) Actual waveforms -20
.............................................. . Fig 2.5 : Classification of multiharmonic tuning behavior -22
............................................................ . Fig 2.6. Basic circuit used in MHT optimization 24
............................ . Fig 2.7. Drain current for different gate width (gate length=0.25pm) 26
........................................................ Fig . 2.8. Modified BSIM3V3 CMOS device mode1 -27
Fig . 2.9. Schematic used in ADS simulator ...................................................................... 28
........................................ Fig . 2.10. PAE and Pout contours for SS mode (Pin=12dBm) -30
........................................ Fig . 2.1 1 : PAE and Pout contours for OS mode (Pin= 12dBm) 30
........................................ . Fig 2.12. PAE and Pout contours for SO mode (Pin= 12dBm) 31
....................................... Fig . 2.1 3: PAE and Pout contours for 00 mode (Pin= 12dBm) 31
Fig . 2.14. PAE. Pout and Gain vs . Pin of the MOSFET ................................................... 32
.................................... Fig . 2.15. Drain voltage and current wavefoms in the 00 mode 32
.... Fig . 2.16. PAE and Pout contours vs . phases of TL of the MOSFET power amplifier 33
...................................................... Fig . 3.1 : RF MOS ce11 layout with substrate contacts 39
....................................................................................... Fig . 3.2. RF input pad structure -39
..................................................................................... . Fig 3.3. Layout of test structures 40
............................................................................... Fig . 3.4. Micrograph of the test chip -40
....... Fig . 3.5. Ids-Vgs transfer characteristics of the MOSFET (W/L=200Opm/0.25pm) 41
....................................................... Fig . 3.6. Breakdown characteristics of the MOSFET 42
................... Fig . 3.7. IDs-VDs characteristics of the MOSFET (W/L=200qim/0.25pm) 42
.............................................................. . Fig 3.8. fT andf,, of the CMOS power device 44
................................. Fig . 3.9. Harmonic on-wafer load-pull measurement system setup 46
............................. . Fig 3.10. Measured characteristics of the MOSFET power amplifier 48
. . ......................................................... Fig A 1 : Harmonic Load Pull Measurement Setup -54
List of Tables
Page
............................................................... Table 1 . 1 : Characteristics of Wireless Standards 4
Table 2.1 : Optimal ZL(oO) of the MOSFET
..................................................... optimized for maximum PAE (Pin= 12dBm) 29
............... Table 2.2. Simulated characteristics of the power device and power amplifier 35
Table 3.1 : Measured and Simulated fT and f,, ............................................................... 44
............................. Table 3.2. Measured and simulated performance of each MHT mode 47
.................. Table 3.3 : Measured and Simulated Parameters of power MOSFET and PA 49
Chapter 1: Introduction 1
Introduction
In addition to persona1 communication products such as pagers and cellular phones,
wireless technology has impacted many other rapidly growing markets, for instance,
wireless local area networks (WLANs), global positioning systems (GPS), and RF
identification systems (RFIDs). A wide variety of system standards have been adopted to
support these applications. The explosive market of wireless communications is motivating
extensive research and design effort to develop communications devices with increasingly
higher performance, lower cost and low power consurnption.
Wherever there are wireless communications, there are transmitters; wherever there are
transmitters, there are RF power amplifiers. Power amplifiers (PAS) are used to amplie the
signal being transmitted to the necessary level needed to drive the antema at a particular
power output level, so that it can be received and decoded by the receiver within a certain
geographical area. Power amplifiers typically dominate the power consumption of the
transmitters (or transceivers), thus have critical impact on system performance and cost,
especially in low-voltage, low-power portable applications.
Advances in conventional CMOS technology have made this technology a promising
alternative for low-cost, low-voltage implementation and integration of wireless transceiver
building blocks, such as Digital Signal Process (DSP) cores, Low Noise Amplifiers,
Mixers, and other front-end ICs. However, RF power amplifiers, the bottleneck of wireless
Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers University of Toronto
Chapter 1: Introduction 2
transceivers, are still being implemented in expensive GaAs technologies or specific RF
LDMOS technologies, which prevent the integration of the power amplifiers into
transceivers. To enable a single-chip transceiver implementation and furiher reduce system
costs, it is highly desirable to develop RF power amplifiers h l ly compatible with
conventional CMOS technologies.
1.1 RF Power Amplifiers for Wireless Communications
Power Amplifiers in Wireless Transceivers
Digital Voice Modulator
Voice Coding Pulse O) ,* cornPressior * Interieaving * Shaping
I 1 Amplifier
Carrier
Down Converter
Ampl ifter Carrier
De-interleaving Voice 1
Decoding * Decompression - DAC - i Audio Speaker
Amplifier (b)
Fig. 1 . 1 : Block diagram of a generic digital RF transceiver (a) transmitter, (b) receiver
A generic digital RF transceiver is s h o w in Fig. 1 .la. On the transmitter side (Fig.
l.la), the voice signal is first digitized by an analog-to-digital converter (ADC) and
Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers University of Toronto
Chapter 1: Introduction 3
compressed to reduce the bit rate and hence the required bandwidth. Then, the data
undergoes "coding" and "interleaving" to format the data such that the receiver can detect
and minimize errors by performing the reverse process. Since rectangular pulses are usually
not optimum for modulation, the data is "shaped and modulated by the RF carrier
fiequency. After filtering, the signal is applied to the power amplifier which drives the
antenna. As illustrated in Fig. 1.1 b, on the receiver side, the signal received by the antenna
is amplified, downconverted, and digitized. Subsequently, demodulation, equalization,
decoding, de-interleaving, and decompression are performed in the digital domain. The
resulting data is then converted to an analog signal by a digital-to-analog converter (DAC),
amplified, and applied to the speaker.
