1 - Presentation of Two-Port Networks

Library of Congress Cotalging-in- Publication Data Gonzalez, Guillermo ‘Microwave transistor ampifers:analysis and design / Guillermo Gonzalen.— 2nded Inddes bibliographical references and index. ISBNO-13-296538-4 'nanitoramplivs 2, Merowaveamplifers 1. THe. ‘TKraG3 196 G1 er Se 2D sgt Acquisitions Editor: Erie Svendsen Production Edivor: Rose Kernan ‘Cover Designer: Wendy Alling Judy Buyer: Julia Meehan Editorial Assistant: Kathryn E. assino (© 1997, 1984 by Prentice-Hall, Ine. Simon & Schuster/A Viacom Company Upper Saddle River, NI 07458 All rights reserved. No part of this book may be reproduced, in any form or by any means, without permission in writing from the publisher. “The author and publisher ofthis book have used their best efforts in preparing this book. ‘These efforts include the development, research, and testing of the theories and programs to determine their effectiveness. The author and publisher shall not be liable in any event for incidental or consequential damages with, or arising out of, the furnishing, performance, or use ofthese programs. Printed inthe United States of America 10 9 ISBN O0-13-254335-4 Prentice-Hall International (UK) Limited, London Prentice-Hall of Australia Pty. Limited, Sydney Prentice-Hall Canada Inc, Toronto Prentice-Hall Hispanoamericana, .A., Mexico Prentice-Hall of India Private Limited, New Delhi Tok ‘Simon & Schuster Asia Pie, Lid. Singapore. Bipolar Transistors, 71 CONTENTS PREFACE vii 1 REPRESENTATIONS OF TWO-PORT NETWORKS: 1 1.1 Introduction, 1 1.2 The Impedance, Admittance, Hybrid, and ABCD Matrices, 1 13. Transmission-Line Concepts, 4 The Lossy Transmission Line, 20 14 The Scattering Matrix and the Chain Scattering Matrix, 22 15 Shifting Reference Planes, 26 1.6 Properties of Scattering Parameters, 28 ‘Scattering Matrix of n-Port Networks, 42 1.7 Power Waves and Generalized Scattering Parameters, 45 Generalized Scattering Matrix for n-Port Networks, 52 1.8 Two-Port Network Parameters Conversions, 60 1.9 Measurement of Scattering Parameters, 61 1.10 Scattering Parameters of Transistors, 66 1.11 Characteristics of Microwave Transistors, 71w Contents Field Effect Transistors, 79 Problems, 86 2. MATCHING NETWORKS AND SIGNAL FLOW GRAPHS 92 2.1 Introduction, 92 22 The Smith Chart, 93 23 The Normalized Impedance and Admittance Smith Chart, 105 24 — Impedance Matching Networks, 112 2.5 Microstrip Matching Networks, 141 Microstrip Lines, 141 Design of Maiching Networks, 152 2.6 Signal Flow Graphs and Applications, 175 ‘Applications of Signal Flow Graphs, 180 2.7 Power-Gain Expressions: Alternate Derivations, 185 Power-Gain Expressions in Terms of Sp parameters, 192 28 VSWR Calculations, 194 Problems, 200 3 MICROWAVE TRANSISTOR AMPLIFIER DESIGN 212 3 Introduction, 212 3.2 Power Gain Equations, 213 33 Stability Considerations, 217 3.4 Constant-Gain Circles: Unilateral Case, 228 Unconditionally Stable Case, Sil <1, 231 Potentially Unstable Case, \Sil > 1, 234 3.5 Unilateral Figure of Merit, 238 3.6 Simultaneous Conjugate Match: Bilateral Case, 240 3.7 Operating and Available Power-Gain Circles, 247 Operating Power-Gain Circles, 247 Unconditionally stable bilateral case, 247 Potentially unstable bilateral case, 252 ‘Available Power-Gain Circle, 257 Unconditionally stable bilateral case, 257 Potentially unstable bilateral case, 260 38 Constant VSWR Circles, 260 3.9 DCBias Networks, 273 BIT Bias Networks, 273 GaAs FET Bias Networks, 280 Contents ‘4 NOISE, BROADBAND, AND HIGH-POWER DESIGN METHODS 41 Introduction, 294 4.2 Noise in Two-Port Networks, 295 43 Constant Noise Figure Circles, 299 44 Broadband Amplifier Design, 323 Balanced Amplifiers, 327 Feedback Amplifiers, 333 45 Amplifier Tuning, 348 4.6 Bandwidth Analysis, 348 4.7 High-Power Amplifier Design, 352 Class-A Operation, 353 Cass-B and Class-C Operation, 356 Intermodulation Distortion, 362 Power Combiners, 364 Design Examples, 365 48 Two-Stage Amplifier Design, 372 Problems, 374 5 MICROWAVE TRANSISTOR OSCILLATOR DESIGN 5.1 Introduction, 384 5.2 Oscillation Conditions, 384 Feedback Oscillators, 384 One-Port Negative-Resistance Oscillators, 388 5.3 Two-Port Negative-Resistance Oscillators, 397 54 Oscillator Design Using Large-Signal Measurements, 404 55 Oscillator Configurations, 411 Dielectric Resonator Oscillators, 414 YIG Oscillators, 422 Varactor-Tuned Oscillators, 425 Problems, 428 APPENDIX COMPUTER-AIDED DESIGNS APPENDIX A A.1. CIRCLE EQUATIONS: BILINEAR TRANSFORMATION ‘A.2, DERIVATION OF THE INPUT AND OUTPUT STABILITY CIRCLES [EQUATIONS (3.3.5) AND (3.3.6)] 294 384 433 eh i nppennX B STABILITY CONDITIONS gg |APPENDIX C APPENDIX D APPENDIX E APPENDIX F ‘APPENDIX G APPENDIX H ‘APPENDIX | APPENDIX J APPENDIX K APPENDIX L APPENDIX M INDEX, UNCONDITIONAL STABILITY CONDITIONS: K>1andBy>0, DERIVATION OF THE UNILATERAL CONSTANT- GAIN CIRCLES [EQUATION (3.4.10)], E.1. ANALYSIS OF (3.6.5) AND (3.6.6) FOR [P| <1 AND val <1, .2, CONDITION FOR A SIMULTANEOUS CONJUGATE MATCH, DERIVATION OF Grex [EQUATION (3.6.10)], DERIVATION OF THE CONSTANT OPERATING POWER.-GAIN CIRCLES, EXPRESSIONS FORT, CONSTANT VSWR CIRCLES, MAPPING OF CIRCLES IN THET, ANDI, PLANES, NOISE CONCEPTS, NOISE FIGURE OF AN AMPLIFIER, CONDITIONS FOR A STABLE OSCILLATION. Contents 466 467 472 474 476 ee eas PREFACE In this second edition, I have expanded considerably the material in the origi- nal book. New material has been included throughout the book to cover scattering parameter techniques in detail. Additional design procedures have been included for amplifiers and oscillators. At the suggestion of mauy readers, detailed derivations are included for several relations that were either referenced or included in the problems sections of the first edition. The revisions and ad- ditions make the text as self-contained as possible. Many new examples and problems have been added. ‘The main objective of this book has remained to present a unified treat- ‘ment of the analysis and design of microwave transistor amplifiers using scattering parameters techniques. The term microwave frequencies is used to refer to those frequencies whose wavelengths are in the centimeter range (ie.,1 to 100 em). However, the design procedures and analysis presented in this book are not limited to the microwave frequencies. In fact, they can be used in any frequency range where the scattering parameters of a transistor are known. This book is intended to be used in a senior-graduate-level course in microwave transistor amplifiers and oscillators, or by practicing microwave en- gineers. It is assumed that the reader has completed the undergraduate network theory, electronics, and electromagnetic courses, or equivalent courses. ‘The transmission-line theory needed is fully covered in this book, especially the use of Smith charts as a design tool. ‘The main transistors used in microwave amplifiers and oscillators are the silicon bipolar junction transistor (BJT) and the gallium-arsenide metal- semiconductor field-effect transistor (GaAs MESFET). The high-electronvill Preface mobility transistor is also discussed (HEMT). The BJT performs very well up to approximately 4 GHz.In this frequency range the BJT are reliable, low cost, have a high gain, and a low noise figure. The GaAs MESFET performance above approximately 4 GHz is superior to that of the BIT, as well as for very low-noise applications below 4 GHz. Microwave transistors are conveniently represented by two-port networks and characterized by scattering parameters, The scattering parameters are popular because they are easy to measure with modern network analyzers, their use in microwave transistor amplifier design is conceptually simple, and they provide meaningful design information. Furthermore, flow graph theory is readily applicable. Chapter 1 to 4 present the basic principles and techniques used in microwave transistor amplifier analysis and design. These chapters provide the foun- dation for a well-designed microwave transistor amplifier. Specifically, in Chapter 1 a detailed review of transmission-line concepts is given. The section ‘on transmission lines provides a comprehensive presentation of transmission line theory under a sinusoidal excitation. A new section dealing with power waves and ge -d scattering parameters is included. The measurement of the scattering parameters is discussed. The section on the characteristics of microwave transistors has been updated. Chapter 2 begins with a discussion of Smith charts. The design of matching networks in the Smith charts using lumped elements and microstrip transmission lines is discussed. Two- and three-lumped-elements matching networks are discussed in detail, as well as a variety of microstrip matching configurations. Signal flow graphs are used to derive the gain relations, and a section on the derivation of power relations directly from the incident and reflected waves is included. Power-gain expressions in terms of the power scattering parameters are given, Relations for the calculations of VSWRs and mismatch factors are discussed. Chapter 3 deals with microwave transistor amplifier design. A detailed derivation of the stability conditions has been included. Design procedures are presented involving the transducer power gain, the operating power gain, and the available power gain. Constant-gain circles for unilateral and bilateral devices are derived. A section dealing with the trade-offs between VSWRs and gain has been included. The last part of the chapter discusses the selection and design of various de bias networks. Chapter 4 treats the topics of low-noise amplifiers and the trade-offs between low-noise performance, gain, and VSWRs. New material on the design of broadband amplifiers, balance amplifiers, couplers, and feedback amplifiers is included. The section of power amplifiers has been expanded significantly. Chapter 5 discusses oscillators. Basically, an oscillator consists of an amplifier with the proper amount of feedback to make it oscillate. The negative- resistance approach to oscillator design is studied in detail. A variety of oscillators are designed using BJTs, GaAs FETs, dielectric resonators (DRs), Proface ix Several new appendixes that supplement the text material are included. For example, Appendix A discusses circles equations, the bilinear transformation, and transformation properties of circles. The derivation of the stability conditions is provided in Appendix B. Appendix I shows the derivation of the constant input and output VSWR circles. Appendix K discusses noise concepts, and Appendix L shows the derivation of the noise figure relation for a two- port network. The number of problems has been increased significantly. The problems form an integral part of the text, and even if they are not solved, they should be read, Many of the design calculations in this book can be conveniently made using a simple computer-aided design (CAD) program or a programmable cal- culator. In the first edition of the book, a listing of the program UM-MAAD (University of Miami Microwave Amplifiers Analysis and Design) was given. ‘The latest version of UM-MAAD can be obtained from the author by sending him, at the address provided at the end of this Preface, a 3.5" disk and a stamped, self-addressed envelope. This CAD program, written in FORTRAN- 77s simple to use. For completeness, some design examples using a large-scale CAD program are given in the Appendix titled “Computer-Aided Designs” The exam- plesillustrate some of the simulations and optimizations that can be performed using a large-scale CAD program. It is only after the problem is fully under- stood that CAD techniques should be used. Otherwise, the natural human ten- dency to be erroneous and inefficient can substantially increase the cost of a design. The large-scale CAD program used in the Appendix titled “Computer: Aided Designs” is the Hewlett-Packard HP85150B Microwave and RF Design Systems (referred to as the HP MDS program).A copy of this program was do- nated by Hewlett-Packard to the University of Miami in 1990 for teaching and research purposes. Many of my students and users of the first edition have provided input to this book. I wish to thank all of my former students for their helpful comments, especially the invaluable suggestions and constructive criticisms from Branko Avanic, William Sanfiel, Deniz Ergener, Ching Y. Kung, Claudio J. Traslavina, Levent Y. Erbora, Augusto E. Rodriguez, Edgar Duque, Sergio Bustamante, and Wil Ponce, and Mahes M. Ekanayake. The contributions of Dr. Branko Avanicto the solutions of the problems were very helpful, as were his many inputs. Dr. Kamal Premaratne,a friend and colleague, has provided comments and inputs to both the first and second editions of this book, especially in the chapter on oscillators. Over the years, I have also received support and encouragement from several colleagues—namely, Professors Tzay Young, Reuven Lask, Kamal Ya- coub, Manuel A. Huerta, and James C. Nearing. T also wish to give special thanks to Dr. Les Besser. In the 1970s, he first and varactor diodes. aa A7 diledlataiel design. Over the years, he has motivated my thinking and influenced the way that I approach some design problems. Dr. Besser has also done alot to promote education in the microwave electronics field. He teaches a variety of excellent courses (Besser Associates, 4600 El Camino Real #210, Los Altos, CA 94022) which are very well known and highly regarded in the microwave industry. Finally, my deepest appreciation goes to my wife, Pat, my children, Donna and Alex, and my late parents, Ricardo and Raquel, for their love, encouragement, and patience. Guillermo Gonzalez, Ph.D. University of Miami Department of Electrical and Computer Engineering Coral Gables, Florida 33124 REPRESENTATIONS OF TWO-PORT NETWORKS 1.1 INTRODUCTION In order to characterize the behavior of a two-port network, measured data of both its transfer and impedance functions must be obtained. At low frequencies, the z, y, h, or ABCD parameters are examples of network functions used in the description of two-port networks. These parameters cannot be measured accurately at higher frequencies because the required short- and open-circuit tests are difficult to achieve over a broadband range of microwave frequencies. ‘Aset of parameters that is very useful in the microwave range are