Wireless Communication Modulation Schemes
The type of power amplifier in a wireless system depends on the type of modulation
standard used in the system. Digital modulation with binary baseband waveforms can be
performed by one of the following methods [ 1 ] : Amplitude Shift Keying (ASK), Frequency
Shift Key ing (FSK), Phase Shi fi Key ing (PSK), or Quadrature Amplitude Modulation
(QAW
In many applications, "quadrature modulation" is used to reduce the bandwidth
requirement [2]. Quadrature modulation includes two broad categories: quadrature phase
shift keying (QPSK) and minimum shift keying (MSK). QPSK includes specific options
such as Offset QPSK (OQPSK) and d4-QPSK. MSK has a widely used subset known as
Gaussian MSK (GMSK).
These modulation schemes fa11 into two general categories: linear modulation and
constant envelope modulation, depending on the envelope shape of the modulated
waveforms. In linear modulation schemes, such as QPSK and QAM, the abrupt phase
changes in the modulated waveform result in envelope variations if a filtcr limits the
Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers University of Toronto
Chapter 1: introduction 4 - - -
bandwidth. Such variations in turn require a linear power amplifier to avoid spectral
degradation. On the other hand, constant envelope modulation signals, such as FSK and
GMSK signals, c m be processed by high efficiency noniinear power arnplifiers. Table 1.1
lists some of the characteristics of several wireless standards such as Digital European
Cordless Telephone (DECT), Personal Handyphone System (PHs), Personal
Communications Services at l9OOMHz (PCS 1900) and Universal Digital Portable
Communications (UDPC). The power amplifiers to be studied in this thesis target non-
linear power arnplifiers used in conjunction with constant envelope modulation schemes.
Table 1.1 : Charactenstics of Wireless Standards [3,4]
1 Standard II DECT 1 P H s 1 PCS1900 1 UDPC 1
1.2 RF Power Amplifiers
Modulation
Envelope
Power Amp
Power Amplifier Metrics
The most commonly used metric to characterize the efficiency of a power amplifier is
the Power Added Eficiency (PAE), which is defined as
FSK
constant
non-linear
PAE = POUT -Pm p~~
where Pour is the RF power delivered to the load, PIN is the available input power and PDc
is the total power taken fiom the DC supply.
d4-QPSK
variable
linear
Another metric of efficiency is the drain (or collector) efficiency given by
- - - -
Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers University of Toronto
GMSK
constant
non- linear
d4-QPSK
variable
linear
Chapter 1: Introduction 5 - - -- -
Linearity is another concem in power arnplifier design. When the input power is small,
the gain (POUdPIlv) of a power arnplifier is almost constant. When the input power
increases, the output power is a compressive function of the input due to the nonlinearity of
the amplifier, that is, the gain approaches zero for sufficiently high input levels. The
nonlinearity of a power arnplifier can be characterized by the "1-dB compression point",
defined as the input signal power level that causes the smail-signal gain to drop by 1 dB.
Conventional Power Amplifiers (PAS)
Power amplifiers have been traditionally categorized as Class A, B, AB, C, D, E, F and
S 151. They cm be classified into three families:
Unsaturated PA: The output power is a function of the input power. This family
includes Class A, B, AB, C, and F. The power transistor in these PAS operates as a
current source. Class A, AB, and B PAS may be used as linear PAS, whereas Class C
power amplifiers are more nonlinear in nature.
Saturated PA (or switching-mode PA): This family includes Class D, E, and S. In
these classes, the power transistor operates as a switch. The transistor "on" voltage
is usually as close to zero as possible.
Mixed-mode PA: This family includes Class AB, B or C where the power transistors
are over driven into gain compression at fûll output to improve efficiency. The
transistor saturates during part of the "on" portion of the RF cycle, acting as a
switch; for the rest period of the "on" portion, the transistor operates as a current
source.
Fig. 1.2 shows a generic schematic of Class A, B, AB, or C power arnplifier using a
MOSFET and a tuned load. The primary distinction between these classes is the gate bias
voltage of the transistor that determines the fraction (conduction angle a) of the RF cycle
for which the transistor conducts. For Class A power amplifiers, the transistor is on for the
Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers University of Toronto
Chapter 1: Introduction 6
entire cycle (a=2n, device biased far above threshold VI), whereas it is on for half the cycle
for Class B PAS (a-, device biased ai threshold), is on for greater than half the cycle for
Class AB PAS (a-, device biased slighily above threshold), and is on for less than half the
cycle for Class C PAS (a<x, device biased beloiv threshold).
Fig. 1.2: A generic schematic of Class A,B,AB or C Power Amplifier
In these classical PAS, the drain current or voltage current waveforms are assumed to be
sinusoidal. The LC tank in the power amplifier is tuned at the fundamental fiequency and
shorts harmonic components thus producing a sinewave voltage. The drain current c m be
modeled as a truncated sinewave with a DC offset component In order to obtain the
maximum voltage swing at the drain, an optimum load resistance Rapt must be present at the
drain. Fig. 1.3 shows the variation of output power and drain eficiency as a function of
conduction angle.
The sinusoidal waveform assumption limits the eficiency and output power of
classical power amplifiers. In reality, the existence of higher order harmonies in the
waveform can be exploited to improve the performance. One way is to overdrive these
Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers University of Toronto
Chapter 1: Introduction 7
1.5 Po
Po(C1ass A) - -
1 .& +-- RF Power -
- - -
0.5 0% 27~ x Conduction angle
Class A AB B C
Fig. 1.3: RF power and drain efficiency as a function of conduction angle. (optimum load, harmonic short and zero-Vsat assumed)
classical PAS such that the voltage or current waveform of the transistor is clipped
significantly at both ends [SI. Although undesirable distortion is introduced, there can be
usefiil trade-off between eficiency enhancement and linearity that can be utilized in low or
intermediate envelope amplitude applications. Appropriate harmonic termination at the
output can also help enhancing the performance. An example involves replacing the
sinewave voltage at the device output with a flatter, square-like periodic waveform. This
can result in important benefits in both power and efficiency, both of which can be traded
effectively for linearity. This method results in Class F power amplifiers [ 5 ] and other
subtypes [6]. An implementation of Class F is show in Fig. 1.4, where the basic structure is
that of a Class B power amplifier but with a quarterwave SCSS (short-circuit shunt stub)
used as an even harmonic trap. A maximum esciency of 88.4% can be achieved for such a
power amplifier using an ideal transistor.
In Class D, E, and S, the transistor acts as a switch. An ideal switch has either zero
voltage across it or zero current through it, thus it consumes no power. Therefore switching
Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers University of Toronto
Chapter 1: Introduction 8
RF Shon Vdc
l ' t l
Even Harmonic Trap W4) (SCSS) Series LC @Io
Input Match
RL
I
Fig. 1.4: Class F power amplifier with even harmonic trap
type power amplifiers have ideally 100940 eff~ciency. Class D power amplifiers need a high-
side device which makes them unsuitable for RF applications, both in terms of parasitic
reactances and drive requirements. Class E power amplifiers are attracting attention in RF
communications applications [7,8]. Theoretical efficiency of around 90% can be achieved,
but with an RF output power lower than that obtainable from the same device in a
conventional Class AB power amplifier, and with a peak voltage as high as 3.6 times the
DC supply. Class S power amplifiers are similar to a pulse-width modulator with a low pass
filter at its output and are not suitable for high frequency operation.
Multiharmonic Tuning of Power Amplifiers
The conventional power amplifier theones are generally based on waveforrn analysis
for ideal power devices. However, real devices exhibit parasitics such as finite on-state
resistance and output capacitance. Conventional PA theories cannot be used to analysis the
effects of these parasitics on the overall PA performance, thus preventing improvements in
PA performance.
Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers University of Toronto
Chapter 1: Introduction 9
Multiharmonic tuning techniques, combined with harmonic balance simulations, have
recently been proposed for GaAs MESFET and HEMT power amplifiers to provide M e r
insights in improving power amplifier performance[9-141. The key idea is to control the
shape and overlap of the drain voltage ancilor current waveforms using appropriate muhi-
harmonic terminations. Two kinds of multihannonic tuning behavior were reported:
The first one is the widely known Class F. The drain voltage waveform is square-
like. Staudinger et al. [9,10] studied the multiharmonic tuning effect on a set of
GaAs MESFET power amplifiers. Fig. 1.5 shows the output power and efficiency
contours versus the phases of TL at 200 and 3a0. The variations of eficiency and
Po,, are insensitive to the phase of TL(3mn); in another words, the results exhibit a
strong dependence of efficiency and output power on the second harmonic
termination and a weaker dependence on the third and higher hannonic
terminations. These results suggest that significant improvements in linearity,
eficiency and output power c m be achieved with proper harmonic terminations.
Specifically, the best performance can be obtained when the 2nd harmonic is SHORT
and the 3rd harmonic is OPEN (Class F operation).
ma(r,@ - D.B AM@@ =O) - ~.t( (a) (b)
Fig. 1.5: Power and Eniciency contours vs. phases of load reflection coefficients (rL) of the MESFET PA [IO]
Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers University of Toronto
Chapter 1: Introduction 10
The second type is the opposite to the first one, Le., the 2nd harmonic is OPEN and
the 3rd harmonic is SHORT. Or in generai, the drain current waveform is square-
like, containing only odd harmonics, and the drain voltage waveform containing
only even harmonics, as shown in Fig. 1.6. This type is referred as Inverse Class F.
Drain Voltage
Fig. 1.6: Inverse Class F waveforms
This mode was first reported in power amplifiers using GaAs MESFET or HEMT
power devices [11,12]. A more complete analysis of this type of behavior was
performed recently [13,14]. The eficiency of power amplifiers in this mode can be
expectcd to be higher than Class F. However, the peak drain voltage swing is larger
than that in Class F and hence power devices with higher breakdown voltages are
required.
Muitiharmonic Tuning Behavior of MOSFET RF Power Amplifiers University of Toronto
Chapter 1: Introduction 11
Unresolved Issues
Even though the multiharmonic tuning technique has been proposed and successfully
used in power amplifier design, there are still some unresolved issues. First, there is no
complete classification of multiharmonic tuning behavior. Second, the relationship between
conventional power amplifier modes (Class AB, E, F, etc.) and multiharmonic w i n g modes
is vague. Finally, there is no systematic procedure to find the best mode and corresponding
optimal multiharmonic impedances for power amplifier design. These issues restrict the
practical design of power amplifiers using multiharmonic tuning techniques.
1.3 Objectives and Outline of the Thesis
As the channel length of a conventional MOS device is reduced into the deep
submicron range, its high frequency performance improves and it becomes attractive for
low-cost integrated implementations of RF power ampli fiers [16- 191.
The objective of this thesis is to study the multihmonic tuning behavior of RF power
amplifiers using power MOSFETs implemented in conventional bulk silicon CMOS
technologies. The MOSFET power amplifiers are targeted to operate at 1.88GHz fiom a 2V
supply with an output power of 2OOmW suitable for wireless applications.
In Chapter 2 the multiharmonic tuning behavior is analyzed and classified into four
basic modes and conventional power amplifier modes are characterized using this
classification. A systematic multiharmonic tuning optimization procedure is proposed to
find the best mode and corresponding optimal harmonic terminations. The multiharmonic
tuning behavior of bulk silicon MOSFET RF power amplifiers is then studied by
simulation. The power devices used are implemented in a 0.25pm CMOS process. Device
characterization and measured power amplifier performance are then presented in Chapter 3
[20]. Finally, Chapter 4 summarizes the results and outlines fbture work.
Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers University of Toronto
Chapter 1: Introduction 12
References
V. K. Garg and J. E. Wilkes, "Wireless and Persona1 Communications Systems", Pren- tice-Hall, New Jersey, 1996.
B. Razavi, "RF Microelectronics", Prentice-Hall, New Jersey, 1998.
R. Pandya, "Mobile and Persona1 Communication Services and Systems", IEEE Press, New York, 2000.
L. E. Larson, "RF and Microwave Circuit Design for Wireless Communications", Artech House, Boston, 1996.
S. C. Cripps, "RF Power Amplifiers for Wireless Communications", Artech House, Boston, 1999.
B. Ingruber, W. Pritzl, D. Smely, M-Wachutka and G Magerl, "High-Eficiency Har- monic-Control Amplifier", IEEE Tram. Microwave Theory and Techniques, Vo1.46, pp.857-862, 1998.
N. O. Sokal, "Class-E switching-mode high-eficiency tuned RFImicrowave power amplifier improved design equations," IEEE MTT-S Int. Microwave Symp. Dig., vol. 2, pp. 779-782,2000.
T. Sowlati, C.A.T. Salama, J. Sitch, G. Rabjohn, and D. Smith, "Low Voltage, High Eficiency GaAs Class E Power Amplifiers for Wireless Transmitters", IEEE Journal of Solid-State Circuits, Vo1.30, pp. 1074- 1085, 1995.
J. Staudinger, "Multiharmonic load termination effects on GaAs MESFET power amplifiers," Microwave Journal, pp.60-77, April 1996.
1101 J. Staudinger and G Norris, "The effect of harmonic load terminations on RF power amplifier linearity for sinusoidal and pi/4 DQPSK stimuli," IEEE MTT-S Tech. for Wireless Applications Dig., pp.23-28, 1997.
11 13 J. C. Pedro, L. R. Gomes and N. B. Carvalho, "Design techniques for highly efficient Class-F amplifiers driven by low voltage supplies," IEEE Int. Cod. on Electronics, Circuits and System, vol. 3, pp.201-204, 1998.
[12] Y. Tkachenko, A. Klimashov, C. J. Wei, Y. Zhao and D. Bartle, "E-PHEMT for single supply, no drain switch and high eficiency cellular telephone power amplifiers," GaAs IC Symp. Proc., pp. 127-1 30, 1999.
[13] C. J. Wei, P. DiCarlo, Y. A. Tkachenko, R. McMorrow and D. Bartle, "Analysis and experimental waveform study on Inverse Class-F mode of microwave power FETs," IEEE MTT-S Int. Microwave Syrnp. Dig., vol. 1, pp. 525-528,2000.
[14] A. houe, T. Heirna, A. Ohta, R. Hattori and Y. Mitsui, "Analysis of Class-F and Inverse Class-F mode amplifiers," IEEE MTT-S Int. Microwave Symp. Dig., vol. 2, pp. 775-778,2000.
-- - - - -
Multiharrnonic Tuning Behavior of MOSFET RF Power Amplifiers University of Toronto
Chapter 1: Introduction 13
[15] E. Morifuji, H.S. Momose, T. Ohguro, T. Yoshitomi, H. Kimijima, F. Matsuoka, M. Kinugawa, Y. Katsumata and H. Iwai, "Future Perspective and Scaling Down Roadmap for RF CMOS," IEEE Symposium on VLSI Technology, Proceedings, pp. 163 - 164, 1999.
1161 G Hefiman, "GaAs, MOS, Bipolar Vie For Power Applications," Microwave & RF, pp.3 1-39, July 1998.
[17] D. Su and W. McFarland, "A 2.5-V 1-W Monolithic CMOS RF Power Amplifier," IEEE Custom Integrated Circuits Conference, Proceedings, pp. 1 89-1 92, 1997.
[18] T. Ohguro, M. Saito, E. Morifuji, K. Murakami, K. Matsuzaki, T. Yoshitomi, T. Morim- oto, H.S. Momose, Y. Katsumata and H. Iwai, "High eficiency 2 GHz power Si-MOS- FET design under low supply voltage down to IV," IEEE Int. Electron Devices Meeting, Technical Digest, pp.83-86, 1996.
[19] 1. Yoshida, M. Katsueda, Y. Maruyarna, and 1. Kohjiro, "A Highly Escient 1.9-GHz Si High-Power MOS Amplifier," IEEE Trans. Electron Devices, Vo1.45, pp.953-956, 1998.
[20] Y.-C. Zhang and C. A. T. Salama, "Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers," IEEE MTT-S [nt. Microwave Symp., May 2001.
Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers University of Toronto
Chapter 2: Theoretical Multiharrnonic Tuning Behavior of MOSFET RF Power Amplifiers 14
CHAPTER 2
Theoretical Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers
2.1 Introduction
In this chapter the conventional power amplifier modes and multiharmonic tuning
modes are analyzed. Multiharmonic tuning is then classified into four basic modes and
conventional power arnplifiers are characterized using this classification. A systematic
multiharmonic tuning optimization procedure is proposed to find the optimal fundamental
and harrnonic terminations. The multiharmonic tuning behavior of MOSFET RF power
arnplifiers is then studied using the proposed multiharmonic tuning optimization procedure.
2.2 Analysis of Power Amplifier Modes
2.2.1 Conventional Power Amplifier Modes
In this section each power amplifier mode is first analyzed in a traditionai way and then
analyzed fiom a multiharmonic tuning point of view and the requirements of load
impedance at the harmonic fiequencies are given.
Class A, B, AB and C
Class A, B, AB and C power amplifiers are the conventional power arnplifiers. In these
PAS the drain voltage waveform is assumed to be sinusoidal. The drain current can be
Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers University of Toronto
Chapter 2: Theoretical Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers 15
modeled as a complete or truncated sinewave of amplitude Im and with a DC component Iq
as shown in Fig. 2.1.