the scattering parameters (S parameters). These parameters are defined in terms of traveling waves and completely characterize the behavior of two-port networks, In the 1970s the popularity of S parameters increased because of the ap: pearance of new network analyzers, which performed S-parameter measurements with ease. The S parameters are simple to use in analysis, and flow graph theory is directly applicable. Although the principal use of S parameters in this textis in the characterization of two-port networks, they can also be used in the characterization of n-port networks. Power waves and the use of generalized scattering parameters (Sp parameters) in the analysis of two-port networks are also discussed, 1.2 THE IMPEDANCE, ADMITTANCE, HYBRID, AND ABCD MATRICES At low frequencies the two-port network shown in Fig. 1.2.1 can be represented in several ways. The most common representations are the impedance matrix (z parameters), the admittance matrix (y parameters), the hybrid matrix2 Representations of Two-Port Networks Chap. 1 A ~ — 3 Twopot “ network ve a4 |_5 vague 121 Tworport network sepe- pani Pon? ‘entation (parameters), and the chain or ABCD matrix (chain or ABCD parameters). ‘These parameters are defined as follows: z Parameters: “ = eyah + Zrale (Ie alt] ibe] ()- te lc) (i]-[e all] ‘The previous two-port representations are very useful at low frequencies because the parameters are readily measured using short- and open-circuit tests at the terminals of the two-port network. For example, or in matrix form y Parameters: h Parameters: ABCD Parameters: is measured with an ac open circuit at port 2 (i.e. i “The z, y, and ABCD parameters are also useful in the computer analysis of circuits, When two-port networks are connected in series, as shown in Fig. 1.2.2, we can find the overall z parameters by adding the individual z parameters, namely vu] a [vit] _ [eit ch + zh] [i wel [at 4] [eh teh ch + eh) le ‘Sec. 1.2 The Impedance, Admittance, Hybrid, and ABCD Matrices atom 3 + el ZA “il =| [z4 [os f Na] [an 4 -,_-lla igure 1.22. Series connection using z parameters and a typical application. When two-port networks are connected in shunt, as shown in Fig, 1.2.° we can find the overall y parameters by adding the individual y parameter namely , i) fata) Pitot xb t rb] ly a) [a+ 8)” patyh yh ton] When cascading two-port networks the chain or ABCD matrix can b used as follows (see Fig, 1.2.4) - Gee La) ale 33] w= # i en a * mt wa e vA val ® No wh ve ya ve Figure 12.3 Shunt connection using y parameters and atypical application.4 Representations of Two-Port Networks Chap. 1 iy a 4 ve o Figure 1.24 Cascade connection using ABCD parameters anda typical application. because vj = v? and —if = i. The relation (1.2.1) shows that the overall ABCD matrix is equal to the product (i.e., matrix multiplication) of the individual ABCD matrices. ‘At microwave frequencies the z, y, h, or ABCD parameters are very difficult (if not impossible) to measure. The reason is that short and open circuits, to ac signals are difficult to implement over a broadband, at microwave frequencies. Also, an active two-port (e.g.,a microwave transistor) might oscillate under short- or open-circuit conditions. Therefore, a new representation of the ‘two-port network at microwave frequencies is needed, The appropriate representation is called the scattering matrix and the scattering parameters are defined in terms of traveling waves, 1.3 TRANSMISSION-LINE CONCEPTS ‘There are many varieties of transmission lines. Three common types of trans~ mission lines are shown in Fig. 1.3.1: the two-wire transmission line, the coaxial transmission line, and the microstrip transmission line. The microstrip tran mission line is the most appropriate for the construction of microwave ampli fiers. The theory of operation of these transmission lines can be explained using a distributed circuit model for the transmission lines. Such a model provides practical results for the voltage and current along the transmission lines with- ‘out having to resort to Maxwell's equations. ‘An electrical model for a transmission line is shown in Fig. 1.3.2a. Viewed as a two-port network, the transmission line receives power from the source at the input port (source end) and delivers power to the load at the output port (load end), The length / of the transmission line is divided into many identical Sec. 1.3 Transmission-Line Concepts ° | @) ) © Figure 131 (a) Two-wite transmission line:(b) coaxial transmission tine(c) miro- stip transmission line sections Ax. Each section Ax is modeled by a resistance R per unit length (R in ‘Qim), an inductance L per unit length (L in H/m), a capacitance C per unit length (C in F/m), and a conductance G per unit length (G in S/m). These pa rameters are assumed to be constant along the transmission line (i.e. the trans mission line is uniform). Figure 1.3.2b shows a section Ax of the transmission line and the voltage and currents at the input and output ports of the section 4x of transmission line Observe that the voltages and currents along the transmission line are func tions of position and time. At the input of the section Ax the voltage and cur rent are v(x,t) and i(x,t), while at the output the voltage and current arc vor + 4x1) and i(x + AX). Applying Kirchhotf’s voltage law to the model in Fig, 1.3.2b gives Moet) = Ma + Ane) = Rass) + Lae 2D Dividing by 4x we can write a Lrrrt—~———CC : ar ‘Ax ‘As dx approaches zero the left-hand side is recognized as the partial derivative of v(t) with respect to x. Hence, taking the limit as 4x90 we obtain VK) _ _ pirg yy — Diet ed — Rages) ~ LSE (3.1 Similarly, applying Kirchhoffs current law to the model in Fig. 1.3.2b give ave + xs (xd) — i(x + Ax) = Gaxv(x + 4x) + cay 2M ae Dividing by 4x and taking the limit as 4x0 results in iat) aux) mee eK -ce 32 ax Gx) — CTF, a6 Representations of Two-Port Networks Chap. 1 Figure 132. (2) A distributed circuit model for a transmission line; (b) a section Ax of the transmission ine; (c) section Ar ina lossless transmission line ‘The partial differential equations (1.3.1) and (1.3.2) describe the voltages and currents along the transmission lines. Of particular practical interest in microwave electronics is the lossless transmission line—that is, a transmission line where R and G are negligible, or simply R = G = 0.A section Ax in a lossless transmission line is shown in Fig. 1.3.2c. Ina lossless transmission line, R = G = Oand (1.3.1) and (1.3.2) reduce to avers) _ _, dite 7 L (1.3.3) and ‘Sec. 1.3 Transmission-Line Concepts 7 dies ce aa We are interested in the solution of (1.3.3) and (1.3.4) for a sinusoidal ex. citation of the line. Phasor analysis will be used to obtain the steady-state solu. tion for v(x,0) and i(x,t). Fora sinusoidal excitation [i.e for w(t) = V; cos wi] the steady-state volt ages and currents along the transmission line are also sinusoidal functions of time whose dependence on position and time can be expressed in the general form Mx) = flz)cos(wt + p(x)) a 3) and W(x.) = g(x)eos(wt + 9(x)) (1.3.6) where f(x) and g(x) are real functions of position and where g(x) and n(x) de: seribe the positional dependence of the phase. Using Euler’s formula, namely e = cos + jsinO where cos 0 is the real part of e# (i.e., Ree] = cos 8) and sin is the imaginary part of e (i.e, Ife] = sin 6), we can express (1.3.5) and (1.3.6) in the form W(xt) = fladeos(or + g()) = flxyRefel—r6] = ReffeweM] (13.7 and I(x) = g(x)cos(wt + n(x)) = g@)Refe"*")] = Relg(xje™Mer"] (1.3.8 In (1.3.7) and (1.3.8) we can define the phasor quantities Vix) = flyer and T(x) = goer" and write (1.3.7) and (1.3.8) in the form ver) = Re[Vixye!] (39 and (xd) = Re[x)e"] (3.10 ‘The phasors V(x) and I(x) are complex functions of position and expres the variations of the voltage and current as a function of position along th: transmission line. Substituting (1.3.9) and (1.3.10) into (1.3.3) gives a a Hy RelMexdet] = ~L 5, Rell(xe] Can8 Representations of Two-Port Networks Chap. 1 Since the Re operator commutes with the 3/@x and /2r operators, we can write (1.3.11) in the form a 2] = une a LRefI(x)jwe™] (1.3.12) Observing that V(x) is only a function of position, it follows that ave) _ ave) ox dx and (1.3.12) can be written as ae + int) ] =0 (13.13) Equation (1.3.13) must be satisfied for all times t. Therefore, the term in paren- theses must be zero, namely we —joL I(x) (13.14) ‘Actually, this last step might not be so obvious to the reader. It can be proved by observing that for wt = 0,e” = 1, and (1.3.13) gives ee + ints] =0 (13.15) and for wt = zi2, ef? = j, and (1.3.13) gives no (2 mare) =0 which is equivalent to wf + sotto] =0 (13:16) Equations (1.3.15) and (1.3.16) show that both the real and imaginary parts of (dV{x)/dx) + joL I(x) must be zero. Hence, (1.3.14) is verified. ‘A comparison of (13.3) and the corresponding phasor equation in (1.3.14) shows that (1.3.14) follows by simply replacing in (1.3.3) the 8/@r operator by jo, and v(x.) and i(x,t) by the phasors V(x) and I(x), respectively. The use of phasors allows us to change the partial differential equation (1.3.3) into the ordinary differential equation (1.3.14). Similarly, substituting (1.3.9) and (1.3.10) into (1.3.4) results in the fol- 3.7) Sec. 1.3 Transmission-Line Concepts 8 Equations (1.3.14) and (1.3.17) are the differential equations satisfied by the phasors V(x) and (x) along a lossless transmission line. These equations can be solved for V(x) and I(x) as follows. Differentiating (1.3.14) with respect to x gives e mo ~jot HO (13.18) Substituting (1.3.17) into (1. 3.18) gives ae —joL(—jwC)V(x) = -« LCV) or P 2M) + ee) =0 (3.19) where B=oVLC (1320) ‘The parameter (in radians per meter) is known as the propagation constant. ‘The general solution to the second-order differential equation in (1.3.19) is V(x) = Ae" + Belt™ 321), where the constants A and B are, in general, complex constants. They can be evaluated using the boundary conditions at the input and output ports of the transmission line. ‘The general solution /(x) can be obtained from (1.3.14) and (1.3.21). Thatis, hava) 2 a1 gyi 13) = Trax” jok dx + Be) . (1322) = Fac — ney Defining the characteristic impedance of the transmission line as ok __ol £ == I 1.3.23) 2 ( we can express (1.3.22) in the form (1.3.24) From (1.3.23) we observe that Z, in a lossless transmission line is real. ‘The time-dependent form of the voltage and current along the transmission line follows from (1.3.9) and (1.3.10). That is.0 Representations of Two-Port Networks Chap. 1 v(a,) = Re[Vix)el’] = Re[AeH*-) + BeMBxte) (1.3.25) and i(x) = Re[I(x)e™] = Re} Sertarnen Zest (1.3.26) In the simple case that A and B are real, (1.3.25) and (1.3.26) can be expressed in the form v(x,t) = A cos(wt — Bx) + Bcos(wt + Bx) (1.3.27) and i(eg) = Scostor - pr) - Zecostor + By) (1.3.28) ‘The functions A cos(wt ~ Br) and B cos(wt + Bx) are known as wave functions. Let us examine the wave characteristics of these functions. Let M(t) = A cos(wt — Bx). The function vi(x,¢) can be analyzed by first fixing the position x and observing the behavior of 1(x,¢) as a function of time. To this end we let x = Oand plot (0,0) = A cos at in Fig. 1.3.3a. The function i(0,¢) is seen to vary sinusoidally as time increases. In Fig. 1.3.3a the time interval between two consecutive equal values of the periodic signal is defined as the period 7. Hence, in Fig. 1.3.3a the period Tis given by ot|,.p = 2" or 2a =. Next, we examine the behavior of vi(x,t) as a function of position. To this end we let ¢ = 0 and plot vi(x,0) = A cos fx in Fig. 1.3.3b. In Fig. 1.3.3 the distance between two consecutive equal values of the signal is defined as the wavelength. Hence, the wavelength is given by Blan = or aa (13.29) If we now look at two values of (x,t) as a function of time we can calculate the speed of propagation of the wave. For example, at times and fz (where >t) we have V,(eh) = A cos(Bx — ot,) and a ‘Sec. 1.3 Transmission- w(0.) w(x) A A 27 ot er Be “A i “A i ot] see mn © valet) wbx) © Figure 13.3 (a) Plot of (04) = A cos wt; (b) plot of w(x.0) = A cos Bx: (c) caleu- lation ofthe phase velocity Wilh) = A cos(Br - wt) A plot of (x,t) and 11(x,¢2) is shown in Fig. 1.3.3c.A point of constant phase a at 1 = nand at = fy is shown in the figure. Since at point a the phase of vi(x1.11) is identical to the phase of w(x2,2) we can write Wish) = W@n) or2 Representations of Two-Port Networks Chap. 1 cos(Ary — wh) = c0s(Bx2 ~ ct) Hence, Ar ~ 04 = Bra ~ of e . 1.3.30) aan 8 a ‘The quantity on the left is recognized as the speed of propagation of the point a, known as the phase velocity vp. That is, o_o 1 B wVLc VLC g (1.3.31) we obtain (13.31) Substituting (1.3.20) into (1.3.29) and 2x 2a == = v,T Bo ovic f ? In other words, the wavelength is the distance traveled by the wave in a time interval equal to one period. Furthermore, from (1.3.30) since @ > and w/f is a positive quantity, it follows that x2 — x1 must be positive or x2 > x1. This shows that the point of constant phase is moving toward the right in Fig. 1.3.3c (ie., toward the load in a transmission line). In other words, the function A cos(wt — Bx) represents a traveling wave moving at a velocity vp toward the load. This wave is called an ‘outgoing wave when viewed from the source, or an incident wave when viewed from the load. ‘The analysis of B cos(wt + Ax) will show that this function represents a traveling wave moving at a velocity vp toward the left (i.e., toward the source in a transmission line). This wave is called an incoming wave when viewed from the source, or a reflected wave when viewed from the load. The nomenclature incident wave and reflected wave is preferred in transmission-line work. ‘Since the wave A cos(wt — Ax) is associated with the phasor Ae~i*, and Bcos(wt + Bx) with Bei*, we also refer to the phasor Ae~#* as the incident wave (in phasor form) and Bei" as the reflected wave (in phasor form). The quantity Bx is known as the electrical length of the line. In general, the voltage and current in a transmission line are composed of an incident and a reflected wave. ‘A transmission line of characteristic impedance Zz, length J, and terminated in a load Z, is shown in Fig. 1.3.4a. The source end is located at x = 0 and the load end at x ‘The incident and reflected waves are also shown where the wavy arrow symbol is used to denote a wave. Itis convenient in some transmission-line problems to show the load end at a distance zero and the source end at a distance / from the load. Letting din (1.3.21) and (1.3.24) we can write (see Fig. 1.3.4b) 4 Sec. 13 Transmission-Line Concepts a Tw(d) —— a Agile we Aeis Z x 2 z Via) = Are + Byer a)= Bot 3 rte Ay= Aer 8, = Boh © Figure 1.34 Transmission line withthe source end at x = 0 and load end at x (b) transmission line with the load end at d = O and source end at d= L (a) = Aye + Bye (13.32) and A dy = At eine ‘it x Go) Zz," Zz (1.3.33) where Ai = Ae-i#' and By = Bel”, The change x = 1 ~ d (such that d =~ x)is done in some transmission-line problems in order to measure positive distances as one moves from the load (at d = 0) toward the source (at d= /). From (1.3.32) and (1.3.33) we observe that in terms of the variable d the phasor Aye! represents the incident wave and Bye~* the reflected wave (see Fig. 1.3.4b). Figure 1.3.4 summarizes the two nomenclatures commonly used in the analysis of transmission lines. ‘The reflection coefficient is defined as the ratio of the incident to the reflected wave along a transmission line. In Fig. 1.3.4b the reflection coefficient at any position d, denoted by Pin(d),is Bye By ina 33 Til) = ae = Ge 334) The load reflection coefficient, denoted by To, is the value of Pix(d) at d=0,Thatis4 Representations of Two-Port Networks Chap. 1 Hence, (1.3.34) can be expressed in the form. Tyy(d) = Tye 74 (13.35) Also, (1.3.32) and (1.3.33) can be expressed in the forms V(d) = Ay(el™ + Tye) = Aye + Tye) (1.3.36) and (a) = Saco — Tye7/*4) = Seine = Tye 24) (1.3.37) ‘The value of the complex constant A; is obtained by using a known value of V(d) (ie., a boundary condition), usually the value of V(d) at the source end (i.c.,at d= 1), Of course, the value of V(d) at d = [ depends on the source amplitude and phase and source impedance connected to the line at d = 1. As we will show shortly [see (1.3.39)] the complex constant Io is simply evaluated in terms of Z, and Zo Referring to Fig. 1.3.5, the input impedance of the transmission line at any position dis defined as _ Md) _ elt + Tye Zl) = Tg) = 20S Pew (1.3.38) In (1.3.38) the constant Po can be evaluated using the boundary condition atthe load, namely that the value of the input impedance at d = O must be equal to Z1, That is, Zyy(0) = Z,, Then, from (1.3.38) 1+ ZnO) = 2. = Ze or n+ 2, Tid) ta) To. Wd) Zo a — a0 “zuta) = Vie) gure 135 Input impedance of the Kd) ‘transmission line at any position Sec. 1.3 Transmission-Line Concepts 6 Equation (1.3.39) shows that Py = 0 when Zz = Zo. That is, there is no reflection from the load when Zz = Z..A line with Zz = Z. is called a properly terminated or matched transmission line. Substituting (1.3.39) into (1.3.38) gives (Z, + Z,)e™ + (Z, ~ Z,)e (2, + Z,)e™ = (2, — Z,)e°™ Z,, 00s fd + jZ, sin Ba °Z, cos Bad + jZ, sin Ba ay Zt i2,tan fl °Z, + iZ, tan Bd 2x(d) = 2, (1.3.40) Equation (1.3.40) gives the value of the input impedance at any location d along the transmission line. At d = 0, (1.3.40) reduces to Zin(0) = Zu, as expected. At the input of the line the input impedance follows from (1.3.40) with d= 1.A very important property of a transmission line, and one that is used extensively in the design of microwave amplifiers, is the ability of the transmission line to change a load impedance to another value of impedance at its input. , ‘The addition of the two waves traveling in opposite directions in a transmission line produces a standing-wave pattern—that is,a sinusoidal function of time whose amplitude is a function of position. From (1.3.36) the magnitude of the voltage along the line is given by [May] = [Ai] | + Poe] (341) From (1.3.41), it follows that the maximum value of |V(d)| along the line has the value [VG nx = [41] + [Fold (1.3.42) and the minimum value of |(a)is [Y@)]ein = [Ai] C= [Tol) (13.43) ‘These valuesare used to define the voltage standing-wave ratio (VSWR), namely Vase VSWR = TVIaia (13.44) We now analyze four important cases of transmission lines: the matched line, the short-circuited line, the open-circuited line, and the quarter-wave line. In the properly terminated or matched transmission line shown in Fig. 1.3.6, we obtain from (1.3.39), (1.3.40), and (1.3.44) that To = 0, Zin(d) = Ze, and VSWR = 1. In other words, there is no reflected wave (since T = 0), the input impedance is Z, at any location d, and the VSWR has its minimum value of one.16 Representations of Two-Port Networks Chap. 1 Zn(0) = Zo Zn(€) = Zotan Be C1 Z ZZ Z %=0 dee d=0 dee d=0 Znld) = Zp Zee(d) = Zotan Bd @ OO) -—— o— Base 2 a oo! dze d=0 ae d=0 Zed) = ~Ho cot Bo 4 © Co) Figure136 (a)The matched transmission line;(b) the short- bls) bel fz) Parti pore Figure 142. Procedure for meting Su. The characteris inedanes ofthe transmission lines are Zo: and Z.2. oe ‘The chain scattering parameters, also called the scattering transfer parameters or T parameters, are used when cascading networks. They are defined in such away that the input waves.a( ih) and by(l,)in Fig. 1.4.1 are the dependent variables and the output waves aa(lz) and b(/2) are the independent variables. That is. ah) ) | Ti Tz} | Ox) [o3 Ty, Tal [ast ae ‘The relationship between the S and T parameters can be developed from (1.4.9) and (1.4.10). Namely, Th 1 and Tm 7p Tale Si Sp] fT? Th (1.4.12) Sie elp|er Ta Ty Ty ‘The T parameters are useful in the analysis of cascade connections of two: port networks. Figure 1.4.3 shows that the output waves of the first network (N,) are identical to the input waves of the second network (N,), namely [2] Le] Since26 Representations of Two-Port Networks Chap. 1 Me Ny ° a mm re |e ee || 1 i ~ bay Talla a || . ole Sell bn Tall = by Figure 14.3 Cascade connection of two-port networks ay] [Th Ti] [bo bu) [Th Te] [a ayy) [Ti Tie] | boy byl [Th Te] [an] it follows that the chain scattering matrix of the cascade connection can be written in terms of the individual chain scattering matrix as follows: eu) [Th Tal[TH Tal [es =| E 4.4.13) [*] [fm Ta\|Th Talla, : Hence, the overall T matrix is obtained by multiplying (7*] and [P)]. Equation (1.4.13) is useful in the analysis and design of microwave amplifiers using computer-aided design techniques. and 1.5 SHIFTING REFERENCE PLANES: In practice we often need to attach transmission lines to the two-port network. Since the S parameters are measured using traveling waves, we need to specify the positions where the measurements are made. The positions are called ref erence planes. For example, in Fig. 1.5.1 we can measure the S parameters at the reference planes located at port 1’ and port 2’ and relate them to the S parameters at port 1 and port 2 of the two-port network. At the reference planes at port 1 and port 2 in Fig. 1.5.1,we write the scattering matrix as md] _ [Su Sa] fay [23] [= el] _ and at port 1’ and port 2' as 0] _ [Sin Sis] fox [23] [s ses] C In (1.5.1) the parameters, Si, Sia,S21,and Sa denote the S parameters at the unprimed reference planes (i.e.,at x1 = i and x2 = /2); and in (1.5.2) the parame- Sec. 1.5 Shifting Reference Planes a ae Reference planes Bae a fo aa) \ i I ms a0) ate) Two-port TY atta) Yael) by) 2 be) ae Abate) bald) — ——— I 7 1 \ \ { ' Port Port Pont 2 Port 2 a= neh nee m=0 Figure 15.1 Model for shifting reference planes. ters $)1,5'2,Sysand Sip denote the $ parameters at the primed reference planes (ie. at = O and x2 = 0). ‘The transmission lines and the two-port network in Fig. 1.5.1 when viewed from port 1’ and port 2’ constitute a new two-port network whose S parameters are defined by (1.5.2). The angles 0; and 62 are the electrical lengths of the transmission lines between the primed and unprimed reference planes. From our knowledge of traveling waves on a lossless transmission line we can write bill) = bie ay(l) = ay(0)e-” byl) = bse’ and a,() = a,(0)e#* where the factor e*/ accounts for the phase difference of the waves at the different reference planes. Substituting the previous relations into (1.5.1) gives i405] _T SyeP®— SyeHO09) Fay eo) 7 [Sime a Jes) (1s) ——S Sir Sia] _[ SuePA Sere 5 [fs = ~[S%ern Sye7P% ] and 8, S a eee [x $2] = [Steen sae | (iss) Equations (1.5.4) and (1.5.5) provide the relationship between S parameters at ‘two sets of reference planes.28 Representations of Two-Port Networks Chap. 1 1.6 PROPERTIES OF SCATTERING PARAMETERS Consider the two-port network shown in Fig. 1.6.1 where the transmission lines are assumed to be lossless and the characteristic impedances are real. This is the typical situation at microwave frequencies where 50-Q transmission lines (ie., = Zea = 50.Q) and 50-Q terminations are commonly used. From (1.4.1) and (1.42) the voltages and currents along the transmission lines can be written as Vax) = Vi Qe) + Vi) (1.6.1) and 1a) = 1308) — 17 ey = ED - ED (162) where i = 1 or 2. ‘The phasor notation in (1.6.1) and (1.6.2) is used for peak values, and for root mean square (rms) scaled values the following rms phasor notation is used: Vies(®) = Virma(®) + Viral) and Togs) = Ube) — Firms) als the rms value of the phasor and their peak value where for sinusoidal are related by V2. That is, Ve) Vismal) = iar Zo Wx) reork Valea) Zee Port 1" Port 1 Port 2 m=0 mae maak Figure 16.1. Tworport network, Sec. 1.6 Properties of Scattering Parameters. 29 [Note: In the first edition of this book it was assumed that the voltages and currents in (1.6.1) and (1.6.2) were scaled to rms values.) Using the normalized notation (similar to the notation introduced in Section 1.4) Vix) Vx) = Vz. i) = VZ Ne) ofa) = “HEP = Vz 155) and VZu1; (x) bx) VZu wwe can express (1.6.1) and (1.6.2) in the form v(x) = a(x) + B(x) and ix) = ax) — B(x) Italso follows that the normalized incident and reflected voltages at the ith port in terms of V(x) and I(x) are a(x) = Vix) + ZoiA(x)) (1.6.3) ne and (a) = = [Vfx) - Zee llx) 1.6.4) we (x)] re Observe that in (1.6.3) and (1.6.4) the normalizing impedances Zo1 and Z.2 are the characteristic impedances of the transmission lines. ‘The average power associated with the incident wave on the primed ith port (ie.,at x1 = O and x2 = 0) can be expressed in several ways, namely PF (O) = Re[Vinm(0) Uama(O))"] = 3 RLV? (0) (17(0))"] =pe{vio Or @|-4 a mo m0)? (165) Similarly, the average reflected power is30 Representations of Two-Port Networks Chap. 1 P75) = Re[Vizms(0) (Cizms(0))*] = fret V7 U7 )"] ded iro G0] = 116 F POP = Pram(OF (166) Since the line is lossless [i.e., P} (0) = P} (J) and P7 (0) (i). (.6.5) and (1.6.6) show that the quantities }]a,(0)|? = 4]a,(x)|? and 4]6,(0)|? = 3[,(x)|? represent the power associated with the incident and reflected waves, respectively. Now consider the network in Fig. 1.6.2, in which port 1 is excited by the sinusoidal voltage source represented by the phasor E; with source impedance Z; = Zn, and port 2' is matched (ie., Z2 = Zo2). When the output port is matched we say that the network is terminated in its normalizing impedance. Again, we point out that in practice Zo and Z.2 are real (usually 50-Q transmission lines) and Z; and Z; are pure resistors (usually 50 Q). In Fig. 1.6.2 Zn is the input impedance of the two-port at x; = h. Since port 2 in Fig. 1.6.2 terminated in its normalizing impedance, at x2 = 0 we can write Va(0) = ~Za1,(0) (67) and from (1,6.3) it follows that 10) = 7p MO + Za WE Fah * Za 'fO] oO Hence, there is no reflected wave from the load Z2 = Z.2. Atx: = 0,we have Vi(0) = Ey ~ 2 (0) (1.6.8) an 4 fa t Z=Z0 wo) es) ke 10) a Towen | MAG) SO) aE = 0 C= Pot, Pon 2 Pon! per’ Figure 162 Two-port network excited by the sinusoidal voltage source represented by the phasor £; with source impedance Z; = Zy, and terminated in the impedance 23 = Zea Sec. 1.6 Properties of Scattering Parameters 31 Substituting (1.6.8) into (1.6.3) gives ~AL 1 = 7 or -/AE lao? = Hence, from (1.6.5) (1.69) Equation (1.6.9) shows that the power in the incident wave }|a,(0) |? represents the power available from the source E, with internal resistance Z; = Z.1, We call this quantity Pavs [ie.,Pays = P{ (0) when Z; = Z.1}-Since the line is los less [ie., }]a,(0)|? = $]a,(l,)|*] the quantity $]a,(/,)|? represents the power available from the source at port 1. The power available from the source is independent of the input impedance Zn of the two-port network. If Z; isnot equal to Z,1 then (1.6.8) must be modified accordingly, and the resulting incident power [ie., the value of }]a,(0)|] will not be equal to the power available from the source. ‘Substituting (1.6.8) into (1.6.9) gives (MO) + Zh OV} [M(O) + ZA (O)}* BZ, Eyer = P= ot = Bla? = 37 1 8Z,, UMP + ZhOVIO + Z.VOMTO) + Z|} (1.6.10) larly, from (1.6.4) we obtain 1 i 2 o 2 [o,(0)|? = az, HM? = Zh(OVFO) ~ ZAK (0)Y,(0) + Z|hOP I] (1.6.11) Subtracting (1.6.11) from (1.6.10) gives 1 i , Flas? ~ 5 [oun = FOTO) + MOYO) = PREMIO) which represents the power delivered to port 1’, or to port 1 since the line is lossless. We call this quantity P1(0)[i.e., P,(0) a,(0)|? — $]b,(0) 7]. There- fore, it follows that 132 Representations of Two-Port Networks Chap. 1 FOP = Pas ~ Pr) (16:12) ‘The quantity }|b,(0)|? represents the reflected power from port 1 or port 1’ Equation (1.6.12) can also be written as. E oyG)P = Pass — PW) Equation (1.6.12) shows that the generator sends the available power avs = 4|a,(0)|? toward the input port 1. This power is independent of the put impedance Zn. If the input impedance Zn is matched to the transmission line (i.e., if Zr = Zoi), then the reflected power is zero. However, if Zr: # Zor, part of the incident power }|a,(0)|* is reflected back to the generator. The reflected power is given by }{b,(0)|? and the net power delivered to port 1 is Py) = Pull) = Pass ~ 3 (OO? From (1.6.7) and (1.6.4) we obtain 8400) = 7 MO) ~ ZahlON = PEt Pals) ~ Za) = -VZah) Therefore, FoF = 5 OP Za represents the power delivered to the load Z; = Zc2. We will denote the power delivered to the load Zs by P2(0) [i.e., P2(0) = 3|62(0)|")- In order to calculate the S parameters of the two-port network in Fig. 1.6.2 at the unprimed reference planes (ie.,at ports 1 and 2) we first observe that Si: at x1 = his given by Sas 29 | =i) (1.6.13) (4) Jaageo Vi Ch) Isao which from (1.3.39) can be expressed as, 5, = 2a ~2a (1.6.14) "Ent Zn Equation (1.6.13) or (1.6.14) shows that Si; is the reflection coefficient of port 1 with port 2 terminated in its normalizing impedance Z2 (i.