Fig. 2.1 : Wavefonns of Class A/B/C (The value of a de fines the class of operation)
In general, the drain current I (8 ) can be expressed as
where a ( O 5 a < 2n ) is the conduction angle (as defined on page 5). The Fourier series
expansion o f this waveform is given by
where
is the DC component and
Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers University of Toronto
C hapter 2: Theoretical Multiharmonic Tuning Behavior of MOSFET RF Pow er Amplifiers 16 --
where i l is the fundamental fiequency current and in (n>l) is the n-th order harmonic
component.
The DC component of the drain current can be used to calculate the average power
drawn fiom the power supply. With an appropnately designed load network, only the
fiindamental component of the drain current will be fed to the load resistor.
In order to obtain the maximum voltage swing at the drain, an optimal load resistance
RL must be present at the drain. For a given a, RL is given by
where Vdd is the drain supply voltage, V, is the MOSFET saturation (knee) voltage and
iI(a) is the amplitude of the fundamental fiequency current for the specified a.
The output power is
Given the above, the drain efficiency is
Vdd - Vsat , a - sina '1 =
V d d +in a - a cos 7 a>
Thus ideal Class AB power amplifiers have drain efficiency ranging fiom 50% to 78% and
ideal Class C power ampli fiers have an efficiency ranging from 78% to 100%.
As stated previously, the classical PA modes assume a sinusoidal drain voltage
waveform at the fundamental fiequency. The drain voltage V(0) can be expressed as
Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers University of Toronto
Chapter 2: Theoretical Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers 17
where ZL is the load impedance, ZLO is the DC load resistance, ZLi is the load
impedance at the fundamental fiequency, and ZLn ( n N ) is the load impedance at the n-th
order harrnonic fiequency. The load impedance at al1 harmonic fiequencies (ZLn (n>l))
MUST be ZERO (short circuit) in order to produce the required sinusoidal waveform.
Class F
The sinusoidal drain voltage wavefom assurnption in the above PA modes restricts the
efficiency and output power. If the drain voltage waveform is shaped to reduce the overlap
of the voltage and current waveforms, a significant improvement in efficiency and power
can be achieved [ 1 1.
In the Class F mode, improvements of both efficiency and output power are
accomplished by using odd harmonics to cause the drain voltage waveform to approximate
a square waveform and using even harmonics to cause the drain current to approximate a
half sinusoid.
Drain Voltage
Fig. 2.2: Waveforms of ideal Class F mode
The ideal wavefoms are depicted in Fig. 2.2. The total peak-to-peak drain voltage is
twice the supply voltage Vdd. The amplitude of the fundamental component is given by
Fourier analysis as
Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers University of Toronto
Chapter 2: Theoretical Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers 18
The load network is designed so that only the fundamental fiequency power can be
delivered to the load. Thus the output power is given by
where RL is the fundamental load resistance presented at the drain. Note that the value of Vi
exceeds Vdd which indicates higher output power can be delivered to the sarne fundamental
load RL compared to Class B mode. The eficiency of the ideal Class F mode is 100%.
The presence of harmonics in the voltage/current waveforms requires the correct load
impedances at the harmonic fiequencies. For ideal Class F mode, since only odd harmonics
exist in the voltage waveform and only even harmonics exist in the current waveform, the
load impedance should be M F N T E (open circuit) at ODD harmonics and be ZERO (short
circuit) at EVEN harmonics.
It is practically diKicult to realize the waveforms of ideal Class F mode. Other
alternatives of Class F mode utilize only the third harmonic frequency (and the fifih
harmonic frequency) to maximally flatten the drain voltage waveform to approximate a
square waveform [Il]. They require INFMTE load impedance at the THlRD (and FIFTH)
harmonics and ZERO load impedance at EVEN harmonics. The eficiency ranges from 88.4%
(for using only the third harmonic) to 92% (for using both the third and fifth harrnonics).
Class E
Power amplifier eficiency is maximized by minimizing power dissipation, while
providing a desired output power. The largest power dissipation occurs usually in the RF
power transistor and is determined by the average integration of the product of transistor
Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers Universi@ of Toronto
Chapter 2: Theoretical Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers 19
voltage and transistor current over the RF period. This power dissipation can be minimized
by avoiding simultaneous occurrence of high voltage and high current. Fig. 2.3 shows the
conceptual transistor voltage and current waveforms needed to meet the high eficiency
requirements. For this case,
Drain Current
7 ; switch : a I I
I on state ; I
I I
Drain I
I I I
1 I 1 Voltage I I I
1 I I I I I I I I ~ i m e D
0 x 2n 3x 4n
Fig. 2.3: Conceptual "target" waveforms of Class E
(a) The voltage across the switch (transistor) at the turn on time must be zero. (b) the
slope of the voltage across the switch (transistor) at the turn on time must be minimized,
and (c) the rise of the voltage across the switch at turn-off must be delayed till the current
drops to zero.
Thus the power delivered to the load RL equals the power provided by the power
supply source, achieving 100% efficiency. A Class E power amplifier containing a switch
and a load network meets the above critena [IO].
Fig. 2.4(a) shows the basic Class E circuit implementation (named "low-order Class
E") which consists of an ideal switch (transistor) shunted by a capacitor C I , a senes LC
circuit Lr-C2a resonating at the fundamental frequency fo, an additional reactance Ca to
adjust the phase of the voltage of capacitor C I , and the RF load resistor RL [9]. An RF choke
LI provides power supply to the switch. The wavefonns generated by such circuit
approximate the conceptual waveforms shown in Fig. 2.4(b). Note that these actual
Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers University of Toronto
Chapter 2: Theoretical Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers 20
waveforms meet al1 three criteria listed
is part of a sinusoid, the current through
'A' I C2a C2b
above. Even though the current through the switch
the load resistor is sinusoidal.