¢., a2 0). ‘The evaluation of Si: at x1 = 0 (denoted by Sj,) can be done using (1.5.4) e., Shy = Sue~P04); or alternately we can calculate the input impedance at x; = 0, and its associated reflection coefficient would be Si,. Sec. 1.6 Properties of Scattering Parameters 3 If we consider the quantity |S1,[? we find from (1.6.9) and (1.6.12) that |sylt = DE | Baus = Puy LaF foero Pas or Pill) = Py(O) = Pays(l ~ [Sul?) (1.6.15) ‘The previous relations show that |S),|? represents the ratio of the power reflected from port 1 to the power available at port 1. If |Si1| > 1, the power reflected is larger than the power available at port 1. Therefore, in this case port 1 acts as a source of power and oscillations can occur. ‘The evaluation of S:; at the unprimed reference plane is as follows: gs, = 2) ~VZa1h) 2 ai(h) fast-0 WZ FA) 2 LV Zab VZo1 F(A) ‘The last step in (1.6.16) follows because fa(h) = 13(h) 13h) = 0). Equation (1.6.16) can be manipulated further to make the evaluation of ‘Su simpler. To this end we replace the network in Fig. 1.6.2 by the equivalent network shown in Fig. 1.6.3. The equivalent network was obtained by finding the Thévenin’s equivalent at ports 1 and 2. At port | the Thévenin’s voltage is, called Ey,ru and the Thévenin’s resistance (obtained by setting E: = 0) is simply Zo. The Thévenin voltage (i.e. the open-circuit voltage at xi = /; in Fig. 1,62) can be shown to be Em = Eie#%, At the matched port 2, the Thévenin’s equiv- (1.6.16) Psay-0 ah) = 13 (h)[since alent resistance is Zoo. From (1.6.3), i(h) is given by +) =4@ 21 TO) = Ye az, lMalld + Zac (1617) Zn + 7 | at aie) | Two-pon — Jantcal=0, &m@ weed) nemo ey AOE Za Port Pont2 nels ele Figure 1.6.3 Two-port network with Thévenin’s equivalent at ports 1 and 2.4 Representations of Two-Port Networks Chap. 1 where from Fig. 1.6.3, Vi(h) is given by V4) = Ena ~ Zotlh) (1.6.18) Substituting (1.6.18) into (1.6.17) gives E Th) “a (1.6.19) Also, at port 2 of Fig. 1.6.3 we have Valb) Ih) = te (1.6.20) ‘Substituting (1.6.19) and (1.6.20) into (1.6.16) results in the following practical expression for Sz Sy = a Md (1.6.21) VZ,, Fim ot Equation (1.6.21) shows that Sai represents a forward voltage transmission coefficient from port 1 to port 2. To evaluate Saat the primed reference planes, one uses (1.5.4). From (1.6.21), |Szil? can be expressed as 1 2, 2 Iva] ben Tole, which shows that |Sos? represents the ratio of the power delivered to the load Zea (ie, Px) to the power available from the source E1,rx (i.e., Pavs). The ratio P./Pyys is known as the transducer power gain Gr. Hence, the transducer power gain is given by sal Gr= [Sal? (1.6.22) f2 = Zy it follows from (1.6.21) that Y tZi= and the transducer power gain is given by vy P a Gr= Sal? = E, | If we analyze the network shown in Fig. 1.6.4 in which the sinusoidal excitation represented by the phasor E> with source impedance Z2 = Zx2is placed Sec. 1.6 Properties of Scattering Parameters 38 zn Hy po aes bales) network bay 22 & Port 1’ Pon 1 Por 2 Port 2° x20 aa %=0 Figure 1.64 Two-port network excited by the voltage source F with internal im- Pedance Z: = Za and terminated inthe impedance Z, = at port 2' and port 1’ is matched (i.e., Z; = Zor) we find that at the unprimed reference planes b(t) Zn -Z, = Pa) = 2Zn=Z: : Sa al) Jauiyeo Zr + Z,y (1.6.23) and = bath) - als) |a¢,ye0 where Ea.ra is the Thévenin’s voltage at port 2. Equation (1.6.23) shows that S12 is the reflection coefficient of port 2 with port 1 terminated in its normalizing impedance Z; = Zor [ie., ai(h) = 0], and Siz represents a reverse voltage transmission coefficient from port 2 to port 1. Equation (1.5.4) can be used to evaluate Sip and Sip at the primed reference planes. ‘The quantity |Sz2|? represents the ratio of the power reflected from port 2 to the power available at port 2. If [S| > 1, the power reflected is larger than the power available at port 2 and oscillation can occur. The quantity |Si2? represents a reverse transducer power gain. In fact, 2 a 2) gary [Exral?h [82,5 [Swf = The S parameters of a transistor are commonly measured with Zoi = Zor =Zq and Z; = Zz = Zo in Fig. 1.6.2. These S parameters are said to be measured in a Z, system, If this transistor is then used in the circuit of Fig. 1.6.2 with arbitrary terminations Z; and Z2 the gain Gr is no longer given by (1.6.22). In Sections 2.6 [see (2.6.13)] we will show that for arbitrary values of Z: and Zz36 Representations of Two-Port Networks Chap. 1 the gain Gr can be expressed in terms of Zi, Za, and the S parameters of the transistor measured in a Z, system, Example 1.6.1 Evaluate the S parameters, in a Z, system, of (a) a series impedance Z and (b) a shunt admittance ¥. Solution. (a) The two-port network of a series impedance is shown in Fig.1.6.5a, and the network connected to transmission lines with characteristic impedances Zo, excited by a source represented by the phasor E; with source impedance Z; = Zo, and terminated in the normalizing impedance Z2 = Ze is shown in Fig. 1.6.5b (ie, in a Ze system). The Thévenin’s equivalent circuit is shown in Fig. 1.6.5. ‘The S parameters are evaluated at the reference planes denoted by port 1 and port 2, From (1.6.14) we find that s, : lcaeo where from Fig. 1.65¢ Zn = Z + Zs. Therefore, Z Za, (1.6.24) ‘Since V2(b) is given by (see Fig. 1.6.5¢) EZ. Aumze Os we find from (1.621) (with Zn = Zea = Zo) that Sy = 2. (1625) Bim | 2+ 2Z, From symmetry, we observe that Sua = Sis and Siz = Sz. If the two-port network consists of a series inductor with Z = 100 Q, then in a 50-2 system (1e., with Z, = 50 Q) it follows that the S parameters of the series inductor from (1.6.24) and (1.6.25) are tsy=[ 0778" o707|—asr o7o7|=45° 0.70745" (b) The two-port network of a shunt admittance is shown in Fg. 1.6.6a, and the terminated network is shown in Fig. 1.6.6b. The Thévenin’s equivalent circuit is shown in Fig, 1.6.6c. In this case and from (1.6.14) (1.626) Sec. 1.6 Properties of Scattering Parameters a z ° ° Port Port 2 @ ' ) cy Figure 165 (a) A two-port network consisting of a series impedance two-port network excited by a source and terminated inthe normalizing i Zz= Z« (€) a Taévenin’s equivalent circuit. Since from Fig. 1.6.6¢ ExmZn _E, a x ea areas, we obtain from (1.6.21) that 5, = 2b) a Eira (1.627) Again, from symmetry we observe that Se38 Representations of Two-Port Networks Chap. 1 | Pont Por 2 © Zn 2 + + ¥ Vales) Zo Em : Pott Pore © Figure 1.66 (a) A two-port network consisting of a shunt admittance ¥; (b) the two-port network excited by a source and terminated in the normalizing impedance 2 Zei(¢) A Toévenin’s equivalent circuit, If the two-port network consists of a shunt 100-@ resistor (ic, ¥ = 1/100 = 10 1m), then in a 50-@ system it follows that the S parameters of the shunt resistor from (0.6.26) and (1.627) are Example 1.6.2 Evaluate the S parameters of the short-circuited shunt stub shown in Fig. 1.6.a. Sec. 1.6 Properties of Scattering Parameters 39 '=-¥ocot st Pont Pon? ©) Figure 16.7. (a) A short-circuited shunt stub, (b) equivalent shunt admittance. Solution: The input impedance of a short-circuited shunt stub of length / was derived in Section 1.3, namely Zy (J) = jZ» tan 6l. Hence, its input admittance (denoted by Yis "yeh = -jr,cot Al where ¥, = Zo. ‘The short-circuited shunt stub can be replaced by its equivalent shunt admittance ¥ as shown in Fig. 1.6.7. Then, using the results of Example 1.6.1(b), the § parameters of the shunt admittance in Fig. 1.6.7b in a Z, system are readily found. That is, from (2.6.26) we obtain 2 eY. -1 274 Z,¥ 1+ ptanBl and from (1.627) —— jeot pi Example 163 In the two-port network shown in Fig. 1.6.8: (@) Find Zy(0). (b) Evaluate ai(0),bx(0),ai(4/8),by(48), and a3(0), (©) Evaluate Vi(0), (WB), (0), and (2/8). (4) Evaluate the average input power atx (¢) Evaluate Sy at x1 = Oand.x; = 2/8. (f) Evaluate the input VSWR and the output VSWR. (@) If the scattering parameters of the two-port network measured at the x1 xz = AB reference planes are Su. = 0.487 [63.4°, Siz = 0.01 [40°, Sx = 5 [135° and ‘Sz = 046 [40°,caleulate the power delivered to. the load Z Oand at x40 Representations of Two-Port Networks == Chap. 1 2x40) Zn =50+/509 son | [ a(x) aba) Two-pot Ze Za500 network = 500 soa babe) ale) x= a= 08 2208 w=0 Pont Port Pont2 Pon" Figure 1.68 Circuit diagram for Example 1.63. Solution: (a)Except for notational changes, the input impedance calculation is, the same at that in Example 1.3.1. That is, substituting Zn = 50 + j50 Q for Zin (1.3.40) the input impedance at x1 = 0 [which corresponds to d = 4/8 in (1.3.40)] is = 5p (504/80) + j80 tan 45° _ = 9050 + (50 + j50) tan 45° (b) Since at x1 = 0 the voltage Vi(0) is Vi(0) = 10 [0° ~ 50/(0) it follows from (4.63) that Zp) = Zin (Dla am 100 — 502 0) = sag 0) + ZA = Ja 0 ~ SOK) + SOKO] = Je ‘Then, with f) = A/8 it follows that 6, = fh, = 2/4 and (8) = 0,0." = J From (1.64), 1 —ino- 840) = Fy WAUO) ~ ZohLO = 7aplHO- ASHI] (1.628) “The current (0) given by E 10 ee eee : 10) = gah” we TOT Rs SUES A (1829) Substituting (1.6.29) into (1.6.28) gives 1 ae (0) = sbgg [to ~ 2602063184357 = 03162657 ‘The value of the reflected wave b(2/8) follows from by GiB) = by Oe = 0.316|=26:57°(1|45*) = 0316[18.43° Sec. 1.6 — Properties of Scattering Parameters a ‘At the output port the line is matched. Hence, ax(0) = 0. An alternative way of calculating the reflected wave b,(4/8) is to calculate Sy, at 21 = AB using (1.6.14), namely Zp = 50 _ (50 + j80) ~ 50 Sn FOSS Gora ry OTSA (1630) ‘Then (208) st oa (5-4) anna (©) V\(O) can be evaluated in several ways. We can use a voltage divider equation at 10 obtain _ EiZn(0)__ _10(100 = 50) = BZnO_ = 107|-8:3° MO = 5 Z(0) ~ 30+ C00 — j50) ~ 7071-8 V ‘The current 1y(0) was evaluated in (1.6.29). Hence, Vi(0) can also be evaluated using \(0) = 3(0)Zin(0). Alternatively, Vi(0) can be evaluated using the fact that = a,(0) + 6,0) V0) = VZ, [a,(0) + 6,(0)] + 0316 2627] =107| = 813° ‘The voltage Vs(2/8) i given by V(Al8) = VZ;[ay(W8) + by(2/8)] = V30 [xa [case + 0.316L18. «| = 632|-2657° V and (@ Tne input power at x1 = Ois (0) = ERe{V(O)1F(0)] = 3 Re(7.07|=-813°(0.063|—18435")] = 0.2 W 2 and the power at x1 = 18 is 1 P\QI8) = 5 rel yicsy/3/8)] = 5 Refo.32|=26.57"(0.089[71.57°)] = 02 W As expected, Pi(0) = Pi(2/8) since the line is lossless. ‘Another way of calculating Pi(0) and Py(W/8) is to use (1.6.15)-SinceChap. 1 1 \2 Pas = 3 lal it follows from (1.6.15) that P,(0) = P\(W8) = Pays(1 ~ [Su*) = 0.251 ~ (0.447) = 0.2 W (c) Sivat x; = 2/8 was evaluated in (1.6.30). Similarly, at the input ofthe line x1 = 0 (ic., at the primed plane), denoting the value as Sj), we obtain Zyy(0) = Zp _ (00 = j50) = $0 _ 9 449) 96 57° Za(0) + Z,~ (4O0= j50) +50 ~ 047-2657 Observe that Si: and Si, are related by [see (1.55)] PIS $e 0,447)63.43° (0 The output transmission line is matched (ie., Z2 = Z» = 50 @). Therefore, the load reflection coefficient associated with Z; = 50 2 is zero, and the output VSWR is unity ‘AL the input, the reflection coefficient associated with Zr is Su. Hence, using (1.3.44) with [fy replaced by [Si gives an input VSWR of (g) Since the output is matched (ie. Z2= Zor = 50 @) it follows that a2(0) = ‘a,(4/8) = 0. Hence, from (1.6.16) we have that b(AI8) = S028) or bls) = sus dg 3.54|90° Then (0) = by(A/8)e-™* = 445" ‘The power delivered to Z2is x0) = $ [os0))? = $eset = 627 Scattering Matrix of n-Port Networks ‘The extension of the previous formulation to an n-port network is simple. Consider the n-port network shown in Fig. 1.6.9. The transmission lines are assumed to be lossless with characteristic impedance Za: (i = 1 to n). Then it follows that we can write the scattering matrix of the 7 port, at the unprimed reference planes, in the form [5] = [5] fa) (1.631) ‘Sec. 1.6 Properties of Scattering Parameters 43 2n network Figure 169 Anv-portnetwork. Porta” Porn where [a] = bias "1 CV) + (Z,) UD (1.6.32) w= F2s"1qv) - dey (1633) Su Sz + Sw Sn Sn + + Sw and (257)‘The [a], [5], [V], and [7] are column matrices. That is, by Y q by Va 4 [b=]. M=|- Ca 4, by Vy, i ‘The S parameters of the n-port networks are easily measured. For example, from (1.6.31) Sis at x1 = h is given by 5, = td na Za ah) |e-o=2,3...9 Zn + Zor where Zn is the impedance seen at port 1 with the other ports matched. In other words, $1: can be measured by connecting a voltage source E; with source impedance Zo1 to port 1" and terminating all other ports in their normalizing impedances (.e., using matched terminations so that a, = O for j = 2,3,....m). ‘A transistor can be considered to be a three-port device, as shown in Fig. 1.6.10. In this case the scattering matrix, also called the indefinite scattering matrix, is by} [Su Sz Sis}}ar b}=|Sy Sa Sa ||ar (1.634) bs} [Su Sx Sas} Las, The name “indefinite scattering matrix” is used because no defi choice is made to ground a particular port. For example, the meaning of Si: in (1.6.34) is bb 1 leieo.esn0 That is, to measure Sy, reference resistances of 50 Q are used at ports 2 and 3. In a two-port common-emitter configuration, Si: is measured with the emitter grounded. Therefore, the value of Si: in (1.6.35) will be different from the value of Si in a two-port common-emitter configuration. Similarly, the parameters Siz, Sz1,and Sn in (1.6.34) will be different from the Siz,Sai,and Szz in ‘a two-port common-emitter configuration, Su (1.6.35) ~ b wee Rigure 16.10 Atransistorasa three-port 3 3 network 1.7 POWER WAVES AND GENERALIZED ‘SCATTERING PARAMETERS In the previous two sections we showed that in a transmission line it is quite natural to represent the voltage and currents in terms of traveling waves—that is, an incident wave a(x) and a reflected wave b(x). In this section we discuss the analysis of lumped circuits (from one-port to n-port lumped circuits) in terms of a new set of waves, called power waves. ‘We first consider the one-port circuit shown in Fig. 1.7.1, where both the source and load impedances are, in general, complex. The normalized incident and reflected traveling waves defined in (1.44) and (1.4.5) do not appear suit- able for the analysis of the circuit in Fig. 1.7.1a since there is no transmission line, and therefore the characteristic impedance is not defined. In addition, the concept of a load reflection coefficient according to (1.3.39) or (1.4.3) has no meaning. However, we can introduce a new set of waves, called the power waves. The power waves ap and bp are defined by a linear transformation anal- ogous to (1.4.7) and (1.48). That is, let 1 wet) 7.) and (1.7.2) 1 7 we 2 where R, = Re{Z,]. The definitions in (1.7.1) and (1.7.2) are such that the quantity }|a,|? is equal to the power available from the source, and the reflected power Wave b, is zero when the load impedance is conjugately matched to the source impedance (i.e., when Z, = Z3). In Fig. 1.7.1b a power-wave representation of the one-port network is shown. Let us look further into the significance of these power waves. From Fig. 1.7.1a the relation between Vand /is, V=E,-ZI (1.73) ! ct [ a v zu wey la on & ® Co) Figure 17.1. (a) A lumped one-port network, (b) power-waves representation46 Representations of Two-Port Networks Chap. 1 Substituting (1.7.3) into (1.7.1) gives L E, ~~ 21+29 =e 174 atk —_ Hence, or (1.75) which is recognized as the power available from the source (i.e., Pavs }1a,|°). In the circuit of Fig. 1.7.1a maximum power is delivered to the load when Z; = Z*.In fact, the reader should recall that the power delivered to the load (P:), where E, Z, + Z,| : 1 P= ZUPRelzI => RelZ] attains its maximum value when Z: = Z's, and Pismax (or Pavs) is given by Prine = Pavs = 5 which is identical to (1.7.5). Since }|a,|? represents the power available from the source, then 3|b,|* should represent the reflected power. In fact, from (1.7.1) and (1.7.2) we obtain 1 1 - te, V= 220 (V~ Z30* 1 aR Yt ZV + 2D" ~ aR 1 i zRev"n which is recognized as the power dissipated in Z, (ie., Pt). Therefore, (1.7.6) From (1.7.6), we can also write line lel? = Pavs ~ Pa, Hence, the reflected power }|b,|? is equal to the available power from the source minus the power dissipated in the load. ‘Under conjugate matched conditions the power available from the source is delivered to Z and the power reflected 3 |b, |" is zero (a very important prop- ‘Sec. 1.7 Power Waves and Generalized Scattering Parameters ar erty of the power waves). In fact, in the conjugate matched circuit shown in Fig. 1.7.2 we have vezi and from (1.7.2) the reflected power wave is 1 8, Zi1- ZN) =0 1.77 ao Rete D (1.7.7) as expected. ‘A power-wave reflection coefficient Ip (also called a generalized reflection coefficient) can be defined as the ratio of bp to ap. Hence, from (1.7.1) and (1.72) we obtain V-Z VeZd (1.7.8) Solving (1.7.1) and (1.7.2) for Vand J gives (Zila, + Z,b,) (1.7.9) and 1=@,-5,) 1.7.10) TE ( 1 a) a) |e . neo Y | [zee Figure 1.72 Conjugate matched network.43 Representations of Two-Port Networks Chap. 1 ‘One should observe that the addition and subtraction of the incident and reflected traveling waves in (1.4.4) and (1.4.5) is equal to the normalized voltage and current along the transmission line, while from (1.7.9) and (1.7.10) the voltage and currents are not given by a simple addition or subtraction of the power waves. From (1.7.9) and (1.7.10) we can also define incident and reflected voltages and currents (denoted by V;, Vj,/j,and J;) and relate them to the power waves. That is, let VeVEtV, (7.1) and (1.7.12) where (1.7.13) (1.7.14) (1.7.15) and Ve Z (1.7.16) From (1.7.13) and (1.7.14) a voltage reflection coefficient 'v can be defined as ae 2 Zp 2B V;. Zia, Zi? Zi Z+2, (17.17) v Furthermore, from (1.7.15) and (1.7.16) a current reflection coefficient Ty can be defined as which is identical to the power-wave reflection coefficient. ‘An alternative way of introducing the power waves would be to define incident and reflected voltages (Vj and V,) in Fig, 1.7.1a and then define the power waves ap and by in terms of V; and V;. We will not go through this development of the power waves since it would be somewhat repetitious. How- ever, a better understanding of the previous concepts is gained by calculating directly the incident and reflected voltages in Fig. 1.7-1a, and then verifying their relationship with the power waves [ic, (1.7.13) and (1.7.14)]. 7 Power Waves and Generalized Scattering Parameters 43 In Fig. 1.7.1a the voltage V;, which is defined as the voltage V when 21 = Zz, is given by (1.7.18) Using (1.7.4), (1.7.18) can be expressed in the form 2 Ee v; =e VR, % which is recognized as (1.7.13). In Fig. 1.7.1a, Vis given by v= oe Z.+Z, Since V= Vj + V> and Vj is given by (1.7.18), it follows that the reflected voltage is BoE pee Eee Earp iag PE +Z, 2, 2R, 2, +Z, (1.7.19) Equation (1.7.19) can be expressed in several ways, Using (1.7.17) and (1.7.18), (1.7.19) can be expressed in the form V; = VsTy Using (1.7.13) and (1.7.17), Vj can be written as 2g (Ze) . 4: VR, "Zi 4, which is recognized as (1.7.14). In terms of incident and reflected voltages, the analysis of the circuit in Fig. 1.7.2, where Z; = Zp, produces: V= V; and Vj; = 0 since there is no reflected wave. Itis simple to verify that in Fig. 1.7.1a the current /is E, "74, and the incident and reflected currents are al 2R, and nT, where [7 is the current reflection coefficient.50 Representations of Two-Port Networks Chap. 1 ‘When the normalizing impedance Z, is real and positive, the expressions for the power waves ap and b, are identical to the traveling waves a(x) and b(x). In fact, when Z, is equal to a real and positive normalizing impedance Z,, the relations derived for power waves are identical to those derived for traveling waves. Specifically, for Z; real and equal to Zz, (1.7.1) and (1.7.2) are identical to (1.4.7) and (1.4.8). Also, Vand Jin (1.7.11) and (1.7.12) are identical to V(x) and I(x) in (14.1) and (1.4.2). Finally, 'p, as well as P'yand I, are all equal to To. That is, for Z, = Z¥ = Zo, aryer,= 22% NY ae From a mathematical point of view the relations of a(x) and b(x) as functions of V(x) and /(x) are linear transformations such that V(x) and /(x) follow from the inverse transformation. Similarly, the relations for ap and bp as functions of Vand J fice., (1.7.1) and (1.7.2)] are different linear transformations, where their inverse transformations in (1.7.9) and (1.7.10) give Vand Jin aone- port network. In the case of a two-port network, as shown in Fig. 1.7.3, we can define generalized scattering parameters (S, parameters), denoted by Spit, Spr2, Spat, and S,22, in terms of power waves as follows: Tj Pps = Sputpr + Sprotpe (1.7.20) Ppa = Spardpr + Spr2Mpr (1.7.21) where : Opn = Re t Zh) (1.7.22) VR L ap = (Va + Zale) (1.7.23) n= VR, a 1 bn = RM Zh) (1.7.24) 1 4 cs om a ey ~~ Two-pot a a Ys network v» 4 c ~ (Spparameters) . bps ie & Figure 17.3 Two-port network representation in terms of generalized scatering parameters, ‘Sec. 1.7 Power Waves and Generalized Scattering Parameters 51 and -—L., bn 7m Me Zit) (1.7.25) The resistors Ry and Re are Ri = Re[Zi] and R; = Re{Z:].The values of the S, parameters depend on the terminal impedances Z; and Z2. These impedances are called the reference impedances. There is no way of directly measuring the generalized scattering parameters. For example, from (1.7.20) Hence, Su can be determined if apy and bp: can be measured with ay: = 0. In Fig. 1.7.4 the circuit in Fig. 1.7.3 is shown with £2 = 0, the output voltage is V2 = ~/22,,and from (1.7.23) it follows that aj2 = 0. Unfortunately, ap and bp: are not traveling waves and, therefore, cannot be separated and measured using a directional coupler. ‘The evaluation of Sp in Fig. 1.7.4 follows from (1.7.8); that is, - Zy- Zt Sm (1.726) ‘The input power to the two-port network can be expressed as 1 1 Pox = 3 lapil? — 5 [bal? = Pavs(l ~ |Spui!?) While the Sp parameters cannot be measured, the $ parameters of the two-port in Fig. 1.7.3 can be easily measured and the S, parameters can be calculated in terms of the S parameters. Recall that the S parameters in a Z, system are measured by inserting transmission lines with characteristic imped ances Z, at both ports in Fig. 1.7.3, changing Z; and Z2 to Z, (usually Z. = 50 ),and taking the appropriate measurement of the traveling waves ai(x),2(*), h i 4 [ + a oa Tworport “ ‘etwork a os (S,parameters) — a _ bn be Figure 1.74 Two-port network with F: = 02 Representations of Two-Port Networks Chap. 1 (x), and ba(x). The relations between the S, parameters and the S parameters are derived in Section 2.7. Power-gain expressions for the network in Fig. 1.7.4 are easily written in terms of S, parameters. For example, aya = 0 in Fig. 1.7.4, and the power delivered to the load Z2, denoted by P1, is Since from (1.7.21), when ajo = 0, bya = Saxap1 we can express P, in the form 1 PL = Elsptla, ‘The power available from the source is }|a,,|*; therefore the transducer power gain for the network in Fig. 1.7.4 is given by (1.7.27) Equation (1.7.27) gives the transducer power gain of a two-port network with arbitrary terminations Z, and Zz. If in Fig. 1.7.4 we let Zi = Z> = Zp then it also follows from (1.6.22) that Gr = |Szi[?, where So) is measured using the normalizing impedance Z,. In other words, for Z1 = Z2 = Z. it follows that Gr = |SyasP? = [Sil (1.7.28) Equation (1.7.28) expresses the fact that in a system where the S parameters are measured with Z; = Z; = Zo the transducer power gain is simply [S21 Since for Z; = Zz = Z, power waves and traveling waves give identical results, it follows that the transducer power gain is also given by |S,zi)2. For source and load impedances such that Z1 # Zo and Z; Zoit follows that|Spu|? + |Sei?. In Sections 2.6 and 2.7 we will learn that the gain Gr in Fig. 1.74 (as well as other power expressions) can be conveniently expressed in terms of Zi, Zz, and the S parameters of the transistor measured in a Z, system. Generalized Scattering Matrix for n-Port Networks In the n-port network shown in Fig. 1.7.5 the normalizing incident and reflected power waves are defined as Fe) av) + zu (1.7.29) and 1 zl (Ro?) (Vv) - [Z*] 1) (1.7.30)} Sec. 1.7 Power Waves and Gener mPont network Figure 17.5 Anmport network. where 9 by Ope bon (e,] Son bon q 0 [2 = 0 and 53Observe that in the case of a one-port network, (1.7.29) and (1.7.30) reduce to (1.7.1) and (1.7.2), and for a two-port network the equations reduce to (1.7.22) to (1.7.25). Solving (1.7.29) and (1.7.30) for [V] and [J] results in (V} = [vp] + [Yp1 (= U5]- U5] where. vg) = (2103) (Vs) = ZI] (a) = R15) and [e,] = [R'") 5] “The generalized scattering matrix of the n-port network is defined as, {6,] = [5,] [ap] (1.731) where Sou Spr Spin Sa if (S]= S, nt pn ‘The definition in (1.7.31) shows that different reference impedances Z; (i= 1 ton) produce different values of the generalized scattering parameters, ‘Therefore, the generalized scattering parameters are defined in terms of spe~ cific reference impedances. If the reference impedances are pure re (denoted by Z; = Za), the formulation in (1.7.29), (1.7.30), and (1.7.31) is i tical to (1.6.31) to (1.6.33) Sec. 1.7 Power Waves and Generalized Scattering Parameters: = ‘The S, parameters of the n-port network are defined by (1.7.31). For example, the parameter Spu is recognized as the input power reflection coefficient of port i, determined with all other ports terminated in their reference impedances (i.e., with Ex = 0 then Vk = ~Zalk and a= 0,k # i,k = 0,1,2,...41). Therefore, Pn Vi - Zi, Zn ~ ZF PH dy legnoakeik=Oten Mit 2d ZntZ The quantity |5,,,|? can be shown to be the transducer power gain from port ito port k. Example 1.7.1 Caiculate the power waves and the power delivered to the load in the circuit shown in Fig. 1.7.6. Solution: In Fig. 1.7.6, the voltage Vand the current J are given by EZ, , 10100 = j$0) = 59) 96.570 : Z,_+Z, 100 - j80 + 100 + j50 Soo and : ao 200 05.4, Z+z, From (1.7.1) [see also (1.7.4)] with R, = Re[Z,] = 100 @ and V= E,~ IZ, it follows thatthe power wave ais 1 E, 10. wazpe eee WR + 2D 7m VIO Since a conjugate match condition exists in Fig. 1.76, the reflected power wave by is zero [see (1.7.1): Therefore, the power delivered o the load is equal to Pavs and given by (1.7.5), namely 100 + j50.0 v Z,=100-j50.0 = 1010" Figure 1.7.6 Circuit for Example 17.1,56 Representations of Two-Port Networks Chap. 1 “The circuit in Fig, 1.7.6 is identical to the equivalent circuit in Example 1.3.1 (see Fig. 1.3.80). Traveling waves were used in the solution of Example 1.3. Example 1.7.2 Calculate the power delivered to the load in the circuit shown in Fig. 1.7.7a using (a) power waves and (b) traveling waves. Solution: The transmission line in Fig, 1.7.7a is matched since Z = Zo. There- fore,Ty = O and the input impedance at x = 0 is Zin(x = 0) = Zin(d = 4/4) = $0Q.The equivalent circuit at x = is shown in Fig. 1.7.7b. We can analyze the equivalent circuit atthe input of the transmission line, shown in Fig. 1.7.7b,either in terms of power waves or traveling waves. For power-waves analysis we use the notation in Fig. 1.7.7b, and for traveling-waves analysis we use the notation in Fig. 1.7-7c, (a) Power-Waves Analysis ‘The equivalent lumped one-port network in Fig. 1.7.6b can be analyzed in terms of power waves as follows. Since v= 2000 tov 3 100 and zeta i 10) Z= 1000 Z,= 1000 v Zn(0) = 50.0 + vo) = Zn(0) = 50.0 > 01 a (b) to) Figure 177-9) Teasmisionsine seit for Example 1726) the equal ci cee 2) edipaerneer cot) equlem cca == Oforuaehng waves analysis. ‘Sec. 1.7 Power Waves and Generalized Scattering Parameters 87 30. 