Switch Voltage
r Load Current . -
Fig. 2.4: (a) Schematic of basic (low-order) Class E power amplifier (b) Actual waveforrns
The design equations for the Class E amplifier obtained by time domain analysis are
surnrnarized below [9,1 O]
RFC with X usually > 1 OXc
Q$L RnfO
where fo is the fundamental frequency, QL (>1.7879) is the network loaded Q factor
(chosen by the designer as a trade-off), CI includes Co, of the switch transistor, C2 includes
in series with C2b.
The harmonic impedances do not play a direct part in the conventional time domain
analysis of Class E mode. However, the network topology itself provides the hannonic
terminations, which is analyzed below. In Fig. 2.4(a), if the impedance of the RF choke LI
Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers University of Toronto
Chapter 2: Theoretical Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers 21
is assumed to be infinite and Cl is treated as the output capacitance of the switch, the load
impedances ZL at reference plane A-A' is then given by
Substituting (2.15) into (2.16), ZL at the fündarnental frequency fo is
which is equivalent to a resistor RL in series with an inductor L less than L2,
And ZL at the harmonic frequencies nf, (m 1) is
zL( nfo) = j2nnf,L2 - \ 1 - 1 }+RL (2.18)
n2[ 1 + 1.1 10 /(QL - 1.7879)]
which is equivalent to a resistor RL in series with an inductor L approximately equal to L,.
At radio frequencies, the absolute value of ZL(nfo) is very large and can be assurned to be
INFNTE (open circuit).
As a conclusion, the load irnpedances of Class E mode at al1 harmonic fiequencies are
INFINITE (open circuit).
2.2.2 Classification of Multiharmonic Tuning Behavior
Based on the analysis of conventional power amplifier mdoes and the characteristics of
the multiharmonic tuning power amplifiers, this section proposes a complete classification
of multiharmonic tuning behavior. As shown in Fig. 2.5, this behavior can be classified into
four basic types according to the phases of the harmonic load reflection coefficients rL (the
magnitudes are assumed to be close to 1 .O). The four basic modes are:
SS mode: Soth odd and even hannonics are SHORT. The drain VOLTAGE waveform
is sinusoidal, containing no harmonic components. This type includes conventional
Class A, AB, B and C.
Multiharrnonic Tuning Behavior of MOSFET RF Power Amplifiers University of Toronto
Chapter 2: Theoretical Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers 22
180" 1) (short) .- OS mode SS mode
(Class F) (Clé ss A, B,C)
00 mode SO mode (Class E)
f \ \ J
-1 80" 0" 180" (short) (open) (short)
Phase of TL at odd harmonics
Fie. 2.5: Classification of multiharmonic tuning behavior
OS mode: Odd harmonics are OPEN and even hannonics are SHORT. The drain
VOLTAGE waveform is square-like, containing only odd harmonic components. This
type corresponds to the Class F [2,3].
SO mode: Odd hannonics are SHORT and even harmonics are OPEN. The drain
CURRENT waveform is square-like, containing only odd harmonics. The peak
voltage is higher than twice the power supply. This type corresponds to the Inverse
Class F 16-71.
00 mode: Both odd and even harmonics are OPEN. The drain CURRENT waveforrn
is sinusoidal, containing no harmonics, resulting in little or no energy wasted at
harmonic frequencies and yielding high eficiency. The peak voltage is also higher
than twice the power supply. This type includes the low-order Class E [9]. This
mode has not been reported in previous work on multiharmonic tuning.
Compared to the low-order Class E, the 00 mode has a few advantages. First, it
does not require rectangular gate driving waveforms, easing the design of the
Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers University of Toronto
Chapter 2: Theoretical Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers 23
dnving stage; second, it exhibits lower peak drain voltage, relieving the voltage
stress on the power device; third, the power device can be biased at a conduction
angle other than 180°, enabling trade-offs between efficiency and output power.
However, the 00 mdoe may result in a Iower eficiency than the low-order Class E.
In Fig. 2.5 it is interesting to note that the OS mode (Inverse Class F) is the dual of the SO
mode (Class F), and the 00 mode (Class E) is the dual of the SS mode (Class A, B, C).
This classification reveals that there are other possible high efficiency multiharmonic
tuning modes besides the reported Class F and Inverse Class F. Furthemore, the
relationship between the conventional power amplifier modes and multiharmonic tuning
modes is also shown in Fig. 2.5'.
2.3 Multiharmonic h i n g Optimization Procedure
Traditionally, the load-pull measurements have been used to find the optimal load
terminations (at fundamental and harrnonic fiequencies) for power amplifier design. But the
experiments are time consuming. The proposed simulation procedure [8] c m help find the
optimal load impedances. However, this procedure does not consider the effect of harmonic
teminations on the optimal fundamental load impedance.
This section describes a systematic optimization procedure to find the optimal load
tenninations. Fig. 2.6 shows the basic circuit used in harmonic balance (HB) simulations
using an HB simulator such as Agilent Technologies' ADS [16]. On the load side, only the
2nd and 3rd harmonics are considered and the magnitudes of their reflection coefficients are
set to be 0.99. Because higher order harmonics have smaller amplitudes and hence do not
have a significant effect, they can be ignored by terminating them to 5 M l On the source
side, al1 harmonics are also terminated to 5CKZ The optimization takes into account the
1. The SS mode can be further classified into Clam A, AB, B or C, according to the conduction angle. Furthermore, all multiharmonic tuning modes can be biased at a conduction angle between O" and 360".
Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers University of Toronto
Chapter 2: Theoretical Multiharmonic Tuning Behavior of MOSFET RF Power Ampliaers 24
mutual effects of the harmonic terminations on the choice of the optimal fundamental load
impedance.