50+ 100 from (1.7.1) and (1.7.2) with R, = Re{Z,] = 100 Q we obtain 2a =15 aig + 20 = sgn + 10002 and 1 1 WR 210 10 ~ ‘The power available from the source is 1 Pas Ze (ve zp 100(0.2)) = ~0.5 te 5 (15)? = 1125 w Observe that since Zix(0) = 50 & is not conjugate matched to the source Z,.there H(05)? = 0.125 Wand the isa reflected power wave by. The reflected power is $[b,|? power dissipated in Zin(0) is 1 2 Since the line is lossless, the power delivered to the load in Fig. 1.7.7a is 1 W. ‘The previous calculation for Pas can also be made by observing that the power reflection coefficient is Pe P— 5 Jo,[* = 1425 ~ 0.125 = 1 Therefore, : Pall ~ Ig) = 1321 = (8) <1 ‘Asa check to the calculations of ay and by we can use (1.79) and (1.7.10) to ver ily the values of Vand 1. That is, with Z, = Z* = 100 @, Vand / in texms of a, and dp are 1 V = Sipg 0011.5) + 100(-0.5)] = 10 and 02a ooo re T= Sigglts ~ 05) Which, of course are the correct values of Vand J (b) Traveling-Wave Analysis Referring to Fig. 1.7.7e, the voltage and currents at the input of the line are denoted by (0) and 1(0), respectively. Their values are (0) = 10 V and 1(0) = 02.58 Representations of Two-Port Networks Chay From (1.4.7) and (1.4.8) with Ze = 50 Q, it follows that (0) = WE +Z,10)) a + 50(0.2)] = v2 and 0) 1 1 . FV MO) ~ 2,10] = 57a [10 ~ 80002)] = As expected, there is no reflected wave [i.e., b(x) = 0] in a matched transmission line. Observe that {|a(0)]? = 1W is not equal to the power available from the source (Pavs = 1.125 W).In fact, only when Z, = Z, is }a(0)|? equal to Pays ‘The power dissipated by Zin(0) (and in the load since the line is lossless) is ns = 5 Lacon? ~ 5 [OCP = 3 ¢vay8 = 1 ‘The previous calculation for Pry can also be made by observing that Ty ‘Therefore, Wir = (0) ~ [Tol*) ‘The values of V(0) and 1(0) can be verified using (1.4.4) and (1.4.5), namely M0) = VZ,v,(0) = V50[a(0) + 6(0)] = V50[V2] = 10V and 110) = & We [a(0) - 6(0)) We [v3] =02A which, of course, are the correct values of the voltage and current: ‘The incident wave along the line is given by a(x) = a(Oen#* = V2e-iP and, since b(x) = 0, V(x) is given by Va) = VZ_ a(x) = V50 (V/2e"*) = 1027 With d = 2/4 — x we can write the voltage in terms of d as Vid) = 100-294 Example 173 Calculate the generalized parameters Sj and S,x1 at 1 GHz in the lossless, recip- rocal, two-port network shown in Fig. 1.7.8a. Then calculate Syz2 and Spiz using Fig. 1.786. Solution: At 1 GHz the impedance of the inductor is Z, = j10 ®, and it follows that Vi =0.167[0° Ei, 41 = 0.0118|=45° Ey, V2=0.118|=45° Ej, and I -0.0118|=45° E, From (1.7.22) to (1.725) we obtain: a = 0.071|0° Bi, by: = Sec. 1.7. Power Waves and Generalized Scattering Parameters 59 50+ /50.0 %=100 at L=1.59 0H = itt * eenom | % vy | + Z=100 = 80+/500 & ————) Two-por network © Figure 1.78 (a) A two-port network consisting of a series 139 nlf inductor with reference impedances Z; and Z: and with the voltage Eat the input; (b) same net- ‘work with the voltage E> atthe output 0.061 78.69" Es, aye = 0, and bp2 = 0.037|=45" £). Therefore. in Fig, 1.7.8a, Zn = 10+ {102 and from (1.7.26) 2, = Zt _ 10 +10 - (50) fel Zn = Zi _ 10 + f0 = (50 ~ j50) | 5 Gnlgeeo” Zn > Z, 104 f10 + (50+ j50) ~ OPE (1 > 32 and 2, 5. Sin ra cat 0.525|-45° “oon From Fig. 1.7.8b we have Zra = $0 + j60 and it follows that Vj = 0.833|0° f= ~0.0118|=45* En Vs = 0,925.19" Bs, and [2 = 0.0118|=45° E. From (1.7.22) to (1.7.28) we obtain: api = 0, bp: = 0.083[=45° Ez, apr = 0.158|0° Es, and bps = 0.134[11.32° E>. Therefore it follows thatRepresentations of Two-Port Networks Chap. 1 ° D Zp Zh, 50+ j60-10 . a = D401» ggstirs Sma] Zatz” 50+ 04107 LX 1733) Spleg-o Zr * 2a and S| _ 0.0831=45° _ 9 595) 45 Galeo” OR important commentis that most CAD programs will calculate the generalized Fig. 17.8a they will calculate the S, parameters If the $ parameters of the 1.59-nH inductor in Fig 1.7.8a are desired in a 50-Q system, one lets Z; = Zz = 50 Qin Fig. 1.7.88, ‘and the resulting Sp parameters calculated by the CAD program are the S parameters of the 1.59-nH inductor. 1.8 TWO-PORT NETWORK PARAMETERS CONVERSIONS ‘Ata given frequency a two-port network can be described in terms of several parameters. Therefore, it is desirable to have relations to convert from one set of parameters to another. For example, the z parameters of a two-port network are defined by V)= (0) (181) where v= («| aft m= [i] and a f= ie al In terms of incident and reflected waves, we obtain from (1.8.1) we} + v= bl @e}- i) or (el + ZU = el - DE where Z, is assumed to be real and [2] Sec.1.9 surement of Scattering Par (I= ty payT pT lt 2D el - zd 82) and solving for [z], we obtain (el = [2,1 (4) + ES) (f4) - [s)-* (1.8.3) where (1] is the unit diagonal matrix. Equations (1.8.2) and (1.8.3) give the conversion relations between the S and z parameters. These conversions, as well as. others among the z, y, A, ABCD, and S parameters, are tabulated in Fig. 1.8.1 1.9 MEASUREMENT OF SCATTERING PARAMETERS Measurement of the S parameters of a transistor requires reflection and transmission measurement of traveling waves at both ports. The transistors must be properly biased at the desired Q point, and small-signal conditions must be maintained through the measurements. Two measurement systems that are used to measure S parameters are discussed in this section. The first measurement system uses a vector voltmeter, and the second system uses a network analyzer. ‘The measurement system using a vector voltmeter is illustrated in Fig. 1,9.1a, The vector voltmeter used is the HP8508A, which operates over the frequency range extending from 0.1 MHz to 1 GHz. The probes A and B from the vector voltmeter are inserted into two HP11536A TEEs to measure the ii dent voltage at Ai and the reflected voltage at Bi. ‘The set-up in Fig. 1.9.1a is used to measure Su: and Sa: of a transistor in a '50-Q system. The main components are a signal generator to produce a sinusoidal signal at the frequency where the S parameters of the device are measured, a vector voltmeter, two dual directional couplers, two de bias circuits, a short-circuit calibration section, and a through calibration section. The vector voltmeter measures the magnitude of two voltages (in Volts rms) and the phase difference between them (in degrees). The dual directional couplers are used to separate and measure the incident and reflected voltages. The de bias circuits consist of a coupling capacitor and a large inductor. This large inductor is commonly known as a radio frequency coil (RFC). The RFC presents a very high impedance to the ac signal and zero impedance to the de signal. The transistor gets its base to emitter voltage through the dc bias circuit #1, and its collector to emitter voltage through the de bias circuit #2. A de bias circuit is shown in Fig. 1.9.1b, The voltage Vo and resistor Ra will set the Q-point values of fs and Ic, and Vcc sets the collector to emitter voltage. The voltage Ves can be varied to obtain exactly the desired Q-point value of Ic. A signal proportional to the incident voltage is read with probe A measuring the signal at position Ai, and a signal proportional to the reflected voltage is read with probe B measuring the signal at position B1. If the constant of,poo ow | fle de | fe oe fa 8 i aa ie | Fe aoe fle tle q fel eae | 2 oe |e Me as fs (Suit ga) = 01S 8 Saaasah4 () EE ES Conversion between Common-Base, Common-Emitter, and Common: Collector y Parameters June = Must Yay + Yan + Ym.8 = Sane ine = Oran + ¥22,0) = One t Yd Jaye = ~On.» + Yx,0) = ~ Ore + Yad) Jane = Vane tine + Inve + are Yaas 7 Vite + Ye + ane + Ya.e * Sane Mae = ~One Had = One tad Yap =~ Orne + Ye) = —Orae + Yn. Ye = Yine Yave Fane + are + Vane Yue = Mine Yay + Sao Yann + Yan Ye = One tVi2e) = Ons + Yarn) Yare = ~Oue + Ya.) = “Orne +a) Ye Yie * Vine * Yave + Vane = Sine ©) Figure 18.1 (a) Conversions among the zy. h, ABCD, and S parameters: (b)conversions between y parameters we Pon [ee Fe] Da tase ea et, . aL Figure 1.9.1. (a) Block diagram of an volumeter:(b) the de bias arrangement parameter measuring system using a vectorwo Representations of Two-Port Networks Chap. 1 the attenuation introduced by the directional coupler in the forward and reverse direction is the same), the vector voltmeter will measure the voltages at A: and By,and therefore the voltage ratio between the B and the A probes, which is exactly [S1|.To measure the phase of Su, the system must be calibrated so that the electrical length traveled by the incident voltage and the reflected voltage is equal. This ensures a zero degree phase shift between the incident and reflected voltages at A and B. The short-circuited section shown in Fig. 1.9.1a is used to calibrate the phase between probe A and probe B. With the short-circuited section connected at the position of the transistor, the line stretcher is adjusted until the phase difference in the vector voltmeter reads 180° (which is the phase shift produced by a short circuit). When the transistor jig is connected (with the input port of the transistor at the same position as the short-circuited section), the vector voltmeter will read the phase difference between B and A, which is the angle of Six. The measuring or reference plane is indicated in Fig.1.9.1a. ‘The parameter Sz: is also measured using the set-up in Fig. 1.9.1a. In this case the ratio used is that of the transmitted voltage measured at Bz to the incident voltage measured at Ai. The angle calibration for Sai is accomplished using the through calibration section shown in Fig. 1.9.1. With the through section connected, the line stretcher is adjusted until a zero degree phase shift is read in the vector voltmeter. A little thought shows that ifthe short-circuited calibration section is made exactly half the length of the through calibration section, once the line stretcher is adjusted with the short-circuit section, the vector voltmeter will automatically produce a zero degree phase shift with the through section. ‘The parameters Sz and Siz are measured by reversing the transistor jig and the associated de bias connections. This system provides a fairly inexpen- sive method for measuring S parameters. After the initial set-up of the system it is quite simple to measure the S parameters of a two-port network over a wide frequency range. ‘Ata given Q point and frequency, the S parameters of the transistor will be constant for certain amplitude values of the incident voltage Vj.As the amplitude of V+ is increased, the transistor will be driven into its nonlinear region and the S parameters will vary as a function of the input signal amplitude. “The second method for measuring S parameters uses a network analyzer. Network analyzers generate a calibrated RF signal and have three input measuring channels (see Fig. 1.9.2). These channels are commonly called the R, the ‘A, and the B channels. The R channel is used to measure the incident voltage, and the A and B channels measure reflected and transmitted voltages. The two ‘measuring channels A and B allow the measuring of any two parameters with a single measurement set-up. The RF signal generator of the network analyzer provides the sinusoidal test signal, which can be either a single frequency sinu- soid or a sweep-mode sinusoidal signal. The output of the RF signal generator is calibrated in dBm (using a 50-@ load). Network analyzers can display a variety of rectangular and polar plots Sec. 1.9 Measurement of Scattering Parameters. 65 ‘Smith chart ‘veray Network analyzer 8 6 Short-circuit ection HL Measuring J q Dual drectonal coupler be, Figure 1.9.2 One-port measurements using a network analyzer Jar measurements available in network analyzers are (a) power level at the R input (in dBm), (b) power level at the A input (in dBm), (c) power level at the Binput (in dBm), (d) power ratio between the A and R inputs (4/R), (e) power ratio between the B and R inputs (B/R), f) phase difference between the A and Rinputs (|A/R) and (g) phase difference between the B and R inputs |B/R. The network analyzer can also display in polar form the power ratio and phase of AUR and B/R. This XY-type of display provides magnitude and phase information. In fact, with proper calibration a Smith chart can be superimposed on the screen of the network analyzer and reflection coefficient data can be read directly from the screen. Figure 1.9.2 illustrates a measurement set-up that can be used for one- port measurements, such as the input impedance or reflection coefficient of a ‘one-port network. A short-circuited calibration section is used to calibrate the network analyzer display, Network analyzers contain internal circuitry for phase calibration; therefore, the phase difference between the channels can be calibrated by simply rotating a knob in the analyzer. For example, ifa polar display of A/R is selected, the short-circuited section can be used to set the dot on the screen to the edge of the Smith chart at 180°. In other words, the reflection coefficient of the short circuit at the measuring plane is set at 1[180°. Then, when the one-port is connected for measurements (at the indicated measuring plane), the dot will move to the proper value of reflection coefficient (or impedance) in the Smith chart. Figure 1.9.3 shows a measurement set-up for S parameters. This set-up is similar to that shown in Fig. 1.9.1a. After the proper phase calibration with the R,A, and B channels connected in the position shown in Fig. 1.9.3, selection of AV/Rin polar form will display Si, and B/R will display S2). With the transistor jig reversed, the parameters So2 and Siz are measured. Since one of the most important uses of the network analyzer is to mea- eres fouuency of the sionals anplicd to the RA. and B inputs. The rectangu- sure S$ parameters, an_ S-parameter test set is usually a part of the network6s Representations of Two-Port Networks Chap. 1 ‘Smith chart overiay \ Network analyzer vB up a a 9 868 as son | > 3 A J \ : {Hy — i r 50m ial drostona couple ‘coupler sn * won Trarisor fg with de bias circuits Figure 19.3 A measurement set-up for S parameters using a network analyzer, analyzer, or an external S-parameter test set is interfaced with the network analyzer. An HP8753C network analyzer with an HP85047 S-parameter test set shown in Fig. 1.9.4. Observe that the RF signal from the network analyzer is connected to the RF input of the S-parameter test set. The R, A and B channels of the network analyzer are connected to the corresponding R, A, and B channels of the S-parameter test set. The calibration of the measuring system in Fig. 1.94 is accomplished using a short-circuited section and a through section. ‘The S-parameter test set provides a convenient means for measuring S parameters, Basically, it replaces the two dual directional couplers shown in Fig. 1.933, provides two measuring ports (denoted by port 1 and port 2) to connect the device under test, and provides connections to apply the dc bias to ports 1 and 2. An electronic switch selects the $ parameters to be measured such that the S-parameter test set applies the RF signal to the appropriate port, applies the appropriate 50-Q termination to the other port, and outputs the appropriate A/R and B/R signals to the network analyzer. 