Fig. 2.6: Basic circuit used in MHT optimization
DUT
The procedure used to find the optimal harmonic terminations is as follows:
Tuner 2
Estimate the load resistance RL at the fundarnental frequency (oo) according to the
available power supply Vdd and the required output power Pour:
where V,, is the saturation voltage, usually estimated to be O. 1 VdK0.2 Vdd.
Then determine the fundamental source impedance Z'(on) for good input match.
I w r
For each multiharmonic tuning mode, set the phases of rL(20d and rL(3a/9,
perforrn load-pull simulations to find the optimal fundarnental load impedance
ZLo(mo) Retuning Zs(oo) may be necessary.
2s
For each ZLo(wo) obtained in Step 2, simultaneously sweep the phases of TL(2qJ
and rL(3wd. Compare the results and chose the best mode and corresponding
load impedance.
Z-param ZL block
TL
- -
-- -
Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers University of Toronto
-
Chapter 2: Theoreticai Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers 25
2.4 Multiharmonic Tuning Behavior of MOSFET PAS
As previously mentioned, Class F (OS mode) or Inverse Class F (SO mode) have been
found to achieve the highest efficiency for GaAs MESFET or HEMT power amplifiers[2-
71. An important difference between GaAs devices and silicon MOSFETs for power
amplifier applications is that GaAs devices have smaller drain-source capacitances (Cd*)
because of the semi-insulating substrate. For devices with large drain-source capacitance,
such as power MOSFETs implemented on bulk silicon substrate, the question is which
multiharmonic tuning mode yields the highest eficiency.
In order to investigate this issue, power MOSFETs implemented in a conventional
0.25pm bulk silicon CMOS technology were used. The technology features single poly,
five metal layers and a minimum drawn channel length of 0.25pm7 suitable for RF and high
speed mixed signal circuits.
2.4.1 Device Design
Device Sizing
Determining the size of the power MOSFET is an iterative process where the initial
guess originated from the 1-V characteristic of the power transistor and is based on the
traditional load line theory [ I l .
The first step is to calculate the maximum drain current according to the power supply.
The drain supply voltage Vdd was chosen to be 2V for low voltage applications. (The
maximum Vdd is limited by the drain-gate oxide breakdown voltage BVgd of the MOSFET.
BVgd for the 0.25pm CMOS technology was estimated to be above SV) To obtain an output
power of 250mW, the optimum load resistance ROpl was firstly estimated to be
Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers University of Toronto
Chapter 2: Theoretical Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers 26
And the peak current through the load was then given by
According to the load line theory, the maximum drain current Idm is about twice I& i.e.,
Idm = 2 Ipk = 580mA (2.22)
Ih occurs when the gate voltage reaches its peak value VP. The maximum amplitude of
the driving (gate) signal was around Vdd and thus the peak gate voltage was given by
Vg,, Va + Vdd = 2.7V (2.23)
where Vgg is the gate biasing voltage (which equals 0.7V as described in Section 2.4.2).
The next step is to obtain an initiai guess of the transistor size. Fig. 2.7 shows the drain
current for 0.25pm N-MOSFETs with different gate width at a gate voltage of 2.7V. It can
be seen that a gate width of 900pm can provide 600mA drain current. Therefore the initial
guess of the transistor size was obtained as
Gate Width (urn)
Fig. 2.7: Drain current for different gate width (gate length=0.25pm)
Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers University of Toronto
Chapter 2: Theoretical Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers 27
Once the initial size was seîected the more accurate RF behavior of the MOSFET was
then found by load-pull simulations which indicated that the initial size is not sufficient to
obtain the required output power. Afier a few steps of iteration, the device size was finally
chosen to be
W / ' = 2 0 0 0 ~ m f l . 2 5 ~ m
Device Modeling
The BSIM3V3 device model was used in simulations. It was reported that BSIM3V3
model is basically adequate as a nonlinear physical large signal model for power amplifier
design [12,13]. However, to improve simulation accuracy, the BSIM3V3 model was further
enhanced to take parasitics into account, as shown in Fig. 2.8 and described below.
Fig. 2.8: Modified BSIM3V3 CMOS device model
The gate resistance and substrate resistances were added to the model as shown in
Fig. 2.8 [14]. The gate resistance Rg was calculated to be 0.34Q from the
interdigitated transistor layout and the resistivity of the gate material. The substrate
resistances Kubd was estimated to be 3.5R frorn simulations [14,15]. Cjd and Cjs
are the drain and source junction capacitance, respectively, and were set to be the
same corresponding values in the BSIM3V3 model.
Multiharmonic huiing Behavior of MOSFET RF Power Amplifiers University of Toronto
Chapter 2: Theoretical Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers 28
The layout parasitics such as pad capacitances and in te rco~ec t overlap
capacitances were added in p s t layout simulations.
2.4.2 MHT Optimization for P A .
Before MHT optimization, DC and S-parameter simulations were performed to obtain
some key device parameters as follows: W/L=200ûpm/0.25pm, Vth=O.6V, BVd,=7.0Vy
%,=0.55*, gm=330mS/mrn, Cd,=2.8pFy f ~ 3 1 GHz, fm,=28GHz.
The final circuit schematic used in multiharmonic tuning optimization is shown in Fig.
2.9. The device was biased at Vdd=2.0V and Vgg=0.7V through RF chokes. The RF choke
parasitic resistances (032) were measured and included in the schematic. The DC blocks
separate the RF input/output signals to DC supplies. The probe-pad contact resistance Rpp
(OSSI) (especially the contact resistance at the drain terminal) in the on-wafer load-pull test
setup was also included in the schematic. This value is small but must be included to match
the simulation and experimental results.