1.10 SCATTERING PARAMETERS OF TRANSISTORS ‘The S parameters of microwave transistors are usually available for the transistor in chip and packaged form. Transistors in chip form are used when the ‘Sec. 1.10 Scattering Parameters of Transistors 6 HP 8753WITH HP 85047A Figure 19.4 The HPS7S3C network analyzer with an HP8SO47A S-parameter test set. (From HP Test and Measurements Catalog, 1994; courtesy of Hewlett Packard) are very popular because they come in sealed enclosures and are easy to work with. The parasitic elements introduced by the package produce a degradation in the transistor ac performance. Manufacturers usually measure and provide common-emitter oi ‘common-source $ parameters of transistors as a function of frequency at given de bias. Since the minimum noise figure, linear output power, and maxi mum gain require different de bias settings, the manufacturers usually provid two or three sets of S parameters. ‘Conversions among S parameters or to other parameters can be done us ing the conversion relations in Fig. 1.8.1. For example, to convert from com mon-emitter to common-base S parameters, we first convert the ‘common-emitter S parameters to common-emitter y parameters, then conver the common-emitter y parameters to common-base y parameters, and ther convert the common-base y parameters to common-base S parameters. ‘The common-emitter § parameters of a transistor are shown in Fig 1.10.1. This figure illustrates some of the information typically provided b3 manufacturers. The values of the S parameters are explicity given at variou: recommended Q points as a function of frequency. Figure 1.10.1 also shows best performance in gain, bandwidth, and noise is desired. Packaged transistors that the values of Si1 and Sz. as a function of frequency are conveniently showr\MRF962 COMMON-EMITTER S-PARAMETERS INPUT/OUTPUT REFLECTION FORWARO/REVERSE TRANSMISSION (COEFFICIENTS versus FREQUENCY COEFFICIENTS versus FREQUENCY (Woe = 10V, Ie = 50 mA) (Wee = 10V, le = 50 mA) +50, Vee] ic | ot Siz (voits| (may | (MHz) Zé_| Bal | zo 5 Too tas —[-oom | as °) 10 | S00 36 Fd 500 cy Ps 700 7 26 1000 a a 1800, 54 2. 700 Tie % 2 | 30 2 2 500 & 36 700 B ° 1000 rc 2 1500 5. a To Tis 7 50) 300 94 «0 300 a % 700 % 50 1000 a a 1500, e o 7 100 131 a7 Oo} 1) 300 100 & 500 a 2s 700 % 2 1000 % 50 1500, s x 100 12 co 25 | 500 7 ea 500 a6 33 700 % a 1000 n a 1500 ia 2 37 8 | 300 zg 500 3 700 3 i000 $ 1500 g Figure 1.10.1. S-parar Data Manual, second70 Representations of Two-Port Networks Chap. 1 ina Smith chart (the Smith chart is discussed in detail in Sections 2.2 and 2.3). For example, at Vee = 10 V, Ic = 50mA,and f = 1500 MHz, the value of Si: is listed as 0.79[165°. Using (1.6.14), the impedance associated with Su in a 50-Q system is Zn = ZL Su wm 591 +0 1-3,” 1- 079/165" From the Smith chart plot it is seen that the impedance associated with Si at {f= 1500 MHz is approximately the value of Zn calculated in the preceding equation. ‘The forward and reverse transmission coefficients Sa and Siz are usually given in a polar plot, as shown in Fig. 1.10.1. Observe that the amplitude scale for Sn is different from the amplitude scale for Siz. For example, at Vee = 10 50 mA, and f = 500 MHz we read from the plots: Sz; = 7[80° and 0.036|45°. These readings closely agree with the listed values. For many transistors, plots of Si and Sz2 are given for the transistor in chip and packaged forms. Typical plots of Sy: for a transistor in chip form and packaged form are shown in Fig. 1.10.2. We will learn in Chapter 2 that these plots can be used to develop equivalent circuits for the input and output impedances of the transistor. A typical Bode plot of the behavior of |Say), [Siz|, and |SziSia| is shown in Fig, 1.10.3. It is seen that the parameter |S is constant for frequencies below the beta cutoff frequency (i.e. fg) and then decays at 6 dB/octave. The transducer cutoff frequency (f,) is the frequency where |Sa1| is equal to 1 (0.dB). The parameter |Su increases at approximately 6 dB/octave, levels off around j,,and decays at higher frequencies. 5.97 + j65Q Smith chart Figure 1.102 Si, of a common-emitter transistor in chip and packaged form. Soc. 1.11 Characteristics of Microwave Transistors n Figure 1.10.3 Frequency behavior of [Sai [Sia and |Si2Sn]. (From Ref. [1.1]; ‘courtesy of Hewlett-Packard.) 1.11 CHARACTERISTICS OF MICROWAVE TRANSISTORS Bipolar Transistors Most microwave bipolar junction transistors (BJTs) are planar in form and made from silicon in the npn type. Below 4 GHz, silicon BJTs provide a reliable and low-cost solution to many electronic designs. The transistor dimen- sions are very small in order to permit operation at microwave frequencies. For example, typical emitter widths of 2 am are found in general purpose microwave transistors, and emitter widths from 0.5 zm to 1 jm in low-noise microwave transistors. Base thickness of tenth of a micron are typical. Four common applications where silicon BITS are used at microwave frequencies are small-signal amplifiers, linear power amplifiers, low-noise amplifiers, and oscillators. Manufacturers usually provide a selection of transistors in chip and packaged forms specially made to suit one of the previous applications. Figure 1.11.1 lists typical power gains, output power, and noise figure performances for silicon BJTs in packaged form. For specific and up-to-date information, one should acquire the microwave catalog from several manufacturers. Also, Microwave and RF magazine publishes on a regular basis comprehensive performance characteristics of microwave transistors. BJTs are manufactured using ion implantation and self-alignment techniques to obtain a multifinger emitter-base construction. A typical cross section of a silicon BIT is shown in Fig. 1.11.2a, and the top view of the structure is, shown in Fig. 1.11.2b. This type of structure minimizes the electron transit time from the emitter to the collector while maintaining a large emitter area. An example of a packaged transistor is shown in Fig. 1.11.2c. Power BJTsare also available. The increased power required for these transistors is obtained using an interdigitated construction, as shown in Fig. 1.11.24. ‘The equivalent hybrid-7 model of the intrinsic BJT is shown in Fig. 1.11.3. In the microwave range, the reactance of Cyc is very small in comparison to the resistance of ry and the resistor re is very large. Also, the reactance of Cy is usually smaller than the resistance rv. Therefore, the simplified model shown in Fig. 1.11.4 followsR Representations of Two-Port Networks Chap. 1 Power Output ‘Minimum Frequency Gain Power __Noise Figure Smallsignal 1GHz 2048 20dBm_— 18GB transistors 2GHz © 184B_«20dBm_=—244B (er general 40H: 1aB_I8dBm_=—3.54B purpose transistors) Linear power 1GHz —15dB294Bm = transistors 2GH2 = 12dB_ 243m = 4GHz 8d 27¢Bm = Low-noise 1GHz —18dB18dBm_— 1.268 twansstors 2GHz © 13€B_ARdBm = L748 4GH2 = 94B_«s18dBm=— 284 Osclators 1GHz = — 068m - 2H = 30dBm = 40H — 7am = Figure L111 Typical performances for silicon BJT in packaged form. A more advanced hybrid-2 model will include the distributed nature of rr and Cye, 8 shown in Fig. 1.11.5. ‘The following example illustrates in a simple manner how the S parameters depend on the model used for the transistor. Example L111 (a) A transistor with B = 100i biased at Vee = 10V and Ic = 10 mA, as shown in Fig. 1.11.6a, The small-signal behavior of the transistor at frequencies below 100 kHz can be represented by the hybrid model shown in Fig, 1.11.6b. Calculate, using Fig. 1.11.6b, the $ parameters of the transistor at frequencies below 100 kHz. (b) At 1 MHiz the small-signal behavior of the transistor can be modeled using the hybrid-z model shown in Fig, 1.11.7a, Determine the Si: parameter ofthe transistor at | MHz, Solution, (a) The capacitors C; and Cz are coupling capacitors. Their values are large in order to behave like short circuits to the ac signal. The two radio frequency coils RFCs are large inductors. Their purpose is to behave like an open circuit to the ac sig~ nals and to present zero impedance to the de signal. Vce is equal to Vec since the RFC at the collector is a short circuit at de. Since B = 100, the base current is Je Wxi0? oa ly Foo 7 100 HA ‘Then the value of Reis py = n= 92. 5-97 | ag Ip 100 x 10" ‘To set the Q point accurately, the value of Vea is varied, around its nominal value of 5 V, until a current of 10 mA is read with an ammeter connected in series with the Vee supply. Sec. 1.11 Cheracteristics of Microwave Transistors 3 Metalization g e B E 8 $02 ee Teg sre fee [ere gees) ae to pase 04 um ‘Collector 15 um 1 Collector 5 substrate de 25 um © Emiter rj Enter Base [1 catiector T 3 4 Base — Emitter © @ Figure 1.11.2 (a) Typical silicon BJT eross section; (b) top view: (c) packaged transistor;(d) top view of an interdigitated BIT. ‘The ac model shown in Fig. 1.11.66 is obtained by replacing C; and Cz with short circuits, the RFCs by open circuits, and the transistor by its hybrid model for frequencies below 100 kHz. The value of rin Fig, 1.11.66 is™ Representations of Two-Port Networks Chap. 1 ve E Figure 1.113 Hybrid model of a common-emitter BIT. 26 10"9 _ 4992610"? _ Foe BT = 100 TTS pss = 2008 and the transconductance is 3 -£.—k_, 0x, Bn 7” BERIT ~ 26x10 ~ BMS Figure 1.11.6¢ shows an equivalent model obtained by finding the Thévenin’s equivalent circuits at ports 1 and 2. From (1.6.14), with Zn = re the value of Si in a 50-2 system is fy 50 _ 260 - 50 Tye ¥ 50 260+ 50 = 0677 From Fig. 1.11.60 Exnttve _ Exru(260) = Antes = Hm) _ Woe yet Ze 260+ 50 Ea and ~BaVore(Fee {| 50) ~ —0.385¥y,(50) = -19.25v, ‘Therefore, Sec. 1.11 Character of Microwave Transistors 5 Figure L115 A hybrid model that includes the distributed nature of ry: and Ce ¥ = -1925(0.839)E, ry = ‘Then, using (1.6.21), we obtain 16ASEy x1 Sy = 2M = 2¢-1615) = -323 or 323180" 8 Em > From (1.6.23), with Zr2 = rar = 100 kQ, the value of Sax is a= 50 Sanaa! which shows that r behaves like an open circuit ina 50-0 system. Since there is no transmission from the output tothe input in the model in Fig 1.11.66, we obtain Su = 0. (b) At 1 Miiz the high-requency capacitances affect the frequency response of the amplifier and, therefore the values ofthe §parameters.The use of Miller's theorem allows us to draw the equivalent input circuit show in Fig 111.7 where Cw is the input Miller's capacitance, namely Cu = Cyl = A.) ™ Cyl + gg(50)) = 1X 107R(L + 19.25) = 2025 pF Hence, the input capacitance is Ci = Core + Cy = 10 X 107 + 20.25 x 107! = 30.25 pF ‘The input impedance in Fig. 1.11.7b is 1 1 Teel oC, = 76 ll aero 025 x10 259.4 ~ 128 ‘Therefore, Zyy — 50 _ (2594 = j12.8) ~ 50 = Zn 0 = 0677|-1.13° Ze #30 ~ 2594 — 128) + 50 ~ O7CES ‘The other $ parameters can be evaluated analytically, or using a CAD program. For the circuit in Fig. 1.11.7a, the other S parameters are Siz =0, Sai = 32.2 [179° and Sex = 0.999 |=0.62°,76 Representations of Two-Port Networks Chap. 1 RFC Je= 10 mA, a 500 + wo Eye | Goon em 7 (260 )' a . io { port ag pone Figure 1.11.6 (a) Transistor circuit for Example 1.11 {quencies below 100 kHz; (c) equivalent ac model. ) the ac model for fre- ‘The equivalent circuit for the transistor in packaged form must include additional parasitic elements introduced by the package. One such model is shown in Fig. 1.11.8. The meaning of the extrinsic parasitic elements Ls, Le, Le, Coe, Che, and Cee is self-explanatory. Typical values of the parasitic inductances n” © aoxifz 2-000 Zeon @ Pon2 Figure 1.11.7. (a) The ac model at 1 MHz; (b) an equivalent input citcuit for the model in (8). Cre by Intrinsic be 8 G 8 aut e model Figure 1.118 Modeling the parasitic elements introduced by the package.8 Representations of Two-Port Networks Chap. 1 ‘are 0.2 nH to 1 nH, and of the parasitic capacitances they are 0.01 pF to 0.05 pF. More advanced packaged models include the addition of small resistances in series with the inductors Ls, Le, and L,, and the use of transmission lines to ‘model the distributed nature of the parasitics associated with the package. ‘Small-signal models can also be obtained from the cross-section geometry of the transistor chip and a knowledge of its operation. The reader is referred to the article by Hsu and Snapp [1.2] for the equivalent circuit of the BJT chip shown in Fig, 1.11.9. Typical values for the HP HXTR-2001 transistor chip are given at 15 V and 15 mA. The equivalent circuit for the transistor in packaged form must include additional extrinsic parasitic elements, as illustrated in Fig. L118. ‘Some advanced CAD programs (such as the HP85150B Microwave De- sign System) allow designers to compare the measured S parameters of a transistor with those obtained from a given model and to optimize the model components in order to match the measured S parameters. “There are two figures of merit that are commonly used by manufacturers of microwave BJTs to describe the transistor performance: 1. fr: the gain-bandwidth frequency. It is the frequency where the short- circuit gain |1p(w)| approximates unity. 