DC block
- -
Fig. 2.9: Schematic used in A D S simulator
2-param Tuner 2
model
Multiharmonic Tunïng Behavior of MOSFET RF Power Amplifiers University of Toronto
Chapter 2: Theoretical Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers 29
Using the MHT optimization procedure described in Section 2.3, the fhdamental load
resistance RL to obtain over 200mW output power at 1.88GHz was estirnated to be 6 0 and
Zs(od was determined to be 6+j14.8R The optimal fundamental impedances obtained
from load-pull simulations for each harrnonic tuning mode are shown in Fig. 2.10, Fig. 2.1 1,
Fig. 2.12, Fig. 2.13 and compared in Table 2.1. The highest eficiency was achieved in the
00 mode where ZL(od=5.6+j7.8Q Since the corresponding output power was only
22.4dBm, further trade-off between output power and efficiency must be made to achieve
high eficiency and the required power simultaneously. By inspecting the PAE and power
contours shown in Fig. 2.13, the optimal ZL(Wd was finally chosen to be 6.5+j4.OR
Table 2.1 : Optimal ZL'd of the MOSFET optimized for maximum PAE (Pi,=12dBm)
Fig. 2.14 shows the power, gain and eficiency variation as a function of the input
power for the various modes considered. At low gain compression, the 00 mode and OS
mode (Class F) exhibit higher efficiency than the SO mode (Inverse Class F) while at higher
gain compression, the 00 mode achieves the highest eficiency of 64% with an output
power of 23.6dBm (230rnW) at an input power of 12dBm.
OS mode
SO mode
00 mode
Fig. 2.15 displays the drain voltage and current wavefonns of the 00 mode at
Pi,=12dBm. The peak drain voltage is 4.4V, greater than 2*V& (4V); the drain current
wavefonn is sinusoidal, implying little or no energy wasted at hannonic frequencies.
Fig. 2.16 shows the PAE and Peut contours versus the phases of TL at 2a0 and 3a0 for
Pi,=12dBm. The variations of PAE and P, are insensitive to the phase of TL(3ao)
59.1
68.1
68.9
Multihamonic Tuning Behavior of MOSFET RF Power Amplifiers University of Toronto
6.3+j3.7
5.8+j7.7
5.6-ej7.8
23.1
22.2
22.4
Chapter 2: Theoretical Multihannonic Tuning Behavior of MOSFET RF Power Amplifiers 30
Fig. 2.10: PAE and Po,, contours for SS mode (Pi,= 12dBm)
Fig. 2.1 1 : PAE and Po,, contours for OS mode (Pi,= 12dBm)
Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers University of Toronto
Chapter 2: Theoretical Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers 31
Fig. 2.12: PAE and Po,, contours for SO mode (Pin=l 2dBm)
Fig. 2.13: PAE and Po,, contours for 00 mode (Pin=12dBm)
- --
Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers University of Toronto
Chapter 2: Theoretical Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers 32
7 0 ~ 1 - 00 mode 1 I 4 I
-5 O 5 10 15
Pin (dBm)
Fig. 2.14: PAE, Pout and Gain vs. Pin of the MOSFET
Fig. 2.15: Drain voltage and current waveforrns in the 00 mode
Furthemore, In Fig. 2.16(a) the white region is the region of the highest efficiency and is
somewhat independent of the phases of rL(2mO) and ïL(300), enabling easy
implementation of the load network.
The above simulation results demonstrate that the 00 mode yields the highest
efficiency at high gain compression for bulk silicon MOSFET RF power amplifiers.
Multiharmonic Tuning Behavior of MOSFET RF' Power Amplifiers University of Toronto
Chapter 2: Theoretical Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers 33
PAE scale :%)
51-53 53-55 57-59
EZI 59-61 O 61-63 O 63-65 7
-150 -100 -50 O 50 100 150
Phase of ïL(3a0)
(a) PAE contours
Phase of rL(3m0)
(b) POM contours
Fig. 2.16: PAE and Po,, contours vs. phases of ïL of the MOSFET power amplifier (ZL(wd=6.5+j4.m Pin=l 2dBm)
Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers UniversiQ of Toronto
Chapter 2: Theoretical Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers 34
It is worthy to compare previous results to the multiharmonic tuning behavior of GaAs
MESFET or HEMT power amplifiers[2-71. As mentioned before, an important difference
between GaAs devices and silicon MOSFETs for power amplifier applications is that GaAs
devices have smaller drain-source capacitances (CdJ because of the semi-insulating
substrate. The cornparison suggests that Cds may play an important role for the 00 mode to
achieve higher efficiency than other modes.
The conventional power amplifier modes and multiharmonic tuning modes were
analyzed in this chapter. A complete classification of multiharmonic tuning behavior was
proposed. Multiharrnonic tuning was classified into four basic modes and conventional
power amplifier modes can also be characterized using this classification. A systematic
multiharmonic tuning optirnization procedure was proposed to find the optimal harmonic
tenninations. The multiharmonic tuning behavior of bulk silicon MOSFET RF power
amplifiers was studied. The results showed that the 00 mode yields the highest efficiency
at high gain compression for the bulk silicon MOSFET RF power amplifiers. Table 2.2
sumrnarizes the simulated characteristics of the power device and power amplifier.
Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers University of Toronto
Chapter 2: Theoretical Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers 35
Table 2.2: Simulated characteristics of the power device and power amplifier
-
Operating Voltage (V) 11 2.0
Device Chawteristics
Size WiL (pxdpm)
Threshold Voltage (V)
Vduës
2000/0.25
Ron at Vgs=2.5V (Cl) (1 0.55
Best mode for highest PAE 11 O 0 mode
Peak PAE (%) 11 64
Corresponding Po,, (am) 11 23.6
Corresponding Pi, (dBm) II 12
Multiharmonic Tuning Behavior of MOSFET RF Power Amplifiers University of Toronto