2. fax: the maximum frequency of oscillation. Its the frequency where the ‘maximum available power gain of the transistor (called Gams) is equal tol. Garmax and fax can be measured by conjugately matching the source impedance to the transistor input impedance, and the load to the transistor output impedance, Of course, the transistor must be unconditionally stable (ie.,n0 oscillations). Gamax is higher than the transducer gain |SuiP because of the matching conditions. These concepts are discussed in detail in Chapter 3. ‘The frequency dependence of hy(w) is given by Ie WhO" TS ithe where hyeis the low-frequency short-circuit current gain and fy is the beta cutoff frequency, namely 1 i I Dare dGue* Crd) 2 -Cve “The frequencies fr and fous for the intrinsic BJT model, shown in Fig. 1.11.4, are given by ~ Sn fm (uit) Sec. 1.11 Characteristics of Microwave Transistors 79 Froax = (1.2) Also, fy and fr are related by qa) ‘The frequency fr can also be expressed in terms of the total signal time delay from emitter to collector as [1.3] i fr © Uitte where ta: i the emitter-to-collector time delay, namely Tete th “The parameter zs represents the base delay time and r; the base-to-collector depletion-layer delay time. Figure 1.11.10 illustrates the meaning of fr, fi, and fox. Observe the gain rolloff at the rate of 6 dB/octave. ‘Two sources of noise in a microwave BJT are thermal noise and shot noise. Thermal noise is caused by the thermal agitation of the carriers in the ohmic resistance of the emitter, base, and collector. Shot noise is a current- dependent effect caused by fluctuations in the electron and hole currents due to bias conditions. Flicker noise (or 1/f noise) is not a problem in silicon BJTs at microwave frequencies since this noise is only significant at low frequencies (below 10 kHz 10 50 kHz). ‘Above 4 GHz the performance of silicon BJTs diminishes significantly. For example, at 8 GHz a typical power gain is low (around 7 dB) with a high noise figure (greater than 5 dB). BITs using gallium arsenide (GaAs) semiconductors have been developed. These devices are called heterojunction bipolar transistors (HBJTs). The increased velocity of the electrons in the n-doped GaAs results in higher gains and better microwave performance. With HBJT, power gains of 20 dB with ‘minimum noise figure of 4 dB can be obtained at 10 GHz. Field Effect Transistors ‘The fabrication of gallium arsenide field-effect transistors (GaAs FETs) is made in the metal semiconductor field effect transistor structure (MESFET). ‘That is, the gate terminal is constructed using a Schottky barrier gate. The microwave FETs are made with GaAs because the electron mobility is greater than that of silicon. The high electron mobility results in excellent frequency re- ce ang noice performance, especially above4 GHz. e, especially above 4 GHz80 Representations of Two-Port Networks Chap. 1 Fo _ Foe lee n-Collector Re st Collector ‘substrate @ E © Figure 1119 (a) The transistor cross-section geometry; (b) small-signal model de- ‘eloped from the ross-section geometry of a BIT. In fact, GaAs MESFETs have become the transistor of choice above 4 GHz, and for a variety of very low noise applications below 4 GHz. Figure 1.11.11 shows typical power gains, output power, and noise figure performances for GaAs MESFETs in packaged form. Certainly, at frequencies above 4 GHz, Sec. 1.11 Characteristics of Microwave Transistors a Values for the HXTR2001 at Vee = 15 V and le Csr = 0.066 pF—Base bond pad capacitance Cop = 0.06 pF—Emitter bond pad capacitance Ca: = 0.07 pF—Collector to base distributed capacitor Cor = 0.056 pF Coa = 0.032 pF Cre = 48 pF—Base to emitter junction capacitance Rac = 0.2 2—Base contact resistance Rai = 0.2 Q—Base distributed resistance Rec= 022 Ror =352 Ru = 44a Re = 5 Q—Collector resistance Re = 1.7 Q—Emitter-base diode resistance 99—Low-frequency common-base current gain gett eh f= 22.7 GHx—Base cutoff frequency f 10.8 ps—Collector delay time depletion region itp Figure L119 Continued GaAs MESFETs have higher power gains, lower noise figures, and higher output power capabilities than silicon BITS. Molecular-beam epitaxy is used in the construction of GaAs MESFETs. Figure 1.11.12a shows a typical crosssection of a GaAs MESFET chip. Power GaAs MESFETs are also available. Basically, these devices are constructed by connecting several low-power transistors in paralle!—that is, using an interdigitated construction with a large gate width (to increase gq) and with 8 to 12 source, gate, and drain fingers to obtain a large output power. The top view of an interdigitated chip is shown in Fig, 1.11.12 ‘The high-frequency model of the intrinsic GaAs FET in a common- source configuration is shown in Fig. 1.11.13. The capacitor C; represents the % fica oo Figure 111.10 Frequency characteristics of Gamuu|SuP and Ji82 Ropresentations of Two-Port Networks Chap. 1 Power Output Minimum Frequency Gain‘ Power__—_Noise Figure ‘Smallsignal 4GHz —144B_—.20dBm 0848 transistors 6GHz dB 20dBm 1oaB, (or general 12GHz —9dB_—18dBm 1848 purpose transistors) Linear power 4GHz = 84B_—-30dBm = transistors 6GHz = 45dB -30dBm - 12GHz = 454B 25 dBm = Low-noise 4GHz —15dB—18dBm 05 aB transistors 6GHe = 3uB kum 065 uB GHz -9SdB_ 16S dBm 11dB. Oscillators 4GHe — 043m - 6GHz = 104Bm - 12GHz = 068m = Figure 1.11.11 Typical performances for GaAs MESFETs in packaged form. ‘Schottky s 8 barr =D a mm Depletion at GaAs. 27 region, 0.3 wm GaAs layer GaAs substrate 100 yum @) D = drain P =] G=gate ~ S = source Cy Figure 1.11.12 (a) A typical cross section of a GaAs MESFET chip; (b) top view. ‘Sec. 1.11 Characteristics of Microwave Transistors 83 Depletion s rogion G oD ‘a GaAs layer Cos GaAs substrate >> Cpe ) Figure 1.11.13. (a) Cross-sectional geometry of a GaAs FET; (b) GaAs FET high: frequency model for the common-source configuration, gate-to-source capacitance, and the resistor ris the small gate-to-source channel resistance (j.e., the charging resistance of Ci). Typical values for the parameters in Fig. 1.11.13 for a GaAs MESFET having a gate length of 1 um and a gate width of 250,am are C, = 0.3 pF, Cyu = 0.02 pF (gate to drain capacitance), Ca, = 0.05 pF (drain to source capacitance), ras = 600 2 (drain to source resistance, gm = 40 mS (transconductance), and r= 2.5 Q. ‘The model for the transistor in packaged form is shown in Fig. 1.11.14. ‘This model includes the extrinsic parasitic elements Ry, Ly, Ry La, Ry and Le ‘Typical values of parasitic inductances are 0.1 nH to 0.9 nH, and values of parasitic resistors are 0.2 Q to 0.19. When the GaAs FET feedback capacitance is very small and can be neglected, there is no reverse transmission from the output to the input port of, the transistor and the transistor becomes unilateral (i.., Si2 = 0). The simplified unilateral high-frequency model for the intrinsic GaAs FET is shown in Fig. LIL15.a4 Representations of Two-Port Networks Chap. 1 ly Ry yg Fy a s Figure 1.11.14 A microwave GaAs FET common-source model that includes par- atic elements. ‘The short length of the gate determines the frequency response of the GaAs FET. For the model shown in Fig. 1.11.15, the frequencies fr and finsx are given by fn 11.4) fr 2xC, am and fr fra 15 fous = 2 V7, (115) Since fris limited by the electron transit time (.) through the channel, fr can be expressed in the form (1.3) frase (1.116) Sec. 1.11 Charact of Microwave Transistors 85 where L Lat % Here L is the gate length and v, is the electron saturation drift velocity. Another expression for fas, found experimentally, is [1.3] 33 x 10° fee eae Equation (1.11.6) shows the importance of the short gate lengths required for operation at microwave frequencies. For example, with L = 0.5 um and v= 1X 10” emis, it follows that fr = 31.8 GHz. Further improvements in the performance of GaAs microwave transistors have been obtained using a heterojunction chip structure. A GaAs transistor fabricated using a heterojunction is known as a high-electron mobility transistor (HEMT) or as a modulation doped field effect transistor (MODFET).A typical heterojunction is formed at the interface of a layer of n- type GaAs and a layer of n-type AlGaAs. A typical cross section of a HEMT is shown in Fig. 1.11.16. The opergtion of HEMT devices is discussed in detail by Pengelly [1.4]. In the HEMT the electrons travel in the n* AlGaAs channel, resulting in a very high electron mobility. These devices operate at frequencies well above 30 GHz. The intrinsic noise sources in a GaAs FET are the thermal-gencrated channel noise and the induced noise at the gate. The induced noise at the gate is produced by the channel noise voltage. There is no shot noise in the GaAs MESFET. However, the flicker noise (1/f noise) is significant below 10 MHz to 50 MHz and therefore can affect the performance of some oscillators. The extrinsic noise sources (see Fig. 1.11.14) are associated with the resistances Ry and R, and the gate bonding pad resistance. Regardless of the transistor used, the S parameters of the device provide all the required information for design purposes. The small-signal model s D 6 [nt Gass | if GaAs AlGaAs GaAs layer GaAs substrate Figure L1LI6 A cross section ofa typical HEMT.26 Representations of Two-Port Networks = Chap. provides an inside to the operation of the transistor and is helpful for noise characterization of the device. PROBLEMS 1L1 Show that the maximum and minimum values of |V(d)| along a transmission line are given by (1.3.42) and (1.3.43), respectively. 12 InExample 1.3.1 write the expressionsfor vd.) andi(d,) alongthe transmission ine. 13 In the circuit shown in Fig. P1.3, (a) Calculate the load reflection coefficient, Zin(2/8), and the VSWR. (b) Evaluate V(U8), 1(U8), PUB), V0), 10), and PCO). Zn(a/8) Z=500 Z.=100+/1000 d= 48 ‘a=0 Figure P13 1.4 Show that for equal reference resistance (Zor = Zea) the scattering matrix in (1449) can be written in the form Vi=SuVi + SaVa Vr {+ SaVi and Ty = Sali + Silt J, = Sql + Salt LS Verify the S- and T-parameter conversions given in (1.4.11) and (1.4.12). 116 Show that the Thévenin voltage E;,ru in Fig. 1.6.3 is given by Erru = Eve" 1.7 (a) Find the ABCD matrix of the series impedance Z and shunt admittance Y of Example 1.6.1. (b) Use Fig. 18.1 to convert the ABCD parameters obtained in part (a) to S parameters. Compare the answers with the results in Example 1.6.1, 1L8 Find the scattering matrix and the chain scattering matrix of a transmission line of length ! and characteristic impedance Zo, 1.9 Find the scattering matrix and the chain scattering matrix of an open-circuited shunt stub of length / and characteristic impedance Zo, Problems: 87 1.10 Determine the S parameters of the two-port network shown in Fig. PI.10. 10092 i 509 son eo + = ef Port Por 2 Figure P1.10 1.11 Find the S parameters of the 1-to-n turns ratio transformer shown in Fig. PL.11 at ports 1-2 and 1'-2'. Figure PILI 112 Show that the overall Sx parameter of two cascade two-port networks with scattering matrices [54] and [Se], respectively, is given by 113 (a) Calculate the reflection coefficient from the 100-9 load and the VSWR in the ‘4/4 line in Fig. P1.13. (©) Calculate V(0), Pays, and the power delivered to the 100-9 load 1.14 (a) Find the value of the source impedance that results in maximum power delivered to the load in Fig. P1.14. Evaluate the maximum power delivered to the load. (b) Using the value of Z, from part (a), find the Thévenin’s equivalent circuit at the load end and evaluate the power delivered to the load 14S In Example 1.63,calculate V;(0) and /o(0) and evaluate the n ‘and evaluate the power delivered to the load using | /,(0)?(50) and $| V,(0)|?/50. 116 In the network shown in Fig. PL16,Zn = 150 + 1502, Z: heh=as,Representations of Two-Port Networks = Chap. 7 Figure PLAS aoe <0) e500 2750+ /500 2n(0) Zn 4 , 2 Two-port a a y= 1010" x20 neh wend Figure PLIG (a) Find Zin(0). {(b) Find as(0), (0), 01(4/8), bx(2/8), nd a3(0). (©) Evaluate Vi(0), Vi(/8), 14(0), and 12/8). (@) Evaluate the average input power at x: = n= 28, and the average input power at Problems 89 (g) Evaluate the input and output VSWR. (h) Find the electrical length and the length in centimeters of the 1/8 transmission line at f = 1 GHz, (D Ifthe scattering parameters of the two-port network at a; = hand x: = hare Sia = 0, Su = 3 [60°, and Sa = 0.7 (30°, calculate the power delivered to Zs. 117 In the network shown in Fig. PL.16, Zn = 150 + /150 QZ, = 75.9, 2: = 1002, Za= 75 Q,h = AB, and be = 4. (@) Find Zn(0). (®) Determine the VSWR in the 4/8 and 4/4 lines. (©) Evaluate ay(0) (0), ai(48), (4/8), and ax(0), (@) Evaluate Pi(0) and Pi(1/8). (e) Evaluate Pays and i |a,(0)[2-Ate they equal? Explain. LAB (a) Calculate dp bp Vj, V5 .J3,and J; for the circuit in Fig. P1.18. (b) Evaluate V and J using the values of ap and bp from pert (a) (©) Evaluate Vand Fusing the values of V;- and V; from part (a) 119 (a) Calculate the incident and reflected voltages V,; and Vz; in the circuit shown in Fig. 17.6. (©) Evaluate a, and 6, using the values of V; and V, {i.e,using (1.7.13) and(17.14)] 1.20 (a) Calculate the incident and reflected voltages V;? and V, in the circuit shown in Fig. 1.7.76. (b) Evaluate ay and by using the valuesof Vj and Vz [i.e using (1.7.13) and (1.7.14). 1.21 Consider the circuit shown in Fig. P1.21. son poo 4 Z aC ) & - Port 1 Port 2 - Figure PL2i90 Representations of Two-Port Networks Chap. 1 (8) Evaluate Sp, Sy, Sp1,and Spa with Z: = Za = 2002. (b) Evaluate Spi, Spa, Spas, and Spra with Zi = Zz = 50.Q. (©) Evaluate Si1,Sz, Sz, and Sizin a 50-Q system. 1122 For the two-port network in Fig. 1.7.8, (a) Evaluate the S, parameters of the 1.59-nHl inductor if Zi = Z2 = 50. (b) Evaluate the S parameters of the 1.59-nH inductor in a $0-Q system. 41.23 Show that in the indefinite scattering matrix given in (1.6.34) the sum of the coefficient of any row is equal to 1 and the sum of the coefficient of any column is equal to 1. Hint: Since (1.6.34) = as = O,as shown in Fig. P1.23a. Then bi = Si at Pwe can write valid for any values of a1, a, and as, consider the case where bz = Suai,and bs = Ssiaz,and Heh+ht+hy “Therefoe;it follows that Su + Su + Sy 71 ‘The circuit shown in Fig. PL23b can be used to show thatthe sum ofthe eoeff- sin any ow is equal to 1. Be BR ees » Figure P23 1.24 In Fig. 1.811, verify the conversions between (@) zand y parameters (b) z and ABCD parameters. References ct 1.25 (a) Show that [Ss] = ~(b] + D1 - () and 1 = (¥.] 4) - fs) (A) + Es)“ where (b) Verify the conversion between S and y parameters in Fig. 18.1. 1.26 In the network shown in Fig. P1.26, the S parameters of the BJT and the value of L are known, Explain how the overall S parameters ofthe two-port can be calculated. Figure P1.26 127 The common-emitter $ parameters of a GaAs FET at f= 10 GHz are Sy = 0.73[=128° 73[73° 04s|114° Say = 0.75[=52° Determine the common-base and common-collector S parameters. 128 (a) Derive the equations for fr and fy, for a BIT, given in (1.11.1) and (1.11.3), (b) Derive the equation for fr for a GaAs FET, given in (1.11.4). REFERENCES [1.1] “S Parameter Design,” Hewlett-Packard Application Note 154, April 1972. [1.2] T. H. Hsu and C. P. Snapp, “Low-Noise Microwave Bipolar Transistor with Sub- Half-Micrometer Emitter Width,” IEEE Transaction on Electron Devices, Vol. ED-25, June 1978, pp. 723-730, [1.3] D. V. Morgan and M. J. Howes, editors, Microwave Solid State Devices and Appli- cations, Peter Peregrinus Ltd., New York, 1980. [1.4] R. S.Pengelly, Microwave Field-E fect Transistors—Theory, Design, and Applica- tions, 2nd edition, Research Studies Press, Letchworth, England, 1986.

1 - Presentation of Two-Port Networks

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

1 - Presentation of Two-Port Networks

Hochgeladen von

Copyright:

Verfügbare Formate

Das könnte Ihnen auch gefallen